From 41f425adbce9013a0e91a9670bf43fa141ab9414 Mon Sep 17 00:00:00 2001 From: Matt Roper Date: Thu, 17 Nov 2022 09:33:58 -0800 Subject: [PATCH 01/13] drm/i915/gt: Manage uncore->lock while waiting on MCR register The GT MCR code currently relies on uncore->lock to avoid race conditions on the steering control register during MCR operations. The *_fw() versions of MCR operations expect the caller to already hold uncore->lock, while the non-fw variants manage the lock internally. However the sole callsite of intel_gt_mcr_wait_for_reg_fw() does not currently obtain the forcewake lock, allowing a potential race condition (and triggering an assertion on lockdep builds). Furthermore, since 'wait for register value' requests may not return immediately, it is undesirable to hold a fundamental lock like uncore->lock for the entire wait and block all other MMIO for the duration; rather the lock is only needed around the MCR read operations and can be released during the delays. Convert intel_gt_mcr_wait_for_reg_fw() to a non-fw variant that will manage uncore->lock internally. This does have the side effect of causing an unnecessary lookup in the forcewake table on each read operation, but since the caller is still holding the relevant forcewake domain, this will ultimately just incremenent the reference count and won't actually cause any additional MMIO traffic. In the future we plan to switch to a dedicated MCR lock to protect the steering critical section rather than using the overloaded and high-traffic uncore->lock; on MTL and beyond the new lock can be implemented on top of the hardware-provided synchonization mechanism for steering. Fixes: 3068bec83eea ("drm/i915/gt: Add intel_gt_mcr_wait_for_reg_fw()") Cc: Lucas De Marchi Signed-off-by: Matt Roper Reviewed-by: Lucas De Marchi Link: https://patchwork.freedesktop.org/patch/msgid/20221117173358.1980230-1-matthew.d.roper@intel.com (cherry picked from commit 192bb40f030a41ca95c5cff8c9340b725bc7ba8b) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_gt.c | 6 +++--- drivers/gpu/drm/i915/gt/intel_gt_mcr.c | 18 ++++++++++-------- drivers/gpu/drm/i915/gt/intel_gt_mcr.h | 12 ++++++------ 3 files changed, 19 insertions(+), 17 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c index 0325f071046ca..b5ad9caa55372 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt.c +++ b/drivers/gpu/drm/i915/gt/intel_gt.c @@ -1035,9 +1035,9 @@ get_reg_and_bit(const struct intel_engine_cs *engine, const bool gen8, static int wait_for_invalidate(struct intel_gt *gt, struct reg_and_bit rb) { if (GRAPHICS_VER_FULL(gt->i915) >= IP_VER(12, 50)) - return intel_gt_mcr_wait_for_reg_fw(gt, rb.mcr_reg, rb.bit, 0, - TLB_INVAL_TIMEOUT_US, - TLB_INVAL_TIMEOUT_MS); + return intel_gt_mcr_wait_for_reg(gt, rb.mcr_reg, rb.bit, 0, + TLB_INVAL_TIMEOUT_US, + TLB_INVAL_TIMEOUT_MS); else return __intel_wait_for_register_fw(gt->uncore, rb.reg, rb.bit, 0, TLB_INVAL_TIMEOUT_US, diff --git a/drivers/gpu/drm/i915/gt/intel_gt_mcr.c b/drivers/gpu/drm/i915/gt/intel_gt_mcr.c index 830edffe88cce..d9a8ff9e5e578 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_mcr.c +++ b/drivers/gpu/drm/i915/gt/intel_gt_mcr.c @@ -730,17 +730,19 @@ void intel_gt_mcr_get_ss_steering(struct intel_gt *gt, unsigned int dss, * * Return: 0 if the register matches the desired condition, or -ETIMEDOUT. */ -int intel_gt_mcr_wait_for_reg_fw(struct intel_gt *gt, - i915_mcr_reg_t reg, - u32 mask, - u32 value, - unsigned int fast_timeout_us, - unsigned int slow_timeout_ms) +int intel_gt_mcr_wait_for_reg(struct intel_gt *gt, + i915_mcr_reg_t reg, + u32 mask, + u32 value, + unsigned int fast_timeout_us, + unsigned int slow_timeout_ms) { - u32 reg_value = 0; -#define done (((reg_value = intel_gt_mcr_read_any_fw(gt, reg)) & mask) == value) int ret; + lockdep_assert_not_held(>->uncore->lock); + +#define done ((intel_gt_mcr_read_any(gt, reg) & mask) == value) + /* Catch any overuse of this function */ might_sleep_if(slow_timeout_ms); GEM_BUG_ON(fast_timeout_us > 20000); diff --git a/drivers/gpu/drm/i915/gt/intel_gt_mcr.h b/drivers/gpu/drm/i915/gt/intel_gt_mcr.h index 3fb0502bff22d..ae93b20e1c178 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_mcr.h +++ b/drivers/gpu/drm/i915/gt/intel_gt_mcr.h @@ -37,12 +37,12 @@ void intel_gt_mcr_report_steering(struct drm_printer *p, struct intel_gt *gt, void intel_gt_mcr_get_ss_steering(struct intel_gt *gt, unsigned int dss, unsigned int *group, unsigned int *instance); -int intel_gt_mcr_wait_for_reg_fw(struct intel_gt *gt, - i915_mcr_reg_t reg, - u32 mask, - u32 value, - unsigned int fast_timeout_us, - unsigned int slow_timeout_ms); +int intel_gt_mcr_wait_for_reg(struct intel_gt *gt, + i915_mcr_reg_t reg, + u32 mask, + u32 value, + unsigned int fast_timeout_us, + unsigned int slow_timeout_ms); /* * Helper for for_each_ss_steering loop. On pre-Xe_HP platforms, subslice From eaa96d2c22989d7d2b54a4cfcf885bd5986be6a7 Mon Sep 17 00:00:00 2001 From: Tvrtko Ursulin Date: Fri, 18 Nov 2022 11:52:49 +0000 Subject: [PATCH 02/13] drm/i915: Fix workarounds on Gen2-3 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit In 3653727560d0 ("drm/i915: Simplify internal helper function signature") I broke the old platforms by not noticing engine workaround init does not initialize the list on old platforms. Fix it by always initializing which already does the right thing by mostly not doing anything if there aren't any workarounds on the list. Signed-off-by: Tvrtko Ursulin Fixes: 3653727560d0 ("drm/i915: Simplify internal helper function signature") Reported-by: Ville Syrjälä Cc: Mika Kuoppala Reviewed-by: Matt Roper Link: https://patchwork.freedesktop.org/patch/msgid/20221118115249.2683946-1-tvrtko.ursulin@linux.intel.com (cherry picked from commit 71feb6f901ecba962177a0a029dc545c91a4b396) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_workarounds.c | 5 +---- 1 file changed, 1 insertion(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_workarounds.c b/drivers/gpu/drm/i915/gt/intel_workarounds.c index 1bd8d63ad4f3f..2afb4f80a954d 100644 --- a/drivers/gpu/drm/i915/gt/intel_workarounds.c +++ b/drivers/gpu/drm/i915/gt/intel_workarounds.c @@ -3011,7 +3011,7 @@ general_render_compute_wa_init(struct intel_engine_cs *engine, struct i915_wa_li static void engine_init_workarounds(struct intel_engine_cs *engine, struct i915_wa_list *wal) { - if (I915_SELFTEST_ONLY(GRAPHICS_VER(engine->i915) < 4)) + if (GRAPHICS_VER(engine->i915) < 4) return; engine_fake_wa_init(engine, wal); @@ -3036,9 +3036,6 @@ void intel_engine_init_workarounds(struct intel_engine_cs *engine) { struct i915_wa_list *wal = &engine->wa_list; - if (GRAPHICS_VER(engine->i915) < 4) - return; - wa_init_start(wal, engine->gt, "engine", engine->name); engine_init_workarounds(engine, wal); wa_init_finish(wal); From ddb97ea7cdb6462d7a719c649f58858b083f7eed Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Fri, 18 Nov 2022 20:52:01 +0200 Subject: [PATCH 03/13] drm/i915: Remove non-existent pipes from bigjoiner pipe mask MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit bigjoiner_pipes() doesn't consider that: - RKL only has three pipes - some pipes may be fused off This means that intel_atomic_check_bigjoiner() won't reject all configurations that would need a non-existent pipe. Instead we just keep on rolling witout actually having reserved the slave pipe we need. It's possible that we don't outright explode anywhere due to this since eg. for_each_intel_crtc_in_pipe_mask() will only walk the crtcs we've registered even though the passed in pipe_mask asks for more of them. But clearly the thing won't do what is expected of it when the required pipes are not present. Fix the problem by consulting the device info pipe_mask already in bigjoiner_pipes(). Cc: stable@vger.kernel.org Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20221118185201.10469-1-ville.syrjala@linux.intel.com Reviewed-by: Arun R Murthy (cherry picked from commit f1c87a94a1087a26f41007ee83264033007421b5) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_display.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index b3e23708d1942..6c2686ecb62a2 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -3733,12 +3733,16 @@ static bool ilk_get_pipe_config(struct intel_crtc *crtc, static u8 bigjoiner_pipes(struct drm_i915_private *i915) { + u8 pipes; + if (DISPLAY_VER(i915) >= 12) - return BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D); + pipes = BIT(PIPE_A) | BIT(PIPE_B) | BIT(PIPE_C) | BIT(PIPE_D); else if (DISPLAY_VER(i915) >= 11) - return BIT(PIPE_B) | BIT(PIPE_C); + pipes = BIT(PIPE_B) | BIT(PIPE_C); else - return 0; + pipes = 0; + + return pipes & RUNTIME_INFO(i915)->pipe_mask; } static bool transcoder_ddi_func_is_enabled(struct drm_i915_private *dev_priv, From dfa5e6ef3ccefff9fa8a70d9f5fa6ef6244aa312 Mon Sep 17 00:00:00 2001 From: Jani Nikula Date: Tue, 22 Nov 2022 16:16:16 +0200 Subject: [PATCH 04/13] drm/i915/guc: make default_lists const data The default_lists array should be in rodata. Fixes: dce2bd542337 ("drm/i915/guc: Add Gen9 registers for GuC error state capture.") Cc: Alan Previn Cc: Umesh Nerlige Ramappa Cc: Lucas De Marchi Signed-off-by: Jani Nikula Reviewed-by: Lucas De Marchi Link: https://patchwork.freedesktop.org/patch/msgid/20221122141616.3469214-1-jani.nikula@intel.com (cherry picked from commit 8b7f7a9b10b704ba7d73199ff0f01354e0bad7a5) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c index 1d49a7ec0bd8f..1c1b85073b4bd 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_capture.c @@ -170,7 +170,7 @@ static const struct __guc_mmio_reg_descr empty_regs_list[] = { } /* List of lists */ -static struct __guc_mmio_reg_descr_group default_lists[] = { +static const struct __guc_mmio_reg_descr_group default_lists[] = { MAKE_REGLIST(default_global_regs, PF, GLOBAL, 0), MAKE_REGLIST(default_rc_class_regs, PF, ENGINE_CLASS, GUC_RENDER_CLASS), MAKE_REGLIST(xe_lpd_rc_inst_regs, PF, ENGINE_INSTANCE, GUC_RENDER_CLASS), From 71b6b2557058d4374d11e792ec550a1a098fb7a9 Mon Sep 17 00:00:00 2001 From: Umesh Nerlige Ramappa Date: Thu, 10 Nov 2022 17:19:12 +0000 Subject: [PATCH 05/13] i915/uncore: Acquire fw before loop in intel_uncore_read64_2x32 PMU reads the GT timestamp as a 2x32 mmio read and since upper and lower 32 bit registers are read in a loop, there is a latency involved between getting the GT timestamp and the CPU timestamp. As part of the resolution, refactor intel_uncore_read64_2x32 to acquire forcewake and uncore lock prior to reading upper and lower regs. Signed-off-by: Umesh Nerlige Ramappa Reviewed-by: Tvrtko Ursulin Reviewed-by: Ashutosh Dixit Signed-off-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20221110171913.670286-2-umesh.nerlige.ramappa@intel.com (cherry picked from commit e746f84b8e813816951b63485134927ed6763a1b) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/intel_uncore.h | 44 ++++++++++++++++++++--------- 1 file changed, 30 insertions(+), 14 deletions(-) diff --git a/drivers/gpu/drm/i915/intel_uncore.h b/drivers/gpu/drm/i915/intel_uncore.h index 5449146a06247..e9e38490815d0 100644 --- a/drivers/gpu/drm/i915/intel_uncore.h +++ b/drivers/gpu/drm/i915/intel_uncore.h @@ -382,20 +382,6 @@ __uncore_write(write_notrace, 32, l, false) */ __uncore_read(read64, 64, q, true) -static inline u64 -intel_uncore_read64_2x32(struct intel_uncore *uncore, - i915_reg_t lower_reg, i915_reg_t upper_reg) -{ - u32 upper, lower, old_upper, loop = 0; - upper = intel_uncore_read(uncore, upper_reg); - do { - old_upper = upper; - lower = intel_uncore_read(uncore, lower_reg); - upper = intel_uncore_read(uncore, upper_reg); - } while (upper != old_upper && loop++ < 2); - return (u64)upper << 32 | lower; -} - #define intel_uncore_posting_read(...) ((void)intel_uncore_read_notrace(__VA_ARGS__)) #define intel_uncore_posting_read16(...) ((void)intel_uncore_read16_notrace(__VA_ARGS__)) @@ -455,6 +441,36 @@ static inline void intel_uncore_rmw_fw(struct intel_uncore *uncore, intel_uncore_write_fw(uncore, reg, val); } +static inline u64 +intel_uncore_read64_2x32(struct intel_uncore *uncore, + i915_reg_t lower_reg, i915_reg_t upper_reg) +{ + u32 upper, lower, old_upper, loop = 0; + enum forcewake_domains fw_domains; + unsigned long flags; + + fw_domains = intel_uncore_forcewake_for_reg(uncore, lower_reg, + FW_REG_READ); + + fw_domains |= intel_uncore_forcewake_for_reg(uncore, upper_reg, + FW_REG_READ); + + spin_lock_irqsave(&uncore->lock, flags); + intel_uncore_forcewake_get__locked(uncore, fw_domains); + + upper = intel_uncore_read_fw(uncore, upper_reg); + do { + old_upper = upper; + lower = intel_uncore_read_fw(uncore, lower_reg); + upper = intel_uncore_read_fw(uncore, upper_reg); + } while (upper != old_upper && loop++ < 2); + + intel_uncore_forcewake_put__locked(uncore, fw_domains); + spin_unlock_irqrestore(&uncore->lock, flags); + + return (u64)upper << 32 | lower; +} + static inline int intel_uncore_write_and_verify(struct intel_uncore *uncore, i915_reg_t reg, u32 val, u32 mask, u32 expected_val) From 264bc5d3ebc5d695339ff84d8635773c23e34ba9 Mon Sep 17 00:00:00 2001 From: Umesh Nerlige Ramappa Date: Thu, 10 Nov 2022 17:19:13 +0000 Subject: [PATCH 06/13] drm/i915/selftest: Bump up sample period for busy stats selftest Engine busyness samples around a 10ms period is failing with busyness ranging approx. from 87% to 115% as shown below. The expected range is +/- 5% of the sample period. Fail 10% of the time. rcs0: reported 11716042ns [91%] busyness while spinning [for 12805719ns] When determining busyness of active engine, the GuC based engine busyness implementation relies on a 64 bit timestamp register read. The latency incurred by this register read causes the failure. On DG1, when the test fails, the observed latencies range from 900us - 1.5ms. Optimizing the 2x32 read by acquiring the lock and forcewake prior to all reg reads reduces the rate of failure to around 2%, but does not eliminate it. In order to make the selftest more robust and always account for such latencies, increase the sample period to 100 ms. This eliminates the issue as seen in a 1000 runs. v2: (Ashutosh) - Add error to commit msg - Include gitlab bug - Update commit for inclusion of 2x32 optimized read Closes: https://gitlab.freedesktop.org/drm/intel/-/issues/4418 Signed-off-by: Umesh Nerlige Ramappa Acked-by: Tvrtko Ursulin Reviewed-by: Ashutosh Dixit Signed-off-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20221110171913.670286-3-umesh.nerlige.ramappa@intel.com (cherry picked from commit 529d95a6067b74da9d4d5d9ab3009b35c98c5fce) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/selftest_engine_pm.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/selftest_engine_pm.c b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c index 0dcb3ed44a737..87c94314cf674 100644 --- a/drivers/gpu/drm/i915/gt/selftest_engine_pm.c +++ b/drivers/gpu/drm/i915/gt/selftest_engine_pm.c @@ -317,7 +317,7 @@ static int live_engine_busy_stats(void *arg) ENGINE_TRACE(engine, "measuring busy time\n"); preempt_disable(); de = intel_engine_get_busy_time(engine, &t[0]); - mdelay(10); + mdelay(100); de = ktime_sub(intel_engine_get_busy_time(engine, &t[1]), de); preempt_enable(); dt = ktime_sub(t[1], t[0]); From 8e5f37828145fa662f30f890f74924d55255d68d Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Thu, 10 Nov 2022 16:56:51 -0800 Subject: [PATCH 07/13] drm/i915/huc: fix leak of debug object in huc load fence on driver unload The fence is always initialized in huc_init_early, but the cleanup in huc_fini is only being run if HuC is enabled. This causes a leaking of the debug object when HuC is disabled/not supported, which can in turn trigger a warning if we try to register a new debug offset at the same address on driver reload. To fix the issue, make sure to always run the cleanup code. Reported-by: Tvrtko Ursulin Reported-by: Brian Norris Fixes: 27536e03271d ("drm/i915/huc: track delayed HuC load with a fence") Signed-off-by: Daniele Ceraolo Spurio Cc: Tvrtko Ursulin Cc: Brian Norris Cc: Alan Previn Cc: John Harrison Tested-by: Brian Norris Reviewed-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20221111005651.4160369-1-daniele.ceraolospurio@intel.com (cherry picked from commit 02224691cb0f367acb476911bddfa21e2d596ca5) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/uc/intel_huc.c | 12 +++++++----- drivers/gpu/drm/i915/gt/uc/intel_uc.c | 1 + 2 files changed, 8 insertions(+), 5 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc.c b/drivers/gpu/drm/i915/gt/uc/intel_huc.c index be855811d85df..0976e9101346f 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_huc.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_huc.c @@ -329,13 +329,15 @@ int intel_huc_init(struct intel_huc *huc) void intel_huc_fini(struct intel_huc *huc) { - if (!intel_uc_fw_is_loadable(&huc->fw)) - return; - + /* + * the fence is initialized in init_early, so we need to clean it up + * even if HuC loading is off. + */ delayed_huc_load_complete(huc); - i915_sw_fence_fini(&huc->delayed_load.fence); - intel_uc_fw_fini(&huc->fw); + + if (intel_uc_fw_is_loadable(&huc->fw)) + intel_uc_fw_fini(&huc->fw); } void intel_huc_suspend(struct intel_huc *huc) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_uc.c b/drivers/gpu/drm/i915/gt/uc/intel_uc.c index 1d28286e6f066..2a508b137e900 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_uc.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_uc.c @@ -722,6 +722,7 @@ int intel_uc_runtime_resume(struct intel_uc *uc) static const struct intel_uc_ops uc_ops_off = { .init_hw = __uc_check_hw, + .fini = __uc_fini, /* to clean-up the init_early initialization */ }; static const struct intel_uc_ops uc_ops_on = { From 6a5347293992e0412bc748dae11228a7081393fa Mon Sep 17 00:00:00 2001 From: Janusz Krzysztofik Date: Mon, 21 Nov 2022 15:56:54 +0100 Subject: [PATCH 08/13] drm/i915: Fix negative value passed as remaining time Commit b97060a99b01 ("drm/i915/guc: Update intel_gt_wait_for_idle to work with GuC") extended the API of intel_gt_retire_requests_timeout() with an extra argument 'remaining_timeout', intended for passing back unconsumed portion of requested timeout when 0 (success) is returned. However, when request retirement happens to succeed despite an error returned by a call to dma_fence_wait_timeout(), that error code (a negative value) is passed back instead of remaining time. If we then pass that negative value forward as requested timeout to intel_uc_wait_for_idle(), an explicit BUG will be triggered. If request retirement succeeds but an error code is passed back via remaininig_timeout, we may have no clue on how much of the initial timeout might have been left for spending it on waiting for GuC to become idle. OTOH, since all pending requests have been successfully retired, that error code has been already ignored by intel_gt_retire_requests_timeout(), then we shouldn't fail. Assume no more time has been left on error and pass 0 timeout value to intel_uc_wait_for_idle() to give it a chance to return success if GuC is already idle. v3: Don't fail on any error passed back via remaining_timeout. v2: Fix the issue on the caller side, not the provider. Fixes: b97060a99b01 ("drm/i915/guc: Update intel_gt_wait_for_idle to work with GuC") Signed-off-by: Janusz Krzysztofik Cc: stable@vger.kernel.org # v5.15+ Reviewed-by: Andrzej Hajda Signed-off-by: Tvrtko Ursulin Link: https://patchwork.freedesktop.org/patch/msgid/20221121145655.75141-2-janusz.krzysztofik@linux.intel.com (cherry picked from commit f235dbd5b768e238d365fd05d92de5a32abc1c1f) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_gt.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/intel_gt.c b/drivers/gpu/drm/i915/gt/intel_gt.c index b5ad9caa55372..7ef0edb2e37cd 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt.c +++ b/drivers/gpu/drm/i915/gt/intel_gt.c @@ -677,8 +677,13 @@ int intel_gt_wait_for_idle(struct intel_gt *gt, long timeout) return -EINTR; } - return timeout ? timeout : intel_uc_wait_for_idle(>->uc, - remaining_timeout); + if (timeout) + return timeout; + + if (remaining_timeout < 0) + remaining_timeout = 0; + + return intel_uc_wait_for_idle(>->uc, remaining_timeout); } int intel_gt_init(struct intel_gt *gt) From 35aba5f51a39fb95351844ffb14ec02b8970e19f Mon Sep 17 00:00:00 2001 From: Janusz Krzysztofik Date: Mon, 21 Nov 2022 15:56:55 +0100 Subject: [PATCH 09/13] drm/i915: Never return 0 if not all requests retired Users of intel_gt_retire_requests_timeout() expect 0 return value on success. However, we have no protection from passing back 0 potentially returned by a call to dma_fence_wait_timeout() when it succedes right after its timeout has expired. Replace 0 with -ETIME before potentially using the timeout value as return code, so -ETIME is returned if there are still some requests not retired after timeout, 0 otherwise. v3: Use conditional expression, more compact but also better reflecting intention standing behind the change. v2: Move the added lines down so flush_submission() is not affected. Fixes: f33a8a51602c ("drm/i915: Merge wait_for_timelines with retire_request") Signed-off-by: Janusz Krzysztofik Reviewed-by: Andrzej Hajda Cc: stable@vger.kernel.org # v5.5+ Signed-off-by: Tvrtko Ursulin Link: https://patchwork.freedesktop.org/patch/msgid/20221121145655.75141-3-janusz.krzysztofik@linux.intel.com (cherry picked from commit f301a29f143760ce8d3d6b6a8436d45d3448cde6) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/intel_gt_requests.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/gt/intel_gt_requests.c b/drivers/gpu/drm/i915/gt/intel_gt_requests.c index edb881d756309..1dfd01668c79c 100644 --- a/drivers/gpu/drm/i915/gt/intel_gt_requests.c +++ b/drivers/gpu/drm/i915/gt/intel_gt_requests.c @@ -199,7 +199,7 @@ out_active: spin_lock(&timelines->lock); if (remaining_timeout) *remaining_timeout = timeout; - return active_count ? timeout : 0; + return active_count ? timeout ?: -ETIME : 0; } static void retire_work_handler(struct work_struct *work) From 1382901f75a5a7dc8eac05059fd0c7816def4eae Mon Sep 17 00:00:00 2001 From: Xia Fukun Date: Fri, 25 Nov 2022 14:34:28 +0800 Subject: [PATCH 10/13] drm/i915/bios: fix a memory leak in generate_lfp_data_ptrs When (size != 0 || ptrs->lvds_ entries != 3), the program tries to free() the ptrs. However, the ptrs is not created by calling kzmalloc(), but is obtained by pointer offset operation. This may lead to memory leaks or undefined behavior. Fix this by replacing the arguments of kfree() with ptrs_block. Fixes: a87d0a847607 ("drm/i915/bios: Generate LFP data table pointers if the VBT lacks them") Signed-off-by: Xia Fukun Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20221125063428.69486-1-xiafukun@huawei.com (cherry picked from commit 7674cd0b7d28b952151c3df26bbfa7e07eb2b4ec) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_bios.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/display/intel_bios.c b/drivers/gpu/drm/i915/display/intel_bios.c index c2987f2c2b2e2..572a4e3769f38 100644 --- a/drivers/gpu/drm/i915/display/intel_bios.c +++ b/drivers/gpu/drm/i915/display/intel_bios.c @@ -414,7 +414,7 @@ static void *generate_lfp_data_ptrs(struct drm_i915_private *i915, ptrs->lvds_entries++; if (size != 0 || ptrs->lvds_entries != 3) { - kfree(ptrs); + kfree(ptrs_block); return NULL; } From f9cdf4130671d767071607d0a7568c9bd36a68d0 Mon Sep 17 00:00:00 2001 From: Mikko Kovanen Date: Sat, 26 Nov 2022 13:27:13 +0000 Subject: [PATCH 11/13] drm/i915/dsi: fix VBT send packet port selection for dual link DSI intel_dsi->ports contains bitmask of enabled ports and correspondingly logic for selecting port for VBT packet sending must use port specific bitmask when deciding appropriate port. Fixes: 08c59dde71b7 ("drm/i915/dsi: fix VBT send packet port selection for ICL+") Cc: stable@vger.kernel.org Signed-off-by: Mikko Kovanen Reviewed-by: Jani Nikula Signed-off-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/DBBPR09MB466592B16885D99ABBF2393A91119@DBBPR09MB4665.eurprd09.prod.outlook.com (cherry picked from commit 8d58bb7991c45f6b60710cc04c9498c6ea96db90) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_dsi_vbt.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_dsi_vbt.c b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c index 75e8cc4337c93..fce69fa446d58 100644 --- a/drivers/gpu/drm/i915/display/intel_dsi_vbt.c +++ b/drivers/gpu/drm/i915/display/intel_dsi_vbt.c @@ -137,9 +137,9 @@ static enum port intel_dsi_seq_port_to_port(struct intel_dsi *intel_dsi, return ffs(intel_dsi->ports) - 1; if (seq_port) { - if (intel_dsi->ports & PORT_B) + if (intel_dsi->ports & BIT(PORT_B)) return PORT_B; - else if (intel_dsi->ports & PORT_C) + else if (intel_dsi->ports & BIT(PORT_C)) return PORT_C; } From 2d299f5bfe212b68bde981153a00926e817ee82a Mon Sep 17 00:00:00 2001 From: Daniele Ceraolo Spurio Date: Wed, 23 Nov 2022 15:54:17 -0800 Subject: [PATCH 12/13] drm/i915/huc: always init the delayed load fence The fence is only tracking if the HuC load is in progress or not and doesn't distinguish between already loaded, not supported or disabled, so we can always initialize it to completed, no matter the actual support. We already do that for most platforms, but we skip it on GTs that lack VCS engines (e.g. MTL root GT), so fix that. Note that the cleanup is already unconditional. While at it, move the init/fini to helper functions. Fixes: 8e5f37828145 ("drm/i915/huc: fix leak of debug object in huc load fence on driver unload") Signed-off-by: Daniele Ceraolo Spurio Cc: John Harrison Cc: Alan Previn Reviewed-by: John Harrison Link: https://patchwork.freedesktop.org/patch/msgid/20221123235417.1475709-1-daniele.ceraolospurio@intel.com (cherry picked from commit 14347a9c889fbdbae81e500f6c6e313f5d8e5271) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/uc/intel_huc.c | 47 +++++++++++++++++++------- 1 file changed, 34 insertions(+), 13 deletions(-) diff --git a/drivers/gpu/drm/i915/gt/uc/intel_huc.c b/drivers/gpu/drm/i915/gt/uc/intel_huc.c index 0976e9101346f..410905da8e974 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_huc.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_huc.c @@ -211,6 +211,30 @@ void intel_huc_unregister_gsc_notifier(struct intel_huc *huc, struct bus_type *b huc->delayed_load.nb.notifier_call = NULL; } +static void delayed_huc_load_init(struct intel_huc *huc) +{ + /* + * Initialize fence to be complete as this is expected to be complete + * unless there is a delayed HuC load in progress. + */ + i915_sw_fence_init(&huc->delayed_load.fence, + sw_fence_dummy_notify); + i915_sw_fence_commit(&huc->delayed_load.fence); + + hrtimer_init(&huc->delayed_load.timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); + huc->delayed_load.timer.function = huc_delayed_load_timer_callback; +} + +static void delayed_huc_load_fini(struct intel_huc *huc) +{ + /* + * the fence is initialized in init_early, so we need to clean it up + * even if HuC loading is off. + */ + delayed_huc_load_complete(huc); + i915_sw_fence_fini(&huc->delayed_load.fence); +} + static bool vcs_supported(struct intel_gt *gt) { intel_engine_mask_t mask = gt->info.engine_mask; @@ -241,6 +265,15 @@ void intel_huc_init_early(struct intel_huc *huc) intel_uc_fw_init_early(&huc->fw, INTEL_UC_FW_TYPE_HUC); + /* + * we always init the fence as already completed, even if HuC is not + * supported. This way we don't have to distinguish between HuC not + * supported/disabled or already loaded, and can focus on if the load + * is currently in progress (fence not complete) or not, which is what + * we care about for stalling userspace submissions. + */ + delayed_huc_load_init(huc); + if (!vcs_supported(gt)) { intel_uc_fw_change_status(&huc->fw, INTEL_UC_FIRMWARE_NOT_SUPPORTED); return; @@ -255,17 +288,6 @@ void intel_huc_init_early(struct intel_huc *huc) huc->status.mask = HUC_FW_VERIFIED; huc->status.value = HUC_FW_VERIFIED; } - - /* - * Initialize fence to be complete as this is expected to be complete - * unless there is a delayed HuC reload in progress. - */ - i915_sw_fence_init(&huc->delayed_load.fence, - sw_fence_dummy_notify); - i915_sw_fence_commit(&huc->delayed_load.fence); - - hrtimer_init(&huc->delayed_load.timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL); - huc->delayed_load.timer.function = huc_delayed_load_timer_callback; } #define HUC_LOAD_MODE_STRING(x) (x ? "GSC" : "legacy") @@ -333,8 +355,7 @@ void intel_huc_fini(struct intel_huc *huc) * the fence is initialized in init_early, so we need to clean it up * even if HuC loading is off. */ - delayed_huc_load_complete(huc); - i915_sw_fence_fini(&huc->delayed_load.fence); + delayed_huc_load_fini(huc); if (intel_uc_fw_is_loadable(&huc->fw)) intel_uc_fw_fini(&huc->fw); From 7d579eff3e55a256fdd71624543747ed4d7fa38b Mon Sep 17 00:00:00 2001 From: Gustavo Sousa Date: Thu, 24 Nov 2022 13:21:23 -0300 Subject: [PATCH 13/13] drm/i915/dmc: Update DG2 DMC version to v2.08 Release notes: 1. Fixes for Register noclaims and few restore. Fixes: c4cf059d9c2c ("drm/i915/dmc: Update DG2 DMC firmware to v2.07") Signed-off-by: Gustavo Sousa Reviewed-by: Anusha Srivatsa Signed-off-by: Rodrigo Vivi Link: https://patchwork.freedesktop.org/patch/msgid/20221124162123.16870-1-gustavo.sousa@intel.com (cherry picked from commit 6ee6692520133a14b0d0f3ddddf8c44783cfee06) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_dmc.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_dmc.c b/drivers/gpu/drm/i915/display/intel_dmc.c index 081a4d0083b13..eff3add706117 100644 --- a/drivers/gpu/drm/i915/display/intel_dmc.c +++ b/drivers/gpu/drm/i915/display/intel_dmc.c @@ -52,8 +52,8 @@ #define DISPLAY_VER12_DMC_MAX_FW_SIZE ICL_DMC_MAX_FW_SIZE -#define DG2_DMC_PATH DMC_PATH(dg2, 2, 07) -#define DG2_DMC_VERSION_REQUIRED DMC_VERSION(2, 07) +#define DG2_DMC_PATH DMC_PATH(dg2, 2, 08) +#define DG2_DMC_VERSION_REQUIRED DMC_VERSION(2, 8) MODULE_FIRMWARE(DG2_DMC_PATH); #define ADLP_DMC_PATH DMC_PATH(adlp, 2, 16)