d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
dd7f80a359
freebsd-dev/sys/dev/fb/vesa.c

1965 lines
47 KiB
C
Raw Normal View History

Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/*-
* Copyright (c) 1998 Kazutaka YOKOTA and Michael Smith
Add my copyright here. It seems I have contributed enough code. :-)
2010-03-23 23:19:23 +00:00
* Copyright (c) 2009-2010 Jung-uk Kim <jkim@FreeBSD.org>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer as
* the first lines of this file unmodified.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHORS ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_vga.h"
#include "opt_vesa.h"
#ifndef VGA_NO_MODE_CHANGE
#include <sys/param.h>
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
#include <sys/bus.h>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#include <sys/systm.h>
#include <sys/kernel.h>
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
#include <sys/lock.h>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#include <sys/module.h>
#include <sys/malloc.h>
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
#include <sys/mutex.h>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#include <sys/fbio.h>
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
#include <sys/sysctl.h>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#include <vm/vm.h>
#include <vm/vm_extern.h>
#include <vm/vm_kern.h>
#include <vm/vm_param.h>
#include <vm/pmap.h>
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
#include <machine/pc/bios.h>
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
#include <dev/fb/vesa.h>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#include <dev/fb/fbreg.h>
#include <dev/fb/vgareg.h>
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
#include <dev/pci/pcivar.h>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#include <isa/isareg.h>
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Move sys/dev/x86bios to sys/compat/x86bios. It may not be optimal but it is clearly better than the old place. OK'ed by: delphij, paradox (ddkprog yahoo com)
2009-09-23 20:49:14 +00:00
#include <compat/x86bios/x86bios.h>
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
#define VESA_BIOS_OFFSET 0xc0000
#define VESA_PALETTE_SIZE (256 * 4)
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#define VESA_VIA_CLE266 "VIA CLE266\r\n"
#ifndef VESA_DEBUG
#define VESA_DEBUG 0
#endif
/* VESA video adapter state buffer stub */
struct adp_state {
int sig;
#define V_STATE_SIG 0x61736576
u_char regs[1];
};
typedef struct adp_state adp_state_t;
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
static struct mtx vesa_lock;
Probe supported states for save/restore function. Some VBE implementation refuses to save/restore states if an unsupported bit is set.
2012-02-23 19:05:57 +00:00
static int vesa_state;
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
static void *vesa_state_buf = NULL;
static uint32_t vesa_state_buf_offs = 0;
static ssize_t vesa_state_buf_size = 0;
static u_char *vesa_palette = NULL;
static uint32_t vesa_palette_offs = 0;
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
static void *vesa_bios = NULL;
static uint32_t vesa_bios_offs = VESA_BIOS_OFFSET;
static uint32_t vesa_bios_int10 = 0;
static size_t vesa_bios_size = 0;
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* VESA video adapter */
static video_adapter_t *vesa_adp = NULL;
Mark all SYSCTL_NODEs static that have no corresponding SYSCTL_DECLs. The SYSCTL_NODE macro defines a list that stores all child-elements of that node. If there's no SYSCTL_DECL macro anywhere else, there's no reason why it shouldn't be static.
2011-11-07 15:43:11 +00:00
static SYSCTL_NODE(_debug, OID_AUTO, vesa, CTLFLAG_RD, NULL, "VESA debugging");
Disable video ROM shadowing by default as I originally intended. I found a case where video ROM for an IGP is too tightly coupled with system BIOS to get relocated.
2010-09-02 17:22:12 +00:00
static int vesa_shadow_rom = 0;
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
TUNABLE_INT("debug.vesa.shadow_rom", &vesa_shadow_rom);
SYSCTL_INT(_debug_vesa, OID_AUTO, shadow_rom, CTLFLAG_RDTUN, &vesa_shadow_rom,
Disable video ROM shadowing by default as I originally intended. I found a case where video ROM for an IGP is too tightly coupled with system BIOS to get relocated.
2010-09-02 17:22:12 +00:00
0, "Enable video BIOS shadow");
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* VESA functions */
#if 0
static int vesa_nop(void);
#endif
static int vesa_error(void);
static vi_probe_t vesa_probe;
static vi_init_t vesa_init;
static vi_get_info_t vesa_get_info;
static vi_query_mode_t vesa_query_mode;
static vi_set_mode_t vesa_set_mode;
static vi_save_font_t vesa_save_font;
static vi_load_font_t vesa_load_font;
static vi_show_font_t vesa_show_font;
static vi_save_palette_t vesa_save_palette;
static vi_load_palette_t vesa_load_palette;
static vi_set_border_t vesa_set_border;
static vi_save_state_t vesa_save_state;
static vi_load_state_t vesa_load_state;
static vi_set_win_org_t vesa_set_origin;
static vi_read_hw_cursor_t vesa_read_hw_cursor;
static vi_set_hw_cursor_t vesa_set_hw_cursor;
static vi_set_hw_cursor_shape_t vesa_set_hw_cursor_shape;
static vi_blank_display_t vesa_blank_display;
static vi_mmap_t vesa_mmap;
static vi_ioctl_t vesa_ioctl;
static vi_clear_t vesa_clear;
static vi_fill_rect_t vesa_fill_rect;
static vi_bitblt_t vesa_bitblt;
static vi_diag_t vesa_diag;
static int vesa_bios_info(int level);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static video_switch_t vesavidsw = {
vesa_probe,
vesa_init,
vesa_get_info,
vesa_query_mode,
vesa_set_mode,
vesa_save_font,
vesa_load_font,
vesa_show_font,
vesa_save_palette,
vesa_load_palette,
vesa_set_border,
vesa_save_state,
vesa_load_state,
vesa_set_origin,
vesa_read_hw_cursor,
vesa_set_hw_cursor,
vesa_set_hw_cursor_shape,
vesa_blank_display,
vesa_mmap,
vesa_ioctl,
vesa_clear,
vesa_fill_rect,
vesa_bitblt,
vesa_error,
vesa_error,
vesa_diag,
};
static video_switch_t *prevvidsw;
/* VESA BIOS video modes */
#define VESA_MAXMODES 64
#define EOT (-1)
#define NA (-2)
#define MODE_TABLE_DELTA 8
static int vesa_vmode_max = 0;
static video_info_t vesa_vmode_empty = { EOT };
static video_info_t *vesa_vmode = &vesa_vmode_empty;
static int vesa_init_done = FALSE;
static int has_vesa_bios = FALSE;
static struct vesa_info *vesa_adp_info = NULL;
static u_int16_t *vesa_vmodetab = NULL;
static char *vesa_oemstr = NULL;
static char *vesa_venderstr = NULL;
static char *vesa_prodstr = NULL;
static char *vesa_revstr = NULL;
/* local macros and functions */
#define BIOS_SADDRTOLADDR(p) ((((p) & 0xffff0000) >> 12) + ((p) & 0x0000ffff))
static int int10_set_mode(int mode);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
static int vesa_bios_post(void);
Use M_WAITOK for VESA BIOS initialization consistently.
2010-06-23 23:34:56 +00:00
static int vesa_bios_get_mode(int mode, struct vesa_mode *vmode, int flags);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int vesa_bios_set_mode(int mode);
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
#if 0
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int vesa_bios_get_dac(void);
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
#endif
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int vesa_bios_set_dac(int bits);
static int vesa_bios_save_palette(int start, int colors, u_char *palette,
int bits);
static int vesa_bios_save_palette2(int start, int colors, u_char *r, u_char *g,
u_char *b, int bits);
static int vesa_bios_load_palette(int start, int colors, u_char *palette,
int bits);
static int vesa_bios_load_palette2(int start, int colors, u_char *r, u_char *g,
u_char *b, int bits);
#define STATE_SIZE 0
#define STATE_SAVE 1
#define STATE_LOAD 2
Probe supported states for save/restore function. Some VBE implementation refuses to save/restore states if an unsupported bit is set.
2012-02-23 19:05:57 +00:00
static ssize_t vesa_bios_state_buf_size(int);
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
static int vesa_bios_save_restore(int code, void *p);
Do not penalize correct or correctable VESA mode tables by calling another VBE function. Most problems should be corrected by the mode table sanity check and we only need the paranoid in extremely rare cases.
2010-03-25 17:51:05 +00:00
#ifdef MODE_TABLE_BROKEN
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int vesa_bios_get_line_length(void);
Do not penalize correct or correctable VESA mode tables by calling another VBE function. Most problems should be corrected by the mode table sanity check and we only need the paranoid in extremely rare cases.
2010-03-25 17:51:05 +00:00
#endif
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int vesa_bios_set_line_length(int pixel, int *bytes, int *lines);
#if 0
static int vesa_bios_get_start(int *x, int *y);
#endif
static int vesa_bios_set_start(int x, int y);
static int vesa_map_gen_mode_num(int type, int color, int mode);
static int vesa_translate_flags(u_int16_t vflags);
static int vesa_translate_mmodel(u_int8_t vmodel);
Be extremely careful when we determine bytes per scan line information. First, we compare mode table data against minimum value. If the mode table does not make sense, we set the minimum in the mode info. When we actually set the mode, we try VESA BIOS function and compare it against the previous value. If it makes more sense, update the information.
2010-03-23 23:10:17 +00:00
static int vesa_get_bpscanline(struct vesa_mode *vmode);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int vesa_bios_init(void);
static void vesa_clear_modes(video_info_t *info, int color);
#if 0
static int vesa_get_origin(video_adapter_t *adp, off_t *offset);
#endif
/* INT 10 BIOS calls */
static int
int10_set_mode(int mode)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AL = mode;
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
}
static int
vesa_bios_post(void)
{
x86regs_t regs;
devclass_t dc;
device_t *devs;
device_t dev;
int count, i, is_pci;
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
if (x86bios_get_orm(vesa_bios_offs) == NULL)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
dev = NULL;
is_pci = 0;
/* Find the matching PCI video controller. */
dc = devclass_find("vgapci");
if (dc != NULL && devclass_get_devices(dc, &devs, &count) == 0) {
Remove unnecessary pointer increment. A wrong pointer may be passed to free(9) and it can cause kernel panic when there are multiple graphics controllers in the system. Tested by: Brandon Gooch (jamesbrandongooch at gmail dot com) MFC after: 3 days
2010-05-18 18:28:17 +00:00
for (i = 0; i < count; i++)
if (device_get_flags(devs[i]) != 0 &&
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
x86bios_match_device(vesa_bios_offs, devs[i])) {
Remove unnecessary pointer increment. A wrong pointer may be passed to free(9) and it can cause kernel panic when there are multiple graphics controllers in the system. Tested by: Brandon Gooch (jamesbrandongooch at gmail dot com) MFC after: 3 days
2010-05-18 18:28:17 +00:00
dev = devs[i];
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
is_pci = 1;
break;
}
free(devs, M_TEMP);
}
/* Try VGA if a PCI device is not found. */
if (dev == NULL) {
dc = devclass_find(VGA_DRIVER_NAME);
if (dc != NULL)
dev = devclass_get_device(dc, 0);
}
if (bootverbose)
printf("%s: calling BIOS POST\n",
dev == NULL ? "VESA" : device_get_nameunit(dev));
x86bios_init_regs(&regs);
if (is_pci) {
regs.R_AH = pci_get_bus(dev);
regs.R_AL = (pci_get_slot(dev) << 3) |
(pci_get_function(dev) & 0x07);
}
regs.R_DL = 0x80;
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
x86bios_call(&regs, X86BIOS_PHYSTOSEG(vesa_bios_offs + 3),
X86BIOS_PHYSTOOFF(vesa_bios_offs + 3));
- When we restore VESA state, try BIOS POST earlier. VESA restore state function may not work properly if we don't. Turn off hardware cursor as vesa_set_mode() does. - Add VBE 3.0 specific fields in VESA mode structure and pack it. Note the padding is 190 bytes although VBE 3.0 says 189 bytes. It must be wrong because the size of structure becomes 255 bytes and the specification says it must be 256 bytes in total. In fact, an example code in the spec. does it right, though. While we are at it, fix some i386-isms. - Remove state buffer size limitation. It is no longer necessary since sys/compat/x86bios/x86bios.c r198251. - Move int 0x10 vector test into vesa_bios_post() as we always do it anyway.
2009-10-23 18:41:00 +00:00
if (x86bios_get_intr(0x10) == 0)
return (1);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
/* VESA BIOS calls */
static int
Use M_WAITOK for VESA BIOS initialization consistently.
2010-06-23 23:34:56 +00:00
vesa_bios_get_mode(int mode, struct vesa_mode *vmode, int flags)
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
uint32_t offs;
void *buf;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Use M_WAITOK for VESA BIOS initialization consistently.
2010-06-23 23:34:56 +00:00
buf = x86bios_alloc(&offs, sizeof(*vmode), flags);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (buf == NULL)
return (1);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f01;
regs.R_CX = mode;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
regs.R_ES = X86BIOS_PHYSTOSEG(offs);
regs.R_DI = X86BIOS_PHYSTOOFF(offs);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f) {
x86bios_free(buf, sizeof(*vmode));
return (1);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
}
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
bcopy(buf, vmode, sizeof(*vmode));
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_free(buf, sizeof(*vmode));
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int
vesa_bios_set_mode(int mode)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f02;
regs.R_BX = mode;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_AX != 0x004f);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
#if 0
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int
vesa_bios_get_dac(void)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f08;
regs.R_BL = 1;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (6);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_BH);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
#endif
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int
vesa_bios_set_dac(int bits)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f08;
/* regs.R_BL = 0; */
regs.R_BH = bits;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (6);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_BH);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_bios_save_palette(int start, int colors, u_char *palette, int bits)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
int i;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f09;
regs.R_BL = 1;
regs.R_CX = colors;
regs.R_DX = start;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
regs.R_ES = X86BIOS_PHYSTOSEG(vesa_palette_offs);
regs.R_DI = X86BIOS_PHYSTOOFF(vesa_palette_offs);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
bits = 8 - bits;
mtx_lock(&vesa_lock);
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
if (regs.R_AX != 0x004f) {
mtx_unlock(&vesa_lock);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
for (i = 0; i < colors; ++i) {
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
palette[i * 3] = vesa_palette[i * 4 + 2] << bits;
palette[i * 3 + 1] = vesa_palette[i * 4 + 1] << bits;
palette[i * 3 + 2] = vesa_palette[i * 4] << bits;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_unlock(&vesa_lock);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_bios_save_palette2(int start, int colors, u_char *r, u_char *g, u_char *b,
int bits)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
int i;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f09;
regs.R_BL = 1;
regs.R_CX = colors;
regs.R_DX = start;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
regs.R_ES = X86BIOS_PHYSTOSEG(vesa_palette_offs);
regs.R_DI = X86BIOS_PHYSTOOFF(vesa_palette_offs);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
bits = 8 - bits;
mtx_lock(&vesa_lock);
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
if (regs.R_AX != 0x004f) {
mtx_unlock(&vesa_lock);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
for (i = 0; i < colors; ++i) {
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
r[i] = vesa_palette[i * 4 + 2] << bits;
g[i] = vesa_palette[i * 4 + 1] << bits;
b[i] = vesa_palette[i * 4] << bits;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_unlock(&vesa_lock);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_bios_load_palette(int start, int colors, u_char *palette, int bits)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
int i;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f09;
/* regs.R_BL = 0; */
regs.R_CX = colors;
regs.R_DX = start;
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
regs.R_ES = X86BIOS_PHYSTOSEG(vesa_palette_offs);
regs.R_DI = X86BIOS_PHYSTOOFF(vesa_palette_offs);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
bits = 8 - bits;
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_lock(&vesa_lock);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
for (i = 0; i < colors; ++i) {
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
vesa_palette[i * 4] = palette[i * 3 + 2] >> bits;
vesa_palette[i * 4 + 1] = palette[i * 3 + 1] >> bits;
vesa_palette[i * 4 + 2] = palette[i * 3] >> bits;
vesa_palette[i * 4 + 3] = 0;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_unlock(&vesa_lock);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_AX != 0x004f);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_bios_load_palette2(int start, int colors, u_char *r, u_char *g, u_char *b,
int bits)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
int i;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f09;
/* regs.R_BL = 0; */
regs.R_CX = colors;
regs.R_DX = start;
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
regs.R_ES = X86BIOS_PHYSTOSEG(vesa_palette_offs);
regs.R_DI = X86BIOS_PHYSTOOFF(vesa_palette_offs);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
bits = 8 - bits;
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_lock(&vesa_lock);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
for (i = 0; i < colors; ++i) {
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
vesa_palette[i * 4] = b[i] >> bits;
vesa_palette[i * 4 + 1] = g[i] >> bits;
vesa_palette[i * 4 + 2] = r[i] >> bits;
vesa_palette[i * 4 + 3] = 0;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_unlock(&vesa_lock);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_AX != 0x004f);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
- When we restore VESA state, try BIOS POST earlier. VESA restore state function may not work properly if we don't. Turn off hardware cursor as vesa_set_mode() does. - Add VBE 3.0 specific fields in VESA mode structure and pack it. Note the padding is 190 bytes although VBE 3.0 says 189 bytes. It must be wrong because the size of structure becomes 255 bytes and the specification says it must be 256 bytes in total. In fact, an example code in the spec. does it right, though. While we are at it, fix some i386-isms. - Remove state buffer size limitation. It is no longer necessary since sys/compat/x86bios/x86bios.c r198251. - Move int 0x10 vector test into vesa_bios_post() as we always do it anyway.
2009-10-23 18:41:00 +00:00
static ssize_t
Probe supported states for save/restore function. Some VBE implementation refuses to save/restore states if an unsupported bit is set.
2012-02-23 19:05:57 +00:00
vesa_bios_state_buf_size(int state)
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f04;
/* regs.R_DL = STATE_SIZE; */
Probe supported states for save/restore function. Some VBE implementation refuses to save/restore states if an unsupported bit is set.
2012-02-23 19:05:57 +00:00
regs.R_CX = state;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (0);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_BX * 64);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
vesa_bios_save_restore(int code, void *p)
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (code != STATE_SAVE && code != STATE_LOAD)
return (1);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f04;
regs.R_DL = code;
Probe supported states for save/restore function. Some VBE implementation refuses to save/restore states if an unsupported bit is set.
2012-02-23 19:05:57 +00:00
regs.R_CX = vesa_state;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
regs.R_ES = X86BIOS_PHYSTOSEG(vesa_state_buf_offs);
regs.R_BX = X86BIOS_PHYSTOOFF(vesa_state_buf_offs);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_lock(&vesa_lock);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
switch (code) {
case STATE_SAVE:
x86bios_intr(&regs, 0x10);
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
bcopy(vesa_state_buf, p, vesa_state_buf_size);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
break;
case STATE_LOAD:
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
bcopy(p, vesa_state_buf, vesa_state_buf_size);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_intr(&regs, 0x10);
break;
}
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_unlock(&vesa_lock);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_AX != 0x004f);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Do not penalize correct or correctable VESA mode tables by calling another VBE function. Most problems should be corrected by the mode table sanity check and we only need the paranoid in extremely rare cases.
2010-03-25 17:51:05 +00:00
#ifdef MODE_TABLE_BROKEN
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int
vesa_bios_get_line_length(void)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f06;
regs.R_BL = 1;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (-1);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_BX);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Do not penalize correct or correctable VESA mode tables by calling another VBE function. Most problems should be corrected by the mode table sanity check and we only need the paranoid in extremely rare cases.
2010-03-25 17:51:05 +00:00
#endif
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int
vesa_bios_set_line_length(int pixel, int *bytes, int *lines)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f06;
/* regs.R_BL = 0; */
regs.R_CX = pixel;
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#if VESA_DEBUG > 1
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
printf("bx:%d, cx:%d, dx:%d\n", regs.R_BX, regs.R_CX, regs.R_DX);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#endif
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (-1);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (bytes != NULL)
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
*bytes = regs.R_BX;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (lines != NULL)
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
*lines = regs.R_DX;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
#if 0
static int
vesa_bios_get_start(int *x, int *y)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f07;
regs.R_BL = 1;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (-1);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
*x = regs.R_CX;
*y = regs.R_DX;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
#endif
static int
vesa_bios_set_start(int x, int y)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f07;
regs.R_BL = 0x80;
regs.R_CX = x;
regs.R_DX = y;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (regs.R_AX != 0x004f);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
/* map a generic video mode to a known mode */
static int
vesa_map_gen_mode_num(int type, int color, int mode)
{
static struct {
int from;
int to;
} mode_map[] = {
{ M_TEXT_132x25, M_VESA_C132x25 },
{ M_TEXT_132x43, M_VESA_C132x43 },
{ M_TEXT_132x50, M_VESA_C132x50 },
{ M_TEXT_132x60, M_VESA_C132x60 },
};
int i;
for (i = 0; i < sizeof(mode_map)/sizeof(mode_map[0]); ++i) {
if (mode_map[i].from == mode)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (mode_map[i].to);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (mode);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_translate_flags(u_int16_t vflags)
{
static struct {
u_int16_t mask;
int set;
int reset;
} ftable[] = {
{ V_MODECOLOR, V_INFO_COLOR, 0 },
{ V_MODEGRAPHICS, V_INFO_GRAPHICS, 0 },
{ V_MODELFB, V_INFO_LINEAR, 0 },
Fix VESA color palette corruption: - VBE 3.0 says palette format resets to 6-bit mode when video mode changes. We simply set 8-bit mode when we switch modes if the adapter supports it. - VBE 3.0 also says if the mode is not VGA compatible, we must use VBE function to save/restore palette. Otherwise, VGA function may be used. Thus, reinstate the save/load palette functions only for non-VGA compatible modes regardless of its palette format. - Let vesa(4) set VESA modes even if vga(4) claims to support it. - Reset default palette if VESA pixel mode is set initially. - Fix more style nits.
2009-11-03 20:22:09 +00:00
{ V_MODENONVGA, V_INFO_NONVGA, 0 },
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
};
int flags;
int i;
for (flags = 0, i = 0; i < sizeof(ftable)/sizeof(ftable[0]); ++i) {
flags |= (vflags & ftable[i].mask) ?
ftable[i].set : ftable[i].reset;
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (flags);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_translate_mmodel(u_int8_t vmodel)
{
static struct {
u_int8_t vmodel;
int mmodel;
} mtable[] = {
{ V_MMTEXT, V_INFO_MM_TEXT },
{ V_MMCGA, V_INFO_MM_CGA },
{ V_MMHGC, V_INFO_MM_HGC },
{ V_MMEGA, V_INFO_MM_PLANAR },
{ V_MMPACKED, V_INFO_MM_PACKED },
{ V_MMDIRCOLOR, V_INFO_MM_DIRECT },
};
int i;
for (i = 0; mtable[i].mmodel >= 0; ++i) {
if (mtable[i].vmodel == vmodel)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (mtable[i].mmodel);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (V_INFO_MM_OTHER);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Be extremely careful when we determine bytes per scan line information. First, we compare mode table data against minimum value. If the mode table does not make sense, we set the minimum in the mode info. When we actually set the mode, we try VESA BIOS function and compare it against the previous value. If it makes more sense, update the information.
2010-03-23 23:10:17 +00:00
static int
vesa_get_bpscanline(struct vesa_mode *vmode)
{
int bpsl;
if ((vmode->v_modeattr & V_MODEGRAPHICS) != 0) {
/* Find the minimum length. */
switch (vmode->v_bpp / vmode->v_planes) {
case 1:
bpsl = vmode->v_width / 8;
break;
case 2:
bpsl = vmode->v_width / 4;
break;
case 4:
bpsl = vmode->v_width / 2;
break;
default:
bpsl = vmode->v_width * ((vmode->v_bpp + 7) / 8);
bpsl /= vmode->v_planes;
break;
}
/* Use VBE 3.0 information if it looks sane. */
if ((vmode->v_modeattr & V_MODELFB) != 0 &&
vesa_adp_info->v_version >= 0x0300 &&
vmode->v_linbpscanline > bpsl)
return (vmode->v_linbpscanline);
/* Return the minimum if the mode table looks absurd. */
if (vmode->v_bpscanline < bpsl)
return (bpsl);
}
return (vmode->v_bpscanline);
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
#define VESA_MAXSTR 256
#define VESA_STRCPY(dst, src) do { \
char *str; \
int i; \
dst = malloc(VESA_MAXSTR, M_DEVBUF, M_WAITOK); \
str = x86bios_offset(BIOS_SADDRTOLADDR(src)); \
for (i = 0; i < VESA_MAXSTR - 1 && str[i] != '\0'; i++) \
dst[i] = str[i]; \
dst[i] = '\0'; \
} while (0)
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
static int
vesa_bios_init(void)
{
struct vesa_mode vmode;
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
struct vesa_info *buf;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
video_info_t *p;
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Reject some VESA graphics modes if the controller does not have enough memory to support them. Some adapters have expansible memory slots but video mode table is static. In this case, unusable modes may be reported. Submitted by: paradox (ddkprog yahoo com) (initial patch)
2009-09-25 18:08:38 +00:00
size_t bsize;
Map and report actual video memory we need.
2010-02-06 00:52:42 +00:00
size_t msize;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
void *vmbuf;
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
uint8_t *vbios;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
uint32_t offs;
uint16_t vers;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
int is_via_cle266;
int modes;
int i;
if (vesa_init_done)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
has_vesa_bios = FALSE;
vesa_adp_info = NULL;
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
vesa_bios_offs = VESA_BIOS_OFFSET;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vesa_vmode_max = 0;
vesa_vmode[0].vi_mode = EOT;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
/*
* If the VBE real mode interrupt vector is not found, try BIOS POST.
*/
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
vesa_bios_int10 = x86bios_get_intr(0x10);
if (vesa_bios_int10 == 0) {
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (vesa_bios_post() != 0)
return (1);
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
vesa_bios_int10 = x86bios_get_intr(0x10);
if (vesa_bios_int10 == 0)
return (1);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
/*
* Shadow video ROM.
*/
Make sure the interrupt entry point is within the video ROM range. We must not change interrupt vector if it is not pointing the ROM itself. Actually, we just fail shadowing altogether if that is the case because the shadowed copy will be useless for sure and POST may not be relocatable or useful. While I'm here, fix a debugging message under bootverbose, really. r211829 fixed one case but broke another. Mea Culpa.
2010-08-31 20:21:52 +00:00
offs = vesa_bios_int10;
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
if (vesa_shadow_rom) {
vbios = x86bios_get_orm(vesa_bios_offs);
if (vbios != NULL) {
vesa_bios_size = vbios[2] * 512;
Fix segment:offset calculation of interrupt vector for relocated video BIOS when the original offset is bigger than size of one page. X86BIOS macros cannot be used here because it is assumed address is only linear in a page. Tested by: netchild
2010-09-13 19:58:46 +00:00
if (((VESA_BIOS_OFFSET << 12) & 0xffff0000) ==
(vesa_bios_int10 & 0xffff0000) &&
vesa_bios_size > (vesa_bios_int10 & 0xffff)) {
Make sure the interrupt entry point is within the video ROM range. We must not change interrupt vector if it is not pointing the ROM itself. Actually, we just fail shadowing altogether if that is the case because the shadowed copy will be useless for sure and POST may not be relocatable or useful. While I'm here, fix a debugging message under bootverbose, really. r211829 fixed one case but broke another. Mea Culpa.
2010-08-31 20:21:52 +00:00
vesa_bios = x86bios_alloc(&vesa_bios_offs,
vesa_bios_size, M_WAITOK);
memcpy(vesa_bios, vbios, vesa_bios_size);
Fix segment:offset calculation of interrupt vector for relocated video BIOS when the original offset is bigger than size of one page. X86BIOS macros cannot be used here because it is assumed address is only linear in a page. Tested by: netchild
2010-09-13 19:58:46 +00:00
offs = ((vesa_bios_offs << 12) & 0xffff0000) +
(vesa_bios_int10 & 0xffff);
Make sure the interrupt entry point is within the video ROM range. We must not change interrupt vector if it is not pointing the ROM itself. Actually, we just fail shadowing altogether if that is the case because the shadowed copy will be useless for sure and POST may not be relocatable or useful. While I'm here, fix a debugging message under bootverbose, really. r211829 fixed one case but broke another. Mea Culpa.
2010-08-31 20:21:52 +00:00
x86bios_set_intr(0x10, offs);
Fix segment:offset calculation of interrupt vector for relocated video BIOS when the original offset is bigger than size of one page. X86BIOS macros cannot be used here because it is assumed address is only linear in a page. Tested by: netchild
2010-09-13 19:58:46 +00:00
}
Make sure the interrupt entry point is within the video ROM range. We must not change interrupt vector if it is not pointing the ROM itself. Actually, we just fail shadowing altogether if that is the case because the shadowed copy will be useless for sure and POST may not be relocatable or useful. While I'm here, fix a debugging message under bootverbose, really. r211829 fixed one case but broke another. Mea Culpa.
2010-08-31 20:21:52 +00:00
}
if (vesa_bios == NULL)
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
printf("VESA: failed to shadow video ROM\n");
}
if (bootverbose)
printf("VESA: INT 0x10 vector 0x%04x:0x%04x\n",
Fix a debugging message under bootverbose. This address is not linear.
2010-08-25 22:48:18 +00:00
(offs >> 16) & 0xffff, offs & 0xffff);
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f00;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Let x86bios_alloc() pass contigmalloc(9) flags. Use it to set M_WAITOK from VESA BIOS initialization. All other malloc(9) uses in the function is blocking any way.
2010-06-23 17:20:51 +00:00
vmbuf = x86bios_alloc(&offs, sizeof(*buf), M_WAITOK);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
regs.R_ES = X86BIOS_PHYSTOSEG(offs);
regs.R_DI = X86BIOS_PHYSTOOFF(offs);
bcopy("VBE2", vmbuf, 4); /* try for VBE2 data */
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f || bcmp("VESA", vmbuf, 4) != 0)
goto fail;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
vesa_adp_info = buf = malloc(sizeof(*buf), M_DEVBUF, M_WAITOK);
bcopy(vmbuf, buf, sizeof(*buf));
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (bootverbose) {
printf("VESA: information block\n");
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
hexdump(buf, sizeof(*buf), NULL, HD_OMIT_CHARS);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
vers = buf->v_version = le16toh(buf->v_version);
buf->v_oemstr = le32toh(buf->v_oemstr);
buf->v_flags = le32toh(buf->v_flags);
buf->v_modetable = le32toh(buf->v_modetable);
buf->v_memsize = le16toh(buf->v_memsize);
buf->v_revision = le16toh(buf->v_revision);
buf->v_venderstr = le32toh(buf->v_venderstr);
buf->v_prodstr = le32toh(buf->v_prodstr);
buf->v_revstr = le32toh(buf->v_revstr);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (vers < 0x0102) {
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
printf("VESA: VBE version %d.%d is not supported; "
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
"version 1.2 or later is required.\n",
((vers & 0xf000) >> 12) * 10 + ((vers & 0x0f00) >> 8),
((vers & 0x00f0) >> 4) * 10 + (vers & 0x000f));
Plug a possible memory leak. Submitted by: Yamagi Burmeister (lists at yamagi dot org)
2010-07-06 18:08:55 +00:00
goto fail;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
VESA_STRCPY(vesa_oemstr, buf->v_oemstr);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (vers >= 0x0200) {
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
VESA_STRCPY(vesa_venderstr, buf->v_venderstr);
VESA_STRCPY(vesa_prodstr, buf->v_prodstr);
VESA_STRCPY(vesa_revstr, buf->v_revstr);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
is_via_cle266 = strncmp(vesa_oemstr, VESA_VIA_CLE266,
sizeof(VESA_VIA_CLE266)) == 0;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
if (buf->v_modetable == 0)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
goto fail;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
msize = (size_t)buf->v_memsize * 64 * 1024;
Map and report actual video memory we need.
2010-02-06 00:52:42 +00:00
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
vesa_vmodetab = x86bios_offset(BIOS_SADDRTOLADDR(buf->v_modetable));
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
for (i = 0, modes = 0; (i < (M_VESA_MODE_MAX - M_VESA_BASE + 1)) &&
(vesa_vmodetab[i] != 0xffff); ++i) {
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
vesa_vmodetab[i] = le16toh(vesa_vmodetab[i]);
Use M_WAITOK for VESA BIOS initialization consistently.
2010-06-23 23:34:56 +00:00
if (vesa_bios_get_mode(vesa_vmodetab[i], &vmode, M_WAITOK))
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
continue;
- When we restore VESA state, try BIOS POST earlier. VESA restore state function may not work properly if we don't. Turn off hardware cursor as vesa_set_mode() does. - Add VBE 3.0 specific fields in VESA mode structure and pack it. Note the padding is 190 bytes although VBE 3.0 says 189 bytes. It must be wrong because the size of structure becomes 255 bytes and the specification says it must be 256 bytes in total. In fact, an example code in the spec. does it right, though. While we are at it, fix some i386-isms. - Remove state buffer size limitation. It is no longer necessary since sys/compat/x86bios/x86bios.c r198251. - Move int 0x10 vector test into vesa_bios_post() as we always do it anyway.
2009-10-23 18:41:00 +00:00
vmode.v_modeattr = le16toh(vmode.v_modeattr);
vmode.v_wgran = le16toh(vmode.v_wgran);
vmode.v_wsize = le16toh(vmode.v_wsize);
vmode.v_waseg = le16toh(vmode.v_waseg);
vmode.v_wbseg = le16toh(vmode.v_wbseg);
vmode.v_posfunc = le32toh(vmode.v_posfunc);
vmode.v_bpscanline = le16toh(vmode.v_bpscanline);
vmode.v_width = le16toh(vmode.v_width);
vmode.v_height = le16toh(vmode.v_height);
vmode.v_lfb = le32toh(vmode.v_lfb);
vmode.v_offscreen = le32toh(vmode.v_offscreen);
vmode.v_offscreensize = le16toh(vmode.v_offscreensize);
vmode.v_linbpscanline = le16toh(vmode.v_linbpscanline);
vmode.v_maxpixelclock = le32toh(vmode.v_maxpixelclock);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* reject unsupported modes */
#if 0
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
if ((vmode.v_modeattr &
(V_MODESUPP | V_MODEOPTINFO | V_MODENONVGA)) !=
(V_MODESUPP | V_MODEOPTINFO))
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
continue;
#else
if ((vmode.v_modeattr & V_MODEOPTINFO) == 0) {
#if VESA_DEBUG > 1
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
printf("Rejecting VESA %s mode: %d x %d x %d bpp "
" attr = %x\n",
vmode.v_modeattr & V_MODEGRAPHICS ?
"graphics" : "text",
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vmode.v_width, vmode.v_height, vmode.v_bpp,
vmode.v_modeattr);
#endif
continue;
}
#endif
Be extremely careful when we determine bytes per scan line information. First, we compare mode table data against minimum value. If the mode table does not make sense, we set the minimum in the mode info. When we actually set the mode, we try VESA BIOS function and compare it against the previous value. If it makes more sense, update the information.
2010-03-23 23:10:17 +00:00
bsize = vesa_get_bpscanline(&vmode) * vmode.v_height;
Map and report actual video memory we need.
2010-02-06 00:52:42 +00:00
if ((vmode.v_modeattr & V_MODEGRAPHICS) != 0)
bsize *= vmode.v_planes;
/* Does it have enough memory to support this mode? */
if (msize < bsize) {
#if VESA_DEBUG > 1
printf("Rejecting VESA %s mode: %d x %d x %d bpp "
" attr = %x, not enough memory\n",
vmode.v_modeattr & V_MODEGRAPHICS ?
"graphics" : "text",
vmode.v_width, vmode.v_height, vmode.v_bpp,
vmode.v_modeattr);
#endif
continue;
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* expand the array if necessary */
if (modes >= vesa_vmode_max) {
vesa_vmode_max += MODE_TABLE_DELTA;
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
p = malloc(sizeof(*vesa_vmode) * (vesa_vmode_max + 1),
M_DEVBUF, M_WAITOK);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#if VESA_DEBUG > 1
printf("vesa_bios_init(): modes:%d, vesa_mode_max:%d\n",
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
modes, vesa_vmode_max);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#endif
if (modes > 0) {
bcopy(vesa_vmode, p, sizeof(*vesa_vmode)*modes);
free(vesa_vmode, M_DEVBUF);
}
vesa_vmode = p;
}
#if VESA_DEBUG > 1
printf("Found VESA %s mode: %d x %d x %d bpp\n",
vmode.v_modeattr & V_MODEGRAPHICS ? "graphics" : "text",
vmode.v_width, vmode.v_height, vmode.v_bpp);
#endif
if (is_via_cle266) {
if ((vmode.v_width & 0xff00) >> 8 == vmode.v_height - 1) {
vmode.v_width &= 0xff;
vmode.v_waseg = 0xb8000 >> 4;
}
}
/* copy some fields */
bzero(&vesa_vmode[modes], sizeof(vesa_vmode[modes]));
vesa_vmode[modes].vi_mode = vesa_vmodetab[i];
vesa_vmode[modes].vi_width = vmode.v_width;
vesa_vmode[modes].vi_height = vmode.v_height;
vesa_vmode[modes].vi_depth = vmode.v_bpp;
vesa_vmode[modes].vi_planes = vmode.v_planes;
vesa_vmode[modes].vi_cwidth = vmode.v_cwidth;
vesa_vmode[modes].vi_cheight = vmode.v_cheight;
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
vesa_vmode[modes].vi_window = (vm_offset_t)vmode.v_waseg << 4;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* XXX window B */
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
vesa_vmode[modes].vi_window_size = vmode.v_wsize * 1024;
vesa_vmode[modes].vi_window_gran = vmode.v_wgran * 1024;
Map and report actual video memory we need.
2010-02-06 00:52:42 +00:00
if (vmode.v_modeattr & V_MODELFB)
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vesa_vmode[modes].vi_buffer = vmode.v_lfb;
Map and report actual video memory we need.
2010-02-06 00:52:42 +00:00
vesa_vmode[modes].vi_buffer_size = bsize;
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
vesa_vmode[modes].vi_mem_model =
vesa_translate_mmodel(vmode.v_memmodel);
Re-add accidentally removed pixel format for direct memory model.
2010-02-23 01:02:11 +00:00
switch (vesa_vmode[modes].vi_mem_model) {
case V_INFO_MM_DIRECT:
if ((vmode.v_modeattr & V_MODELFB) != 0 &&
vers >= 0x0300) {
vesa_vmode[modes].vi_pixel_fields[0] =
vmode.v_linredfieldpos;
vesa_vmode[modes].vi_pixel_fields[1] =
vmode.v_lingreenfieldpos;
vesa_vmode[modes].vi_pixel_fields[2] =
vmode.v_linbluefieldpos;
vesa_vmode[modes].vi_pixel_fields[3] =
vmode.v_linresfieldpos;
vesa_vmode[modes].vi_pixel_fsizes[0] =
vmode.v_linredmasksize;
vesa_vmode[modes].vi_pixel_fsizes[1] =
vmode.v_lingreenmasksize;
vesa_vmode[modes].vi_pixel_fsizes[2] =
vmode.v_linbluemasksize;
vesa_vmode[modes].vi_pixel_fsizes[3] =
vmode.v_linresmasksize;
} else {
vesa_vmode[modes].vi_pixel_fields[0] =
vmode.v_redfieldpos;
vesa_vmode[modes].vi_pixel_fields[1] =
vmode.v_greenfieldpos;
vesa_vmode[modes].vi_pixel_fields[2] =
vmode.v_bluefieldpos;
vesa_vmode[modes].vi_pixel_fields[3] =
vmode.v_resfieldpos;
vesa_vmode[modes].vi_pixel_fsizes[0] =
vmode.v_redmasksize;
vesa_vmode[modes].vi_pixel_fsizes[1] =
vmode.v_greenmasksize;
vesa_vmode[modes].vi_pixel_fsizes[2] =
vmode.v_bluemasksize;
vesa_vmode[modes].vi_pixel_fsizes[3] =
vmode.v_resmasksize;
}
/* FALLTHROUGH */
case V_INFO_MM_PACKED:
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
vesa_vmode[modes].vi_pixel_size = (vmode.v_bpp + 7) / 8;
Re-add accidentally removed pixel format for direct memory model.
2010-02-23 01:02:11 +00:00
break;
}
Remove dead code and fix style(9) bugs.
2010-02-03 22:20:57 +00:00
vesa_vmode[modes].vi_flags =
vesa_translate_flags(vmode.v_modeattr) | V_INFO_VESA;
Reject some VESA graphics modes if the controller does not have enough memory to support them. Some adapters have expansible memory slots but video mode table is static. In this case, unusable modes may be reported. Submitted by: paradox (ddkprog yahoo com) (initial patch)
2009-09-25 18:08:38 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
++modes;
}
vesa_vmode[modes].vi_mode = EOT;
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (bootverbose)
printf("VESA: %d mode(s) found\n", modes);
has_vesa_bios = (modes > 0);
if (!has_vesa_bios)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
goto fail;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
x86bios_free(vmbuf, sizeof(*buf));
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
Probe supported states for save/restore function. Some VBE implementation refuses to save/restore states if an unsupported bit is set.
2012-02-23 19:05:57 +00:00
/* Probe supported save/restore states. */
for (i = 0; i < 4; i++)
if (vesa_bios_state_buf_size(1 << i) > 0)
vesa_state |= 1 << i;
if (vesa_state != 0)
vesa_state_buf_size = vesa_bios_state_buf_size(vesa_state);
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
vesa_palette = x86bios_alloc(&vesa_palette_offs,
VESA_PALETTE_SIZE + vesa_state_buf_size, M_WAITOK);
if (vesa_state_buf_size > 0) {
vesa_state_buf = vesa_palette + VESA_PALETTE_SIZE;
vesa_state_buf_offs = vesa_palette_offs + VESA_PALETTE_SIZE;
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
return (0);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
fail:
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
if (vesa_bios != NULL) {
x86bios_set_intr(0x10, vesa_bios_int10);
vesa_bios_offs = VESA_BIOS_OFFSET;
x86bios_free(vesa_bios, vesa_bios_size);
vesa_bios = NULL;
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (vmbuf != NULL)
x86bios_free(vmbuf, sizeof(buf));
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
if (vesa_adp_info != NULL) {
free(vesa_adp_info, M_DEVBUF);
vesa_adp_info = NULL;
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (vesa_oemstr != NULL) {
free(vesa_oemstr, M_DEVBUF);
vesa_oemstr = NULL;
}
if (vesa_venderstr != NULL) {
free(vesa_venderstr, M_DEVBUF);
vesa_venderstr = NULL;
}
if (vesa_prodstr != NULL) {
free(vesa_prodstr, M_DEVBUF);
vesa_prodstr = NULL;
}
if (vesa_revstr != NULL) {
free(vesa_revstr, M_DEVBUF);
vesa_revstr = NULL;
}
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static void
vesa_clear_modes(video_info_t *info, int color)
{
while (info->vi_mode != EOT) {
if ((info->vi_flags & V_INFO_COLOR) != color)
info->vi_mode = NA;
++info;
}
}
/* entry points */
static int
vesa_configure(int flags)
{
video_adapter_t *adp;
int adapters;
int error;
int i;
if (vesa_init_done)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (flags & VIO_PROBE_ONLY)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/*
* If the VESA module has already been loaded, abort loading
* the module this time.
*/
for (i = 0; (adp = vid_get_adapter(i)) != NULL; ++i) {
if (adp->va_flags & V_ADP_VESA)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (ENXIO);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (adp->va_type == KD_VGA)
break;
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/*
* The VGA adapter is not found. This is because either
* 1) the VGA driver has not been initialized, or 2) the VGA card
* is not present. If 1) is the case, we shall defer
* initialization for now and try again later.
*/
if (adp == NULL) {
vga_sub_configure = vesa_configure;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (ENODEV);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
/* count number of registered adapters */
for (++i; vid_get_adapter(i) != NULL; ++i)
;
adapters = i;
/* call VESA BIOS */
vesa_adp = adp;
if (vesa_bios_init()) {
vesa_adp = NULL;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (ENXIO);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
vesa_adp->va_flags |= V_ADP_VESA;
/* remove conflicting modes if we have more than one adapter */
if (adapters > 1) {
vesa_clear_modes(vesa_vmode,
(vesa_adp->va_flags & V_ADP_COLOR) ?
V_INFO_COLOR : 0);
}
if ((error = vesa_load_ioctl()) == 0) {
prevvidsw = vidsw[vesa_adp->va_index];
vidsw[vesa_adp->va_index] = &vesavidsw;
vesa_init_done = TRUE;
} else {
vesa_adp = NULL;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (error);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
#if 0
static int
vesa_nop(void)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
#endif
static int
vesa_error(void)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_probe(int unit, video_adapter_t **adpp, void *arg, int flags)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->probe)(unit, adpp, arg, flags));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_init(int unit, video_adapter_t *adp, int flags)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->init)(unit, adp, flags));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_get_info(video_adapter_t *adp, int mode, video_info_t *info)
{
int i;
if ((*prevvidsw->get_info)(adp, mode, info) == 0)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (adp != vesa_adp)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
mode = vesa_map_gen_mode_num(vesa_adp->va_type,
vesa_adp->va_flags & V_ADP_COLOR, mode);
for (i = 0; vesa_vmode[i].vi_mode != EOT; ++i) {
if (vesa_vmode[i].vi_mode == NA)
continue;
if (vesa_vmode[i].vi_mode == mode) {
*info = vesa_vmode[i];
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_query_mode(video_adapter_t *adp, video_info_t *info)
{
int i;
if ((*prevvidsw->query_mode)(adp, info) == 0)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (adp != vesa_adp)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (ENODEV);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
for (i = 0; vesa_vmode[i].vi_mode != EOT; ++i) {
if ((info->vi_width != 0)
&& (info->vi_width != vesa_vmode[i].vi_width))
continue;
if ((info->vi_height != 0)
&& (info->vi_height != vesa_vmode[i].vi_height))
continue;
if ((info->vi_cwidth != 0)
&& (info->vi_cwidth != vesa_vmode[i].vi_cwidth))
continue;
if ((info->vi_cheight != 0)
&& (info->vi_cheight != vesa_vmode[i].vi_cheight))
continue;
if ((info->vi_depth != 0)
&& (info->vi_depth != vesa_vmode[i].vi_depth))
continue;
if ((info->vi_planes != 0)
&& (info->vi_planes != vesa_vmode[i].vi_planes))
continue;
/* pixel format, memory model */
if ((info->vi_flags != 0)
&& (info->vi_flags != vesa_vmode[i].vi_flags))
continue;
*info = vesa_vmode[i];
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (ENODEV);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_set_mode(video_adapter_t *adp, int mode)
{
video_info_t info;
if (adp != vesa_adp)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->set_mode)(adp, mode));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
mode = vesa_map_gen_mode_num(adp->va_type,
adp->va_flags & V_ADP_COLOR, mode);
#if VESA_DEBUG > 0
printf("VESA: set_mode(): %d(%x) -> %d(%x)\n",
adp->va_mode, adp->va_mode, mode, mode);
#endif
/*
* If the current mode is a VESA mode and the new mode is not,
* restore the state of the adapter first by setting one of the
* standard VGA mode, so that non-standard, extended SVGA registers
* are set to the state compatible with the standard VGA modes.
* Otherwise (*prevvidsw->set_mode)() may not be able to set up
* the new mode correctly.
*/
if (VESA_MODE(adp->va_mode)) {
Fix VESA color palette corruption: - VBE 3.0 says palette format resets to 6-bit mode when video mode changes. We simply set 8-bit mode when we switch modes if the adapter supports it. - VBE 3.0 also says if the mode is not VGA compatible, we must use VBE function to save/restore palette. Otherwise, VGA function may be used. Thus, reinstate the save/load palette functions only for non-VGA compatible modes regardless of its palette format. - Let vesa(4) set VESA modes even if vga(4) claims to support it. - Reset default palette if VESA pixel mode is set initially. - Fix more style nits.
2009-11-03 20:22:09 +00:00
if (!VESA_MODE(mode) &&
(*prevvidsw->get_info)(adp, mode, &info) == 0) {
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
if ((adp->va_flags & V_ADP_DAC8) != 0) {
vesa_bios_set_dac(6);
adp->va_flags &= ~V_ADP_DAC8;
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
int10_set_mode(adp->va_initial_bios_mode);
if (adp->va_info.vi_flags & V_INFO_LINEAR)
Replace some homegrown functions with better/correct ones.
2010-02-05 18:00:24 +00:00
pmap_unmapdev(adp->va_buffer,
Map entire video memory again. This is a partial backout of r203535. Although we do not use them all directly, it seems VGA render may access unmapped memory region and cause kernel panic.
2010-03-23 22:35:52 +00:00
vesa_adp_info->v_memsize * 64 * 1024);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/*
* Once (*prevvidsw->get_info)() succeeded,
* (*prevvidsw->set_mode)() below won't fail...
*/
}
}
/* we may not need to handle this mode after all... */
Fix VESA color palette corruption: - VBE 3.0 says palette format resets to 6-bit mode when video mode changes. We simply set 8-bit mode when we switch modes if the adapter supports it. - VBE 3.0 also says if the mode is not VGA compatible, we must use VBE function to save/restore palette. Otherwise, VGA function may be used. Thus, reinstate the save/load palette functions only for non-VGA compatible modes regardless of its palette format. - Let vesa(4) set VESA modes even if vga(4) claims to support it. - Reset default palette if VESA pixel mode is set initially. - Fix more style nits.
2009-11-03 20:22:09 +00:00
if (!VESA_MODE(mode) && (*prevvidsw->set_mode)(adp, mode) == 0)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* is the new mode supported? */
if (vesa_get_info(adp, mode, &info))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* assert(VESA_MODE(mode)); */
#if VESA_DEBUG > 0
printf("VESA: about to set a VESA mode...\n");
#endif
/* don't use the linear frame buffer for text modes. XXX */
if (!(info.vi_flags & V_INFO_GRAPHICS))
info.vi_flags &= ~V_INFO_LINEAR;
Make sure the VESA mode number is between 256 and 511 inclusive.
2012-02-16 22:46:00 +00:00
if ((info.vi_flags & V_INFO_LINEAR) != 0)
mode |= 0x4000;
if (vesa_bios_set_mode(mode))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
/* Palette format is reset by the above VBE function call. */
adp->va_flags &= ~V_ADP_DAC8;
if ((vesa_adp_info->v_flags & V_DAC8) != 0 &&
(info.vi_flags & V_INFO_GRAPHICS) != 0 &&
vesa_bios_set_dac(8) > 6)
adp->va_flags |= V_ADP_DAC8;
Fix VESA color palette corruption: - VBE 3.0 says palette format resets to 6-bit mode when video mode changes. We simply set 8-bit mode when we switch modes if the adapter supports it. - VBE 3.0 also says if the mode is not VGA compatible, we must use VBE function to save/restore palette. Otherwise, VGA function may be used. Thus, reinstate the save/load palette functions only for non-VGA compatible modes regardless of its palette format. - Let vesa(4) set VESA modes even if vga(4) claims to support it. - Reset default palette if VESA pixel mode is set initially. - Fix more style nits.
2009-11-03 20:22:09 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (adp->va_info.vi_flags & V_INFO_LINEAR)
Map entire video memory again. This is a partial backout of r203535. Although we do not use them all directly, it seems VGA render may access unmapped memory region and cause kernel panic.
2010-03-23 22:35:52 +00:00
pmap_unmapdev(adp->va_buffer,
vesa_adp_info->v_memsize * 64 * 1024);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#if VESA_DEBUG > 0
printf("VESA: mode set!\n");
#endif
Make sure the VESA mode number is between 256 and 511 inclusive.
2012-02-16 22:46:00 +00:00
vesa_adp->va_mode = mode & 0x1ff; /* Mode number is 9-bit. */
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vesa_adp->va_flags &= ~V_ADP_COLOR;
vesa_adp->va_flags |=
(info.vi_flags & V_INFO_COLOR) ? V_ADP_COLOR : 0;
vesa_adp->va_crtc_addr =
(vesa_adp->va_flags & V_ADP_COLOR) ? COLOR_CRTC : MONO_CRTC;
Be extremely careful when we determine bytes per scan line information. First, we compare mode table data against minimum value. If the mode table does not make sense, we set the minimum in the mode info. When we actually set the mode, we try VESA BIOS function and compare it against the previous value. If it makes more sense, update the information.
2010-03-23 23:10:17 +00:00
vesa_adp->va_line_width = info.vi_buffer_size / info.vi_height;
if ((info.vi_flags & V_INFO_GRAPHICS) != 0)
vesa_adp->va_line_width /= info.vi_planes;
Do not penalize correct or correctable VESA mode tables by calling another VBE function. Most problems should be corrected by the mode table sanity check and we only need the paranoid in extremely rare cases.
2010-03-25 17:51:05 +00:00
#ifdef MODE_TABLE_BROKEN
Be extremely careful when we determine bytes per scan line information. First, we compare mode table data against minimum value. If the mode table does not make sense, we set the minimum in the mode info. When we actually set the mode, we try VESA BIOS function and compare it against the previous value. If it makes more sense, update the information.
2010-03-23 23:10:17 +00:00
/* If VBE function returns bigger bytes per scan line, use it. */
Do not penalize correct or correctable VESA mode tables by calling another VBE function. Most problems should be corrected by the mode table sanity check and we only need the paranoid in extremely rare cases.
2010-03-25 17:51:05 +00:00
{
int bpsl = vesa_bios_get_line_length();
if (bpsl > vesa_adp->va_line_width) {
vesa_adp->va_line_width = bpsl;
info.vi_buffer_size = bpsl * info.vi_height;
if ((info.vi_flags & V_INFO_GRAPHICS) != 0)
info.vi_buffer_size *= info.vi_planes;
}
Be extremely careful when we determine bytes per scan line information. First, we compare mode table data against minimum value. If the mode table does not make sense, we set the minimum in the mode info. When we actually set the mode, we try VESA BIOS function and compare it against the previous value. If it makes more sense, update the information.
2010-03-23 23:10:17 +00:00
}
Do not penalize correct or correctable VESA mode tables by calling another VBE function. Most problems should be corrected by the mode table sanity check and we only need the paranoid in extremely rare cases.
2010-03-25 17:51:05 +00:00
#endif
Be extremely careful when we determine bytes per scan line information. First, we compare mode table data against minimum value. If the mode table does not make sense, we set the minimum in the mode info. When we actually set the mode, we try VESA BIOS function and compare it against the previous value. If it makes more sense, update the information.
2010-03-23 23:10:17 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (info.vi_flags & V_INFO_LINEAR) {
#if VESA_DEBUG > 1
printf("VESA: setting up LFB\n");
#endif
vesa_adp->va_buffer =
Replace some homegrown functions with better/correct ones.
2010-02-05 18:00:24 +00:00
(vm_offset_t)pmap_mapdev_attr(info.vi_buffer,
Map entire video memory again. This is a partial backout of r203535. Although we do not use them all directly, it seems VGA render may access unmapped memory region and cause kernel panic.
2010-03-23 22:35:52 +00:00
vesa_adp_info->v_memsize * 64 * 1024, PAT_WRITE_COMBINING);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vesa_adp->va_window = vesa_adp->va_buffer;
Map and report actual video memory we need.
2010-02-06 00:52:42 +00:00
vesa_adp->va_window_size = info.vi_buffer_size / info.vi_planes;
vesa_adp->va_window_gran = info.vi_buffer_size / info.vi_planes;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
} else {
vesa_adp->va_buffer = 0;
Correct virtual address of frame buffer for non-linear mode. Reported by: Marc UBM Bocklet (ubm dot freebsd at googlemail dot com)
2010-02-03 22:07:50 +00:00
vesa_adp->va_window = (vm_offset_t)x86bios_offset(info.vi_window);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vesa_adp->va_window_size = info.vi_window_size;
vesa_adp->va_window_gran = info.vi_window_gran;
}
Map and report actual video memory we need.
2010-02-06 00:52:42 +00:00
vesa_adp->va_buffer_size = info.vi_buffer_size;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vesa_adp->va_window_orig = 0;
vesa_adp->va_disp_start.x = 0;
vesa_adp->va_disp_start.y = 0;
#if VESA_DEBUG > 0
Reject some VESA graphics modes if the controller does not have enough memory to support them. Some adapters have expansible memory slots but video mode table is static. In this case, unusable modes may be reported. Submitted by: paradox (ddkprog yahoo com) (initial patch)
2009-09-25 18:08:38 +00:00
printf("vesa_set_mode(): vi_width:%d, line_width:%d\n",
info.vi_width, vesa_adp->va_line_width);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#endif
bcopy(&info, &vesa_adp->va_info, sizeof(vesa_adp->va_info));
/* move hardware cursor out of the way */
(*vidsw[vesa_adp->va_index]->set_hw_cursor)(vesa_adp, -1, -1);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_save_font(video_adapter_t *adp, int page, int fontsize, int fontwidth,
u_char *data, int ch, int count)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->save_font)(adp, page, fontsize, fontwidth, data,
ch, count));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_load_font(video_adapter_t *adp, int page, int fontsize, int fontwidth,
u_char *data, int ch, int count)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->load_font)(adp, page, fontsize, fontwidth, data,
ch, count));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_show_font(video_adapter_t *adp, int page)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->show_font)(adp, page));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_save_palette(video_adapter_t *adp, u_char *palette)
{
int bits;
Fall back to VGA palette functions if VESA function failed and DAC is still in 6-bit mode. Although we have to check non-VGA compatibility bit here, it seems there are too many broken VESA BIOSes out to rely on it.
2010-03-23 22:50:22 +00:00
if (adp == vesa_adp && VESA_MODE(adp->va_mode)) {
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
bits = (adp->va_flags & V_ADP_DAC8) != 0 ? 8 : 6;
Fall back to VGA palette functions if VESA function failed and DAC is still in 6-bit mode. Although we have to check non-VGA compatibility bit here, it seems there are too many broken VESA BIOSes out to rely on it.
2010-03-23 22:50:22 +00:00
if (vesa_bios_save_palette(0, 256, palette, bits) == 0)
return (0);
Use VESA palette load/save functions if VESA BIOS says the current palette format is higher than 6-bit instead of relying VGA compatibility flag. This fixes palette problem of NVIDIA GeForce 6600. Reduce code differences between palette load/save functions while we are here. Tested by: danfe
2010-01-27 17:00:42 +00:00
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->save_palette)(adp, palette));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_load_palette(video_adapter_t *adp, u_char *palette)
{
int bits;
Fall back to VGA palette functions if VESA function failed and DAC is still in 6-bit mode. Although we have to check non-VGA compatibility bit here, it seems there are too many broken VESA BIOSes out to rely on it.
2010-03-23 22:50:22 +00:00
if (adp == vesa_adp && VESA_MODE(adp->va_mode)) {
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
bits = (adp->va_flags & V_ADP_DAC8) != 0 ? 8 : 6;
Fall back to VGA palette functions if VESA function failed and DAC is still in 6-bit mode. Although we have to check non-VGA compatibility bit here, it seems there are too many broken VESA BIOSes out to rely on it.
2010-03-23 22:50:22 +00:00
if (vesa_bios_load_palette(0, 256, palette, bits) == 0)
return (0);
Use VESA palette load/save functions if VESA BIOS says the current palette format is higher than 6-bit instead of relying VGA compatibility flag. This fixes palette problem of NVIDIA GeForce 6600. Reduce code differences between palette load/save functions while we are here. Tested by: danfe
2010-01-27 17:00:42 +00:00
}
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->load_palette)(adp, palette));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_set_border(video_adapter_t *adp, int color)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->set_border)(adp, color));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_save_state(video_adapter_t *adp, void *p, size_t size)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (adp != vesa_adp)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->save_state)(adp, p, size));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
if (vesa_state_buf_size == 0)
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (size == 0)
- When we restore VESA state, try BIOS POST earlier. VESA restore state function may not work properly if we don't. Turn off hardware cursor as vesa_set_mode() does. - Add VBE 3.0 specific fields in VESA mode structure and pack it. Note the padding is 190 bytes although VBE 3.0 says 189 bytes. It must be wrong because the size of structure becomes 255 bytes and the specification says it must be 256 bytes in total. In fact, an example code in the spec. does it right, though. While we are at it, fix some i386-isms. - Remove state buffer size limitation. It is no longer necessary since sys/compat/x86bios/x86bios.c r198251. - Move int 0x10 vector test into vesa_bios_post() as we always do it anyway.
2009-10-23 18:41:00 +00:00
return (offsetof(adp_state_t, regs) + vesa_state_buf_size);
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
if (size < (offsetof(adp_state_t, regs) + vesa_state_buf_size))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
((adp_state_t *)p)->sig = V_STATE_SIG;
bzero(((adp_state_t *)p)->regs, vesa_state_buf_size);
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
return (vesa_bios_save_restore(STATE_SAVE, ((adp_state_t *)p)->regs));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_load_state(video_adapter_t *adp, void *p)
{
Set the initial mode for the adapter after executing VESA BIOS POST. There is no need to set initial mode for BIOS.
2012-02-16 22:51:44 +00:00
int mode;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if ((adp != vesa_adp) || (((adp_state_t *)p)->sig != V_STATE_SIG))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->load_state)(adp, p));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
if (vesa_state_buf_size == 0)
- When we restore VESA state, try BIOS POST earlier. VESA restore state function may not work properly if we don't. Turn off hardware cursor as vesa_set_mode() does. - Add VBE 3.0 specific fields in VESA mode structure and pack it. Note the padding is 190 bytes although VBE 3.0 says 189 bytes. It must be wrong because the size of structure becomes 255 bytes and the specification says it must be 256 bytes in total. In fact, an example code in the spec. does it right, though. While we are at it, fix some i386-isms. - Remove state buffer size limitation. It is no longer necessary since sys/compat/x86bios/x86bios.c r198251. - Move int 0x10 vector test into vesa_bios_post() as we always do it anyway.
2009-10-23 18:41:00 +00:00
return (1);
Save/restore VGA state from vga_pci.c instead of relying on vga_isa.c. It was not working because we were saving its state after the device was powered down. Simplify vesa_load_state() as the culprit is fixed now.
2009-11-05 22:58:50 +00:00
/* Try BIOS POST to restore a sane state. */
(void)vesa_bios_post();
Set the initial mode for the adapter after executing VESA BIOS POST. There is no need to set initial mode for BIOS.
2012-02-16 22:51:44 +00:00
mode = adp->va_mode;
(void)vesa_set_mode(adp, adp->va_initial_mode);
Fix a typo introduced in r231843.
2012-02-23 18:59:32 +00:00
if (mode != adp->va_initial_mode)
Set the initial mode for the adapter after executing VESA BIOS POST. There is no need to set initial mode for BIOS.
2012-02-16 22:51:44 +00:00
(void)vesa_set_mode(adp, mode);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
return (vesa_bios_save_restore(STATE_LOAD, ((adp_state_t *)p)->regs));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
#if 0
static int
vesa_get_origin(video_adapter_t *adp, off_t *offset)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f05;
regs.R_BL = 0x10;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (1);
Style(9) fixes. Submitted by: swell.k at gmail.com
2009-09-21 08:53:26 +00:00
*offset = regs.DX * adp->va_window_gran;
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
#endif
static int
vesa_set_origin(video_adapter_t *adp, off_t offset)
{
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
x86regs_t regs;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/*
* This function should return as quickly as possible to
* maintain good performance of the system. For this reason,
* error checking is kept minimal and let the VESA BIOS to
* detect error.
*/
if (adp != vesa_adp)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->set_win_org)(adp, offset));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* if this is a linear frame buffer, do nothing */
if (adp->va_info.vi_flags & V_INFO_LINEAR)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* XXX */
if (adp->va_window_gran == 0)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f05;
regs.R_DX = offset / adp->va_window_gran;
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (regs.R_AX != 0x004f)
return (1);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
x86bios_init_regs(&regs);
regs.R_AX = 0x4f05;
regs.R_BL = 1;
regs.R_DX = offset / adp->va_window_gran;
- Use FreeBSD function naming convention. - Change x86biosCall() to more appropriate x86bios_intr().[1] Discussed with: delphij, paradox (ddkprog yahoo com) Submitted by: paradox (ddkprog yahoo com)[1]
2009-09-24 19:24:42 +00:00
x86bios_intr(&regs, 0x10);
- Teach vesa(4) and dpms(4) about x86emu. [1] - Add vesa kernel options for amd64. - Connect libvgl library and splash kernel modules to amd64 build. - Connect manual page dpms(4) to amd64 build. - Remove old vesa/dpms files. Submitted by: paradox <ddkprog yahoo com> [1], swell k at gmail.com (with some minor tweaks)
2009-09-09 09:50:31 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
adp->va_window_orig = (offset/adp->va_window_gran)*adp->va_window_gran;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0); /* XXX */
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_read_hw_cursor(video_adapter_t *adp, int *col, int *row)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->read_hw_cursor)(adp, col, row));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_set_hw_cursor(video_adapter_t *adp, int col, int row)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->set_hw_cursor)(adp, col, row));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_set_hw_cursor_shape(video_adapter_t *adp, int base, int height,
int celsize, int blink)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->set_hw_cursor_shape)(adp, base, height, celsize,
blink));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_blank_display(video_adapter_t *adp, int mode)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* XXX: use VESA DPMS */
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->blank_display)(adp, mode));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
Update d_mmap() to accept vm_ooffset_t and vm_memattr_t. This replaces d_mmap() with the d_mmap2() implementation and also changes the type of offset to vm_ooffset_t. Purge d_mmap2(). All driver modules will need to be rebuilt since D_VERSION is also bumped. Reviewed by: jhb@ MFC after: Not in this lifetime...
2009-12-29 21:51:28 +00:00
vesa_mmap(video_adapter_t *adp, vm_ooffset_t offset, vm_paddr_t *paddr,
int prot, vm_memattr_t *memattr)
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#if VESA_DEBUG > 0
Update d_mmap() to accept vm_ooffset_t and vm_memattr_t. This replaces d_mmap() with the d_mmap2() implementation and also changes the type of offset to vm_ooffset_t. Purge d_mmap2(). All driver modules will need to be rebuilt since D_VERSION is also bumped. Reviewed by: jhb@ MFC after: Not in this lifetime...
2009-12-29 21:51:28 +00:00
printf("vesa_mmap(): window:0x%tx, buffer:0x%tx, offset:0x%jx\n",
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
adp->va_info.vi_window, adp->va_info.vi_buffer, offset);
#endif
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if ((adp == vesa_adp) &&
(adp->va_info.vi_flags & V_INFO_LINEAR) != 0) {
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* va_window_size == va_buffer_size/vi_planes */
/* XXX: is this correct? */
if (offset > adp->va_window_size - PAGE_SIZE)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (-1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
*paddr = adp->va_info.vi_buffer + offset;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Update d_mmap() to accept vm_ooffset_t and vm_memattr_t. This replaces d_mmap() with the d_mmap2() implementation and also changes the type of offset to vm_ooffset_t. Purge d_mmap2(). All driver modules will need to be rebuilt since D_VERSION is also bumped. Reviewed by: jhb@ MFC after: Not in this lifetime...
2009-12-29 21:51:28 +00:00
return ((*prevvidsw->mmap)(adp, offset, paddr, prot, memattr));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_clear(video_adapter_t *adp)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->clear)(adp));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_fill_rect(video_adapter_t *adp, int val, int x, int y, int cx, int cy)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->fill_rect)(adp, val, x, y, cx, cy));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_bitblt(video_adapter_t *adp,...)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* FIXME */
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
get_palette(video_adapter_t *adp, int base, int count,
u_char *red, u_char *green, u_char *blue, u_char *trans)
{
u_char *r;
u_char *g;
u_char *b;
int bits;
int error;
Use VESA palette load/save functions if VESA BIOS says the current palette format is higher than 6-bit instead of relying VGA compatibility flag. This fixes palette problem of NVIDIA GeForce 6600. Reduce code differences between palette load/save functions while we are here. Tested by: danfe
2010-01-27 17:00:42 +00:00
if (base < 0 || base >= 256 || count < 0 || count > 256)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if ((base + count) > 256)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Use VESA palette load/save functions if VESA BIOS says the current palette format is higher than 6-bit instead of relying VGA compatibility flag. This fixes palette problem of NVIDIA GeForce 6600. Reduce code differences between palette load/save functions while we are here. Tested by: danfe
2010-01-27 17:00:42 +00:00
if (!VESA_MODE(adp->va_mode))
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
bits = (adp->va_flags & V_ADP_DAC8) != 0 ? 8 : 6;
Fix VESA color palette corruption: - VBE 3.0 says palette format resets to 6-bit mode when video mode changes. We simply set 8-bit mode when we switch modes if the adapter supports it. - VBE 3.0 also says if the mode is not VGA compatible, we must use VBE function to save/restore palette. Otherwise, VGA function may be used. Thus, reinstate the save/load palette functions only for non-VGA compatible modes regardless of its palette format. - Let vesa(4) set VESA modes even if vga(4) claims to support it. - Reset default palette if VESA pixel mode is set initially. - Fix more style nits.
2009-11-03 20:22:09 +00:00
r = malloc(count * 3, M_DEVBUF, M_WAITOK);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
g = r + count;
b = g + count;
error = vesa_bios_save_palette2(base, count, r, g, b, bits);
if (error == 0) {
copyout(r, red, count);
copyout(g, green, count);
copyout(b, blue, count);
if (trans != NULL) {
bzero(r, count);
copyout(r, trans, count);
}
}
free(r, M_DEVBUF);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (error);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
set_palette(video_adapter_t *adp, int base, int count,
u_char *red, u_char *green, u_char *blue, u_char *trans)
{
u_char *r;
u_char *g;
u_char *b;
int bits;
int error;
Use VESA palette load/save functions if VESA BIOS says the current palette format is higher than 6-bit instead of relying VGA compatibility flag. This fixes palette problem of NVIDIA GeForce 6600. Reduce code differences between palette load/save functions while we are here. Tested by: danfe
2010-01-27 17:00:42 +00:00
if (base < 0 || base >= 256 || count < 0 || count > 256)
return (1);
if ((base + count) > 256)
return (1);
if (!VESA_MODE(adp->va_mode))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
bits = (adp->va_flags & V_ADP_DAC8) != 0 ? 8 : 6;
Fix VESA color palette corruption: - VBE 3.0 says palette format resets to 6-bit mode when video mode changes. We simply set 8-bit mode when we switch modes if the adapter supports it. - VBE 3.0 also says if the mode is not VGA compatible, we must use VBE function to save/restore palette. Otherwise, VGA function may be used. Thus, reinstate the save/load palette functions only for non-VGA compatible modes regardless of its palette format. - Let vesa(4) set VESA modes even if vga(4) claims to support it. - Reset default palette if VESA pixel mode is set initially. - Fix more style nits.
2009-11-03 20:22:09 +00:00
r = malloc(count * 3, M_DEVBUF, M_WAITOK);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
g = r + count;
b = g + count;
copyin(red, r, count);
copyin(green, g, count);
copyin(blue, b, count);
error = vesa_bios_load_palette2(base, count, r, g, b, bits);
free(r, M_DEVBUF);
Fix VESA color palette corruption: - VBE 3.0 says palette format resets to 6-bit mode when video mode changes. We simply set 8-bit mode when we switch modes if the adapter supports it. - VBE 3.0 also says if the mode is not VGA compatible, we must use VBE function to save/restore palette. Otherwise, VGA function may be used. Thus, reinstate the save/load palette functions only for non-VGA compatible modes regardless of its palette format. - Let vesa(4) set VESA modes even if vga(4) claims to support it. - Reset default palette if VESA pixel mode is set initially. - Fix more style nits.
2009-11-03 20:22:09 +00:00
return (error);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_ioctl(video_adapter_t *adp, u_long cmd, caddr_t arg)
{
int bytes;
if (adp != vesa_adp)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
switch (cmd) {
case FBIO_SETWINORG: /* set frame buffer window origin */
if (!VESA_MODE(adp->va_mode))
return (*prevvidsw->ioctl)(adp, cmd, arg);
return (vesa_set_origin(adp, *(off_t *)arg) ? ENODEV : 0);
case FBIO_SETDISPSTART: /* set display start address */
if (!VESA_MODE(adp->va_mode))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (vesa_bios_set_start(((video_display_start_t *)arg)->x,
((video_display_start_t *)arg)->y))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (ENODEV);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
adp->va_disp_start.x = ((video_display_start_t *)arg)->x;
adp->va_disp_start.y = ((video_display_start_t *)arg)->y;
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
case FBIO_SETLINEWIDTH: /* set line length in pixel */
if (!VESA_MODE(adp->va_mode))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (vesa_bios_set_line_length(*(u_int *)arg, &bytes, NULL))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (ENODEV);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
adp->va_line_width = bytes;
#if VESA_DEBUG > 1
printf("new line width:%d\n", adp->va_line_width);
#endif
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
case FBIO_GETPALETTE: /* get color palette */
if (get_palette(adp, ((video_color_palette_t *)arg)->index,
((video_color_palette_t *)arg)->count,
((video_color_palette_t *)arg)->red,
((video_color_palette_t *)arg)->green,
((video_color_palette_t *)arg)->blue,
((video_color_palette_t *)arg)->transparent))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
case FBIO_SETPALETTE: /* set color palette */
if (set_palette(adp, ((video_color_palette_t *)arg)->index,
((video_color_palette_t *)arg)->count,
((video_color_palette_t *)arg)->red,
((video_color_palette_t *)arg)->green,
((video_color_palette_t *)arg)->blue,
((video_color_palette_t *)arg)->transparent))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
case FBIOGETCMAP: /* get color palette */
if (get_palette(adp, ((struct fbcmap *)arg)->index,
((struct fbcmap *)arg)->count,
((struct fbcmap *)arg)->red,
((struct fbcmap *)arg)->green,
((struct fbcmap *)arg)->blue, NULL))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
case FBIOPUTCMAP: /* set color palette */
if (set_palette(adp, ((struct fbcmap *)arg)->index,
((struct fbcmap *)arg)->count,
((struct fbcmap *)arg)->red,
((struct fbcmap *)arg)->green,
((struct fbcmap *)arg)->blue, NULL))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
default:
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return ((*prevvidsw->ioctl)(adp, cmd, arg));
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
}
static int
vesa_diag(video_adapter_t *adp, int level)
{
int error;
/* call the previous handler first */
error = (*prevvidsw->diag)(adp, level);
if (error)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (error);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (adp != vesa_adp)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (level <= 0)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_bios_info(int level)
{
#if VESA_DEBUG > 1
struct vesa_mode vmode;
int i;
#endif
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
uint16_t vers;
vers = vesa_adp_info->v_version;
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
if (bootverbose) {
/* general adapter information */
printf(
"VESA: v%d.%d, %dk memory, flags:0x%x, mode table:%p (%x)\n",
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
(vers >> 12) * 10 + ((vers & 0x0f00) >> 8),
((vers & 0x00f0) >> 4) * 10 + (vers & 0x000f),
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
vesa_adp_info->v_memsize * 64, vesa_adp_info->v_flags,
vesa_vmodetab, vesa_adp_info->v_modetable);
/* OEM string */
if (vesa_oemstr != NULL)
printf("VESA: %s\n", vesa_oemstr);
}
if (level <= 0)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (vers >= 0x0200 && bootverbose) {
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* vender name, product name, product revision */
printf("VESA: %s %s %s\n",
(vesa_venderstr != NULL) ? vesa_venderstr : "unknown",
(vesa_prodstr != NULL) ? vesa_prodstr : "unknown",
(vesa_revstr != NULL) ? vesa_revstr : "?");
}
#if VESA_DEBUG > 1
/* mode information */
for (i = 0;
(i < (M_VESA_MODE_MAX - M_VESA_BASE + 1))
&& (vesa_vmodetab[i] != 0xffff); ++i) {
Use M_WAITOK for VESA BIOS initialization consistently.
2010-06-23 23:34:56 +00:00
if (vesa_bios_get_mode(vesa_vmodetab[i], &vmode, M_NOWAIT))
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
continue;
/* print something for diagnostic purpose */
printf("VESA: mode:0x%03x, flags:0x%04x",
vesa_vmodetab[i], vmode.v_modeattr);
if (vmode.v_modeattr & V_MODEOPTINFO) {
if (vmode.v_modeattr & V_MODEGRAPHICS) {
printf(", G %dx%dx%d %d, ",
vmode.v_width, vmode.v_height,
vmode.v_bpp, vmode.v_planes);
} else {
printf(", T %dx%d, ",
vmode.v_width, vmode.v_height);
}
printf("font:%dx%d, ",
vmode.v_cwidth, vmode.v_cheight);
printf("pages:%d, mem:%d",
vmode.v_ipages + 1, vmode.v_memmodel);
}
if (vmode.v_modeattr & V_MODELFB) {
printf("\nVESA: LFB:0x%x, off:0x%x, off_size:0x%x",
vmode.v_lfb, vmode.v_offscreen,
vmode.v_offscreensize*1024);
}
printf("\n");
printf("VESA: window A:0x%x (%x), window B:0x%x (%x), ",
vmode.v_waseg, vmode.v_waattr,
vmode.v_wbseg, vmode.v_wbattr);
printf("size:%dk, gran:%dk\n",
vmode.v_wsize, vmode.v_wgran);
}
#endif /* VESA_DEBUG > 1 */
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
/* module loading */
static int
vesa_load(void)
{
int error;
if (vesa_init_done)
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (0);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_init(&vesa_lock, "VESA lock", NULL, MTX_DEF);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/* locate a VGA adapter */
vesa_adp = NULL;
error = vesa_configure(0);
if (error == 0)
vesa_bios_info(bootverbose);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (error);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_unload(void)
{
u_char palette[256*3];
int error;
/* if the adapter is currently in a VESA mode, don't unload */
if ((vesa_adp != NULL) && VESA_MODE(vesa_adp->va_mode))
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (EBUSY);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
/*
* FIXME: if there is at least one vty which is in a VESA mode,
* we shouldn't be unloading! XXX
*/
if ((error = vesa_unload_ioctl()) == 0) {
if (vesa_adp != NULL) {
Yet another attempt to make palette loading more safer: - Add a separate palette data for 8-bit DAC mode when SC_PIXEL_MODE is set and fill it up with default gray-scale palette data for text. Now we don't have to set `hint.sc.0.vesa_mode' to get the default palette data. - Add a new adapter flag, V_ADP_DAC8 to track whether the controller is using 8-bit palette format and load correct palette when switching modes. - Set 8-bit DAC mode only for non-VGA compatible graphics mode.
2010-02-23 21:51:14 +00:00
if ((vesa_adp->va_flags & V_ADP_DAC8) != 0) {
vesa_bios_save_palette(0, 256, palette, 8);
vesa_bios_set_dac(6);
vesa_adp->va_flags &= ~V_ADP_DAC8;
vesa_bios_load_palette(0, 256, palette, 6);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
vesa_adp->va_flags &= ~V_ADP_VESA;
vidsw[vesa_adp->va_index] = prevvidsw;
}
}
Add an experimental feature to shadow video BIOS. Long ago, this trick was supported by many BIOSes to improve performance of VESA BIOS calls for real mode OSes but it is not our intention here. However, this may help some platforms where the video ROMs are inaccessible after suspend, for example. Note it may consume up to 64K bytes of contiguous memory depending on video controller model when it is enabled. This feature can be disabled by setting zero to 'debug.vesa.shadow_rom' loader tunable via loader(8) or loader.conf(5). The default is 1 (enabled), for now.
2010-08-25 22:09:02 +00:00
if (vesa_bios != NULL) {
x86bios_set_intr(0x10, vesa_bios_int10);
x86bios_free(vesa_bios, vesa_bios_size);
}
Allocate memory for VBE info block with malloc(9), not as static local.
2010-04-07 21:38:42 +00:00
if (vesa_adp_info != NULL)
free(vesa_adp_info, M_DEVBUF);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
if (vesa_oemstr != NULL)
free(vesa_oemstr, M_DEVBUF);
if (vesa_venderstr != NULL)
free(vesa_venderstr, M_DEVBUF);
if (vesa_prodstr != NULL)
free(vesa_prodstr, M_DEVBUF);
if (vesa_revstr != NULL)
free(vesa_revstr, M_DEVBUF);
if (vesa_vmode != &vesa_vmode_empty)
free(vesa_vmode, M_DEVBUF);
Preallocate buffers for palette and state. Do not save DAC registers as we reset DAC mode and restore palette data while we are resuming always.
2010-07-13 19:48:20 +00:00
if (vesa_palette != NULL)
x86bios_free(vesa_palette,
VESA_PALETTE_SIZE + vesa_state_buf_size);
Protect shared palette and state buffer with a mutex. Remove defunct spltty() calls while I am here.
2010-08-07 05:46:04 +00:00
mtx_destroy(&vesa_lock);
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (error);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static int
vesa_mod_event(module_t mod, int type, void *data)
{
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
switch (type) {
case MOD_LOAD:
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (vesa_load());
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
case MOD_UNLOAD:
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (vesa_unload());
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
Rewrite x86bios and update its dependent drivers. - Do not map entire real mode memory (1MB). Instead, we map IVT/BDA and ROM area separately. Most notably, ROM area is mapped as device memory (uncacheable) as it should be. User memory is dynamically allocated and free'ed with contigmalloc(9) and contigfree(9). Remove now redundant and potentially dangerous x86bios_alloc.c. If this emulator ever grows to support non-PC hardware, we may implement it with rman(9) later. - Move all host-specific initializations from x86emu_util.c to x86bios.c and remove now unnecessary x86emu_util.c. Currently, non-PC hardware is not supported. We may use bus_space(9) later when the KPI is fixed. - Replace all bzero() calls for emulated registers with more obviously named x86bios_init_regs(). This function also initializes DS and SS properly. - Add x86bios_get_intr(). This function checks if the interrupt vector is available for the platform. It is not necessary for PC-compatible hardware but it may be needed later. ;-) - Do not try turning off monitor if DPMS does not support the state. - Allocate stable memory for VESA OEM strings instead of just holding pointers to them. They may or may not be accessible always. Fix a memory leak of video mode table while I am here. - Add (experimental) BIOS POST call for vesa(4). This function calls VGA BIOS POST code from the current VGA option ROM. Some video controllers cannot save and restore the state properly even if it is claimed to be supported. Usually the symptom is blank display after resuming from suspend state. If the video mode does not match the previous mode after restoring, we try BIOS POST and force the known good initial state. Some magic was taken from NetBSD (and it was taken from vbetool, I believe.) - Add a loader tunable for vgapci(4) to give a hint to dpms(4) and vesa(4) to identify who owns the VESA BIOS. This is very useful for multi-display adapter setup. By default, the POST video controller is automatically probed and the tunable "hw.pci.default_vgapci_unit" is set to corresponding vgapci unit number. You may override it from loader but it is very unlikely to be necessary. Unfortunately only AGP/PCI/PCI-E controllers can be matched because ISA controller does not have necessary device IDs. - Fix a long standing bug in state save/restore function. The state buffer pointer should be ES:BX, not ES:DI according to VBE 3.0. If it ever worked, that's because BX was always zero. :-) - Clean up register initializations more clearer per VBE 3.0. - Fix a lot of style issues with vesa(4).
2009-10-19 20:58:10 +00:00
return (EOPNOTSUPP);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
}
static moduledata_t vesa_mod = {
"vesa",
vesa_mod_event,
NULL,
};
DECLARE_MODULE(vesa, vesa_mod, SI_SUB_DRIVERS, SI_ORDER_MIDDLE);
Collapase interrupt supporting functions to a new module, and switch from x86emu to this new module. This changeset also brings a fix for bugs introduced with the initial x86emu commit, which prevents the user from using some display mode or cause instant reboots during mode switch. Submitted by: paradox <ddkprog yahoo com>
2009-09-21 08:17:57 +00:00
MODULE_DEPEND(vesa, x86bios, 1, 1, 1);
Copy the following files to new places, a subsequent commit would remove them from the old place. This commit necessary so that the tree would not enter a broken state. sys/i386/isa/vesa.c -> dev/fb/vesa.c sys/i386/include/pc/vesa.h -> dev/fb/vesa.h sys/i386/isa/dpms.c -> dev/dpms/dpms.c
2009-09-09 07:45:08 +00:00
#endif /* VGA_NO_MODE_CHANGE */
Reference in New Issue Copy Permalink