Re: [PATCH v1] drm/xe: Force EXEC_QUEUE_FLAG_KERNEL for kernel internal VMs

[Matthew Brost <matthew.brost@intel.com>]

On Wed, Feb 11, 2026 at 09:44:33AM -0800, Matthew Brost wrote:
> On Wed, Feb 11, 2026 at 06:14:41PM +0100, Piórkowski, Piotr wrote:
> > From: Piotr Piórkowski <piotr.piorkowski@intel.com>
> > 
> > VMs created without an associated xe_file originate from kernel
> > contexts and should use kernel exec queues. Ensure such VMs
> > create bind exec queues with EXEC_QUEUE_FLAG_KERNEL set.
> > Let's ensure bind exec queues created for kernel VMs are always
> > marked with EXEC_QUEUE_FLAG_KERNEL.
> > 
> > Signed-off-by: Piotr Piórkowski <piotr.piorkowski@intel.com>
> > ---
> >  drivers/gpu/drm/xe/xe_vm.c | 3 +++
> >  1 file changed, 3 insertions(+)
> > 
> > diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
> > index 8fe54a998385..708e8c7becf8 100644
> > --- a/drivers/gpu/drm/xe/xe_vm.c
> > +++ b/drivers/gpu/drm/xe/xe_vm.c
> > @@ -1657,6 +1657,9 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags, struct xe_file *xef)
> >  			if (!vm->pt_root[id])
> >  				continue;
> >  
> > +			if (!xef) /* Not from userspace */
> > +				create_flags |= EXEC_QUEUE_FLAG_KERNEL;
> 
> This LGTM, but for completness we should probably remove setting
> EXEC_QUEUE_FLAG_KERNEL at all callers too.

I misread this patch moving a little to fast... I was thinking this was
a test in exec_queue create on xef not a test xe_vm_create.

This patch is correct as is. With that: 

Reviewed-by: Matthew Brost <matthew.brost@intel.com>

> 
> Matt
> 
> > +
> >  			q = xe_exec_queue_create_bind(xe, tile, vm, create_flags, 0);
> >  			if (IS_ERR(q)) {
> >  				err = PTR_ERR(q);
> > -- 
> > 2.34.1
> >