drm/xe/vf: Defer fixups if migrated twice fast

If another VF migration happened during post-migration recovery,
then the current worker should be finished to allow the next
one start swiftly and cleanly.

Check for defer in two places: before fixups, and before
sending RESFIX_DONE.

Signed-off-by: Tomasz Lis <tomasz.lis@intel.com>
Reviewed-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20241104213449.1455694-6-tomasz.lis@intel.com
Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>

diff --git a/drivers/gpu/drm/xe/xe_sriov_vf.c b/drivers/gpu/drm/xe/xe_sriov_vf.c
index ad29bd8..c1275e6 100644 (file)
--- a/drivers/gpu/drm/xe/xe_sriov_vf.c
+++ b/drivers/gpu/drm/xe/xe_sriov_vf.c
@@ -157,6 +157,19 @@ static int vf_post_migration_requery_guc(struct xe_device *xe)
        return ret;
 }
 
+/*
+ * vf_post_migration_imminent - Check if post-restore recovery is coming.
+ * @xe: the &xe_device struct instance
+ *
+ * Return: True if migration recovery worker will soon be running. Any worker currently
+ * executing does not affect the result.
+ */
+static bool vf_post_migration_imminent(struct xe_device *xe)
+{
+       return xe->sriov.vf.migration.gt_flags != 0 ||
+       work_pending(&xe->sriov.vf.migration.worker);
+}
+
 /*
  * Notify all GuCs about resource fixups apply finished.
  */
@@ -166,8 +179,14 @@ static void vf_post_migration_notify_resfix_done(struct xe_device *xe)
        unsigned int id;
 
        for_each_gt(gt, xe, id) {
+               if (vf_post_migration_imminent(xe))
+                       goto skip;
                xe_gt_sriov_vf_notify_resfix_done(gt);
        }
+       return;
+
+skip:
+       drm_dbg(&xe->drm, "another recovery imminent, skipping notifications\n");
 }
 
 static void vf_post_migration_recovery(struct xe_device *xe)
@@ -177,6 +196,8 @@ static void vf_post_migration_recovery(struct xe_device *xe)
        drm_dbg(&xe->drm, "migration recovery in progress\n");
        xe_pm_runtime_get(xe);
        err = vf_post_migration_requery_guc(xe);
+       if (vf_post_migration_imminent(xe))
+               goto defer;
        if (unlikely(err))
                goto fail;
 
@@ -185,6 +206,10 @@ static void vf_post_migration_recovery(struct xe_device *xe)
        xe_pm_runtime_put(xe);
        drm_notice(&xe->drm, "migration recovery ended\n");
        return;
+defer:
+       xe_pm_runtime_put(xe);
+       drm_dbg(&xe->drm, "migration recovery deferred\n");
+       return;
 fail:
        xe_pm_runtime_put(xe);
        drm_err(&xe->drm, "migration recovery failed (%pe)\n", ERR_PTR(err));