If we rewind the RING_TAIL on a context, due to a preemption event, we must force the context restore for the RING_TAIL update to be properly handled. Rather than note which preemption events may cause us to rewind the tail, compare the new request's tail with the previously submitted RING_TAIL, as it turns out that timeslicing was causing unexpected rewinds.
<0>[ 1283.366080] <idle>-0 0d.s2 1280851190us : __execlists_submission_tasklet: 0000:00:02.0 rcs0: expired last=130:4698, prio=3, hint=3 <0>[ 1283.366114] <idle>-0 0d.s2 1280851192us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 66:119966, current 119964 <0>[ 1283.366148] <idle>-0 0d.s2 1280851195us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 130:4698, current 4695 <0>[ 1283.366183] <idle>-0 0d.s2 1280851198us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 130:4696, current 4695 <0>[ 1283.366216] <idle>-0 0d.s2 1280851208us : __i915_request_submit: 0000:00:02.0 rcs0: fence 130:4696, current 4695 <0>[ 1283.366250] <idle>-0 0d.s2 1280851213us : __i915_request_submit: 0000:00:02.0 rcs0: fence 134:1508, current 1506 <0>[ 1283.366278] <idle>-0 0d.s2 1280851219us : trace_ports: 0000:00:02.0 rcs0: submit { 130:4696*, 134:1508 } <0>[ 1283.366306] synmark2-5425 2..s. 1280851239us : process_csb: 0000:00:02.0 rcs0: cs-irq head=5, tail=0 <0>[ 1283.366333] synmark2-5425 2..s. 1280851240us : process_csb: 0000:00:02.0 rcs0: csb[0]: status=0x00008002:0x00000000 <0>[ 1283.366361] synmark2-5425 2..s. 1280851243us : trace_ports: 0000:00:02.0 rcs0: preempted { 130:4698, 66:119966 } <0>[ 1283.366387] synmark2-5425 2..s. 1280851246us : trace_ports: 0000:00:02.0 rcs0: promote { 130:4696*, 134:1508 } <0>[ 1283.366441] synmark2-5425 2.... 1280851462us : __i915_request_commit: 0000:00:02.0 rcs0: fence 130:4700, current 4695 <0>[ 1283.366502] synmark2-5425 2.... 1280852111us : __i915_request_commit: 0000:00:02.0 rcs0: fence 130:4702, current 4695 <0>[ 1283.366529] synmark2-5425 2.Ns1 1280852296us : process_csb: 0000:00:02.0 rcs0: cs-irq head=0, tail=2 <0>[ 1283.366557] synmark2-5425 2.Ns1 1280852297us : process_csb: 0000:00:02.0 rcs0: csb[1]: status=0x00000814:0x00000000 <0>[ 1283.366585] synmark2-5425 2.Ns1 1280852299us : trace_ports: 0000:00:02.0 rcs0: completed { 130:4696!, 134:1508 } <0>[ 1283.366612] synmark2-5425 2.Ns1 1280852301us : process_csb: 0000:00:02.0 rcs0: csb[2]: status=0x00000818:0x00000040 <0>[ 1283.366639] synmark2-5425 2.Ns1 1280852302us : trace_ports: 0000:00:02.0 rcs0: completed { 134:1508, 0:0 } <0>[ 1283.366666] synmark2-5425 2.Ns1 1280852313us : process_csb: process_csb:2336 GEM_BUG_ON(!i915_request_completed(*execlists->active) && !reset_in_progress(execlists))
Fixes: 8ee36e048c98 ("drm/i915/execlists: Minimalistic timeslicing") Referenecs: 82c69bf58650 ("drm/i915/gt: Detect if we miss WaIdleLiteRestore") Signed-off-by: Chris Wilson chris@chris-wilson.co.uk Cc: Mika Kuoppala mika.kuoppala@linux.intel.com Reviewed-by: Mika Kuoppala mika.kuoppala@linux.intel.com Cc: stable@vger.kernel.org # v5.4+ --- drivers/gpu/drm/i915/gt/intel_lrc.c | 18 ++++++++---------- drivers/gpu/drm/i915/gt/intel_ring.c | 1 + drivers/gpu/drm/i915/gt/intel_ring.h | 8 ++++++++ drivers/gpu/drm/i915/gt/intel_ring_types.h | 1 + 4 files changed, 18 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index ed1e4d883d47..b5baaea2c535 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -1321,7 +1321,7 @@ static u64 execlists_update_context(struct i915_request *rq) { struct intel_context *ce = rq->context; u64 desc = ce->lrc_desc; - u32 tail; + u32 tail, prev;
/* * WaIdleLiteRestore:bdw,skl @@ -1334,9 +1334,15 @@ static u64 execlists_update_context(struct i915_request *rq) * subsequent resubmissions (for lite restore). Should that fail us, * and we try and submit the same tail again, force the context * reload. + * + * If we need to return to a preempted context, we need to skip the + * lite-restore and force it to reload the RING_TAIL. Otherwise, the + * HW has a tendency to ignore us rewinding the TAIL to the end of + * an earlier request. */ tail = intel_ring_set_tail(rq->ring, rq->tail); - if (unlikely(ce->lrc_reg_state[CTX_RING_TAIL] == tail)) + prev = ce->lrc_reg_state[CTX_RING_TAIL]; + if (unlikely(intel_ring_direction(rq->ring, tail, prev) <= 0)) desc |= CTX_DESC_FORCE_RESTORE; ce->lrc_reg_state[CTX_RING_TAIL] = tail; rq->tail = rq->wa_tail; @@ -1854,14 +1860,6 @@ static void execlists_dequeue(struct intel_engine_cs *engine) */ __unwind_incomplete_requests(engine);
- /* - * If we need to return to the preempted context, we - * need to skip the lite-restore and force it to - * reload the RING_TAIL. Otherwise, the HW has a - * tendency to ignore us rewinding the TAIL to the - * end of an earlier request. - */ - last->context->lrc_desc |= CTX_DESC_FORCE_RESTORE; last = NULL; } else if (need_timeslice(engine, last) && timer_expired(&engine->execlists.timer)) { diff --git a/drivers/gpu/drm/i915/gt/intel_ring.c b/drivers/gpu/drm/i915/gt/intel_ring.c index 366013367526..8cda1b7e17ba 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring.c +++ b/drivers/gpu/drm/i915/gt/intel_ring.c @@ -143,6 +143,7 @@ intel_engine_create_ring(struct intel_engine_cs *engine, int size)
kref_init(&ring->ref); ring->size = size; + ring->wrap = BITS_PER_TYPE(ring->size) - ilog2(size);
/* * Workaround an erratum on the i830 which causes a hang if diff --git a/drivers/gpu/drm/i915/gt/intel_ring.h b/drivers/gpu/drm/i915/gt/intel_ring.h index ea2839d9e044..5bdce24994aa 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring.h +++ b/drivers/gpu/drm/i915/gt/intel_ring.h @@ -56,6 +56,14 @@ static inline u32 intel_ring_wrap(const struct intel_ring *ring, u32 pos) return pos & (ring->size - 1); }
+static inline int intel_ring_direction(const struct intel_ring *ring, + u32 next, u32 prev) +{ + typecheck(typeof(ring->size), next); + typecheck(typeof(ring->size), prev); + return (next - prev) << ring->wrap; +} + static inline bool intel_ring_offset_valid(const struct intel_ring *ring, unsigned int pos) diff --git a/drivers/gpu/drm/i915/gt/intel_ring_types.h b/drivers/gpu/drm/i915/gt/intel_ring_types.h index d9f17f38e0cc..3cd7fec7fd8d 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring_types.h +++ b/drivers/gpu/drm/i915/gt/intel_ring_types.h @@ -45,6 +45,7 @@ struct intel_ring {
u32 space; u32 size; + u32 wrap; u32 effective_size; };
If we rewind the RING_TAIL on a context, due to a preemption event, we must force the context restore for the RING_TAIL update to be properly handled. Rather than note which preemption events may cause us to rewind the tail, compare the new request's tail with the previously submitted RING_TAIL, as it turns out that timeslicing was causing unexpected rewinds.
<idle>-0 0d.s2 1280851190us : __execlists_submission_tasklet: 0000:00:02.0 rcs0: expired last=130:4698, prio=3, hint=3 <idle>-0 0d.s2 1280851192us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 66:119966, current 119964 <idle>-0 0d.s2 1280851195us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 130:4698, current 4695 ^---- Note we unwind 2 requests from the same context
<idle>-0 0d.s2 1280851198us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 130:4696, current 4695 <idle>-0 0d.s2 1280851208us : __i915_request_submit: 0000:00:02.0 rcs0: fence 130:4696, current 4695 <idle>-0 0d.s2 1280851213us : __i915_request_submit: 0000:00:02.0 rcs0: fence 134:1508, current 1506 ^---- But to apply the new timeslice, we have to replay the first request before he new client can start
<idle>-0 0d.s2 1280851219us : trace_ports: 0000:00:02.0 rcs0: submit { 130:4696*, 134:1508 } synmark2-5425 2..s. 1280851239us : process_csb: 0000:00:02.0 rcs0: cs-irq head=5, tail=0 synmark2-5425 2..s. 1280851240us : process_csb: 0000:00:02.0 rcs0: csb[0]: status=0x00008002:0x00000000 ^---- Preemption event for the ELSP update; note the lite-restore
synmark2-5425 2..s. 1280851243us : trace_ports: 0000:00:02.0 rcs0: preempted { 130:4698, 66:119966 } synmark2-5425 2..s. 1280851246us : trace_ports: 0000:00:02.0 rcs0: promote { 130:4696*, 134:1508 } synmark2-5425 2.... 1280851462us : __i915_request_commit: 0000:00:02.0 rcs0: fence 130:4700, current 4695 synmark2-5425 2.... 1280852111us : __i915_request_commit: 0000:00:02.0 rcs0: fence 130:4702, current 4695 synmark2-5425 2.Ns1 1280852296us : process_csb: 0000:00:02.0 rcs0: cs-irq head=0, tail=2 synmark2-5425 2.Ns1 1280852297us : process_csb: 0000:00:02.0 rcs0: csb[1]: status=0x00000814:0x00000000 synmark2-5425 2.Ns1 1280852299us : trace_ports: 0000:00:02.0 rcs0: completed { 130:4696!, 134:1508 } synmark2-5425 2.Ns1 1280852301us : process_csb: 0000:00:02.0 rcs0: csb[2]: status=0x00000818:0x00000040 synmark2-5425 2.Ns1 1280852302us : trace_ports: 0000:00:02.0 rcs0: completed { 134:1508, 0:0 } synmark2-5425 2.Ns1 1280852313us : process_csb: process_csb:2336 GEM_BUG_ON(!i915_request_completed(*execlists->active) && !reset_in_progress(execlists))
Fixes: 8ee36e048c98 ("drm/i915/execlists: Minimalistic timeslicing") Referenecs: 82c69bf58650 ("drm/i915/gt: Detect if we miss WaIdleLiteRestore") Signed-off-by: Chris Wilson chris@chris-wilson.co.uk Cc: Mika Kuoppala mika.kuoppala@linux.intel.com Reviewed-by: Mika Kuoppala mika.kuoppala@linux.intel.com Cc: stable@vger.kernel.org # v5.4+ --- drivers/gpu/drm/i915/gt/intel_lrc.c | 18 ++++++++---------- drivers/gpu/drm/i915/gt/intel_ring.c | 1 + drivers/gpu/drm/i915/gt/intel_ring.h | 8 ++++++++ drivers/gpu/drm/i915/gt/intel_ring_types.h | 1 + 4 files changed, 18 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index ed1e4d883d47..b5baaea2c535 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -1321,7 +1321,7 @@ static u64 execlists_update_context(struct i915_request *rq) { struct intel_context *ce = rq->context; u64 desc = ce->lrc_desc; - u32 tail; + u32 tail, prev;
/* * WaIdleLiteRestore:bdw,skl @@ -1334,9 +1334,15 @@ static u64 execlists_update_context(struct i915_request *rq) * subsequent resubmissions (for lite restore). Should that fail us, * and we try and submit the same tail again, force the context * reload. + * + * If we need to return to a preempted context, we need to skip the + * lite-restore and force it to reload the RING_TAIL. Otherwise, the + * HW has a tendency to ignore us rewinding the TAIL to the end of + * an earlier request. */ tail = intel_ring_set_tail(rq->ring, rq->tail); - if (unlikely(ce->lrc_reg_state[CTX_RING_TAIL] == tail)) + prev = ce->lrc_reg_state[CTX_RING_TAIL]; + if (unlikely(intel_ring_direction(rq->ring, tail, prev) <= 0)) desc |= CTX_DESC_FORCE_RESTORE; ce->lrc_reg_state[CTX_RING_TAIL] = tail; rq->tail = rq->wa_tail; @@ -1854,14 +1860,6 @@ static void execlists_dequeue(struct intel_engine_cs *engine) */ __unwind_incomplete_requests(engine);
- /* - * If we need to return to the preempted context, we - * need to skip the lite-restore and force it to - * reload the RING_TAIL. Otherwise, the HW has a - * tendency to ignore us rewinding the TAIL to the - * end of an earlier request. - */ - last->context->lrc_desc |= CTX_DESC_FORCE_RESTORE; last = NULL; } else if (need_timeslice(engine, last) && timer_expired(&engine->execlists.timer)) { diff --git a/drivers/gpu/drm/i915/gt/intel_ring.c b/drivers/gpu/drm/i915/gt/intel_ring.c index 366013367526..8cda1b7e17ba 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring.c +++ b/drivers/gpu/drm/i915/gt/intel_ring.c @@ -143,6 +143,7 @@ intel_engine_create_ring(struct intel_engine_cs *engine, int size)
kref_init(&ring->ref); ring->size = size; + ring->wrap = BITS_PER_TYPE(ring->size) - ilog2(size);
/* * Workaround an erratum on the i830 which causes a hang if diff --git a/drivers/gpu/drm/i915/gt/intel_ring.h b/drivers/gpu/drm/i915/gt/intel_ring.h index ea2839d9e044..5bdce24994aa 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring.h +++ b/drivers/gpu/drm/i915/gt/intel_ring.h @@ -56,6 +56,14 @@ static inline u32 intel_ring_wrap(const struct intel_ring *ring, u32 pos) return pos & (ring->size - 1); }
+static inline int intel_ring_direction(const struct intel_ring *ring, + u32 next, u32 prev) +{ + typecheck(typeof(ring->size), next); + typecheck(typeof(ring->size), prev); + return (next - prev) << ring->wrap; +} + static inline bool intel_ring_offset_valid(const struct intel_ring *ring, unsigned int pos) diff --git a/drivers/gpu/drm/i915/gt/intel_ring_types.h b/drivers/gpu/drm/i915/gt/intel_ring_types.h index d9f17f38e0cc..3cd7fec7fd8d 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring_types.h +++ b/drivers/gpu/drm/i915/gt/intel_ring_types.h @@ -45,6 +45,7 @@ struct intel_ring {
u32 space; u32 size; + u32 wrap; u32 effective_size; };
If we rewind the RING_TAIL on a context, due to a preemption event, we must force the context restore for the RING_TAIL update to be properly handled. Rather than note which preemption events may cause us to rewind the tail, compare the new request's tail with the previously submitted RING_TAIL, as it turns out that timeslicing was causing unexpected rewinds.
<idle>-0 0d.s2 1280851190us : __execlists_submission_tasklet: 0000:00:02.0 rcs0: expired last=130:4698, prio=3, hint=3 <idle>-0 0d.s2 1280851192us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 66:119966, current 119964 <idle>-0 0d.s2 1280851195us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 130:4698, current 4695 <idle>-0 0d.s2 1280851198us : __i915_request_unsubmit: 0000:00:02.0 rcs0: fence 130:4696, current 4695 ^---- Note we unwind 2 requests from the same context
<idle>-0 0d.s2 1280851208us : __i915_request_submit: 0000:00:02.0 rcs0: fence 130:4696, current 4695 <idle>-0 0d.s2 1280851213us : __i915_request_submit: 0000:00:02.0 rcs0: fence 134:1508, current 1506 ^---- But to apply the new timeslice, we have to replay the first request before the new client can start
<idle>-0 0d.s2 1280851219us : trace_ports: 0000:00:02.0 rcs0: submit { 130:4696*, 134:1508 } synmark2-5425 2..s. 1280851239us : process_csb: 0000:00:02.0 rcs0: cs-irq head=5, tail=0 synmark2-5425 2..s. 1280851240us : process_csb: 0000:00:02.0 rcs0: csb[0]: status=0x00008002:0x00000000 ^---- Preemption event for the ELSP update; note the lite-restore
synmark2-5425 2..s. 1280851243us : trace_ports: 0000:00:02.0 rcs0: preempted { 130:4698, 66:119966 } synmark2-5425 2..s. 1280851246us : trace_ports: 0000:00:02.0 rcs0: promote { 130:4696*, 134:1508 } synmark2-5425 2.... 1280851462us : __i915_request_commit: 0000:00:02.0 rcs0: fence 130:4700, current 4695 synmark2-5425 2.... 1280852111us : __i915_request_commit: 0000:00:02.0 rcs0: fence 130:4702, current 4695 synmark2-5425 2.Ns1 1280852296us : process_csb: 0000:00:02.0 rcs0: cs-irq head=0, tail=2 synmark2-5425 2.Ns1 1280852297us : process_csb: 0000:00:02.0 rcs0: csb[1]: status=0x00000814:0x00000000 synmark2-5425 2.Ns1 1280852299us : trace_ports: 0000:00:02.0 rcs0: completed { 130:4696!, 134:1508 } synmark2-5425 2.Ns1 1280852301us : process_csb: 0000:00:02.0 rcs0: csb[2]: status=0x00000818:0x00000040 synmark2-5425 2.Ns1 1280852302us : trace_ports: 0000:00:02.0 rcs0: completed { 134:1508, 0:0 } synmark2-5425 2.Ns1 1280852313us : process_csb: process_csb:2336 GEM_BUG_ON(!i915_request_completed(*execlists->active) && !reset_in_progress(execlists))
Fixes: 8ee36e048c98 ("drm/i915/execlists: Minimalistic timeslicing") Referenecs: 82c69bf58650 ("drm/i915/gt: Detect if we miss WaIdleLiteRestore") Signed-off-by: Chris Wilson chris@chris-wilson.co.uk Cc: Mika Kuoppala mika.kuoppala@linux.intel.com Reviewed-by: Mika Kuoppala mika.kuoppala@linux.intel.com Cc: stable@vger.kernel.org # v5.4+ --- drivers/gpu/drm/i915/gt/intel_lrc.c | 18 ++++++++---------- drivers/gpu/drm/i915/gt/intel_ring.c | 1 + drivers/gpu/drm/i915/gt/intel_ring.h | 8 ++++++++ drivers/gpu/drm/i915/gt/intel_ring_types.h | 1 + 4 files changed, 18 insertions(+), 10 deletions(-)
diff --git a/drivers/gpu/drm/i915/gt/intel_lrc.c b/drivers/gpu/drm/i915/gt/intel_lrc.c index ed1e4d883d47..b5baaea2c535 100644 --- a/drivers/gpu/drm/i915/gt/intel_lrc.c +++ b/drivers/gpu/drm/i915/gt/intel_lrc.c @@ -1321,7 +1321,7 @@ static u64 execlists_update_context(struct i915_request *rq) { struct intel_context *ce = rq->context; u64 desc = ce->lrc_desc; - u32 tail; + u32 tail, prev;
/* * WaIdleLiteRestore:bdw,skl @@ -1334,9 +1334,15 @@ static u64 execlists_update_context(struct i915_request *rq) * subsequent resubmissions (for lite restore). Should that fail us, * and we try and submit the same tail again, force the context * reload. + * + * If we need to return to a preempted context, we need to skip the + * lite-restore and force it to reload the RING_TAIL. Otherwise, the + * HW has a tendency to ignore us rewinding the TAIL to the end of + * an earlier request. */ tail = intel_ring_set_tail(rq->ring, rq->tail); - if (unlikely(ce->lrc_reg_state[CTX_RING_TAIL] == tail)) + prev = ce->lrc_reg_state[CTX_RING_TAIL]; + if (unlikely(intel_ring_direction(rq->ring, tail, prev) <= 0)) desc |= CTX_DESC_FORCE_RESTORE; ce->lrc_reg_state[CTX_RING_TAIL] = tail; rq->tail = rq->wa_tail; @@ -1854,14 +1860,6 @@ static void execlists_dequeue(struct intel_engine_cs *engine) */ __unwind_incomplete_requests(engine);
- /* - * If we need to return to the preempted context, we - * need to skip the lite-restore and force it to - * reload the RING_TAIL. Otherwise, the HW has a - * tendency to ignore us rewinding the TAIL to the - * end of an earlier request. - */ - last->context->lrc_desc |= CTX_DESC_FORCE_RESTORE; last = NULL; } else if (need_timeslice(engine, last) && timer_expired(&engine->execlists.timer)) { diff --git a/drivers/gpu/drm/i915/gt/intel_ring.c b/drivers/gpu/drm/i915/gt/intel_ring.c index 366013367526..8cda1b7e17ba 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring.c +++ b/drivers/gpu/drm/i915/gt/intel_ring.c @@ -143,6 +143,7 @@ intel_engine_create_ring(struct intel_engine_cs *engine, int size)
kref_init(&ring->ref); ring->size = size; + ring->wrap = BITS_PER_TYPE(ring->size) - ilog2(size);
/* * Workaround an erratum on the i830 which causes a hang if diff --git a/drivers/gpu/drm/i915/gt/intel_ring.h b/drivers/gpu/drm/i915/gt/intel_ring.h index ea2839d9e044..5bdce24994aa 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring.h +++ b/drivers/gpu/drm/i915/gt/intel_ring.h @@ -56,6 +56,14 @@ static inline u32 intel_ring_wrap(const struct intel_ring *ring, u32 pos) return pos & (ring->size - 1); }
+static inline int intel_ring_direction(const struct intel_ring *ring, + u32 next, u32 prev) +{ + typecheck(typeof(ring->size), next); + typecheck(typeof(ring->size), prev); + return (next - prev) << ring->wrap; +} + static inline bool intel_ring_offset_valid(const struct intel_ring *ring, unsigned int pos) diff --git a/drivers/gpu/drm/i915/gt/intel_ring_types.h b/drivers/gpu/drm/i915/gt/intel_ring_types.h index d9f17f38e0cc..3cd7fec7fd8d 100644 --- a/drivers/gpu/drm/i915/gt/intel_ring_types.h +++ b/drivers/gpu/drm/i915/gt/intel_ring_types.h @@ -45,6 +45,7 @@ struct intel_ring {
u32 space; u32 size; + u32 wrap; u32 effective_size; };
linux-stable-mirror@lists.linaro.org