Re: [alsa-devel] Timing issues between ALSA and i915 drivers

On Thu, 17 Jan 2019 20:53:06 +0100, Pierre-Louis Bossart wrote:

...

I could use some feedback on HDMI audio issues exposed during the 4.21 merge window. By accident (misleading documentation) we ended up enabling the Skylake driver instead of the HDaudio legacy, and broke audio on a number of Skylake and ApolloLake devices where the HDMI/iDISP codec was not detected (bit 2 not set in the codec_mask). Linus' Dell XPS13 9350 was the first to be impacted of course...

After debugging a bit, this issue can be resolved by either

a) compiling both SOUND and DRM as built-ins (y instead of m)

b) moving the calls snd_hdac_i915_init() to the probe function instead of the worker queue (code at https://github.com/plbossart/sound/commits/fix/skl-hdmi)

Both solutions point to timing issues.

During internal reviews I was alerted to the fact that the suggested fix essentially reverts patch ab1b732d53c18 ('ASoC: Intel: Skylake: Move i915 registration to worker thread') which was introduced to solve DRM lockup issues.

I tried to narrow down the issue further and my current understanding is that the Skylake driver performs link reset operations without the display power turned on - which does not look like a very smart thing to do in hindsight.

In other words, it's not really when snd_hdac_i915_init() is called that matters as I assumed initially, but more when snd_hdac_display_power() is invoked. There are two cases where this happens, and for each of them turning the display power on results in HDMI detection. The attached diffs split the initialization from the power on, which provides a better understanding of the issue.

OK, this makes some sense, and that's the very reason we have HDA_CODEC_IDX_CONTROLLER for snd_hdac_display_power(). IIRC, we needed to power on the display for probing of the legacy HDA, too. Once after that, for the normal operation, the display power is needed only when you output the HDMI stream.

What would be really useful at this point is a confirmation that snd_hdac_i915_init() cannot be called in the initial probe but does need to be executed in a work queue. That would really impact the way the initialization sequence is reworked on the Skylake side as well as modify the way the SOF driver deals with i915 initialization.

It's needed to be called in a work queue, yes.

Basically you shouldn't call request_module() in the driver's probe callback. When the probe callback is called from the module loading, it blocks the module loading itself, hence loading yet another module can't work. A situation might be easier than the past (which deadlocked), but still it's advised to use either the request_module_nowait() with the callback or call request_module() asynchronously from probe.

thanks,

Takashi

Thanks

-Pierre

diff --git a/sound/soc/intel/skylake/skl.c b/sound/soc/intel/skylake/skl.c index 60c94836bf5b..56556d06a17f 100644 --- a/sound/soc/intel/skylake/skl.c +++ b/sound/soc/intel/skylake/skl.c @@ -767,23 +767,6 @@ static const struct hdac_bus_ops bus_core_ops = { .get_response = snd_hdac_bus_get_response, };

-static int skl_i915_init(struct hdac_bus *bus) -{

• int err;
• /*

• * The HDMI codec is in GPU so we need to ensure that it is powered
  

• * up and ready for probe
  

• */
  

• err = snd_hdac_i915_init(bus);
• if (err < 0)

• return err;
  

• snd_hdac_display_power(bus, HDA_CODEC_IDX_CONTROLLER, true);
• return 0;

-}

static void skl_probe_work(struct work_struct *work) { struct skl *skl = container_of(work, struct skl, probe_work); @@ -791,12 +774,6 @@ static void skl_probe_work(struct work_struct *work) struct hdac_ext_link *hlink = NULL; int err;

• if (IS_ENABLED(CONFIG_SND_SOC_HDAC_HDMI)) {

• err = skl_i915_init(bus);
  

• if (err < 0)
  

• 	return;
  

• }
• err = skl_init_chip(bus, true); if (err < 0) { dev_err(bus->dev, "Init chip failed with err: %d\n", err);

@@ -899,6 +876,11 @@ static int skl_first_init(struct hdac_bus *bus) unsigned short gcap; int cp_streams, pb_streams, start_idx;

• err = snd_hdac_i915_init(bus);
• if (err < 0)

• return err;
  

• err = pci_request_regions(pci, "Skylake HD audio"); if (err < 0) return err;

@@ -910,7 +892,10 @@ static int skl_first_init(struct hdac_bus *bus) return -ENXIO; }

• snd_hdac_bus_reset_link(bus, true);

• /* this bus_reset_link is unnecessary, and without the display

• *  power turned on prevents HDMI from being detected
  

• */
  

• //snd_hdac_bus_reset_link(bus, true);

  snd_hdac_bus_parse_capabilities(bus);

@@ -962,7 +947,14 @@ static int skl_first_init(struct hdac_bus *bus) /* initialize chip */ skl_init_pci(skl);

• return skl_init_chip(bus, true);

• /* turning the display power here works */
• snd_hdac_display_power(bus, HDA_CODEC_IDX_CONTROLLER, true);
• err = skl_init_chip(bus, true);
• /* turning the display power here does not work, HDMI not detected */
• return err;

}

static int skl_probe(struct pci_dev *pci,