Hi Bjorn,
On 2/19/26 2:51 AM, Bjorn Helgaas wrote:
I'm not sure if you also fixed the kernel test robot warnings
(https://lore.kernel.org/all/[email protected],
https://lore.kernel.org/all/[email protected]).
I think I did, but the () are still missing, will do another round.
I assume this series fixes some problem where VGA isn't enabled
correctly in some topology? It wasn't obvious to me from the commit
logs, but it would be good to include that if so.
It's the missing error handling if VGA cannot be enabled because a
bridge on the way doesn't support VGA forwarding and hardwires the VGA
bit to zero.
The i915 driver has a workaround for two internal state machines that go
out of sync when voltage regulators are reprogrammed, so it needs to
perform a few VGA accesses on initialization, or there is a risk of a
bus error later when control is returned to vgacon.
This isn't a concern if VGA accesses *never* work, so we can simply skip
it in this case, but we need to know that there was an error.
The i915 counterpart to this is
https://patchwork.freedesktop.org/series/161721/
which changes code introduced in December, so if the pci/vgaarb patches
are backported, that one also needs to be backported manually.
That's what the __must_check are for, the warnings they generate go away
when the i915 patch is applied, so it probably makes sense to do that first.
So, for v3:
- expand commit messages with the text above
- recheck commit message style
Does it make sense to Cc stable on these?
Simon
