Re: [PATCH v1 1/5] drm/xe/pf: Split the bulk priority provisioning function

["Bernatowicz, Marcin" <marcin.bernatowicz@linux.intel.com>]

LGTM,

Reviewed-by: Marcin Bernatowicz <marcin.bernatowicz@linux.intel.com>

On 10/27/2026 11:30 AM, Piórkowski, Piotr wrote:
> From: Michal Wajdeczko <michal.wajdeczko@intel.com>
>
> During VGPU profile provisioning we will want to reuse the current
> code that converts the scheduling priority into the schedule-if-idle
> policy, while already holding the master mutex.
>
> Move that code into a static helper that expects the caller to hold
> the master mutex, and let the public function take that mutex. As a
> side effect, the schedule-if-idle policy is now updated on all GTs
> under a single lock, instead of taking the lock separately for each
> GT.
>
> Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
> Signed-off-by: Piotr Piórkowski <piotr.piorkowski@intel.com>
> ---
>   drivers/gpu/drm/xe/xe_sriov_pf_provision.c | 35 ++++++++++++++--------
>   1 file changed, 22 insertions(+), 13 deletions(-)
>
> diff --git a/drivers/gpu/drm/xe/xe_sriov_pf_provision.c b/drivers/gpu/drm/xe/xe_sriov_pf_provision.c
> index c37e422b5a5c..91670063f706 100644
> --- a/drivers/gpu/drm/xe/xe_sriov_pf_provision.c
> +++ b/drivers/gpu/drm/xe/xe_sriov_pf_provision.c
> @@ -376,18 +376,7 @@ int xe_sriov_pf_provision_query_vf_pt(struct xe_device *xe, unsigned int vfid, u
>   	return !count ? -ENODATA : 0;
>   }
>   
> -/**
> - * xe_sriov_pf_provision_bulk_apply_priority() - Change scheduling priority of all VFs and PF.
> - * @xe: the PF &xe_device
> - * @prio: scheduling priority to set
> - *
> - * Change the scheduling priority provisioning on all tiles/GTs.
> - *
> - * This function can only be called on PF.
> - *
> - * Return: 0 on success or a negative error code on failure.
> - */
> -int xe_sriov_pf_provision_bulk_apply_priority(struct xe_device *xe, u32 prio)
> +static int pf_bulk_apply_priority(struct xe_device *xe, u32 prio)
>   {
>   	bool sched_if_idle;
>   	struct xe_gt *gt;
> @@ -395,6 +384,8 @@ int xe_sriov_pf_provision_bulk_apply_priority(struct xe_device *xe, u32 prio)
>   	int result = 0;
>   	int err;
>   
> +	lockdep_assert_held(xe_sriov_pf_master_mutex(xe));
> +
>   	/*
>   	 * Currently, priority changes that involves VFs are only allowed using
>   	 * the 'sched_if_idle' policy KLV, so only LOW and NORMAL are supported.
> @@ -403,13 +394,31 @@ int xe_sriov_pf_provision_bulk_apply_priority(struct xe_device *xe, u32 prio)
>   	sched_if_idle = prio == GUC_SCHED_PRIORITY_NORMAL;
>   
>   	for_each_gt(gt, xe, id) {
> -		err = xe_gt_sriov_pf_policy_set_sched_if_idle(gt, sched_if_idle);
> +		err = xe_gt_sriov_pf_policy_set_sched_if_idle_locked(gt, sched_if_idle);
>   		result = result ?: err;
>   	}
>   
>   	return result;
>   }
>   
> +/**
> + * xe_sriov_pf_provision_bulk_apply_priority() - Change scheduling priority of all VFs and PF.
> + * @xe: the PF &xe_device
> + * @prio: scheduling priority to set
> + *
> + * Change the scheduling priority provisioning on all tiles/GTs.
> + *
> + * This function can only be called on PF.
> + *
> + * Return: 0 on success or a negative error code on failure.
> + */
> +int xe_sriov_pf_provision_bulk_apply_priority(struct xe_device *xe, u32 prio)
> +{
> +	guard(mutex)(xe_sriov_pf_master_mutex(xe));
> +
> +	return pf_bulk_apply_priority(xe, prio);
> +}
> +
>   /**
>    * xe_sriov_pf_provision_apply_vf_priority() - Change VF's scheduling priority.
>    * @xe: the PF &xe_device