Re: [PATCH v2] drm/xe/guc: Guard page-fault ack with runtime PM check

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

On Fri, Sep 04, 2026 at 01:35:06PM +0530, Varun Gupta wrote:
> During VM teardown, the VM's runtime PM reference is dropped
> asynchronously, allowing the device to autosuspend while stale page
> faults belonging to the now-dead VM are still queued. When the
> page-fault worker later tries to ack one of these, it calls into
> guc_ct_send_locked() on an already-suspended device, tripping:
> 
>   Assertion `!xe_pm_runtime_suspended(xe)` failed!
>   WARNING at xe_device.c:1267 xe_device_assert_mem_access+0x11c/0x140 [xe]
> 
> A live VM/exec queue always holds a PM reference while it has
> outstanding work, so if the device is suspended at ack time, the
> owning context is already gone and the fault is stale.
> 
> Take a runtime PM reference across the entire page-fault ack batch
> preventing mid-batch suspends.
> 
> v2:
>  - Hold PM ref across the entire batch (begin/end) instead of per-ack.
>    This prevents the device from autosuspending mid-batch, which would
>    leave write_only acks written but the end flush skipped, and skip
>    counter++, desyncing the cadence check.(Himal)
>  - Add a comment explaining stale faults.(Himal)
> 

I'd guard the entire page fault fault handler with PM a ref rather than
the ack part.

e.g. At the very top of xe_pagefault_queue_work:

guard(xe_pm_runtime)(xe);

Matt

> Fixes: f289f7807119 ("drm/xe: Add xe_guc_pagefault layer")
> Reported-by: Szymon Markiewicz <szymon.markiewicz@intel.com>
> Signed-off-by: Varun Gupta <varun.gupta@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_guc_pagefault.c | 12 ++++++++++++
>  1 file changed, 12 insertions(+)
> 
> diff --git a/drivers/gpu/drm/xe/xe_guc_pagefault.c b/drivers/gpu/drm/xe/xe_guc_pagefault.c
> index 8f8210a732e9..9c9cd6e056fb 100644
> --- a/drivers/gpu/drm/xe/xe_guc_pagefault.c
> +++ b/drivers/gpu/drm/xe/xe_guc_pagefault.c
> @@ -9,12 +9,21 @@
>  #include "xe_guc_pagefault.h"
>  #include "xe_pagefault.h"
>  #include "xe_pagefault_types.h"
> +#include "xe_pm.h"
>  
>  #define XE_GUC_PAGEFAULT_FLUSH_PERIOD	BIT(4)	/* Sixteen */
>  
>  static void guc_ack_fault_begin(void *private)
>  {
>  	struct xe_guc *guc = private;
> +	struct xe_device *xe = guc_to_xe(guc);
> +
> +	/*
> +	 * Live VMs hold a PM ref, so faults during suspend are stale.
> +	 * Hold a PM ref across the entire batch to safely drain them
> +	 * and prevent mid-batch autosuspend from desyncing CT flushes.
> +	 */
> +	xe_pm_runtime_get(xe);
>  
>  	xe_guc_ct_lock(&guc->ct);
>  
> @@ -62,10 +71,13 @@ static void guc_ack_fault(struct xe_pagefault *pf, int err)
>  static void guc_ack_fault_end(void *private)
>  {
>  	struct xe_guc *guc = private;
> +	struct xe_device *xe = guc_to_xe(guc);
>  
>  	if ((guc->pagefault_ack_counter & (XE_GUC_PAGEFAULT_FLUSH_PERIOD - 1)) != 1)
>  		xe_guc_ct_send_flush(&guc->ct);
>  	xe_guc_ct_unlock(&guc->ct);
> +
> +	xe_pm_runtime_put(xe);
>  }
>  
>  static const struct xe_pagefault_ops guc_pagefault_ops = {
> -- 
> 2.43.0
>