d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
6244c6e7db
freebsd-dev/sys/mips/cavium/octeon_ebt3000_cf.c

739 lines
21 KiB
C
Raw Normal View History

Add Cavium's standard copyright to those files that are currently lacking a copyright/license statement. All these files were in the Cavium FreeBSD source drop and appear to be written by Cavium (some are nearly verbatim copies of files from the cnusers' 1.9.0 SDK, which also uses this copyright).
2010-01-28 20:46:40 +00:00
/***********************license start***************
* Copyright (c) 2003-2008 Cavium Networks (support@cavium.com). All rights
* reserved.
*
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are
* met:
*
* * Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
*
* * 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.
*
* * Neither the name of Cavium Networks nor the names of
* its contributors may be used to endorse or promote products
* derived from this software without specific prior written
* permission.
*
* TO THE MAXIMUM EXTENT PERMITTED BY LAW, THE SOFTWARE IS PROVIDED "AS IS"
* AND WITH ALL FAULTS AND CAVIUM NETWORKS MAKES NO PROMISES, REPRESENTATIONS
* OR WARRANTIES, EITHER EXPRESS, IMPLIED, STATUTORY, OR OTHERWISE, WITH
* RESPECT TO THE SOFTWARE, INCLUDING ITS CONDITION, ITS CONFORMITY TO ANY
* REPRESENTATION OR DESCRIPTION, OR THE EXISTENCE OF ANY LATENT OR PATENT
* DEFECTS, AND CAVIUM SPECIFICALLY DISCLAIMS ALL IMPLIED (IF ANY) WARRANTIES
* OF TITLE, MERCHANTABILITY, NONINFRINGEMENT, FITNESS FOR A PARTICULAR
* PURPOSE, LACK OF VIRUSES, ACCURACY OR COMPLETENESS, QUIET ENJOYMENT, QUIET
* POSSESSION OR CORRESPONDENCE TO DESCRIPTION. THE ENTIRE RISK ARISING OUT
* OF USE OR PERFORMANCE OF THE SOFTWARE LIES WITH YOU.
*
*
* For any questions regarding licensing please contact marketing@caviumnetworks.com
*
***********************license end**************************************/
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/*
* octeon_ebt3000_cf.c
*
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include <sys/param.h>
#include <sys/bio.h>
#include <sys/systm.h>
#include <sys/sysctl.h>
Update the port of FreeBSD to Cavium Octeon to use the Cavium Simple Executive library: o) Increase inline unit / large function growth limits for MIPS to accommodate the needs of the Simple Executive, which uses a shocking amount of inlining. o) Remove TARGET_OCTEON and use CPU_CNMIPS to do things required by cnMIPS and the Octeon SoC. o) Add OCTEON_VENDOR_LANNER to use Lanner's allocation of vendor-specific board numbers, specifically to support the MR320. o) Add OCTEON_BOARD_CAPK_0100ND to hard-wire configuration for the CAPK-0100nd, which improperly uses an evaluation board's board number and breaks board detection at runtime. This board is sold by Portwell as the CAM-0100. o) Add support for the RTC available on some Octeon boards. o) Add support for the Octeon PCI bus. Note that rman_[sg]et_virtual for IO ports can not work unless building for n64. o) Clean up the CompactFlash driver to use Simple Executive macros and structures where possible (it would be advisable to use the Simple Executive API to set the PIO mode, too, but that is not done presently.) Also use structures from FreeBSD's ATA layer rather than structures copied from Linux. o) Print available Octeon SoC features on boot. o) Add support for the Octeon timecounter. o) Use the Simple Executive's routines rather than local copies for doing reads and writes to 64-bit addresses and use its macros for various device addresses rather than using local copies. o) Rename octeon_board_real to octeon_is_simulation to reduce differences with Cavium-provided code originally written for Linux. Also make it use the same simplified test that the Simple Executive and Linux both use rather than our complex one. o) Add support for the Octeon CIU, which is the main interrupt unit, as a bus to use normal interrupt allocation and setup routines. o) Use the Simple Executive's bootmem facility to allocate physical memory for the kernel, rather than assuming we know which addresses we can steal. NB: This may reduce the amount of RAM the kernel reports you as having if you are leaving large temporary allocations made by U-Boot allocated when starting FreeBSD. o) Add a port of the Cavium-provided Ethernet driver for Linux. This changes Ethernet interface naming from rgmxN to octeN. The new driver has vast improvements over the old one, both in performance and functionality, but does still have some features which have not been ported entirely and there may be unimplemented code that can be hit in everyday use. I will make every effort to correct those as they are reported. o) Support loading the kernel on non-contiguous cores. o) Add very conservative support for harvesting randomness from the Octeon random number device. o) Turn SMP on by default. o) Clean up the style of the Octeon kernel configurations a little and make them compile with -march=octeon. o) Add support for the Lanner MR320 and the CAPK-0100nd to the Simple Executive. o) Modify the Simple Executive to build on FreeBSD and to build without executive-config.h or cvmx-config.h. In the future we may want to revert part of these changes and supply executive-config.h and cvmx-config.h and access to the options contained in those files via kernel configuration files. o) Modify the Simple Executive USB routines to support getting and setting of the USB PID.
2010-07-20 19:25:11 +00:00
#include <sys/ata.h>
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
#include <sys/bus.h>
#include <sys/kernel.h>
#include <sys/module.h>
#include <sys/rman.h>
#include <sys/power.h>
#include <sys/smp.h>
#include <sys/time.h>
#include <sys/timetc.h>
#include <sys/malloc.h>
#include <geom/geom.h>
#include <machine/clock.h>
#include <machine/locore.h>
#include <machine/md_var.h>
#include <machine/cpuregs.h>
Cope with the move and if_timer going way.
2010-01-11 04:29:26 +00:00
#include <mips/cavium/octeon_pcmap_regs.h>
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
Update the port of FreeBSD to Cavium Octeon to use the Cavium Simple Executive library: o) Increase inline unit / large function growth limits for MIPS to accommodate the needs of the Simple Executive, which uses a shocking amount of inlining. o) Remove TARGET_OCTEON and use CPU_CNMIPS to do things required by cnMIPS and the Octeon SoC. o) Add OCTEON_VENDOR_LANNER to use Lanner's allocation of vendor-specific board numbers, specifically to support the MR320. o) Add OCTEON_BOARD_CAPK_0100ND to hard-wire configuration for the CAPK-0100nd, which improperly uses an evaluation board's board number and breaks board detection at runtime. This board is sold by Portwell as the CAM-0100. o) Add support for the RTC available on some Octeon boards. o) Add support for the Octeon PCI bus. Note that rman_[sg]et_virtual for IO ports can not work unless building for n64. o) Clean up the CompactFlash driver to use Simple Executive macros and structures where possible (it would be advisable to use the Simple Executive API to set the PIO mode, too, but that is not done presently.) Also use structures from FreeBSD's ATA layer rather than structures copied from Linux. o) Print available Octeon SoC features on boot. o) Add support for the Octeon timecounter. o) Use the Simple Executive's routines rather than local copies for doing reads and writes to 64-bit addresses and use its macros for various device addresses rather than using local copies. o) Rename octeon_board_real to octeon_is_simulation to reduce differences with Cavium-provided code originally written for Linux. Also make it use the same simplified test that the Simple Executive and Linux both use rather than our complex one. o) Add support for the Octeon CIU, which is the main interrupt unit, as a bus to use normal interrupt allocation and setup routines. o) Use the Simple Executive's bootmem facility to allocate physical memory for the kernel, rather than assuming we know which addresses we can steal. NB: This may reduce the amount of RAM the kernel reports you as having if you are leaving large temporary allocations made by U-Boot allocated when starting FreeBSD. o) Add a port of the Cavium-provided Ethernet driver for Linux. This changes Ethernet interface naming from rgmxN to octeN. The new driver has vast improvements over the old one, both in performance and functionality, but does still have some features which have not been ported entirely and there may be unimplemented code that can be hit in everyday use. I will make every effort to correct those as they are reported. o) Support loading the kernel on non-contiguous cores. o) Add very conservative support for harvesting randomness from the Octeon random number device. o) Turn SMP on by default. o) Clean up the style of the Octeon kernel configurations a little and make them compile with -march=octeon. o) Add support for the Lanner MR320 and the CAPK-0100nd to the Simple Executive. o) Modify the Simple Executive to build on FreeBSD and to build without executive-config.h or cvmx-config.h. In the future we may want to revert part of these changes and supply executive-config.h and cvmx-config.h and access to the options contained in those files via kernel configuration files. o) Modify the Simple Executive USB routines to support getting and setting of the USB PID.
2010-07-20 19:25:11 +00:00
#include <contrib/octeon-sdk/cvmx.h>
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* ATA Commands */
#define CMD_READ_SECTOR 0x20
#define CMD_WRITE_SECTOR 0x30
#define CMD_IDENTIFY 0xEC
/* The ATA Task File */
#define TF_DATA 0x00
#define TF_ERROR 0x01
#define TF_PRECOMP 0x01
#define TF_SECTOR_COUNT 0x02
#define TF_SECTOR_NUMBER 0x03
#define TF_CYL_LSB 0x04
#define TF_CYL_MSB 0x05
#define TF_DRV_HEAD 0x06
#define TF_STATUS 0x07
#define TF_COMMAND 0x07
/* Status Register */
#define STATUS_BSY 0x80 /* Drive is busy */
#define STATUS_RDY 0x40 /* Drive is ready */
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
#define STATUS_DF 0x20 /* Device fault */
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
#define STATUS_DRQ 0x08 /* Data can be transferred */
/* Miscelaneous */
#define SECTOR_SIZE 512
#define WAIT_DELAY 1000
#define NR_TRIES 1000
#define SWAP_SHORT(x) ((x << 8) | (x >> 8))
#define MODEL_STR_SIZE 40
/* Globals */
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
/*
* There's three bus types supported by this driver.
*
* CF_8 -- Traditional PC Card IDE interface on an 8-bit wide bus. We assume
* the bool loader has configure attribute memory properly. We then access
* the device like old-school 8-bit IDE card (which is all a traditional PC Card
* interface really is).
* CF_16 -- Traditional PC Card IDE interface on a 16-bit wide bus. Registers on
* this bus are 16-bits wide too. When accessing registers in the task file, you
* have to do it in 16-bit chunks, and worry about masking out what you don't want
* or ORing together the traditional 8-bit values. We assume the bootloader does
* the right attribute memory initialization dance.
* CF_TRUE_IDE_8 - CF Card wired to True IDE mode. There's no Attribute memory
* space at all. Instead all the traditional 8-bit registers are there, but
* on a 16-bit bus where addr0 isn't wired. This means we need to read/write them
* 16-bit chunks, but only the lower 8 bits are valid. We do not (and can not)
* access this like CF_16 with the comingled registers. Yet we can't access
* this like CF_8 because of the register offset. Except the TF_DATA register
* appears to be full width?
*/
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
void *base_addr;
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
int bus_type;
#define CF_8 1 /* 8-bit bus, no offsets - PC Card */
#define CF_16 2 /* 16-bit bus, registers shared - PC Card */
#define CF_TRUE_IDE_8 3 /* 16-bit bus, only lower 8-bits, TrueIDE */
const char *const cf_type[] = {
"impossible type",
"CF 8-bit",
"CF 16-bit",
"True IDE"
};
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* Device parameters */
struct drive_param{
union {
char buf[SECTOR_SIZE];
Update the port of FreeBSD to Cavium Octeon to use the Cavium Simple Executive library: o) Increase inline unit / large function growth limits for MIPS to accommodate the needs of the Simple Executive, which uses a shocking amount of inlining. o) Remove TARGET_OCTEON and use CPU_CNMIPS to do things required by cnMIPS and the Octeon SoC. o) Add OCTEON_VENDOR_LANNER to use Lanner's allocation of vendor-specific board numbers, specifically to support the MR320. o) Add OCTEON_BOARD_CAPK_0100ND to hard-wire configuration for the CAPK-0100nd, which improperly uses an evaluation board's board number and breaks board detection at runtime. This board is sold by Portwell as the CAM-0100. o) Add support for the RTC available on some Octeon boards. o) Add support for the Octeon PCI bus. Note that rman_[sg]et_virtual for IO ports can not work unless building for n64. o) Clean up the CompactFlash driver to use Simple Executive macros and structures where possible (it would be advisable to use the Simple Executive API to set the PIO mode, too, but that is not done presently.) Also use structures from FreeBSD's ATA layer rather than structures copied from Linux. o) Print available Octeon SoC features on boot. o) Add support for the Octeon timecounter. o) Use the Simple Executive's routines rather than local copies for doing reads and writes to 64-bit addresses and use its macros for various device addresses rather than using local copies. o) Rename octeon_board_real to octeon_is_simulation to reduce differences with Cavium-provided code originally written for Linux. Also make it use the same simplified test that the Simple Executive and Linux both use rather than our complex one. o) Add support for the Octeon CIU, which is the main interrupt unit, as a bus to use normal interrupt allocation and setup routines. o) Use the Simple Executive's bootmem facility to allocate physical memory for the kernel, rather than assuming we know which addresses we can steal. NB: This may reduce the amount of RAM the kernel reports you as having if you are leaving large temporary allocations made by U-Boot allocated when starting FreeBSD. o) Add a port of the Cavium-provided Ethernet driver for Linux. This changes Ethernet interface naming from rgmxN to octeN. The new driver has vast improvements over the old one, both in performance and functionality, but does still have some features which have not been ported entirely and there may be unimplemented code that can be hit in everyday use. I will make every effort to correct those as they are reported. o) Support loading the kernel on non-contiguous cores. o) Add very conservative support for harvesting randomness from the Octeon random number device. o) Turn SMP on by default. o) Clean up the style of the Octeon kernel configurations a little and make them compile with -march=octeon. o) Add support for the Lanner MR320 and the CAPK-0100nd to the Simple Executive. o) Modify the Simple Executive to build on FreeBSD and to build without executive-config.h or cvmx-config.h. In the future we may want to revert part of these changes and supply executive-config.h and cvmx-config.h and access to the options contained in those files via kernel configuration files. o) Modify the Simple Executive USB routines to support getting and setting of the USB PID.
2010-07-20 19:25:11 +00:00
struct ata_params driveid;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
} u;
char model[MODEL_STR_SIZE];
uint32_t nr_sectors;
uint16_t sector_size;
uint16_t heads;
uint16_t tracks;
uint16_t sec_track;
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
};
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
/* Device softc */
struct cf_priv {
device_t dev;
struct drive_param drive_param;
struct bio_queue_head cf_bq;
struct g_geom *cf_geom;
struct g_provider *cf_provider;
};
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* GEOM class implementation */
static g_access_t cf_access;
static g_start_t cf_start;
static g_ioctl_t cf_ioctl;
struct g_class g_cf_class = {
.name = "CF",
.version = G_VERSION,
.start = cf_start,
.access = cf_access,
.ioctl = cf_ioctl,
};
Declare the CF GEOM class so that g_modevent will get called, the class will be recorded, etc. This fixes libgeom on Octeon.
2010-11-02 23:43:44 +00:00
DECLARE_GEOM_CLASS(g_cf_class, g_cf);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* Device methods */
static int cf_probe(device_t);
static void cf_identify(driver_t *, device_t);
static int cf_attach(device_t);
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
static void cf_attach_geom(void *, int);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* ATA methods */
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
static int cf_cmd_identify(struct cf_priv *);
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
static int cf_cmd_write(uint32_t, uint32_t, void *);
static int cf_cmd_read(uint32_t, uint32_t, void *);
static int cf_wait_busy(void);
static int cf_send_cmd(uint32_t, uint8_t);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* Miscelenous */
static void cf_swap_ascii(unsigned char[], char[]);
/* ------------------------------------------------------------------- *
* cf_access() *
* ------------------------------------------------------------------- */
static int cf_access (struct g_provider *pp, int r, int w, int e)
{
return (0);
}
/* ------------------------------------------------------------------- *
* cf_start() *
* ------------------------------------------------------------------- */
static void cf_start (struct bio *bp)
{
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
struct cf_priv *cf_priv;
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
int error;
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
cf_priv = bp->bio_to->geom->softc;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/*
* Handle actual I/O requests. The request is passed down through
* the bio struct.
*/
Don't assume that bio_cmd is a bitfield. Differential Revision: https://reviews.freebsd.org/D5591
2016-03-10 06:25:47 +00:00
switch (bp->bio_cmd) {
case BIO_GETATTR:
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
if (g_handleattr_int(bp, "GEOM::fwsectors", cf_priv->drive_param.sec_track))
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
return;
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
if (g_handleattr_int(bp, "GEOM::fwheads", cf_priv->drive_param.heads))
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
return;
g_io_deliver(bp, ENOIOCTL);
return;
Don't assume that bio_cmd is a bitfield. Differential Revision: https://reviews.freebsd.org/D5591
2016-03-10 06:25:47 +00:00
case BIO_READ:
error = cf_cmd_read(bp->bio_length / cf_priv->drive_param.sector_size,
bp->bio_offset / cf_priv->drive_param.sector_size, bp->bio_data);
break;
case BIO_WRITE:
error = cf_cmd_write(bp->bio_length / cf_priv->drive_param.sector_size,
bp->bio_offset/cf_priv->drive_param.sector_size, bp->bio_data);
break;
default:
printf("%s: unrecognized bio_cmd %x.\n", __func__, bp->bio_cmd);
error = ENOTSUP;
break;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Don't assume that bio_cmd is a bitfield. Differential Revision: https://reviews.freebsd.org/D5591
2016-03-10 06:25:47 +00:00
if (error != 0) {
g_io_deliver(bp, error);
return;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Don't assume that bio_cmd is a bitfield. Differential Revision: https://reviews.freebsd.org/D5591
2016-03-10 06:25:47 +00:00
bp->bio_resid = 0;
bp->bio_completed = bp->bio_length;
g_io_deliver(bp, 0);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
static int cf_ioctl (struct g_provider *pp, u_long cmd, void *data, int fflag, struct thread *td)
{
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
return (0);
}
static uint8_t cf_inb_8(int port)
{
/*
* Traditional 8-bit PC Card/CF bus access.
*/
if (bus_type == CF_8) {
volatile uint8_t *task_file = (volatile uint8_t *)base_addr;
return task_file[port];
}
/*
* True IDE access. lower 8 bits on a 16-bit bus (see above).
*/
volatile uint16_t *task_file = (volatile uint16_t *)base_addr;
return task_file[port] & 0xff;
}
static void cf_outb_8(int port, uint8_t val)
{
/*
* Traditional 8-bit PC Card/CF bus access.
*/
if (bus_type == CF_8) {
volatile uint8_t *task_file = (volatile uint8_t *)base_addr;
task_file[port] = val;
Add missing return that broke 8-bit CF support in refactoring in r222671. Tested on a Cavium CN5860-EVB-NIC4. This was broken for over a year.
2012-10-27 23:36:41 +00:00
return;
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
}
/*
* True IDE access. lower 8 bits on a 16-bit bus (see above).
*/
volatile uint16_t *task_file = (volatile uint16_t *)base_addr;
task_file[port] = val & 0xff;
}
static uint8_t cf_inb_16(int port)
{
volatile uint16_t *task_file = (volatile uint16_t *)base_addr;
uint16_t val = task_file[port / 2];
if (port & 1)
return (val >> 8) & 0xff;
return val & 0xff;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
static uint16_t cf_inw_16(int port)
{
volatile uint16_t *task_file = (volatile uint16_t *)base_addr;
uint16_t val = task_file[port / 2];
return val;
}
static void cf_outw_16(int port, uint16_t val)
{
volatile uint16_t *task_file = (volatile uint16_t *)base_addr;
task_file[port / 2] = val;
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* ------------------------------------------------------------------- *
* cf_cmd_read() *
* ------------------------------------------------------------------- *
*
* Read nr_sectors from the device starting from start_sector.
*/
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
static int cf_cmd_read (uint32_t nr_sectors, uint32_t start_sector, void *buf)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
{
unsigned long lba;
uint32_t count;
uint16_t *ptr_16;
uint8_t *ptr_8;
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
int error;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
ptr_8 = (uint8_t*)buf;
ptr_16 = (uint16_t*)buf;
lba = start_sector;
while (nr_sectors--) {
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
error = cf_send_cmd(lba, CMD_READ_SECTOR);
if (error != 0) {
printf("%s: cf_send_cmd(CMD_READ_SECTOR) failed: %d\n", __func__, error);
return (error);
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
switch (bus_type)
{
case CF_8:
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
for (count = 0; count < SECTOR_SIZE; count++) {
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
*ptr_8++ = cf_inb_8(TF_DATA);
if ((count & 0xf) == 0)
(void)cf_inb_8(TF_STATUS);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
case CF_TRUE_IDE_8:
case CF_16:
default:
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
for (count = 0; count < SECTOR_SIZE; count+=2) {
uint16_t temp;
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
temp = cf_inw_16(TF_DATA);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
*ptr_16++ = SWAP_SHORT(temp);
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
if ((count & 0xf) == 0)
(void)cf_inb_16(TF_STATUS);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
lba++;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
return (0);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
/* ------------------------------------------------------------------- *
* cf_cmd_write() *
* ------------------------------------------------------------------- *
*
* Write nr_sectors to the device starting from start_sector.
*/
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
static int cf_cmd_write (uint32_t nr_sectors, uint32_t start_sector, void *buf)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
{
uint32_t lba;
uint32_t count;
uint16_t *ptr_16;
uint8_t *ptr_8;
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
int error;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
lba = start_sector;
ptr_8 = (uint8_t*)buf;
ptr_16 = (uint16_t*)buf;
while (nr_sectors--) {
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
error = cf_send_cmd(lba, CMD_WRITE_SECTOR);
if (error != 0) {
printf("%s: cf_send_cmd(CMD_WRITE_SECTOR) failed: %d\n", __func__, error);
return (error);
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
switch (bus_type)
{
case CF_8:
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
for (count = 0; count < SECTOR_SIZE; count++) {
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
cf_outb_8(TF_DATA, *ptr_8++);
if ((count & 0xf) == 0)
(void)cf_inb_8(TF_STATUS);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
case CF_TRUE_IDE_8:
case CF_16:
default:
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
for (count = 0; count < SECTOR_SIZE; count+=2) {
uint16_t temp = *ptr_16++;
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
cf_outw_16(TF_DATA, SWAP_SHORT(temp));
if ((count & 0xf) == 0)
(void)cf_inb_16(TF_STATUS);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
lba++;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
return (0);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
/* ------------------------------------------------------------------- *
* cf_cmd_identify() *
* ------------------------------------------------------------------- *
*
* Read parameters and other information from the drive and store
* it in the drive_param structure
*
*/
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
static int cf_cmd_identify(struct cf_priv *cf_priv)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
{
int count;
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
int error;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
error = cf_send_cmd(0, CMD_IDENTIFY);
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
if (error != 0) {
printf("%s: identify failed: %d\n", __func__, error);
return (error);
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
switch (bus_type)
{
case CF_8:
for (count = 0; count < SECTOR_SIZE; count++)
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
cf_priv->drive_param.u.buf[count] = cf_inb_8(TF_DATA);
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
case CF_TRUE_IDE_8:
case CF_16:
default:
for (count = 0; count < SECTOR_SIZE; count += 2) {
uint16_t temp;
temp = cf_inw_16(TF_DATA);
/* endianess will be swapped below */
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
cf_priv->drive_param.u.buf[count] = (temp & 0xff);
cf_priv->drive_param.u.buf[count + 1] = (temp & 0xff00) >> 8;
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
}
break;
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
cf_swap_ascii(cf_priv->drive_param.u.driveid.model, cf_priv->drive_param.model);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
cf_priv->drive_param.sector_size = 512; //= SWAP_SHORT (cf_priv->drive_param.u.driveid.sector_bytes);
cf_priv->drive_param.heads = SWAP_SHORT (cf_priv->drive_param.u.driveid.current_heads);
cf_priv->drive_param.tracks = SWAP_SHORT (cf_priv->drive_param.u.driveid.current_cylinders);
cf_priv->drive_param.sec_track = SWAP_SHORT (cf_priv->drive_param.u.driveid.current_sectors);
cf_priv->drive_param.nr_sectors = (uint32_t)SWAP_SHORT (cf_priv->drive_param.u.driveid.lba_size_1) |
((uint32_t)SWAP_SHORT (cf_priv->drive_param.u.driveid.lba_size_2));
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
if (bootverbose) {
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
printf(" model %s\n", cf_priv->drive_param.model);
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
printf(" heads %d tracks %d sec_tracks %d sectors %d\n",
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
cf_priv->drive_param.heads, cf_priv->drive_param.tracks,
cf_priv->drive_param.sec_track, cf_priv->drive_param.nr_sectors);
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
return (0);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
/* ------------------------------------------------------------------- *
* cf_send_cmd() *
* ------------------------------------------------------------------- *
*
* Send command to read/write one sector specified by lba.
*
*/
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
static int cf_send_cmd (uint32_t lba, uint8_t cmd)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
{
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
switch (bus_type)
{
case CF_8:
case CF_TRUE_IDE_8:
while (cf_inb_8(TF_STATUS) & STATUS_BSY)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
DELAY(WAIT_DELAY);
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
cf_outb_8(TF_SECTOR_COUNT, 1);
cf_outb_8(TF_SECTOR_NUMBER, lba & 0xff);
cf_outb_8(TF_CYL_LSB, (lba >> 8) & 0xff);
cf_outb_8(TF_CYL_MSB, (lba >> 16) & 0xff);
cf_outb_8(TF_DRV_HEAD, ((lba >> 24) & 0xff) | 0xe0);
cf_outb_8(TF_COMMAND, cmd);
break;
case CF_16:
default:
while (cf_inb_16(TF_STATUS) & STATUS_BSY)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
DELAY(WAIT_DELAY);
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
cf_outw_16(TF_SECTOR_COUNT, 1 | ((lba & 0xff) << 8));
cf_outw_16(TF_CYL_LSB, ((lba >> 8) & 0xff) | (((lba >> 16) & 0xff) << 8));
cf_outw_16(TF_DRV_HEAD, (((lba >> 24) & 0xff) | 0xe0) | (cmd << 8));
break;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
return (cf_wait_busy());
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
/* ------------------------------------------------------------------- *
* cf_wait_busy() *
* ------------------------------------------------------------------- *
*
* Wait until the drive finishes a given command and data is
* ready to be transferred. This is done by repeatedly checking
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
* the BSY bit of the status register. When the controller is ready for
* data transfer, it clears the BSY bit and sets the DRQ bit.
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
*
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
* If the DF bit is ever set, we return error.
*
* This code originally spun on DRQ. If that behavior turns out to be
* necessary, a flag can be added or this function can be called
* repeatedly as long as it is returning ENXIO.
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
*/
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
static int cf_wait_busy (void)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
{
uint8_t status;
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
switch (bus_type)
{
case CF_8:
case CF_TRUE_IDE_8:
status = cf_inb_8(TF_STATUS);
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
while ((status & STATUS_BSY) == STATUS_BSY) {
if ((status & STATUS_DF) != 0) {
printf("%s: device fault (status=%x)\n", __func__, status);
return (EIO);
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
DELAY(WAIT_DELAY);
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
status = cf_inb_8(TF_STATUS);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
case CF_16:
default:
status = cf_inb_16(TF_STATUS);
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
while ((status & STATUS_BSY) == STATUS_BSY) {
if ((status & STATUS_DF) != 0) {
printf("%s: device fault (status=%x)\n", __func__, status);
return (EIO);
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
DELAY(WAIT_DELAY);
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
status = cf_inb_16(TF_STATUS);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
/* DRQ is only for when read data is actually available; check BSY */
/* Some vendors do assert DRQ, but not all. Check BSY instead. */
if (status & STATUS_BSY) {
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
printf("%s: device not ready (status=%x)\n", __func__, status);
return (ENXIO);
}
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
Check for device faults and for failures to set DRQ when expected, rather than spinning forever. This fixes booting with CF ejected. NB: I've made the driver pretty chatty about errors in case there's hardware that operates differently to mine, so we can easily track down any issues. Reviewed by: imp Sponsored by: Packet Forensics
2010-03-06 05:49:15 +00:00
return (0);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
/* ------------------------------------------------------------------- *
* cf_swap_ascii() *
* ------------------------------------------------------------------- *
*
* The ascii string returned by the controller specifying
* the model of the drive is byte-swaped. This routine
* corrects the byte ordering.
*
*/
static void cf_swap_ascii (unsigned char str1[], char str2[])
{
int i;
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
for(i = 0; i < MODEL_STR_SIZE; i++)
str2[i] = str1[i ^ 1];
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
/* ------------------------------------------------------------------- *
* cf_probe() *
* ------------------------------------------------------------------- */
static int cf_probe (device_t dev)
{
Actually check board type rather than using a specialized octeon_is_simulation function.
2012-10-30 06:36:14 +00:00
if (cvmx_sysinfo_get()->board_type == CVMX_BOARD_TYPE_SIM)
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
return (ENXIO);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
if (device_get_unit(dev) != 0) {
panic("can't attach more devices\n");
}
device_set_desc(dev, "Octeon Compact Flash Driver");
Devices that rely on hints or identify routines for discovery need to return BUS_PROBE_NOWILDCARD from their probe routines to avoid claiming wildcard devices on their parent bus. Do a sweep through the MIPS tree. MFC after: 2 weeks
2013-10-29 14:07:31 +00:00
return (BUS_PROBE_NOWILDCARD);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
/* ------------------------------------------------------------------- *
* cf_identify() *
* ------------------------------------------------------------------- *
*
* Find the bootbus region for the CF to determine
* 16 or 8 bit and check to see if device is
* inserted.
*
*/
static void cf_identify (driver_t *drv, device_t parent)
{
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
int bus_region;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
int count = 0;
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
cvmx_mio_boot_reg_cfgx_t cfg;
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
uint64_t phys_base;
Actually check board type rather than using a specialized octeon_is_simulation function.
2012-10-30 06:36:14 +00:00
if (cvmx_sysinfo_get()->board_type == CVMX_BOARD_TYPE_SIM)
Make compile.
2009-06-14 06:49:13 +00:00
return;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
phys_base = cvmx_sysinfo_get()->compact_flash_common_base_addr;
If the CF physical base is 0, attach no CF devices. This fixes a warning about a 0 passed to cvmx_phys_to_ptr on systems without a CF interface, such as the RSYS4GBE.
2012-10-31 04:23:36 +00:00
if (phys_base == 0)
return;
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
base_addr = cvmx_phys_to_ptr(phys_base);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
for (bus_region = 0; bus_region < 8; bus_region++)
{
Update the port of FreeBSD to Cavium Octeon to use the Cavium Simple Executive library: o) Increase inline unit / large function growth limits for MIPS to accommodate the needs of the Simple Executive, which uses a shocking amount of inlining. o) Remove TARGET_OCTEON and use CPU_CNMIPS to do things required by cnMIPS and the Octeon SoC. o) Add OCTEON_VENDOR_LANNER to use Lanner's allocation of vendor-specific board numbers, specifically to support the MR320. o) Add OCTEON_BOARD_CAPK_0100ND to hard-wire configuration for the CAPK-0100nd, which improperly uses an evaluation board's board number and breaks board detection at runtime. This board is sold by Portwell as the CAM-0100. o) Add support for the RTC available on some Octeon boards. o) Add support for the Octeon PCI bus. Note that rman_[sg]et_virtual for IO ports can not work unless building for n64. o) Clean up the CompactFlash driver to use Simple Executive macros and structures where possible (it would be advisable to use the Simple Executive API to set the PIO mode, too, but that is not done presently.) Also use structures from FreeBSD's ATA layer rather than structures copied from Linux. o) Print available Octeon SoC features on boot. o) Add support for the Octeon timecounter. o) Use the Simple Executive's routines rather than local copies for doing reads and writes to 64-bit addresses and use its macros for various device addresses rather than using local copies. o) Rename octeon_board_real to octeon_is_simulation to reduce differences with Cavium-provided code originally written for Linux. Also make it use the same simplified test that the Simple Executive and Linux both use rather than our complex one. o) Add support for the Octeon CIU, which is the main interrupt unit, as a bus to use normal interrupt allocation and setup routines. o) Use the Simple Executive's bootmem facility to allocate physical memory for the kernel, rather than assuming we know which addresses we can steal. NB: This may reduce the amount of RAM the kernel reports you as having if you are leaving large temporary allocations made by U-Boot allocated when starting FreeBSD. o) Add a port of the Cavium-provided Ethernet driver for Linux. This changes Ethernet interface naming from rgmxN to octeN. The new driver has vast improvements over the old one, both in performance and functionality, but does still have some features which have not been ported entirely and there may be unimplemented code that can be hit in everyday use. I will make every effort to correct those as they are reported. o) Support loading the kernel on non-contiguous cores. o) Add very conservative support for harvesting randomness from the Octeon random number device. o) Turn SMP on by default. o) Clean up the style of the Octeon kernel configurations a little and make them compile with -march=octeon. o) Add support for the Lanner MR320 and the CAPK-0100nd to the Simple Executive. o) Modify the Simple Executive to build on FreeBSD and to build without executive-config.h or cvmx-config.h. In the future we may want to revert part of these changes and supply executive-config.h and cvmx-config.h and access to the options contained in those files via kernel configuration files. o) Modify the Simple Executive USB routines to support getting and setting of the USB PID.
2010-07-20 19:25:11 +00:00
cfg.u64 = cvmx_read_csr(CVMX_MIO_BOOT_REG_CFGX(bus_region));
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
if (cfg.s.base == phys_base >> 16)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
{
o Clean up some ID printfs, and put under bootverbose o Remove redundant lookups of base address in cf_identify o Fix some indenting issues o Fix an identification bug that uses DRQ to checlk for ident block returned. The correct spec is to look for BSY to be cleared. Reviewed by: imp, marcel Obtained from: Juniper Networks, Inc Author: Andrew Duane
2011-10-04 18:03:55 +00:00
if (cvmx_sysinfo_get()->compact_flash_attribute_base_addr == 0)
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
bus_type = CF_TRUE_IDE_8;
else
bus_type = (cfg.s.width) ? CF_16 : CF_8;
printf("Compact flash found in bootbus region %d (%s).\n", bus_region, cf_type[bus_type]);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
break;
}
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
switch (bus_type)
{
case CF_8:
case CF_TRUE_IDE_8:
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* Check if CF is inserted */
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
while (cf_inb_8(TF_STATUS) & STATUS_BSY) {
if ((count++) == NR_TRIES ) {
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
printf("Compact Flash not present\n");
return;
}
DELAY(WAIT_DELAY);
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
case CF_16:
default:
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
/* Check if CF is inserted */
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
while (cf_inb_16(TF_STATUS) & STATUS_BSY) {
if ((count++) == NR_TRIES ) {
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
printf("Compact Flash not present\n");
return;
}
DELAY(WAIT_DELAY);
}
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
break;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
}
BUS_ADD_CHILD(parent, 0, "cf", 0);
}
/* ------------------------------------------------------------------- *
* cf_attach_geom() *
* ------------------------------------------------------------------- */
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
static void cf_attach_geom (void *arg, int flag)
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
{
struct cf_priv *cf_priv;
cf_priv = (struct cf_priv *) arg;
cf_priv->cf_geom = g_new_geomf(&g_cf_class, "cf%d", device_get_unit(cf_priv->dev));
cf_priv->cf_geom->softc = cf_priv;
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
cf_priv->cf_provider = g_new_providerf(cf_priv->cf_geom, cf_priv->cf_geom->name);
cf_priv->cf_provider->sectorsize = cf_priv->drive_param.sector_size;
cf_priv->cf_provider->mediasize = cf_priv->drive_param.nr_sectors * cf_priv->cf_provider->sectorsize;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
g_error_provider(cf_priv->cf_provider, 0);
}
/* ------------------------------------------------------------------- *
* cf_attach() *
* ------------------------------------------------------------------- */
static int cf_attach (device_t dev)
{
struct cf_priv *cf_priv;
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
int error;
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
Actually check board type rather than using a specialized octeon_is_simulation function.
2012-10-30 06:36:14 +00:00
if (cvmx_sysinfo_get()->board_type == CVMX_BOARD_TYPE_SIM)
Add support for True IDE mode to the Octeon CF driver. This mode is signalled when the attribute address for the CF is 0 in the octeon sysinfo structure. In this mode, the DATA port is 16-bits, but the other ports are 8-bits, but on a 16-bit bus (so you have to access it a short at a time, but only believe the lower byte). See the code for more details on this slightly odd arrangement. I'm still not 100% happy with the abstractions here on many levels (starting with the globals for these settings, on down to no bus_space use, etc), but the driver had these problems before the change. Also, clean up the code a bit to make this support easier, and the code a bit easier to read. I tried to follow existing style, but may have missed a few spots. Add some comments. Fix probe/attach routine to return a proper error for the simulator. With this change, my EBH5200 eval board now recognizes the CF well enough to boot to the login prompt. Before it would say it never became ready. My CN3010-EVB-HS5 still boots properly. My older CN3860-based board won't load the 64-bit kernel, either before or after the change, and I didn't chase that down.
2011-06-04 07:06:05 +00:00
return (ENXIO);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
cf_priv = device_get_softc(dev);
cf_priv->dev = dev;
o) Don't keep drive parameters in a global, put them in the softc. o) Don't report a bogus stripesize. o) Eliminate a nearby gratuitous "proxy" indirection.
2012-03-21 10:27:12 +00:00
error = cf_cmd_identify(cf_priv);
if (error != 0) {
device_printf(dev, "cf_cmd_identify failed: %d\n", error);
return (error);
}
g_post_event(cf_attach_geom, cf_priv, M_WAITOK, NULL);
Merge in Cavium's CF driver. This too is in the wrong place and will be moved.
2009-06-14 03:47:44 +00:00
bioq_init(&cf_priv->cf_bq);
return 0;
}
static device_method_t cf_methods[] = {
/* Device interface */
DEVMETHOD(device_probe, cf_probe),
DEVMETHOD(device_identify, cf_identify),
DEVMETHOD(device_attach, cf_attach),
DEVMETHOD(device_detach, bus_generic_detach),
DEVMETHOD(device_shutdown, bus_generic_shutdown),
{ 0, 0 }
};
static driver_t cf_driver = {
"cf",
cf_methods,
sizeof(struct cf_priv)
};
static devclass_t cf_devclass;
DRIVER_MODULE(cf, nexus, cf_driver, cf_devclass, 0, 0);
Reference in New Issue Copy Permalink