[PATCH 0/3] drm/radeon: fix up some badness when probe fails

Daniel Palmer posted 3 patches 3 months, 3 weeks ago
drivers/gpu/drm/radeon/radeon_drv.c | 34 ++++++++---------------------
drivers/gpu/drm/radeon/radeon_kms.c |  1 -
2 files changed, 9 insertions(+), 26 deletions(-)
[PATCH 0/3] drm/radeon: fix up some badness when probe fails
Posted by Daniel Palmer 3 months, 3 weeks ago
I have been trying to get a Radeon 9250 running on an Amiga 4000[0].

On that setup it fails to find the BIOS and the probe fails which is
expected but then a bunch of WARN_ON()s etc are triggered.

I though maybe this is "m68k problems" so I bought an old x86 board and
there if I have a different primary VGA card the BIOS part of the
probe fails in the same way and the same scary messages[1] are showing
up in the console.

It seems like the probe failure path wasn't tested when some previous
cleaning up happened.

I'll fix the issues with not finding the BIOS if the card wasn't
initialised in the normal x86 way later.

0 - https://lore.kernel.org/lkml/20251007092313.755856-1-daniel@thingy.jp/
1 - https://gist.github.com/fifteenhex/b971bd62c49383a0558395c62c05ce3b

Daniel Palmer (3):
  drm/radeon: Clean up pdev->dev instances in probe
  drm/radeon: Do not kfree() devres managed rdev
  drm/radeon: Remove calls to drm_put_dev()

 drivers/gpu/drm/radeon/radeon_drv.c | 34 ++++++++---------------------
 drivers/gpu/drm/radeon/radeon_kms.c |  1 -
 2 files changed, 9 insertions(+), 26 deletions(-)

-- 
2.51.0
Re: [PATCH 0/3] drm/radeon: fix up some badness when probe fails
Posted by Alex Deucher 3 months, 2 weeks ago
Applied.  Thanks!

Alex

On Sat, Oct 18, 2025 at 1:51 AM Daniel Palmer <daniel@0x0f.com> wrote:
>
> I have been trying to get a Radeon 9250 running on an Amiga 4000[0].
>
> On that setup it fails to find the BIOS and the probe fails which is
> expected but then a bunch of WARN_ON()s etc are triggered.
>
> I though maybe this is "m68k problems" so I bought an old x86 board and
> there if I have a different primary VGA card the BIOS part of the
> probe fails in the same way and the same scary messages[1] are showing
> up in the console.
>
> It seems like the probe failure path wasn't tested when some previous
> cleaning up happened.
>
> I'll fix the issues with not finding the BIOS if the card wasn't
> initialised in the normal x86 way later.
>
> 0 - https://lore.kernel.org/lkml/20251007092313.755856-1-daniel@thingy.jp/
> 1 - https://gist.github.com/fifteenhex/b971bd62c49383a0558395c62c05ce3b
>
> Daniel Palmer (3):
>   drm/radeon: Clean up pdev->dev instances in probe
>   drm/radeon: Do not kfree() devres managed rdev
>   drm/radeon: Remove calls to drm_put_dev()
>
>  drivers/gpu/drm/radeon/radeon_drv.c | 34 ++++++++---------------------
>  drivers/gpu/drm/radeon/radeon_kms.c |  1 -
>  2 files changed, 9 insertions(+), 26 deletions(-)
>
> --
> 2.51.0
>