From: Tomasz Lis <tomasz.lis@intel.com>
To: intel-xe@lists.freedesktop.org
Cc: "Michał Winiarski" <michal.winiarski@intel.com>,
"Michał Wajdeczko" <michal.wajdeczko@intel.com>,
"Piotr Piórkowski" <piotr.piorkowski@intel.com>,
"Matthew Brost" <matthew.brost@intel.com>
Subject: [PATCH v1] drm/xe/ct: Separate waiting for retry from ct send function
Date: Thu, 9 Oct 2025 19:08:44 +0200 [thread overview]
Message-ID: <20251009170844.178199-1-tomasz.lis@intel.com> (raw)
The function `guc_ct_send_locked()` is really quite simple, but still
looks complex due to exposed internals. It is sending a message,
and in case of lack of space, waiting for a proper moment to send
a retry.
Clear separation of send function and wait function will help with
readability. This is a cosmetic change only, no functional difference
is expected.
This patch introduces `guc_ct_send_wait_for_retry()`, and uses it to
greatly simplify `guc_ct_send_locked()`.
Signed-off-by: Tomasz Lis <tomasz.lis@intel.com>
---
drivers/gpu/drm/xe/xe_guc_ct.c | 64 +++++++++++++++++++++-------------
1 file changed, 39 insertions(+), 25 deletions(-)
diff --git a/drivers/gpu/drm/xe/xe_guc_ct.c b/drivers/gpu/drm/xe/xe_guc_ct.c
index 3ae1e8db143a..e68953ef3a00 100644
--- a/drivers/gpu/drm/xe/xe_guc_ct.c
+++ b/drivers/gpu/drm/xe/xe_guc_ct.c
@@ -991,22 +991,15 @@ static void kick_reset(struct xe_guc_ct *ct)
static int dequeue_one_g2h(struct xe_guc_ct *ct);
-static int guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, u32 len,
- u32 g2h_len, u32 num_g2h,
- struct g2h_fence *g2h_fence)
+/*
+ * wait before retry of sending h2g message
+ * Return: true if ready for retry, false if the wait timeouted
+ */
+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)
{
struct xe_device *xe = ct_to_xe(ct);
- struct xe_gt *gt = ct_to_gt(ct);
- unsigned int sleep_period_ms = 1;
- int ret;
-
- xe_gt_assert(gt, !g2h_len || !g2h_fence);
- lockdep_assert_held(&ct->lock);
- xe_device_assert_mem_access(ct_to_xe(ct));
-
-try_again:
- ret = __guc_ct_send_locked(ct, action, len, g2h_len, num_g2h,
- g2h_fence);
/*
* We wait to try to restore credits for about 1 second before bailing.
@@ -1015,24 +1008,22 @@ static int guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, u32 len,
* the case of G2H we process any G2H in the channel, hopefully freeing
* credits as we consume the G2H messages.
*/
- if (unlikely(ret == -EBUSY &&
- !h2g_has_room(ct, len + GUC_CTB_HDR_LEN))) {
+ if (!h2g_has_room(ct, len + GUC_CTB_HDR_LEN)) {
struct guc_ctb *h2g = &ct->ctbs.h2g;
- if (sleep_period_ms == 1024)
- goto broken;
+ if (*sleep_period_ms == 1024)
+ 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;
-
- goto try_again;
- } else if (unlikely(ret == -EBUSY)) {
+ msleep(*sleep_period_ms);
+ *sleep_period_ms <<= 1;
+ } else {
struct xe_device *xe = ct_to_xe(ct);
struct guc_ctb *g2h = &ct->ctbs.g2h;
+ int ret;
trace_xe_guc_ct_g2h_flow_control(xe, g2h->info.head,
desc_read(xe, g2h, tail),
@@ -1046,7 +1037,7 @@ static int guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, u32 len,
(desc_read(ct_to_xe(ct), (&ct->ctbs.g2h), tail) != ct->ctbs.g2h.info.head)
if (!wait_event_timeout(ct->wq, !ct->g2h_outstanding ||
g2h_avail(ct), HZ))
- goto broken;
+ return false;
#undef g2h_avail
ret = dequeue_one_g2h(ct);
@@ -1054,9 +1045,32 @@ static int guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, u32 len,
if (ret != -ECANCELED)
xe_gt_err(ct_to_gt(ct), "CTB receive failed (%pe)",
ERR_PTR(ret));
- goto broken;
+ return false;
}
+ }
+ return true;
+}
+
+static int guc_ct_send_locked(struct xe_guc_ct *ct, const u32 *action, u32 len,
+ u32 g2h_len, u32 num_g2h,
+ struct g2h_fence *g2h_fence)
+{
+ struct xe_gt *gt = ct_to_gt(ct);
+ unsigned int sleep_period_ms = 1;
+ int ret;
+
+ xe_gt_assert(gt, !g2h_len || !g2h_fence);
+ lockdep_assert_held(&ct->lock);
+ xe_device_assert_mem_access(ct_to_xe(ct));
+try_again:
+ ret = __guc_ct_send_locked(ct, action, len, g2h_len, num_g2h,
+ g2h_fence);
+
+ if (unlikely(ret == -EBUSY)) {
+ if (!guc_ct_send_wait_for_retry(ct, len, g2h_len, g2h_fence,
+ &sleep_period_ms))
+ goto broken;
goto try_again;
}
--
2.25.1
next reply other threads:[~2025-10-09 17:07 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-10-09 17:08 Tomasz Lis [this message]
2025-10-09 17:14 ` ✓ CI.KUnit: success for drm/xe/ct: Separate waiting for retry from ct send function Patchwork
2025-10-09 17:27 ` [PATCH v1] " Matthew Brost
2025-10-09 17:50 ` ✓ Xe.CI.BAT: success for " Patchwork
2025-10-10 0:56 ` ✓ Xe.CI.Full: " 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=20251009170844.178199-1-tomasz.lis@intel.com \
--to=tomasz.lis@intel.com \
--cc=intel-xe@lists.freedesktop.org \
--cc=matthew.brost@intel.com \
--cc=michal.wajdeczko@intel.com \
--cc=michal.winiarski@intel.com \
--cc=piotr.piorkowski@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