Commit Graph

197 Commits

Author SHA1 Message Date
Jung-uk Kim
f6ff063a6e 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
Jung-uk Kim
90017de999 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
Jung-uk Kim
c74dea9cc1 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
Jung-uk Kim
2b13785931 Fix a debugging message under bootverbose. This address is not linear. 2010-08-25 22:48:18 +00:00
Jung-uk Kim
a73f383ee9 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
Jung-uk Kim
90f06c5eae 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
Jung-uk Kim
708358cdb6 Initialize a variable before its use. 2010-07-13 19:58:06 +00:00
Jung-uk Kim
917b850dce 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
Jung-uk Kim
061b7e5463 Plug a possible memory leak.
Submitted by:	Yamagi Burmeister (lists at yamagi dot org)
2010-07-06 18:08:55 +00:00
Jung-uk Kim
c3cbd4125e Fix a possible null pointer dereference. A patch for -STABLE was
Submitted by:	Yamagi Burmeister (lists at yamagi dot org)
2010-07-06 18:05:05 +00:00
Jung-uk Kim
28ef508f86 Use M_WAITOK for VESA BIOS initialization consistently. 2010-06-23 23:34:56 +00:00
Jung-uk Kim
362487c0ba 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
Konstantin Belousov
95882b9865 Remove unused i586 optimized bcopy/bzero/etc implementations that utilize
FPU registers for copying. Remove the switch table and jumps from
bcopy/bzero/... to the actual implementation.
As a side-effect, i486-optimized bzero is removed.

Reviewed by:	bde
Tested by:	pho (previous version)
2010-06-23 10:40:28 +00:00
Jung-uk Kim
a46dd0b9e2 Restore the previous VESA mode after BIOS POST, just in case. 2010-05-18 22:30:55 +00:00
Jung-uk Kim
842795abcc 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
Jung-uk Kim
004e8bae43 Allocate memory for VBE info block with malloc(9), not as static local. 2010-04-07 21:38:42 +00:00
Jung-uk Kim
bb6ba5f6b9 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
Jung-uk Kim
1e161437f8 Teach VGA framebuffer about 8-bit palette format for VESA. 2010-03-24 15:37:47 +00:00
Jung-uk Kim
b266b96af9 Add my copyright here. It seems I have contributed enough code. :-) 2010-03-23 23:19:23 +00:00
Jung-uk Kim
3b56b13e8e 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
Jung-uk Kim
51cb3184fd 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
Jung-uk Kim
5fb3ac06b7 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
Jung-uk Kim
8d521790d0 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
Jung-uk Kim
0dc682e029 Re-add accidentally removed pixel format for direct memory model. 2010-02-23 01:02:11 +00:00
Jung-uk Kim
6a6289a2b8 Fix FBIO_ADPINFO ioctl on amd64. 2010-02-23 01:00:19 +00:00
Jung-uk Kim
316bcf56a2 Map and report actual video memory we need. 2010-02-06 00:52:42 +00:00
Jung-uk Kim
f801fa102d Replace some homegrown functions with better/correct ones. 2010-02-05 18:00:24 +00:00
Jung-uk Kim
bfd1f292d4 Remove dead code and fix style(9) bugs. 2010-02-03 22:20:57 +00:00
Jung-uk Kim
c5813a500a Use bytes per scan line from mode table. The previous implementation did
not reflect actual number of bytes when it was not exactly width * bpp * 8.
2010-02-03 22:17:30 +00:00
Jung-uk Kim
a95cde0132 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
Jung-uk Kim
dd6155ac34 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
Robert Noland
cfd7bacef2 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
Jung-uk Kim
2f28bf307c - Remove a redundant check for dpms(4).
- Test a cheaper function first.
2009-11-12 18:16:35 +00:00
Jung-uk Kim
bf71c25f87 - Partially revert hackish r198964 and r199002.
- Add a proxy driver vgapm to help vgapci to save/load VGA state.
- Move device_set_desc() to the right place while we are here.

Reviewed by:	jhb
2009-11-12 17:56:56 +00:00
Jung-uk Kim
b66e2b8e50 Remove duplicate suspend/resume code from vga_pci.c and let vga(4) register
itself to an associated PCI device if it exists.  It is little bit hackish
but it should fix build without frame buffer driver since r198964.
Fix some style(9) nits in vga_isa.c while we are here.
2009-11-06 20:32:26 +00:00
Jung-uk Kim
2259d74c68 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
Jung-uk Kim
00b9e39e68 Do not probe video mode if we are not going to use it. 2009-11-04 17:30:48 +00:00
Jung-uk Kim
421cd2f2fb Restore color palette format if we reset video mode. 2009-11-04 01:00:28 +00:00
Jung-uk Kim
7ac3e951e3 Save/restore VGA color palette while suspending and resuming. 2009-11-04 00:58:20 +00:00
Jung-uk Kim
761eeb5fff 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
Jung-uk Kim
9871fde43f Remove a redundant byte swapping in the previous commit. 2009-10-23 19:02:53 +00:00
Jung-uk Kim
a0ce33465f - 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
Jung-uk Kim
3219f535d9 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
Jung-uk Kim
54680451c3 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
Xin LI
57f20544f5 Fix LINT build. 2009-09-25 02:19:57 +00:00
Jung-uk Kim
4d849a58c4 Some broken VESA BIOSes, e.g., IBM T23, return wrong value from
vesa_bios_get_line_length() in graphics mode.  Always calculate the value
from known video info instead.

Submitted by:	paradox (ddkprog yahoo com)
2009-09-24 20:10:00 +00:00
Jung-uk Kim
a867274808 - 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
Jung-uk Kim
19de5df5e5 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
Xin LI
f27b147e40 Initialize registers to zero before calling the interrupt handlers inside
emulator.  This fixes VESA related freeze observed on some systems.

Submitted by:	paradox <ddkprog yahoo com>
2009-09-23 02:45:02 +00:00
Xin LI
13c937f899 Style(9) fixes.
Submitted by:	swell.k at gmail.com
2009-09-21 08:53:26 +00:00