[Intel-gfx] [CI] drm/i915/audio: Skip the cdclk modeset if no pipes attached

[Intel-gfx] [CI] drm/i915/audio: Skip the cdclk modeset if no pipes attached

[Chris Wilson]

If the display is not driving any pipes, we cannot change the bclk and
doing so risks chasing NULL pointers:

<6> [278.907105] snd_hda_intel 0000:00:0e.0: DSP detected with PCI class/subclass/prog-if info 0x040100
<6> [278.909936] snd_hda_intel 0000:00:0e.0: bound 0000:00:02.0 (ops i915_audio_component_bind_ops [i915])
<7> [278.910078] i915 0000:00:02.0: [drm:intel_power_well_enable [i915]] enabling power well 2
<1> [278.910228] BUG: kernel NULL pointer dereference, address: 0000000000000080
<1> [278.910243] #PF: supervisor read access in kernel mode
<1> [278.910251] #PF: error_code(0x0000) - not-present page
<6> [278.910260] PGD 0 P4D 0
<4> [278.910267] Oops: 0000 [#1] PREEMPT SMP PTI
<4> [278.910276] CPU: 0 PID: 5 Comm: kworker/0:0 Tainted: G     U            5.5.0-CI-CI_DRM_7853+ #1
<4> [278.910289] Hardware name: Intel Corp. Geminilake/GLK RVP2 LP4SD (07), BIOS GELKRVPA.X64.0062.B30.1708222146 08/22/2017
<4> [278.910312] Workqueue: events azx_probe_work [snd_hda_intel]
<4> [278.910327] RIP: 0010:__ww_mutex_lock.constprop.15+0x5e/0x1090
<4> [278.910338] Code: 75 88 be a7 03 00 00 65 48 8b 04 25 28 00 00 00 48 89 45 c8 31 c0 4c 89 c3 e8 5e b3 6d ff 44 8b 3d 2f 24 37 02 45 85 ff 75 0a <4d> 3b 6d 58 0f 85 3f 07 00 00 48 85 db 74 22 49 8b 95 80 00 00 00
<4> [278.910362] RSP: 0018:ffffc9000008bc10 EFLAGS: 00010246
<4> [278.910371] RAX: 0000000000000246 RBX: ffffc9000008bd30 RCX: 0000000000000001
<4> [278.910382] RDX: 0000000000000000 RSI: ffffffff82647c60 RDI: ffff88817b27d848
<4> [278.910393] RBP: ffffc9000008bcc0 R08: 0000000000000000 R09: 0000000000000001
<4> [278.910404] R10: ffffc9000008bce0 R11: 0000000000000000 R12: ffffffff8168f0fc
<4> [278.910414] R13: 0000000000000028 R14: ffffc9000008bd60 R15: 0000000000000000
<4> [278.910425] FS:  0000000000000000(0000) GS:ffff88817bc00000(0000) knlGS:0000000000000000
<4> [278.910437] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
<4> [278.910446] CR2: 0000000000000080 CR3: 00000001650da000 CR4: 0000000000340ef0
<4> [278.910456] Call Trace:
<4> [278.910468]  ? mark_held_locks+0x49/0x70
<4> [278.910479]  ? ww_mutex_lock+0x39/0x70
<4> [278.910487]  ww_mutex_lock+0x39/0x70
<4> [278.910497]  drm_modeset_lock+0x6c/0x120
<4> [278.910575]  glk_force_audio_cdclk+0x7d/0x140 [i915]
<4> [278.910656]  i915_audio_component_get_power+0xf2/0x110 [i915]
<4> [278.910673]  snd_hdac_display_power+0x7d/0x120 [snd_hda_core]
<4> [278.910686]  azx_probe_work+0x88/0x7e0 [snd_hda_intel]

v2: Reorder glk_force_audio_cdclk() parameters to fit in with the norm.

Closes: https://gitlab.freedesktop.org/drm/intel/issues/1095
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
 drivers/gpu/drm/i915/display/intel_audio.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/i915/display/intel_audio.c b/drivers/gpu/drm/i915/display/intel_audio.c
index 13833a7a8f27..30fb7c887ff0 100644
--- a/drivers/gpu/drm/i915/display/intel_audio.c
+++ b/drivers/gpu/drm/i915/display/intel_audio.c
@@ -811,15 +811,13 @@ void intel_init_audio_hooks(struct drm_i915_private *dev_priv)
 }
 
 static int glk_force_audio_cdclk_commit(struct intel_atomic_state *state,
+					struct intel_crtc *crtc,
 					bool enable)
 {
-	struct drm_i915_private *dev_priv = to_i915(state->base.dev);
 	struct intel_cdclk_state *cdclk_state;
-	struct intel_crtc *crtc;
 	int ret;
 
 	/* need to hold at least one crtc lock for the global state */
-	crtc = intel_get_crtc_for_pipe(dev_priv, PIPE_A);
 	ret = drm_modeset_lock(&crtc->base.mutex, state->base.acquire_ctx);
 	if (ret)
 		return ret;
@@ -843,8 +841,13 @@ static void glk_force_audio_cdclk(struct drm_i915_private *dev_priv,
 {
 	struct drm_modeset_acquire_ctx ctx;
 	struct drm_atomic_state *state;
+	struct intel_crtc *crtc;
 	int ret;
 
+	crtc = intel_get_crtc_for_pipe(dev_priv, PIPE_A);
+	if (!crtc)
+		return;
+
 	drm_modeset_acquire_init(&ctx, 0);
 	state = drm_atomic_state_alloc(&dev_priv->drm);
 	if (drm_WARN_ON(&dev_priv->drm, !state))
@@ -853,7 +856,8 @@ static void glk_force_audio_cdclk(struct drm_i915_private *dev_priv,
 	state->acquire_ctx = &ctx;
 
 retry:
-	ret = glk_force_audio_cdclk_commit(to_intel_atomic_state(state), enable);
+	ret = glk_force_audio_cdclk_commit(to_intel_atomic_state(state), crtc,
+					   enable);
 	if (ret == -EDEADLK) {
 		drm_atomic_state_clear(state);
 		drm_modeset_backoff(&ctx);
-- 
2.25.0

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx

[Intel-gfx] ✗ Fi.CI.BAT: failure for drm/i915/audio: Skip the cdclk modeset if no pipes attached (rev2)

[Patchwork]

== Series Details ==

Series: drm/i915/audio: Skip the cdclk modeset if no pipes attached (rev2)
URL   : https://patchwork.freedesktop.org/series/72863/
State : failure

== Summary ==

Applying: drm/i915/audio: Skip the cdclk modeset if no pipes attached
Using index info to reconstruct a base tree...
M	drivers/gpu/drm/i915/display/intel_audio.c
Falling back to patching base and 3-way merge...
No changes -- Patch already applied.

_______________________________________________
Intel-gfx mailing list
Intel-gfx@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/intel-gfx