We don't actually check for INTEL_OUTPUT_DP_MST at all in here, as a result we skip assigning a DPLL to any DP MST ports, which makes link training fail, after which things just keep going downhill from there.
Consequently, this fixes DisplayPort MST causing kernel panics, machine check errors, etc. on Skylake.
Signed-off-by: Lyude cpaul@redhat.com --- drivers/gpu/drm/i915/intel_ddi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index e6408e5..54a165b 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -1589,7 +1589,8 @@ skl_ddi_pll_select(struct intel_crtc *intel_crtc, DPLL_CFGCR2_KDIV(wrpll_params.kdiv) | DPLL_CFGCR2_PDIV(wrpll_params.pdiv) | wrpll_params.central_freq; - } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT) { + } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT || + intel_encoder->type == INTEL_OUTPUT_DP_MST) { switch (crtc_state->port_clock / 2) { case 81000: ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, 0);
Assuming any connector that isn't DP, MST, or HDMI is eDP definitely seems likely to cover up other bugs in the future.
Signed-off-by: Lyude cpaul@redhat.com --- drivers/gpu/drm/i915/intel_ddi.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index 54a165b..e2adc380 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -1604,8 +1604,10 @@ skl_ddi_pll_select(struct intel_crtc *intel_crtc, }
cfgcr1 = cfgcr2 = 0; - } else /* eDP */ + } else if (intel_encoder->type == INTEL_OUTPUT_EDP) { return true; + } else + return false;
memset(&crtc_state->dpll_hw_state, 0, sizeof(crtc_state->dpll_hw_state));
On Tue, 02 Feb 2016, Lyude cpaul@redhat.com wrote:
We don't actually check for INTEL_OUTPUT_DP_MST at all in here, as a result we skip assigning a DPLL to any DP MST ports, which makes link training fail, after which things just keep going downhill from there.
Apparently broken since
commit 82d35437018950c25200051733fb2fbf63cca86c Author: Satheeshakrishna M satheeshakrishna.m@intel.com Date: Thu Nov 13 14:55:20 2014 +0000
drm/i915/skl: Implementation of SKL DPLL programming
Consequently, this fixes DisplayPort MST causing kernel panics, machine check errors, etc. on Skylake.
Please copy-paste some traces for the commit message so it'll be easier to match bug reports and fixes later on.
BR, Jani.
Signed-off-by: Lyude cpaul@redhat.com
drivers/gpu/drm/i915/intel_ddi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index e6408e5..54a165b 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -1589,7 +1589,8 @@ skl_ddi_pll_select(struct intel_crtc *intel_crtc, DPLL_CFGCR2_KDIV(wrpll_params.kdiv) | DPLL_CFGCR2_PDIV(wrpll_params.pdiv) | wrpll_params.central_freq;
- } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT) {
- } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
intel_encoder->type == INTEL_OUTPUT_DP_MST) {
We don't actually check for INTEL_OUTPUT_DP_MST at all in here, as a result we skip assigning a DPLL to any DP MST ports, which makes link training fail:
[ 1442.933896] [drm:intel_power_well_enable] enabling DDI D power well [ 1442.933905] [drm:skl_set_power_well] Enabling DDI D power well [ 1442.933957] [drm:intel_mst_pre_enable_dp] 0 [ 1442.935474] [drm:intel_dp_set_signal_levels] Using signal levels 00000000 [ 1442.935477] [drm:intel_dp_set_signal_levels] Using vswing level 0 [ 1442.935480] [drm:intel_dp_set_signal_levels] Using pre-emphasis level 0 [ 1442.936190] [drm:intel_dp_set_signal_levels] Using signal levels 05000000 [ 1442.936193] [drm:intel_dp_set_signal_levels] Using vswing level 1 [ 1442.936195] [drm:intel_dp_set_signal_levels] Using pre-emphasis level 1 [ 1442.936858] [drm:intel_dp_set_signal_levels] Using signal levels 08000000 [ 1442.936862] [drm:intel_dp_set_signal_levels] Using vswing level 2 … [ 1442.998253] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up [ 1442.998512] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to train DP, aborting
After which the pipe state goes completely out of sync:
[ 70.075596] [drm:check_crtc_state] [CRTC:25] [ 70.075696] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in ddi_pll_sel (expected 0x00000000, found 0x00000001) [ 70.075747] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in shared_dpll (expected -1, found 0) [ 70.075798] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in dpll_hw_state.ctrl1 (expected 0x00000000, found 0x00000021) [ 70.075840] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in dpll_hw_state.cfgcr1 (expected 0x00000000, found 0x80400173) [ 70.075884] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in dpll_hw_state.cfgcr2 (expected 0x00000000, found 0x000003a5) [ 70.075954] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in base.adjusted_mode.crtc_clock (expected 262750, found 72256) [ 70.075999] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in port_clock (expected 540000, found 148500)
And if you're especially lucky, it keeps going downhill:
[ 83.309256] Kernel panic - not syncing: Timeout: Not all CPUs entered broadcast exception handler [ 83.309265] [ 83.309265] ================================= [ 83.309266] [ INFO: inconsistent lock state ] [ 83.309267] 4.5.0-rc1Lyude-Test #265 Not tainted [ 83.309267] --------------------------------- [ 83.309268] inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. [ 83.309270] Xorg/1194 [HC0[1]:SC0[0]:HE1:SE1] takes: [ 83.309293] (&(&dev_priv->uncore.lock)->rlock){?.-...}, at: [<ffffffffa02a6073>] gen9_write32+0x63/0x400 [i915] [ 83.309293] {IN-HARDIRQ-W} state was registered at: [ 83.309297] [<ffffffff810e84f4>] __lock_acquire+0x9c4/0x1d00 [ 83.309299] [<ffffffff810ea1be>] lock_acquire+0xce/0x1c0 [ 83.309302] [<ffffffff8177d936>] _raw_spin_lock_irqsave+0x56/0x90 [ 83.309321] [<ffffffffa02a5492>] gen9_read32+0x52/0x3d0 [i915] [ 83.309332] [<ffffffffa024beea>] gen8_irq_handler+0x27a/0x6a0 [i915] [ 83.309337] [<ffffffff810fdbc1>] handle_irq_event_percpu+0x41/0x300 [ 83.309339] [<ffffffff810fdeb9>] handle_irq_event+0x39/0x60 [ 83.309341] [<ffffffff811010b4>] handle_edge_irq+0x74/0x130 [ 83.309344] [<ffffffff81009073>] handle_irq+0x73/0x120 [ 83.309346] [<ffffffff817805f1>] do_IRQ+0x61/0x120 [ 83.309348] [<ffffffff8177e6d6>] ret_from_intr+0x0/0x20 [ 83.309351] [<ffffffff815f5105>] cpuidle_enter_state+0x105/0x330 [ 83.309353] [<ffffffff815f5367>] cpuidle_enter+0x17/0x20 [ 83.309356] [<ffffffff810dbe1a>] call_cpuidle+0x2a/0x50 [ 83.309358] [<ffffffff810dc1dd>] cpu_startup_entry+0x26d/0x3a0 [ 83.309360] [<ffffffff817701da>] rest_init+0x13a/0x140 [ 83.309363] [<ffffffff81f2af8e>] start_kernel+0x475/0x482 [ 83.309365] [<ffffffff81f2a315>] x86_64_start_reservations+0x2a/0x2c [ 83.309367] [<ffffffff81f2a452>] x86_64_start_kernel+0x13b/0x14a
Fixes: 82d354370189 ("drm/i915/skl: Implementation of SKL DPLL programming") Signed-off-by: Lyude cpaul@redhat.com --- Changes * Add "Fixes" line with link to commit that introduced this problem * Add some stack traces to aid in future debugging and fixes
drivers/gpu/drm/i915/intel_ddi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index e6408e5..54a165b 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -1589,7 +1589,8 @@ skl_ddi_pll_select(struct intel_crtc *intel_crtc, DPLL_CFGCR2_KDIV(wrpll_params.kdiv) | DPLL_CFGCR2_PDIV(wrpll_params.pdiv) | wrpll_params.central_freq; - } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT) { + } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT || + intel_encoder->type == INTEL_OUTPUT_DP_MST) { switch (crtc_state->port_clock / 2) { case 81000: ctrl1 |= DPLL_CTRL1_LINK_RATE(DPLL_CTRL1_LINK_RATE_810, 0);
On Tue, Feb 02, 2016 at 10:49:43AM -0500, Lyude wrote:
We don't actually check for INTEL_OUTPUT_DP_MST at all in here, as a result we skip assigning a DPLL to any DP MST ports, which makes link training fail:
[ 1442.933896] [drm:intel_power_well_enable] enabling DDI D power well [ 1442.933905] [drm:skl_set_power_well] Enabling DDI D power well [ 1442.933957] [drm:intel_mst_pre_enable_dp] 0 [ 1442.935474] [drm:intel_dp_set_signal_levels] Using signal levels 00000000 [ 1442.935477] [drm:intel_dp_set_signal_levels] Using vswing level 0 [ 1442.935480] [drm:intel_dp_set_signal_levels] Using pre-emphasis level 0 [ 1442.936190] [drm:intel_dp_set_signal_levels] Using signal levels 05000000 [ 1442.936193] [drm:intel_dp_set_signal_levels] Using vswing level 1 [ 1442.936195] [drm:intel_dp_set_signal_levels] Using pre-emphasis level 1 [ 1442.936858] [drm:intel_dp_set_signal_levels] Using signal levels 08000000 [ 1442.936862] [drm:intel_dp_set_signal_levels] Using vswing level 2 … [ 1442.998253] [drm:intel_dp_link_training_clock_recovery [i915]] *ERROR* too many full retries, give up [ 1442.998512] [drm:intel_dp_start_link_train [i915]] *ERROR* failed to train DP, aborting
After which the pipe state goes completely out of sync:
[ 70.075596] [drm:check_crtc_state] [CRTC:25] [ 70.075696] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in ddi_pll_sel (expected 0x00000000, found 0x00000001) [ 70.075747] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in shared_dpll (expected -1, found 0) [ 70.075798] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in dpll_hw_state.ctrl1 (expected 0x00000000, found 0x00000021) [ 70.075840] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in dpll_hw_state.cfgcr1 (expected 0x00000000, found 0x80400173) [ 70.075884] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in dpll_hw_state.cfgcr2 (expected 0x00000000, found 0x000003a5) [ 70.075954] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in base.adjusted_mode.crtc_clock (expected 262750, found 72256) [ 70.075999] [drm:intel_pipe_config_compare [i915]] *ERROR* mismatch in port_clock (expected 540000, found 148500)
And if you're especially lucky, it keeps going downhill:
[ 83.309256] Kernel panic - not syncing: Timeout: Not all CPUs entered broadcast exception handler [ 83.309265] [ 83.309265] ================================= [ 83.309266] [ INFO: inconsistent lock state ] [ 83.309267] 4.5.0-rc1Lyude-Test #265 Not tainted [ 83.309267] --------------------------------- [ 83.309268] inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage. [ 83.309270] Xorg/1194 [HC0[1]:SC0[0]:HE1:SE1] takes: [ 83.309293] (&(&dev_priv->uncore.lock)->rlock){?.-...}, at: [<ffffffffa02a6073>] gen9_write32+0x63/0x400 [i915] [ 83.309293] {IN-HARDIRQ-W} state was registered at: [ 83.309297] [<ffffffff810e84f4>] __lock_acquire+0x9c4/0x1d00 [ 83.309299] [<ffffffff810ea1be>] lock_acquire+0xce/0x1c0 [ 83.309302] [<ffffffff8177d936>] _raw_spin_lock_irqsave+0x56/0x90 [ 83.309321] [<ffffffffa02a5492>] gen9_read32+0x52/0x3d0 [i915] [ 83.309332] [<ffffffffa024beea>] gen8_irq_handler+0x27a/0x6a0 [i915] [ 83.309337] [<ffffffff810fdbc1>] handle_irq_event_percpu+0x41/0x300 [ 83.309339] [<ffffffff810fdeb9>] handle_irq_event+0x39/0x60 [ 83.309341] [<ffffffff811010b4>] handle_edge_irq+0x74/0x130 [ 83.309344] [<ffffffff81009073>] handle_irq+0x73/0x120 [ 83.309346] [<ffffffff817805f1>] do_IRQ+0x61/0x120 [ 83.309348] [<ffffffff8177e6d6>] ret_from_intr+0x0/0x20 [ 83.309351] [<ffffffff815f5105>] cpuidle_enter_state+0x105/0x330 [ 83.309353] [<ffffffff815f5367>] cpuidle_enter+0x17/0x20 [ 83.309356] [<ffffffff810dbe1a>] call_cpuidle+0x2a/0x50 [ 83.309358] [<ffffffff810dc1dd>] cpu_startup_entry+0x26d/0x3a0 [ 83.309360] [<ffffffff817701da>] rest_init+0x13a/0x140 [ 83.309363] [<ffffffff81f2af8e>] start_kernel+0x475/0x482 [ 83.309365] [<ffffffff81f2a315>] x86_64_start_reservations+0x2a/0x2c [ 83.309367] [<ffffffff81f2a452>] x86_64_start_kernel+0x13b/0x14a
Fixes: 82d354370189 ("drm/i915/skl: Implementation of SKL DPLL programming") Signed-off-by: Lyude cpaul@redhat.com
Somehow CI is ignoring your patch, but then we don't have any dp mst skl machines in CI anyway yet (would be obvious we'd have noticed this too). So applied both, with cc: stable for patch 1.
Thanks, Daniel
Changes
- Add "Fixes" line with link to commit that introduced this problem
- Add some stack traces to aid in future debugging and fixes
drivers/gpu/drm/i915/intel_ddi.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/i915/intel_ddi.c b/drivers/gpu/drm/i915/intel_ddi.c index e6408e5..54a165b 100644 --- a/drivers/gpu/drm/i915/intel_ddi.c +++ b/drivers/gpu/drm/i915/intel_ddi.c @@ -1589,7 +1589,8 @@ skl_ddi_pll_select(struct intel_crtc *intel_crtc, DPLL_CFGCR2_KDIV(wrpll_params.kdiv) | DPLL_CFGCR2_PDIV(wrpll_params.pdiv) | wrpll_params.central_freq;
- } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT) {
- } else if (intel_encoder->type == INTEL_OUTPUT_DISPLAYPORT ||
intel_encoder->type == INTEL_OUTPUT_DP_MST) {-- 2.5.0
Intel-gfx mailing list Intel-gfx@lists.freedesktop.org http://lists.freedesktop.org/mailman/listinfo/intel-gfx
dri-devel@lists.freedesktop.org
-
Daniel Vetter -
Jani Nikula -
Lyude