On Wed, Nov 24, 2021 at 08:57:39AM -0800, Scott Bruce wrote:

It looks like the current work in progress branch for 5.15.5-rc1 picked up fe1bda72c1e8 "ASoC: SOF: Intel: hda-dai: fix potential locking issue" without also picking 868ddfcef31f "ALSA: hda: hdac_ext_stream: fix potential locking issues" from master. This causes builds to fail on x86_64 using clang:

... CC [M] drivers/misc/eeprom/at25.o CC [M] drivers/block/pktcdvd.o LTO [M] sound/soc/codecs/snd-soc-max98088.lto.o sound/soc/sof/intel/hda-dai.c:111:4: error: implicit declaration of function 'snd_hdac_ext_stream_decouple_locked' [-Werror,-Wimplicit-function-declaration] snd_hdac_ext_stream_decouple_locked(bus, res, true); ^ sound/soc/sof/intel/hda-dai.c:111:4: note: did you mean 'snd_hdac_ext_stream_decouple'? ./include/sound/hdaudio_ext.h:91:6: note: 'snd_hdac_ext_stream_decouple' declared here void snd_hdac_ext_stream_decouple(struct hdac_bus *bus, ^ 1 error generated. make[4]: *** [scripts/Makefile.build:277: sound/soc/sof/intel/hda-dai.o] Error 1 make[3]: *** [scripts/Makefile.build:540: sound/soc/sof/intel] Error 2 make[2]: *** [scripts/Makefile.build:540: sound/soc/sof] Error 2 make[2]: *** Waiting for unfinished jobs.... AR drivers/misc/cb710/built-in.a CC [M] drivers/misc/cb710/core.o CC [M] net/netfilter/xt_nfacct.o ...

After cherry-picking the second commit (868ddfcef31f) builds complete normally.

Now added, thanks.

In case it's relevant there's a third commit in this same series that's also not included: 1465d06a6d85, "ALSA: hda: hdac_stream: fix potential locking issue in snd_hdac_stream_assign()".

I've queued that up now as well, thanks!

greg k-h