mirror of
https://github.com/AuxXxilium/linux_dsm_epyc7002.git
synced 2024-12-23 13:00:33 +07:00
5ba32c7be8
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 -- the unexpected RING_TAIL rewind <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:8ee36e048c
("drm/i915/execlists: Minimalistic timeslicing") Referenecs:82c69bf586
("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+ Link: https://patchwork.freedesktop.org/patch/msgid/20200207211452.2860634-1-chris@chris-wilson.co.uk
140 lines
3.9 KiB
C
140 lines
3.9 KiB
C
/*
|
|
* SPDX-License-Identifier: MIT
|
|
*
|
|
* Copyright © 2019 Intel Corporation
|
|
*/
|
|
|
|
#ifndef INTEL_RING_H
|
|
#define INTEL_RING_H
|
|
|
|
#include "i915_gem.h" /* GEM_BUG_ON */
|
|
#include "i915_request.h"
|
|
#include "intel_ring_types.h"
|
|
|
|
struct intel_engine_cs;
|
|
|
|
struct intel_ring *
|
|
intel_engine_create_ring(struct intel_engine_cs *engine, int size);
|
|
|
|
u32 *intel_ring_begin(struct i915_request *rq, unsigned int num_dwords);
|
|
int intel_ring_cacheline_align(struct i915_request *rq);
|
|
|
|
unsigned int intel_ring_update_space(struct intel_ring *ring);
|
|
|
|
int intel_ring_pin(struct intel_ring *ring);
|
|
void intel_ring_unpin(struct intel_ring *ring);
|
|
void intel_ring_reset(struct intel_ring *ring, u32 tail);
|
|
|
|
void intel_ring_free(struct kref *ref);
|
|
|
|
static inline struct intel_ring *intel_ring_get(struct intel_ring *ring)
|
|
{
|
|
kref_get(&ring->ref);
|
|
return ring;
|
|
}
|
|
|
|
static inline void intel_ring_put(struct intel_ring *ring)
|
|
{
|
|
kref_put(&ring->ref, intel_ring_free);
|
|
}
|
|
|
|
static inline void intel_ring_advance(struct i915_request *rq, u32 *cs)
|
|
{
|
|
/* Dummy function.
|
|
*
|
|
* This serves as a placeholder in the code so that the reader
|
|
* can compare against the preceding intel_ring_begin() and
|
|
* check that the number of dwords emitted matches the space
|
|
* reserved for the command packet (i.e. the value passed to
|
|
* intel_ring_begin()).
|
|
*/
|
|
GEM_BUG_ON((rq->ring->vaddr + rq->ring->emit) != cs);
|
|
}
|
|
|
|
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)
|
|
{
|
|
if (pos & -ring->size) /* must be strictly within the ring */
|
|
return false;
|
|
|
|
if (!IS_ALIGNED(pos, 8)) /* must be qword aligned */
|
|
return false;
|
|
|
|
return true;
|
|
}
|
|
|
|
static inline u32 intel_ring_offset(const struct i915_request *rq, void *addr)
|
|
{
|
|
/* Don't write ring->size (equivalent to 0) as that hangs some GPUs. */
|
|
u32 offset = addr - rq->ring->vaddr;
|
|
GEM_BUG_ON(offset > rq->ring->size);
|
|
return intel_ring_wrap(rq->ring, offset);
|
|
}
|
|
|
|
static inline void
|
|
assert_ring_tail_valid(const struct intel_ring *ring, unsigned int tail)
|
|
{
|
|
GEM_BUG_ON(!intel_ring_offset_valid(ring, tail));
|
|
|
|
/*
|
|
* "Ring Buffer Use"
|
|
* Gen2 BSpec "1. Programming Environment" / 1.4.4.6
|
|
* Gen3 BSpec "1c Memory Interface Functions" / 2.3.4.5
|
|
* Gen4+ BSpec "1c Memory Interface and Command Stream" / 5.3.4.5
|
|
* "If the Ring Buffer Head Pointer and the Tail Pointer are on the
|
|
* same cacheline, the Head Pointer must not be greater than the Tail
|
|
* Pointer."
|
|
*
|
|
* We use ring->head as the last known location of the actual RING_HEAD,
|
|
* it may have advanced but in the worst case it is equally the same
|
|
* as ring->head and so we should never program RING_TAIL to advance
|
|
* into the same cacheline as ring->head.
|
|
*/
|
|
#define cacheline(a) round_down(a, CACHELINE_BYTES)
|
|
GEM_BUG_ON(cacheline(tail) == cacheline(ring->head) &&
|
|
tail < ring->head);
|
|
#undef cacheline
|
|
}
|
|
|
|
static inline unsigned int
|
|
intel_ring_set_tail(struct intel_ring *ring, unsigned int tail)
|
|
{
|
|
/* Whilst writes to the tail are strictly order, there is no
|
|
* serialisation between readers and the writers. The tail may be
|
|
* read by i915_request_retire() just as it is being updated
|
|
* by execlists, as although the breadcrumb is complete, the context
|
|
* switch hasn't been seen.
|
|
*/
|
|
assert_ring_tail_valid(ring, tail);
|
|
ring->tail = tail;
|
|
return tail;
|
|
}
|
|
|
|
static inline unsigned int
|
|
__intel_ring_space(unsigned int head, unsigned int tail, unsigned int size)
|
|
{
|
|
/*
|
|
* "If the Ring Buffer Head Pointer and the Tail Pointer are on the
|
|
* same cacheline, the Head Pointer must not be greater than the Tail
|
|
* Pointer."
|
|
*/
|
|
GEM_BUG_ON(!is_power_of_2(size));
|
|
return (head - tail - CACHELINE_BYTES) & (size - 1);
|
|
}
|
|
|
|
#endif /* INTEL_RING_H */
|