[BUGFIX] drm/i915: Fix for VBT expected size

[BUGFIX] drm/i915: Fix for VBT expected size

[Mika Kahola]

Depending on the VBT BDB version the maximum size
can be up to 38 bytes.

This fix increases the maximum of the VBT expected size
from 33 bytes to 38 bytes and by doing so cures the kernel
hang on BSW box.

Signed-off-by: Mika Kahola <mika.kahola@intel.com>
---
 drivers/gpu/drm/i915/intel_bios.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h
index f7ad6a5..788463d 100644
--- a/drivers/gpu/drm/i915/intel_bios.h
+++ b/drivers/gpu/drm/i915/intel_bios.h
@@ -246,7 +246,7 @@ struct common_child_dev_config {
 union child_device_config {
 	/* This one is safe to be used anywhere, but the code should still check
 	 * the BDB version. */
-	u8 raw[33];
+	u8 raw[38];
 	/* This one should only be kept for legacy code. */
 	struct old_child_dev_config old;
 	/* This one should also be safe to use anywhere, even without version
-- 
1.9.1

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [BUGFIX] drm/i915: Fix for VBT expected size

[Daniel Vetter]

On Tue, Aug 11, 2015 at 04:49:33PM +0300, Mika Kahola wrote:
> Depending on the VBT BDB version the maximum size
> can be up to 38 bytes.
> 
> This fix increases the maximum of the VBT expected size
> from 33 bytes to 38 bytes and by doing so cures the kernel
> hang on BSW box.
> 
> Signed-off-by: Mika Kahola <mika.kahola@intel.com>

We already have David's patch in -fixes, how does this relate? How does it
blow up? Is this a regression? If so which commit created it? Where's the
bugzilla link from QA?
-Daniel

> ---
>  drivers/gpu/drm/i915/intel_bios.h | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h
> index f7ad6a5..788463d 100644
> --- a/drivers/gpu/drm/i915/intel_bios.h
> +++ b/drivers/gpu/drm/i915/intel_bios.h
> @@ -246,7 +246,7 @@ struct common_child_dev_config {
>  union child_device_config {
>  	/* This one is safe to be used anywhere, but the code should still check
>  	 * the BDB version. */
> -	u8 raw[33];
> +	u8 raw[38];
>  	/* This one should only be kept for legacy code. */
>  	struct old_child_dev_config old;
>  	/* This one should also be safe to use anywhere, even without version
> -- 
> 1.9.1
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [BUGFIX] drm/i915: Fix for VBT expected size

[Jani Nikula]

On Wed, 12 Aug 2015, Daniel Vetter <daniel@ffwll.ch> wrote:
> On Tue, Aug 11, 2015 at 04:49:33PM +0300, Mika Kahola wrote:
>> Depending on the VBT BDB version the maximum size
>> can be up to 38 bytes.
>> 
>> This fix increases the maximum of the VBT expected size
>> from 33 bytes to 38 bytes and by doing so cures the kernel
>> hang on BSW box.
>> 
>> Signed-off-by: Mika Kahola <mika.kahola@intel.com>
>
> We already have David's patch in -fixes, how does this relate? How does it
> blow up? Is this a regression? If so which commit created it? Where's the
> bugzilla link from QA?

There's no bugzilla link from QA because Mika found the bug and I told
him to just send the patch.

BR,
Jani.


> -Daniel
>
>> ---
>>  drivers/gpu/drm/i915/intel_bios.h | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>> 
>> diff --git a/drivers/gpu/drm/i915/intel_bios.h b/drivers/gpu/drm/i915/intel_bios.h
>> index f7ad6a5..788463d 100644
>> --- a/drivers/gpu/drm/i915/intel_bios.h
>> +++ b/drivers/gpu/drm/i915/intel_bios.h
>> @@ -246,7 +246,7 @@ struct common_child_dev_config {
>>  union child_device_config {
>>  	/* This one is safe to be used anywhere, but the code should still check
>>  	 * the BDB version. */
>> -	u8 raw[33];
>> +	u8 raw[38];
>>  	/* This one should only be kept for legacy code. */
>>  	struct old_child_dev_config old;
>>  	/* This one should also be safe to use anywhere, even without version
>> -- 
>> 1.9.1
>> 
>> _______________________________________________
>> Intel-gfx mailing list
>> Intel-gfx@lists.freedesktop.org
>> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
>
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> http://blog.ffwll.ch
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx

-- 
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [BUGFIX] drm/i915: Fix for VBT expected size

[Daniel Vetter]

On Wed, Aug 12, 2015 at 04:42:53PM +0300, Jani Nikula wrote:
> On Wed, 12 Aug 2015, Daniel Vetter <daniel@ffwll.ch> wrote:
> > On Tue, Aug 11, 2015 at 04:49:33PM +0300, Mika Kahola wrote:
> >> Depending on the VBT BDB version the maximum size
> >> can be up to 38 bytes.
> >> 
> >> This fix increases the maximum of the VBT expected size
> >> from 33 bytes to 38 bytes and by doing so cures the kernel
> >> hang on BSW box.
> >> 
> >> Signed-off-by: Mika Kahola <mika.kahola@intel.com>
> >
> > We already have David's patch in -fixes, how does this relate? How does it
> > blow up? Is this a regression? If so which commit created it? Where's the
> > bugzilla link from QA?
> 
> There's no bugzilla link from QA because Mika found the bug and I told
> him to just send the patch.

so bsw doesn't boot and QA didn't notice? That's fail too, just different
kind of fail ...
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [BUGFIX] drm/i915: Fix for VBT expected size

[Mika Kahola]

On Wed, 2015-08-12 at 15:43 +0200, Daniel Vetter wrote:
> On Wed, Aug 12, 2015 at 04:42:53PM +0300, Jani Nikula wrote:
> > On Wed, 12 Aug 2015, Daniel Vetter <daniel@ffwll.ch> wrote:
> > > On Tue, Aug 11, 2015 at 04:49:33PM +0300, Mika Kahola wrote:
> > >> Depending on the VBT BDB version the maximum size
> > >> can be up to 38 bytes.
> > >> 
> > >> This fix increases the maximum of the VBT expected size
> > >> from 33 bytes to 38 bytes and by doing so cures the kernel
> > >> hang on BSW box.
> > >> 
> > >> Signed-off-by: Mika Kahola <mika.kahola@intel.com>
> > >
> > > We already have David's patch in -fixes, how does this relate? How does it
> > > blow up? Is this a regression? If so which commit created it? Where's the
> > > bugzilla link from QA?
> > 
> > There's no bugzilla link from QA because Mika found the bug and I told
> > him to just send the patch.
> 
> so bsw doesn't boot and QA didn't notice? That's fail too, just different
> kind of fail ...
> -Daniel
Well, there is now a bug report now which can be found

https://bugs.freedesktop.org/show_bug.cgi?id=91613

In addition, David refined his patch series, which relates to this issue
and therefore my bugfix can be ignored.
 
https://patchwork.freedesktop.org/patch/56911/

-Mika-

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx