linux/drivers/firewire/fw-card.c

556 lines
14 KiB
C
Raw Normal View History

/*
* Copyright (C) 2005-2007 Kristian Hoegsberg <krh@bitplanet.net>
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation; either version 2 of the License, or
* (at your option) any later version.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*
* You should have received a copy of the GNU General Public License
* along with this program; if not, write to the Free Software Foundation,
* Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
*/
#include <linux/completion.h>
#include <linux/crc-itu-t.h>
firewire: fix crash in automatic module unloading "modprobe firewire-ohci; sleep .1; modprobe -r firewire-ohci" used to result in crashes like this: BUG: unable to handle kernel paging request at ffffffff8807b455 IP: [<ffffffff8807b455>] PGD 203067 PUD 207063 PMD 7c170067 PTE 0 Oops: 0010 [1] PREEMPT SMP CPU 0 Modules linked in: i915 drm cpufreq_ondemand acpi_cpufreq freq_table applesmc input_polldev led_class coretemp hwmon eeprom snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss button thermal processor sg snd_hda_intel snd_pcm snd_timer snd snd_page_alloc sky2 i2c_i801 rtc [last unloaded: crc_itu_t] Pid: 9, comm: events/0 Not tainted 2.6.25-rc2 #3 RIP: 0010:[<ffffffff8807b455>] [<ffffffff8807b455>] RSP: 0018:ffff81007dcdde88 EFLAGS: 00010246 RAX: ffff81007dc95040 RBX: ffff81007dee5390 RCX: 0000000000005e13 RDX: 0000000000008c8b RSI: 0000000000000001 RDI: ffff81007dee5388 RBP: ffff81007dc5eb40 R08: 0000000000000002 R09: ffffffff8022d05c R10: ffffffff8023b34c R11: ffffffff8041a353 R12: ffff81007dee5388 R13: ffffffff8807b455 R14: ffffffff80593bc0 R15: 0000000000000000 FS: 0000000000000000(0000) GS:ffffffff8055a000(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: ffffffff8807b455 CR3: 0000000000201000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process events/0 (pid: 9, threadinfo ffff81007dcdc000, task ffff81007dc95040) Stack: ffffffff8023b396 ffffffff88082524 0000000000000000 ffffffff8807d9ae ffff81007dc5eb40 ffff81007dc9dce0 ffff81007dc5eb40 ffff81007dc5eb80 ffff81007dc9dce0 ffffffffffffffff ffffffff8023be87 0000000000000000 Call Trace: [<ffffffff8023b396>] ? run_workqueue+0xdf/0x1df [<ffffffff8023be87>] ? worker_thread+0xd8/0xe3 [<ffffffff8023e917>] ? autoremove_wake_function+0x0/0x2e [<ffffffff8023bdaf>] ? worker_thread+0x0/0xe3 [<ffffffff8023e813>] ? kthread+0x47/0x74 [<ffffffff804198e0>] ? trace_hardirqs_on_thunk+0x35/0x3a [<ffffffff8020c008>] ? child_rip+0xa/0x12 [<ffffffff8020b6e3>] ? restore_args+0x0/0x3d [<ffffffff8023e68a>] ? kthreadd+0x14c/0x171 [<ffffffff8023e68a>] ? kthreadd+0x14c/0x171 [<ffffffff8023e7cc>] ? kthread+0x0/0x74 [<ffffffff8020bffe>] ? child_rip+0x0/0x12 Code: Bad RIP value. RIP [<ffffffff8807b455>] RSP <ffff81007dcdde88> CR2: ffffffff8807b455 ---[ end trace c7366c6657fe5bed ]--- Note that this crash happened _after_ firewire-core was unloaded. The shared workqueue tried to run firewire-core's device initialization jobs or similar jobs. The fix makes sure that firewire-ohci and hence firewire-core is not unloaded before all device shutdown jobs have been completed. This is determined by the count of device initializations minus device releases. Also skip useless retries in the node initialization job if the node is to be shut down. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> Signed-off-by: Jarod Wilson <jwilson@redhat.com>
2008-02-27 21:14:27 +00:00
#include <linux/delay.h>
#include <linux/device.h>
#include <linux/errno.h>
#include <linux/kref.h>
#include <linux/module.h>
#include <linux/mutex.h>
#include "fw-transaction.h"
#include "fw-topology.h"
#include "fw-device.h"
int fw_compute_block_crc(u32 *block)
{
__be32 be32_block[256];
int i, length;
length = (*block >> 16) & 0xff;
for (i = 0; i < length; i++)
be32_block[i] = cpu_to_be32(block[i + 1]);
*block |= crc_itu_t(0, (u8 *) be32_block, length * 4);
return length;
}
static DEFINE_MUTEX(card_mutex);
static LIST_HEAD(card_list);
static LIST_HEAD(descriptor_list);
static int descriptor_count;
#define BIB_CRC(v) ((v) << 0)
#define BIB_CRC_LENGTH(v) ((v) << 16)
#define BIB_INFO_LENGTH(v) ((v) << 24)
#define BIB_LINK_SPEED(v) ((v) << 0)
#define BIB_GENERATION(v) ((v) << 4)
#define BIB_MAX_ROM(v) ((v) << 8)
#define BIB_MAX_RECEIVE(v) ((v) << 12)
#define BIB_CYC_CLK_ACC(v) ((v) << 16)
#define BIB_PMC ((1) << 27)
#define BIB_BMC ((1) << 28)
#define BIB_ISC ((1) << 29)
#define BIB_CMC ((1) << 30)
#define BIB_IMC ((1) << 31)
static u32 *
generate_config_rom(struct fw_card *card, size_t *config_rom_length)
{
struct fw_descriptor *desc;
static u32 config_rom[256];
int i, j, length;
/*
* Initialize contents of config rom buffer. On the OHCI
* controller, block reads to the config rom accesses the host
* memory, but quadlet read access the hardware bus info block
* registers. That's just crack, but it means we should make
* sure the contents of bus info block in host memory mathces
* the version stored in the OHCI registers.
*/
memset(config_rom, 0, sizeof(config_rom));
config_rom[0] = BIB_CRC_LENGTH(4) | BIB_INFO_LENGTH(4) | BIB_CRC(0);
config_rom[1] = 0x31333934;
config_rom[2] =
BIB_LINK_SPEED(card->link_speed) |
BIB_GENERATION(card->config_rom_generation++ % 14 + 2) |
BIB_MAX_ROM(2) |
BIB_MAX_RECEIVE(card->max_receive) |
BIB_BMC | BIB_ISC | BIB_CMC | BIB_IMC;
config_rom[3] = card->guid >> 32;
config_rom[4] = card->guid;
/* Generate root directory. */
i = 5;
config_rom[i++] = 0;
config_rom[i++] = 0x0c0083c0; /* node capabilities */
j = i + descriptor_count;
/* Generate root directory entries for descriptors. */
list_for_each_entry (desc, &descriptor_list, link) {
if (desc->immediate > 0)
config_rom[i++] = desc->immediate;
config_rom[i] = desc->key | (j - i);
i++;
j += desc->length;
}
/* Update root directory length. */
config_rom[5] = (i - 5 - 1) << 16;
/* End of root directory, now copy in descriptors. */
list_for_each_entry (desc, &descriptor_list, link) {
memcpy(&config_rom[i], desc->data, desc->length * 4);
i += desc->length;
}
/* Calculate CRCs for all blocks in the config rom. This
* assumes that CRC length and info length are identical for
* the bus info block, which is always the case for this
* implementation. */
for (i = 0; i < j; i += length + 1)
length = fw_compute_block_crc(config_rom + i);
*config_rom_length = j;
return config_rom;
}
static void
update_config_roms(void)
{
struct fw_card *card;
u32 *config_rom;
size_t length;
list_for_each_entry (card, &card_list, link) {
config_rom = generate_config_rom(card, &length);
card->driver->set_config_rom(card, config_rom, length);
}
}
int
fw_core_add_descriptor(struct fw_descriptor *desc)
{
size_t i;
/*
* Check descriptor is valid; the length of all blocks in the
* descriptor has to add up to exactly the length of the
* block.
*/
i = 0;
while (i < desc->length)
i += (desc->data[i] >> 16) + 1;
if (i != desc->length)
return -EINVAL;
mutex_lock(&card_mutex);
list_add_tail(&desc->link, &descriptor_list);
descriptor_count++;
if (desc->immediate > 0)
descriptor_count++;
update_config_roms();
mutex_unlock(&card_mutex);
return 0;
}
void
fw_core_remove_descriptor(struct fw_descriptor *desc)
{
mutex_lock(&card_mutex);
list_del(&desc->link);
descriptor_count--;
if (desc->immediate > 0)
descriptor_count--;
update_config_roms();
mutex_unlock(&card_mutex);
}
static const char gap_count_table[] = {
63, 5, 7, 8, 10, 13, 16, 18, 21, 24, 26, 29, 32, 35, 37, 40
};
struct bm_data {
struct fw_transaction t;
struct {
__be32 arg;
__be32 data;
} lock;
u32 old;
int rcode;
struct completion done;
};
static void
complete_bm_lock(struct fw_card *card, int rcode,
void *payload, size_t length, void *data)
{
struct bm_data *bmd = data;
if (rcode == RCODE_COMPLETE)
bmd->old = be32_to_cpu(*(__be32 *) payload);
bmd->rcode = rcode;
complete(&bmd->done);
}
static void
fw_card_bm_work(struct work_struct *work)
{
struct fw_card *card = container_of(work, struct fw_card, work.work);
struct fw_device *root_device;
struct fw_node *root_node, *local_node;
struct bm_data bmd;
unsigned long flags;
int root_id, new_root_id, irm_id, gap_count, generation, grace;
bool do_reset = false;
spin_lock_irqsave(&card->lock, flags);
local_node = card->local_node;
root_node = card->root_node;
if (local_node == NULL) {
spin_unlock_irqrestore(&card->lock, flags);
return;
}
fw_node_get(local_node);
fw_node_get(root_node);
generation = card->generation;
root_device = root_node->data;
if (root_device)
fw_device_get(root_device);
root_id = root_node->node_id;
grace = time_after(jiffies, card->reset_jiffies + DIV_ROUND_UP(HZ, 10));
if (card->bm_generation + 1 == generation ||
(card->bm_generation != generation && grace)) {
/*
* This first step is to figure out who is IRM and
* then try to become bus manager. If the IRM is not
* well defined (e.g. does not have an active link
* layer or does not responds to our lock request, we
* will have to do a little vigilante bus management.
* In that case, we do a goto into the gap count logic
* so that when we do the reset, we still optimize the
* gap count. That could well save a reset in the
* next generation.
*/
irm_id = card->irm_node->node_id;
if (!card->irm_node->link_on) {
new_root_id = local_node->node_id;
fw_notify("IRM has link off, making local node (%02x) root.\n",
new_root_id);
goto pick_me;
}
bmd.lock.arg = cpu_to_be32(0x3f);
bmd.lock.data = cpu_to_be32(local_node->node_id);
spin_unlock_irqrestore(&card->lock, flags);
init_completion(&bmd.done);
fw_send_request(card, &bmd.t, TCODE_LOCK_COMPARE_SWAP,
irm_id, generation,
SCODE_100, CSR_REGISTER_BASE + CSR_BUS_MANAGER_ID,
&bmd.lock, sizeof(bmd.lock),
complete_bm_lock, &bmd);
wait_for_completion(&bmd.done);
if (bmd.rcode == RCODE_GENERATION) {
/*
* Another bus reset happened. Just return,
* the BM work has been rescheduled.
*/
goto out;
}
if (bmd.rcode == RCODE_COMPLETE && bmd.old != 0x3f)
/* Somebody else is BM, let them do the work. */
goto out;
spin_lock_irqsave(&card->lock, flags);
if (bmd.rcode != RCODE_COMPLETE) {
/*
* The lock request failed, maybe the IRM
* isn't really IRM capable after all. Let's
* do a bus reset and pick the local node as
* root, and thus, IRM.
*/
new_root_id = local_node->node_id;
fw_notify("BM lock failed, making local node (%02x) root.\n",
new_root_id);
goto pick_me;
}
} else if (card->bm_generation != generation) {
/*
* OK, we weren't BM in the last generation, and it's
* less than 100ms since last bus reset. Reschedule
* this task 100ms from now.
*/
spin_unlock_irqrestore(&card->lock, flags);
schedule_delayed_work(&card->work, DIV_ROUND_UP(HZ, 10));
goto out;
}
/*
* We're bus manager for this generation, so next step is to
* make sure we have an active cycle master and do gap count
* optimization.
*/
card->bm_generation = generation;
if (root_device == NULL) {
/*
* Either link_on is false, or we failed to read the
* config rom. In either case, pick another root.
*/
new_root_id = local_node->node_id;
} else if (atomic_read(&root_device->state) != FW_DEVICE_RUNNING) {
/*
* If we haven't probed this device yet, bail out now
* and let's try again once that's done.
*/
spin_unlock_irqrestore(&card->lock, flags);
goto out;
} else if (root_device->cmc) {
/*
* FIXME: I suppose we should set the cmstr bit in the
* STATE_CLEAR register of this node, as described in
* 1394-1995, 8.4.2.6. Also, send out a force root
* packet for this node.
*/
new_root_id = root_id;
} else {
/*
* Current root has an active link layer and we
* successfully read the config rom, but it's not
* cycle master capable.
*/
new_root_id = local_node->node_id;
}
pick_me:
/*
* Pick a gap count from 1394a table E-1. The table doesn't cover
* the typically much larger 1394b beta repeater delays though.
*/
if (!card->beta_repeaters_present &&
root_node->max_hops < ARRAY_SIZE(gap_count_table))
gap_count = gap_count_table[root_node->max_hops];
else
gap_count = 63;
/*
* Finally, figure out if we should do a reset or not. If we have
* done less than 5 resets with the same physical topology and we
* have either a new root or a new gap count setting, let's do it.
*/
if (card->bm_retries++ < 5 &&
(card->gap_count != gap_count || new_root_id != root_id))
do_reset = true;
spin_unlock_irqrestore(&card->lock, flags);
if (do_reset) {
fw_notify("phy config: card %d, new root=%x, gap_count=%d\n",
card->index, new_root_id, gap_count);
fw_send_phy_config(card, new_root_id, generation, gap_count);
fw_core_initiate_bus_reset(card, 1);
}
out:
if (root_device)
fw_device_put(root_device);
fw_node_put(root_node);
fw_node_put(local_node);
}
static void
flush_timer_callback(unsigned long data)
{
struct fw_card *card = (struct fw_card *)data;
fw_flush_transactions(card);
}
void
fw_card_initialize(struct fw_card *card, const struct fw_card_driver *driver,
struct device *device)
{
static atomic_t index = ATOMIC_INIT(-1);
card->index = atomic_inc_return(&index);
card->driver = driver;
card->device = device;
card->current_tlabel = 0;
card->tlabel_mask = 0;
card->color = 0;
card->broadcast_channel = BROADCAST_CHANNEL_INITIAL;
kref_init(&card->kref);
init_completion(&card->done);
INIT_LIST_HEAD(&card->transaction_list);
spin_lock_init(&card->lock);
setup_timer(&card->flush_timer,
flush_timer_callback, (unsigned long)card);
card->local_node = NULL;
INIT_DELAYED_WORK(&card->work, fw_card_bm_work);
}
EXPORT_SYMBOL(fw_card_initialize);
int
fw_card_add(struct fw_card *card,
u32 max_receive, u32 link_speed, u64 guid)
{
u32 *config_rom;
size_t length;
card->max_receive = max_receive;
card->link_speed = link_speed;
card->guid = guid;
mutex_lock(&card_mutex);
config_rom = generate_config_rom(card, &length);
list_add_tail(&card->link, &card_list);
mutex_unlock(&card_mutex);
return card->driver->enable(card, config_rom, length);
}
EXPORT_SYMBOL(fw_card_add);
/*
* The next few functions implements a dummy driver that use once a
* card driver shuts down an fw_card. This allows the driver to
* cleanly unload, as all IO to the card will be handled by the dummy
* driver instead of calling into the (possibly) unloaded module. The
* dummy driver just fails all IO.
*/
static int
dummy_enable(struct fw_card *card, u32 *config_rom, size_t length)
{
BUG();
return -1;
}
static int
dummy_update_phy_reg(struct fw_card *card, int address,
int clear_bits, int set_bits)
{
return -ENODEV;
}
static int
dummy_set_config_rom(struct fw_card *card,
u32 *config_rom, size_t length)
{
/*
* We take the card out of card_list before setting the dummy
* driver, so this should never get called.
*/
BUG();
return -1;
}
static void
dummy_send_request(struct fw_card *card, struct fw_packet *packet)
{
packet->callback(packet, card, -ENODEV);
}
static void
dummy_send_response(struct fw_card *card, struct fw_packet *packet)
{
packet->callback(packet, card, -ENODEV);
}
static int
dummy_cancel_packet(struct fw_card *card, struct fw_packet *packet)
{
return -ENOENT;
}
static int
dummy_enable_phys_dma(struct fw_card *card,
int node_id, int generation)
{
return -ENODEV;
}
static struct fw_card_driver dummy_driver = {
.enable = dummy_enable,
.update_phy_reg = dummy_update_phy_reg,
.set_config_rom = dummy_set_config_rom,
.send_request = dummy_send_request,
.cancel_packet = dummy_cancel_packet,
.send_response = dummy_send_response,
.enable_phys_dma = dummy_enable_phys_dma,
};
void
fw_card_release(struct kref *kref)
{
struct fw_card *card = container_of(kref, struct fw_card, kref);
complete(&card->done);
}
void
fw_core_remove_card(struct fw_card *card)
{
card->driver->update_phy_reg(card, 4,
PHY_LINK_ACTIVE | PHY_CONTENDER, 0);
fw_core_initiate_bus_reset(card, 1);
mutex_lock(&card_mutex);
list_del(&card->link);
mutex_unlock(&card_mutex);
/* Set up the dummy driver. */
card->driver = &dummy_driver;
fw_destroy_nodes(card);
/* Wait for all users, especially device workqueue jobs, to finish. */
fw_card_put(card);
wait_for_completion(&card->done);
firewire: fix crash in automatic module unloading "modprobe firewire-ohci; sleep .1; modprobe -r firewire-ohci" used to result in crashes like this: BUG: unable to handle kernel paging request at ffffffff8807b455 IP: [<ffffffff8807b455>] PGD 203067 PUD 207063 PMD 7c170067 PTE 0 Oops: 0010 [1] PREEMPT SMP CPU 0 Modules linked in: i915 drm cpufreq_ondemand acpi_cpufreq freq_table applesmc input_polldev led_class coretemp hwmon eeprom snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss button thermal processor sg snd_hda_intel snd_pcm snd_timer snd snd_page_alloc sky2 i2c_i801 rtc [last unloaded: crc_itu_t] Pid: 9, comm: events/0 Not tainted 2.6.25-rc2 #3 RIP: 0010:[<ffffffff8807b455>] [<ffffffff8807b455>] RSP: 0018:ffff81007dcdde88 EFLAGS: 00010246 RAX: ffff81007dc95040 RBX: ffff81007dee5390 RCX: 0000000000005e13 RDX: 0000000000008c8b RSI: 0000000000000001 RDI: ffff81007dee5388 RBP: ffff81007dc5eb40 R08: 0000000000000002 R09: ffffffff8022d05c R10: ffffffff8023b34c R11: ffffffff8041a353 R12: ffff81007dee5388 R13: ffffffff8807b455 R14: ffffffff80593bc0 R15: 0000000000000000 FS: 0000000000000000(0000) GS:ffffffff8055a000(0000) knlGS:0000000000000000 CS: 0010 DS: 0018 ES: 0018 CR0: 000000008005003b CR2: ffffffff8807b455 CR3: 0000000000201000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400 Process events/0 (pid: 9, threadinfo ffff81007dcdc000, task ffff81007dc95040) Stack: ffffffff8023b396 ffffffff88082524 0000000000000000 ffffffff8807d9ae ffff81007dc5eb40 ffff81007dc9dce0 ffff81007dc5eb40 ffff81007dc5eb80 ffff81007dc9dce0 ffffffffffffffff ffffffff8023be87 0000000000000000 Call Trace: [<ffffffff8023b396>] ? run_workqueue+0xdf/0x1df [<ffffffff8023be87>] ? worker_thread+0xd8/0xe3 [<ffffffff8023e917>] ? autoremove_wake_function+0x0/0x2e [<ffffffff8023bdaf>] ? worker_thread+0x0/0xe3 [<ffffffff8023e813>] ? kthread+0x47/0x74 [<ffffffff804198e0>] ? trace_hardirqs_on_thunk+0x35/0x3a [<ffffffff8020c008>] ? child_rip+0xa/0x12 [<ffffffff8020b6e3>] ? restore_args+0x0/0x3d [<ffffffff8023e68a>] ? kthreadd+0x14c/0x171 [<ffffffff8023e68a>] ? kthreadd+0x14c/0x171 [<ffffffff8023e7cc>] ? kthread+0x0/0x74 [<ffffffff8020bffe>] ? child_rip+0x0/0x12 Code: Bad RIP value. RIP [<ffffffff8807b455>] RSP <ffff81007dcdde88> CR2: ffffffff8807b455 ---[ end trace c7366c6657fe5bed ]--- Note that this crash happened _after_ firewire-core was unloaded. The shared workqueue tried to run firewire-core's device initialization jobs or similar jobs. The fix makes sure that firewire-ohci and hence firewire-core is not unloaded before all device shutdown jobs have been completed. This is determined by the count of device initializations minus device releases. Also skip useless retries in the node initialization job if the node is to be shut down. Signed-off-by: Stefan Richter <stefanr@s5r6.in-berlin.de> Signed-off-by: Jarod Wilson <jwilson@redhat.com>
2008-02-27 21:14:27 +00:00
cancel_delayed_work_sync(&card->work);
WARN_ON(!list_empty(&card->transaction_list));
del_timer_sync(&card->flush_timer);
}
EXPORT_SYMBOL(fw_core_remove_card);
int
fw_core_initiate_bus_reset(struct fw_card *card, int short_reset)
{
int reg = short_reset ? 5 : 1;
int bit = short_reset ? PHY_BUS_SHORT_RESET : PHY_BUS_RESET;
return card->driver->update_phy_reg(card, reg, 0, bit);
}
EXPORT_SYMBOL(fw_core_initiate_bus_reset);