d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
609729b7a6
freebsd-skq/sys/dev/exca/exca.c

908 lines
25 KiB
C
Raw Normal View History

Start each of the license/copyright comments with /*-, minor shuffle of lines
2005-01-06 01:43:34 +00:00
/*-
Properly calculate the offset in mapping the memory of pccards. This allows my 3com cards to work again. It appears that this code was once there, but I removed it when I added the alignment issues. MFC After: 5 days PR: 70639 (and likely others)
2005-01-11 00:32:43 +00:00
* Copyright (c) 2002-2005 M Warner Losh. All rights reserved.
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*
* This software may be derived from NetBSD i82365.c and other files with
* the following copyright:
*
* Copyright (c) 1997 Marc Horowitz. All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by Marc Horowitz.
* 4. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
*/
Fix location of $FreeBSD$ from last commit.
2003-08-27 23:30:01 +00:00
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
#include <sys/param.h>
#include <sys/systm.h>
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
#include <sys/condvar.h>
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
#include <sys/errno.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/queue.h>
Make oldcard and newcard kernel module work.
2002-05-30 17:38:00 +00:00
#include <sys/module.h>
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
#include <sys/lock.h>
#include <sys/mutex.h>
Make oldcard and newcard kernel module work.
2002-05-30 17:38:00 +00:00
#include <sys/conf.h>
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
#include <sys/bus.h>
#include <machine/bus.h>
#include <sys/rman.h>
#include <machine/resource.h>
#include <dev/pccard/pccardreg.h>
#include <dev/pccard/pccardvar.h>
#include <dev/exca/excareg.h>
#include <dev/exca/excavar.h>
#ifdef EXCA_DEBUG
#define DEVPRINTF(dev, fmt, args...) device_printf((dev), (fmt), ## args)
#define DPRINTF(fmt, args...) printf(fmt, ## args)
#else
#define DEVPRINTF(dev, fmt, args...)
#define DPRINTF(fmt, args...)
#endif
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
static const char *chip_names[] =
{
"CardBus socket",
"Intel i82365SL-A/B or clone",
"Intel i82365sl-DF step",
"VLSI chip",
"Cirrus Logic PD6710",
"Cirrus logic PD6722",
"Cirrus Logic PD6729",
"Vadem 365",
"Vadem 465",
"Vadem 468",
"Vadem 469",
"Ricoh RF5C296",
"Ricoh RF5C396",
"IBM clone",
"IBM KING PCMCIA Controller"
};
static exca_getb_fn exca_mem_getb;
static exca_putb_fn exca_mem_putb;
static exca_getb_fn exca_io_getb;
static exca_putb_fn exca_io_putb;
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
/* memory */
#define EXCA_MEMINFO(NUM) { \
EXCA_SYSMEM_ADDR ## NUM ## _START_LSB, \
EXCA_SYSMEM_ADDR ## NUM ## _START_MSB, \
EXCA_SYSMEM_ADDR ## NUM ## _STOP_LSB, \
EXCA_SYSMEM_ADDR ## NUM ## _STOP_MSB, \
EXCA_SYSMEM_ADDR ## NUM ## _WIN, \
EXCA_CARDMEM_ADDR ## NUM ## _LSB, \
EXCA_CARDMEM_ADDR ## NUM ## _MSB, \
Fix a couple of places where preprocessor concatination was misused and something wound up grafted to a comma.
2002-06-01 16:21:26 +00:00
EXCA_ADDRWIN_ENABLE_MEM ## NUM, \
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
}
static struct mem_map_index_st {
int sysmem_start_lsb;
int sysmem_start_msb;
int sysmem_stop_lsb;
int sysmem_stop_msb;
int sysmem_win;
int cardmem_lsb;
int cardmem_msb;
int memenable;
} mem_map_index[] = {
EXCA_MEMINFO(0),
EXCA_MEMINFO(1),
EXCA_MEMINFO(2),
EXCA_MEMINFO(3),
EXCA_MEMINFO(4)
};
#undef EXCA_MEMINFO
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
static uint8_t
exca_mem_getb(struct exca_softc *sc, int reg)
{
return (bus_space_read_1(sc->bst, sc->bsh, sc->offset + reg));
}
static void
exca_mem_putb(struct exca_softc *sc, int reg, uint8_t val)
{
Don't use return(foo(...)); in function returning void. Found by: FlexeLint
2003-05-31 16:51:41 +00:00
bus_space_write_1(sc->bst, sc->bsh, sc->offset + reg, val);
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
}
static uint8_t
exca_io_getb(struct exca_softc *sc, int reg)
{
bus_space_write_1(sc->bst, sc->bsh, EXCA_REG_INDEX, reg + sc->offset);
return (bus_space_read_1(sc->bst, sc->bsh, EXCA_REG_DATA));
}
static void
exca_io_putb(struct exca_softc *sc, int reg, uint8_t val)
{
bus_space_write_1(sc->bst, sc->bsh, EXCA_REG_INDEX, reg + sc->offset);
bus_space_write_1(sc->bst, sc->bsh, EXCA_REG_DATA, val);
}
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
/*
* Helper function. This will map the requested memory slot. We setup the
* map before we call this function. This is used to initially force the
* mapping, as well as later restore the mapping after it has been destroyed
* in some fashion (due to a power event typically).
*/
static void
exca_do_mem_map(struct exca_softc *sc, int win)
{
struct mem_map_index_st *map;
struct pccard_mem_handle *mem;
Properly calculate the offset in mapping the memory of pccards. This allows my 3com cards to work again. It appears that this code was once there, but I removed it when I added the alignment issues. MFC After: 5 days PR: 70639 (and likely others)
2005-01-11 00:32:43 +00:00
uint32_t offset;
Cleanup sloppy ifdef.
2006-09-24 00:26:33 +00:00
int mem8 = 1 /* mem->kind == PCCARD_A_MEM_ATTR */;
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
map = &mem_map_index[win];
mem = &sc->mem[win];
Properly calculate the offset in mapping the memory of pccards. This allows my 3com cards to work again. It appears that this code was once there, but I removed it when I added the alignment issues. MFC After: 5 days PR: 70639 (and likely others)
2005-01-11 00:32:43 +00:00
offset = ((mem->cardaddr >> EXCA_CARDMEM_ADDRX_SHIFT) -
(mem->addr >> EXCA_SYSMEM_ADDRX_SHIFT)) & 0x3fff;
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, map->sysmem_start_lsb,
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
(mem->addr >> EXCA_SYSMEM_ADDRX_SHIFT) & 0xff);
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, map->sysmem_start_msb,
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
((mem->addr >> (EXCA_SYSMEM_ADDRX_SHIFT + 8)) &
MFP4: Minor tweaks.
2005-10-14 04:48:11 +00:00
EXCA_SYSMEM_ADDRX_START_MSB_ADDR_MASK) |
(mem8 ? 0 : EXCA_SYSMEM_ADDRX_START_MSB_DATASIZE_16BIT));
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, map->sysmem_stop_lsb,
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
((mem->addr + mem->realsize - 1) >>
EXCA_SYSMEM_ADDRX_SHIFT) & 0xff);
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, map->sysmem_stop_msb,
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
(((mem->addr + mem->realsize - 1) >>
(EXCA_SYSMEM_ADDRX_SHIFT + 8)) &
Other part of the cbb isa/pci split. This likely is causing problems wrt cbb module.
2004-08-15 23:17:54 +00:00
EXCA_SYSMEM_ADDRX_STOP_MSB_ADDR_MASK) |
EXCA_SYSMEM_ADDRX_STOP_MSB_WAIT2);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, map->sysmem_win,
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
(mem->addr >> EXCA_MEMREG_WIN_SHIFT) & 0xff);
Properly calculate the offset in mapping the memory of pccards. This allows my 3com cards to work again. It appears that this code was once there, but I removed it when I added the alignment issues. MFC After: 5 days PR: 70639 (and likely others)
2005-01-11 00:32:43 +00:00
exca_putb(sc, map->cardmem_lsb, offset & 0xff);
exca_putb(sc, map->cardmem_msb, (((offset >> 8) & 0xff) &
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
EXCA_CARDMEM_ADDRX_MSB_ADDR_MASK) |
When we went to the set_flags interface for the memory resource, we switched from PCCARD_MEM_FOO to PCCARD_A_MEM_FOO, yet we didn't change exca in all the right places. Do so now. Also use PCCARD_WIDTH_AUTO rather than the magic cookie 0.
2003-08-29 05:27:36 +00:00
((mem->kind == PCCARD_A_MEM_ATTR) ?
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
EXCA_CARDMEM_ADDRX_MSB_REGACTIVE_ATTR : 0));
A number of minor fixes to the mapping of memory. This should help reading the CIS on some cards. However, not all just yet. This makes at least some of the xircom cards that weren't working to work. It doesn't make my home and away card work, however. o Don't get the card offset wrong. This is the biggest hassle for reading the CIS. The old code was just so wrong I can't believe that it worked at all. o Don't set the bit that allows/forces 16-bit memory access to the memory. It is hard coded with 0x80. o Don't need to slow down memory access with wait-states. OLDCARD didn't need them and it doesn't hurt anything. o remove bogus grousying in comment.
2003-08-28 05:01:11 +00:00
#ifdef EXCA_DEBUG
When we went to the set_flags interface for the memory resource, we switched from PCCARD_MEM_FOO to PCCARD_A_MEM_FOO, yet we didn't change exca in all the right places. Do so now. Also use PCCARD_WIDTH_AUTO rather than the magic cookie 0.
2003-08-29 05:27:36 +00:00
if (mem->kind == PCCARD_A_MEM_ATTR)
A number of minor fixes to the mapping of memory. This should help reading the CIS on some cards. However, not all just yet. This makes at least some of the xircom cards that weren't working to work. It doesn't make my home and away card work, however. o Don't get the card offset wrong. This is the biggest hassle for reading the CIS. The old code was just so wrong I can't believe that it worked at all. o Don't set the bit that allows/forces 16-bit memory access to the memory. It is hard coded with 0x80. o Don't need to slow down memory access with wait-states. OLDCARD didn't need them and it doesn't hurt anything. o remove bogus grousying in comment.
2003-08-28 05:01:11 +00:00
printf("attribtue memory\n");
else
printf("common memory\n");
#endif
Other part of the cbb isa/pci split. This likely is causing problems wrt cbb module.
2004-08-15 23:17:54 +00:00
exca_setb(sc, EXCA_ADDRWIN_ENABLE, map->memenable |
EXCA_ADDRWIN_ENABLE_MEMCS16);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
DELAY(100);
#ifdef EXCA_DEBUG
{
int r1, r2, r3, r4, r5, r6, r7;
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
r1 = exca_getb(sc, map->sysmem_start_msb);
r2 = exca_getb(sc, map->sysmem_start_lsb);
r3 = exca_getb(sc, map->sysmem_stop_msb);
r4 = exca_getb(sc, map->sysmem_stop_lsb);
r5 = exca_getb(sc, map->cardmem_msb);
r6 = exca_getb(sc, map->cardmem_lsb);
r7 = exca_getb(sc, map->sysmem_win);
Other part of the cbb isa/pci split. This likely is causing problems wrt cbb module.
2004-08-15 23:17:54 +00:00
printf("exca_do_mem_map win %d: %02x%02x %02x%02x "
"%02x%02x %02x (%08x+%06x.%06x*%06x)\n",
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
win, r1, r2, r3, r4, r5, r6, r7,
mem->addr, mem->size, mem->realsize,
A number of minor fixes to the mapping of memory. This should help reading the CIS on some cards. However, not all just yet. This makes at least some of the xircom cards that weren't working to work. It doesn't make my home and away card work, however. o Don't get the card offset wrong. This is the biggest hassle for reading the CIS. The old code was just so wrong I can't believe that it worked at all. o Don't set the bit that allows/forces 16-bit memory access to the memory. It is hard coded with 0x80. o Don't need to slow down memory access with wait-states. OLDCARD didn't need them and it doesn't hurt anything. o remove bogus grousying in comment.
2003-08-28 05:01:11 +00:00
mem->cardaddr);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
}
#endif
}
/*
* public interface to map a resource. kind is the type of memory to
* map (either common or attribute). Memory created via this interface
* starts out at card address 0. Since the only way to set this is
* to set it on a struct resource after it has been mapped, we're safe
* in maping this assumption. Note that resources can be remapped using
* exca_do_mem_map so that's how the card address can be set later.
*/
int
exca_mem_map(struct exca_softc *sc, int kind, struct resource *res)
{
int win;
for (win = 0; win < EXCA_MEM_WINS; win++) {
if ((sc->memalloc & (1 << win)) == 0) {
sc->memalloc |= (1 << win);
break;
}
}
if (win >= EXCA_MEM_WINS)
- Fix exca_(io|mem)_map() to return proper errno values. - Change exca_activate_resource() to call BUS_ACTIVATE_RESOURCE() before calling exca_(io|mem)_map() since the latter use rman_get_bus(tag|handle) and the recent changes to nexus(4) mean that you need to activate a resource before reading the bus tag and handle. This was true before, but now the nexus(4) drivers on x86 and ia64 are more forceful about it. Reviewed by: imp
2007-03-23 17:15:07 +00:00
return (ENOSPC);
CARDMEM is the offset of the address ON THE CARD (eg which page). Instead, use EXCA_MEMREG_WIN_SHIFT which is the amount we shift the bus address by to write into upper memory (eg above 24MB). Use the latter in this case.
2003-09-24 22:13:25 +00:00
if (((rman_get_start(res) >> EXCA_MEMREG_WIN_SHIFT) & 0xff) != 0 &&
MFp4: o Protect .h against multiple includes. o eliminate the pointers to the read/write routines. The bus_space_read routines can cope since we have the offset field. o Print a warning if the requested map address is > 16M and your chipset doesn't support the extended ExCA registers.
2002-07-26 08:01:08 +00:00
(sc->flags & EXCA_HAS_MEMREG_WIN) == 0) {
device_printf(sc->dev, "Does not support mapping above 24M.");
- Fix exca_(io|mem)_map() to return proper errno values. - Change exca_activate_resource() to call BUS_ACTIVATE_RESOURCE() before calling exca_(io|mem)_map() since the latter use rman_get_bus(tag|handle) and the recent changes to nexus(4) mean that you need to activate a resource before reading the bus tag and handle. This was true before, but now the nexus(4) drivers on x86 and ia64 are more forceful about it. Reviewed by: imp
2007-03-23 17:15:07 +00:00
return (EINVAL);
MFp4: o Protect .h against multiple includes. o eliminate the pointers to the read/write routines. The bus_space_read routines can cope since we have the offset field. o Print a warning if the requested map address is > 16M and your chipset doesn't support the extended ExCA registers.
2002-07-26 08:01:08 +00:00
}
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
sc->mem[win].cardaddr = 0;
sc->mem[win].memt = rman_get_bustag(res);
sc->mem[win].memh = rman_get_bushandle(res);
sc->mem[win].addr = rman_get_start(res);
sc->mem[win].size = rman_get_end(res) - sc->mem[win].addr + 1;
sc->mem[win].realsize = sc->mem[win].size + EXCA_MEM_PAGESIZE - 1;
sc->mem[win].realsize = sc->mem[win].realsize -
(sc->mem[win].realsize % EXCA_MEM_PAGESIZE);
sc->mem[win].kind = kind;
A number of minor fixes to the mapping of memory. This should help reading the CIS on some cards. However, not all just yet. This makes at least some of the xircom cards that weren't working to work. It doesn't make my home and away card work, however. o Don't get the card offset wrong. This is the biggest hassle for reading the CIS. The old code was just so wrong I can't believe that it worked at all. o Don't set the bit that allows/forces 16-bit memory access to the memory. It is hard coded with 0x80. o Don't need to slow down memory access with wait-states. OLDCARD didn't need them and it doesn't hurt anything. o remove bogus grousying in comment.
2003-08-28 05:01:11 +00:00
DPRINTF("exca_mem_map window %d bus %x+%x card addr %x\n",
win, sc->mem[win].addr, sc->mem[win].size, sc->mem[win].cardaddr);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
exca_do_mem_map(sc, win);
return (0);
}
/*
* Private helper function. This turns off a given memory map that is in
* use. We do this by just clearing the enable bit in the pcic. If we needed
* to make memory unmapping/mapping pairs faster, we would have to store
* more state information about the pcic and then use that to intelligently
* to the map/unmap. However, since we don't do that sort of thing often
* (generally just at configure time), it isn't a case worth optimizing.
*/
static void
exca_mem_unmap(struct exca_softc *sc, int window)
{
if (window < 0 || window >= EXCA_MEM_WINS)
panic("exca_mem_unmap: window out of range");
exca_clrb(sc, EXCA_ADDRWIN_ENABLE, mem_map_index[window].memenable);
sc->memalloc &= ~(1 << window);
}
/*
- Fix exca_(io|mem)_map() to return proper errno values. - Change exca_activate_resource() to call BUS_ACTIVATE_RESOURCE() before calling exca_(io|mem)_map() since the latter use rman_get_bus(tag|handle) and the recent changes to nexus(4) mean that you need to activate a resource before reading the bus tag and handle. This was true before, but now the nexus(4) drivers on x86 and ia64 are more forceful about it. Reviewed by: imp
2007-03-23 17:15:07 +00:00
* Find the map that we're using to hold the resource. This works well
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
* so long as the client drivers don't do silly things like map the same
* area mutliple times, or map both common and attribute memory at the
* same time. This latter restriction is a bug. We likely should just
* store a pointer to the res in the mem[x] data structure.
*/
static int
exca_mem_findmap(struct exca_softc *sc, struct resource *res)
{
int win;
for (win = 0; win < EXCA_MEM_WINS; win++) {
if (sc->mem[win].memt == rman_get_bustag(res) &&
sc->mem[win].addr == rman_get_start(res) &&
sc->mem[win].size == rman_get_size(res))
return (win);
}
return (-1);
}
/*
* Set the memory flag. This means that we are setting if the memory
* is coming from attribute memory or from common memory on the card.
* CIS entries are generally in attribute memory (although they can
* reside in common memory). Generally, this is the only use for attribute
* memory. However, some cards require their drivers to dance in both
* common and/or attribute memory and this interface (and setting the
* offset interface) exist for such cards.
*/
int
exca_mem_set_flags(struct exca_softc *sc, struct resource *res, uint32_t flags)
{
int win;
win = exca_mem_findmap(sc, res);
if (win < 0) {
device_printf(sc->dev,
"set_res_flags: specified resource not active\n");
return (ENOENT);
}
sc->mem[win].kind = flags;
exca_do_mem_map(sc, win);
return (0);
}
/*
* Given a resource, go ahead and unmap it if we can find it in the
* resrouce list that's used.
*/
int
exca_mem_unmap_res(struct exca_softc *sc, struct resource *res)
{
int win;
win = exca_mem_findmap(sc, res);
if (win < 0)
return (ENOENT);
exca_mem_unmap(sc, win);
return (0);
}
/*
* Set the offset of the memory. We use this for reading the CIS and
Minor commentary correction
2005-10-17 08:20:44 +00:00
* frobbing the pccard's pccard registers (CCR, etc). Some drivers
* need to access arbitrary attribute and common memory during their
* initialization and operation.
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
*/
int
exca_mem_set_offset(struct exca_softc *sc, struct resource *res,
uint32_t cardaddr, uint32_t *deltap)
{
int win;
uint32_t delta;
win = exca_mem_findmap(sc, res);
if (win < 0) {
device_printf(sc->dev,
"set_memory_offset: specified resource not active\n");
return (ENOENT);
}
A number of minor fixes to the mapping of memory. This should help reading the CIS on some cards. However, not all just yet. This makes at least some of the xircom cards that weren't working to work. It doesn't make my home and away card work, however. o Don't get the card offset wrong. This is the biggest hassle for reading the CIS. The old code was just so wrong I can't believe that it worked at all. o Don't set the bit that allows/forces 16-bit memory access to the memory. It is hard coded with 0x80. o Don't need to slow down memory access with wait-states. OLDCARD didn't need them and it doesn't hurt anything. o remove bogus grousying in comment.
2003-08-28 05:01:11 +00:00
sc->mem[win].cardaddr = cardaddr & ~(EXCA_MEM_PAGESIZE - 1);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
delta = cardaddr % EXCA_MEM_PAGESIZE;
if (deltap)
*deltap = delta;
sc->mem[win].realsize = sc->mem[win].size + delta +
EXCA_MEM_PAGESIZE - 1;
sc->mem[win].realsize = sc->mem[win].realsize -
(sc->mem[win].realsize % EXCA_MEM_PAGESIZE);
exca_do_mem_map(sc, win);
return (0);
}
/* I/O */
#define EXCA_IOINFO(NUM) { \
EXCA_IOADDR ## NUM ## _START_LSB, \
EXCA_IOADDR ## NUM ## _START_MSB, \
EXCA_IOADDR ## NUM ## _STOP_LSB, \
EXCA_IOADDR ## NUM ## _STOP_MSB, \
Fix a couple of places where preprocessor concatination was misused and something wound up grafted to a comma.
2002-06-01 16:21:26 +00:00
EXCA_ADDRWIN_ENABLE_IO ## NUM, \
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
EXCA_IOCTL_IO ## NUM ## _WAITSTATE \
| EXCA_IOCTL_IO ## NUM ## _ZEROWAIT \
| EXCA_IOCTL_IO ## NUM ## _IOCS16SRC_MASK \
| EXCA_IOCTL_IO ## NUM ## _DATASIZE_MASK, \
{ \
EXCA_IOCTL_IO ## NUM ## _IOCS16SRC_CARD, \
EXCA_IOCTL_IO ## NUM ## _IOCS16SRC_DATASIZE \
| EXCA_IOCTL_IO ## NUM ## _DATASIZE_8BIT, \
EXCA_IOCTL_IO ## NUM ## _IOCS16SRC_DATASIZE \
| EXCA_IOCTL_IO ## NUM ## _DATASIZE_16BIT, \
} \
}
static struct io_map_index_st {
int start_lsb;
int start_msb;
int stop_lsb;
int stop_msb;
int ioenable;
int ioctlmask;
int ioctlbits[3]; /* indexed by PCCARD_WIDTH_* */
} io_map_index[] = {
EXCA_IOINFO(0),
EXCA_IOINFO(1),
};
#undef EXCA_IOINFO
static void
exca_do_io_map(struct exca_softc *sc, int win)
{
struct io_map_index_st *map;
struct pccard_io_handle *io;
map = &io_map_index[win];
io = &sc->io[win];
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, map->start_lsb, io->addr & 0xff);
exca_putb(sc, map->start_msb, (io->addr >> 8) & 0xff);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, map->stop_lsb, (io->addr + io->size - 1) & 0xff);
exca_putb(sc, map->stop_msb, ((io->addr + io->size - 1) >> 8) & 0xff);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
exca_clrb(sc, EXCA_IOCTL, map->ioctlmask);
exca_setb(sc, EXCA_IOCTL, map->ioctlbits[io->width]);
exca_setb(sc, EXCA_ADDRWIN_ENABLE, map->ioenable);
#ifdef EXCA_DEBUG
{
int r1, r2, r3, r4;
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
r1 = exca_getb(sc, map->start_msb);
r2 = exca_getb(sc, map->start_lsb);
r3 = exca_getb(sc, map->stop_msb);
r4 = exca_getb(sc, map->stop_lsb);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
DPRINTF("exca_do_io_map window %d: %02x%02x %02x%02x "
"(%08x+%08x)\n", win, r1, r2, r3, r4,
io->addr, io->size);
}
#endif
}
int
exca_io_map(struct exca_softc *sc, int width, struct resource *r)
{
int win;
#ifdef EXCA_DEBUG
static char *width_names[] = { "auto", "io8", "io16"};
#endif
for (win=0; win < EXCA_IO_WINS; win++) {
if ((sc->ioalloc & (1 << win)) == 0) {
sc->ioalloc |= (1 << win);
break;
}
}
if (win >= EXCA_IO_WINS)
- Fix exca_(io|mem)_map() to return proper errno values. - Change exca_activate_resource() to call BUS_ACTIVATE_RESOURCE() before calling exca_(io|mem)_map() since the latter use rman_get_bus(tag|handle) and the recent changes to nexus(4) mean that you need to activate a resource before reading the bus tag and handle. This was true before, but now the nexus(4) drivers on x86 and ia64 are more forceful about it. Reviewed by: imp
2007-03-23 17:15:07 +00:00
return (ENOSPC);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
sc->io[win].iot = rman_get_bustag(r);
sc->io[win].ioh = rman_get_bushandle(r);
sc->io[win].addr = rman_get_start(r);
sc->io[win].size = rman_get_end(r) - sc->io[win].addr + 1;
sc->io[win].flags = 0;
sc->io[win].width = width;
DPRINTF("exca_io_map window %d %s port %x+%x\n",
win, width_names[width], sc->io[win].addr,
sc->io[win].size);
exca_do_io_map(sc, win);
return (0);
}
static void
exca_io_unmap(struct exca_softc *sc, int window)
{
if (window >= EXCA_IO_WINS)
panic("exca_io_unmap: window out of range");
exca_clrb(sc, EXCA_ADDRWIN_ENABLE, io_map_index[window].ioenable);
sc->ioalloc &= ~(1 << window);
sc->io[window].iot = 0;
sc->io[window].ioh = 0;
sc->io[window].addr = 0;
sc->io[window].size = 0;
sc->io[window].flags = 0;
sc->io[window].width = 0;
}
static int
exca_io_findmap(struct exca_softc *sc, struct resource *res)
{
int win;
for (win = 0; win < EXCA_IO_WINS; win++) {
if (sc->io[win].iot == rman_get_bustag(res) &&
sc->io[win].addr == rman_get_start(res) &&
sc->io[win].size == rman_get_size(res))
return (win);
}
return (-1);
}
int
exca_io_unmap_res(struct exca_softc *sc, struct resource *res)
{
int win;
win = exca_io_findmap(sc, res);
if (win < 0)
return (ENOENT);
exca_io_unmap(sc, win);
return (0);
}
/* Misc */
/*
* If interrupts are enabled, then we should be able to just wait for
* an interrupt routine to wake us up. Busy waiting shouldn't be
* necessary. Sadly, not all legacy ISA cards support an interrupt
* for the busy state transitions, at least according to their datasheets,
* so we busy wait a while here..
*/
static void
exca_wait_ready(struct exca_softc *sc)
{
int i;
DEVPRINTF(sc->dev, "exca_wait_ready: status 0x%02x\n",
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_getb(sc, EXCA_IF_STATUS));
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
for (i = 0; i < 10000; i++) {
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
if (exca_getb(sc, EXCA_IF_STATUS) & EXCA_IF_STATUS_READY)
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
return;
DELAY(500);
}
device_printf(sc->dev, "ready never happened, status = %02x\n",
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_getb(sc, EXCA_IF_STATUS));
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
}
/*
* Reset the card. Ideally, we'd do a lot of this via interrupts.
* However, many PC Cards will deassert the ready signal. This means
* that they are asserting an interrupt. This makes it hard to
* do anything but a busy wait here. One could argue that these
* such cards are broken, or that the bridge that allows this sort
* of interrupt through isn't quite what you'd want (and may be a standards
'PC Card' instead of other variants
2005-09-22 06:01:44 +00:00
* violation). However, such arguing would leave a huge class of PC Cards
Update comments to remind me about power bugs in oldcard.
2002-10-07 06:18:50 +00:00
* and bridges out of reach for use in the system.
*
* Maybe I should reevaluate the above based on the power bug I fixed
* in OLDCARD.
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
*/
void
exca_reset(struct exca_softc *sc, device_t child)
{
int win;
/* enable socket i/o */
exca_setb(sc, EXCA_PWRCTL, EXCA_PWRCTL_OE);
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, EXCA_INTR, EXCA_INTR_ENABLE);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
/* hold reset for 30ms */
DELAY(30*1000);
/* clear the reset flag */
exca_setb(sc, EXCA_INTR, EXCA_INTR_RESET);
'PC Card' instead of other variants
2005-09-22 06:01:44 +00:00
/* wait 20ms as per PC Card standard (r2.01) section 4.3.6 */
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
DELAY(20*1000);
exca_wait_ready(sc);
/* disable all address windows */
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_putb(sc, EXCA_ADDRWIN_ENABLE, 0);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
Card type stuff was backwards, so assume io for the moment. mem needs a lot of help and with luck bms will help me fix that.
2003-10-07 04:29:04 +00:00
exca_setb(sc, EXCA_INTR, EXCA_INTR_CARDTYPE_IO);
DEVPRINTF(sc->dev, "card type is io\n");
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
/* reinstall all the memory and io mappings */
for (win = 0; win < EXCA_MEM_WINS; ++win)
if (sc->memalloc & (1 << win))
exca_do_mem_map(sc, win);
for (win = 0; win < EXCA_IO_WINS; ++win)
if (sc->ioalloc & (1 << win))
exca_do_io_map(sc, win);
}
/*
* Initialize the exca_softc data structure for the first time.
*/
void
MFp4: o Protect .h against multiple includes. o eliminate the pointers to the read/write routines. The bus_space_read routines can cope since we have the offset field. o Print a warning if the requested map address is > 16M and your chipset doesn't support the extended ExCA registers.
2002-07-26 08:01:08 +00:00
exca_init(struct exca_softc *sc, device_t dev,
bus_space_tag_t bst, bus_space_handle_t bsh, uint32_t offset)
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
{
sc->dev = dev;
sc->memalloc = 0;
sc->ioalloc = 0;
sc->bst = bst;
sc->bsh = bsh;
sc->offset = offset;
sc->flags = 0;
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
sc->getb = exca_mem_getb;
sc->putb = exca_mem_putb;
}
/*
* Is this socket valid?
*/
static int
exca_valid_slot(struct exca_softc *exca)
{
uint8_t c;
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
/* Assume the worst */
exca->chipset = EXCA_BOGUS;
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
/*
* see if there's a PCMCIA controller here
* Intel PCMCIA controllers use 0x82 and 0x83
* IBM clone chips use 0x88 and 0x89, apparently
*/
c = exca_getb(exca, EXCA_IDENT);
MFP4: Minor tweaks.
2005-10-14 04:48:11 +00:00
DEVPRINTF(exca->dev, "Ident is %x\n", c);
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
if ((c & EXCA_IDENT_IFTYPE_MASK) != EXCA_IDENT_IFTYPE_MEM_AND_IO)
return (0);
if ((c & EXCA_IDENT_ZERO) != 0)
return (0);
switch (c & EXCA_IDENT_REV_MASK) {
/*
* 82365 or clones.
*/
case EXCA_IDENT_REV_I82365SLR0:
case EXCA_IDENT_REV_I82365SLR1:
exca->chipset = EXCA_I82365;
/*
* Check for Vadem chips by unlocking their extra
* registers and looking for valid ID. Bit 3 in
* the ID register is normally 0, except when
* EXCA_VADEMREV is set. Other bridges appear
* to ignore this frobbing.
*/
bus_space_write_1(exca->bst, exca->bsh, EXCA_REG_INDEX,
EXCA_VADEM_COOKIE1);
bus_space_write_1(exca->bst, exca->bsh, EXCA_REG_INDEX,
EXCA_VADEM_COOKIE2);
exca_setb(exca, EXCA_VADEM_VMISC, EXCA_VADEM_REV);
c = exca_getb(exca, EXCA_IDENT);
if (c & 0x08) {
switch (c & 7) {
case 1:
exca->chipset = EXCA_VG365;
break;
case 2:
exca->chipset = EXCA_VG465;
break;
case 3:
exca->chipset = EXCA_VG468;
break;
default:
exca->chipset = EXCA_VG469;
break;
}
exca_clrb(exca, EXCA_VADEM_VMISC, EXCA_VADEM_REV);
break;
}
/*
* Check for RICOH RF5C[23]96 PCMCIA Controller
*/
c = exca_getb(exca, EXCA_RICOH_ID);
if (c == EXCA_RID_396) {
exca->chipset = EXCA_RF5C396;
break;
} else if (c == EXCA_RID_296) {
exca->chipset = EXCA_RF5C296;
break;
}
/*
* Check for Cirrus logic chips.
*/
exca_putb(exca, EXCA_CIRRUS_CHIP_INFO, 0);
c = exca_getb(exca, EXCA_CIRRUS_CHIP_INFO);
if ((c & EXCA_CIRRUS_CHIP_INFO_CHIP_ID) ==
EXCA_CIRRUS_CHIP_INFO_CHIP_ID) {
c = exca_getb(exca, EXCA_CIRRUS_CHIP_INFO);
if ((c & EXCA_CIRRUS_CHIP_INFO_CHIP_ID) == 0) {
if (c & EXCA_CIRRUS_CHIP_INFO_SLOTS)
exca->chipset = EXCA_PD6722;
else
exca->chipset = EXCA_PD6710;
break;
}
}
break;
case EXCA_IDENT_REV_I82365SLDF:
/*
* Intel i82365sl-DF step or maybe a vlsi 82c146
* we detected the vlsi case earlier, so if the controller
* isn't set, we know it is a i82365sl step D.
*/
exca->chipset = EXCA_I82365SL_DF;
break;
case EXCA_IDENT_REV_IBM1:
case EXCA_IDENT_REV_IBM2:
exca->chipset = EXCA_IBM;
break;
case EXCA_IDENT_REV_IBM_KING:
exca->chipset = EXCA_IBM_KING;
break;
default:
return (0);
}
return (1);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
}
/*
* Probe the expected slots. We maybe should set the ID for each of these
* slots too while we're at it. But maybe that belongs to a separate
* function.
*
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
* The caller must guarantee that at least EXCA_NSLOTS are present in exca.
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
*/
int
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca_probe_slots(device_t dev, struct exca_softc *exca, bus_space_tag_t iot,
bus_space_handle_t ioh)
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
{
int err;
int i;
err = ENXIO;
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
for (i = 0; i < EXCA_NSLOTS; i++) {
MFp4: o Protect .h against multiple includes. o eliminate the pointers to the read/write routines. The bus_space_read routines can cope since we have the offset field. o Print a warning if the requested map address is > 16M and your chipset doesn't support the extended ExCA registers.
2002-07-26 08:01:08 +00:00
exca_init(&exca[i], dev, iot, ioh, i * EXCA_SOCKET_SIZE);
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
exca->getb = exca_io_getb;
exca->putb = exca_io_putb;
Set the name of the chipset found
2005-10-19 00:25:39 +00:00
if (exca_valid_slot(&exca[i])) {
device_set_desc(dev, chip_names[exca[i].chipset]);
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
err = 0;
Set the name of the chipset found
2005-10-19 00:25:39 +00:00
}
First draft of common code between pccbb and pcic drivers for talking to ExCA register sets. These registers exist in both ISA and PCI devices in a couple different ways, and this will provide a common base for future building. This code is a rehash of the pccbb 16-bit code, which was a rehash of the pcic code, which was a rehash of the netbsd i82365 code. More hashing to come.
2002-01-29 06:48:38 +00:00
}
return (err);
}
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
void
exca_insert(struct exca_softc *exca)
{
exca->pccarddev should always be non-null now. Only call it when the device is actually attached.
2007-02-24 15:56:06 +00:00
if (device_is_attached(exca->pccarddev)) {
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
if (CARD_ATTACH_CARD(exca->pccarddev) != 0)
device_printf(exca->dev,
"PC Card card activation failed\n");
} else {
device_printf(exca->dev,
"PC Card inserted, but no pccard bus.\n");
}
}
void
exca_removal(struct exca_softc *exca)
{
It turns out that it is easier to not NULL out pccard and cardbus device pointers. They don't change as the children device drivers come and go. Rather, check to see if the device is attached where we would have checked ! NULL. This solves many asymmetries in the code that likely could lead to crashes when loading/unloading cbb without one or more of the expected children's driver not present.
2007-02-15 07:22:27 +00:00
if (device_is_attached(exca->pccarddev))
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
CARD_DETACH_CARD(exca->pccarddev);
}
int
exca_activate_resource(struct exca_softc *exca, device_t child, int type,
int rid, struct resource *res)
{
int err;
- Fix exca_(io|mem)_map() to return proper errno values. - Change exca_activate_resource() to call BUS_ACTIVATE_RESOURCE() before calling exca_(io|mem)_map() since the latter use rman_get_bus(tag|handle) and the recent changes to nexus(4) mean that you need to activate a resource before reading the bus tag and handle. This was true before, but now the nexus(4) drivers on x86 and ia64 are more forceful about it. Reviewed by: imp
2007-03-23 17:15:07 +00:00
if (rman_get_flags(res) & RF_ACTIVE)
return (0);
err = BUS_ACTIVATE_RESOURCE(device_get_parent(exca->dev), child,
type, rid, res);
if (err)
return (err);
switch (type) {
case SYS_RES_IOPORT:
err = exca_io_map(exca, PCCARD_WIDTH_AUTO, res);
break;
case SYS_RES_MEMORY:
err = exca_mem_map(exca, PCCARD_A_MEM_COM, res);
break;
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
}
- Fix exca_(io|mem)_map() to return proper errno values. - Change exca_activate_resource() to call BUS_ACTIVATE_RESOURCE() before calling exca_(io|mem)_map() since the latter use rman_get_bus(tag|handle) and the recent changes to nexus(4) mean that you need to activate a resource before reading the bus tag and handle. This was true before, but now the nexus(4) drivers on x86 and ia64 are more forceful about it. Reviewed by: imp
2007-03-23 17:15:07 +00:00
if (err)
BUS_DEACTIVATE_RESOURCE(device_get_parent(exca->dev), child,
type, rid, res);
return (err);
MFP4: Merge some common code from cbb into exca.
2003-06-07 20:46:39 +00:00
}
int
exca_deactivate_resource(struct exca_softc *exca, device_t child, int type,
int rid, struct resource *res)
{
if (rman_get_flags(res) & RF_ACTIVE) { /* if activated */
switch (type) {
case SYS_RES_IOPORT:
if (exca_io_unmap_res(exca, res))
return (ENOENT);
break;
case SYS_RES_MEMORY:
if (exca_mem_unmap_res(exca, res))
return (ENOENT);
break;
}
}
return (BUS_DEACTIVATE_RESOURCE(device_get_parent(exca->dev), child,
type, rid, res));
}
Other part of the cbb isa/pci split. This likely is causing problems wrt cbb module.
2004-08-15 23:17:54 +00:00
#if 0
static struct resource *
exca_alloc_resource(struct exca_softc *sc, device_t child, int type, int *rid,
u_long start, u_long end, u_long count, uint flags)
{
struct resource *res = NULL;
int tmp;
switch (type) {
case SYS_RES_MEMORY:
if (start < cbb_start_mem)
start = cbb_start_mem;
if (end < start)
end = start;
flags = (flags & ~RF_ALIGNMENT_MASK) |
rman_make_alignment_flags(CBB_MEMALIGN);
break;
case SYS_RES_IOPORT:
if (start < cbb_start_16_io)
start = cbb_start_16_io;
if (end < start)
end = start;
break;
case SYS_RES_IRQ:
tmp = rman_get_start(sc->irq_res);
if (start > tmp || end < tmp || count != 1) {
device_printf(child, "requested interrupt %ld-%ld,"
"count = %ld not supported by cbb\n",
start, end, count);
return (NULL);
}
flags |= RF_SHAREABLE;
start = end = rman_get_start(sc->irq_res);
break;
}
res = BUS_ALLOC_RESOURCE(up, child, type, rid,
start, end, count, flags & ~RF_ACTIVE);
if (res == NULL)
return (NULL);
cbb_insert_res(sc, res, type, *rid);
if (flags & RF_ACTIVE) {
if (bus_activate_resource(child, type, *rid, res) != 0) {
bus_release_resource(child, type, *rid, res);
return (NULL);
}
}
return (res);
}
static int
exca_release_resource(struct exca_softc *sc, device_t child, int type,
int rid, struct resource *res)
{
int error;
if (rman_get_flags(res) & RF_ACTIVE) {
error = bus_deactivate_resource(child, type, rid, res);
if (error != 0)
return (error);
}
cbb_remove_res(sc, res);
return (BUS_RELEASE_RESOURCE(device_get_parent(brdev), child,
type, rid, res));
}
#endif
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
static int
exca_modevent(module_t mod, int cmd, void *arg)
Make oldcard and newcard kernel module work.
2002-05-30 17:38:00 +00:00
{
return 0;
}
Massive overhaul of exca to help with the isa efforts: o chip_name arrays ifdef'd out. o use the OLDCARD-like get/put functions so we can support differnt types of mappings. o Write the beggings of is this a valid exca device and introduce more chipset support. # this is partially a wip, but also needed because some other cahnges I've # made require some of these changes.
2003-02-14 06:21:18 +00:00
Make oldcard and newcard kernel module work.
2002-05-30 17:38:00 +00:00
DEV_MODULE(exca, exca_modevent, NULL);
MODULE_VERSION(exca, 1);
Reference in New Issue Copy Permalink