HDaudio probe: need to introduce workqueue to avoid deadlocks.

[plbossart]

Timing issues reported on Skylake expose a conceptual issue in the way the HDaudio probe is handled.

Currently the probe flow is

hda_dsp_probe
-- hda_init_caps
---- hda_codec_i915_init
------ snd_hdac_i915_init
-------- request_module("i915");
---- hda_codec_probe_bus
------ hda_codec_probe
-------- request_module("Hdaudio xyz")

As explained by Takashi Iwai on the alsa-devel mailing list:

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.

The request is to move init_caps to a work queue, and cancel it on remove.

Note that this will impact how hardware initialization issues are handled. We already know that errors on driver_probe happening when platform_device_register_data() is called will not be propagated, and the probe might actually be executed later. with a workqueue errors will be even harder to be made available to higher level PCI drivers.

[plbossart]

fixed