* [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts
@ 2021-08-26 10:52 ` Tvrtko Ursulin
0 siblings, 0 replies; 9+ messages in thread
From: Tvrtko Ursulin @ 2021-08-26 10:52 UTC (permalink / raw)
To: Intel-gfx
Cc: dri-devel, Tvrtko Ursulin, Chris Wilson, Zhen Han, Matthew Brost
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
When a non-persistent context exits we currently mark it as banned in
order to trigger fast termination of any outstanding GPU jobs it may have
left running.
In doing so we apply a very strict 1ms limit in which the left over job
has to preempt before we issues an engine resets.
Some workloads are not able to cleanly preempt in that time window and it
can be argued that it would instead be better to give them a bit more
grace since avoiding engine resets is generally preferrable.
To achieve this the patch splits handling of banned contexts from simply
closed non-persistent ones and then applies different timeouts for both
and also extends the criteria which determines if a request should be
scheduled back in after preemption or not.
20ms preempt timeout grace is given to exited non-persistent contexts
which have been empirically tested to satisfy customers requirements
and still provides reasonably quick cleanup post exit.
v2:
* Streamline fast path checks.
v3:
* Simplify by using only schedulable status.
* Increase timeout to 20ms.
v4:
* Fix live_execlists selftest.
v5:
* Fix logic in kill_engines.
v6:
* Rebase.
v7:
* Add GuC support.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Zhen Han <zhen.han@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 22 +++++++++++-----
drivers/gpu/drm/i915/gt/intel_context.c | 25 ++++++++++++++++++
drivers/gpu/drm/i915/gt/intel_context.h | 26 ++++++++++++++-----
drivers/gpu/drm/i915/gt/intel_context_types.h | 3 ++-
.../drm/i915/gt/intel_execlists_submission.c | 13 +++++++---
.../gpu/drm/i915/gt/intel_ring_submission.c | 7 ++---
.../gpu/drm/i915/gt/uc/intel_guc_submission.c | 13 ++++++----
drivers/gpu/drm/i915/i915_request.c | 2 +-
8 files changed, 84 insertions(+), 27 deletions(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index fd169cf2f75a..6ae803cb4de3 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -1072,7 +1072,8 @@ static struct intel_engine_cs *active_engine(struct intel_context *ce)
return engine;
}
-static void kill_engines(struct i915_gem_engines *engines, bool ban)
+static void
+kill_engines(struct i915_gem_engines *engines, bool ban, bool persistent)
{
struct i915_gem_engines_iter it;
struct intel_context *ce;
@@ -1086,8 +1087,15 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
*/
for_each_gem_engine(ce, engines, it) {
struct intel_engine_cs *engine;
+ bool skip = false;
- if (ban && intel_context_ban(ce, NULL))
+ if (ban)
+ skip = intel_context_ban(ce, NULL);
+ else if (!persistent)
+ skip = intel_context_exit_nonpersistent(ce, NULL);
+
+ /* Already banned or non-persistent closed. */
+ if (skip)
continue;
/*
@@ -1100,7 +1108,7 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
engine = active_engine(ce);
/* First attempt to gracefully cancel the context */
- if (engine && !__cancel_engine(engine) && ban)
+ if (engine && !__cancel_engine(engine) && (ban || !persistent))
/*
* If we are unable to send a preemptive pulse to bump
* the context from the GPU, we have to resort to a full
@@ -1112,8 +1120,6 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
static void kill_context(struct i915_gem_context *ctx)
{
- bool ban = (!i915_gem_context_is_persistent(ctx) ||
- !ctx->i915->params.enable_hangcheck);
struct i915_gem_engines *pos, *next;
spin_lock_irq(&ctx->stale.lock);
@@ -1126,7 +1132,8 @@ static void kill_context(struct i915_gem_context *ctx)
spin_unlock_irq(&ctx->stale.lock);
- kill_engines(pos, ban);
+ kill_engines(pos, !ctx->i915->params.enable_hangcheck,
+ i915_gem_context_is_persistent(ctx));
spin_lock_irq(&ctx->stale.lock);
GEM_BUG_ON(i915_sw_fence_signaled(&pos->fence));
@@ -1172,7 +1179,8 @@ static void engines_idle_release(struct i915_gem_context *ctx,
kill:
if (list_empty(&engines->link)) /* raced, already closed */
- kill_engines(engines, true);
+ kill_engines(engines, true,
+ i915_gem_context_is_persistent(ctx));
i915_sw_fence_commit(&engines->fence);
}
diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
index 745e84c72c90..b9880ffe5da7 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.c
+++ b/drivers/gpu/drm/i915/gt/intel_context.c
@@ -533,6 +533,31 @@ struct i915_request *intel_context_find_active_request(struct intel_context *ce)
return active;
}
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq)
+{
+ bool ret = intel_context_set_banned(ce);
+
+ trace_intel_context_ban(ce);
+
+ if (ce->ops->revoke)
+ ce->ops->revoke(ce, rq,
+ INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS);
+
+ return ret;
+}
+
+bool intel_context_exit_nonpersistent(struct intel_context *ce,
+ struct i915_request *rq)
+{
+ bool ret = intel_context_set_exiting(ce);
+
+ if (ce->ops->revoke)
+ ce->ops->revoke(ce, rq,
+ INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS);
+
+ return ret;
+}
+
#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
#include "selftest_context.c"
#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
index c41098950746..8b77f3e7449a 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.h
+++ b/drivers/gpu/drm/i915/gt/intel_context.h
@@ -25,6 +25,9 @@
##__VA_ARGS__); \
} while (0)
+#define INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS (20)
+#define INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS (1)
+
struct i915_gem_ww_ctx;
void intel_context_init(struct intel_context *ce,
@@ -254,18 +257,27 @@ static inline bool intel_context_set_banned(struct intel_context *ce)
return test_and_set_bit(CONTEXT_BANNED, &ce->flags);
}
-static inline bool intel_context_ban(struct intel_context *ce,
- struct i915_request *rq)
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq);
+
+static inline bool intel_context_is_schedulable(const struct intel_context *ce)
{
- bool ret = intel_context_set_banned(ce);
+ return !test_bit(CONTEXT_EXITING, &ce->flags) &&
+ !test_bit(CONTEXT_BANNED, &ce->flags);
+}
- trace_intel_context_ban(ce);
- if (ce->ops->ban)
- ce->ops->ban(ce, rq);
+static inline bool intel_context_is_exiting(const struct intel_context *ce)
+{
+ return test_bit(CONTEXT_EXITING, &ce->flags);
+}
- return ret;
+static inline bool intel_context_set_exiting(struct intel_context *ce)
+{
+ return test_and_set_bit(CONTEXT_EXITING, &ce->flags);
}
+bool intel_context_exit_nonpersistent(struct intel_context *ce,
+ struct i915_request *rq);
+
static inline bool
intel_context_force_single_submission(const struct intel_context *ce)
{
diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
index e54351a170e2..4a0a1e736972 100644
--- a/drivers/gpu/drm/i915/gt/intel_context_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
@@ -37,7 +37,7 @@ struct intel_context_ops {
int (*alloc)(struct intel_context *ce);
- void (*ban)(struct intel_context *ce, struct i915_request *rq);
+ void (*revoke)(struct intel_context *ce, struct i915_request *rq, unsigned int preempt_timeout_ms);
int (*pre_pin)(struct intel_context *ce, struct i915_gem_ww_ctx *ww, void **vaddr);
int (*pin)(struct intel_context *ce, void *vaddr);
@@ -112,6 +112,7 @@ struct intel_context {
#define CONTEXT_FORCE_SINGLE_SUBMISSION 7
#define CONTEXT_NOPREEMPT 8
#define CONTEXT_LRCA_DIRTY 9
+#define CONTEXT_EXITING 10
struct {
u64 timeout_us;
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
index de5f9c86b9a4..43823b9d4013 100644
--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -478,7 +478,7 @@ __execlists_schedule_in(struct i915_request *rq)
!intel_engine_has_heartbeat(engine)))
intel_context_set_banned(ce);
- if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
+ if (unlikely(!intel_context_is_schedulable(ce) || bad_request(rq)))
reset_active(rq, engine);
if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
@@ -1222,12 +1222,19 @@ static void record_preemption(struct intel_engine_execlists *execlists)
static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
const struct i915_request *rq)
{
+ struct intel_context *ce;
+
if (!rq)
return 0;
+ ce = rq->context;
+
/* Force a fast reset for terminated contexts (ignoring sysfs!) */
- if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq)))
- return 1;
+ if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
+ return INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS;
+ /* Longer grace for closed non-persistent contexts to avoid resets. */
+ else if (unlikely(intel_context_is_exiting(ce)))
+ return INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS;
return READ_ONCE(engine->props.preempt_timeout_ms);
}
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
index 3c65efcb7bed..a28862bbe51e 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
@@ -587,8 +587,9 @@ static void ring_context_reset(struct intel_context *ce)
clear_bit(CONTEXT_VALID_BIT, &ce->flags);
}
-static void ring_context_ban(struct intel_context *ce,
- struct i915_request *rq)
+static void ring_context_revoke(struct intel_context *ce,
+ struct i915_request *rq,
+ unsigned int preempt_timeout_ms)
{
struct intel_engine_cs *engine;
@@ -623,7 +624,7 @@ static const struct intel_context_ops ring_context_ops = {
.cancel_request = ring_context_cancel_request,
- .ban = ring_context_ban,
+ .revoke = ring_context_revoke,
.pre_pin = ring_context_pre_pin,
.pin = ring_context_pin,
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
index 87d8dc8f51b9..b0a8629b26e3 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
@@ -1618,7 +1618,9 @@ static void __guc_context_set_preemption_timeout(struct intel_guc *guc,
intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
}
-static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
+static void
+guc_context_revoke(struct intel_context *ce, struct i915_request *rq,
+ unsigned int preempt_timeout_ms)
{
struct intel_guc *guc = ce_to_guc(ce);
struct intel_runtime_pm *runtime_pm =
@@ -1655,7 +1657,8 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
* gets kicked off the HW ASAP.
*/
with_intel_runtime_pm(runtime_pm, wakeref) {
- __guc_context_set_preemption_timeout(guc, guc_id, 1);
+ __guc_context_set_preemption_timeout(guc, guc_id,
+ preempt_timeout_ms);
__guc_context_sched_disable(guc, ce, guc_id);
}
} else {
@@ -1663,7 +1666,7 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
with_intel_runtime_pm(runtime_pm, wakeref)
__guc_context_set_preemption_timeout(guc,
ce->guc_id,
- 1);
+ preempt_timeout_ms);
spin_unlock_irqrestore(&ce->guc_state.lock, flags);
}
}
@@ -1979,7 +1982,7 @@ static const struct intel_context_ops guc_context_ops = {
.unpin = guc_context_unpin,
.post_unpin = guc_context_post_unpin,
- .ban = guc_context_ban,
+ .revoke = guc_context_revoke,
.cancel_request = guc_context_cancel_request,
@@ -2176,7 +2179,7 @@ static const struct intel_context_ops virtual_guc_context_ops = {
.unpin = guc_context_unpin,
.post_unpin = guc_context_post_unpin,
- .ban = guc_context_ban,
+ .revoke = guc_context_revoke,
.cancel_request = guc_context_cancel_request,
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index ce446716d092..b1a9bec83339 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -583,7 +583,7 @@ bool __i915_request_submit(struct i915_request *request)
goto active;
}
- if (unlikely(intel_context_is_banned(request->context)))
+ if (unlikely(!intel_context_is_schedulable(request->context)))
i915_request_set_error_once(request, -EIO);
if (unlikely(fatal_error(request->fence.error)))
--
2.30.2
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [Intel-gfx] [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts
@ 2021-08-26 10:52 ` Tvrtko Ursulin
0 siblings, 0 replies; 9+ messages in thread
From: Tvrtko Ursulin @ 2021-08-26 10:52 UTC (permalink / raw)
To: Intel-gfx
Cc: dri-devel, Tvrtko Ursulin, Chris Wilson, Zhen Han, Matthew Brost
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
When a non-persistent context exits we currently mark it as banned in
order to trigger fast termination of any outstanding GPU jobs it may have
left running.
In doing so we apply a very strict 1ms limit in which the left over job
has to preempt before we issues an engine resets.
Some workloads are not able to cleanly preempt in that time window and it
can be argued that it would instead be better to give them a bit more
grace since avoiding engine resets is generally preferrable.
To achieve this the patch splits handling of banned contexts from simply
closed non-persistent ones and then applies different timeouts for both
and also extends the criteria which determines if a request should be
scheduled back in after preemption or not.
20ms preempt timeout grace is given to exited non-persistent contexts
which have been empirically tested to satisfy customers requirements
and still provides reasonably quick cleanup post exit.
v2:
* Streamline fast path checks.
v3:
* Simplify by using only schedulable status.
* Increase timeout to 20ms.
v4:
* Fix live_execlists selftest.
v5:
* Fix logic in kill_engines.
v6:
* Rebase.
v7:
* Add GuC support.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Zhen Han <zhen.han@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 22 +++++++++++-----
drivers/gpu/drm/i915/gt/intel_context.c | 25 ++++++++++++++++++
drivers/gpu/drm/i915/gt/intel_context.h | 26 ++++++++++++++-----
drivers/gpu/drm/i915/gt/intel_context_types.h | 3 ++-
.../drm/i915/gt/intel_execlists_submission.c | 13 +++++++---
.../gpu/drm/i915/gt/intel_ring_submission.c | 7 ++---
.../gpu/drm/i915/gt/uc/intel_guc_submission.c | 13 ++++++----
drivers/gpu/drm/i915/i915_request.c | 2 +-
8 files changed, 84 insertions(+), 27 deletions(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index fd169cf2f75a..6ae803cb4de3 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -1072,7 +1072,8 @@ static struct intel_engine_cs *active_engine(struct intel_context *ce)
return engine;
}
-static void kill_engines(struct i915_gem_engines *engines, bool ban)
+static void
+kill_engines(struct i915_gem_engines *engines, bool ban, bool persistent)
{
struct i915_gem_engines_iter it;
struct intel_context *ce;
@@ -1086,8 +1087,15 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
*/
for_each_gem_engine(ce, engines, it) {
struct intel_engine_cs *engine;
+ bool skip = false;
- if (ban && intel_context_ban(ce, NULL))
+ if (ban)
+ skip = intel_context_ban(ce, NULL);
+ else if (!persistent)
+ skip = intel_context_exit_nonpersistent(ce, NULL);
+
+ /* Already banned or non-persistent closed. */
+ if (skip)
continue;
/*
@@ -1100,7 +1108,7 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
engine = active_engine(ce);
/* First attempt to gracefully cancel the context */
- if (engine && !__cancel_engine(engine) && ban)
+ if (engine && !__cancel_engine(engine) && (ban || !persistent))
/*
* If we are unable to send a preemptive pulse to bump
* the context from the GPU, we have to resort to a full
@@ -1112,8 +1120,6 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
static void kill_context(struct i915_gem_context *ctx)
{
- bool ban = (!i915_gem_context_is_persistent(ctx) ||
- !ctx->i915->params.enable_hangcheck);
struct i915_gem_engines *pos, *next;
spin_lock_irq(&ctx->stale.lock);
@@ -1126,7 +1132,8 @@ static void kill_context(struct i915_gem_context *ctx)
spin_unlock_irq(&ctx->stale.lock);
- kill_engines(pos, ban);
+ kill_engines(pos, !ctx->i915->params.enable_hangcheck,
+ i915_gem_context_is_persistent(ctx));
spin_lock_irq(&ctx->stale.lock);
GEM_BUG_ON(i915_sw_fence_signaled(&pos->fence));
@@ -1172,7 +1179,8 @@ static void engines_idle_release(struct i915_gem_context *ctx,
kill:
if (list_empty(&engines->link)) /* raced, already closed */
- kill_engines(engines, true);
+ kill_engines(engines, true,
+ i915_gem_context_is_persistent(ctx));
i915_sw_fence_commit(&engines->fence);
}
diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
index 745e84c72c90..b9880ffe5da7 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.c
+++ b/drivers/gpu/drm/i915/gt/intel_context.c
@@ -533,6 +533,31 @@ struct i915_request *intel_context_find_active_request(struct intel_context *ce)
return active;
}
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq)
+{
+ bool ret = intel_context_set_banned(ce);
+
+ trace_intel_context_ban(ce);
+
+ if (ce->ops->revoke)
+ ce->ops->revoke(ce, rq,
+ INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS);
+
+ return ret;
+}
+
+bool intel_context_exit_nonpersistent(struct intel_context *ce,
+ struct i915_request *rq)
+{
+ bool ret = intel_context_set_exiting(ce);
+
+ if (ce->ops->revoke)
+ ce->ops->revoke(ce, rq,
+ INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS);
+
+ return ret;
+}
+
#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
#include "selftest_context.c"
#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
index c41098950746..8b77f3e7449a 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.h
+++ b/drivers/gpu/drm/i915/gt/intel_context.h
@@ -25,6 +25,9 @@
##__VA_ARGS__); \
} while (0)
+#define INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS (20)
+#define INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS (1)
+
struct i915_gem_ww_ctx;
void intel_context_init(struct intel_context *ce,
@@ -254,18 +257,27 @@ static inline bool intel_context_set_banned(struct intel_context *ce)
return test_and_set_bit(CONTEXT_BANNED, &ce->flags);
}
-static inline bool intel_context_ban(struct intel_context *ce,
- struct i915_request *rq)
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq);
+
+static inline bool intel_context_is_schedulable(const struct intel_context *ce)
{
- bool ret = intel_context_set_banned(ce);
+ return !test_bit(CONTEXT_EXITING, &ce->flags) &&
+ !test_bit(CONTEXT_BANNED, &ce->flags);
+}
- trace_intel_context_ban(ce);
- if (ce->ops->ban)
- ce->ops->ban(ce, rq);
+static inline bool intel_context_is_exiting(const struct intel_context *ce)
+{
+ return test_bit(CONTEXT_EXITING, &ce->flags);
+}
- return ret;
+static inline bool intel_context_set_exiting(struct intel_context *ce)
+{
+ return test_and_set_bit(CONTEXT_EXITING, &ce->flags);
}
+bool intel_context_exit_nonpersistent(struct intel_context *ce,
+ struct i915_request *rq);
+
static inline bool
intel_context_force_single_submission(const struct intel_context *ce)
{
diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
index e54351a170e2..4a0a1e736972 100644
--- a/drivers/gpu/drm/i915/gt/intel_context_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
@@ -37,7 +37,7 @@ struct intel_context_ops {
int (*alloc)(struct intel_context *ce);
- void (*ban)(struct intel_context *ce, struct i915_request *rq);
+ void (*revoke)(struct intel_context *ce, struct i915_request *rq, unsigned int preempt_timeout_ms);
int (*pre_pin)(struct intel_context *ce, struct i915_gem_ww_ctx *ww, void **vaddr);
int (*pin)(struct intel_context *ce, void *vaddr);
@@ -112,6 +112,7 @@ struct intel_context {
#define CONTEXT_FORCE_SINGLE_SUBMISSION 7
#define CONTEXT_NOPREEMPT 8
#define CONTEXT_LRCA_DIRTY 9
+#define CONTEXT_EXITING 10
struct {
u64 timeout_us;
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
index de5f9c86b9a4..43823b9d4013 100644
--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -478,7 +478,7 @@ __execlists_schedule_in(struct i915_request *rq)
!intel_engine_has_heartbeat(engine)))
intel_context_set_banned(ce);
- if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
+ if (unlikely(!intel_context_is_schedulable(ce) || bad_request(rq)))
reset_active(rq, engine);
if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
@@ -1222,12 +1222,19 @@ static void record_preemption(struct intel_engine_execlists *execlists)
static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
const struct i915_request *rq)
{
+ struct intel_context *ce;
+
if (!rq)
return 0;
+ ce = rq->context;
+
/* Force a fast reset for terminated contexts (ignoring sysfs!) */
- if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq)))
- return 1;
+ if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
+ return INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS;
+ /* Longer grace for closed non-persistent contexts to avoid resets. */
+ else if (unlikely(intel_context_is_exiting(ce)))
+ return INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS;
return READ_ONCE(engine->props.preempt_timeout_ms);
}
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
index 3c65efcb7bed..a28862bbe51e 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
@@ -587,8 +587,9 @@ static void ring_context_reset(struct intel_context *ce)
clear_bit(CONTEXT_VALID_BIT, &ce->flags);
}
-static void ring_context_ban(struct intel_context *ce,
- struct i915_request *rq)
+static void ring_context_revoke(struct intel_context *ce,
+ struct i915_request *rq,
+ unsigned int preempt_timeout_ms)
{
struct intel_engine_cs *engine;
@@ -623,7 +624,7 @@ static const struct intel_context_ops ring_context_ops = {
.cancel_request = ring_context_cancel_request,
- .ban = ring_context_ban,
+ .revoke = ring_context_revoke,
.pre_pin = ring_context_pre_pin,
.pin = ring_context_pin,
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
index 87d8dc8f51b9..b0a8629b26e3 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
@@ -1618,7 +1618,9 @@ static void __guc_context_set_preemption_timeout(struct intel_guc *guc,
intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
}
-static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
+static void
+guc_context_revoke(struct intel_context *ce, struct i915_request *rq,
+ unsigned int preempt_timeout_ms)
{
struct intel_guc *guc = ce_to_guc(ce);
struct intel_runtime_pm *runtime_pm =
@@ -1655,7 +1657,8 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
* gets kicked off the HW ASAP.
*/
with_intel_runtime_pm(runtime_pm, wakeref) {
- __guc_context_set_preemption_timeout(guc, guc_id, 1);
+ __guc_context_set_preemption_timeout(guc, guc_id,
+ preempt_timeout_ms);
__guc_context_sched_disable(guc, ce, guc_id);
}
} else {
@@ -1663,7 +1666,7 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
with_intel_runtime_pm(runtime_pm, wakeref)
__guc_context_set_preemption_timeout(guc,
ce->guc_id,
- 1);
+ preempt_timeout_ms);
spin_unlock_irqrestore(&ce->guc_state.lock, flags);
}
}
@@ -1979,7 +1982,7 @@ static const struct intel_context_ops guc_context_ops = {
.unpin = guc_context_unpin,
.post_unpin = guc_context_post_unpin,
- .ban = guc_context_ban,
+ .revoke = guc_context_revoke,
.cancel_request = guc_context_cancel_request,
@@ -2176,7 +2179,7 @@ static const struct intel_context_ops virtual_guc_context_ops = {
.unpin = guc_context_unpin,
.post_unpin = guc_context_post_unpin,
- .ban = guc_context_ban,
+ .revoke = guc_context_revoke,
.cancel_request = guc_context_cancel_request,
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index ce446716d092..b1a9bec83339 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -583,7 +583,7 @@ bool __i915_request_submit(struct i915_request *request)
goto active;
}
- if (unlikely(intel_context_is_banned(request->context)))
+ if (unlikely(!intel_context_is_schedulable(request->context)))
i915_request_set_error_once(request, -EIO);
if (unlikely(fatal_error(request->fence.error)))
--
2.30.2
^ permalink raw reply related [flat|nested] 9+ messages in thread
* Re: [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts
2021-08-26 10:52 ` [Intel-gfx] " Tvrtko Ursulin
@ 2021-08-26 13:32 ` Daniel Vetter
-1 siblings, 0 replies; 9+ messages in thread
From: Daniel Vetter @ 2021-08-26 13:32 UTC (permalink / raw)
To: Tvrtko Ursulin
Cc: Intel-gfx, dri-devel, Tvrtko Ursulin, Chris Wilson, Zhen Han,
Matthew Brost
On Thu, Aug 26, 2021 at 11:52:14AM +0100, Tvrtko Ursulin wrote:
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>
> When a non-persistent context exits we currently mark it as banned in
> order to trigger fast termination of any outstanding GPU jobs it may have
> left running.
>
> In doing so we apply a very strict 1ms limit in which the left over job
> has to preempt before we issues an engine resets.
>
> Some workloads are not able to cleanly preempt in that time window and it
> can be argued that it would instead be better to give them a bit more
> grace since avoiding engine resets is generally preferrable.
>
> To achieve this the patch splits handling of banned contexts from simply
> closed non-persistent ones and then applies different timeouts for both
> and also extends the criteria which determines if a request should be
> scheduled back in after preemption or not.
>
> 20ms preempt timeout grace is given to exited non-persistent contexts
> which have been empirically tested to satisfy customers requirements
> and still provides reasonably quick cleanup post exit.
>
> v2:
> * Streamline fast path checks.
>
> v3:
> * Simplify by using only schedulable status.
> * Increase timeout to 20ms.
>
> v4:
> * Fix live_execlists selftest.
>
> v5:
> * Fix logic in kill_engines.
>
> v6:
> * Rebase.
>
> v7:
> * Add GuC support.
>
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Zhen Han <zhen.han@intel.com>
> Cc: Matthew Brost <matthew.brost@intel.com>
> ---
> drivers/gpu/drm/i915/gem/i915_gem_context.c | 22 +++++++++++-----
> drivers/gpu/drm/i915/gt/intel_context.c | 25 ++++++++++++++++++
> drivers/gpu/drm/i915/gt/intel_context.h | 26 ++++++++++++++-----
> drivers/gpu/drm/i915/gt/intel_context_types.h | 3 ++-
> .../drm/i915/gt/intel_execlists_submission.c | 13 +++++++---
> .../gpu/drm/i915/gt/intel_ring_submission.c | 7 ++---
> .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 13 ++++++----
> drivers/gpu/drm/i915/i915_request.c | 2 +-
> 8 files changed, 84 insertions(+), 27 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
> index fd169cf2f75a..6ae803cb4de3 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
> @@ -1072,7 +1072,8 @@ static struct intel_engine_cs *active_engine(struct intel_context *ce)
> return engine;
> }
>
> -static void kill_engines(struct i915_gem_engines *engines, bool ban)
> +static void
> +kill_engines(struct i915_gem_engines *engines, bool ban, bool persistent)
> {
> struct i915_gem_engines_iter it;
> struct intel_context *ce;
> @@ -1086,8 +1087,15 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
> */
> for_each_gem_engine(ce, engines, it) {
> struct intel_engine_cs *engine;
> + bool skip = false;
>
> - if (ban && intel_context_ban(ce, NULL))
> + if (ban)
> + skip = intel_context_ban(ce, NULL);
> + else if (!persistent)
> + skip = intel_context_exit_nonpersistent(ce, NULL);
> +
> + /* Already banned or non-persistent closed. */
> + if (skip)
> continue;
>
> /*
> @@ -1100,7 +1108,7 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
> engine = active_engine(ce);
>
> /* First attempt to gracefully cancel the context */
> - if (engine && !__cancel_engine(engine) && ban)
> + if (engine && !__cancel_engine(engine) && (ban || !persistent))
> /*
> * If we are unable to send a preemptive pulse to bump
> * the context from the GPU, we have to resort to a full
> @@ -1112,8 +1120,6 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
>
> static void kill_context(struct i915_gem_context *ctx)
> {
> - bool ban = (!i915_gem_context_is_persistent(ctx) ||
> - !ctx->i915->params.enable_hangcheck);
> struct i915_gem_engines *pos, *next;
>
> spin_lock_irq(&ctx->stale.lock);
> @@ -1126,7 +1132,8 @@ static void kill_context(struct i915_gem_context *ctx)
>
> spin_unlock_irq(&ctx->stale.lock);
>
> - kill_engines(pos, ban);
> + kill_engines(pos, !ctx->i915->params.enable_hangcheck,
> + i915_gem_context_is_persistent(ctx));
>
> spin_lock_irq(&ctx->stale.lock);
> GEM_BUG_ON(i915_sw_fence_signaled(&pos->fence));
> @@ -1172,7 +1179,8 @@ static void engines_idle_release(struct i915_gem_context *ctx,
>
> kill:
> if (list_empty(&engines->link)) /* raced, already closed */
> - kill_engines(engines, true);
> + kill_engines(engines, true,
> + i915_gem_context_is_persistent(ctx));
>
> i915_sw_fence_commit(&engines->fence);
> }
> diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
> index 745e84c72c90..b9880ffe5da7 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context.c
> +++ b/drivers/gpu/drm/i915/gt/intel_context.c
> @@ -533,6 +533,31 @@ struct i915_request *intel_context_find_active_request(struct intel_context *ce)
> return active;
> }
>
> +bool intel_context_ban(struct intel_context *ce, struct i915_request *rq)
> +{
> + bool ret = intel_context_set_banned(ce);
> +
> + trace_intel_context_ban(ce);
> +
> + if (ce->ops->revoke)
> + ce->ops->revoke(ce, rq,
> + INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS);
> +
> + return ret;
> +}
> +
> +bool intel_context_exit_nonpersistent(struct intel_context *ce,
> + struct i915_request *rq)
> +{
> + bool ret = intel_context_set_exiting(ce);
> +
> + if (ce->ops->revoke)
> + ce->ops->revoke(ce, rq,
> + INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS);
> +
> + return ret;
> +}
> +
> #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
> #include "selftest_context.c"
> #endif
> diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
> index c41098950746..8b77f3e7449a 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context.h
> +++ b/drivers/gpu/drm/i915/gt/intel_context.h
> @@ -25,6 +25,9 @@
> ##__VA_ARGS__); \
> } while (0)
>
> +#define INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS (20)
> +#define INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS (1)
> +
> struct i915_gem_ww_ctx;
>
> void intel_context_init(struct intel_context *ce,
> @@ -254,18 +257,27 @@ static inline bool intel_context_set_banned(struct intel_context *ce)
> return test_and_set_bit(CONTEXT_BANNED, &ce->flags);
> }
>
> -static inline bool intel_context_ban(struct intel_context *ce,
> - struct i915_request *rq)
> +bool intel_context_ban(struct intel_context *ce, struct i915_request *rq);
> +
> +static inline bool intel_context_is_schedulable(const struct intel_context *ce)
> {
> - bool ret = intel_context_set_banned(ce);
> + return !test_bit(CONTEXT_EXITING, &ce->flags) &&
> + !test_bit(CONTEXT_BANNED, &ce->flags);
> +}
>
> - trace_intel_context_ban(ce);
> - if (ce->ops->ban)
> - ce->ops->ban(ce, rq);
> +static inline bool intel_context_is_exiting(const struct intel_context *ce)
> +{
> + return test_bit(CONTEXT_EXITING, &ce->flags);
> +}
>
> - return ret;
> +static inline bool intel_context_set_exiting(struct intel_context *ce)
> +{
> + return test_and_set_bit(CONTEXT_EXITING, &ce->flags);
Uh more unordered atomics without barriers and without comments explaining
the barriers and pointing at the counter-party barrier. This needs to be
fixed.
In general it would be really good if all the context status bits are at
least properly documented. As-is this is an extremely challenging maze,
and I'm not sure it's really correct in all cases.
-Daniel
> }
>
> +bool intel_context_exit_nonpersistent(struct intel_context *ce,
> + struct i915_request *rq);
> +
> static inline bool
> intel_context_force_single_submission(const struct intel_context *ce)
> {
> diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
> index e54351a170e2..4a0a1e736972 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context_types.h
> +++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
> @@ -37,7 +37,7 @@ struct intel_context_ops {
>
> int (*alloc)(struct intel_context *ce);
>
> - void (*ban)(struct intel_context *ce, struct i915_request *rq);
> + void (*revoke)(struct intel_context *ce, struct i915_request *rq, unsigned int preempt_timeout_ms);
>
> int (*pre_pin)(struct intel_context *ce, struct i915_gem_ww_ctx *ww, void **vaddr);
> int (*pin)(struct intel_context *ce, void *vaddr);
> @@ -112,6 +112,7 @@ struct intel_context {
> #define CONTEXT_FORCE_SINGLE_SUBMISSION 7
> #define CONTEXT_NOPREEMPT 8
> #define CONTEXT_LRCA_DIRTY 9
> +#define CONTEXT_EXITING 10
>
> struct {
> u64 timeout_us;
> diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> index de5f9c86b9a4..43823b9d4013 100644
> --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> @@ -478,7 +478,7 @@ __execlists_schedule_in(struct i915_request *rq)
> !intel_engine_has_heartbeat(engine)))
> intel_context_set_banned(ce);
>
> - if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
> + if (unlikely(!intel_context_is_schedulable(ce) || bad_request(rq)))
> reset_active(rq, engine);
>
> if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
> @@ -1222,12 +1222,19 @@ static void record_preemption(struct intel_engine_execlists *execlists)
> static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
> const struct i915_request *rq)
> {
> + struct intel_context *ce;
> +
> if (!rq)
> return 0;
>
> + ce = rq->context;
> +
> /* Force a fast reset for terminated contexts (ignoring sysfs!) */
> - if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq)))
> - return 1;
> + if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
> + return INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS;
> + /* Longer grace for closed non-persistent contexts to avoid resets. */
> + else if (unlikely(intel_context_is_exiting(ce)))
> + return INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS;
>
> return READ_ONCE(engine->props.preempt_timeout_ms);
> }
> diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
> index 3c65efcb7bed..a28862bbe51e 100644
> --- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
> +++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
> @@ -587,8 +587,9 @@ static void ring_context_reset(struct intel_context *ce)
> clear_bit(CONTEXT_VALID_BIT, &ce->flags);
> }
>
> -static void ring_context_ban(struct intel_context *ce,
> - struct i915_request *rq)
> +static void ring_context_revoke(struct intel_context *ce,
> + struct i915_request *rq,
> + unsigned int preempt_timeout_ms)
> {
> struct intel_engine_cs *engine;
>
> @@ -623,7 +624,7 @@ static const struct intel_context_ops ring_context_ops = {
>
> .cancel_request = ring_context_cancel_request,
>
> - .ban = ring_context_ban,
> + .revoke = ring_context_revoke,
>
> .pre_pin = ring_context_pre_pin,
> .pin = ring_context_pin,
> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> index 87d8dc8f51b9..b0a8629b26e3 100644
> --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> @@ -1618,7 +1618,9 @@ static void __guc_context_set_preemption_timeout(struct intel_guc *guc,
> intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
> }
>
> -static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
> +static void
> +guc_context_revoke(struct intel_context *ce, struct i915_request *rq,
> + unsigned int preempt_timeout_ms)
> {
> struct intel_guc *guc = ce_to_guc(ce);
> struct intel_runtime_pm *runtime_pm =
> @@ -1655,7 +1657,8 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
> * gets kicked off the HW ASAP.
> */
> with_intel_runtime_pm(runtime_pm, wakeref) {
> - __guc_context_set_preemption_timeout(guc, guc_id, 1);
> + __guc_context_set_preemption_timeout(guc, guc_id,
> + preempt_timeout_ms);
> __guc_context_sched_disable(guc, ce, guc_id);
> }
> } else {
> @@ -1663,7 +1666,7 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
> with_intel_runtime_pm(runtime_pm, wakeref)
> __guc_context_set_preemption_timeout(guc,
> ce->guc_id,
> - 1);
> + preempt_timeout_ms);
> spin_unlock_irqrestore(&ce->guc_state.lock, flags);
> }
> }
> @@ -1979,7 +1982,7 @@ static const struct intel_context_ops guc_context_ops = {
> .unpin = guc_context_unpin,
> .post_unpin = guc_context_post_unpin,
>
> - .ban = guc_context_ban,
> + .revoke = guc_context_revoke,
>
> .cancel_request = guc_context_cancel_request,
>
> @@ -2176,7 +2179,7 @@ static const struct intel_context_ops virtual_guc_context_ops = {
> .unpin = guc_context_unpin,
> .post_unpin = guc_context_post_unpin,
>
> - .ban = guc_context_ban,
> + .revoke = guc_context_revoke,
>
> .cancel_request = guc_context_cancel_request,
>
> diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
> index ce446716d092..b1a9bec83339 100644
> --- a/drivers/gpu/drm/i915/i915_request.c
> +++ b/drivers/gpu/drm/i915/i915_request.c
> @@ -583,7 +583,7 @@ bool __i915_request_submit(struct i915_request *request)
> goto active;
> }
>
> - if (unlikely(intel_context_is_banned(request->context)))
> + if (unlikely(!intel_context_is_schedulable(request->context)))
> i915_request_set_error_once(request, -EIO);
>
> if (unlikely(fatal_error(request->fence.error)))
> --
> 2.30.2
>
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
^ permalink raw reply [flat|nested] 9+ messages in thread
* Re: [Intel-gfx] [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts
@ 2021-08-26 13:32 ` Daniel Vetter
0 siblings, 0 replies; 9+ messages in thread
From: Daniel Vetter @ 2021-08-26 13:32 UTC (permalink / raw)
To: Tvrtko Ursulin
Cc: Intel-gfx, dri-devel, Tvrtko Ursulin, Chris Wilson, Zhen Han,
Matthew Brost
On Thu, Aug 26, 2021 at 11:52:14AM +0100, Tvrtko Ursulin wrote:
> From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
>
> When a non-persistent context exits we currently mark it as banned in
> order to trigger fast termination of any outstanding GPU jobs it may have
> left running.
>
> In doing so we apply a very strict 1ms limit in which the left over job
> has to preempt before we issues an engine resets.
>
> Some workloads are not able to cleanly preempt in that time window and it
> can be argued that it would instead be better to give them a bit more
> grace since avoiding engine resets is generally preferrable.
>
> To achieve this the patch splits handling of banned contexts from simply
> closed non-persistent ones and then applies different timeouts for both
> and also extends the criteria which determines if a request should be
> scheduled back in after preemption or not.
>
> 20ms preempt timeout grace is given to exited non-persistent contexts
> which have been empirically tested to satisfy customers requirements
> and still provides reasonably quick cleanup post exit.
>
> v2:
> * Streamline fast path checks.
>
> v3:
> * Simplify by using only schedulable status.
> * Increase timeout to 20ms.
>
> v4:
> * Fix live_execlists selftest.
>
> v5:
> * Fix logic in kill_engines.
>
> v6:
> * Rebase.
>
> v7:
> * Add GuC support.
>
> Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
> Cc: Chris Wilson <chris@chris-wilson.co.uk>
> Cc: Zhen Han <zhen.han@intel.com>
> Cc: Matthew Brost <matthew.brost@intel.com>
> ---
> drivers/gpu/drm/i915/gem/i915_gem_context.c | 22 +++++++++++-----
> drivers/gpu/drm/i915/gt/intel_context.c | 25 ++++++++++++++++++
> drivers/gpu/drm/i915/gt/intel_context.h | 26 ++++++++++++++-----
> drivers/gpu/drm/i915/gt/intel_context_types.h | 3 ++-
> .../drm/i915/gt/intel_execlists_submission.c | 13 +++++++---
> .../gpu/drm/i915/gt/intel_ring_submission.c | 7 ++---
> .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 13 ++++++----
> drivers/gpu/drm/i915/i915_request.c | 2 +-
> 8 files changed, 84 insertions(+), 27 deletions(-)
>
> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
> index fd169cf2f75a..6ae803cb4de3 100644
> --- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
> +++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
> @@ -1072,7 +1072,8 @@ static struct intel_engine_cs *active_engine(struct intel_context *ce)
> return engine;
> }
>
> -static void kill_engines(struct i915_gem_engines *engines, bool ban)
> +static void
> +kill_engines(struct i915_gem_engines *engines, bool ban, bool persistent)
> {
> struct i915_gem_engines_iter it;
> struct intel_context *ce;
> @@ -1086,8 +1087,15 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
> */
> for_each_gem_engine(ce, engines, it) {
> struct intel_engine_cs *engine;
> + bool skip = false;
>
> - if (ban && intel_context_ban(ce, NULL))
> + if (ban)
> + skip = intel_context_ban(ce, NULL);
> + else if (!persistent)
> + skip = intel_context_exit_nonpersistent(ce, NULL);
> +
> + /* Already banned or non-persistent closed. */
> + if (skip)
> continue;
>
> /*
> @@ -1100,7 +1108,7 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
> engine = active_engine(ce);
>
> /* First attempt to gracefully cancel the context */
> - if (engine && !__cancel_engine(engine) && ban)
> + if (engine && !__cancel_engine(engine) && (ban || !persistent))
> /*
> * If we are unable to send a preemptive pulse to bump
> * the context from the GPU, we have to resort to a full
> @@ -1112,8 +1120,6 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
>
> static void kill_context(struct i915_gem_context *ctx)
> {
> - bool ban = (!i915_gem_context_is_persistent(ctx) ||
> - !ctx->i915->params.enable_hangcheck);
> struct i915_gem_engines *pos, *next;
>
> spin_lock_irq(&ctx->stale.lock);
> @@ -1126,7 +1132,8 @@ static void kill_context(struct i915_gem_context *ctx)
>
> spin_unlock_irq(&ctx->stale.lock);
>
> - kill_engines(pos, ban);
> + kill_engines(pos, !ctx->i915->params.enable_hangcheck,
> + i915_gem_context_is_persistent(ctx));
>
> spin_lock_irq(&ctx->stale.lock);
> GEM_BUG_ON(i915_sw_fence_signaled(&pos->fence));
> @@ -1172,7 +1179,8 @@ static void engines_idle_release(struct i915_gem_context *ctx,
>
> kill:
> if (list_empty(&engines->link)) /* raced, already closed */
> - kill_engines(engines, true);
> + kill_engines(engines, true,
> + i915_gem_context_is_persistent(ctx));
>
> i915_sw_fence_commit(&engines->fence);
> }
> diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
> index 745e84c72c90..b9880ffe5da7 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context.c
> +++ b/drivers/gpu/drm/i915/gt/intel_context.c
> @@ -533,6 +533,31 @@ struct i915_request *intel_context_find_active_request(struct intel_context *ce)
> return active;
> }
>
> +bool intel_context_ban(struct intel_context *ce, struct i915_request *rq)
> +{
> + bool ret = intel_context_set_banned(ce);
> +
> + trace_intel_context_ban(ce);
> +
> + if (ce->ops->revoke)
> + ce->ops->revoke(ce, rq,
> + INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS);
> +
> + return ret;
> +}
> +
> +bool intel_context_exit_nonpersistent(struct intel_context *ce,
> + struct i915_request *rq)
> +{
> + bool ret = intel_context_set_exiting(ce);
> +
> + if (ce->ops->revoke)
> + ce->ops->revoke(ce, rq,
> + INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS);
> +
> + return ret;
> +}
> +
> #if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
> #include "selftest_context.c"
> #endif
> diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
> index c41098950746..8b77f3e7449a 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context.h
> +++ b/drivers/gpu/drm/i915/gt/intel_context.h
> @@ -25,6 +25,9 @@
> ##__VA_ARGS__); \
> } while (0)
>
> +#define INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS (20)
> +#define INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS (1)
> +
> struct i915_gem_ww_ctx;
>
> void intel_context_init(struct intel_context *ce,
> @@ -254,18 +257,27 @@ static inline bool intel_context_set_banned(struct intel_context *ce)
> return test_and_set_bit(CONTEXT_BANNED, &ce->flags);
> }
>
> -static inline bool intel_context_ban(struct intel_context *ce,
> - struct i915_request *rq)
> +bool intel_context_ban(struct intel_context *ce, struct i915_request *rq);
> +
> +static inline bool intel_context_is_schedulable(const struct intel_context *ce)
> {
> - bool ret = intel_context_set_banned(ce);
> + return !test_bit(CONTEXT_EXITING, &ce->flags) &&
> + !test_bit(CONTEXT_BANNED, &ce->flags);
> +}
>
> - trace_intel_context_ban(ce);
> - if (ce->ops->ban)
> - ce->ops->ban(ce, rq);
> +static inline bool intel_context_is_exiting(const struct intel_context *ce)
> +{
> + return test_bit(CONTEXT_EXITING, &ce->flags);
> +}
>
> - return ret;
> +static inline bool intel_context_set_exiting(struct intel_context *ce)
> +{
> + return test_and_set_bit(CONTEXT_EXITING, &ce->flags);
Uh more unordered atomics without barriers and without comments explaining
the barriers and pointing at the counter-party barrier. This needs to be
fixed.
In general it would be really good if all the context status bits are at
least properly documented. As-is this is an extremely challenging maze,
and I'm not sure it's really correct in all cases.
-Daniel
> }
>
> +bool intel_context_exit_nonpersistent(struct intel_context *ce,
> + struct i915_request *rq);
> +
> static inline bool
> intel_context_force_single_submission(const struct intel_context *ce)
> {
> diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
> index e54351a170e2..4a0a1e736972 100644
> --- a/drivers/gpu/drm/i915/gt/intel_context_types.h
> +++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
> @@ -37,7 +37,7 @@ struct intel_context_ops {
>
> int (*alloc)(struct intel_context *ce);
>
> - void (*ban)(struct intel_context *ce, struct i915_request *rq);
> + void (*revoke)(struct intel_context *ce, struct i915_request *rq, unsigned int preempt_timeout_ms);
>
> int (*pre_pin)(struct intel_context *ce, struct i915_gem_ww_ctx *ww, void **vaddr);
> int (*pin)(struct intel_context *ce, void *vaddr);
> @@ -112,6 +112,7 @@ struct intel_context {
> #define CONTEXT_FORCE_SINGLE_SUBMISSION 7
> #define CONTEXT_NOPREEMPT 8
> #define CONTEXT_LRCA_DIRTY 9
> +#define CONTEXT_EXITING 10
>
> struct {
> u64 timeout_us;
> diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> index de5f9c86b9a4..43823b9d4013 100644
> --- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> +++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
> @@ -478,7 +478,7 @@ __execlists_schedule_in(struct i915_request *rq)
> !intel_engine_has_heartbeat(engine)))
> intel_context_set_banned(ce);
>
> - if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
> + if (unlikely(!intel_context_is_schedulable(ce) || bad_request(rq)))
> reset_active(rq, engine);
>
> if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
> @@ -1222,12 +1222,19 @@ static void record_preemption(struct intel_engine_execlists *execlists)
> static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
> const struct i915_request *rq)
> {
> + struct intel_context *ce;
> +
> if (!rq)
> return 0;
>
> + ce = rq->context;
> +
> /* Force a fast reset for terminated contexts (ignoring sysfs!) */
> - if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq)))
> - return 1;
> + if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
> + return INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS;
> + /* Longer grace for closed non-persistent contexts to avoid resets. */
> + else if (unlikely(intel_context_is_exiting(ce)))
> + return INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS;
>
> return READ_ONCE(engine->props.preempt_timeout_ms);
> }
> diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
> index 3c65efcb7bed..a28862bbe51e 100644
> --- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
> +++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
> @@ -587,8 +587,9 @@ static void ring_context_reset(struct intel_context *ce)
> clear_bit(CONTEXT_VALID_BIT, &ce->flags);
> }
>
> -static void ring_context_ban(struct intel_context *ce,
> - struct i915_request *rq)
> +static void ring_context_revoke(struct intel_context *ce,
> + struct i915_request *rq,
> + unsigned int preempt_timeout_ms)
> {
> struct intel_engine_cs *engine;
>
> @@ -623,7 +624,7 @@ static const struct intel_context_ops ring_context_ops = {
>
> .cancel_request = ring_context_cancel_request,
>
> - .ban = ring_context_ban,
> + .revoke = ring_context_revoke,
>
> .pre_pin = ring_context_pre_pin,
> .pin = ring_context_pin,
> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> index 87d8dc8f51b9..b0a8629b26e3 100644
> --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
> @@ -1618,7 +1618,9 @@ static void __guc_context_set_preemption_timeout(struct intel_guc *guc,
> intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
> }
>
> -static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
> +static void
> +guc_context_revoke(struct intel_context *ce, struct i915_request *rq,
> + unsigned int preempt_timeout_ms)
> {
> struct intel_guc *guc = ce_to_guc(ce);
> struct intel_runtime_pm *runtime_pm =
> @@ -1655,7 +1657,8 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
> * gets kicked off the HW ASAP.
> */
> with_intel_runtime_pm(runtime_pm, wakeref) {
> - __guc_context_set_preemption_timeout(guc, guc_id, 1);
> + __guc_context_set_preemption_timeout(guc, guc_id,
> + preempt_timeout_ms);
> __guc_context_sched_disable(guc, ce, guc_id);
> }
> } else {
> @@ -1663,7 +1666,7 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
> with_intel_runtime_pm(runtime_pm, wakeref)
> __guc_context_set_preemption_timeout(guc,
> ce->guc_id,
> - 1);
> + preempt_timeout_ms);
> spin_unlock_irqrestore(&ce->guc_state.lock, flags);
> }
> }
> @@ -1979,7 +1982,7 @@ static const struct intel_context_ops guc_context_ops = {
> .unpin = guc_context_unpin,
> .post_unpin = guc_context_post_unpin,
>
> - .ban = guc_context_ban,
> + .revoke = guc_context_revoke,
>
> .cancel_request = guc_context_cancel_request,
>
> @@ -2176,7 +2179,7 @@ static const struct intel_context_ops virtual_guc_context_ops = {
> .unpin = guc_context_unpin,
> .post_unpin = guc_context_post_unpin,
>
> - .ban = guc_context_ban,
> + .revoke = guc_context_revoke,
>
> .cancel_request = guc_context_cancel_request,
>
> diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
> index ce446716d092..b1a9bec83339 100644
> --- a/drivers/gpu/drm/i915/i915_request.c
> +++ b/drivers/gpu/drm/i915/i915_request.c
> @@ -583,7 +583,7 @@ bool __i915_request_submit(struct i915_request *request)
> goto active;
> }
>
> - if (unlikely(intel_context_is_banned(request->context)))
> + if (unlikely(!intel_context_is_schedulable(request->context)))
> i915_request_set_error_once(request, -EIO);
>
> if (unlikely(fatal_error(request->fence.error)))
> --
> 2.30.2
>
--
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/i915: Be more gentle when exiting non-persistent contexts (rev2)
2021-08-26 10:52 ` [Intel-gfx] " Tvrtko Ursulin
(?)
(?)
@ 2021-08-26 14:57 ` Patchwork
-1 siblings, 0 replies; 9+ messages in thread
From: Patchwork @ 2021-08-26 14:57 UTC (permalink / raw)
To: Tvrtko Ursulin; +Cc: intel-gfx
== Series Details ==
Series: drm/i915: Be more gentle when exiting non-persistent contexts (rev2)
URL : https://patchwork.freedesktop.org/series/93420/
State : warning
== Summary ==
$ dim checkpatch origin/drm-tip
660ff9cdf697 drm/i915: Be more gentle when exiting non-persistent contexts
-:213: WARNING:LONG_LINE: line length of 107 exceeds 100 columns
#213: FILE: drivers/gpu/drm/i915/gt/intel_context_types.h:40:
+ void (*revoke)(struct intel_context *ce, struct i915_request *rq, unsigned int preempt_timeout_ms);
total: 0 errors, 1 warnings, 0 checks, 247 lines checked
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Intel-gfx] ✓ Fi.CI.BAT: success for drm/i915: Be more gentle when exiting non-persistent contexts (rev2)
2021-08-26 10:52 ` [Intel-gfx] " Tvrtko Ursulin
` (2 preceding siblings ...)
(?)
@ 2021-08-26 15:28 ` Patchwork
-1 siblings, 0 replies; 9+ messages in thread
From: Patchwork @ 2021-08-26 15:28 UTC (permalink / raw)
To: Tvrtko Ursulin; +Cc: intel-gfx
[-- Attachment #1: Type: text/plain, Size: 3354 bytes --]
== Series Details ==
Series: drm/i915: Be more gentle when exiting non-persistent contexts (rev2)
URL : https://patchwork.freedesktop.org/series/93420/
State : success
== Summary ==
CI Bug Log - changes from CI_DRM_10525 -> Patchwork_20903
====================================================
Summary
-------
**SUCCESS**
No regressions found.
External URL: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/index.html
Known issues
------------
Here are the changes found in Patchwork_20903 that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@amdgpu/amd_cs_nop@sync-fork-compute0:
- fi-kbl-soraka: NOTRUN -> [SKIP][1] ([fdo#109271]) +8 similar issues
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/fi-kbl-soraka/igt@amdgpu/amd_cs_nop@sync-fork-compute0.html
* igt@i915_module_load@reload:
- fi-tgl-1115g4: [PASS][2] -> [DMESG-WARN][3] ([i915#4002]) +1 similar issue
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/fi-tgl-1115g4/igt@i915_module_load@reload.html
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/fi-tgl-1115g4/igt@i915_module_load@reload.html
#### Possible fixes ####
* igt@kms_prop_blob@basic:
- fi-tgl-1115g4: [DMESG-WARN][4] ([i915#4002]) -> [PASS][5]
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/fi-tgl-1115g4/igt@kms_prop_blob@basic.html
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/fi-tgl-1115g4/igt@kms_prop_blob@basic.html
#### Warnings ####
* igt@gem_exec_suspend@basic-s3:
- fi-tgl-1115g4: [FAIL][6] ([i915#1888]) -> [DMESG-WARN][7] ([i915#4002])
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/fi-tgl-1115g4/igt@gem_exec_suspend@basic-s3.html
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/fi-tgl-1115g4/igt@gem_exec_suspend@basic-s3.html
* igt@kms_psr@primary_page_flip:
- fi-tgl-1115g4: [SKIP][8] ([i915#1072] / [i915#1385]) -> [SKIP][9] ([i915#1072])
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/fi-tgl-1115g4/igt@kms_psr@primary_page_flip.html
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/fi-tgl-1115g4/igt@kms_psr@primary_page_flip.html
[fdo#109271]: https://bugs.freedesktop.org/show_bug.cgi?id=109271
[i915#1072]: https://gitlab.freedesktop.org/drm/intel/issues/1072
[i915#1385]: https://gitlab.freedesktop.org/drm/intel/issues/1385
[i915#1888]: https://gitlab.freedesktop.org/drm/intel/issues/1888
[i915#4002]: https://gitlab.freedesktop.org/drm/intel/issues/4002
Participating hosts (40 -> 34)
------------------------------
Missing (6): fi-ilk-m540 bat-adls-5 fi-hsw-4200u fi-bsw-cyan fi-bdw-samus bat-jsl-1
Build changes
-------------
* Linux: CI_DRM_10525 -> Patchwork_20903
CI-20190529: 20190529
CI_DRM_10525: 059309d37ac2de5d93cf6d71fd7fe33c9c2c66ea @ git://anongit.freedesktop.org/gfx-ci/linux
IGT_6186: 250081b306c6fa8f95405fab6a7604f1968dd4ec @ https://gitlab.freedesktop.org/drm/igt-gpu-tools.git
Patchwork_20903: 660ff9cdf6970f83c2b76f96b4daeaeebf1a7b98 @ git://anongit.freedesktop.org/gfx-ci/linux
== Linux commits ==
660ff9cdf697 drm/i915: Be more gentle when exiting non-persistent contexts
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/index.html
[-- Attachment #2: Type: text/html, Size: 4267 bytes --]
^ permalink raw reply [flat|nested] 9+ messages in thread
* [Intel-gfx] ✗ Fi.CI.IGT: failure for drm/i915: Be more gentle when exiting non-persistent contexts (rev2)
2021-08-26 10:52 ` [Intel-gfx] " Tvrtko Ursulin
` (3 preceding siblings ...)
(?)
@ 2021-08-26 23:22 ` Patchwork
-1 siblings, 0 replies; 9+ messages in thread
From: Patchwork @ 2021-08-26 23:22 UTC (permalink / raw)
To: Tvrtko Ursulin; +Cc: intel-gfx
[-- Attachment #1: Type: text/plain, Size: 30287 bytes --]
== Series Details ==
Series: drm/i915: Be more gentle when exiting non-persistent contexts (rev2)
URL : https://patchwork.freedesktop.org/series/93420/
State : failure
== Summary ==
CI Bug Log - changes from CI_DRM_10525_full -> Patchwork_20903_full
====================================================
Summary
-------
**FAILURE**
Serious unknown changes coming with Patchwork_20903_full absolutely need to be
verified manually.
If you think the reported changes have nothing to do with the changes
introduced in Patchwork_20903_full, please notify your bug team to allow them
to document this new failure mode, which will reduce false positives in CI.
Possible new issues
-------------------
Here are the unknown changes that may have been introduced in Patchwork_20903_full:
### IGT changes ###
#### Possible regressions ####
* igt@i915_suspend@debugfs-reader:
- shard-skl: [PASS][1] -> [DMESG-WARN][2]
[1]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl5/igt@i915_suspend@debugfs-reader.html
[2]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl3/igt@i915_suspend@debugfs-reader.html
Known issues
------------
Here are the changes found in Patchwork_20903_full that come from known issues:
### IGT changes ###
#### Issues hit ####
* igt@drm_import_export@flink:
- shard-glk: [PASS][3] -> [INCOMPLETE][4] ([i915#2369])
[3]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-glk7/igt@drm_import_export@flink.html
[4]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-glk6/igt@drm_import_export@flink.html
* igt@feature_discovery@psr2:
- shard-iclb: [PASS][5] -> [SKIP][6] ([i915#658])
[5]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-iclb2/igt@feature_discovery@psr2.html
[6]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb4/igt@feature_discovery@psr2.html
* igt@gem_ctx_persistence@idempotent:
- shard-snb: NOTRUN -> [SKIP][7] ([fdo#109271] / [i915#1099])
[7]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-snb7/igt@gem_ctx_persistence@idempotent.html
* igt@gem_ctx_sseu@mmap-args:
- shard-tglb: NOTRUN -> [SKIP][8] ([i915#280])
[8]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@gem_ctx_sseu@mmap-args.html
* igt@gem_exec_fair@basic-deadline:
- shard-apl: NOTRUN -> [FAIL][9] ([i915#2846])
[9]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@gem_exec_fair@basic-deadline.html
* igt@gem_exec_fair@basic-flow@rcs0:
- shard-tglb: [PASS][10] -> [FAIL][11] ([i915#2842]) +1 similar issue
[10]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-tglb6/igt@gem_exec_fair@basic-flow@rcs0.html
[11]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@gem_exec_fair@basic-flow@rcs0.html
* igt@gem_exec_fair@basic-none-solo@rcs0:
- shard-tglb: NOTRUN -> [FAIL][12] ([i915#2842])
[12]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@gem_exec_fair@basic-none-solo@rcs0.html
* igt@gem_exec_fair@basic-none-vip@rcs0:
- shard-kbl: [PASS][13] -> [FAIL][14] ([i915#2842]) +1 similar issue
[13]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-kbl7/igt@gem_exec_fair@basic-none-vip@rcs0.html
[14]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl1/igt@gem_exec_fair@basic-none-vip@rcs0.html
* igt@gem_exec_params@secure-non-master:
- shard-tglb: NOTRUN -> [SKIP][15] ([fdo#112283])
[15]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@gem_exec_params@secure-non-master.html
* igt@gem_exec_whisper@basic-queues-priority-all:
- shard-glk: [PASS][16] -> [DMESG-WARN][17] ([i915#118] / [i915#95])
[16]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-glk6/igt@gem_exec_whisper@basic-queues-priority-all.html
[17]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-glk9/igt@gem_exec_whisper@basic-queues-priority-all.html
* igt@gem_pread@exhaustion:
- shard-apl: NOTRUN -> [WARN][18] ([i915#2658]) +1 similar issue
[18]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl7/igt@gem_pread@exhaustion.html
* igt@gem_render_copy@yf-tiled-to-vebox-linear:
- shard-skl: NOTRUN -> [SKIP][19] ([fdo#109271]) +19 similar issues
[19]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl5/igt@gem_render_copy@yf-tiled-to-vebox-linear.html
- shard-iclb: NOTRUN -> [SKIP][20] ([i915#768])
[20]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@gem_render_copy@yf-tiled-to-vebox-linear.html
* igt@gem_userptr_blits@input-checking:
- shard-snb: NOTRUN -> [DMESG-WARN][21] ([i915#3002])
[21]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-snb6/igt@gem_userptr_blits@input-checking.html
* igt@gem_userptr_blits@readonly-pwrite-unsync:
- shard-tglb: NOTRUN -> [SKIP][22] ([i915#3297])
[22]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@gem_userptr_blits@readonly-pwrite-unsync.html
- shard-iclb: NOTRUN -> [SKIP][23] ([i915#3297])
[23]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@gem_userptr_blits@readonly-pwrite-unsync.html
* igt@gem_userptr_blits@vma-merge:
- shard-snb: NOTRUN -> [FAIL][24] ([i915#2724])
[24]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-snb2/igt@gem_userptr_blits@vma-merge.html
* igt@gen3_render_tiledy_blits:
- shard-tglb: NOTRUN -> [SKIP][25] ([fdo#109289])
[25]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@gen3_render_tiledy_blits.html
* igt@i915_module_load@reload-with-fault-injection:
- shard-skl: [PASS][26] -> [DMESG-WARN][27] ([i915#1982]) +1 similar issue
[26]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl5/igt@i915_module_load@reload-with-fault-injection.html
[27]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl4/igt@i915_module_load@reload-with-fault-injection.html
* igt@i915_pm_dc@dc6-dpms:
- shard-kbl: NOTRUN -> [FAIL][28] ([i915#454])
[28]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl3/igt@i915_pm_dc@dc6-dpms.html
* igt@i915_pm_dc@dc6-psr:
- shard-iclb: [PASS][29] -> [FAIL][30] ([i915#454])
[29]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-iclb1/igt@i915_pm_dc@dc6-psr.html
[30]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb2/igt@i915_pm_dc@dc6-psr.html
* igt@i915_pm_lpsp@kms-lpsp@kms-lpsp-dp:
- shard-apl: NOTRUN -> [SKIP][31] ([fdo#109271] / [i915#1937])
[31]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl2/igt@i915_pm_lpsp@kms-lpsp@kms-lpsp-dp.html
* igt@i915_pm_rpm@gem-mmap-type@fixed:
- shard-apl: NOTRUN -> [SKIP][32] ([fdo#109271] / [i915#3976])
[32]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl7/igt@i915_pm_rpm@gem-mmap-type@fixed.html
* igt@i915_pm_rpm@modeset-non-lpsp-stress-no-wait:
- shard-tglb: NOTRUN -> [SKIP][33] ([fdo#111644] / [i915#1397] / [i915#2411])
[33]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@i915_pm_rpm@modeset-non-lpsp-stress-no-wait.html
* igt@kms_big_fb@x-tiled-max-hw-stride-32bpp-rotate-0-hflip:
- shard-skl: NOTRUN -> [SKIP][34] ([fdo#109271] / [i915#3777])
[34]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl5/igt@kms_big_fb@x-tiled-max-hw-stride-32bpp-rotate-0-hflip.html
* igt@kms_big_fb@x-tiled-max-hw-stride-32bpp-rotate-180-hflip:
- shard-apl: NOTRUN -> [SKIP][35] ([fdo#109271] / [i915#3777]) +3 similar issues
[35]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl7/igt@kms_big_fb@x-tiled-max-hw-stride-32bpp-rotate-180-hflip.html
* igt@kms_big_fb@yf-tiled-max-hw-stride-64bpp-rotate-0:
- shard-apl: NOTRUN -> [SKIP][36] ([fdo#109271]) +406 similar issues
[36]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_big_fb@yf-tiled-max-hw-stride-64bpp-rotate-0.html
* igt@kms_big_fb@yf-tiled-max-hw-stride-64bpp-rotate-0-hflip:
- shard-tglb: NOTRUN -> [SKIP][37] ([fdo#111615])
[37]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@kms_big_fb@yf-tiled-max-hw-stride-64bpp-rotate-0-hflip.html
* igt@kms_ccs@pipe-a-bad-pixel-format-y_tiled_gen12_rc_ccs_cc:
- shard-skl: NOTRUN -> [SKIP][38] ([fdo#109271] / [i915#3886]) +2 similar issues
[38]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl4/igt@kms_ccs@pipe-a-bad-pixel-format-y_tiled_gen12_rc_ccs_cc.html
* igt@kms_ccs@pipe-a-crc-sprite-planes-basic-y_tiled_gen12_rc_ccs_cc:
- shard-kbl: NOTRUN -> [SKIP][39] ([fdo#109271] / [i915#3886]) +1 similar issue
[39]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl2/igt@kms_ccs@pipe-a-crc-sprite-planes-basic-y_tiled_gen12_rc_ccs_cc.html
* igt@kms_ccs@pipe-a-missing-ccs-buffer-y_tiled_gen12_rc_ccs_cc:
- shard-apl: NOTRUN -> [SKIP][40] ([fdo#109271] / [i915#3886]) +20 similar issues
[40]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_ccs@pipe-a-missing-ccs-buffer-y_tiled_gen12_rc_ccs_cc.html
* igt@kms_ccs@pipe-b-missing-ccs-buffer-y_tiled_gen12_rc_ccs_cc:
- shard-iclb: NOTRUN -> [SKIP][41] ([fdo#109278] / [i915#3886])
[41]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_ccs@pipe-b-missing-ccs-buffer-y_tiled_gen12_rc_ccs_cc.html
* igt@kms_ccs@pipe-c-random-ccs-data-y_tiled_gen12_mc_ccs:
- shard-tglb: NOTRUN -> [SKIP][42] ([i915#3689] / [i915#3886])
[42]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@kms_ccs@pipe-c-random-ccs-data-y_tiled_gen12_mc_ccs.html
* igt@kms_chamelium@dp-audio:
- shard-tglb: NOTRUN -> [SKIP][43] ([fdo#109284] / [fdo#111827]) +1 similar issue
[43]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@kms_chamelium@dp-audio.html
* igt@kms_chamelium@hdmi-hpd-fast:
- shard-iclb: NOTRUN -> [SKIP][44] ([fdo#109284] / [fdo#111827]) +1 similar issue
[44]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_chamelium@hdmi-hpd-fast.html
* igt@kms_chamelium@vga-edid-read:
- shard-kbl: NOTRUN -> [SKIP][45] ([fdo#109271] / [fdo#111827]) +2 similar issues
[45]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl2/igt@kms_chamelium@vga-edid-read.html
* igt@kms_chamelium@vga-hpd:
- shard-apl: NOTRUN -> [SKIP][46] ([fdo#109271] / [fdo#111827]) +37 similar issues
[46]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_chamelium@vga-hpd.html
* igt@kms_color@pipe-d-ctm-green-to-red:
- shard-iclb: NOTRUN -> [SKIP][47] ([fdo#109278] / [i915#1149])
[47]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_color@pipe-d-ctm-green-to-red.html
* igt@kms_color_chamelium@pipe-a-ctm-blue-to-red:
- shard-snb: NOTRUN -> [SKIP][48] ([fdo#109271] / [fdo#111827]) +14 similar issues
[48]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-snb7/igt@kms_color_chamelium@pipe-a-ctm-blue-to-red.html
* igt@kms_content_protection@atomic:
- shard-apl: NOTRUN -> [TIMEOUT][49] ([i915#1319])
[49]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_content_protection@atomic.html
* igt@kms_content_protection@atomic-dpms:
- shard-tglb: NOTRUN -> [SKIP][50] ([fdo#111828]) +1 similar issue
[50]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@kms_content_protection@atomic-dpms.html
- shard-iclb: NOTRUN -> [SKIP][51] ([fdo#109300] / [fdo#111066])
[51]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_content_protection@atomic-dpms.html
* igt@kms_content_protection@uevent:
- shard-apl: NOTRUN -> [FAIL][52] ([i915#2105])
[52]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_content_protection@uevent.html
* igt@kms_cursor_crc@pipe-a-cursor-32x32-sliding:
- shard-tglb: NOTRUN -> [SKIP][53] ([i915#3319])
[53]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@kms_cursor_crc@pipe-a-cursor-32x32-sliding.html
* igt@kms_cursor_crc@pipe-b-cursor-size-change:
- shard-apl: NOTRUN -> [FAIL][54] ([i915#3444])
[54]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_cursor_crc@pipe-b-cursor-size-change.html
* igt@kms_cursor_crc@pipe-c-cursor-32x10-onscreen:
- shard-tglb: NOTRUN -> [SKIP][55] ([i915#3359])
[55]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@kms_cursor_crc@pipe-c-cursor-32x10-onscreen.html
* igt@kms_cursor_crc@pipe-c-cursor-512x512-random:
- shard-tglb: NOTRUN -> [SKIP][56] ([fdo#109279] / [i915#3359])
[56]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@kms_cursor_crc@pipe-c-cursor-512x512-random.html
* igt@kms_cursor_legacy@flip-vs-cursor-atomic-transitions-varying-size:
- shard-skl: [PASS][57] -> [FAIL][58] ([i915#2346] / [i915#533])
[57]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl8/igt@kms_cursor_legacy@flip-vs-cursor-atomic-transitions-varying-size.html
[58]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl7/igt@kms_cursor_legacy@flip-vs-cursor-atomic-transitions-varying-size.html
* igt@kms_fbcon_fbt@fbc-suspend:
- shard-tglb: [PASS][59] -> [INCOMPLETE][60] ([i915#456])
[59]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-tglb8/igt@kms_fbcon_fbt@fbc-suspend.html
[60]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb6/igt@kms_fbcon_fbt@fbc-suspend.html
* igt@kms_flip@2x-single-buffer-flip-vs-dpms-off-vs-modeset-interruptible:
- shard-iclb: NOTRUN -> [SKIP][61] ([fdo#109274])
[61]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_flip@2x-single-buffer-flip-vs-dpms-off-vs-modeset-interruptible.html
* igt@kms_flip@flip-vs-expired-vblank-interruptible@c-dp1:
- shard-apl: NOTRUN -> [FAIL][62] ([i915#79])
[62]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl7/igt@kms_flip@flip-vs-expired-vblank-interruptible@c-dp1.html
* igt@kms_flip@flip-vs-suspend-interruptible@c-dp1:
- shard-apl: [PASS][63] -> [DMESG-WARN][64] ([i915#180]) +1 similar issue
[63]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-apl8/igt@kms_flip@flip-vs-suspend-interruptible@c-dp1.html
[64]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl6/igt@kms_flip@flip-vs-suspend-interruptible@c-dp1.html
* igt@kms_flip_scaled_crc@flip-32bpp-ytile-to-32bpp-ytilegen12rcccs:
- shard-apl: NOTRUN -> [SKIP][65] ([fdo#109271] / [i915#2672]) +1 similar issue
[65]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_flip_scaled_crc@flip-32bpp-ytile-to-32bpp-ytilegen12rcccs.html
* igt@kms_frontbuffer_tracking@fbc-suspend:
- shard-kbl: [PASS][66] -> [DMESG-WARN][67] ([i915#180]) +4 similar issues
[66]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-kbl4/igt@kms_frontbuffer_tracking@fbc-suspend.html
[67]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl7/igt@kms_frontbuffer_tracking@fbc-suspend.html
* igt@kms_frontbuffer_tracking@fbcpsr-2p-primscrn-indfb-plflip-blt:
- shard-snb: NOTRUN -> [SKIP][68] ([fdo#109271]) +350 similar issues
[68]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-snb7/igt@kms_frontbuffer_tracking@fbcpsr-2p-primscrn-indfb-plflip-blt.html
* igt@kms_frontbuffer_tracking@fbcpsr-2p-primscrn-pri-indfb-draw-mmap-gtt:
- shard-iclb: NOTRUN -> [SKIP][69] ([fdo#109280]) +1 similar issue
[69]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_frontbuffer_tracking@fbcpsr-2p-primscrn-pri-indfb-draw-mmap-gtt.html
- shard-tglb: NOTRUN -> [SKIP][70] ([fdo#111825]) +1 similar issue
[70]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@kms_frontbuffer_tracking@fbcpsr-2p-primscrn-pri-indfb-draw-mmap-gtt.html
* igt@kms_frontbuffer_tracking@psr-1p-primscrn-spr-indfb-draw-mmap-cpu:
- shard-kbl: NOTRUN -> [SKIP][71] ([fdo#109271]) +48 similar issues
[71]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl3/igt@kms_frontbuffer_tracking@psr-1p-primscrn-spr-indfb-draw-mmap-cpu.html
* igt@kms_hdr@bpc-switch-dpms:
- shard-skl: [PASS][72] -> [FAIL][73] ([i915#1188])
[72]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl6/igt@kms_hdr@bpc-switch-dpms.html
[73]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl4/igt@kms_hdr@bpc-switch-dpms.html
* igt@kms_pipe_crc_basic@compare-crc-sanitycheck-pipe-d:
- shard-apl: NOTRUN -> [SKIP][74] ([fdo#109271] / [i915#533]) +4 similar issues
[74]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl7/igt@kms_pipe_crc_basic@compare-crc-sanitycheck-pipe-d.html
* igt@kms_pipe_crc_basic@suspend-read-crc-pipe-b:
- shard-skl: [PASS][75] -> [INCOMPLETE][76] ([i915#198])
[75]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl6/igt@kms_pipe_crc_basic@suspend-read-crc-pipe-b.html
[76]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl2/igt@kms_pipe_crc_basic@suspend-read-crc-pipe-b.html
* igt@kms_plane_alpha_blend@pipe-a-alpha-opaque-fb:
- shard-apl: NOTRUN -> [FAIL][77] ([fdo#108145] / [i915#265]) +3 similar issues
[77]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl8/igt@kms_plane_alpha_blend@pipe-a-alpha-opaque-fb.html
* igt@kms_plane_alpha_blend@pipe-b-alpha-7efc:
- shard-kbl: NOTRUN -> [FAIL][78] ([fdo#108145] / [i915#265])
[78]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl2/igt@kms_plane_alpha_blend@pipe-b-alpha-7efc.html
* igt@kms_plane_alpha_blend@pipe-b-alpha-transparent-fb:
- shard-apl: NOTRUN -> [FAIL][79] ([i915#265])
[79]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl2/igt@kms_plane_alpha_blend@pipe-b-alpha-transparent-fb.html
* igt@kms_plane_alpha_blend@pipe-c-constant-alpha-min:
- shard-skl: [PASS][80] -> [FAIL][81] ([fdo#108145] / [i915#265]) +1 similar issue
[80]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl4/igt@kms_plane_alpha_blend@pipe-c-constant-alpha-min.html
[81]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl10/igt@kms_plane_alpha_blend@pipe-c-constant-alpha-min.html
* igt@kms_plane_cursor@pipe-d-viewport-size-64:
- shard-iclb: NOTRUN -> [SKIP][82] ([fdo#109278]) +4 similar issues
[82]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_plane_cursor@pipe-d-viewport-size-64.html
* igt@kms_plane_lowres@pipe-d-tiling-none:
- shard-tglb: NOTRUN -> [SKIP][83] ([i915#3536])
[83]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@kms_plane_lowres@pipe-d-tiling-none.html
* igt@kms_prime@basic-crc@first-to-second:
- shard-tglb: NOTRUN -> [SKIP][84] ([i915#1836])
[84]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@kms_prime@basic-crc@first-to-second.html
* igt@kms_psr2_sf@overlay-primary-update-sf-dmg-area-1:
- shard-tglb: NOTRUN -> [SKIP][85] ([i915#2920])
[85]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@kms_psr2_sf@overlay-primary-update-sf-dmg-area-1.html
- shard-skl: NOTRUN -> [SKIP][86] ([fdo#109271] / [i915#658])
[86]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl5/igt@kms_psr2_sf@overlay-primary-update-sf-dmg-area-1.html
- shard-iclb: NOTRUN -> [SKIP][87] ([i915#658])
[87]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_psr2_sf@overlay-primary-update-sf-dmg-area-1.html
* igt@kms_psr2_sf@overlay-primary-update-sf-dmg-area-5:
- shard-apl: NOTRUN -> [SKIP][88] ([fdo#109271] / [i915#658]) +9 similar issues
[88]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl1/igt@kms_psr2_sf@overlay-primary-update-sf-dmg-area-5.html
* igt@kms_psr2_su@frontbuffer:
- shard-iclb: [PASS][89] -> [SKIP][90] ([fdo#109642] / [fdo#111068] / [i915#658])
[89]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-iclb2/igt@kms_psr2_su@frontbuffer.html
[90]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb4/igt@kms_psr2_su@frontbuffer.html
* igt@kms_psr@psr2_sprite_mmap_gtt:
- shard-iclb: NOTRUN -> [SKIP][91] ([fdo#109441])
[91]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb3/igt@kms_psr@psr2_sprite_mmap_gtt.html
* igt@kms_psr@psr2_sprite_plane_move:
- shard-iclb: [PASS][92] -> [SKIP][93] ([fdo#109441]) +3 similar issues
[92]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-iclb2/igt@kms_psr@psr2_sprite_plane_move.html
[93]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb4/igt@kms_psr@psr2_sprite_plane_move.html
* igt@kms_sysfs_edid_timing:
- shard-apl: NOTRUN -> [FAIL][94] ([IGT#2])
[94]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl6/igt@kms_sysfs_edid_timing.html
* igt@kms_vblank@pipe-d-wait-idle:
- shard-kbl: NOTRUN -> [SKIP][95] ([fdo#109271] / [i915#533])
[95]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl2/igt@kms_vblank@pipe-d-wait-idle.html
* igt@kms_writeback@writeback-check-output:
- shard-iclb: NOTRUN -> [SKIP][96] ([i915#2437])
[96]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb6/igt@kms_writeback@writeback-check-output.html
- shard-skl: NOTRUN -> [SKIP][97] ([fdo#109271] / [i915#2437])
[97]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl5/igt@kms_writeback@writeback-check-output.html
- shard-tglb: NOTRUN -> [SKIP][98] ([i915#2437])
[98]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@kms_writeback@writeback-check-output.html
* igt@kms_writeback@writeback-fb-id:
- shard-apl: NOTRUN -> [SKIP][99] ([fdo#109271] / [i915#2437])
[99]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl6/igt@kms_writeback@writeback-fb-id.html
* igt@nouveau_crc@pipe-d-ctx-flip-detection:
- shard-tglb: NOTRUN -> [SKIP][100] ([i915#2530])
[100]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@nouveau_crc@pipe-d-ctx-flip-detection.html
* igt@perf@polling-parameterized:
- shard-skl: [PASS][101] -> [FAIL][102] ([i915#1542])
[101]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl8/igt@perf@polling-parameterized.html
[102]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl8/igt@perf@polling-parameterized.html
* igt@prime_nv_api@i915_nv_reimport_twice_check_flink_name:
- shard-tglb: NOTRUN -> [SKIP][103] ([fdo#109291])
[103]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@prime_nv_api@i915_nv_reimport_twice_check_flink_name.html
* igt@prime_nv_test@i915_import_gtt_mmap:
- shard-iclb: NOTRUN -> [SKIP][104] ([fdo#109291])
[104]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-iclb3/igt@prime_nv_test@i915_import_gtt_mmap.html
* igt@sysfs_clients@fair-1:
- shard-apl: NOTRUN -> [SKIP][105] ([fdo#109271] / [i915#2994]) +7 similar issues
[105]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-apl8/igt@sysfs_clients@fair-1.html
- shard-skl: NOTRUN -> [SKIP][106] ([fdo#109271] / [i915#2994])
[106]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl4/igt@sysfs_clients@fair-1.html
* igt@sysfs_clients@sema-10:
- shard-tglb: NOTRUN -> [SKIP][107] ([i915#2994]) +1 similar issue
[107]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@sysfs_clients@sema-10.html
* igt@sysfs_clients@sema-50:
- shard-kbl: NOTRUN -> [SKIP][108] ([fdo#109271] / [i915#2994])
[108]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl2/igt@sysfs_clients@sema-50.html
#### Possible fixes ####
* igt@fbdev@unaligned-write:
- {shard-rkl}: [SKIP][109] ([i915#2582]) -> [PASS][110]
[109]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-1/igt@fbdev@unaligned-write.html
[110]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@fbdev@unaligned-write.html
* igt@gem_eio@in-flight-contexts-1us:
- shard-tglb: [TIMEOUT][111] ([i915#3063]) -> [PASS][112]
[111]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-tglb7/igt@gem_eio@in-flight-contexts-1us.html
[112]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb2/igt@gem_eio@in-flight-contexts-1us.html
* igt@gem_eio@unwedge-stress:
- shard-tglb: [TIMEOUT][113] ([i915#2369] / [i915#3063] / [i915#3648]) -> [PASS][114]
[113]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-tglb1/igt@gem_eio@unwedge-stress.html
[114]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-tglb7/igt@gem_eio@unwedge-stress.html
* igt@gem_mmap_gtt@cpuset-big-copy-xy:
- shard-glk: [FAIL][115] ([i915#1888] / [i915#307]) -> [PASS][116]
[115]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-glk4/igt@gem_mmap_gtt@cpuset-big-copy-xy.html
[116]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-glk1/igt@gem_mmap_gtt@cpuset-big-copy-xy.html
* igt@i915_pm_rpm@gem-execbuf:
- {shard-rkl}: [SKIP][117] ([fdo#109308]) -> [PASS][118] +1 similar issue
[117]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-1/igt@i915_pm_rpm@gem-execbuf.html
[118]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@i915_pm_rpm@gem-execbuf.html
* igt@i915_pm_rpm@modeset-lpsp-stress-no-wait:
- {shard-rkl}: [SKIP][119] ([i915#1397]) -> [PASS][120]
[119]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-2/igt@i915_pm_rpm@modeset-lpsp-stress-no-wait.html
[120]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@i915_pm_rpm@modeset-lpsp-stress-no-wait.html
* igt@i915_suspend@fence-restore-untiled:
- shard-kbl: [DMESG-WARN][121] ([i915#180]) -> [PASS][122] +1 similar issue
[121]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-kbl4/igt@i915_suspend@fence-restore-untiled.html
[122]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-kbl3/igt@i915_suspend@fence-restore-untiled.html
* igt@kms_async_flips@alternate-sync-async-flip:
- shard-skl: [FAIL][123] ([i915#2521]) -> [PASS][124]
[123]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-skl2/igt@kms_async_flips@alternate-sync-async-flip.html
[124]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-skl10/igt@kms_async_flips@alternate-sync-async-flip.html
* igt@kms_big_fb@x-tiled-32bpp-rotate-180:
- {shard-rkl}: [SKIP][125] ([i915#3638]) -> [PASS][126] +3 similar issues
[125]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-2/igt@kms_big_fb@x-tiled-32bpp-rotate-180.html
[126]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@kms_big_fb@x-tiled-32bpp-rotate-180.html
- shard-glk: [DMESG-WARN][127] ([i915#118] / [i915#95]) -> [PASS][128]
[127]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-glk4/igt@kms_big_fb@x-tiled-32bpp-rotate-180.html
[128]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-glk7/igt@kms_big_fb@x-tiled-32bpp-rotate-180.html
* igt@kms_big_fb@x-tiled-max-hw-stride-32bpp-rotate-0:
- {shard-rkl}: [SKIP][129] ([i915#3721]) -> [PASS][130]
[129]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-1/igt@kms_big_fb@x-tiled-max-hw-stride-32bpp-rotate-0.html
[130]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@kms_big_fb@x-tiled-max-hw-stride-32bpp-rotate-0.html
* igt@kms_big_fb@y-tiled-16bpp-rotate-90:
- {shard-rkl}: [SKIP][131] ([fdo#111614]) -> [PASS][132]
[131]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-2/igt@kms_big_fb@y-tiled-16bpp-rotate-90.html
[132]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@kms_big_fb@y-tiled-16bpp-rotate-90.html
* igt@kms_color@pipe-a-ctm-0-75:
- {shard-rkl}: [SKIP][133] ([i915#1149] / [i915#1849]) -> [PASS][134] +3 similar issues
[133]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-2/igt@kms_color@pipe-a-ctm-0-75.html
[134]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@kms_color@pipe-a-ctm-0-75.html
* igt@kms_cursor_crc@pipe-a-cursor-256x256-rapid-movement:
- {shard-rkl}: [SKIP][135] ([fdo#112022]) -> [PASS][136] +8 similar issues
[135]: https://intel-gfx-ci.01.org/tree/drm-tip/CI_DRM_10525/shard-rkl-2/igt@kms_cursor_crc@pipe-a-cursor-256x256-rapid-movement.html
[136]: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/shard-rkl-6/igt@kms_cursor_crc@pipe-a-cursor-256x256-rapid-movement.html
== Logs ==
For more details see: https://intel-gfx-ci.01.org/tree/drm-tip/Patchwork_20903/index.html
[-- Attachment #2: Type: text/html, Size: 33693 bytes --]
^ permalink raw reply [flat|nested] 9+ messages in thread
* [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts
@ 2022-02-24 13:03 Tvrtko Ursulin
0 siblings, 0 replies; 9+ messages in thread
From: Tvrtko Ursulin @ 2022-02-24 13:03 UTC (permalink / raw)
To: Intel-gfx
Cc: Matthew Brost, Tvrtko Ursulin, dri-devel, Zhen Han, Chris Wilson,
John Harrison
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
When a non-persistent context exits we currently mark it as banned in
order to trigger fast termination of any outstanding GPU jobs it may have
left running.
In doing so we apply a very strict 1ms limit in which the left over job
has to preempt before we issues an engine resets.
Some workloads are not able to cleanly preempt in that time window and it
can be argued that it would instead be better to give them a bit more
grace since avoiding engine resets is generally preferrable.
To achieve this the patch splits handling of banned contexts from simply
closed non-persistent ones and then applies different timeouts for both
and also extends the criteria which determines if a request should be
scheduled back in after preemption or not.
20ms preempt timeout grace is given to exited non-persistent contexts
which have been empirically tested to satisfy customers requirements
and still provides reasonably quick cleanup post exit.
v2:
* Streamline fast path checks.
v3:
* Simplify by using only schedulable status.
* Increase timeout to 20ms.
v4:
* Fix live_execlists selftest.
v5:
* Fix logic in kill_engines.
v6:
* Rebase.
v7:
* Add GuC support.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Zhen Han <zhen.han@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: John Harrison <John.C.Harrison@Intel.com>
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 22 +++++++++++-----
drivers/gpu/drm/i915/gt/intel_context.c | 25 ++++++++++++++++++
drivers/gpu/drm/i915/gt/intel_context.h | 26 ++++++++++++++-----
drivers/gpu/drm/i915/gt/intel_context_types.h | 4 ++-
.../drm/i915/gt/intel_execlists_submission.c | 13 +++++++---
.../gpu/drm/i915/gt/intel_ring_submission.c | 7 ++---
.../gpu/drm/i915/gt/uc/intel_guc_submission.c | 15 ++++++-----
drivers/gpu/drm/i915/i915_request.c | 2 +-
8 files changed, 86 insertions(+), 28 deletions(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index bc6d59df064d..3a61ec753894 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -1332,7 +1332,8 @@ static struct intel_engine_cs *active_engine(struct intel_context *ce)
return engine;
}
-static void kill_engines(struct i915_gem_engines *engines, bool ban)
+static void
+kill_engines(struct i915_gem_engines *engines, bool ban, bool persistent)
{
struct i915_gem_engines_iter it;
struct intel_context *ce;
@@ -1346,8 +1347,15 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
*/
for_each_gem_engine(ce, engines, it) {
struct intel_engine_cs *engine;
+ bool skip = false;
- if (ban && intel_context_ban(ce, NULL))
+ if (ban)
+ skip = intel_context_ban(ce, NULL);
+ else if (!persistent)
+ skip = intel_context_exit_nonpersistent(ce, NULL);
+
+ /* Already banned or non-persistent closed. */
+ if (skip)
continue;
/*
@@ -1360,7 +1368,7 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
engine = active_engine(ce);
/* First attempt to gracefully cancel the context */
- if (engine && !__cancel_engine(engine) && ban)
+ if (engine && !__cancel_engine(engine) && (ban || !persistent))
/*
* If we are unable to send a preemptive pulse to bump
* the context from the GPU, we have to resort to a full
@@ -1372,8 +1380,6 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
static void kill_context(struct i915_gem_context *ctx)
{
- bool ban = (!i915_gem_context_is_persistent(ctx) ||
- !ctx->i915->params.enable_hangcheck);
struct i915_gem_engines *pos, *next;
spin_lock_irq(&ctx->stale.lock);
@@ -1386,7 +1392,8 @@ static void kill_context(struct i915_gem_context *ctx)
spin_unlock_irq(&ctx->stale.lock);
- kill_engines(pos, ban);
+ kill_engines(pos, !ctx->i915->params.enable_hangcheck,
+ i915_gem_context_is_persistent(ctx));
spin_lock_irq(&ctx->stale.lock);
GEM_BUG_ON(i915_sw_fence_signaled(&pos->fence));
@@ -1432,7 +1439,8 @@ static void engines_idle_release(struct i915_gem_context *ctx,
kill:
if (list_empty(&engines->link)) /* raced, already closed */
- kill_engines(engines, true);
+ kill_engines(engines, true,
+ i915_gem_context_is_persistent(ctx));
i915_sw_fence_commit(&engines->fence);
}
diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
index 5d0ec7c49b6a..27cd71c13097 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.c
+++ b/drivers/gpu/drm/i915/gt/intel_context.c
@@ -576,6 +576,31 @@ void intel_context_bind_parent_child(struct intel_context *parent,
child->parallel.parent = parent;
}
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq)
+{
+ bool ret = intel_context_set_banned(ce);
+
+ trace_intel_context_ban(ce);
+
+ if (ce->ops->revoke)
+ ce->ops->revoke(ce, rq,
+ INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS);
+
+ return ret;
+}
+
+bool intel_context_exit_nonpersistent(struct intel_context *ce,
+ struct i915_request *rq)
+{
+ bool ret = intel_context_set_exiting(ce);
+
+ if (ce->ops->revoke)
+ ce->ops->revoke(ce, rq,
+ INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS);
+
+ return ret;
+}
+
#if IS_ENABLED(CONFIG_DRM_I915_SELFTEST)
#include "selftest_context.c"
#endif
diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
index d8c74bbf9aae..11c95e9d76ab 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.h
+++ b/drivers/gpu/drm/i915/gt/intel_context.h
@@ -25,6 +25,9 @@
##__VA_ARGS__); \
} while (0)
+#define INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS (20)
+#define INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS (1)
+
struct i915_gem_ww_ctx;
void intel_context_init(struct intel_context *ce,
@@ -309,18 +312,27 @@ static inline bool intel_context_set_banned(struct intel_context *ce)
return test_and_set_bit(CONTEXT_BANNED, &ce->flags);
}
-static inline bool intel_context_ban(struct intel_context *ce,
- struct i915_request *rq)
+bool intel_context_ban(struct intel_context *ce, struct i915_request *rq);
+
+static inline bool intel_context_is_schedulable(const struct intel_context *ce)
{
- bool ret = intel_context_set_banned(ce);
+ return !test_bit(CONTEXT_EXITING, &ce->flags) &&
+ !test_bit(CONTEXT_BANNED, &ce->flags);
+}
- trace_intel_context_ban(ce);
- if (ce->ops->ban)
- ce->ops->ban(ce, rq);
+static inline bool intel_context_is_exiting(const struct intel_context *ce)
+{
+ return test_bit(CONTEXT_EXITING, &ce->flags);
+}
- return ret;
+static inline bool intel_context_set_exiting(struct intel_context *ce)
+{
+ return test_and_set_bit(CONTEXT_EXITING, &ce->flags);
}
+bool intel_context_exit_nonpersistent(struct intel_context *ce,
+ struct i915_request *rq);
+
static inline bool
intel_context_force_single_submission(const struct intel_context *ce)
{
diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
index 30cd81ad8911..34073430cf8a 100644
--- a/drivers/gpu/drm/i915/gt/intel_context_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
@@ -37,7 +37,8 @@ struct intel_context_ops {
int (*alloc)(struct intel_context *ce);
- void (*ban)(struct intel_context *ce, struct i915_request *rq);
+ void (*revoke)(struct intel_context *ce, struct i915_request *rq,
+ unsigned int preempt_timeout_ms);
int (*pre_pin)(struct intel_context *ce, struct i915_gem_ww_ctx *ww, void **vaddr);
int (*pin)(struct intel_context *ce, void *vaddr);
@@ -119,6 +120,7 @@ struct intel_context {
#define CONTEXT_GUC_INIT 10
#define CONTEXT_PERMA_PIN 11
#define CONTEXT_IS_PARKING 12
+#define CONTEXT_EXITING 13
struct {
u64 timeout_us;
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
index 961d795220a3..953d1e7453c6 100644
--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -481,7 +481,7 @@ __execlists_schedule_in(struct i915_request *rq)
!intel_engine_has_heartbeat(engine)))
intel_context_set_banned(ce);
- if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
+ if (unlikely(!intel_context_is_schedulable(ce) || bad_request(rq)))
reset_active(rq, engine);
if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
@@ -1225,12 +1225,19 @@ static void record_preemption(struct intel_engine_execlists *execlists)
static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
const struct i915_request *rq)
{
+ struct intel_context *ce;
+
if (!rq)
return 0;
+ ce = rq->context;
+
/* Force a fast reset for terminated contexts (ignoring sysfs!) */
- if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq)))
- return 1;
+ if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
+ return INTEL_CONTEXT_BANNED_PREEMPT_TIMEOUT_MS;
+ /* Longer grace for closed non-persistent contexts to avoid resets. */
+ else if (unlikely(intel_context_is_exiting(ce)))
+ return INTEL_CONTEXT_EXITING_PREEMPT_TIMEOUT_MS;
return READ_ONCE(engine->props.preempt_timeout_ms);
}
diff --git a/drivers/gpu/drm/i915/gt/intel_ring_submission.c b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
index 6d7ec3bf1f32..a3dcc227a702 100644
--- a/drivers/gpu/drm/i915/gt/intel_ring_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_ring_submission.c
@@ -596,8 +596,9 @@ static void ring_context_reset(struct intel_context *ce)
clear_bit(CONTEXT_VALID_BIT, &ce->flags);
}
-static void ring_context_ban(struct intel_context *ce,
- struct i915_request *rq)
+static void ring_context_revoke(struct intel_context *ce,
+ struct i915_request *rq,
+ unsigned int preempt_timeout_ms)
{
struct intel_engine_cs *engine;
@@ -632,7 +633,7 @@ static const struct intel_context_ops ring_context_ops = {
.cancel_request = ring_context_cancel_request,
- .ban = ring_context_ban,
+ .revoke = ring_context_revoke,
.pre_pin = ring_context_pre_pin,
.pin = ring_context_pin,
diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
index b3a429a92c0d..23fdbc7b34ab 100644
--- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
+++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
@@ -2602,7 +2602,9 @@ static void __guc_context_set_preemption_timeout(struct intel_guc *guc,
intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
}
-static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
+static void
+guc_context_revoke(struct intel_context *ce, struct i915_request *rq,
+ unsigned int preempt_timeout_ms)
{
struct intel_guc *guc = ce_to_guc(ce);
struct intel_runtime_pm *runtime_pm =
@@ -2641,7 +2643,8 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
* gets kicked off the HW ASAP.
*/
with_intel_runtime_pm(runtime_pm, wakeref) {
- __guc_context_set_preemption_timeout(guc, guc_id, 1);
+ __guc_context_set_preemption_timeout(guc, guc_id,
+ preempt_timeout_ms);
__guc_context_sched_disable(guc, ce, guc_id);
}
} else {
@@ -2649,7 +2652,7 @@ static void guc_context_ban(struct intel_context *ce, struct i915_request *rq)
with_intel_runtime_pm(runtime_pm, wakeref)
__guc_context_set_preemption_timeout(guc,
ce->guc_id.id,
- 1);
+ preempt_timeout_ms);
spin_unlock_irqrestore(&ce->guc_state.lock, flags);
}
}
@@ -2998,7 +3001,7 @@ static const struct intel_context_ops guc_context_ops = {
.unpin = guc_context_unpin,
.post_unpin = guc_context_post_unpin,
- .ban = guc_context_ban,
+ .revoke = guc_context_revoke,
.cancel_request = guc_context_cancel_request,
@@ -3247,7 +3250,7 @@ static const struct intel_context_ops virtual_guc_context_ops = {
.unpin = guc_virtual_context_unpin,
.post_unpin = guc_context_post_unpin,
- .ban = guc_context_ban,
+ .revoke = guc_context_revoke,
.cancel_request = guc_context_cancel_request,
@@ -3336,7 +3339,7 @@ static const struct intel_context_ops virtual_parent_context_ops = {
.unpin = guc_parent_context_unpin,
.post_unpin = guc_context_post_unpin,
- .ban = guc_context_ban,
+ .revoke = guc_context_revoke,
.cancel_request = guc_context_cancel_request,
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index 582770360ad1..13c555d190a0 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -611,7 +611,7 @@ bool __i915_request_submit(struct i915_request *request)
goto active;
}
- if (unlikely(intel_context_is_banned(request->context)))
+ if (unlikely(!intel_context_is_schedulable(request->context)))
i915_request_set_error_once(request, -EIO);
if (unlikely(fatal_error(request->fence.error)))
--
2.32.0
^ permalink raw reply related [flat|nested] 9+ messages in thread
* [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts
@ 2021-08-05 12:05 Tvrtko Ursulin
0 siblings, 0 replies; 9+ messages in thread
From: Tvrtko Ursulin @ 2021-08-05 12:05 UTC (permalink / raw)
To: Intel-gfx; +Cc: dri-devel, Tvrtko Ursulin, Chris Wilson, Zhen Han
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
When a non-persistent context exits we currently mark it as banned in
order to trigger fast termination of any outstanding GPU jobs it may have
left running.
In doing so we apply a very strict 1ms limit in which the left over job
has to preempt before we issues an engine resets.
Some workloads are not able to cleanly preempt in that time window and it
can be argued that it would instead be better to give them a bit more
grace since avoiding engine resets is generally preferrable.
To achieve this the patch splits handling of banned contexts from simply
closed non-persistent ones and then applies different timeouts for both
and also extends the criteria which determines if a request should be
scheduled back in after preemption or not.
15ms preempt timeout grace is given to exited non-persistent contexts
which have been empirically tested to satisfy customers requirements
and still provides reasonably quick cleanup post exit.
v2:
* Streamline fast path checks.
v3:
* Simplify by using only schedulable status.
* Increase timeout to 20ms.
v4:
* Fix live_execlists selftest.
v5:
* Fix logic in kill_engines.
v6:
* Rebase.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Zhen Han <zhen.han@intel.com>
---
drivers/gpu/drm/i915/gem/i915_gem_context.c | 22 +++++++++++++------
drivers/gpu/drm/i915/gt/intel_context.c | 2 ++
drivers/gpu/drm/i915/gt/intel_context.h | 17 +++++++++++++-
drivers/gpu/drm/i915/gt/intel_context_types.h | 1 +
.../drm/i915/gt/intel_execlists_submission.c | 11 ++++++++--
drivers/gpu/drm/i915/gt/selftest_execlists.c | 20 +++++++++++------
drivers/gpu/drm/i915/i915_request.c | 2 +-
7 files changed, 57 insertions(+), 18 deletions(-)
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_context.c b/drivers/gpu/drm/i915/gem/i915_gem_context.c
index cff72679ad7c..21fe5d4057ab 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_context.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_context.c
@@ -1065,7 +1065,8 @@ static struct intel_engine_cs *active_engine(struct intel_context *ce)
return engine;
}
-static void kill_engines(struct i915_gem_engines *engines, bool ban)
+static void
+kill_engines(struct i915_gem_engines *engines, bool ban, bool persistent)
{
struct i915_gem_engines_iter it;
struct intel_context *ce;
@@ -1079,8 +1080,15 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
*/
for_each_gem_engine(ce, engines, it) {
struct intel_engine_cs *engine;
+ bool skip = false;
+
+ if (ban)
+ skip = intel_context_ban(ce, NULL);
+ else if (!persistent)
+ skip = !intel_context_clear_schedulable(ce);
- if (ban && intel_context_ban(ce, NULL))
+ /* Already previously banned or made non-schedulable? */
+ if (skip)
continue;
/*
@@ -1093,7 +1101,7 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
engine = active_engine(ce);
/* First attempt to gracefully cancel the context */
- if (engine && !__cancel_engine(engine) && ban)
+ if (engine && !__cancel_engine(engine) && (ban || !persistent))
/*
* If we are unable to send a preemptive pulse to bump
* the context from the GPU, we have to resort to a full
@@ -1105,8 +1113,6 @@ static void kill_engines(struct i915_gem_engines *engines, bool ban)
static void kill_context(struct i915_gem_context *ctx)
{
- bool ban = (!i915_gem_context_is_persistent(ctx) ||
- !ctx->i915->params.enable_hangcheck);
struct i915_gem_engines *pos, *next;
spin_lock_irq(&ctx->stale.lock);
@@ -1119,7 +1125,8 @@ static void kill_context(struct i915_gem_context *ctx)
spin_unlock_irq(&ctx->stale.lock);
- kill_engines(pos, ban);
+ kill_engines(pos, !ctx->i915->params.enable_hangcheck,
+ i915_gem_context_is_persistent(ctx));
spin_lock_irq(&ctx->stale.lock);
GEM_BUG_ON(i915_sw_fence_signaled(&pos->fence));
@@ -1165,7 +1172,8 @@ static void engines_idle_release(struct i915_gem_context *ctx,
kill:
if (list_empty(&engines->link)) /* raced, already closed */
- kill_engines(engines, true);
+ kill_engines(engines, true,
+ i915_gem_context_is_persistent(ctx));
i915_sw_fence_commit(&engines->fence);
}
diff --git a/drivers/gpu/drm/i915/gt/intel_context.c b/drivers/gpu/drm/i915/gt/intel_context.c
index 745e84c72c90..bc1701ef1578 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.c
+++ b/drivers/gpu/drm/i915/gt/intel_context.c
@@ -382,6 +382,8 @@ intel_context_init(struct intel_context *ce, struct intel_engine_cs *engine)
ce->ring = NULL;
ce->ring_size = SZ_4K;
+ __set_bit(CONTEXT_SCHEDULABLE, &ce->flags);
+
ewma_runtime_init(&ce->runtime.avg);
ce->vm = i915_vm_get(engine->gt->vm);
diff --git a/drivers/gpu/drm/i915/gt/intel_context.h b/drivers/gpu/drm/i915/gt/intel_context.h
index c41098950746..5b50716654dd 100644
--- a/drivers/gpu/drm/i915/gt/intel_context.h
+++ b/drivers/gpu/drm/i915/gt/intel_context.h
@@ -251,7 +251,22 @@ static inline bool intel_context_is_banned(const struct intel_context *ce)
static inline bool intel_context_set_banned(struct intel_context *ce)
{
- return test_and_set_bit(CONTEXT_BANNED, &ce->flags);
+ bool banned = test_and_set_bit(CONTEXT_BANNED, &ce->flags);
+
+ if (!banned)
+ clear_bit(CONTEXT_SCHEDULABLE, &ce->flags);
+
+ return banned;
+}
+
+static inline bool intel_context_clear_schedulable(struct intel_context *ce)
+{
+ return test_and_clear_bit(CONTEXT_SCHEDULABLE, &ce->flags);
+}
+
+static inline bool intel_context_is_schedulable(const struct intel_context *ce)
+{
+ return test_bit(CONTEXT_SCHEDULABLE, &ce->flags);
}
static inline bool intel_context_ban(struct intel_context *ce,
diff --git a/drivers/gpu/drm/i915/gt/intel_context_types.h b/drivers/gpu/drm/i915/gt/intel_context_types.h
index e54351a170e2..3306c70c9c54 100644
--- a/drivers/gpu/drm/i915/gt/intel_context_types.h
+++ b/drivers/gpu/drm/i915/gt/intel_context_types.h
@@ -112,6 +112,7 @@ struct intel_context {
#define CONTEXT_FORCE_SINGLE_SUBMISSION 7
#define CONTEXT_NOPREEMPT 8
#define CONTEXT_LRCA_DIRTY 9
+#define CONTEXT_SCHEDULABLE 10 /* Unless banned or non-persistent closed. */
struct {
u64 timeout_us;
diff --git a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
index de5f9c86b9a4..778f3cda3c71 100644
--- a/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
+++ b/drivers/gpu/drm/i915/gt/intel_execlists_submission.c
@@ -478,7 +478,7 @@ __execlists_schedule_in(struct i915_request *rq)
!intel_engine_has_heartbeat(engine)))
intel_context_set_banned(ce);
- if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
+ if (unlikely(!intel_context_is_schedulable(ce) || bad_request(rq)))
reset_active(rq, engine);
if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
@@ -1222,12 +1222,19 @@ static void record_preemption(struct intel_engine_execlists *execlists)
static unsigned long active_preempt_timeout(struct intel_engine_cs *engine,
const struct i915_request *rq)
{
+ struct intel_context *ce;
+
if (!rq)
return 0;
+ ce = rq->context;
+
/* Force a fast reset for terminated contexts (ignoring sysfs!) */
- if (unlikely(intel_context_is_banned(rq->context) || bad_request(rq)))
+ if (unlikely(intel_context_is_banned(ce) || bad_request(rq)))
return 1;
+ /* Longer grace for closed non-persistent contexts to avoid resets. */
+ else if (unlikely(!intel_context_is_schedulable(ce)))
+ return 20;
return READ_ONCE(engine->props.preempt_timeout_ms);
}
diff --git a/drivers/gpu/drm/i915/gt/selftest_execlists.c b/drivers/gpu/drm/i915/gt/selftest_execlists.c
index f12ffe797639..da36c015caf4 100644
--- a/drivers/gpu/drm/i915/gt/selftest_execlists.c
+++ b/drivers/gpu/drm/i915/gt/selftest_execlists.c
@@ -2050,6 +2050,12 @@ struct live_preempt_cancel {
struct preempt_client a, b;
};
+static void context_clear_banned(struct intel_context *ce)
+{
+ clear_bit(CONTEXT_BANNED, &ce->flags);
+ set_bit(CONTEXT_SCHEDULABLE, &ce->flags);
+}
+
static int __cancel_active0(struct live_preempt_cancel *arg)
{
struct i915_request *rq;
@@ -2068,7 +2074,7 @@ static int __cancel_active0(struct live_preempt_cancel *arg)
if (IS_ERR(rq))
return PTR_ERR(rq);
- clear_bit(CONTEXT_BANNED, &rq->context->flags);
+ context_clear_banned(rq->context);
i915_request_get(rq);
i915_request_add(rq);
if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
@@ -2112,7 +2118,7 @@ static int __cancel_active1(struct live_preempt_cancel *arg)
if (IS_ERR(rq[0]))
return PTR_ERR(rq[0]);
- clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
+ context_clear_banned(rq[0]->context);
i915_request_get(rq[0]);
i915_request_add(rq[0]);
if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
@@ -2128,7 +2134,7 @@ static int __cancel_active1(struct live_preempt_cancel *arg)
goto out;
}
- clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
+ context_clear_banned(rq[1]->context);
i915_request_get(rq[1]);
err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
i915_request_add(rq[1]);
@@ -2183,7 +2189,7 @@ static int __cancel_queued(struct live_preempt_cancel *arg)
if (IS_ERR(rq[0]))
return PTR_ERR(rq[0]);
- clear_bit(CONTEXT_BANNED, &rq[0]->context->flags);
+ context_clear_banned(rq[0]->context);
i915_request_get(rq[0]);
i915_request_add(rq[0]);
if (!igt_wait_for_spinner(&arg->a.spin, rq[0])) {
@@ -2197,7 +2203,7 @@ static int __cancel_queued(struct live_preempt_cancel *arg)
goto out;
}
- clear_bit(CONTEXT_BANNED, &rq[1]->context->flags);
+ context_clear_banned(rq[1]->context);
i915_request_get(rq[1]);
err = i915_request_await_dma_fence(rq[1], &rq[0]->fence);
i915_request_add(rq[1]);
@@ -2273,7 +2279,7 @@ static int __cancel_hostile(struct live_preempt_cancel *arg)
if (IS_ERR(rq))
return PTR_ERR(rq);
- clear_bit(CONTEXT_BANNED, &rq->context->flags);
+ context_clear_banned(rq->context);
i915_request_get(rq);
i915_request_add(rq);
if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
@@ -2329,7 +2335,7 @@ static int __cancel_fail(struct live_preempt_cancel *arg)
if (IS_ERR(rq))
return PTR_ERR(rq);
- clear_bit(CONTEXT_BANNED, &rq->context->flags);
+ context_clear_banned(rq->context);
i915_request_get(rq);
i915_request_add(rq);
if (!igt_wait_for_spinner(&arg->a.spin, rq)) {
diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c
index ce446716d092..b1a9bec83339 100644
--- a/drivers/gpu/drm/i915/i915_request.c
+++ b/drivers/gpu/drm/i915/i915_request.c
@@ -583,7 +583,7 @@ bool __i915_request_submit(struct i915_request *request)
goto active;
}
- if (unlikely(intel_context_is_banned(request->context)))
+ if (unlikely(!intel_context_is_schedulable(request->context)))
i915_request_set_error_once(request, -EIO);
if (unlikely(fatal_error(request->fence.error)))
--
2.30.2
^ permalink raw reply related [flat|nested] 9+ messages in thread
end of thread, other threads:[~2022-02-24 13:03 UTC | newest]
Thread overview: 9+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-08-26 10:52 [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts Tvrtko Ursulin
2021-08-26 10:52 ` [Intel-gfx] " Tvrtko Ursulin
2021-08-26 13:32 ` Daniel Vetter
2021-08-26 13:32 ` [Intel-gfx] " Daniel Vetter
2021-08-26 14:57 ` [Intel-gfx] ✗ Fi.CI.CHECKPATCH: warning for drm/i915: Be more gentle when exiting non-persistent contexts (rev2) Patchwork
2021-08-26 15:28 ` [Intel-gfx] ✓ Fi.CI.BAT: success " Patchwork
2021-08-26 23:22 ` [Intel-gfx] ✗ Fi.CI.IGT: failure " Patchwork
-- strict thread matches above, loose matches on Subject: below --
2022-02-24 13:03 [PATCH] drm/i915: Be more gentle when exiting non-persistent contexts Tvrtko Ursulin
2021-08-05 12:05 Tvrtko Ursulin
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.