Skip to contentSkip to conversation

rtsx_pci_sdmmc lockdep splat

DISCUSSION Borislav Petkov 2 participants 4 messages Apr 15, 2014 – Apr 21, 2014 MMC, SD and SDIO

lore
Share & tools
lore
#0 lore raw

Hi all,

I'm seeing the following when I insert an sd card. Kernel is 3.15-rc1+.

Thanks.

[  682.675345] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_get_cd: RTSX_BIPR = 0x00010000
[  682.679723] mmc0: mmc_rescan_try_freq: trying to init card at 400000 Hz
[  682.683778] mmc0: clock 0Hz busmode 2 powermode 1 cs 0 Vdd 21 width 0 timing 0
[  682.689135] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_switch_voltage: signal_voltage = 0
[  682.701892] mmc0: clock 400000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 0 timing 0
[  682.718034] mmc0: starting CMD52 arg 00000c00 flags 00000195
[  682.720975] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sd_send_cmd: SD/MMC CMD 52, arg = 0x00000c00
[  682.723749] 
[  682.726488] ======================================================
[  682.729295] [ INFO: possible circular locking dependency detected ]
[  682.732138] 3.15.0-rc1+ #3 Not tainted
[  682.734989] -------------------------------------------------------
[  682.737905] kworker/u4:0/6 is trying to acquire lock:
[  682.740853]  (&(&pcr->lock)->rlock){-.-...}, at: [<ffffffffa0004387>] rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.743945] 
[  682.743945] but task is already holding lock:
[  682.749951]  (&(&host->lock)->rlock#3){-.....}, at: [<ffffffffa00bc577>] sdmmc_request+0x57/0x579 [rtsx_pci_sdmmc]
[  682.753076] 
[  682.753076] which lock already depends on the new lock.
[  682.753076] 
[  682.762208] 
[  682.762208] the existing dependency chain (in reverse order) is:
[  682.768191] 
[  682.768191] -> #1 (&(&host->lock)->rlock#3){-.....}:
[  682.774134]        [<ffffffff810bdfd6>] lock_acquire+0x96/0x130
[  682.777063]        [<ffffffff8153729b>] _raw_spin_lock+0x3b/0x50
[  682.779959]        [<ffffffffa00bb91d>] sd_isr_done_transfer+0x2d/0x80 [rtsx_pci_sdmmc]
[  682.782861]        [<ffffffffa0005679>] rtsx_pci_isr+0x119/0x220 [rtsx_pci]
[  682.785736]        [<ffffffff810cdea7>] handle_irq_event_percpu+0x47/0x1e0
[  682.788584]        [<ffffffff810ce088>] handle_irq_event+0x48/0x70
[  682.791387]        [<ffffffff810d0a9f>] handle_edge_irq+0x8f/0x150
[  682.794195]        [<ffffffff81004cee>] handle_irq+0x1e/0x30
[  682.797026]        [<ffffffff8153a8da>] do_IRQ+0x6a/0x110
[  682.799831]        [<ffffffff8153812f>] ret_from_intr+0x0/0x13
[  682.802645]        [<ffffffff81099ea8>] __atomic_notifier_call_chain+0x88/0xb0
[  682.805465]        [<ffffffff81099ee6>] atomic_notifier_call_chain+0x16/0x20
[  682.808246]        [<ffffffff8133af46>] do_con_write.part.18+0x176/0xa10
[  682.811028]        [<ffffffff8133b895>] con_write+0x75/0x80
[  682.813758]        [<ffffffff81324ac5>] n_tty_write+0x325/0x4d0
[  682.816497]        [<ffffffff81321540>] tty_write+0x150/0x2b0
[  682.819195]        [<ffffffff8118d571>] do_loop_readv_writev+0x61/0x90
[  682.821901]        [<ffffffff8118ef25>] do_readv_writev+0x205/0x210
[  682.824583]        [<ffffffff8118efb0>] vfs_writev+0x30/0x60
[  682.827226]        [<ffffffff8118f0f2>] SyS_writev+0x52/0xc0
[  682.829821]        [<ffffffff81538c92>] system_call_fastpath+0x16/0x1b
[  682.832383] 
[  682.832383] -> #0 (&(&pcr->lock)->rlock){-.-...}:
[  682.837324]        [<ffffffff810bd570>] __lock_acquire+0x1690/0x1950
[  682.839790]        [<ffffffff810bdfd6>] lock_acquire+0x96/0x130
[  682.842229]        [<ffffffff815373f6>] _raw_spin_lock_irqsave+0x46/0x60
[  682.844650]        [<ffffffffa0004387>] rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.847029]        [<ffffffffa00bbc0b>] sd_send_cmd+0xcb/0x3c0 [rtsx_pci_sdmmc]
[  682.849359]        [<ffffffffa00bc7c0>] sdmmc_request+0x2a0/0x579 [rtsx_pci_sdmmc]
[  682.851649]        [<ffffffffa009082f>] mmc_start_request+0x16f/0x280 [mmc_core]
[  682.853921]        [<ffffffffa00909ca>] __mmc_start_req+0x6a/0xa0 [mmc_core]
[  682.856162]        [<ffffffffa0090f7a>] mmc_wait_for_cmd+0x7a/0xd0 [mmc_core]
[  682.858387]        [<ffffffffa009bed0>] mmc_io_rw_direct_host+0x90/0x120 [mmc_core]
[  682.860608]        [<ffffffffa009c416>] sdio_reset+0x26/0x70 [mmc_core]
[  682.862815]        [<ffffffffa009399d>] mmc_rescan+0x29d/0x350 [mmc_core]
[  682.865007]        [<ffffffff8108deba>] process_one_work+0x21a/0x560
[  682.867175]        [<ffffffff8108e320>] worker_thread+0x120/0x390
[  682.869286]        [<ffffffff81094aac>] kthread+0xfc/0x120
[  682.871341]        [<ffffffff81538bec>] ret_from_fork+0x7c/0xb0
[  682.873404] 
[  682.873404] other info that might help us debug this:
[  682.873404] 
[  682.879197]  Possible unsafe locking scenario:
[  682.879197] 
[  682.883048]        CPU0                    CPU1
[  682.884972]        ----                    ----
[  682.886860]   lock(&(&host->lock)->rlock#3);
[  682.888763]                                lock(&(&pcr->lock)->rlock);
[  682.890686]                                lock(&(&host->lock)->rlock#3);
[  682.892608]   lock(&(&pcr->lock)->rlock);
[  682.894492] 
[  682.894492]  *** DEADLOCK ***
[  682.894492] 
[  682.899913] 4 locks held by kworker/u4:0/6:
[  682.901735]  #0:  ("kmmcd"){.+.+..}, at: [<ffffffff8108de45>] process_one_work+0x1a5/0x560
[  682.903677]  #1:  ((&(&host->detect)->work)){+.+...}, at: [<ffffffff8108de45>] process_one_work+0x1a5/0x560
[  682.905678]  #2:  (&pcr->pcr_mutex){+.+...}, at: [<ffffffffa00bc56f>] sdmmc_request+0x4f/0x579 [rtsx_pci_sdmmc]
[  682.907737]  #3:  (&(&host->lock)->rlock#3){-.....}, at: [<ffffffffa00bc577>] sdmmc_request+0x57/0x579 [rtsx_pci_sdmmc]
[  682.909851] 
[  682.909851] stack backtrace:
[  682.913909] CPU: 1 PID: 6 Comm: kworker/u4:0 Not tainted 3.15.0-rc1+ #3
[  682.915987] Hardware name: LENOVO 30515QG/30515QG, BIOS 8RET30WW (1.12 ) 09/15/2011
[  682.918124] Workqueue: kmmcd mmc_rescan [mmc_core]
[  682.920264]  ffffffff8202f4f0 ffff880119949940 ffffffff8152f1c5 ffffffff81ff6fd0
[  682.922471]  ffff880119949980 ffffffff8152c20e ffff8801199499d0 ffff8801198e59a8
[  682.924690]  0000000000000003 ffff8801198e51e0 0000000000000004 ffff8801198e59a8
[  682.926923] Call Trace:
[  682.929136]  [<ffffffff8152f1c5>] dump_stack+0x4e/0x7a
[  682.931384]  [<ffffffff8152c20e>] print_circular_bug+0x203/0x211
[  682.933625]  [<ffffffff810bd570>] __lock_acquire+0x1690/0x1950
[  682.935869]  [<ffffffff81537591>] ? _raw_spin_unlock_irqrestore+0x41/0x80
[  682.938106]  [<ffffffffa0004387>] ? rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.940345]  [<ffffffff810bdfd6>] lock_acquire+0x96/0x130
[  682.942585]  [<ffffffffa0004387>] ? rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.944835]  [<ffffffff815373f6>] _raw_spin_lock_irqsave+0x46/0x60
[  682.947089]  [<ffffffffa0004387>] ? rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.949350]  [<ffffffffa0004387>] rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.951601]  [<ffffffffa00bbc0b>] sd_send_cmd+0xcb/0x3c0 [rtsx_pci_sdmmc]
[  682.953849]  [<ffffffffa00bc7c0>] sdmmc_request+0x2a0/0x579 [rtsx_pci_sdmmc]
[  682.956116]  [<ffffffffa009082f>] mmc_start_request+0x16f/0x280 [mmc_core]
[  682.958382]  [<ffffffffa00909ca>] __mmc_start_req+0x6a/0xa0 [mmc_core]
[  682.960642]  [<ffffffffa0090f7a>] mmc_wait_for_cmd+0x7a/0xd0 [mmc_core]
[  682.962906]  [<ffffffffa0090940>] ? mmc_start_request+0x280/0x280 [mmc_core]
[  682.965183]  [<ffffffffa009bed0>] mmc_io_rw_direct_host+0x90/0x120 [mmc_core]
[  682.967471]  [<ffffffffa009c416>] sdio_reset+0x26/0x70 [mmc_core]
[  682.969765]  [<ffffffffa009399d>] mmc_rescan+0x29d/0x350 [mmc_core]
[  682.972056]  [<ffffffff8108deba>] process_one_work+0x21a/0x560
[  682.974342]  [<ffffffff8108de45>] ? process_one_work+0x1a5/0x560
[  682.976634]  [<ffffffff8108e320>] worker_thread+0x120/0x390
[  682.978916]  [<ffffffff8108e200>] ? process_one_work+0x560/0x560
[  682.981209]  [<ffffffff81094aac>] kthread+0xfc/0x120
[  682.983499]  [<ffffffff815375fc>] ? _raw_spin_unlock_irq+0x2c/0x60
[  682.985826]  [<ffffffff810949b0>] ? kthread_create_on_node+0x1f0/0x1f0
[  682.988182]  [<ffffffff81538bec>] ret_from_fork+0x7c/0xb0
[  682.990534]  [<ffffffff810949b0>] ? kthread_create_on_node+0x1f0/0x1f0
[  683.033350] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: timeout for request
Signature · 5 lines
Regards/Gruss,
    Boris.

Sent from a fat crate under my desk. Formatting is fine.
--
#1 lore raw

Hi Borislav,

Quoted 5 lines · Borislav Petkov

Hi all,

I'm seeing the following when I insert an sd card. Kernel is 3.15-rc1+.

Thanks.

The patch below may fix this problem.

From: Micky Ching<[email protected]>

To avoid dead lock, we need make sure host->lock is always acquire before pcr->lock. But in irq handler, we acquired pcr->lock in rtsx mfd driver, and sd_isr_done_transfer() is called during pcr->lock already acquired. Since in sd_isr_done_transfer() the only work we do is schdule tasklet, the cmd_tasklet and data_tasklet never conflict, so it is safe to remove spin_lock() here.

---
 drivers/mmc/host/rtsx_pci_sdmmc.c |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/mmc/host/rtsx_pci_sdmmc.c +1 -3
index 453e1d4..40695e0 100644
@@ -108,12 +108,10 @@ static void sd_isr_done_transfer(struct platform_device *pdev)
 {
 	struct realtek_pci_sdmmc *host = platform_get_drvdata(pdev);
 
-	spin_lock(&host->lock);
 	if (host->cmd)
 		tasklet_schedule(&host->cmd_tasklet);
-	if (host->data)
+	else if (host->data)
 		tasklet_schedule(&host->data_tasklet);
-	spin_unlock(&host->lock);
 }
 
 static void sd_request_timeout(unsigned long host_addr)
Signature · 5 lines
1.7.9.5

I have send out this patch, Thanks.
Best Regards.
micky.
Quoted 121 lines
[  682.675345] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_get_cd: RTSX_BIPR = 0x00010000
[  682.679723] mmc0: mmc_rescan_try_freq: trying to init card at 400000 Hz
[  682.683778] mmc0: clock 0Hz busmode 2 powermode 1 cs 0 Vdd 21 width 0 timing 0
[  682.689135] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_switch_voltage: signal_voltage = 0
[  682.701892] mmc0: clock 400000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 0 timing 0
[  682.718034] mmc0: starting CMD52 arg 00000c00 flags 00000195
[  682.720975] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sd_send_cmd: SD/MMC CMD 52, arg = 0x00000c00
[  682.723749]
[  682.726488] ======================================================
[  682.729295] [ INFO: possible circular locking dependency detected ]
[  682.732138] 3.15.0-rc1+ #3 Not tainted
[  682.734989] -------------------------------------------------------
[  682.737905] kworker/u4:0/6 is trying to acquire lock:
[  682.740853]  (&(&pcr->lock)->rlock){-.-...}, at: [<ffffffffa0004387>] rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.743945]
[  682.743945] but task is already holding lock:
[  682.749951]  (&(&host->lock)->rlock#3){-.....}, at: [<ffffffffa00bc577>] sdmmc_request+0x57/0x579 [rtsx_pci_sdmmc]
[  682.753076]
[  682.753076] which lock already depends on the new lock.
[  682.753076]
[  682.762208]
[  682.762208] the existing dependency chain (in reverse order) is:
[  682.768191]
[  682.768191] -> #1 (&(&host->lock)->rlock#3){-.....}:
[  682.774134]        [<ffffffff810bdfd6>] lock_acquire+0x96/0x130
[  682.777063]        [<ffffffff8153729b>] _raw_spin_lock+0x3b/0x50
[  682.779959]        [<ffffffffa00bb91d>] sd_isr_done_transfer+0x2d/0x80 [rtsx_pci_sdmmc]
[  682.782861]        [<ffffffffa0005679>] rtsx_pci_isr+0x119/0x220 [rtsx_pci]
[  682.785736]        [<ffffffff810cdea7>] handle_irq_event_percpu+0x47/0x1e0
[  682.788584]        [<ffffffff810ce088>] handle_irq_event+0x48/0x70
[  682.791387]        [<ffffffff810d0a9f>] handle_edge_irq+0x8f/0x150
[  682.794195]        [<ffffffff81004cee>] handle_irq+0x1e/0x30
[  682.797026]        [<ffffffff8153a8da>] do_IRQ+0x6a/0x110
[  682.799831]        [<ffffffff8153812f>] ret_from_intr+0x0/0x13
[  682.802645]        [<ffffffff81099ea8>] __atomic_notifier_call_chain+0x88/0xb0
[  682.805465]        [<ffffffff81099ee6>] atomic_notifier_call_chain+0x16/0x20
[  682.808246]        [<ffffffff8133af46>] do_con_write.part.18+0x176/0xa10
[  682.811028]        [<ffffffff8133b895>] con_write+0x75/0x80
[  682.813758]        [<ffffffff81324ac5>] n_tty_write+0x325/0x4d0
[  682.816497]        [<ffffffff81321540>] tty_write+0x150/0x2b0
[  682.819195]        [<ffffffff8118d571>] do_loop_readv_writev+0x61/0x90
[  682.821901]        [<ffffffff8118ef25>] do_readv_writev+0x205/0x210
[  682.824583]        [<ffffffff8118efb0>] vfs_writev+0x30/0x60
[  682.827226]        [<ffffffff8118f0f2>] SyS_writev+0x52/0xc0
[  682.829821]        [<ffffffff81538c92>] system_call_fastpath+0x16/0x1b
[  682.832383]
[  682.832383] -> #0 (&(&pcr->lock)->rlock){-.-...}:
[  682.837324]        [<ffffffff810bd570>] __lock_acquire+0x1690/0x1950
[  682.839790]        [<ffffffff810bdfd6>] lock_acquire+0x96/0x130
[  682.842229]        [<ffffffff815373f6>] _raw_spin_lock_irqsave+0x46/0x60
[  682.844650]        [<ffffffffa0004387>] rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.847029]        [<ffffffffa00bbc0b>] sd_send_cmd+0xcb/0x3c0 [rtsx_pci_sdmmc]
[  682.849359]        [<ffffffffa00bc7c0>] sdmmc_request+0x2a0/0x579 [rtsx_pci_sdmmc]
[  682.851649]        [<ffffffffa009082f>] mmc_start_request+0x16f/0x280 [mmc_core]
[  682.853921]        [<ffffffffa00909ca>] __mmc_start_req+0x6a/0xa0 [mmc_core]
[  682.856162]        [<ffffffffa0090f7a>] mmc_wait_for_cmd+0x7a/0xd0 [mmc_core]
[  682.858387]        [<ffffffffa009bed0>] mmc_io_rw_direct_host+0x90/0x120 [mmc_core]
[  682.860608]        [<ffffffffa009c416>] sdio_reset+0x26/0x70 [mmc_core]
[  682.862815]        [<ffffffffa009399d>] mmc_rescan+0x29d/0x350 [mmc_core]
[  682.865007]        [<ffffffff8108deba>] process_one_work+0x21a/0x560
[  682.867175]        [<ffffffff8108e320>] worker_thread+0x120/0x390
[  682.869286]        [<ffffffff81094aac>] kthread+0xfc/0x120
[  682.871341]        [<ffffffff81538bec>] ret_from_fork+0x7c/0xb0
[  682.873404]
[  682.873404] other info that might help us debug this:
[  682.873404]
[  682.879197]  Possible unsafe locking scenario:
[  682.879197]
[  682.883048]        CPU0                    CPU1
[  682.884972]        ----                    ----
[  682.886860]   lock(&(&host->lock)->rlock#3);
[  682.888763]                                lock(&(&pcr->lock)->rlock);
[  682.890686]                                lock(&(&host->lock)->rlock#3);
[  682.892608]   lock(&(&pcr->lock)->rlock);
[  682.894492]
[  682.894492]  *** DEADLOCK ***
[  682.894492]
[  682.899913] 4 locks held by kworker/u4:0/6:
[  682.901735]  #0:  ("kmmcd"){.+.+..}, at: [<ffffffff8108de45>] process_one_work+0x1a5/0x560
[  682.903677]  #1:  ((&(&host->detect)->work)){+.+...}, at: [<ffffffff8108de45>] process_one_work+0x1a5/0x560
[  682.905678]  #2:  (&pcr->pcr_mutex){+.+...}, at: [<ffffffffa00bc56f>] sdmmc_request+0x4f/0x579 [rtsx_pci_sdmmc]
[  682.907737]  #3:  (&(&host->lock)->rlock#3){-.....}, at: [<ffffffffa00bc577>] sdmmc_request+0x57/0x579 [rtsx_pci_sdmmc]
[  682.909851]
[  682.909851] stack backtrace:
[  682.913909] CPU: 1 PID: 6 Comm: kworker/u4:0 Not tainted 3.15.0-rc1+ #3
[  682.915987] Hardware name: LENOVO 30515QG/30515QG, BIOS 8RET30WW (1.12 ) 09/15/2011
[  682.918124] Workqueue: kmmcd mmc_rescan [mmc_core]
[  682.920264]  ffffffff8202f4f0 ffff880119949940 ffffffff8152f1c5 ffffffff81ff6fd0
[  682.922471]  ffff880119949980 ffffffff8152c20e ffff8801199499d0 ffff8801198e59a8
[  682.924690]  0000000000000003 ffff8801198e51e0 0000000000000004 ffff8801198e59a8
[  682.926923] Call Trace:
[  682.929136]  [<ffffffff8152f1c5>] dump_stack+0x4e/0x7a
[  682.931384]  [<ffffffff8152c20e>] print_circular_bug+0x203/0x211
[  682.933625]  [<ffffffff810bd570>] __lock_acquire+0x1690/0x1950
[  682.935869]  [<ffffffff81537591>] ? _raw_spin_unlock_irqrestore+0x41/0x80
[  682.938106]  [<ffffffffa0004387>] ? rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.940345]  [<ffffffff810bdfd6>] lock_acquire+0x96/0x130
[  682.942585]  [<ffffffffa0004387>] ? rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.944835]  [<ffffffff815373f6>] _raw_spin_lock_irqsave+0x46/0x60
[  682.947089]  [<ffffffffa0004387>] ? rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.949350]  [<ffffffffa0004387>] rtsx_pci_add_cmd+0x47/0xb0 [rtsx_pci]
[  682.951601]  [<ffffffffa00bbc0b>] sd_send_cmd+0xcb/0x3c0 [rtsx_pci_sdmmc]
[  682.953849]  [<ffffffffa00bc7c0>] sdmmc_request+0x2a0/0x579 [rtsx_pci_sdmmc]
[  682.956116]  [<ffffffffa009082f>] mmc_start_request+0x16f/0x280 [mmc_core]
[  682.958382]  [<ffffffffa00909ca>] __mmc_start_req+0x6a/0xa0 [mmc_core]
[  682.960642]  [<ffffffffa0090f7a>] mmc_wait_for_cmd+0x7a/0xd0 [mmc_core]
[  682.962906]  [<ffffffffa0090940>] ? mmc_start_request+0x280/0x280 [mmc_core]
[  682.965183]  [<ffffffffa009bed0>] mmc_io_rw_direct_host+0x90/0x120 [mmc_core]
[  682.967471]  [<ffffffffa009c416>] sdio_reset+0x26/0x70 [mmc_core]
[  682.969765]  [<ffffffffa009399d>] mmc_rescan+0x29d/0x350 [mmc_core]
[  682.972056]  [<ffffffff8108deba>] process_one_work+0x21a/0x560
[  682.974342]  [<ffffffff8108de45>] ? process_one_work+0x1a5/0x560
[  682.976634]  [<ffffffff8108e320>] worker_thread+0x120/0x390
[  682.978916]  [<ffffffff8108e200>] ? process_one_work+0x560/0x560
[  682.981209]  [<ffffffff81094aac>] kthread+0xfc/0x120
[  682.983499]  [<ffffffff815375fc>] ? _raw_spin_unlock_irq+0x2c/0x60
[  682.985826]  [<ffffffff810949b0>] ? kthread_create_on_node+0x1f0/0x1f0
[  682.988182]  [<ffffffff81538bec>] ret_from_fork+0x7c/0xb0
[  682.990534]  [<ffffffff810949b0>] ? kthread_create_on_node+0x1f0/0x1f0
[  683.033350] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: timeout for request
#2 lore raw

Hi,

Quoted 21 lines · micky

From: Micky Ching<[email protected]>

To avoid dead lock, we need make sure host->lock is always acquire before pcr->lock. But in irq handler, we acquired pcr->lock in rtsx mfd driver, and sd_isr_done_transfer() is called during pcr->lock already acquired. Since in sd_isr_done_transfer() the only work we do is schdule tasklet, the cmd_tasklet and data_tasklet never conflict, so it is safe to remove spin_lock() here.

---
 drivers/mmc/host/rtsx_pci_sdmmc.c |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/mmc/host/rtsx_pci_sdmmc.c +0 -0
index 453e1d4..40695e0 100644
@@ -108,12 +108,10 @@ static void sd_isr_done_transfer(struct platform_device *pdev)
 {

drivers/mmc/host/rtsx_pci_sdmmc.c static void sd_isr_done_transfer(struct platform_device *pdev) · line 115

Quoted 10 lines
drivers/mmc/host/rtsx_pci_sdmmc.c +1 -3
 	struct realtek_pci_sdmmc *host = platform_get_drvdata(pdev);
-	spin_lock(&host->lock);
 	if (host->cmd)
 		tasklet_schedule(&host->cmd_tasklet);
-	if (host->data)
+	else if (host->data)
 		tasklet_schedule(&host->data_tasklet);
-	spin_unlock(&host->lock);
 }
 static void sd_request_timeout(unsigned long host_addr)

Patch wouldn't apply against rc1, so I had to add it by hand, diff below.

---

drivers/mmc/host/rtsx_pci_sdmmc.c +1 -3
index 5fb994f9a653..d1d65c0d8d2e 100644
@@ -109,12 +109,10 @@ static void sd_isr_done_transfer(struct platform_device *pdev)
 {
 	struct realtek_pci_sdmmc *host = platform_get_drvdata(pdev);
 
-	spin_lock(&host->lock);
 	if (host->cmd)
 		tasklet_schedule(&host->cmd_tasklet);
-	if (host->data)
+	else if (host->data)
 		tasklet_schedule(&host->data_tasklet);
-	spin_unlock(&host->lock);
 }
 
 static void sd_request_timeout(unsigned long host_addr)

--

Still needs more love:

[   95.112797] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_get_cd: RTSX_BIPR = 0x00010000
[   95.115700] mmc0: mmc_rescan_try_freq: trying to init card at 400000 Hz
[   95.118040] mmc0: clock 0Hz busmode 2 powermode 1 cs 0 Vdd 21 width 0 timing 0
[   95.122504] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_switch_voltage: signal_voltage = 0
[   95.139469] mmc0: clock 400000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 0 timing 0
[   95.155461] mmc0: starting CMD52 arg 00000c00 flags 00000195
[   95.158617] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sd_send_cmd: SD/MMC CMD 52, arg = 0x00000c00
[   95.255497] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: timeout for request
[   95.258412] ------------[ cut here ]------------
[   95.261163] WARNING: CPU: 0 PID: 3 at kernel/locking/mutex.c:698 __mutex_unlock_slowpath+0x124/0x180()
[   95.264120] DEBUG_LOCKS_WARN_ON(in_interrupt())
[   95.264222] Modules linked in: cpufreq_stats cpufreq_conservative cpufreq_powersave cpufreq_userspace binfmt_misc uinput ipv6 vfat fat rtsx_pci_sdmmc mmc_core arc4 radeon rtl8192ce snd_hda_codec_conexant snd_hda_codec_generic snd_hda_codec_hdmi snd_hda_intel rtl_pci rtlwifi snd_hda_controller snd_hda_codec mac80211 cfg80211 snd_hwdep snd_pcm drm_kms_helper pcspkr k10temp evdev snd_seq thermal thinkpad_acpi ehci_pci nvram rtsx_pci mfd_core snd_seq_device rtl8192c_common snd_timer ohci_pci ohci_hcd ehci_hcd snd battery ttm ac video button soundcore
[   95.264231] CPU: 0 PID: 3 Comm: ksoftirqd/0 Not tainted 3.15.0-rc1+ #1
[   95.264235] Hardware name: LENOVO 30515QG/30515QG, BIOS 8RET30WW (1.12 ) 09/15/2011
[   95.264246]  0000000000000009 ffff8801198f3c50 ffffffff81522644 ffff8801198f3c98
[   95.264254]  ffff8801198f3c88 ffffffff810701fc ffff8800c0ec1230 ffff8800c28b5c30
[   95.264262]  ffff8800c0ec1000 ffff8800bea267b0 0000000000000000 ffff8801198f3ce8
[   95.264264] Call Trace:
[   95.264277]  [<ffffffff81522644>] dump_stack+0x4e/0x7a
[   95.264288]  [<ffffffff810701fc>] warn_slowpath_common+0x8c/0xc0
[   95.264297]  [<ffffffff8107027c>] warn_slowpath_fmt+0x4c/0x50
[   95.264305]  [<ffffffff815294c4>] __mutex_unlock_slowpath+0x124/0x180
[   95.264311]  [<ffffffff8152952e>] mutex_unlock+0xe/0x10
[   95.264324]  [<ffffffffa0067119>] sd_finish_request+0xd9/0x1c0 [rtsx_pci_sdmmc]
[   95.264332]  [<ffffffff810751a0>] tasklet_action+0x110/0x170
[   95.264338]  [<ffffffff810756ba>] __do_softirq+0x12a/0x320
[   95.264345]  [<ffffffff810758dd>] run_ksoftirqd+0x2d/0x50
[   95.264354]  [<ffffffff8109be75>] smpboot_thread_fn+0x175/0x2b0
[   95.264363]  [<ffffffff8109bd00>] ? SyS_setgroups+0x150/0x150
[   95.264372]  [<ffffffff8109497c>] kthread+0xfc/0x120
[   95.264382]  [<ffffffff81094880>] ? kthread_create_on_node+0x1f0/0x1f0
[   95.264390]  [<ffffffff8152bfec>] ret_from_fork+0x7c/0xb0
[   95.264398]  [<ffffffff81094880>] ? kthread_create_on_node+0x1f0/0x1f0
[   95.264403] ---[ end trace e7e46e1611dc934d ]---
Signature · 5 lines
Regards/Gruss,
    Boris.

Sent from a fat crate under my desk. Formatting is fine.
--
#3 lore raw

Hi Borislav

Thank you for test the patch. It seems we can't use tasklet here due to mutex_lock. And the spinlock may removed, I'm prepare to resend a different patch later.

Best Regards.
micky

Quoted 96 lines · Borislav Petkov

Hi,

Quoted 21 lines · micky

From: Micky Ching<[email protected]>

To avoid dead lock, we need make sure host->lock is always acquire before pcr->lock. But in irq handler, we acquired pcr->lock in rtsx mfd driver, and sd_isr_done_transfer() is called during pcr->lock already acquired. Since in sd_isr_done_transfer() the only work we do is schdule tasklet, the cmd_tasklet and data_tasklet never conflict, so it is safe to remove spin_lock() here.

---
  drivers/mmc/host/rtsx_pci_sdmmc.c |    4 +---
  1 file changed, 1 insertion(+), 3 deletions(-)
drivers/mmc/host/rtsx_pci_sdmmc.c +0 -0
index 453e1d4..40695e0 100644
@@ -108,12 +108,10 @@ static void sd_isr_done_transfer(struct platform_device *pdev)
  {

drivers/mmc/host/rtsx_pci_sdmmc.c static void sd_isr_done_transfer(struct platform_device *pdev) · line 115

Quoted 10 lines
drivers/mmc/host/rtsx_pci_sdmmc.c +1 -3
  	struct realtek_pci_sdmmc *host = platform_get_drvdata(pdev);
-	spin_lock(&host->lock);
  	if (host->cmd)
  		tasklet_schedule(&host->cmd_tasklet);
-	if (host->data)
+	else if (host->data)
  		tasklet_schedule(&host->data_tasklet);
-	spin_unlock(&host->lock);
  }
  static void sd_request_timeout(unsigned long host_addr)

Patch wouldn't apply against rc1, so I had to add it by hand, diff below.

---

drivers/mmc/host/rtsx_pci_sdmmc.c +1 -4
index 5fb994f9a653..d1d65c0d8d2e 100644
@@ -109,12 +109,10 @@ static void sd_isr_done_transfer(struct platform_device *pdev)
  {
  	struct realtek_pci_sdmmc *host = platform_get_drvdata(pdev);
 
-	spin_lock(&host->lock);
  	if (host->cmd)
  		tasklet_schedule(&host->cmd_tasklet);
-	if (host->data)
+	else if (host->data)
  		tasklet_schedule(&host->data_tasklet);
-	spin_unlock(&host->lock);
  }
 
  static void sd_request_timeout(unsigned long host_addr)
--

Still needs more love:

[   95.112797] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_get_cd: RTSX_BIPR = 0x00010000
[   95.115700] mmc0: mmc_rescan_try_freq: trying to init card at 400000 Hz
[   95.118040] mmc0: clock 0Hz busmode 2 powermode 1 cs 0 Vdd 21 width 0 timing 0
[   95.122504] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sdmmc_switch_voltage: signal_voltage = 0
[   95.139469] mmc0: clock 400000Hz busmode 2 powermode 2 cs 0 Vdd 21 width 0 timing 0
[   95.155461] mmc0: starting CMD52 arg 00000c00 flags 00000195
[   95.158617] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: sd_send_cmd: SD/MMC CMD 52, arg = 0x00000c00
[   95.255497] rtsx_pci_sdmmc rtsx_pci_sdmmc.0: timeout for request
[   95.258412] ------------[ cut here ]------------
[   95.261163] WARNING: CPU: 0 PID: 3 at kernel/locking/mutex.c:698 __mutex_unlock_slowpath+0x124/0x180()
[   95.264120] DEBUG_LOCKS_WARN_ON(in_interrupt())
[   95.264222] Modules linked in: cpufreq_stats cpufreq_conservative cpufreq_powersave cpufreq_userspace binfmt_misc uinput ipv6 vfat fat rtsx_pci_sdmmc mmc_core arc4 radeon rtl8192ce snd_hda_codec_conexant snd_hda_codec_generic snd_hda_codec_hdmi snd_hda_intel rtl_pci rtlwifi snd_hda_controller snd_hda_codec mac80211 cfg80211 snd_hwdep snd_pcm drm_kms_helper pcspkr k10temp evdev snd_seq thermal thinkpad_acpi ehci_pci nvram rtsx_pci mfd_core snd_seq_device rtl8192c_common snd_timer ohci_pci ohci_hcd ehci_hcd snd battery ttm ac video button soundcore
[   95.264231] CPU: 0 PID: 3 Comm: ksoftirqd/0 Not tainted 3.15.0-rc1+ #1
[   95.264235] Hardware name: LENOVO 30515QG/30515QG, BIOS 8RET30WW (1.12 ) 09/15/2011
[   95.264246]  0000000000000009 ffff8801198f3c50 ffffffff81522644 ffff8801198f3c98
[   95.264254]  ffff8801198f3c88 ffffffff810701fc ffff8800c0ec1230 ffff8800c28b5c30
[   95.264262]  ffff8800c0ec1000 ffff8800bea267b0 0000000000000000 ffff8801198f3ce8
[   95.264264] Call Trace:
[   95.264277]  [<ffffffff81522644>] dump_stack+0x4e/0x7a
[   95.264288]  [<ffffffff810701fc>] warn_slowpath_common+0x8c/0xc0
[   95.264297]  [<ffffffff8107027c>] warn_slowpath_fmt+0x4c/0x50
[   95.264305]  [<ffffffff815294c4>] __mutex_unlock_slowpath+0x124/0x180
[   95.264311]  [<ffffffff8152952e>] mutex_unlock+0xe/0x10
[   95.264324]  [<ffffffffa0067119>] sd_finish_request+0xd9/0x1c0 [rtsx_pci_sdmmc]
[   95.264332]  [<ffffffff810751a0>] tasklet_action+0x110/0x170
[   95.264338]  [<ffffffff810756ba>] __do_softirq+0x12a/0x320
[   95.264345]  [<ffffffff810758dd>] run_ksoftirqd+0x2d/0x50
[   95.264354]  [<ffffffff8109be75>] smpboot_thread_fn+0x175/0x2b0
[   95.264363]  [<ffffffff8109bd00>] ? SyS_setgroups+0x150/0x150
[   95.264372]  [<ffffffff8109497c>] kthread+0xfc/0x120
[   95.264382]  [<ffffffff81094880>] ? kthread_create_on_node+0x1f0/0x1f0
[   95.264390]  [<ffffffff8152bfec>] ret_from_fork+0x7c/0xb0
[   95.264398]  [<ffffffff81094880>] ? kthread_create_on_node+0x1f0/0x1f0
[   95.264403] ---[ end trace e7e46e1611dc934d ]---