From: Matthew Brost Date: Fri, 12 Dec 2025 18:28:42 +0000 (-0800) Subject: drm/xe: Use usleep_range for accurate long-running workload timeslicing X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=80f9c601d9c4d26f00356c0a9c461650e7089273;p=thirdparty%2Flinux.git drm/xe: Use usleep_range for accurate long-running workload timeslicing msleep is not very accurate in terms of how long it actually sleeps, whereas usleep_range is precise. Replace the timeslice sleep for long-running workloads with the more accurate usleep_range to avoid jitter if the sleep period is less than 20ms. Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs") Cc: stable@vger.kernel.org Signed-off-by: Matthew Brost Reviewed-by: Thomas Hellström Link: https://patch.msgid.link/20251212182847.1683222-3-matthew.brost@intel.com (cherry picked from commit ca415c4d4c17ad676a2c8981e1fcc432221dce79) Signed-off-by: Thomas Hellström --- diff --git a/drivers/gpu/drm/xe/xe_guc_submit.c b/drivers/gpu/drm/xe/xe_guc_submit.c index 311cd047911a..f6ba2b0f074d 100644 --- a/drivers/gpu/drm/xe/xe_guc_submit.c +++ b/drivers/gpu/drm/xe/xe_guc_submit.c @@ -717,6 +717,24 @@ static bool vf_recovery(struct xe_guc *guc) return xe_gt_recovery_pending(guc_to_gt(guc)); } +static inline void relaxed_ms_sleep(unsigned int delay_ms) +{ + unsigned long min_us, max_us; + + if (!delay_ms) + return; + + if (delay_ms > 20) { + msleep(delay_ms); + return; + } + + min_us = mul_u32_u32(delay_ms, 1000); + max_us = min_us + 500; + + usleep_range(min_us, max_us); +} + static int wq_wait_for_space(struct xe_exec_queue *q, u32 wqi_size) { struct xe_guc *guc = exec_queue_to_guc(q); @@ -1587,7 +1605,7 @@ static void __guc_exec_queue_process_msg_suspend(struct xe_sched_msg *msg) since_resume_ms; if (wait_ms > 0 && q->guc->resume_time) - msleep(wait_ms); + relaxed_ms_sleep(wait_ms); set_exec_queue_suspended(q); disable_scheduling(q, false);