Re: [PATCH] drm/i915: Don't kick-off hangcheck after a DRIinterrupt

[Herbert Xu]

On Thu, Jan 20, 2011 at 09:56:01AM +0000, Chris Wilson wrote:
> Hangcheck is only used by GEM and just OOPSes with incomplete DRI
> configuration:
> 
> BUG: unable to handle kernel paging request at fffffffffffffff0
> IP: [<ffffffffa041ee76>] i915_hangcheck_elapsed+0x96/0x270 [i915]
> PGD 13d1067 PUD 13d2067 PMD 0
> Oops: 0000 [#1] PREEMPT SMP
> last sysfs file: /sys/class/net/lo/operstate
> CPU 2
> Modules linked in: snd_pcm_oss snd_mixer_oss vmnet parport_pc parport
> vmblock vmci vmmon i915 drm_kms_helper drm fb fbdev i2c_algo_bit
> cfbcopyarea video backlight output cfbimgblt cfbfillrect autofs4 ipv6
> nfs lockd fscache nfs_acl auth_rpcgss sunrpc coretemp hwmon_vid mo]
> 
> Pid: 0, comm: kworker/0:1 Not tainted 2.6.36.2 #5 P5KPL-CM/System
> Product Name
> RIP: 0010:[<ffffffffa041ee76>]  [<ffffffffa041ee76>]
> i915_hangcheck_elapsed+0x96/0x270 [i915]
> RSP: 0000:ffff880001703e40  EFLAGS: 00010217
> RAX: 0000000000000000 RBX: ffff880117071800 RCX: ffff880118f7c400
> RDX: 000000007dffffc0 RSI: ffff880118f7c028 RDI: ffff880117071800
> RBP: ffff880001703e70 R08: ffff88000170d460 R09: ffff880001712620
> R10: 0000000000000000 R11: 0000000000000001 R12: ffff880118f7c000
> R13: ffff880117071800 R14: 0000000000000000 R15: 000000000e41e9d8
> FS:  0000000000000000(0000) GS:ffff880001700000(0000)
> knlGS:0000000000000000
> CS:  0010 DS: 0000 ES: 0000 CR0: 000000008005003b
> CR2: fffffffffffffff0 CR3: 00000000d83df000 CR4: 00000000000006e0
> DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
> DR3: 0000000000000000 DR6: 00000000ffff0ff0 DR7: 0000000000000400
> Process kworker/0:1 (pid: 0, threadinfo ffff88011b6b2000, task
> ffff88011b67d5c0)
> Stack:
>  7dffffc000012600 ffff880117071800 ffff88011b6ac000 0000000000000102
> <0> ffff880001703eb0 ffffffffa041ede0 ffff880001703ef0 ffffffff81046fad
> <0> ffff88011b6b3fd8 ffff88011b6b3fd8 ffff88011b6adc20 ffff88011b6ad820
> Call Trace:
>  <IRQ>
>  [<ffffffffa041ede0>] ? i915_hangcheck_elapsed+0x0/0x270 [i915]
>  [<ffffffff81046fad>] run_timer_softirq+0x13d/0x260
>  [<ffffffff81063657>] ? clockevents_program_event+0x57/0xa0
>  [<ffffffff81041c76>] __do_softirq+0xa6/0x130
>  [<ffffffff810032cc>] call_softirq+0x1c/0x30
>  [<ffffffff81005375>] do_softirq+0x55/0x90
>  [<ffffffff8104190d>] irq_exit+0x8d/0xb0
>  [<ffffffff8101de8c>] smp_apic_timer_interrupt+0x6c/0xa0
>  [<ffffffff81002d93>] apic_timer_interrupt+0x13/0x20
>  <EOI>
>  [<ffffffff8100b139>] ? mwait_idle+0x79/0x90
>  [<ffffffff81001610>] ? enter_idle+0x20/0x30
>  [<ffffffff81001689>] cpu_idle+0x69/0xc0
>  [<ffffffff812cb19c>] start_secondary+0x183/0x1e7
> Code: 8d 84 24 18 01 00 00 49 39 84 24 18 01 00 00 0f 84 cf 00 00 00 49
> 8b 85 68 03 00 00 49 8d 74 24 28 48 8b 80 20 01 00 00 4c 89 ef <8b> 58
> f0 e8 42 5e 00 00 89 de 89 c7 e8 29 5e 00 00 84 c0 0f 85
> RIP  [<ffffffffa041ee76>] i915_hangcheck_elapsed+0x96/0x270 [i915]
>  RSP <ffff880001703e40>
> CR2: fffffffffffffff0
> ---[ end trace a327d5ceef537f9e ]---
> 
> Reported-by: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
> Signed-off-by: Chris Wilson <chris@xxxxxxxxxxxxxxxxxx>
> ---
>  drivers/gpu/drm/i915/i915_irq.c |    6 +++++-
>  1 files changed, 5 insertions(+), 1 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index 46d649b..39ce40d 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -348,8 +348,12 @@ static void notify_ring(struct drm_device *dev,
>  			struct intel_ring_buffer *ring)
>  {
>  	struct drm_i915_private *dev_priv = dev->dev_private;
> -	u32 seqno = ring->get_seqno(ring);
> +	u32 seqno;
> +
> +	if (ring->obj == NULL)
> +		return;
>  
> +	seqno = ring->get_seqno(ring);
>  	trace_i915_gem_request_complete(dev, seqno);

While the current kernel tree has indeed changed from 2.6.36,
I don't think this is the spot corresponding to my crash.

My spot was in hangcheck_elapsed and as far as I can see it will
crash in the current kernel in pretty much the same way.  In
particular, i915_hangcheck_ring_idle will probably crash on all
three rings.

FWIW after adding the INIT_LIST_HEAD to the init_dri function
my kernel hasn't crashed yet (a couple of hours and counting).

Thanks,
-- 
Email: Herbert Xu <herbert@xxxxxxxxxxxxxxxxxxx>
Home Page: http://gondor.apana.org.au/~herbert/
PGP Key: http://gondor.apana.org.au/~herbert/pubkey.txt
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/