[PATCH] drm/i915: tune down hangcheck noise

[PATCH] drm/i915: tune down hangcheck noise

[Daniel Vetter]

We already have a big splashing *ERROR* for all the relevant cases of
hangs, so this one here is redudant. And it results in an unclean
dmesg when running with simulated hangs. Regression has been
introduced in

commit 05407ff889ceebe383aa5907219f86582ef96b72
Author: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Date:   Thu May 30 09:04:29 2013 +0300

    drm/i915: detect hang using per ring hangcheck_score

Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68641
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
---
 drivers/gpu/drm/i915/i915_irq.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
index a03b445..4c6853f 100644
--- a/drivers/gpu/drm/i915/i915_irq.c
+++ b/drivers/gpu/drm/i915/i915_irq.c
@@ -2027,9 +2027,9 @@ static void i915_hangcheck_elapsed(unsigned long data)
 
 	for_each_ring(ring, dev_priv, i) {
 		if (ring->hangcheck.score > FIRE) {
-			DRM_ERROR("%s on %s\n",
-				  stuck[i] ? "stuck" : "no progress",
-				  ring->name);
+			DRM_INFO("%s on %s\n",
+				 stuck[i] ? "stuck" : "no progress",
+				 ring->name);
 			rings_hung++;
 		}
 	}
-- 
1.8.4.rc3

Re: [PATCH] drm/i915: tune down hangcheck noise

[Paulo Zanoni]

Hi

2013/8/28 Daniel Vetter <daniel.vetter@ffwll.ch>:
> We already have a big splashing *ERROR* for all the relevant cases of
> hangs, so this one here is redudant. And it results in an unclean
> dmesg when running with simulated hangs. Regression has been
> introduced in

Bug https://bugs.freedesktop.org/show_bug.cgi?id=69747 has a case
where despite the messages, there's no splashing *ERROR* (except for
an unclaimed-register message which should be ignored in this specific
case since it can be fixed without fixing the GPU hang). As a
consequence, QA reported the bug mentioning just the "unclaimed
regsiter" message, not the real GPU hang, while the unclaimed register
is just a consequence of the real GPU hang. ZZ_check_dmesg wouldn't
catch this specific bug if we didn't have the unclaimed register
problem. If the same bug happens on IVB we'll never know because we
won't have the "unclaimed register" message to point it.

>
> commit 05407ff889ceebe383aa5907219f86582ef96b72
> Author: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> Date:   Thu May 30 09:04:29 2013 +0300
>
>     drm/i915: detect hang using per ring hangcheck_score
>
> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68641
> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> ---
>  drivers/gpu/drm/i915/i915_irq.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index a03b445..4c6853f 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -2027,9 +2027,9 @@ static void i915_hangcheck_elapsed(unsigned long data)
>
>         for_each_ring(ring, dev_priv, i) {
>                 if (ring->hangcheck.score > FIRE) {
> -                       DRM_ERROR("%s on %s\n",
> -                                 stuck[i] ? "stuck" : "no progress",
> -                                 ring->name);
> +                       DRM_INFO("%s on %s\n",
> +                                stuck[i] ? "stuck" : "no progress",
> +                                ring->name);
>                         rings_hung++;
>                 }
>         }
> --
> 1.8.4.rc3
>
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx



-- 
Paulo Zanoni

Re: [PATCH] drm/i915: tune down hangcheck noise

[Daniel Vetter]

On Thu, Oct 31, 2013 at 9:14 PM, Paulo Zanoni <przanoni@gmail.com> wrote:
> 2013/8/28 Daniel Vetter <daniel.vetter@ffwll.ch>:
>> We already have a big splashing *ERROR* for all the relevant cases of
>> hangs, so this one here is redudant. And it results in an unclean
>> dmesg when running with simulated hangs. Regression has been
>> introduced in
>
> Bug https://bugs.freedesktop.org/show_bug.cgi?id=69747 has a case
> where despite the messages, there's no splashing *ERROR* (except for
> an unclaimed-register message which should be ignored in this specific
> case since it can be fixed without fixing the GPU hang). As a
> consequence, QA reported the bug mentioning just the "unclaimed
> regsiter" message, not the real GPU hang, while the unclaimed register
> is just a consequence of the real GPU hang. ZZ_check_dmesg wouldn't
> catch this specific bug if we didn't have the unclaimed register
> problem. If the same bug happens on IVB we'll never know because we
> won't have the "unclaimed register" message to point it.

This is just a testcase that uses a simulated gpu hang, so the hang
message is tuned down from ERROR to the info level. So the only real
bug is indeed the unclaimed register writes.
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch

Re: [PATCH] drm/i915: tune down hangcheck noise

[Daniel Vetter]

On Wed, Aug 28, 2013 at 10:59:42AM +0200, Daniel Vetter wrote:
> We already have a big splashing *ERROR* for all the relevant cases of
> hangs, so this one here is redudant. And it results in an unclean
> dmesg when running with simulated hangs. Regression has been
> introduced in
> 
> commit 05407ff889ceebe383aa5907219f86582ef96b72
> Author: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> Date:   Thu May 30 09:04:29 2013 +0300
> 
>     drm/i915: detect hang using per ring hangcheck_score
> 
> Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=68641
> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>

Tested-by: lu hua <huax.lu@intel.com>
> ---
>  drivers/gpu/drm/i915/i915_irq.c | 6 +++---
>  1 file changed, 3 insertions(+), 3 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_irq.c b/drivers/gpu/drm/i915/i915_irq.c
> index a03b445..4c6853f 100644
> --- a/drivers/gpu/drm/i915/i915_irq.c
> +++ b/drivers/gpu/drm/i915/i915_irq.c
> @@ -2027,9 +2027,9 @@ static void i915_hangcheck_elapsed(unsigned long data)
>  
>  	for_each_ring(ring, dev_priv, i) {
>  		if (ring->hangcheck.score > FIRE) {
> -			DRM_ERROR("%s on %s\n",
> -				  stuck[i] ? "stuck" : "no progress",
> -				  ring->name);
> +			DRM_INFO("%s on %s\n",
> +				 stuck[i] ? "stuck" : "no progress",
> +				 ring->name);
>  			rings_hung++;
>  		}
>  	}
> -- 
> 1.8.4.rc3
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch