From e49477f7f78598295551d486ecc7f020d796432e Mon Sep 17 00:00:00 2001 From: Krzysztof Karas Date: Thu, 16 Jan 2025 10:40:46 +0000 Subject: [PATCH 1/6] drm/i915/gt: Use spin_lock_irqsave() in interruptible context spin_lock/unlock() functions used in interrupt contexts could result in a deadlock, as seen in GitLab issue #13399, which occurs when interrupt comes in while holding a lock. Try to remedy the problem by saving irq state before spin lock acquisition. v2: add irqs' state save/restore calls to all locks/unlocks in signal_irq_work() execution (Maciej) v3: use with spin_lock_irqsave() in guc_lrc_desc_unpin() instead of other lock/unlock calls and add Fixes and Cc tags (Tvrtko); change title and commit message Fixes: 2f2cc53b5fe7 ("drm/i915/guc: Close deregister-context race against CT-loss") Closes: https://gitlab.freedesktop.org/drm/i915/kernel/-/issues/13399 Signed-off-by: Krzysztof Karas Cc: # v6.9+ Reviewed-by: Maciej Patelczyk Reviewed-by: Andi Shyti Signed-off-by: Andi Shyti Link: https://patchwork.freedesktop.org/patch/msgid/pusppq5ybyszau2oocboj3mtj5x574gwij323jlclm5zxvimmu@mnfg6odxbpsv (cherry picked from commit c088387ddd6482b40f21ccf23db1125e8fa4af7e) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c | 4 ++-- 1 file changed, 2 insertions(+), 2 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 cc05bd9e43b4..3fce5c000144 100644 --- a/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c +++ b/drivers/gpu/drm/i915/gt/uc/intel_guc_submission.c @@ -3449,10 +3449,10 @@ static inline int guc_lrc_desc_unpin(struct intel_context *ce) */ ret = deregister_context(ce, ce->guc_id.id); if (ret) { - spin_lock(&ce->guc_state.lock); + spin_lock_irqsave(&ce->guc_state.lock, flags); set_context_registered(ce); clr_context_destroyed(ce); - spin_unlock(&ce->guc_state.lock); + spin_unlock_irqrestore(&ce->guc_state.lock, flags); /* * 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. From 879f70382ff3e92fc854589ada3453e3f5f5b601 Mon Sep 17 00:00:00 2001 From: Imre Deak Date: Fri, 14 Feb 2025 16:19:51 +0200 Subject: [PATCH 2/6] drm/i915/dsi: Use TRANS_DDI_FUNC_CTL's own port width macro The format of the port width field in the DDI_BUF_CTL and the TRANS_DDI_FUNC_CTL registers are different starting with MTL, where the x3 lane mode for HDMI FRL has a different encoding in the two registers. To account for this use the TRANS_DDI_FUNC_CTL's own port width macro. Cc: # v6.5+ Fixes: b66a8abaa48a ("drm/i915/display/mtl: Fill port width in DDI_BUF_/TRANS_DDI_FUNC_/PORT_BUF_CTL for HDMI") Reviewed-by: Jani Nikula Signed-off-by: Imre Deak Link: https://patchwork.freedesktop.org/patch/msgid/20250214142001.552916-2-imre.deak@intel.com (cherry picked from commit 76120b3a304aec28fef4910204b81a12db8974da) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/icl_dsi.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/display/icl_dsi.c b/drivers/gpu/drm/i915/display/icl_dsi.c index c977b74f82f0..82bf6c654de2 100644 --- a/drivers/gpu/drm/i915/display/icl_dsi.c +++ b/drivers/gpu/drm/i915/display/icl_dsi.c @@ -809,8 +809,8 @@ gen11_dsi_configure_transcoder(struct intel_encoder *encoder, /* select data lane width */ tmp = intel_de_read(display, TRANS_DDI_FUNC_CTL(display, dsi_trans)); - tmp &= ~DDI_PORT_WIDTH_MASK; - tmp |= DDI_PORT_WIDTH(intel_dsi->lane_count); + tmp &= ~TRANS_DDI_PORT_WIDTH_MASK; + tmp |= TRANS_DDI_PORT_WIDTH(intel_dsi->lane_count); /* select input pipe */ tmp &= ~TRANS_DDI_EDP_INPUT_MASK; From 166ce267ae3f96e439d8ccc838e8ec4d8b4dab73 Mon Sep 17 00:00:00 2001 From: Imre Deak Date: Fri, 14 Feb 2025 16:19:52 +0200 Subject: [PATCH 3/6] drm/i915/ddi: Fix HDMI port width programming in DDI_BUF_CTL Fix the port width programming in the DDI_BUF_CTL register on MTLP+, where this had an off-by-one error. Cc: # v6.5+ Fixes: b66a8abaa48a ("drm/i915/display/mtl: Fill port width in DDI_BUF_/TRANS_DDI_FUNC_/PORT_BUF_CTL for HDMI") Reviewed-by: Jani Nikula Signed-off-by: Imre Deak Link: https://patchwork.freedesktop.org/patch/msgid/20250214142001.552916-3-imre.deak@intel.com (cherry picked from commit b2ecdabe46d23db275f94cd7c46ca414a144818b) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_ddi.c | 2 +- drivers/gpu/drm/i915/i915_reg.h | 2 +- 2 files changed, 2 insertions(+), 2 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c index acb986bc1f33..2b9240ab547d 100644 --- a/drivers/gpu/drm/i915/display/intel_ddi.c +++ b/drivers/gpu/drm/i915/display/intel_ddi.c @@ -3487,7 +3487,7 @@ static void intel_ddi_enable_hdmi(struct intel_atomic_state *state, intel_de_rmw(dev_priv, XELPDP_PORT_BUF_CTL1(dev_priv, port), XELPDP_PORT_WIDTH_MASK | XELPDP_PORT_REVERSAL, port_buf); - buf_ctl |= DDI_PORT_WIDTH(lane_count); + buf_ctl |= DDI_PORT_WIDTH(crtc_state->lane_count); if (DISPLAY_VER(dev_priv) >= 20) buf_ctl |= XE2LPD_DDI_BUF_D2D_LINK_ENABLE; diff --git a/drivers/gpu/drm/i915/i915_reg.h b/drivers/gpu/drm/i915/i915_reg.h index 765e6c0528fb..786c727aea45 100644 --- a/drivers/gpu/drm/i915/i915_reg.h +++ b/drivers/gpu/drm/i915/i915_reg.h @@ -3633,7 +3633,7 @@ enum skl_power_gate { #define DDI_BUF_IS_IDLE (1 << 7) #define DDI_BUF_CTL_TC_PHY_OWNERSHIP REG_BIT(6) #define DDI_A_4_LANES (1 << 4) -#define DDI_PORT_WIDTH(width) (((width) - 1) << 1) +#define DDI_PORT_WIDTH(width) (((width) == 3 ? 4 : ((width) - 1)) << 1) #define DDI_PORT_WIDTH_MASK (7 << 1) #define DDI_PORT_WIDTH_SHIFT 1 #define DDI_INIT_DISPLAY_DETECTED (1 << 0) From 07fb70d82e0df085980246bf17bc12537588795f Mon Sep 17 00:00:00 2001 From: =?UTF-8?q?Ville=20Syrj=C3=A4l=C3=A4?= Date: Wed, 12 Feb 2025 18:43:21 +0200 Subject: [PATCH 4/6] drm/i915: Make sure all planes in use by the joiner have their crtc included MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Any active plane needs to have its crtc included in the atomic state. For planes enabled via uapi that is all handler in the core. But when we use a plane for joiner the uapi code things the plane is disabled and therefore doesn't have a crtc. So we need to pull those in by hand. We do it first thing in intel_joiner_add_affected_crtcs() so that any newly added crtc will subsequently pull in all of its joined crtcs as well. The symptoms from failing to do this are: - duct tape in the form of commit 1d5b09f8daf8 ("drm/i915: Fix NULL ptr deref by checking new_crtc_state") - the plane's hw state will get overwritten by the disabled uapi state if it can't find the uapi counterpart plane in the atomic state from where it should copy the correct state Cc: stable@vger.kernel.org Reviewed-by: Maarten Lankhorst Signed-off-by: Ville Syrjälä Link: https://patchwork.freedesktop.org/patch/msgid/20250212164330.16891-2-ville.syrjala@linux.intel.com (cherry picked from commit 91077d1deb5374eb8be00fb391710f00e751dc4b) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_display.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/drivers/gpu/drm/i915/display/intel_display.c b/drivers/gpu/drm/i915/display/intel_display.c index 4271da219b41..41128469f12a 100644 --- a/drivers/gpu/drm/i915/display/intel_display.c +++ b/drivers/gpu/drm/i915/display/intel_display.c @@ -6628,12 +6628,30 @@ static int intel_async_flip_check_hw(struct intel_atomic_state *state, struct in static int intel_joiner_add_affected_crtcs(struct intel_atomic_state *state) { struct drm_i915_private *i915 = to_i915(state->base.dev); + const struct intel_plane_state *plane_state; struct intel_crtc_state *crtc_state; + struct intel_plane *plane; struct intel_crtc *crtc; u8 affected_pipes = 0; u8 modeset_pipes = 0; int i; + /* + * Any plane which is in use by the joiner needs its crtc. + * Pull those in first as this will not have happened yet + * if the plane remains disabled according to uapi. + */ + for_each_new_intel_plane_in_state(state, plane, plane_state, i) { + crtc = to_intel_crtc(plane_state->hw.crtc); + if (!crtc) + continue; + + crtc_state = intel_atomic_get_crtc_state(&state->base, crtc); + if (IS_ERR(crtc_state)) + return PTR_ERR(crtc_state); + } + + /* Now pull in all joined crtcs */ for_each_new_intel_crtc_in_state(state, crtc, crtc_state, i) { affected_pipes |= crtc_state->joiner_pipes; if (intel_crtc_needs_modeset(crtc_state)) From b9275eabe31e6679ae12c46a4a0a18d622db4570 Mon Sep 17 00:00:00 2001 From: Imre Deak Date: Tue, 18 Feb 2025 00:38:27 +0200 Subject: [PATCH 5/6] drm/i915/dp: Fix error handling during 128b/132b link training At the end of a 128b/132b link training sequence, the HW expects the transcoder training pattern to be set to TPS2 and from that to normal mode (disabling the training pattern). Transitioning from TPS1 directly to normal mode leaves the transcoder in a stuck state, resulting in page-flip timeouts later in the modeset sequence. Atm, in case of a failure during link training, the transcoder may be still set to output the TPS1 pattern. Later the transcoder is then set from TPS1 directly to normal mode in intel_dp_stop_link_train(), leading to modeset failures later as described above. Fix this by setting the training patter to TPS2, if the link training failed at any point. The clue in the specification about the above HW behavior is the explicit mention that TPS2 must be set after the link training sequence (and there isn't a similar requirement specified for the 8b/10b link training), see the Bspec links below. v2: Add bspec aspect/link to the commit log. (Jani) Bspec: 54128, 65448, 68849 Cc: stable@vger.kernel.org # v5.18+ Cc: Jani Nikula Signed-off-by: Imre Deak Acked-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20250217223828.1166093-2-imre.deak@intel.com Signed-off-by: Rodrigo Vivi (cherry picked from commit 8b4bbaf8ddc1f68f3ee96a706f65fdb1bcd9d355) Signed-off-by: Rodrigo Vivi --- .../gpu/drm/i915/display/intel_dp_link_training.c | 15 ++++++++++++++- 1 file changed, 14 insertions(+), 1 deletion(-) diff --git a/drivers/gpu/drm/i915/display/intel_dp_link_training.c b/drivers/gpu/drm/i915/display/intel_dp_link_training.c index 8b1977cfec50..6696a32cdd3e 100644 --- a/drivers/gpu/drm/i915/display/intel_dp_link_training.c +++ b/drivers/gpu/drm/i915/display/intel_dp_link_training.c @@ -1563,7 +1563,7 @@ intel_dp_128b132b_link_train(struct intel_dp *intel_dp, if (wait_for(intel_dp_128b132b_intra_hop(intel_dp, crtc_state) == 0, 500)) { lt_err(intel_dp, DP_PHY_DPRX, "128b/132b intra-hop not clear\n"); - return false; + goto out; } if (intel_dp_128b132b_lane_eq(intel_dp, crtc_state) && @@ -1575,6 +1575,19 @@ intel_dp_128b132b_link_train(struct intel_dp *intel_dp, passed ? "passed" : "failed", crtc_state->port_clock, crtc_state->lane_count); +out: + /* + * Ensure that the training pattern does get set to TPS2 even in case + * of a failure, as is the case at the end of a passing link training + * and what is expected by the transcoder. Leaving TPS1 set (and + * disabling the link train mode in DP_TP_CTL later from TPS1 directly) + * would result in a stuck transcoder HW state and flip-done timeouts + * later in the modeset sequence. + */ + if (!passed) + intel_dp_program_link_training_pattern(intel_dp, crtc_state, + DP_PHY_DPRX, DP_TRAINING_PATTERN_2); + return passed; } From 8058b49bf6fff777bf3f47309c7b15dbef2191af Mon Sep 17 00:00:00 2001 From: Imre Deak Date: Tue, 18 Feb 2025 00:38:28 +0200 Subject: [PATCH 6/6] drm/i915/dp: Fix disabling the transcoder function in 128b/132b mode During disabling the transcoder in DP 128b/132b mode (both in case of an MST master transcoder and in case of SST) the transcoder function must be first disabled without changing any other field in the register (in particular leaving the DDI port and mode select fields unchanged) and clearing the DDI port and mode select fields separately, later during the disabling sequences. Fix the sequence accordingly. Bspec: 54128, 65448, 68849 Cc: Jani Nikula Fixes: 79a6734cd56e ("drm/i915/ddi: disable trancoder port select for 128b/132b SST") Signed-off-by: Imre Deak Reviewed-by: Jani Nikula Link: https://patchwork.freedesktop.org/patch/msgid/20250217223828.1166093-3-imre.deak@intel.com Signed-off-by: Rodrigo Vivi (cherry picked from commit 2ed653c7b843db0670136330480842d76cb65cd8) Signed-off-by: Rodrigo Vivi --- drivers/gpu/drm/i915/display/intel_ddi.c | 6 ++---- 1 file changed, 2 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/display/intel_ddi.c b/drivers/gpu/drm/i915/display/intel_ddi.c index 2b9240ab547d..18c66992aa1d 100644 --- a/drivers/gpu/drm/i915/display/intel_ddi.c +++ b/drivers/gpu/drm/i915/display/intel_ddi.c @@ -658,7 +658,6 @@ void intel_ddi_disable_transcoder_func(const struct intel_crtc_state *crtc_state struct intel_crtc *crtc = to_intel_crtc(crtc_state->uapi.crtc); struct drm_i915_private *dev_priv = to_i915(crtc->base.dev); enum transcoder cpu_transcoder = crtc_state->cpu_transcoder; - bool is_mst = intel_crtc_has_type(crtc_state, INTEL_OUTPUT_DP_MST); u32 ctl; if (DISPLAY_VER(dev_priv) >= 11) @@ -678,8 +677,7 @@ void intel_ddi_disable_transcoder_func(const struct intel_crtc_state *crtc_state TRANS_DDI_PORT_SYNC_MASTER_SELECT_MASK); if (DISPLAY_VER(dev_priv) >= 12) { - if (!intel_dp_mst_is_master_trans(crtc_state) || - (!is_mst && intel_dp_is_uhbr(crtc_state))) { + if (!intel_dp_mst_is_master_trans(crtc_state)) { ctl &= ~(TGL_TRANS_DDI_PORT_MASK | TRANS_DDI_MODE_SELECT_MASK); } @@ -3134,7 +3132,7 @@ static void intel_ddi_post_disable_dp(struct intel_atomic_state *state, intel_dp_set_power(intel_dp, DP_SET_POWER_D3); if (DISPLAY_VER(dev_priv) >= 12) { - if (is_mst) { + if (is_mst || intel_dp_is_uhbr(old_crtc_state)) { enum transcoder cpu_transcoder = old_crtc_state->cpu_transcoder; intel_de_rmw(dev_priv,