Re: [PATCH v4] drm/i915: Fix assert_plane() warning on bootup with external display

["Ville Syrjälä" <ville.syrjala@linux.intel.com>]

On Wed, Jun 27, 2018 at 10:39:56AM -0700, Radhakrishna Sripada wrote:
> On Tue, Jun 26, 2018 at 05:55:59PM -0700, Azhar Shaikh wrote:
> > On KBL, WHL RVPs, booting up with an external display connected, triggers
> > below warning, when the BiOS brings up the external display too.
> > This warning is not seen during hotplug.
> > 
> > [    3.615226] ------------[ cut here ]------------
> > [    3.619829] plane 1A assertion failure (expected on, current off)
> > [    3.632039] WARNING: CPU: 2 PID: 354 at drivers/gpu/drm/i915/intel_display.c:1294 assert_plane+0x71/0xbb
> > [    3.633920] iwlwifi 0000:00:14.3: loaded firmware version 38.c0e03d94.0 op_mode iwlmvm
> > [    3.647157] Modules linked in: iwlwifi cfg80211 btusb btrtl btbcm btintel bluetooth ecdh_generic
> > [    3.647163] CPU: 2 PID: 354 Comm: frecon Not tainted 4.17.0-rc7-50176-g655af12d39c2 #3
> > [    3.647165] Hardware name: Intel Corporation CoffeeLake Client Platform/WhiskeyLake U DDR4 ERB, BIOS CNLSFWR1.R00.X140.B00.1804040304 04/04/2018
> > [    3.684509] RIP: 0010:assert_plane+0x71/0xbb
> > [    3.764451] Call Trace:
> > [    3.766888]  intel_atomic_commit_tail+0xa97/0xb77
> > [    3.771569]  intel_atomic_commit+0x26a/0x279
> > [    3.771572]  drm_atomic_helper_set_config+0x5c/0x76
> > [    3.780670]  __drm_mode_set_config_internal+0x66/0x109
> > [    3.780672]  drm_mode_setcrtc+0x4c9/0x5cc
> > [    3.780674]  ? drm_mode_getcrtc+0x162/0x162
> > [    3.789774]  ? drm_mode_getcrtc+0x162/0x162
> > [    3.798108]  drm_ioctl_kernel+0x8d/0xe4
> > [    3.801926]  drm_ioctl+0x27d/0x368
> > [    3.805311]  ? drm_mode_getcrtc+0x162/0x162
> > [    3.805314]  ? selinux_file_ioctl+0x14e/0x199
> > [    3.805317]  vfs_ioctl+0x21/0x2f
> > [    3.813812]  do_vfs_ioctl+0x491/0x4b4
> > [    3.813813]  ? security_file_ioctl+0x37/0x4b
> > [    3.813816]  ksys_ioctl+0x55/0x75
> > [    3.820672]  __x64_sys_ioctl+0x1a/0x1e
> > [    3.820674]  do_syscall_64+0x51/0x5f
> > [    3.820678]  entry_SYSCALL_64_after_hwframe+0x44/0xa9
> > [    3.828221] RIP: 0033:0x7b5e04953967
> > [    3.835504] RSP: 002b:00007fff2eafb6f8 EFLAGS: 00000246 ORIG_RAX: 0000000000000010
> > [    3.835505] RAX: ffffffffffffffda RBX: 0000000000000002 RCX: 00007b5e04953967
> > [    3.835505] RDX: 00007fff2eafb730 RSI: 00000000c06864a2 RDI: 000000000000000f
> > [    3.835506] RBP: 00007fff2eafb720 R08: 0000000000000000 R09: 0000000000000000
> > [    3.835507] R10: 0000000000000070 R11: 0000000000000246 R12: 000000000000000f
> > [    3.879988] R13: 000056bc9dd7d210 R14: 00007fff2eafb730 R15: 00000000c06864a2
> > [    3.887081] Code: 48 c7 c7 06 71 a5 be 84 c0 48 c7 c2 06 fd a3 be 48 89 f9 48 0f 44 ca 84 db 48 0f 45 d7 48 c7 c7 df d3 a4 be 31 c0 e8 af a0 c0 ff <0f> 0b eb 2b 48 c7 c7 06 fd a3 be 84 c0 48 c7 c2 06 71 a5 be 48
> > [    3.905845] WARNING: CPU: 2 PID: 354 at drivers/gpu/drm/i915/intel_display.c:1294 assert_plane+0x71/0xbb
> > [    3.920964] ---[ end trace dac692f4ac46391a ]---
> > 
> > The warning is seen when mode_setcrtc() is called for pipeB
> > during bootup and before we get a mode_setcrtc() for pipeA,
> > while doing update_crtcs() in intel_atomic_commit_tail().
> > Now since, plane1A is still active after commit, update_crtcs()
> > is done for pipeA and eventually update_plane() for plane1A.
> > 
> > intel_plane_state->ctl for plane1A is not updated since set_modecrtc() is
> > called for pipeB. So intel_plane_state->ctl for plane 1A will be 0x0.
> > So doing an update_plane() for plane1A, will result in clearing
> > PLANE_CTL_ENABLE bit, and hence the warning.
> > 
> > To fix this warning, force all active planes to recompute their states
> > in probe.
> > 
> > Signed-off-by: Azhar Shaikh <azhar.shaikh@intel.com>
> > ---
> > Changes in v4:
> > - Handle locking in intel_initial_commit()
> > - Move the for loop inside intel_initial_commit() so that
> >   drm_atomic_commit() is called only once
> > - Call intel_initial_commit() only for more than one active crtc on boot.
> > - Save the return value of intel_initial_commit() and print a message in
> >   case of an error
> > 
> > Changes in v3:
> > - Add comments
> > 
> > Changes in v2:
> > - Force all planes to recompute their states.(Ville Syrjälä)
> > - Update the commit message
> 
> Include the changelog in the commit message above Signed-off-by
> > 
> >  drivers/gpu/drm/i915/intel_display.c | 81 +++++++++++++++++++++++++++++++++++-
> >  1 file changed, 79 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_display.c b/drivers/gpu/drm/i915/intel_display.c
> > index 3709fa1b6318..40bdb28aa2a5 100644
> > --- a/drivers/gpu/drm/i915/intel_display.c
> > +++ b/drivers/gpu/drm/i915/intel_display.c
> > @@ -15092,12 +15092,76 @@ static void intel_update_fdi_pll_freq(struct drm_i915_private *dev_priv)
> >  	DRM_DEBUG_DRIVER("FDI PLL freq=%d\n", dev_priv->fdi_pll_freq);
> >  }
> >  
> > +static int intel_initial_commit(struct drm_device *dev)
> > +{
> > +	struct drm_atomic_state *state = NULL;
> > +	struct drm_modeset_acquire_ctx ctx;
> > +	struct drm_crtc_state *crtc_state;
> > +	struct intel_crtc *intel_crtc;
> > +	int ret = 0;
> > +
> > +	drm_modeset_acquire_init(&ctx, 0);
> > +
> > +retry:
> > +	state = drm_atomic_state_alloc(dev);
> > +	if (!state)
> > +		goto unlock;
> Locks are not acquired. It should not call unlock. Use another goto to call 
> drm_modeset_acquire_fini

Easier to just do the state_alloc() before drm_modeset_acquire_init().
Then you can just return directly if the alloc fails.

> 
> > +
> > +	state->acquire_ctx = &ctx;
> > +
> > +	ret = drm_modeset_lock_all_ctx(dev, &ctx);

lock_all() not required as everything called later will grab the
required locks on its own. I guess with lock_all() you could simplify
the backoff thing to a simple loop, but that would only work if there
are no futher driver private modeset locks taken later. We don't have
any currently so this is not a problem. And we do have some lock_all()
uses elsewhere which wouldn't be quite correct either.

But probably clearner to not use lock_all() at all and avoid these
headaches.

> > +	if (ret == -EDEADLK) {
> > +		drm_atomic_state_clear(state);
> > +		drm_modeset_backoff(&ctx);
> > +		goto retry;
> > +	} else if (WARN_ON(ret)) {
> > +		goto unlock;
> Same as above. Also call drm_atomic_state_put to free the allocated state.
> 
> > +	}
> > +
> > +	for_each_intel_crtc(dev, intel_crtc) {
> > +		struct drm_crtc *crtc = &intel_crtc->base;
> > +
> > +		crtc_state = drm_atomic_get_crtc_state(state, crtc);
> > +		if (IS_ERR(crtc_state)) {
> > +			ret = PTR_ERR(crtc_state);
> > +			goto out;
> > +		}
> > +
> > +		if (crtc_state->active) {
> > +			ret = drm_atomic_add_affected_planes(state, crtc);
> > +			if (ret)
> > +				goto out;
> > +		}
> > +	}
> > +
> > +	ret = drm_atomic_commit(state);
> > +	if (ret == -EDEADLK) {
> > +		drm_atomic_state_clear(state);
> > +		drm_modeset_backoff(&ctx);
> > +		goto retry;
> > +	}
> > +
> > +out:
> > +	if (state) {
> > +		drm_atomic_state_put(state);
> > +		state = NULL;
> > +	}
> > +
> > +unlock:
> > +	drm_modeset_drop_locks(&ctx);
> > +	drm_modeset_acquire_fini(&ctx);
> 
> Re-order freeing the objects in this case drm_modeset_drop_locks followed by 
> drm_atomic_state_put and drm_modeset_acquire_fini
> 
> Regards,
> Radhakrishna(RK) Sripada
> > +
> > +	return ret;
> > +}
> > +
> >  int intel_modeset_init(struct drm_device *dev)
> >  {
> >  	struct drm_i915_private *dev_priv = to_i915(dev);
> >  	struct i915_ggtt *ggtt = &dev_priv->ggtt;
> >  	enum pipe pipe;
> >  	struct intel_crtc *crtc;
> > +	u16 num_active_crtcs;
> > +	int ret;
> >  
> >  	dev_priv->modeset_wq = alloc_ordered_workqueue("i915_modeset", 0);
> >  
> > @@ -15172,8 +15236,6 @@ int intel_modeset_init(struct drm_device *dev)
> >  		      INTEL_INFO(dev_priv)->num_pipes > 1 ? "s" : "");
> >  
> >  	for_each_pipe(dev_priv, pipe) {
> > -		int ret;
> > -
> >  		ret = intel_crtc_init(dev_priv, pipe);
> >  		if (ret) {
> >  			drm_mode_config_cleanup(dev);
> > @@ -15203,6 +15265,8 @@ int intel_modeset_init(struct drm_device *dev)
> >  
> >  		if (!crtc->active)
> >  			continue;
> > +		else
> > +			num_active_crtcs++;
> >  
> >  		/*
> >  		 * Note that reserving the BIOS fb up front prevents us
> > @@ -15222,6 +15286,19 @@ int intel_modeset_init(struct drm_device *dev)
> >  	}
> >  
> >  	/*
> > +	 * Only in case of more than one active crtc in probe, force
> > +	 * all active planes to recompute their states. So that on
> > +	 * mode_setcrtc after probe, all the intel_plane_state variables
> > +	 * are already calculated and there is no assert_plane warnings
> > +	 * during bootup.
> > +	 */
> > +	if (num_active_crtcs > 1) {
> > +		ret = intel_initial_commit(dev);
> > +		if (ret)
> > +			DRM_DEBUG_KMS("Initial commit in probe failed.\n");
> > +	}
> > +
> > +	/*
> >  	 * Make sure hardware watermarks really match the state we read out.
> >  	 * Note that we need to do this after reconstructing the BIOS fb's
> >  	 * since the watermark calculation done here will use pstate->fb.
> > -- 
> > 1.9.1
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > https://lists.freedesktop.org/mailman/listinfo/intel-gfx
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> https://lists.freedesktop.org/mailman/listinfo/intel-gfx

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