From: Dave Airlie <airlied@xxxxxxxxxx> commit 1f9910b41c857a892b83801feebdc7bdf38c5985 upstream. The fence sync logic doesn't handle a fence sync across devices as it tries to write to a channel offset from one device into the fence bo from a different device, which won't work so well. This patch fixes that to avoid using the sync path in the case where the fences come from different nouveau drm devices. This works fine on a single device as the fence bo is shared across the devices, and mapped into each channels vma space, the channel offsets are therefore okay to pass between sides, so one channel can sync on the seqnos from the other by using the offset into it's vma. Signed-off-by: Dave Airlie <airlied@xxxxxxxxxx> Cc: stable@xxxxxxxxxxxxxxx Reviewed-by: Ben Skeggs <bskeggs@xxxxxxxxxx> [ Fix compilation issue; remove version log from commit messsage. - Danilo ] Signed-off-by: Danilo Krummrich <dakr@xxxxxxxxxx> Link: https://patchwork.freedesktop.org/patch/msgid/20250109005553.623947-1-airlied@xxxxxxxxx Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- drivers/gpu/drm/nouveau/nouveau_fence.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) --- a/drivers/gpu/drm/nouveau/nouveau_fence.c +++ b/drivers/gpu/drm/nouveau/nouveau_fence.c @@ -386,11 +386,13 @@ nouveau_fence_sync(struct nouveau_bo *nv if (f) { struct nouveau_channel *prev; bool must_wait = true; + bool local; rcu_read_lock(); prev = rcu_dereference(f->channel); - if (prev && (prev == chan || - fctx->sync(f, prev, chan) == 0)) + local = prev && prev->cli->drm == chan->cli->drm; + if (local && (prev == chan || + fctx->sync(f, prev, chan) == 0)) must_wait = false; rcu_read_unlock(); if (!must_wait) Patches currently in stable-queue which might be from airlied@xxxxxxxxxx are queue-6.6/nouveau-fence-handle-cross-device-fences-properly.patch
