On Wed, 06 Jul 2022 14:02:21 +0200, Cezary Rojewski wrote: > > Total of 6 fixes and 3 cleanups - cleanups are last. > > All of the fixes address problems that present themselves in situation > when user engages in codec driver reload. Second condition to reproduce > is two-step initialization of HDAudio codec - this is the case only for > ASoC HDAudio bus driver as snd_hda_intel calls only compound function > snd_hda_codec_new(). Once these conditions are met, several > reload/unload scenarios end with null-ptr-deref and page faults. Goal of > the series is to allow codec/bus driver reloading without any errors. > > Amadeusz Sławiński (2): > ALSA: hda: Reset all SIE bits in INTCTL > ALSA: hda: Remove unused macro definition > > Cezary Rojewski (7): > ALSA: hda: Do not unset preset when cleaning up codec > ALSA: hda: Fix null-ptr-deref when i915 fails and hdmi is denylisted > ALSA: hda: Make device usage_count consistent across subsequent > probing > ALSA: hda: Fix put_device() inconsistency in error path > ALSA: hda: Skip event processing for unregistered codecs > ALSA: hda: Fix page fault in snd_hda_codec_shutdown() > ALSA: hda/realtek: Remove redundant init_hook() in alc_default_init() Now applied partially what are applicable and look correct: patches 2, 3, 4, 6, 7, and 8. Patch 5 waits for Mark's PR of ASoC changes, and patch 1 needs a bit more clarification and investigation. thanks, Takashi
