Re: [PATCH 6.18 000/614] 6.18.2-rc1 review

[Ronald Warsow]

On 17.12.25 10:36, Greg Kroah-Hartman wrote:
> On Wed, Dec 17, 2025 at 09:27:49AM +0100, Ronald Warsow wrote:
> > On 17.12.25 06:47, Greg Kroah-Hartman wrote:
> > > On Tue, Dec 16, 2025 at 05:06:56PM +0100, Ronald Warsow wrote:
> > > > Hi
...
> Odd, as you aren't even running the driver that this commit points to,
> right?  You shouldn't be building it, so why does this show up as the
> "bad" commit id?
>
> totally confused,

well I realized I left out several steps to do bisect correct.

I hope this time it's correct:

d84236562448e634208746f0e04f725a509d4648 is the first bad commit
commit d84236562448e634208746f0e04f725a509d4648
Author: Matthew Brost <matthew.brost@xxxxxxxxx>
Date:   Fri Oct 31 16:40:45 2025 -0700

    drm/xe: Enforce correct user fence signaling order using

    [ Upstream commit adda4e855ab6409a3edaa585293f1f2069ab7299 ]

    Prevent application hangs caused by out-of-order fence signaling when
    user fences are attached. Use drm_syncobj (via dma-fence-chain) to
    guarantee that each user fence signals in order, regardless of the
    signaling order of the attached fences. Ensure user fence writebacks to
    user space occur in the correct sequence.

    v7:
     - Skip drm_syncbj create of error (CI)

    Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel 
GPUs")
    Signed-off-by: Matthew Brost <matthew.brost@xxxxxxxxx>
    Reviewed-by: Thomas Hellström <thomas.hellstrom@xxxxxxxxxxxxxxx>
    Link: 
https://patch.msgid.link/20251031234050.3043507-2-matthew.brost@xxxxxxxxx
    Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>

 drivers/gpu/drm/xe/xe_exec_queue.c | 3 +++
 1 file changed, 3 insertions(+)



> greg k-h