From bb699a793110fc29664e80c4ebb158a922151d52 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Fri, 21 Feb 2020 10:09:53 +0000 Subject: [PATCH 1/9] drm/i915/gem: Break up long lists of object reclaim Call cond_resched() between each freed object in case we have a really, really long list, and we don't want to block normal processes. Signed-off-by: Chris Wilson Reviewed-by: Matthew Auld Link: https://patchwork.freedesktop.org/patch/msgid/20200221100953.2587176-1-chris@chris-wilson.co.uk (cherry picked from commit deeee411a97559096523f97655ff16da34cf0573) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/gem/i915_gem_object.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/gpu/drm/i915/gem/i915_gem_object.c b/drivers/gpu/drm/i915/gem/i915_gem_object.c index 35985218bd85..5da9f9e534b9 100644 --- a/drivers/gpu/drm/i915/gem/i915_gem_object.c +++ b/drivers/gpu/drm/i915/gem/i915_gem_object.c @@ -225,6 +225,7 @@ static void __i915_gem_free_objects(struct drm_i915_private *i915, /* But keep the pointer alive for RCU-protected lookups */ call_rcu(&obj->rcu, __i915_gem_free_object_rcu); + cond_resched(); } intel_runtime_pm_put(&i915->runtime_pm, wakeref); } From 33e059a2e4df454359f642f2235af39de9d3e914 Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Jos=C3=A9=20Roberto=20de=20Souza?= Date: Thu, 27 Feb 2020 12:55:40 -0800 Subject: [PATCH 2/9] drm/i915/psr: Force PSR probe only after full initialization MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Commit 60c6a14b489b ("drm/i915/display: Force the state compute phase once to enable PSR") was forcing the state compute too earlier causing errors because not everything was initialized, so here moving to the end of i915_driver_modeset_probe() when the display is all initialized. Also fixing the place where it disarm the force probe as during the atomic check phase errors could happen like the ones due locking and it would cause PSR to never be enabled if that happens. Leaving the disarm to the atomic commit phase, intel_psr_enable() or intel_psr_update() will be called even if the current state do not allow PSR to be enabled. v2: Check if intel_dp is null in intel_psr_force_mode_changed_set() v3: Check intel_dp before get dev_priv v4: - renamed intel_psr_force_mode_changed_set() to intel_psr_set_force_mode_changed() - removed the set parameter from intel_psr_set_force_mode_changed() - not calling intel_psr_set_force_mode_changed() from intel_psr_enable/update(), directly setting it after the same checks that intel_psr_set_force_mode_changed() does - moved intel_psr_set_force_mode_changed() arm call to i915_driver_modeset_probe() as it is a better for a PSR call, all the functions calls happening between the old and the new function call will cause issue [backported to v5.6-rc3] Fixes: 60c6a14b489b ("drm/i915/display: Force the state compute phase once to enable PSR") Closes: https://gitlab.freedesktop.org/drm/intel/issues/1151 Tested-by: Ross Zwisler Reported-by: Ross Zwisler Cc: Gwan-gyeong Mun Cc: Jani Nikula Cc: Anshuman Gupta Reviewed-by: Gwan-gyeong Mun Signed-off-by: José Roberto de Souza Link: https://patchwork.freedesktop.org/patch/msgid/20200221212635.11614-1-jose.souza@intel.com Link: https://patchwork.freedesktop.org/patch/msgid/20200227205540.126135-1-jose.souza@intel.com (cherry picked from commit df1a5bfc16f3275a74f77d73375e69bc62c45c4b) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/display/intel_psr.c | 25 ++++++++++++++++++++---- drivers/gpu/drm/i915/display/intel_psr.h | 1 + drivers/gpu/drm/i915/i915_drv.c | 3 +++ drivers/gpu/drm/i915/i915_drv.h | 2 +- 4 files changed, 26 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c index 89c9cf5f38d2..83025052c965 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.c +++ b/drivers/gpu/drm/i915/display/intel_psr.c @@ -852,10 +852,12 @@ void intel_psr_enable(struct intel_dp *intel_dp, { struct drm_i915_private *dev_priv = dp_to_i915(intel_dp); - if (!crtc_state->has_psr) + if (!CAN_PSR(dev_priv) || dev_priv->psr.dp != intel_dp) return; - if (WARN_ON(!CAN_PSR(dev_priv))) + dev_priv->psr.force_mode_changed = false; + + if (!crtc_state->has_psr) return; WARN_ON(dev_priv->drrs.dp); @@ -1009,6 +1011,8 @@ void intel_psr_update(struct intel_dp *intel_dp, if (!CAN_PSR(dev_priv) || READ_ONCE(psr->dp) != intel_dp) return; + dev_priv->psr.force_mode_changed = false; + mutex_lock(&dev_priv->psr.lock); enable = crtc_state->has_psr && psr_global_enabled(psr->debug); @@ -1534,7 +1538,7 @@ void intel_psr_atomic_check(struct drm_connector *connector, struct drm_crtc_state *crtc_state; if (!CAN_PSR(dev_priv) || !new_state->crtc || - dev_priv->psr.initially_probed) + !dev_priv->psr.force_mode_changed) return; intel_connector = to_intel_connector(connector); @@ -1545,5 +1549,18 @@ void intel_psr_atomic_check(struct drm_connector *connector, crtc_state = drm_atomic_get_new_crtc_state(new_state->state, new_state->crtc); crtc_state->mode_changed = true; - dev_priv->psr.initially_probed = true; +} + +void intel_psr_set_force_mode_changed(struct intel_dp *intel_dp) +{ + struct drm_i915_private *dev_priv; + + if (!intel_dp) + return; + + dev_priv = dp_to_i915(intel_dp); + if (!CAN_PSR(dev_priv) || intel_dp != dev_priv->psr.dp) + return; + + dev_priv->psr.force_mode_changed = true; } diff --git a/drivers/gpu/drm/i915/display/intel_psr.h b/drivers/gpu/drm/i915/display/intel_psr.h index c58a1d438808..274fc6bb6221 100644 --- a/drivers/gpu/drm/i915/display/intel_psr.h +++ b/drivers/gpu/drm/i915/display/intel_psr.h @@ -40,5 +40,6 @@ bool intel_psr_enabled(struct intel_dp *intel_dp); void intel_psr_atomic_check(struct drm_connector *connector, struct drm_connector_state *old_state, struct drm_connector_state *new_state); +void intel_psr_set_force_mode_changed(struct intel_dp *intel_dp); #endif /* __INTEL_PSR_H__ */ diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c index f7385abdd74b..8410330ce4f0 100644 --- a/drivers/gpu/drm/i915/i915_drv.c +++ b/drivers/gpu/drm/i915/i915_drv.c @@ -56,6 +56,7 @@ #include "display/intel_hotplug.h" #include "display/intel_overlay.h" #include "display/intel_pipe_crc.h" +#include "display/intel_psr.h" #include "display/intel_sprite.h" #include "display/intel_vga.h" @@ -330,6 +331,8 @@ static int i915_driver_modeset_probe(struct drm_i915_private *i915) intel_init_ipc(i915); + intel_psr_set_force_mode_changed(i915->psr.dp); + return 0; cleanup_gem: diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h index 077af22b8340..810e3ccd56ec 100644 --- a/drivers/gpu/drm/i915/i915_drv.h +++ b/drivers/gpu/drm/i915/i915_drv.h @@ -505,7 +505,7 @@ struct i915_psr { bool dc3co_enabled; u32 dc3co_exit_delay; struct delayed_work idle_work; - bool initially_probed; + bool force_mode_changed; }; #define QUIRK_LVDS_SSC_DISABLE (1<<1) From c725161924f9a5872a3e53b73345a6026a5c170e Mon Sep 17 00:00:00 2001 From: Matt Roper Date: Thu, 27 Feb 2020 16:43:19 -0800 Subject: [PATCH 3/9] drm/i915: Program MBUS with rmw during initialization It wasn't terribly clear from the bspec's wording, but after discussion with the hardware folks, it turns out that we need to preserve the pre-existing contents of the MBUS ABOX control register when initializing a few specific bits. Bspec: 49213 Bspec: 50096 Fixes: 4cb4585e5a7f ("drm/i915/icl: initialize MBus during display init") Cc: Stanislav Lisovskiy Signed-off-by: Matt Roper Link: https://patchwork.freedesktop.org/patch/msgid/20200204011032.582737-1-matthew.d.roper@intel.com Reviewed-by: Matt Atwood (cherry picked from commit 837b63e6087838d0f1e612d448405419199d8033) Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20200228004320.127142-1-matthew.d.roper@intel.com --- .../gpu/drm/i915/display/intel_display_power.c | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c index 21561acfa3ac..ae532e45e36b 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.c +++ b/drivers/gpu/drm/i915/display/intel_display_power.c @@ -4466,13 +4466,19 @@ static void icl_dbuf_disable(struct drm_i915_private *dev_priv) static void icl_mbus_init(struct drm_i915_private *dev_priv) { - u32 val; + u32 mask, val; - val = MBUS_ABOX_BT_CREDIT_POOL1(16) | - MBUS_ABOX_BT_CREDIT_POOL2(16) | - MBUS_ABOX_B_CREDIT(1) | - MBUS_ABOX_BW_CREDIT(1); + mask = MBUS_ABOX_BT_CREDIT_POOL1_MASK | + MBUS_ABOX_BT_CREDIT_POOL2_MASK | + MBUS_ABOX_B_CREDIT_MASK | + MBUS_ABOX_BW_CREDIT_MASK; + val = I915_READ(MBUS_ABOX_CTL); + val &= ~mask; + val |= MBUS_ABOX_BT_CREDIT_POOL1(16) | + MBUS_ABOX_BT_CREDIT_POOL2(16) | + MBUS_ABOX_B_CREDIT(1) | + MBUS_ABOX_BW_CREDIT(1); I915_WRITE(MBUS_ABOX_CTL, val); } From 4c116e1ae43955a0a38555dfd4d136a222a8996b Mon Sep 17 00:00:00 2001 From: Matt Roper Date: Thu, 27 Feb 2020 16:43:20 -0800 Subject: [PATCH 4/9] drm/i915/tgl: Add Wa_22010178259:tgl MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit We need to explicitly set the TLB Request Timer initial value in the BW_BUDDY registers to 0x8 rather than relying on the hardware default. v2: Apply missing REG_FIELD_PREP to ensure 0x8 is placed in the correct bits during the rmw. (Jose) Bspec: 52890 Bspec: 50044 Fixes: 3fa01d642fa7 ("drm/i915/tgl: Program BW_BUDDY registers during display init") Cc: Stanislav Lisovskiy Cc: José Roberto de Souza Signed-off-by: Matt Roper Link: https://patchwork.freedesktop.org/patch/msgid/20200219215655.2923650-1-matthew.d.roper@intel.com Reviewed-by: José Roberto de Souza (cherry picked from commit 87e04f75928bb5d357ef7df4eedc1a7e2761a833) Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20200228004320.127142-2-matthew.d.roper@intel.com --- drivers/gpu/drm/i915/display/intel_display_power.c | 13 +++++++++++++ drivers/gpu/drm/i915/i915_reg.h | 1 + 2 files changed, 14 insertions(+) diff --git a/drivers/gpu/drm/i915/display/intel_display_power.c b/drivers/gpu/drm/i915/display/intel_display_power.c index ae532e45e36b..46c40db992dd 100644 --- a/drivers/gpu/drm/i915/display/intel_display_power.c +++ b/drivers/gpu/drm/i915/display/intel_display_power.c @@ -4974,8 +4974,21 @@ static void tgl_bw_buddy_init(struct drm_i915_private *dev_priv) I915_WRITE(BW_BUDDY1_CTL, BW_BUDDY_DISABLE); I915_WRITE(BW_BUDDY2_CTL, BW_BUDDY_DISABLE); } else { + u32 val; + I915_WRITE(BW_BUDDY1_PAGE_MASK, table[i].page_mask); I915_WRITE(BW_BUDDY2_PAGE_MASK, table[i].page_mask); + + /* Wa_22010178259:tgl */ + val = I915_READ(BW_BUDDY1_CTL); + val &= ~BW_BUDDY_TLB_REQ_TIMER_MASK; + val |= REG_FIELD_PREP(BW_BUDDY_TLB_REQ_TIMER_MASK, 0x8); + I915_WRITE(BW_BUDDY1_CTL, val); + + val = I915_READ(BW_BUDDY2_CTL); + val &= ~BW_BUDDY_TLB_REQ_TIMER_MASK; + val |= REG_FIELD_PREP(BW_BUDDY_TLB_REQ_TIMER_MASK, 0x8); + I915_WRITE(BW_BUDDY2_CTL, val); } } diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 6cc55c103f67..3575fd30756b 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -7757,6 +7757,7 @@ enum { #define BW_BUDDY1_CTL _MMIO(0x45140) #define BW_BUDDY2_CTL _MMIO(0x45150) #define BW_BUDDY_DISABLE REG_BIT(31) +#define BW_BUDDY_TLB_REQ_TIMER_MASK REG_GENMASK(21, 16) #define BW_BUDDY1_PAGE_MASK _MMIO(0x45144) #define BW_BUDDY2_PAGE_MASK _MMIO(0x45154) From eddf309a8ed42eb3312b17a6934686b018189cd3 Mon Sep 17 00:00:00 2001 From: Lucas De Marchi Date: Mon, 24 Feb 2020 11:12:58 -0800 Subject: [PATCH 5/9] drm/i915/tgl: Add Wa_1608008084 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Wa_1608008084 is an additional WA that applies to writes on FF_MODE2 register. We can't read it back either from CPU or GPU. Since the other bits should be 0, recommendation to handle Wa_1604555607 is to actually just write the timer value. Do a write only and don't try to read it, neither before or after the WA is applied. Fixes: ff690b2111ba ("drm/i915/tgl: Implement Wa_1604555607") Signed-off-by: Lucas De Marchi Reviewed-by: José Roberto de Souza Link: https://patchwork.freedesktop.org/patch/msgid/20200224191258.15668-1-lucas.demarchi@intel.com (cherry picked from commit e94bda14325ccf1a519ffb516738d1201457f97f) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/gt/intel_workarounds.c | 19 +++++++------------ 1 file changed, 7 insertions(+), 12 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 4e292d4bf7b9..173a7f2d109f 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -575,24 +575,19 @@ static void icl_ctx_workarounds_init(struct intel_engine_cs *engine, static void tgl_ctx_workarounds_init(struct intel_engine_cs *engine, struct i915_wa_list *wal) { - u32 val; - /* Wa_1409142259:tgl */ WA_SET_BIT_MASKED(GEN11_COMMON_SLICE_CHICKEN3, GEN12_DISABLE_CPS_AWARE_COLOR_PIPE); - /* Wa_1604555607:tgl */ - val = intel_uncore_read(engine->uncore, FF_MODE2); - val &= ~FF_MODE2_TDS_TIMER_MASK; - val |= FF_MODE2_TDS_TIMER_128; /* - * FIXME: FF_MODE2 register is not readable till TGL B0. We can - * enable verification of WA from the later steppings, which enables - * the read of FF_MODE2. + * Wa_1604555607:gen12 and Wa_1608008084:gen12 + * FF_MODE2 register will return the wrong value when read. The default + * value for this register is zero for all fields and there are no bit + * masks. So instead of doing a RMW we should just write the TDS timer + * value for Wa_1604555607. */ - wa_add(wal, FF_MODE2, FF_MODE2_TDS_TIMER_MASK, val, - IS_TGL_REVID(engine->i915, TGL_REVID_A0, TGL_REVID_A0) ? 0 : - FF_MODE2_TDS_TIMER_MASK); + wa_add(wal, FF_MODE2, FF_MODE2_TDS_TIMER_MASK, + FF_MODE2_TDS_TIMER_128, 0); } static void From 0b1570b7ffe68dfefa07cb092a0723f898bb8184 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Thu, 27 Feb 2020 08:57:14 +0000 Subject: [PATCH 6/9] drm/i915: Protect i915_request_await_start from early waits We need to be extremely careful inside i915_request_await_start() as it needs to walk the list of requests in the foreign timeline with very little protection. As we hold our own timeline mutex, we can not nest inside the signaler's timeline mutex, so all that remains is our RCU protection. However, to be safe we need to tell the compiler that we may be traversing the list only under RCU protection, and furthermore we need to start declaring requests as elements of the timeline from their construction. Fixes: 9ddc8ec027a3 ("drm/i915: Eliminate the trylock for awaiting an earlier request") Fixes: 6a79d848403d ("drm/i915: Lock signaler timeline while navigating") Signed-off-by: Chris Wilson Reviewed-by: Tvrtko Ursulin Link: https://patchwork.freedesktop.org/patch/msgid/20200227085723.1961649-11-chris@chris-wilson.co.uk (cherry picked from commit d22d2d073ef859b346bc32cb25299262e3973769) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_request.c | 41 ++++++++++++++++++++--------- 1 file changed, 28 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_request.c b/drivers/gpu/drm/i915/i915_request.c index f56b046a32de..dcaa85a91090 100644 --- a/drivers/gpu/drm/i915/i915_request.c +++ b/drivers/gpu/drm/i915/i915_request.c @@ -275,7 +275,7 @@ bool i915_request_retire(struct i915_request *rq) spin_unlock_irq(&rq->lock); remove_from_client(rq); - list_del(&rq->link); + list_del_rcu(&rq->link); intel_context_exit(rq->context); intel_context_unpin(rq->context); @@ -721,6 +721,8 @@ __i915_request_create(struct intel_context *ce, gfp_t gfp) rq->infix = rq->ring->emit; /* end of header; start of user payload */ intel_context_mark_active(ce); + list_add_tail_rcu(&rq->link, &tl->requests); + return rq; err_unwind: @@ -777,13 +779,23 @@ i915_request_await_start(struct i915_request *rq, struct i915_request *signal) GEM_BUG_ON(i915_request_timeline(rq) == rcu_access_pointer(signal->timeline)); + if (i915_request_started(signal)) + return 0; + fence = NULL; rcu_read_lock(); spin_lock_irq(&signal->lock); - if (!i915_request_started(signal) && - !list_is_first(&signal->link, - &rcu_dereference(signal->timeline)->requests)) { - struct i915_request *prev = list_prev_entry(signal, link); + do { + struct list_head *pos = READ_ONCE(signal->link.prev); + struct i915_request *prev; + + /* Confirm signal has not been retired, the link is valid */ + if (unlikely(i915_request_started(signal))) + break; + + /* Is signal the earliest request on its timeline? */ + if (pos == &rcu_dereference(signal->timeline)->requests) + break; /* * Peek at the request before us in the timeline. That @@ -791,13 +803,18 @@ i915_request_await_start(struct i915_request *rq, struct i915_request *signal) * after acquiring a reference to it, confirm that it is * still part of the signaler's timeline. */ - if (i915_request_get_rcu(prev)) { - if (list_next_entry(prev, link) == signal) - fence = &prev->fence; - else - i915_request_put(prev); + prev = list_entry(pos, typeof(*prev), link); + if (!i915_request_get_rcu(prev)) + break; + + /* After the strong barrier, confirm prev is still attached */ + if (unlikely(READ_ONCE(prev->link.next) != &signal->link)) { + i915_request_put(prev); + break; } - } + + fence = &prev->fence; + } while (0); spin_unlock_irq(&signal->lock); rcu_read_unlock(); if (!fence) @@ -1242,8 +1259,6 @@ __i915_request_add_to_timeline(struct i915_request *rq) 0); } - list_add_tail(&rq->link, &timeline->requests); - /* * Make sure that no request gazumped us - if it was allocated after * our i915_request_alloc() and called __i915_request_add() before From f4aaa44e8b20f7e0d4ea68d3bca4968b6ae5aaff Mon Sep 17 00:00:00 2001 From: Dan Carpenter Date: Fri, 28 Feb 2020 17:14:13 +0300 Subject: [PATCH 7/9] drm/i915/selftests: Fix return in assert_mmap_offset() The assert_mmap_offset() returns type bool so if we return an error pointer that is "return true;" or success. If we have an error, then we should return false. Fixes: 3d81d589d6e3 ("drm/i915: Test exhaustion of the mmap space") Signed-off-by: Dan Carpenter Reviewed-by: Chris Wilson Signed-off-by: Chris Wilson Link: https://patchwork.freedesktop.org/patch/msgid/20200228141413.qfjf4abr323drlo4@kili.mountain (cherry picked from commit efbf928824820f2738f41271934f6ec2c6ebd587) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c index ef7c74cff28a..43912e9b683d 100644 --- a/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c +++ b/drivers/gpu/drm/i915/gem/selftests/i915_gem_mman.c @@ -570,7 +570,7 @@ static bool assert_mmap_offset(struct drm_i915_private *i915, obj = i915_gem_object_create_internal(i915, size); if (IS_ERR(obj)) - return PTR_ERR(obj); + return false; mmo = mmap_offset_attach(obj, I915_MMAP_OFFSET_GTT, NULL); i915_gem_object_put(obj); From 08f56f8f3799b2ed1c5ac7eed6d86a4926289655 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Mon, 2 Mar 2020 08:57:57 +0000 Subject: [PATCH 8/9] drm/i915/perf: Reintroduce wait on OA configuration completion We still need to wait for the initial OA configuration to happen before we enable OA report writes to the OA buffer. Reported-by: Lionel Landwerlin Fixes: 15d0ace1f876 ("drm/i915/perf: execute OA configuration from command stream") Closes: https://gitlab.freedesktop.org/drm/intel/issues/1356 Testcase: igt/perf/stream-open-close Signed-off-by: Chris Wilson Cc: Lionel Landwerlin Reviewed-by: Lionel Landwerlin Link: https://patchwork.freedesktop.org/patch/msgid/20200302085812.4172450-7-chris@chris-wilson.co.uk (cherry picked from commit 4b4e973d5eb89244b67d3223b60f752d0479f253) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/i915_perf.c | 58 ++++++++++++++++++-------- drivers/gpu/drm/i915/i915_perf_types.h | 3 +- 2 files changed, 43 insertions(+), 18 deletions(-) diff --git a/drivers/gpu/drm/i915/i915_perf.c b/drivers/gpu/drm/i915/i915_perf.c index 0f556d80ba36..3b6b913bd27a 100644 --- a/drivers/gpu/drm/i915/i915_perf.c +++ b/drivers/gpu/drm/i915/i915_perf.c @@ -1954,9 +1954,10 @@ get_oa_vma(struct i915_perf_stream *stream, struct i915_oa_config *oa_config) return i915_vma_get(oa_bo->vma); } -static int emit_oa_config(struct i915_perf_stream *stream, - struct i915_oa_config *oa_config, - struct intel_context *ce) +static struct i915_request * +emit_oa_config(struct i915_perf_stream *stream, + struct i915_oa_config *oa_config, + struct intel_context *ce) { struct i915_request *rq; struct i915_vma *vma; @@ -1964,7 +1965,7 @@ static int emit_oa_config(struct i915_perf_stream *stream, vma = get_oa_vma(stream, oa_config); if (IS_ERR(vma)) - return PTR_ERR(vma); + return ERR_CAST(vma); err = i915_vma_pin(vma, 0, 0, PIN_GLOBAL | PIN_HIGH); if (err) @@ -1989,13 +1990,17 @@ static int emit_oa_config(struct i915_perf_stream *stream, err = rq->engine->emit_bb_start(rq, vma->node.start, 0, I915_DISPATCH_SECURE); + if (err) + goto err_add_request; + + i915_request_get(rq); err_add_request: i915_request_add(rq); err_vma_unpin: i915_vma_unpin(vma); err_vma_put: i915_vma_put(vma); - return err; + return err ? ERR_PTR(err) : rq; } static struct intel_context *oa_context(struct i915_perf_stream *stream) @@ -2003,7 +2008,8 @@ static struct intel_context *oa_context(struct i915_perf_stream *stream) return stream->pinned_ctx ?: stream->engine->kernel_context; } -static int hsw_enable_metric_set(struct i915_perf_stream *stream) +static struct i915_request * +hsw_enable_metric_set(struct i915_perf_stream *stream) { struct intel_uncore *uncore = stream->uncore; @@ -2406,7 +2412,8 @@ static int lrc_configure_all_contexts(struct i915_perf_stream *stream, return oa_configure_all_contexts(stream, regs, ARRAY_SIZE(regs)); } -static int gen8_enable_metric_set(struct i915_perf_stream *stream) +static struct i915_request * +gen8_enable_metric_set(struct i915_perf_stream *stream) { struct intel_uncore *uncore = stream->uncore; struct i915_oa_config *oa_config = stream->oa_config; @@ -2448,7 +2455,7 @@ static int gen8_enable_metric_set(struct i915_perf_stream *stream) */ ret = lrc_configure_all_contexts(stream, oa_config); if (ret) - return ret; + return ERR_PTR(ret); return emit_oa_config(stream, oa_config, oa_context(stream)); } @@ -2460,7 +2467,8 @@ static u32 oag_report_ctx_switches(const struct i915_perf_stream *stream) 0 : GEN12_OAG_OA_DEBUG_DISABLE_CTX_SWITCH_REPORTS); } -static int gen12_enable_metric_set(struct i915_perf_stream *stream) +static struct i915_request * +gen12_enable_metric_set(struct i915_perf_stream *stream) { struct intel_uncore *uncore = stream->uncore; struct i915_oa_config *oa_config = stream->oa_config; @@ -2491,7 +2499,7 @@ static int gen12_enable_metric_set(struct i915_perf_stream *stream) */ ret = gen12_configure_all_contexts(stream, oa_config); if (ret) - return ret; + return ERR_PTR(ret); /* * For Gen12, performance counters are context @@ -2501,7 +2509,7 @@ static int gen12_enable_metric_set(struct i915_perf_stream *stream) if (stream->ctx) { ret = gen12_configure_oar_context(stream, true); if (ret) - return ret; + return ERR_PTR(ret); } return emit_oa_config(stream, oa_config, oa_context(stream)); @@ -2696,6 +2704,20 @@ static const struct i915_perf_stream_ops i915_oa_stream_ops = { .read = i915_oa_read, }; +static int i915_perf_stream_enable_sync(struct i915_perf_stream *stream) +{ + struct i915_request *rq; + + rq = stream->perf->ops.enable_metric_set(stream); + if (IS_ERR(rq)) + return PTR_ERR(rq); + + i915_request_wait(rq, 0, MAX_SCHEDULE_TIMEOUT); + i915_request_put(rq); + + return 0; +} + /** * i915_oa_stream_init - validate combined props for OA stream and init * @stream: An i915 perf stream @@ -2829,7 +2851,7 @@ static int i915_oa_stream_init(struct i915_perf_stream *stream, stream->ops = &i915_oa_stream_ops; perf->exclusive_stream = stream; - ret = perf->ops.enable_metric_set(stream); + ret = i915_perf_stream_enable_sync(stream); if (ret) { DRM_DEBUG("Unable to enable metric set\n"); goto err_enable; @@ -3147,7 +3169,7 @@ static long i915_perf_config_locked(struct i915_perf_stream *stream, return -EINVAL; if (config != stream->oa_config) { - int err; + struct i915_request *rq; /* * If OA is bound to a specific context, emit the @@ -3158,11 +3180,13 @@ static long i915_perf_config_locked(struct i915_perf_stream *stream, * When set globally, we use a low priority kernel context, * so it will effectively take effect when idle. */ - err = emit_oa_config(stream, config, oa_context(stream)); - if (err == 0) + rq = emit_oa_config(stream, config, oa_context(stream)); + if (!IS_ERR(rq)) { config = xchg(&stream->oa_config, config); - else - ret = err; + i915_request_put(rq); + } else { + ret = PTR_ERR(rq); + } } i915_oa_config_put(config); diff --git a/drivers/gpu/drm/i915/i915_perf_types.h b/drivers/gpu/drm/i915/i915_perf_types.h index 45e581455f5d..a0e22f00f6cf 100644 --- a/drivers/gpu/drm/i915/i915_perf_types.h +++ b/drivers/gpu/drm/i915/i915_perf_types.h @@ -339,7 +339,8 @@ struct i915_oa_ops { * counter reports being sampled. May apply system constraints such as * disabling EU clock gating as required. */ - int (*enable_metric_set)(struct i915_perf_stream *stream); + struct i915_request * + (*enable_metric_set)(struct i915_perf_stream *stream); /** * @disable_metric_set: Remove system constraints associated with using From 169c0aa4bc17d37370f55188d9327b99d60fd9d7 Mon Sep 17 00:00:00 2001 From: Chris Wilson Date: Tue, 3 Mar 2020 14:00:09 +0000 Subject: [PATCH 9/9] drm/i915/gt: Drop the timeline->mutex as we wait for retirement As we have pinned the timeline (using tl->active_count), we can safely drop the tl->mutex as we wait for what we believe to be the final request on that timeline. This is useful for ensuring that we do not block the engine heartbeat by hogging the kernel_context's timeline on a dead GPU. References: https://gitlab.freedesktop.org/drm/intel/issues/1364 Fixes: 058179e72e09 ("drm/i915/gt: Replace hangcheck by heartbeats") Fixes: f33a8a51602c ("drm/i915: Merge wait_for_timelines with retire_request") Signed-off-by: Chris Wilson Cc: Mika Kuoppala Reviewed-by: Mika Kuoppala Link: https://patchwork.freedesktop.org/patch/msgid/20200303140009.1494819-1-chris@chris-wilson.co.uk (cherry picked from commit 82126e596d8519baac416aee83cad938f1d23cf8) Signed-off-by: Jani Nikula --- drivers/gpu/drm/i915/gt/intel_gt_requests.c | 14 +++++++++++--- 1 file changed, 11 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.c b/drivers/gpu/drm/i915/gt/intel_gt_requests.c index 8a5054f21bf8..24c99d0838af 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_requests.c +++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.c @@ -147,24 +147,32 @@ long intel_gt_retire_requests_timeout(struct intel_gt *gt, long timeout) fence = i915_active_fence_get(&tl->last_request); if (fence) { + mutex_unlock(&tl->mutex); + timeout = dma_fence_wait_timeout(fence, interruptible, timeout); dma_fence_put(fence); + + /* Retirement is best effort */ + if (!mutex_trylock(&tl->mutex)) { + active_count++; + goto out_active; + } } } if (!retire_requests(tl) || flush_submission(gt)) active_count++; + mutex_unlock(&tl->mutex); - spin_lock(&timelines->lock); +out_active: spin_lock(&timelines->lock); - /* Resume iteration after dropping lock */ + /* Resume list iteration after reacquiring spinlock */ list_safe_reset_next(tl, tn, link); if (atomic_dec_and_test(&tl->active_count)) list_del(&tl->link); - mutex_unlock(&tl->mutex); /* Defer the final release to after the spinlock */ if (refcount_dec_and_test(&tl->kref.refcount)) {