Intel-XE Archive on lore.kernel.org
 help / color / mirror / Atom feed
From: Matthew Brost <matthew.brost@intel.com>
To: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: <intel-xe@lists.freedesktop.org>
Subject: Re: [PATCH v2 3/6] drm/xe/guc: Limit sleep while waiting for H2G credits
Date: Tue, 27 Jan 2026 11:51:45 -0800	[thread overview]
Message-ID: <aXkXUbcvIEcl8mO+@lstrano-desk.jf.intel.com> (raw)
In-Reply-To: <20260127193727.601-4-michal.wajdeczko@intel.com>

On Tue, Jan 27, 2026 at 08:37:23PM +0100, Michal Wajdeczko wrote:
> Instead of endlessly increasing the sleep timeout while waiting
> for the H2G credits, use exponential increase only up to the given
> limit, like it was initially done in the GuC submission code.
> 
> While here, fix the actual timeout to the 1s as it was documented.

Yes, the existing wait is 2s which seems to be overkill.

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

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

> Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_guc_ct.c | 12 +++++++-----
>  1 file changed, 7 insertions(+), 5 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c b/drivers/gpu/drm/xe/xe_guc_ct.c
> index dfbf76037b04..d4111124cd5f 100644
> --- a/drivers/gpu/drm/xe/xe_guc_ct.c
> +++ b/drivers/gpu/drm/xe/xe_guc_ct.c
> @@ -32,6 +32,7 @@
>  #include "xe_guc_tlb_inval.h"
>  #include "xe_map.h"
>  #include "xe_pm.h"
> +#include "xe_sleep.h"
>  #include "xe_sriov_vf.h"
>  #include "xe_trace_guc.h"
>  
> @@ -1101,7 +1102,8 @@ static int dequeue_one_g2h(struct xe_guc_ct *ct);
>   */
>  static bool guc_ct_send_wait_for_retry(struct xe_guc_ct *ct, u32 len,
>  				       u32 g2h_len, struct g2h_fence *g2h_fence,
> -				       unsigned int *sleep_period_ms)
> +				       unsigned int *sleep_period_ms,
> +				       unsigned int *sleep_total_ms)
>  {
>  	struct xe_device *xe = ct_to_xe(ct);
>  
> @@ -1115,15 +1117,14 @@ static bool guc_ct_send_wait_for_retry(struct xe_guc_ct *ct, u32 len,
>  	if (!h2g_has_room(ct, len + GUC_CTB_HDR_LEN)) {
>  		struct guc_ctb *h2g = &ct->ctbs.h2g;
>  
> -		if (*sleep_period_ms == 1024)
> +		if (*sleep_total_ms > 1000)
>  			return false;
>  
>  		trace_xe_guc_ct_h2g_flow_control(xe, h2g->info.head, h2g->info.tail,
>  						 h2g->info.size,
>  						 h2g->info.space,
>  						 len + GUC_CTB_HDR_LEN);
> -		msleep(*sleep_period_ms);
> -		*sleep_period_ms <<= 1;
> +		xe_sleep_exponential_ms(sleep_period_ms, 64);
>  	} else {
>  		struct xe_device *xe = ct_to_xe(ct);
>  		struct guc_ctb *g2h = &ct->ctbs.g2h;
> @@ -1161,6 +1162,7 @@ static int guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, u32 len,
>  {
>  	struct xe_gt *gt = ct_to_gt(ct);
>  	unsigned int sleep_period_ms = 1;
> +	unsigned int sleep_total_ms = 0;
>  	int ret;
>  
>  	xe_gt_assert(gt, !g2h_len || !g2h_fence);
> @@ -1173,7 +1175,7 @@ static int guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, u32 len,
>  
>  	if (unlikely(ret == -EBUSY)) {
>  		if (!guc_ct_send_wait_for_retry(ct, len, g2h_len, g2h_fence,
> -						&sleep_period_ms))
> +						&sleep_period_ms, &sleep_total_ms))
>  			goto broken;
>  		goto try_again;
>  	}
> -- 
> 2.47.1
> 

  reply	other threads:[~2026-01-27 19:51 UTC|newest]

Thread overview: 21+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2026-01-27 19:37 [PATCH v2 0/6] drm/xe/guc: Make MMIO communication more robust Michal Wajdeczko
2026-01-27 19:37 ` [PATCH v2 1/6] drm/xe: Promote relaxed_ms_sleep Michal Wajdeczko
2026-01-27 19:46   ` Matthew Brost
2026-01-27 19:37 ` [PATCH v2 2/6] drm/xe: Move exponential sleep logic to helper Michal Wajdeczko
2026-01-27 19:49   ` Matthew Brost
2026-01-27 19:37 ` [PATCH v2 3/6] drm/xe/guc: Limit sleep while waiting for H2G credits Michal Wajdeczko
2026-01-27 19:51   ` Matthew Brost [this message]
2026-01-27 19:37 ` [PATCH v2 4/6] drm/xe/guc: Drop redundant register read Michal Wajdeczko
2026-01-27 19:37 ` [PATCH v2 5/6] drm/xe/guc: Wait before retrying sending H2G Michal Wajdeczko
2026-01-27 19:56   ` Matthew Brost
2026-01-27 19:37 ` [PATCH v2 6/6] drm/xe/guc: Allow second H2G retry on FLR Michal Wajdeczko
2026-01-27 19:57   ` Matthew Brost
2026-01-27 19:47 ` ✗ CI.checkpatch: warning for drm/xe/guc: Make MMIO communication more robust (rev2) Patchwork
2026-01-27 19:48 ` ✓ CI.KUnit: success " Patchwork
2026-01-27 20:21 ` ✓ Xe.CI.BAT: " Patchwork
2026-01-28  2:54 ` ✗ Xe.CI.Full: failure " Patchwork
2026-01-29 18:57   ` Michal Wajdeczko
2026-02-02 21:34     ` Michal Wajdeczko
2026-01-29 20:23 ` ✗ CI.checkpatch: warning for drm/xe/guc: Make MMIO communication more robust (rev3) Patchwork
2026-01-29 20:25 ` ✓ CI.KUnit: success " Patchwork
2026-01-29 21:01 ` ✓ Xe.CI.BAT: " Patchwork

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=aXkXUbcvIEcl8mO+@lstrano-desk.jf.intel.com \
    --to=matthew.brost@intel.com \
    --cc=intel-xe@lists.freedesktop.org \
    --cc=michal.wajdeczko@intel.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox