Hi,
As you suggest, i've build the kernel 4.9 rt10 from the debian sources,
with lockdep.
Here the trace when system hang :
Mar 6 17:06:46 LaChoze kernel: [ 339.062919] sysrq: SysRq : Show Locks
Held
Mar 6 17:06:46 LaChoze kernel: [ 339.062929]
Mar 6 17:06:46 LaChoze kernel: [ 339.062929] Showing all locks held in
the system:
Mar 6 17:06:46 LaChoze kernel: [ 339.062947] 5 locks held by
irq/1-i8042/126:
Mar 6 17:06:46 LaChoze kernel: [ 339.062948] #0:
(&serio->lock){+.+...}, at: [<ffffffff8794dab8>] serio_interrupt+0x28/0x80
Mar 6 17:06:46 LaChoze kernel: [ 339.062960] #1:
(&dev->event_lock){+.+...}, at: [<ffffffff87953cba>] input_event+0x3a/0x60
Mar 6 17:06:46 LaChoze kernel: [ 339.062967] #2:
(rcu_read_lock){......}, at: [<ffffffff87952ee5>]
input_pass_values.part.5+0x5/0x270
Mar 6 17:06:46 LaChoze kernel: [ 339.062975] #3:
(rcu_read_lock){......}, at: [<ffffffff878c1eb5>] __handle_sysrq+0x5/0x220
Mar 6 17:06:46 LaChoze kernel: [ 339.062982] #4:
(tasklist_lock){+.+...}, at: [<ffffffff874eae1d>]
debug_show_all_locks+0x3d/0x1a0
Mar 6 17:06:46 LaChoze kernel: [ 339.063009] 1 lock held by in:imklog/822:
Mar 6 17:06:46 LaChoze kernel: [ 339.063011] #0:
(&f->f_pos_lock){+.+.+.}, at: [<ffffffff8768883a>] __fdget_pos+0x4a/0x50
Mar 6 17:06:46 LaChoze kernel: [ 339.063040] 2 locks held by Xorg/2365:
Mar 6 17:06:46 LaChoze kernel: [ 339.063041] #0:
(&rdev->exclusive_lock){++++.+}, at: [<ffffffff874f3710>]
rt_down_read+0x10/0x20
Mar 6 17:06:46 LaChoze kernel: [ 339.063049] #1:
(&rdev->pm.mclk_lock){++++++}, at: [<ffffffff874f3710>]
rt_down_read+0x10/0x20
Mar 6 17:06:46 LaChoze kernel: [ 339.063100] 2 locks held by G.Main/3575:
Mar 6 17:06:46 LaChoze kernel: [ 339.063101] #0:
(&rdev->pm.mclk_lock){++++++}, at: [<ffffffff874f3710>]
rt_down_read+0x10/0x20
Mar 6 17:06:46 LaChoze kernel: [ 339.063108] #1:
(reservation_ww_class_mutex){+.+.+.}, at: [<ffffffffc087b7b0>]
ttm_bo_vm_fault+0x60/0x530 [ttm]
Mar 6 17:06:46 LaChoze kernel: [ 339.063124]
Mar 6 17:06:46 LaChoze kernel: [ 339.063125]
=============================================
Mar 6 17:06:46 LaChoze kernel: [ 339.063125]
Mar 6 17:07:11 LaChoze kernel: [ 363.523950] INFO: task Xorg:2365
blocked for more than 120 seconds.
Mar 6 17:07:11 LaChoze kernel: [ 363.523959] Not tainted
4.9.13-rt10 #1
Mar 6 17:07:11 LaChoze kernel: [ 363.523961] "echo 0 >
/proc/sys/kernel/hung_task_timeout_secs" disables this message.
Mar 6 17:07:11 LaChoze kernel: [ 363.523964] Xorg D 0
2365 2363 0x00000004
Mar 6 17:07:11 LaChoze kernel: [ 363.523970] 0000000000000000
ffff96333582b458 ffff9633027bd580 ffff96333582b440
Mar 6 17:07:11 LaChoze kernel: [ 363.523975] ffff96332f2b1c80
ffffba8646bffa50 ffffffff87af622f 00ffba8646bffb08
Mar 6 17:07:11 LaChoze kernel: [ 363.523980] ffff9633027bdbc0
ffff96333582b458 ffff96332841b800 ffff9633027bd580
Mar 6 17:07:11 LaChoze kernel: [ 363.523984] Call Trace:
Mar 6 17:07:11 LaChoze kernel: [ 363.523984] Call Trace:
Mar 6 17:07:11 LaChoze kernel: [ 363.523992] [<ffffffff87af622f>] ?
__schedule+0x32f/0x800
Mar 6 17:07:11 LaChoze kernel: [ 363.523996] [<ffffffff87af674c>]
schedule+0x4c/0xe0
Mar 6 17:07:11 LaChoze kernel: [ 363.523998] [<ffffffff87af8146>]
__rt_mutex_slowlock+0x56/0x1c0
Mar 6 17:07:11 LaChoze kernel: [ 363.524001] [<ffffffff87af8834>]
rt_mutex_slowlock+0x114/0x360
Mar 6 17:07:11 LaChoze kernel: [ 363.524005] [<ffffffff87665615>] ?
get_empty_filp+0x175/0x250
Mar 6 17:07:11 LaChoze kernel: [ 363.524007] [<ffffffff87af8a9a>]
rt_mutex_lock+0x1a/0x20
Mar 6 17:07:11 LaChoze kernel: [ 363.524011] [<ffffffff874f36e3>]
__rt_down_read+0x53/0x70
Mar 6 17:07:11 LaChoze kernel: [ 363.524014] [<ffffffff874f3710>] ?
rt_down_read+0x10/0x20
Mar 6 17:07:11 LaChoze kernel: [ 363.524016] [<ffffffff874f3710>]
rt_down_read+0x10/0x20
Mar 6 17:07:11 LaChoze kernel: [ 363.524043] [<ffffffffc0ae7d97>]
radeon_bo_create+0x157/0x360 [radeon]
Mar 6 17:07:11 LaChoze kernel: [ 363.524064] [<ffffffffc0afb21d>]
radeon_gem_object_create+0xbd/0x1a0 [radeon]
Mar 6 17:07:11 LaChoze kernel: [ 363.524084] [<ffffffffc0afb6ae>]
radeon_gem_create_ioctl+0x6e/0x100 [radeon]
Mar 6 17:07:11 LaChoze kernel: [ 363.524101] [<ffffffffc081ef6a>]
drm_ioctl+0x37a/0x4d0 [drm]
Mar 6 17:07:11 LaChoze kernel: [ 363.524119] [<ffffffffc0afb640>] ?
radeon_gem_pwrite_ioctl+0x30/0x30 [radeon]
Mar 6 17:07:11 LaChoze kernel: [ 363.524122] [<ffffffff877de437>] ?
debug_smp_processor_id+0x17/0x20
Mar 6 17:07:11 LaChoze kernel: [ 363.524125] [<ffffffff8748d8f6>] ?
unpin_current_cpu+0x16/0x70
Mar 6 17:07:11 LaChoze kernel: [ 363.524128] [<ffffffff874bce47>] ?
migrate_enable+0x87/0x150
Mar 6 17:07:11 LaChoze kernel: [ 363.524131] [<ffffffff87afb18f>] ?
rt_spin_unlock+0x2f/0x40
Mar 6 17:07:11 LaChoze kernel: [ 363.524145] [<ffffffffc0ac804f>]
radeon_drm_ioctl+0x4f/0x90 [radeon]
Mar 6 17:07:11 LaChoze kernel: [ 363.524148] [<ffffffff87679fd3>]
do_vfs_ioctl+0xa3/0x6f0
Mar 6 17:07:11 LaChoze kernel: [ 363.524152] [<ffffffff8768789c>] ?
__fget+0x11c/0x210
Mar 6 17:07:11 LaChoze kernel: [ 363.524154] [<ffffffff87687785>] ?
__fget+0x5/0x210
Mar 6 17:07:11 LaChoze kernel: [ 363.524157] [<ffffffff8767a699>]
SyS_ioctl+0x79/0x90
Mar 6 17:07:11 LaChoze kernel: [ 363.524159] [<ffffffff87afb845>]
system_call_fast_compare_end+0xc/0xaf
Mar 6 17:07:11 LaChoze kernel: [ 363.524162]
Mar 6 17:07:11 LaChoze kernel: [ 363.524162] Showing all locks held in
the system:
Mar 6 17:07:11 LaChoze kernel: [ 363.524170] 2 locks held by
khungtaskd/90:
Mar 6 17:07:11 LaChoze kernel: [ 363.524171] #0:
(rcu_read_lock){......}, at: [<ffffffff8756ef6e>] watchdog+0xae/0x640
Mar 6 17:07:11 LaChoze kernel: [ 363.524180] #1:
(tasklist_lock){+.+...}, at: [<ffffffff874eae1d>]
debug_show_all_locks+0x3d/0x1a0
Mar 6 17:07:11 LaChoze kernel: [ 363.524193] 1 lock held by in:imklog/822:
Mar 6 17:07:11 LaChoze kernel: [ 363.524194] #0:
(&f->f_pos_lock){+.+.+.}, at: [<ffffffff8768883a>] __fdget_pos+0x4a/0x50
Mar 6 17:07:11 LaChoze kernel: [ 363.524208] 2 locks held by Xorg/2365:
Mar 6 17:07:11 LaChoze kernel: [ 363.524209] #0:
(&rdev->exclusive_lock){++++.+}, at: [<ffffffff874f3710>]
rt_down_read+0x10/0x20
Mar 6 17:07:11 LaChoze kernel: [ 363.524215] #1:
(&rdev->pm.mclk_lock){++++++}, at: [<ffffffff874f3710>]
rt_down_read+0x10/0x20
Mar 6 17:07:11 LaChoze kernel: [ 363.524263] 2 locks held by G.Main/3575:
Mar 6 17:07:11 LaChoze kernel: [ 363.524264] #0:
(&rdev->pm.mclk_lock){++++++}, at: [<ffffffff874f3710>]
rt_down_read+0x10/0x20
Mar 6 17:07:11 LaChoze kernel: [ 363.524270] #1:
(reservation_ww_class_mutex){+.+.+.}, at: [<ffffffffc087b7b0>]
ttm_bo_vm_fault+0x60/0x530 [ttm]
Mar 6 17:07:11 LaChoze kernel: [ 363.524282]
Mar 6 17:07:11 LaChoze kernel: [ 363.524283]
=============================================
Mar 6 17:07:11 LaChoze kernel: [ 363.524283]
I hope this can help !
On 02/03/2017 22:10, Julia Cartwright wrote:
> On Sun, Feb 26, 2017 at 12:55:01PM +0100, Daniel Vidal wrote:
>> Hi
>>
>> The radeon driver hangs from 3.16 kernel.
>>
>> RT patch for radeon driver only disable ( preempt_disable_rt() ) and
>> reenable ( preempt_enable_rt() ) the whole RT in file radeon_display.c
>>
>> Obviously this is not enough.
>>
>> I think the radeon developers should make some modification in the
>> driver but I do not know which is the right place to indicate it
>>
>> Here my backtrace.
> Two more debugging ideas:
>
> 1. Boot with lockdep enabled, see if that gives you a meaningful
> splat.
>
> 2. Use sysrq-trigger 'd' to dump held locks to console when stuck in
> this state, in order to track down the codepath causing problems.
>
> Julia
Attachment:
signature.asc
Description: OpenPGP digital signature
