From: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
To: Matthew Brost <matthew.brost@intel.com>
Cc: jason.ekstrand@intel.com, daniel.vetter@intel.com,
intel-gfx@lists.freedesktop.org, dri-devel@lists.freedesktop.org
Subject: Re: [Intel-gfx] [RFC PATCH 55/97] drm/i915/guc: Update intel_gt_wait_for_idle to work with GuC
Date: Wed, 26 May 2021 10:21:05 +0100 [thread overview]
Message-ID: <5f84fcc9-5c8c-d44b-3739-5b970aef7eb4@linux.intel.com> (raw)
In-Reply-To: <20210525170718.GB14724@sdutt-i7>
On 25/05/2021 18:07, Matthew Brost wrote:
> On Tue, May 25, 2021 at 11:06:00AM +0100, Tvrtko Ursulin wrote:
>>
>> On 06/05/2021 20:14, Matthew Brost wrote:
>>> When running the GuC the GPU can't be considered idle if the GuC still
>>> has contexts pinned. As such, a call has been added in
>>> intel_gt_wait_for_idle to idle the UC and in turn the GuC by waiting for
>>> the number of unpinned contexts to go to zero.
>>>
>>> Cc: John Harrison <john.c.harrison@intel.com>
>>> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
>>> ---
>>> drivers/gpu/drm/i915/gem/i915_gem_mman.c | 3 +-
>>> drivers/gpu/drm/i915/gt/intel_gt.c | 18 ++++
>>> drivers/gpu/drm/i915/gt/intel_gt.h | 2 +
>>> drivers/gpu/drm/i915/gt/intel_gt_requests.c | 22 ++---
>>> drivers/gpu/drm/i915/gt/intel_gt_requests.h | 7 +-
>>> drivers/gpu/drm/i915/gt/uc/intel_guc.h | 4 +
>>> drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c | 1 +
>>> drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h | 4 +
>>> .../gpu/drm/i915/gt/uc/intel_guc_submission.c | 91 ++++++++++++++++++-
>>> drivers/gpu/drm/i915/gt/uc/intel_uc.h | 5 +
>>> drivers/gpu/drm/i915/i915_debugfs.c | 1 +
>>> drivers/gpu/drm/i915/i915_gem_evict.c | 1 +
>>> .../gpu/drm/i915/selftests/igt_live_test.c | 2 +-
>>> .../gpu/drm/i915/selftests/mock_gem_device.c | 3 +-
>>> 14 files changed, 137 insertions(+), 27 deletions(-)
>>>
>>> diff --git a/drivers/gpu/drm/i915/gem/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
>>> index 8598a1c78a4c..2f5295c9408d 100644
>>> --- a/drivers/gpu/drm/i915/gem/i915_gem_mman.c
>>> +++ b/drivers/gpu/drm/i915/gem/i915_gem_mman.c
>>> @@ -634,7 +634,8 @@ mmap_offset_attach(struct drm_i915_gem_object *obj,
>>> goto insert;
>>> /* Attempt to reap some mmap space from dead objects */
>>> - err = intel_gt_retire_requests_timeout(&i915->gt, MAX_SCHEDULE_TIMEOUT);
>>> + err = intel_gt_retire_requests_timeout(&i915->gt, MAX_SCHEDULE_TIMEOUT,
>>> + NULL);
>>> if (err)
>>> goto err;
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c
>>> index 8d77dcbad059..1742a8561f69 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_gt.c
>>> +++ b/drivers/gpu/drm/i915/gt/intel_gt.c
>>> @@ -574,6 +574,24 @@ static void __intel_gt_disable(struct intel_gt *gt)
>>> GEM_BUG_ON(intel_gt_pm_is_awake(gt));
>>> }
>>> +int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout)
>>> +{
>>> + long rtimeout;
>>> +
>>> + /* If the device is asleep, we have no requests outstanding */
>>> + if (!intel_gt_pm_is_awake(gt))
>>> + return 0;
>>> +
>>> + while ((timeout = intel_gt_retire_requests_timeout(gt, timeout,
>>> + &rtimeout)) > 0) {
>>> + cond_resched();
>>> + if (signal_pending(current))
>>> + return -EINTR;
>>> + }
>>> +
>>> + return timeout ? timeout : intel_uc_wait_for_idle(>->uc, rtimeout);
>>> +}
>>> +
>>> int intel_gt_init(struct intel_gt *gt)
>>> {
>>> int err;
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_gt.h b/drivers/gpu/drm/i915/gt/intel_gt.h
>>> index 7ec395cace69..c775043334bf 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_gt.h
>>> +++ b/drivers/gpu/drm/i915/gt/intel_gt.h
>>> @@ -48,6 +48,8 @@ void intel_gt_driver_release(struct intel_gt *gt);
>>> void intel_gt_driver_late_release(struct intel_gt *gt);
>>> +int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout);
>>> +
>>> void intel_gt_check_and_clear_faults(struct intel_gt *gt);
>>> void intel_gt_clear_error_registers(struct intel_gt *gt,
>>> intel_engine_mask_t engine_mask);
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.c b/drivers/gpu/drm/i915/gt/intel_gt_requests.c
>>> index 647eca9d867a..c6c702f236fa 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_gt_requests.c
>>> +++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.c
>>> @@ -13,6 +13,7 @@
>>> #include "intel_gt_pm.h"
>>> #include "intel_gt_requests.h"
>>> #include "intel_timeline.h"
>>> +#include "uc/intel_uc.h"
>>> static bool retire_requests(struct intel_timeline *tl)
>>> {
>>> @@ -130,7 +131,8 @@ void intel_engine_fini_retire(struct intel_engine_cs *engine)
>>> GEM_BUG_ON(engine->retire);
>>> }
>>> -long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout)
>>> +long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout,
>>> + long *rtimeout)
>>
>> What is 'rtimeout', I know remaining, but it can be more self-descriptive to
>> start with.
>>
>
> 'remaining_timeout' it is.
>
>> It feels a bit churny for what it is. How plausible would be alternatives to
>> either change existing timeout to in/out, or measure sleep internally in
>> this function, or just risk sleeping twice as long by passing the original
>> timeout to uc idle as well?
>>
>
> Originally had it just passing in the same value, got review feedback
> saying I should pass in the adjusted value. Hard to make everyone happy.
Ok.
>
>>> {
>>> struct intel_gt_timelines *timelines = >->timelines;
>>> struct intel_timeline *tl, *tn;
>>> @@ -195,22 +197,10 @@ out_active: spin_lock(&timelines->lock);
>>> if (flush_submission(gt, timeout)) /* Wait, there's more! */
>>> active_count++;
>>> - return active_count ? timeout : 0;
>>> -}
>>> -
>>> -int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout)
>>> -{
>>> - /* If the device is asleep, we have no requests outstanding */
>>> - if (!intel_gt_pm_is_awake(gt))
>>> - return 0;
>>> -
>>> - while ((timeout = intel_gt_retire_requests_timeout(gt, timeout)) > 0) {
>>> - cond_resched();
>>> - if (signal_pending(current))
>>> - return -EINTR;
>>> - }
>>> + if (rtimeout)
>>> + *rtimeout = timeout;
>>> - return timeout;
>>> + return active_count ? timeout : 0;
>>> }
>>> static void retire_work_handler(struct work_struct *work)
>>> diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.h b/drivers/gpu/drm/i915/gt/intel_gt_requests.h
>>> index fcc30a6e4fe9..4419787124e2 100644
>>> --- a/drivers/gpu/drm/i915/gt/intel_gt_requests.h
>>> +++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.h
>>> @@ -10,10 +10,11 @@ struct intel_engine_cs;
>>> struct intel_gt;
>>> struct intel_timeline;
>>> -long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout);
>>> +long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout,
>>> + long *rtimeout);
>>> static inline void intel_gt_retire_requests(struct intel_gt *gt)
>>> {
>>> - intel_gt_retire_requests_timeout(gt, 0);
>>> + intel_gt_retire_requests_timeout(gt, 0, NULL);
>>> }
>>> void intel_engine_init_retire(struct intel_engine_cs *engine);
>>> @@ -21,8 +22,6 @@ void intel_engine_add_retire(struct intel_engine_cs *engine,
>>> struct intel_timeline *tl);
>>> void intel_engine_fini_retire(struct intel_engine_cs *engine);
>>> -int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout);
>>> -
>>> void intel_gt_init_requests(struct intel_gt *gt);
>>> void intel_gt_park_requests(struct intel_gt *gt);
>>> void intel_gt_unpark_requests(struct intel_gt *gt);
>>> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc.h b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
>>> index 485e98f3f304..47eaa69809e8 100644
>>> --- a/drivers/gpu/drm/i915/gt/uc/intel_guc.h
>>> +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc.h
>>> @@ -38,6 +38,8 @@ struct intel_guc {
>>> spinlock_t irq_lock;
>>> unsigned int msg_enabled_mask;
>>> + atomic_t outstanding_submission_g2h;
>>> +
>>> struct {
>>> bool enabled;
>>> void (*reset)(struct intel_guc *guc);
>>> @@ -239,6 +241,8 @@ static inline void intel_guc_disable_msg(struct intel_guc *guc, u32 mask)
>>> spin_unlock_irq(&guc->irq_lock);
>>> }
>>> +int intel_guc_wait_for_idle(struct intel_guc *guc, long timeout);
>>> +
>>> int intel_guc_reset_engine(struct intel_guc *guc,
>>> struct intel_engine_cs *engine);
>>> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
>>> index f1893030ca88..cf701056fa14 100644
>>> --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
>>> +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.c
>>> @@ -111,6 +111,7 @@ void intel_guc_ct_init_early(struct intel_guc_ct *ct)
>>> INIT_LIST_HEAD(&ct->requests.incoming);
>>> INIT_WORK(&ct->requests.worker, ct_incoming_request_worker_func);
>>> tasklet_init(&ct->receive_tasklet, ct_receive_tasklet_func, (unsigned long)ct);
>>> + init_waitqueue_head(&ct->wq);
>>> }
>>> static inline const char *guc_ct_buffer_type_to_str(u32 type)
>>> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
>>> index 660bf37238e2..ab1b79ab960b 100644
>>> --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
>>> +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_ct.h
>>> @@ -10,6 +10,7 @@
>>> #include <linux/spinlock.h>
>>> #include <linux/workqueue.h>
>>> #include <linux/ktime.h>
>>> +#include <linux/wait.h>
>>> #include "intel_guc_fwif.h"
>>> @@ -68,6 +69,9 @@ struct intel_guc_ct {
>>> struct tasklet_struct receive_tasklet;
>>> + /** @wq: wait queue for g2h chanenl */
>>> + wait_queue_head_t wq;
>>> +
>>> struct {
>>> u16 last_fence; /* last fence used to send request */
>>> 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 ae0b386467e3..0ff7dd6d337d 100644
>>> --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
>>> +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c
>>> @@ -253,6 +253,74 @@ static inline void set_lrc_desc_registered(struct intel_guc *guc, u32 id,
>>> xa_store_irq(&guc->context_lookup, id, ce, GFP_ATOMIC);
>>> }
>>> +static int guc_submission_busy_loop(struct intel_guc* guc,
>>> + const u32 *action,
>>> + u32 len,
>>> + u32 g2h_len_dw,
>>> + bool loop)
>>> +{
>>> + int err;
>>> +
>>> + err = intel_guc_send_busy_loop(guc, action, len, g2h_len_dw, loop);
>>> +
>>> + if (!err && g2h_len_dw)
>>> + atomic_inc(&guc->outstanding_submission_g2h);
>>> +
>>> + return err;
>>> +}
>>> +
>>> +static int guc_wait_for_pending_msg(struct intel_guc *guc,
>>> + atomic_t *wait_var,
>>> + bool interruptible,
>>> + long timeout)
>>> +{
>>> + const int state = interruptible ?
>>> + TASK_INTERRUPTIBLE : TASK_UNINTERRUPTIBLE;
>>> + DEFINE_WAIT(wait);
>>> +
>>> + might_sleep();
>>> + GEM_BUG_ON(timeout < 0);
>>> +
>>> + if (!atomic_read(wait_var))
>>> + return 0;
>>> +
>>> + if (!timeout)
>>> + return -ETIME;
>>> +
>>> + for (;;) {
>>> + prepare_to_wait(&guc->ct.wq, &wait, state);
>>> +
>>> + if (!atomic_read(wait_var))
>>> + break;
>>> +
>>> + if (signal_pending_state(state, current)) {
>>> + timeout = -ERESTARTSYS;
>>> + break;
>>> + }
>>> +
>>> + if (!timeout) {
>>> + timeout = -ETIME;
>>> + break;
>>> + }
>>> +
>>> + timeout = io_schedule_timeout(timeout);
>>> + }
>>> + finish_wait(&guc->ct.wq, &wait);
>>> +
>>> + return (timeout < 0) ? timeout : 0;
>>> +}
>>
>> See if it is possible to simplify all this with wait_var_event and
>> wake_up_var.
>>
>
> Let me check on that.
>
>>> +
>>> +int intel_guc_wait_for_idle(struct intel_guc *guc, long timeout)
>>> +{
>>> + bool interruptible = true;
>>> +
>>> + if (unlikely(timeout < 0))
>>> + timeout = -timeout, interruptible = false;
>>> +
>>> + return guc_wait_for_pending_msg(guc, &guc->outstanding_submission_g2h,
>>> + interruptible, timeout);
>>> +}
>>> +
>>> static int guc_add_request(struct intel_guc *guc, struct i915_request *rq)
>>> {
>>> int err;
>>> @@ -279,6 +347,7 @@ static int guc_add_request(struct intel_guc *guc, struct i915_request *rq)
>>> err = intel_guc_send_nb(guc, action, len, g2h_len_dw);
>>> if (!enabled && !err) {
>>> + atomic_inc(&guc->outstanding_submission_g2h);
>>> set_context_enabled(ce);
>>> } else if (!enabled) {
>>> clr_context_pending_enable(ce);
>>> @@ -734,7 +803,7 @@ static int __guc_action_register_context(struct intel_guc *guc,
>>> offset,
>>> };
>>> - return intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
>>> + return guc_submission_busy_loop(guc, action, ARRAY_SIZE(action), 0, true);
>>> }
>>> static int register_context(struct intel_context *ce)
>>> @@ -754,7 +823,7 @@ static int __guc_action_deregister_context(struct intel_guc *guc,
>>> guc_id,
>>> };
>>> - return intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action),
>>> + return guc_submission_busy_loop(guc, action, ARRAY_SIZE(action),
>>> G2H_LEN_DW_DEREGISTER_CONTEXT, true);
>>> }
>>> @@ -871,7 +940,9 @@ static int guc_context_pin(struct intel_context *ce, void *vaddr)
>>> static void guc_context_unpin(struct intel_context *ce)
>>> {
>>> - unpin_guc_id(ce_to_guc(ce), ce);
>>> + struct intel_guc *guc = ce_to_guc(ce);
>>> +
>>> + unpin_guc_id(guc, ce);
>>> lrc_unpin(ce);
>>> }
>>> @@ -894,7 +965,7 @@ static void __guc_context_sched_disable(struct intel_guc *guc,
>>> intel_context_get(ce);
>>> - intel_guc_send_busy_loop(guc, action, ARRAY_SIZE(action),
>>> + guc_submission_busy_loop(guc, action, ARRAY_SIZE(action),
>>> G2H_LEN_DW_SCHED_CONTEXT_MODE_SET, true);
>>> }
>>> @@ -1437,6 +1508,15 @@ g2h_context_lookup(struct intel_guc *guc, u32 desc_idx)
>>> return ce;
>>> }
>>> +static void decr_outstanding_submission_g2h(struct intel_guc *guc)
>>> +{
>>> + if (atomic_dec_and_test(&guc->outstanding_submission_g2h)) {
>>> + smp_mb();
>>> + if (waitqueue_active(&guc->ct.wq))
>>> + wake_up_all(&guc->ct.wq);
>>
>> I keep pointing out this pattern is racy and at least needs comment why it
>> is safe.
>>
>
> There is a comment in wake queue code header saying why this is safe. I
> don't think we need to repeat this here.
Yeah, _describing how to make it safe_, after it starts with:
* NOTE: this function is lockless and requires care, incorrect usage _will_
* lead to sporadic and non-obvious failure.
Then it also says:
* Also note that this 'optimization' trades a spin_lock() for an smp_mb(),
* which (when the lock is uncontended) are of roughly equal cost.
I question the need to optimize this path since it means reader has to figure out if it is safe while a simple wake_up_all after atomic_dec_and_test would have done it.
Is the case of no waiters a predominant one? It at least deserves a comment explaining why the optimisation is important.
Regards,
Tvrtko
>
> Matt
>
>> Regards,
>>
>> Tvrtko
>>
>>> + }
>>> +}
>>> +
>>> int intel_guc_deregister_done_process_msg(struct intel_guc *guc,
>>> const u32 *msg,
>>> u32 len)
>>> @@ -1472,6 +1552,8 @@ int intel_guc_deregister_done_process_msg(struct intel_guc *guc,
>>> lrc_destroy(&ce->ref);
>>> }
>>> + decr_outstanding_submission_g2h(guc);
>>> +
>>> return 0;
>>> }
>>> @@ -1520,6 +1602,7 @@ int intel_guc_sched_done_process_msg(struct intel_guc *guc,
>>> spin_unlock_irqrestore(&ce->guc_state.lock, flags);
>>> }
>>> + decr_outstanding_submission_g2h(guc);
>>> intel_context_put(ce);
>>> return 0;
>>> diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.h b/drivers/gpu/drm/i915/gt/uc/intel_uc.h
>>> index 9c954c589edf..c4cef885e984 100644
>>> --- a/drivers/gpu/drm/i915/gt/uc/intel_uc.h
>>> +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.h
>>> @@ -81,6 +81,11 @@ uc_state_checkers(guc, guc_submission);
>>> #undef uc_state_checkers
>>> #undef __uc_state_checker
>>> +static inline int intel_uc_wait_for_idle(struct intel_uc *uc, long timeout)
>>> +{
>>> + return intel_guc_wait_for_idle(&uc->guc, timeout);
>>> +}
>>> +
>>> #define intel_uc_ops_function(_NAME, _OPS, _TYPE, _RET) \
>>> static inline _TYPE intel_uc_##_NAME(struct intel_uc *uc) \
>>> { \
>>> diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
>>> index 8dd374691102..bb29838d1cd7 100644
>>> --- a/drivers/gpu/drm/i915/i915_debugfs.c
>>> +++ b/drivers/gpu/drm/i915/i915_debugfs.c
>>> @@ -36,6 +36,7 @@
>>> #include "gt/intel_gt_clock_utils.h"
>>> #include "gt/intel_gt.h"
>>> #include "gt/intel_gt_pm.h"
>>> +#include "gt/intel_gt.h"
>>> #include "gt/intel_gt_requests.h"
>>> #include "gt/intel_reset.h"
>>> #include "gt/intel_rc6.h"
>>> diff --git a/drivers/gpu/drm/i915/i915_gem_evict.c b/drivers/gpu/drm/i915/i915_gem_evict.c
>>> index 4d2d59a9942b..2b73ddb11c66 100644
>>> --- a/drivers/gpu/drm/i915/i915_gem_evict.c
>>> +++ b/drivers/gpu/drm/i915/i915_gem_evict.c
>>> @@ -27,6 +27,7 @@
>>> */
>>> #include "gem/i915_gem_context.h"
>>> +#include "gt/intel_gt.h"
>>> #include "gt/intel_gt_requests.h"
>>> #include "i915_drv.h"
>>> diff --git a/drivers/gpu/drm/i915/selftests/igt_live_test.c b/drivers/gpu/drm/i915/selftests/igt_live_test.c
>>> index c130010a7033..1c721542e277 100644
>>> --- a/drivers/gpu/drm/i915/selftests/igt_live_test.c
>>> +++ b/drivers/gpu/drm/i915/selftests/igt_live_test.c
>>> @@ -5,7 +5,7 @@
>>> */
>>> #include "i915_drv.h"
>>> -#include "gt/intel_gt_requests.h"
>>> +#include "gt/intel_gt.h"
>>> #include "../i915_selftest.h"
>>> #include "igt_flush_test.h"
>>> diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
>>> index cf40004bc92a..6c06816e2b99 100644
>>> --- a/drivers/gpu/drm/i915/selftests/mock_gem_device.c
>>> +++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
>>> @@ -51,7 +51,8 @@ void mock_device_flush(struct drm_i915_private *i915)
>>> do {
>>> for_each_engine(engine, gt, id)
>>> mock_engine_flush(engine);
>>> - } while (intel_gt_retire_requests_timeout(gt, MAX_SCHEDULE_TIMEOUT));
>>> + } while (intel_gt_retire_requests_timeout(gt, MAX_SCHEDULE_TIMEOUT,
>>> + NULL));
>>> }
>>> static void mock_device_release(struct drm_device *dev)
>>>
next prev parent reply other threads:[~2021-05-26 9:21 UTC|newest]
Thread overview: 249+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-05-06 19:13 [RFC PATCH 00/97] Basic GuC submission support in the i915 Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 01/97] drm/i915/gt: Move engine setup out of set_default_submission Matthew Brost
2021-05-19 0:25 ` Matthew Brost
2021-05-25 8:44 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:13 ` [RFC PATCH 02/97] drm/i915/gt: Move submission_method into intel_gt Matthew Brost
2021-05-19 3:10 ` Matthew Brost
2021-05-25 8:44 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:13 ` [RFC PATCH 03/97] drm/i915/gt: Move CS interrupt handler to the backend Matthew Brost
2021-05-19 3:31 ` Matthew Brost
2021-05-25 8:45 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:13 ` [RFC PATCH 04/97] drm/i915/guc: skip disabling CTBs before sanitizing the GuC Matthew Brost
2021-05-20 16:47 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 05/97] drm/i915/guc: use probe_error log for CT enablement failure Matthew Brost
2021-05-24 10:30 ` Michal Wajdeczko
2021-05-06 19:13 ` [RFC PATCH 06/97] drm/i915/guc: enable only the user interrupt when using GuC submission Matthew Brost
2021-05-25 0:31 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 07/97] drm/i915/guc: Remove sample_forcewake h2g action Matthew Brost
2021-05-24 10:48 ` Michal Wajdeczko
2021-05-25 0:36 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 08/97] drm/i915/guc: Keep strict GuC ABI definitions Matthew Brost
2021-05-24 23:52 ` Michał Winiarski
2021-05-06 19:13 ` [RFC PATCH 09/97] drm/i915/guc: Stop using fence/status from CTB descriptor Matthew Brost
2021-05-25 2:38 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 10/97] drm/i915: Promote ptrdiff() to i915_utils.h Matthew Brost
2021-05-25 0:42 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 11/97] drm/i915/guc: Only rely on own CTB size Matthew Brost
2021-05-25 2:47 ` Matthew Brost
2021-05-25 12:48 ` [Intel-gfx] " Michal Wajdeczko
2021-05-06 19:13 ` [RFC PATCH 12/97] drm/i915/guc: Don't repeat CTB layout calculations Matthew Brost
2021-05-25 2:53 ` Matthew Brost
2021-05-25 13:07 ` Michal Wajdeczko
2021-05-25 16:56 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 13/97] drm/i915/guc: Replace CTB array with explicit members Matthew Brost
2021-05-25 3:15 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 14/97] drm/i915/guc: Update sizes of CTB buffers Matthew Brost
2021-05-25 2:56 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 15/97] drm/i915/guc: Relax CTB response timeout Matthew Brost
2021-05-25 18:08 ` Matthew Brost
2021-05-25 19:37 ` [Intel-gfx] " Michal Wajdeczko
2021-05-06 19:13 ` [RFC PATCH 16/97] drm/i915/guc: Start protecting access to CTB descriptors Matthew Brost
2021-05-25 3:21 ` Matthew Brost
2021-05-25 3:21 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 17/97] drm/i915/guc: Stop using mutex while sending CTB messages Matthew Brost
2021-05-25 16:14 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 18/97] drm/i915/guc: Don't receive all G2H messages in irq handler Matthew Brost
2021-05-25 18:15 ` Matthew Brost
2021-05-25 19:43 ` [Intel-gfx] " Michal Wajdeczko
2021-05-06 19:13 ` [RFC PATCH 19/97] drm/i915/guc: Always copy CT message to new allocation Matthew Brost
2021-05-25 18:25 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 20/97] drm/i915/guc: Introduce unified HXG messages Matthew Brost
2021-05-11 15:16 ` Daniel Vetter
2021-05-11 17:59 ` Matthew Brost
2021-05-11 22:11 ` Michal Wajdeczko
2021-05-12 8:40 ` Daniel Vetter
2021-05-06 19:13 ` [RFC PATCH 21/97] drm/i915/guc: Update MMIO based communication Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 22/97] drm/i915/guc: Update CTB response status Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 23/97] drm/i915/guc: Support per context scheduling policies Matthew Brost
2021-05-25 1:15 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 24/97] drm/i915/guc: Add flag for mark broken CTB Matthew Brost
2021-05-27 19:44 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 25/97] drm/i915/guc: New definition of the CTB descriptor Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 26/97] drm/i915/guc: New definition of the CTB registration action Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 27/97] drm/i915/guc: New CTB based communication Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 28/97] drm/i915/guc: Kill guc_clients.ct_pool Matthew Brost
2021-05-25 1:01 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 29/97] drm/i915/guc: Update firmware to v60.1.2 Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 30/97] drm/i915/uc: turn on GuC/HuC auto mode by default Matthew Brost
2021-05-24 11:00 ` Michal Wajdeczko
2021-05-06 19:13 ` [RFC PATCH 31/97] drm/i915/guc: Early initialization of GuC send registers Matthew Brost
2021-05-26 20:28 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 32/97] drm/i915: Introduce i915_sched_engine object Matthew Brost
2021-05-11 15:18 ` Daniel Vetter
2021-05-11 17:56 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 33/97] drm/i915: Engine relative MMIO Matthew Brost
2021-05-25 9:05 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:13 ` [RFC PATCH 34/97] drm/i915/guc: Use guc_class instead of engine_class in fw interface Matthew Brost
2021-05-26 20:41 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 35/97] drm/i915/guc: Improve error message for unsolicited CT response Matthew Brost
2021-05-24 11:59 ` Michal Wajdeczko
2021-05-25 17:32 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 36/97] drm/i915/guc: Add non blocking CTB send function Matthew Brost
2021-05-24 12:21 ` Michal Wajdeczko
2021-05-25 17:30 ` Matthew Brost
2021-05-25 9:21 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-25 17:21 ` Matthew Brost
2021-05-26 8:57 ` Tvrtko Ursulin
2021-05-26 18:10 ` Matthew Brost
2021-05-27 10:02 ` Tvrtko Ursulin
2021-05-27 14:35 ` Matthew Brost
2021-05-27 15:11 ` Tvrtko Ursulin
2021-06-07 17:31 ` Matthew Brost
2021-06-08 8:39 ` Tvrtko Ursulin
2021-06-08 8:46 ` Daniel Vetter
2021-06-09 23:10 ` Matthew Brost
2021-06-10 15:27 ` Daniel Vetter
2021-06-24 16:38 ` Matthew Brost
2021-06-24 17:25 ` Daniel Vetter
2021-06-09 13:58 ` Michal Wajdeczko
2021-06-09 23:05 ` Matthew Brost
2021-06-09 14:14 ` Michal Wajdeczko
2021-06-09 23:13 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 37/97] drm/i915/guc: Add stall timer to " Matthew Brost
2021-05-24 12:58 ` Michal Wajdeczko
2021-05-24 18:35 ` Matthew Brost
2021-05-25 14:15 ` Michal Wajdeczko
2021-05-25 16:54 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 38/97] drm/i915/guc: Optimize CTB writes and reads Matthew Brost
2021-05-24 13:31 ` Michal Wajdeczko
2021-05-25 17:39 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 39/97] drm/i915/guc: Increase size of CTB buffers Matthew Brost
2021-05-24 13:43 ` [Intel-gfx] " Michal Wajdeczko
2021-05-24 18:40 ` Matthew Brost
2021-05-25 9:24 ` Tvrtko Ursulin
2021-05-25 17:15 ` Matthew Brost
2021-05-26 9:30 ` Tvrtko Ursulin
2021-05-26 18:20 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 40/97] drm/i915/guc: Module load failure test for CT buffer creation Matthew Brost
2021-05-24 13:45 ` Michal Wajdeczko
2021-05-06 19:13 ` [RFC PATCH 41/97] drm/i915/guc: Add new GuC interface defines and structures Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 42/97] drm/i915/guc: Remove GuC stage descriptor, add lrc descriptor Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 43/97] drm/i915/guc: Add lrc descriptor context lookup array Matthew Brost
2021-05-11 15:26 ` Daniel Vetter
2021-05-11 17:01 ` Matthew Brost
2021-05-11 17:43 ` Daniel Vetter
2021-05-11 19:34 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 44/97] drm/i915/guc: Implement GuC submission tasklet Matthew Brost
2021-05-25 9:43 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-25 17:10 ` Matthew Brost
2021-05-06 19:13 ` [RFC PATCH 45/97] drm/i915/guc: Add bypass tasklet submission path to GuC Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 46/97] drm/i915/guc: Implement GuC context operations for new inteface Matthew Brost
2021-05-29 20:32 ` Michal Wajdeczko
2021-05-06 19:14 ` [RFC PATCH 47/97] drm/i915/guc: Insert fence on context when deregistering Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 48/97] drm/i915/guc: Defer context unpin until scheduling is disabled Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 49/97] drm/i915/guc: Disable engine barriers with GuC during unpin Matthew Brost
2021-05-11 15:37 ` Daniel Vetter
2021-05-11 16:31 ` Matthew Brost
2021-05-26 10:26 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:14 ` [RFC PATCH 50/97] drm/i915/guc: Extend deregistration fence to schedule disable Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 51/97] drm/i915: Disable preempt busywait when using GuC scheduling Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 52/97] drm/i915/guc: Ensure request ordering via completion fences Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 53/97] drm/i915/guc: Disable semaphores when using GuC scheduling Matthew Brost
2021-05-25 9:52 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-25 17:01 ` Matthew Brost
2021-05-26 9:25 ` Tvrtko Ursulin
2021-05-26 18:15 ` Matthew Brost
2021-05-27 8:41 ` Tvrtko Ursulin
2021-05-27 14:38 ` Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 54/97] drm/i915/guc: Ensure G2H response has space in buffer Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 55/97] drm/i915/guc: Update intel_gt_wait_for_idle to work with GuC Matthew Brost
2021-05-25 10:06 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-25 17:07 ` Matthew Brost
2021-05-26 9:21 ` Tvrtko Ursulin [this message]
2021-05-26 18:18 ` Matthew Brost
2021-05-27 9:02 ` Tvrtko Ursulin
2021-05-27 14:37 ` Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 56/97] drm/i915/guc: Update GuC debugfs to support new GuC Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 57/97] drm/i915/guc: Add several request trace points Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 58/97] drm/i915: Add intel_context tracing Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 59/97] drm/i915/guc: GuC virtual engines Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 60/97] drm/i915: Track 'serial' counts for " Matthew Brost
2021-05-25 10:16 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-25 17:52 ` Matthew Brost
2021-05-26 8:40 ` Tvrtko Ursulin
2021-05-26 18:45 ` John Harrison
2021-05-27 8:53 ` Tvrtko Ursulin
2021-05-27 17:01 ` John Harrison
2021-06-01 9:31 ` Tvrtko Ursulin
2021-06-02 1:20 ` John Harrison
2021-06-02 12:04 ` Tvrtko Ursulin
2021-06-02 12:09 ` Tvrtko Ursulin
2021-05-06 19:14 ` [RFC PATCH 61/97] drm/i915: Hold reference to intel_context over life of i915_request Matthew Brost
2021-06-02 12:18 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:14 ` [RFC PATCH 62/97] drm/i915/guc: Disable bonding extension with GuC submission Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 63/97] drm/i915/guc: Direct all breadcrumbs for a class to single breadcrumbs Matthew Brost
2021-06-02 13:31 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:14 ` [RFC PATCH 64/97] drm/i915/guc: Reset implementation for new GuC interface Matthew Brost
2021-06-02 14:33 ` [Intel-gfx] " Tvrtko Ursulin
2021-06-04 3:17 ` Matthew Brost
2021-06-04 8:16 ` Daniel Vetter
2021-06-04 18:02 ` Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 65/97] drm/i915: Reset GPU immediately if submission is disabled Matthew Brost
2021-06-02 14:36 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-06 19:14 ` [RFC PATCH 66/97] drm/i915/guc: Add disable interrupts to guc sanitize Matthew Brost
2021-05-11 8:16 ` [drm/i915/guc] 07336fb545: WARNING:at_drivers/gpu/drm/i915/gt/uc/intel_uc.c:#__uc_sanitize[i915] kernel test robot
2021-05-06 19:14 ` [RFC PATCH 67/97] drm/i915/guc: Suspend/resume implementation for new interface Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 68/97] drm/i915/guc: Handle context reset notification Matthew Brost
2021-05-11 16:25 ` [Intel-gfx] " Daniel Vetter
2021-05-06 19:14 ` [RFC PATCH 69/97] drm/i915/guc: Handle engine reset failure notification Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 70/97] drm/i915/guc: Enable the timer expired interrupt for GuC Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 71/97] drm/i915/guc: Provide mmio list to be saved/restored on engine reset Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 72/97] drm/i915/guc: Don't complain about reset races Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 73/97] drm/i915/guc: Enable GuC engine reset Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 74/97] drm/i915/guc: Capture error state on context reset Matthew Brost
2021-05-11 16:28 ` [Intel-gfx] " Daniel Vetter
2021-05-11 17:12 ` Matthew Brost
2021-05-11 17:45 ` Daniel Vetter
2021-05-06 19:14 ` [RFC PATCH 75/97] drm/i915/guc: Fix for error capture after full GPU reset with GuC Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 76/97] drm/i915/guc: Hook GuC scheduling policies up Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 77/97] drm/i915/guc: Connect reset modparam updates to GuC policy flags Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 78/97] drm/i915/guc: Include scheduling policies in the debugfs state dump Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 79/97] drm/i915/guc: Don't call ring_is_idle in GuC submission Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 80/97] drm/i915/guc: Implement banned contexts for " Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 81/97] drm/i915/guc: Allow flexible number of context ids Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 82/97] drm/i915/guc: Connect the number of guc_ids to debugfs Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 83/97] drm/i915/guc: Don't return -EAGAIN to user when guc_ids exhausted Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 84/97] drm/i915/guc: Don't allow requests not ready to consume all guc_ids Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 85/97] drm/i915/guc: Introduce guc_submit_engine object Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 86/97] drm/i915/guc: Add golden context to GuC ADS Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 87/97] drm/i915/guc: Implement GuC priority management Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 88/97] drm/i915/guc: Support request cancellation Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 89/97] drm/i915/guc: Check return of __xa_store when registering a context Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 90/97] drm/i915/guc: Non-static lrc descriptor registration buffer Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 91/97] drm/i915/guc: Take GT PM ref when deregistering context Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 92/97] drm/i915: Add GT PM delayed worker Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 93/97] drm/i915/guc: Take engine PM when a context is pinned with GuC submission Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 94/97] drm/i915/guc: Don't call switch_to_kernel_context " Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 95/97] drm/i915/guc: Selftest for GuC flow control Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 96/97] drm/i915/guc: Update GuC documentation Matthew Brost
2021-05-06 19:14 ` [RFC PATCH 97/97] drm/i915/guc: Unblock GuC submission on Gen11+ Matthew Brost
2021-05-09 17:12 ` [RFC PATCH 00/97] Basic GuC submission support in the i915 Martin Peres
2021-05-09 23:11 ` Jason Ekstrand
2021-05-10 13:55 ` Martin Peres
2021-05-10 16:25 ` Jason Ekstrand
2021-05-11 8:01 ` Martin Peres
2021-05-10 16:33 ` Daniel Vetter
2021-05-10 18:30 ` [Intel-gfx] " Francisco Jerez
2021-05-11 8:06 ` Martin Peres
2021-05-11 15:26 ` Bloomfield, Jon
2021-05-11 16:39 ` Matthew Brost
2021-05-12 6:26 ` Martin Peres
2021-05-14 16:31 ` Jason Ekstrand
2021-05-25 15:37 ` Alex Deucher
2021-05-11 2:58 ` Dixit, Ashutosh
2021-05-11 7:47 ` Martin Peres
2021-05-14 11:11 ` [Intel-gfx] " Tvrtko Ursulin
2021-05-14 16:36 ` Jason Ekstrand
2021-05-14 16:46 ` Matthew Brost
2021-05-14 16:41 ` Matthew Brost
2021-05-25 10:32 ` Tvrtko Ursulin
2021-05-25 16:45 ` Matthew Brost
2021-06-02 15:27 ` Tvrtko Ursulin
2021-06-02 18:57 ` Daniel Vetter
2021-06-03 3:41 ` Matthew Brost
2021-06-03 4:47 ` Daniel Vetter
2021-06-03 9:49 ` Tvrtko Ursulin
2021-06-03 10:52 ` Tvrtko Ursulin
2021-06-03 4:10 ` Matthew Brost
2021-06-03 8:51 ` Tvrtko Ursulin
2021-06-03 16:34 ` Matthew Brost
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5f84fcc9-5c8c-d44b-3739-5b970aef7eb4@linux.intel.com \
--to=tvrtko.ursulin@linux.intel.com \
--cc=daniel.vetter@intel.com \
--cc=dri-devel@lists.freedesktop.org \
--cc=intel-gfx@lists.freedesktop.org \
--cc=jason.ekstrand@intel.com \
--cc=matthew.brost@intel.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).