Re: [PATCH 3.16.y-ckt 030/168] drm/i915: Handle failure to kick out a conflicting fb driver

Re: [PATCH 3.16.y-ckt 030/168] drm/i915: Handle failure to kick out a conflicting fb driver

[Daniel Vetter]

On Sun, Jan 11, 2015 at 10:49 PM, Ben Hutchings <ben@decadent.org.uk> wrote:
> On Mon, 2014-12-15 at 14:24 +0000, Luis Henriques wrote:
>> 3.16.7-ckt3 -stable review patch.  If anyone has any objections, please let me know.
>>
>> ------------------
>>
>> From: Chris Wilson <chris@chris-wilson.co.uk>
>>
>> commit f96de58fc7e7d3d717c7c63975c3b896c906b5e3 upstream.
>>
>> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
>> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
>
> Should this also be applied to any older stable branches?
>
> i915_kick_out_firmware_fb() was introduced in 3.6 and it has always been
> possible for the alloc_apertures() call to fail.
>
> remove_conflicting_framebuffers() has returned an error code since 3.14
> (but could silently fail before then!) so this should be applicable to
> the 3.14 stable branch too.

tbh I don't know why this patch ended up in a stable kernel, at least
I didn't find anything where we (drm/i915 maintainers) marked it as
such. And there's no bugzilla references added either. Imo the patch
doesn't qualify for stable (it's not a real-world bug afaik).
-Daniel

>
> Ben.
>
>> ---
>>  drivers/gpu/drm/i915/i915_dma.c | 18 +++++++++++++-----
>>  1 file changed, 13 insertions(+), 5 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
>> index d44344140627..0663fb4fa606 100644
>> --- a/drivers/gpu/drm/i915/i915_dma.c
>> +++ b/drivers/gpu/drm/i915/i915_dma.c
>> @@ -1425,15 +1425,16 @@ void i915_master_destroy(struct drm_device *dev, struct drm_master *master)
>>  }
>>
>>  #if IS_ENABLED(CONFIG_FB)
>> -static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
>> +static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
>>  {
>>       struct apertures_struct *ap;
>>       struct pci_dev *pdev = dev_priv->dev->pdev;
>>       bool primary;
>> +     int ret;
>>
>>       ap = alloc_apertures(1);
>>       if (!ap)
>> -             return;
>> +             return -ENOMEM;
>>
>>       ap->ranges[0].base = dev_priv->gtt.mappable_base;
>>       ap->ranges[0].size = dev_priv->gtt.mappable_end;
>> @@ -1441,13 +1442,16 @@ static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
>>       primary =
>>               pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
>>
>> -     remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
>> +     ret = remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
>>
>>       kfree(ap);
>> +
>> +     return ret;
>>  }
>>  #else
>> -static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
>> +static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
>>  {
>> +     return 0;
>>  }
>>  #endif
>>
>> @@ -1664,7 +1668,11 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
>>                       goto out_gtt;
>>               }
>>
>> -             i915_kick_out_firmware_fb(dev_priv);
>> +             ret = i915_kick_out_firmware_fb(dev_priv);
>> +             if (ret) {
>> +                     DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
>> +                     goto out_gtt;
>> +             }
>>       }
>>
>>       pci_set_master(dev->pdev);
>
> --
> Ben Hutchings
> One of the nice things about standards is that there are so many of them.



-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 3.16.y-ckt 030/168] drm/i915: Handle failure to kick out a conflicting fb driver

[Ben Hutchings]

[-- Attachment #1: Type: text/plain, Size: 1793 bytes --]

On Mon, 2015-01-12 at 18:20 +0100, Daniel Vetter wrote:
> On Sun, Jan 11, 2015 at 10:49 PM, Ben Hutchings <ben@decadent.org.uk> wrote:
> > On Mon, 2014-12-15 at 14:24 +0000, Luis Henriques wrote:
> >> 3.16.7-ckt3 -stable review patch.  If anyone has any objections, please let me know.
> >>
> >> ------------------
> >>
> >> From: Chris Wilson <chris@chris-wilson.co.uk>
> >>
> >> commit f96de58fc7e7d3d717c7c63975c3b896c906b5e3 upstream.
> >>
> >> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> >> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
> >> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> >> Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
> >
> > Should this also be applied to any older stable branches?
> >
> > i915_kick_out_firmware_fb() was introduced in 3.6 and it has always been
> > possible for the alloc_apertures() call to fail.
> >
> > remove_conflicting_framebuffers() has returned an error code since 3.14
> > (but could silently fail before then!) so this should be applicable to
> > the 3.14 stable branch too.
> 
> tbh I don't know why this patch ended up in a stable kernel, at least
> I didn't find anything where we (drm/i915 maintainers) marked it as
> such. And there's no bugzilla references added either. Imo the patch
> doesn't qualify for stable (it's not a real-world bug afaik).
[...]

It seems to be a dependency of:

commit 0485c9dc24ec0939b42ca5104c0373297506b555
Author: Daniel Vetter <daniel.vetter@ffwll.ch>
Date:   Fri Nov 14 10:09:49 2014 +0100

    drm/i915: Kick fbdev before vgacon

which was requested for 3.16+.

So, neither of these is needed for earlier versions.  Sorry for the
noise.

Ben.

-- 
Ben Hutchings
The world is coming to an end.	Please log off.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 811 bytes --]

Re: [PATCH 3.16.y-ckt 030/168] drm/i915: Handle failure to kick out a conflicting fb driver

[Luis Henriques]

On Mon, Jan 12, 2015 at 06:20:22PM +0100, Daniel Vetter wrote:
> On Sun, Jan 11, 2015 at 10:49 PM, Ben Hutchings <ben@decadent.org.uk> wrote:
> > On Mon, 2014-12-15 at 14:24 +0000, Luis Henriques wrote:
> >> 3.16.7-ckt3 -stable review patch.  If anyone has any objections, please let me know.
> >>
> >> ------------------
> >>
> >> From: Chris Wilson <chris@chris-wilson.co.uk>
> >>
> >> commit f96de58fc7e7d3d717c7c63975c3b896c906b5e3 upstream.
> >>
> >> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> >> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
> >> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> >> Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
> >
> > Should this also be applied to any older stable branches?
> >
> > i915_kick_out_firmware_fb() was introduced in 3.6 and it has always been
> > possible for the alloc_apertures() call to fail.
> >
> > remove_conflicting_framebuffers() has returned an error code since 3.14
> > (but could silently fail before then!) so this should be applicable to
> > the 3.14 stable branch too.
> 
> tbh I don't know why this patch ended up in a stable kernel, at least
> I didn't find anything where we (drm/i915 maintainers) marked it as
> such. And there's no bugzilla references added either. Imo the patch
> doesn't qualify for stable (it's not a real-world bug afaik).

You're right, this patch was not tagged for stable.

While applying 0485c9dc24ec ("drm/i915: Kick fbdev before vgacon") to
the 3.16 kernel, I found that cherry-picking f96de58fc7e7 would make
it apply cleanly and it didn't shock me to have it in a stable kernel.
That's why I applied it for release 3.16.7-ckt3.  I guess I should
have added a note in the commit text justifying its presence.

Cheers,
--
Luís

> -Daniel
> 
> >
> > Ben.
> >
> >> ---
> >>  drivers/gpu/drm/i915/i915_dma.c | 18 +++++++++++++-----
> >>  1 file changed, 13 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/i915/i915_dma.c b/drivers/gpu/drm/i915/i915_dma.c
> >> index d44344140627..0663fb4fa606 100644
> >> --- a/drivers/gpu/drm/i915/i915_dma.c
> >> +++ b/drivers/gpu/drm/i915/i915_dma.c
> >> @@ -1425,15 +1425,16 @@ void i915_master_destroy(struct drm_device *dev, struct drm_master *master)
> >>  }
> >>
> >>  #if IS_ENABLED(CONFIG_FB)
> >> -static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
> >> +static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
> >>  {
> >>       struct apertures_struct *ap;
> >>       struct pci_dev *pdev = dev_priv->dev->pdev;
> >>       bool primary;
> >> +     int ret;
> >>
> >>       ap = alloc_apertures(1);
> >>       if (!ap)
> >> -             return;
> >> +             return -ENOMEM;
> >>
> >>       ap->ranges[0].base = dev_priv->gtt.mappable_base;
> >>       ap->ranges[0].size = dev_priv->gtt.mappable_end;
> >> @@ -1441,13 +1442,16 @@ static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
> >>       primary =
> >>               pdev->resource[PCI_ROM_RESOURCE].flags & IORESOURCE_ROM_SHADOW;
> >>
> >> -     remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
> >> +     ret = remove_conflicting_framebuffers(ap, "inteldrmfb", primary);
> >>
> >>       kfree(ap);
> >> +
> >> +     return ret;
> >>  }
> >>  #else
> >> -static void i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
> >> +static int i915_kick_out_firmware_fb(struct drm_i915_private *dev_priv)
> >>  {
> >> +     return 0;
> >>  }
> >>  #endif
> >>
> >> @@ -1664,7 +1668,11 @@ int i915_driver_load(struct drm_device *dev, unsigned long flags)
> >>                       goto out_gtt;
> >>               }
> >>
> >> -             i915_kick_out_firmware_fb(dev_priv);
> >> +             ret = i915_kick_out_firmware_fb(dev_priv);
> >> +             if (ret) {
> >> +                     DRM_ERROR("failed to remove conflicting framebuffer drivers\n");
> >> +                     goto out_gtt;
> >> +             }
> >>       }
> >>
> >>       pci_set_master(dev->pdev);
> >
> > --
> > Ben Hutchings
> > One of the nice things about standards is that there are so many of them.
> 
> 
> 
> -- 
> Daniel Vetter
> Software Engineer, Intel Corporation
> +41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx

Re: [PATCH 3.16.y-ckt 030/168] drm/i915: Handle failure to kick out a conflicting fb driver

[Daniel Vetter]

On Mon, Jan 12, 2015 at 6:43 PM, Luis Henriques
<luis.henriques@canonical.com> wrote:
> On Mon, Jan 12, 2015 at 06:20:22PM +0100, Daniel Vetter wrote:
>> On Sun, Jan 11, 2015 at 10:49 PM, Ben Hutchings <ben@decadent.org.uk> wrote:
>> > On Mon, 2014-12-15 at 14:24 +0000, Luis Henriques wrote:
>> >> 3.16.7-ckt3 -stable review patch.  If anyone has any objections, please let me know.
>> >>
>> >> ------------------
>> >>
>> >> From: Chris Wilson <chris@chris-wilson.co.uk>
>> >>
>> >> commit f96de58fc7e7d3d717c7c63975c3b896c906b5e3 upstream.
>> >>
>> >> Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
>> >> Reviewed-by: Jani Nikula <jani.nikula@intel.com>
>> >> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
>> >> Signed-off-by: Luis Henriques <luis.henriques@canonical.com>
>> >
>> > Should this also be applied to any older stable branches?
>> >
>> > i915_kick_out_firmware_fb() was introduced in 3.6 and it has always been
>> > possible for the alloc_apertures() call to fail.
>> >
>> > remove_conflicting_framebuffers() has returned an error code since 3.14
>> > (but could silently fail before then!) so this should be applicable to
>> > the 3.14 stable branch too.
>>
>> tbh I don't know why this patch ended up in a stable kernel, at least
>> I didn't find anything where we (drm/i915 maintainers) marked it as
>> such. And there's no bugzilla references added either. Imo the patch
>> doesn't qualify for stable (it's not a real-world bug afaik).
>
> You're right, this patch was not tagged for stable.
>
> While applying 0485c9dc24ec ("drm/i915: Kick fbdev before vgacon") to
> the 3.16 kernel, I found that cherry-picking f96de58fc7e7 would make
> it apply cleanly and it didn't shock me to have it in a stable kernel.
> That's why I applied it for release 3.16.7-ckt3.  I guess I should
> have added a note in the commit text justifying its presence.

Ah makes sense. Usually we try to tag cc: stable patches with their
depencies (and I looked for that but didn't find anything). As a rule
I only scan stable patches from Greg's kernels since there's way too
much going on (and this time around too much vacation and conferences
on top).
-Daniel
-- 
Daniel Vetter
Software Engineer, Intel Corporation
+41 (0) 79 365 57 48 - http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/intel-gfx