diff mbox

[i-g-t] tests/drv_missed_irq_hang: Fix gem_blt path

Message ID 1465824382-21843-1-git-send-email-mika.kuoppala@intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Mika Kuoppala June 13, 2016, 1:26 p.m. UTC
Don't add SOURCE_DIR to the path for gem_blt as if this
script is invocated on some other directory, the path to
gem_blt will be concatenated two times.

References: https://bugs.freedesktop.org/show_bug.cgi?id=88437
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
---
 tests/drv_missed_irq_hang | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

Comments

Marius Vlad June 15, 2016, 11:56 a.m. UTC | #1
On Mon, Jun 13, 2016 at 04:26:22PM +0300, Mika Kuoppala wrote:
> Don't add SOURCE_DIR to the path for gem_blt as if this
> script is invocated on some other directory, the path to
> gem_blt will be concatenated two times.
> 
> References: https://bugs.freedesktop.org/show_bug.cgi?id=88437
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
> ---
>  tests/drv_missed_irq_hang | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/tests/drv_missed_irq_hang b/tests/drv_missed_irq_hang
> index e76c7db674ac..96a653846005 100755
> --- a/tests/drv_missed_irq_hang
> +++ b/tests/drv_missed_irq_hang
> @@ -6,12 +6,12 @@
>  SOURCE_DIR="$( dirname "${BASH_SOURCE[0]}" )"
>  . $SOURCE_DIR/drm_lib.sh
>  
> -oldpath=`pwd`
> +oldpath=$PWD
>  
>  cd $i915_dfs_path
>  
>  function blt_wait {
> -	$oldpath/$SOURCE_DIR/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
> +	${oldpath}/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
But $SOURCE_DIR will always be '.'. If installed, benchmarks/ will be
underneath tests/ directory. Is that what you're trying to fix?

Posted a fix for this when running under piglit:
https://patchwork.freedesktop.org/patch/87378/
>  }
>  
>  function check_for_missed_irq {
> -- 
> 2.7.4
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx
Chris Wilson June 15, 2016, 12:28 p.m. UTC | #2
On Wed, Jun 15, 2016 at 02:56:39PM +0300, Marius Vlad wrote:
> On Mon, Jun 13, 2016 at 04:26:22PM +0300, Mika Kuoppala wrote:
> > Don't add SOURCE_DIR to the path for gem_blt as if this
> > script is invocated on some other directory, the path to
> > gem_blt will be concatenated two times.
> > 
> > References: https://bugs.freedesktop.org/show_bug.cgi?id=88437
> > Cc: Chris Wilson <chris@chris-wilson.co.uk>
> > Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
> > ---
> >  tests/drv_missed_irq_hang | 4 ++--
> >  1 file changed, 2 insertions(+), 2 deletions(-)
> > 
> > diff --git a/tests/drv_missed_irq_hang b/tests/drv_missed_irq_hang
> > index e76c7db674ac..96a653846005 100755
> > --- a/tests/drv_missed_irq_hang
> > +++ b/tests/drv_missed_irq_hang
> > @@ -6,12 +6,12 @@
> >  SOURCE_DIR="$( dirname "${BASH_SOURCE[0]}" )"
> >  . $SOURCE_DIR/drm_lib.sh
> >  
> > -oldpath=`pwd`
> > +oldpath=$PWD
> >  
> >  cd $i915_dfs_path
> >  
> >  function blt_wait {
> > -	$oldpath/$SOURCE_DIR/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
> > +	${oldpath}/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
> But $SOURCE_DIR will always be '.'. If installed, benchmarks/ will be
> underneath tests/ directory. Is that what you're trying to fix?
> 
> Posted a fix for this when running under piglit:
> https://patchwork.freedesktop.org/patch/87378/

Neither seem like a general solution for finding the testcase, for all of
the installed, not-installed, mixed-installed combinations.

Even more tricky is that we want tighter control over the batch to avoid
a fast GPU never enabling irqs and so never noticing the failed delivery
(either the operation is completed before the wait, or it will complete
within the spin).
-Chris
Mika Kuoppala June 16, 2016, 2:37 p.m. UTC | #3
Chris Wilson <chris@chris-wilson.co.uk> writes:

> On Wed, Jun 15, 2016 at 02:56:39PM +0300, Marius Vlad wrote:
>> On Mon, Jun 13, 2016 at 04:26:22PM +0300, Mika Kuoppala wrote:
>> > Don't add SOURCE_DIR to the path for gem_blt as if this
>> > script is invocated on some other directory, the path to
>> > gem_blt will be concatenated two times.
>> > 
>> > References: https://bugs.freedesktop.org/show_bug.cgi?id=88437
>> > Cc: Chris Wilson <chris@chris-wilson.co.uk>
>> > Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
>> > ---
>> >  tests/drv_missed_irq_hang | 4 ++--
>> >  1 file changed, 2 insertions(+), 2 deletions(-)
>> > 
>> > diff --git a/tests/drv_missed_irq_hang b/tests/drv_missed_irq_hang
>> > index e76c7db674ac..96a653846005 100755
>> > --- a/tests/drv_missed_irq_hang
>> > +++ b/tests/drv_missed_irq_hang
>> > @@ -6,12 +6,12 @@
>> >  SOURCE_DIR="$( dirname "${BASH_SOURCE[0]}" )"
>> >  . $SOURCE_DIR/drm_lib.sh
>> >  
>> > -oldpath=`pwd`
>> > +oldpath=$PWD
>> >  
>> >  cd $i915_dfs_path
>> >  
>> >  function blt_wait {
>> > -	$oldpath/$SOURCE_DIR/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
>> > +	${oldpath}/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
>> But $SOURCE_DIR will always be '.'. If installed, benchmarks/ will be
>> underneath tests/ directory. Is that what you're trying to fix?
>>

My stab at https://bugs.freedesktop.org/show_bug.cgi?id=88437

>> Posted a fix for this when running under piglit:
>> https://patchwork.freedesktop.org/patch/87378/
>

If it fixes the bug above, go for it.

> Neither seem like a general solution for finding the testcase, for all of
> the installed, not-installed, mixed-installed combinations.
>
> Even more tricky is that we want tighter control over the batch to avoid
> a fast GPU never enabling irqs and so never noticing the failed delivery
> (either the operation is completed before the wait, or it will complete
> within the spin).
> -Chris

How about we give up and convert that test C to gain control?

-Mika

>
> -- 
> Chris Wilson, Intel Open Source Technology Centre
Daniel Vetter June 16, 2016, 9:36 p.m. UTC | #4
On Thu, Jun 16, 2016 at 05:37:54PM +0300, Mika Kuoppala wrote:
> Chris Wilson <chris@chris-wilson.co.uk> writes:
> > Neither seem like a general solution for finding the testcase, for all of
> > the installed, not-installed, mixed-installed combinations.
> >
> > Even more tricky is that we want tighter control over the batch to avoid
> > a fast GPU never enabling irqs and so never noticing the failed delivery
> > (either the operation is completed before the wait, or it will complete
> > within the spin).
> > -Chris
> 
> How about we give up and convert that test C to gain control?

+1

Actually: +1k ;-)

Cheers, Daniel
diff mbox

Patch

diff --git a/tests/drv_missed_irq_hang b/tests/drv_missed_irq_hang
index e76c7db674ac..96a653846005 100755
--- a/tests/drv_missed_irq_hang
+++ b/tests/drv_missed_irq_hang
@@ -6,12 +6,12 @@ 
 SOURCE_DIR="$( dirname "${BASH_SOURCE[0]}" )"
 . $SOURCE_DIR/drm_lib.sh
 
-oldpath=`pwd`
+oldpath=$PWD
 
 cd $i915_dfs_path
 
 function blt_wait {
-	$oldpath/$SOURCE_DIR/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
+	${oldpath}/../benchmarks/gem_blt -r 1 -b 64 -t 1 -S > /dev/null
 }
 
 function check_for_missed_irq {