From: <paul.gortmaker@windriver.com>
To: Clark Williams <williams@redhat.com>,
Joseph Salisbury <joseph.salisbury@canonical.com>
Cc: <linux-rt-users@vger.kernel.org>,
Sebastian Andrzej Siewior <bigeasy@linutronix.de>,
Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Subject: [PATCH 1/1] drm/i915: Do not disable preemption for resets
Date: Fri, 18 Aug 2023 22:45:25 -0400 [thread overview]
Message-ID: <20230819024525.2056048-2-paul.gortmaker@windriver.com> (raw)
In-Reply-To: <20230819024525.2056048-1-paul.gortmaker@windriver.com>
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
[commit 40cd2835ced288789a685aa4aa7bc04b492dcd45 in linux-rt-devel]
Commit ade8a0f59844 ("drm/i915: Make all GPU resets atomic") added a
preempt disable section over the hardware reset callback to prepare the
driver for being able to reset from atomic contexts.
In retrospect I can see that the work item at a time was about removing
the struct mutex from the reset path. Code base also briefly entertained
the idea of doing the reset under stop_machine in order to serialize
userspace mmap and temporary glitch in the fence registers (see
eb8d0f5af4ec ("drm/i915: Remove GPU reset dependence on struct_mutex"),
but that never materialized and was soon removed in 2caffbf11762
("drm/i915: Revoke mmaps and prevent access to fence registers across
reset") and replaced with a SRCU based solution.
As such, as far as I can see, today we still have a requirement that
resets must not sleep (invoked from submission tasklets), but no need to
support invoking them from a truly atomic context.
Given that the preemption section is problematic on RT kernels, since the
uncore lock becomes a sleeping lock and so is invalid in such section,
lets try and remove it. Potential downside is that our short waits on GPU
to complete the reset may get extended if CPU scheduling interferes, but
in practice that probably isn't a deal breaker.
In terms of mechanics, since the preemption disabled block is being
removed we just need to replace a few of the wait_for_atomic macros into
busy looping versions which will work (and not complain) when called from
non-atomic sections.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris.p.wilson@intel.com>
Cc: Paul Gortmaker <paul.gortmaker@windriver.com>
Cc: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Link: https://lore.kernel.org/r/20230705093025.3689748-1-tvrtko.ursulin@linux.intel.com
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
[PG: backport from v6.4-rt ; minor context fixup caused by b7d70b8b06ed]
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com>
---
drivers/gpu/drm/i915/gt/intel_reset.c | 12 +++++-------
1 file changed, 5 insertions(+), 7 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_reset.c b/drivers/gpu/drm/i915/gt/intel_reset.c
index 10b930eaa8cb..6108a449cd19 100644
--- a/drivers/gpu/drm/i915/gt/intel_reset.c
+++ b/drivers/gpu/drm/i915/gt/intel_reset.c
@@ -174,13 +174,13 @@ static int i915_do_reset(struct intel_gt *gt,
/* Assert reset for at least 20 usec, and wait for acknowledgement. */
pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE);
udelay(50);
- err = wait_for_atomic(i915_in_reset(pdev), 50);
+ err = _wait_for_atomic(i915_in_reset(pdev), 50, 0);
/* Clear the reset request. */
pci_write_config_byte(pdev, I915_GDRST, 0);
udelay(50);
if (!err)
- err = wait_for_atomic(!i915_in_reset(pdev), 50);
+ err = _wait_for_atomic(!i915_in_reset(pdev), 50, 0);
return err;
}
@@ -200,7 +200,7 @@ static int g33_do_reset(struct intel_gt *gt,
struct pci_dev *pdev = to_pci_dev(gt->i915->drm.dev);
pci_write_config_byte(pdev, I915_GDRST, GRDOM_RESET_ENABLE);
- return wait_for_atomic(g4x_reset_complete(pdev), 50);
+ return _wait_for_atomic(g4x_reset_complete(pdev), 50, 0);
}
static int g4x_do_reset(struct intel_gt *gt,
@@ -217,7 +217,7 @@ static int g4x_do_reset(struct intel_gt *gt,
pci_write_config_byte(pdev, I915_GDRST,
GRDOM_MEDIA | GRDOM_RESET_ENABLE);
- ret = wait_for_atomic(g4x_reset_complete(pdev), 50);
+ ret = _wait_for_atomic(g4x_reset_complete(pdev), 50, 0);
if (ret) {
GT_TRACE(gt, "Wait for media reset failed\n");
goto out;
@@ -225,7 +225,7 @@ static int g4x_do_reset(struct intel_gt *gt,
pci_write_config_byte(pdev, I915_GDRST,
GRDOM_RENDER | GRDOM_RESET_ENABLE);
- ret = wait_for_atomic(g4x_reset_complete(pdev), 50);
+ ret = _wait_for_atomic(g4x_reset_complete(pdev), 50, 0);
if (ret) {
GT_TRACE(gt, "Wait for render reset failed\n");
goto out;
@@ -718,9 +718,7 @@ int __intel_gt_reset(struct intel_gt *gt, intel_engine_mask_t engine_mask)
intel_uncore_forcewake_get(gt->uncore, FORCEWAKE_ALL);
for (retry = 0; ret == -ETIMEDOUT && retry < retries; retry++) {
GT_TRACE(gt, "engine_mask=%x\n", engine_mask);
- preempt_disable();
ret = reset(gt, engine_mask, retry);
- preempt_enable();
}
intel_uncore_forcewake_put(gt->uncore, FORCEWAKE_ALL);
--
2.40.0
next prev parent reply other threads:[~2023-08-19 2:46 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-19 2:45 [PATCH 6.1-rt/5.15-rt 0/1] Backport i915 reset fix from v6.4-rt paul.gortmaker
2023-08-19 2:45 ` paul.gortmaker [this message]
2023-08-23 13:39 ` Sebastian Andrzej Siewior
2023-08-23 17:57 ` Joseph Salisbury
2023-10-18 19:58 ` Joseph Salisbury
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=20230819024525.2056048-2-paul.gortmaker@windriver.com \
--to=paul.gortmaker@windriver.com \
--cc=bigeasy@linutronix.de \
--cc=joseph.salisbury@canonical.com \
--cc=linux-rt-users@vger.kernel.org \
--cc=tvrtko.ursulin@intel.com \
--cc=williams@redhat.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;
as well as URLs for NNTP newsgroup(s).