diff options
author | Jakub Kicinski <kuba@kernel.org> | 2024-03-22 02:14:13 +0300 |
---|---|---|
committer | Jakub Kicinski <kuba@kernel.org> | 2024-03-22 02:15:08 +0300 |
commit | 537c2e91d3549e5d6020bb0576cf9b54a845255f (patch) | |
tree | c09e8a1b7d733cde19b0c72678c28fb2bc97ff6b /drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | |
parent | 237bb5f7f7f55ec5f773469a974c61a49c298625 (diff) | |
parent | cba9ffdb9913dfe6be29f049ce920ce451ce7cc4 (diff) | |
download | linux-537c2e91d3549e5d6020bb0576cf9b54a845255f.tar.xz |
Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Cross-merge networking fixes after downstream PR.
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Diffstat (limited to 'drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c')
-rw-r--r-- | drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 126 |
1 files changed, 120 insertions, 6 deletions
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 a259f1118c5a..f3dcae4b9d45 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -236,6 +236,13 @@ set_context_destroyed(struct intel_context *ce) ce->guc_state.sched_state |= SCHED_STATE_DESTROYED; } +static inline void +clr_context_destroyed(struct intel_context *ce) +{ + lockdep_assert_held(&ce->guc_state.lock); + ce->guc_state.sched_state &= ~SCHED_STATE_DESTROYED; +} + static inline bool context_pending_disable(struct intel_context *ce) { return ce->guc_state.sched_state & SCHED_STATE_PENDING_DISABLE; @@ -613,6 +620,8 @@ static int guc_submission_send_busy_loop(struct intel_guc *guc, u32 g2h_len_dw, bool loop) { + int ret; + /* * We always loop when a send requires a reply (i.e. g2h_len_dw > 0), * so we don't handle the case where we don't get a reply because we @@ -623,7 +632,11 @@ static int guc_submission_send_busy_loop(struct intel_guc *guc, if (g2h_len_dw) atomic_inc(&guc->outstanding_submission_g2h); - return intel_guc_send_busy_loop(guc, action, len, g2h_len_dw, loop); + ret = intel_guc_send_busy_loop(guc, action, len, g2h_len_dw, loop); + if (ret) + atomic_dec(&guc->outstanding_submission_g2h); + + return ret; } int intel_guc_wait_for_pending_msg(struct intel_guc *guc, @@ -1362,7 +1375,45 @@ static void guc_enable_busyness_worker(struct intel_guc *guc) static void guc_cancel_busyness_worker(struct intel_guc *guc) { - cancel_delayed_work_sync(&guc->timestamp.work); + /* + * There are many different call stacks that can get here. Some of them + * hold the reset mutex. The busyness worker also attempts to acquire the + * reset mutex. Synchronously flushing a worker thread requires acquiring + * the worker mutex. Lockdep sees this as a conflict. It thinks that the + * flush can deadlock because it holds the worker mutex while waiting for + * the reset mutex, but another thread is holding the reset mutex and might + * attempt to use other worker functions. + * + * In practice, this scenario does not exist because the busyness worker + * does not block waiting for the reset mutex. It does a try-lock on it and + * immediately exits if the lock is already held. Unfortunately, the mutex + * in question (I915_RESET_BACKOFF) is an i915 implementation which has lockdep + * annotation but not to the extent of explaining the 'might lock' is also a + * 'does not need to lock'. So one option would be to add more complex lockdep + * annotations to ignore the issue (if at all possible). A simpler option is to + * just not flush synchronously when a rest in progress. Given that the worker + * will just early exit and re-schedule itself anyway, there is no advantage + * to running it immediately. + * + * If a reset is not in progress, then the synchronous flush may be required. + * As noted many call stacks lead here, some during suspend and driver unload + * which do require a synchronous flush to make sure the worker is stopped + * before memory is freed. + * + * Trying to pass a 'need_sync' or 'in_reset' flag all the way down through + * every possible call stack is unfeasible. It would be too intrusive to many + * areas that really don't care about the GuC backend. However, there is the + * 'reset_in_progress' flag available, so just use that. + * + * And note that in the case of a reset occurring during driver unload + * (wedge_on_fini), skipping the cancel in _prepare (when the reset flag is set + * is fine because there is another cancel in _finish (when the reset flag is + * not). + */ + if (guc_to_gt(guc)->uc.reset_in_progress) + cancel_delayed_work(&guc->timestamp.work); + else + cancel_delayed_work_sync(&guc->timestamp.work); } static void __reset_guc_busyness_stats(struct intel_guc *guc) @@ -1613,6 +1664,11 @@ static void guc_flush_submissions(struct intel_guc *guc) spin_unlock_irqrestore(&sched_engine->lock, flags); } +void intel_guc_submission_flush_work(struct intel_guc *guc) +{ + flush_work(&guc->submission_state.destroyed_worker); +} + static void guc_flush_destroyed_contexts(struct intel_guc *guc); void intel_guc_submission_reset_prepare(struct intel_guc *guc) @@ -1948,8 +2004,16 @@ void intel_guc_submission_cancel_requests(struct intel_guc *guc) void intel_guc_submission_reset_finish(struct intel_guc *guc) { + /* + * Ensure the busyness worker gets cancelled even on a fatal wedge. + * Note that reset_prepare is not allowed to because it confuses lockdep. + */ + if (guc_submission_initialized(guc)) + guc_cancel_busyness_worker(guc); + /* Reset called during driver load or during wedge? */ if (unlikely(!guc_submission_initialized(guc) || + !intel_guc_is_fw_running(guc) || intel_gt_is_wedged(guc_to_gt(guc)))) { return; } @@ -3283,12 +3347,13 @@ static void guc_context_close(struct intel_context *ce) spin_unlock_irqrestore(&ce->guc_state.lock, flags); } -static inline void guc_lrc_desc_unpin(struct intel_context *ce) +static inline int guc_lrc_desc_unpin(struct intel_context *ce) { struct intel_guc *guc = ce_to_guc(ce); struct intel_gt *gt = guc_to_gt(guc); unsigned long flags; bool disabled; + int ret; GEM_BUG_ON(!intel_gt_pm_is_awake(gt)); GEM_BUG_ON(!ctx_id_mapped(guc, ce->guc_id.id)); @@ -3299,18 +3364,41 @@ static inline void guc_lrc_desc_unpin(struct intel_context *ce) spin_lock_irqsave(&ce->guc_state.lock, flags); disabled = submission_disabled(guc); if (likely(!disabled)) { + /* + * Take a gt-pm ref and change context state to be destroyed. + * NOTE: a G2H IRQ that comes after will put this gt-pm ref back + */ __intel_gt_pm_get(gt); set_context_destroyed(ce); clr_context_registered(ce); } spin_unlock_irqrestore(&ce->guc_state.lock, flags); + if (unlikely(disabled)) { release_guc_id(guc, ce); __guc_context_destroy(ce); - return; + return 0; } - deregister_context(ce, ce->guc_id.id); + /* + * GuC is active, lets destroy this context, but at this point we can still be racing + * with suspend, so we undo everything if the H2G fails in deregister_context so + * that GuC reset will find this context during clean up. + */ + ret = deregister_context(ce, ce->guc_id.id); + if (ret) { + spin_lock(&ce->guc_state.lock); + set_context_registered(ce); + clr_context_destroyed(ce); + spin_unlock(&ce->guc_state.lock); + /* + * As gt-pm is awake at function entry, intel_wakeref_put_async merely decrements + * the wakeref immediately but per function spec usage call this after unlock. + */ + intel_wakeref_put_async(>->wakeref); + } + + return ret; } static void __guc_context_destroy(struct intel_context *ce) @@ -3378,7 +3466,22 @@ static void deregister_destroyed_contexts(struct intel_guc *guc) if (!ce) break; - guc_lrc_desc_unpin(ce); + if (guc_lrc_desc_unpin(ce)) { + /* + * This means GuC's CT link severed mid-way which could happen + * in suspend-resume corner cases. In this case, put the + * context back into the destroyed_contexts list which will + * get picked up on the next context deregistration event or + * purged in a GuC sanitization event (reset/unload/wedged/...). + */ + spin_lock_irqsave(&guc->submission_state.lock, flags); + list_add_tail(&ce->destroyed_link, + &guc->submission_state.destroyed_contexts); + spin_unlock_irqrestore(&guc->submission_state.lock, flags); + /* Bail now since the list might never be emptied if h2gs fail */ + break; + } + } } @@ -3389,6 +3492,17 @@ static void destroyed_worker_func(struct work_struct *w) struct intel_gt *gt = guc_to_gt(guc); intel_wakeref_t wakeref; + /* + * In rare cases we can get here via async context-free fence-signals that + * come very late in suspend flow or very early in resume flows. In these + * cases, GuC won't be ready but just skipping it here is fine as these + * pending-destroy-contexts get destroyed totally at GuC reset time at the + * end of suspend.. OR.. this worker can be picked up later on the next + * context destruction trigger after resume-completes + */ + if (!intel_guc_is_ready(guc)) + return; + with_intel_gt_pm(gt, wakeref) deregister_destroyed_contexts(guc); } |