On Thu, Jan 30, 2025 at 01:45:22PM -0300, Maíra Canal wrote:
Hi Greg,
On 30/01/25 13:26, Greg Kroah-Hartman wrote:
On Thu, Jan 30, 2025 at 12:56:25PM -0300, Maíra Canal wrote:
Hi Greg,
This patch introduced a race-condition that was fixed in 6e64d6b3a3c39655de56682ec83e894978d23412 ("drm/v3d: Assign job pointer to NULL before signaling the fence") - already in torvalds/master. Is it possible to push the two patches together? This way we wouldn't break any devices.
As all 5.15 and newer devices are broken right now, that's not good :(
I agree, it's terrible. I'm currently waiting for stable to pick up the fix (as it is quite quick, thanks for the great work). For a even quicker solution, should I send the fix for the newer stable branches or you can queue 6e64d6b3a3c39655de56682ec83e894978d23412 directly?
I added the "Fixes:" tag to the fix and CCed stable.
You did it all right here, it's just that we can't normally take patches until they show up in a -rc release unless specifically asked to (or it's a trivial one). You hit the window where the longest delay is, not your fault.
thanks,
greg k-h