| Message ID | 20210302003125.1178419-1-pierre-louis.bossart@linux.intel.com |
|---|---|
| Headers | show |
| Series | ALSA/ASoC/SOF/SoundWire: fix Kconfig issues | expand |
On Mon, Mar 01, 2021 at 06:31:18PM -0600, Pierre-Louis Bossart wrote: > In January, Intel kbuild bot and Arnd Bergmann reported multiple > issues with randconfig. This patchset builds on Arnd's suggestions to Acked-by: Mark Brown <broonie@kernel.org>
On 02-03-21, 15:03, Takashi Iwai wrote: > On Tue, 02 Mar 2021 01:31:18 +0100, > Pierre-Louis Bossart wrote: > > > > In January, Intel kbuild bot and Arnd Bergmann reported multiple > > issues with randconfig. This patchset builds on Arnd's suggestions to > > > > a) expose ACPI and PCI devices in separate modules, while sof-acpi-dev > > and sof-pci-dev become helpers. This will result in minor changes > > required for developers/testers, i.e. modprobe snd-sof-pci will no > > longer result in a probe. The SOF CI was already updated to deal with > > this module dependency change and introduction of new modules. > > > > b) Fix SOF/SoundWire/DSP_config dependencies by moving the code > > required to detect SoundWire presence in ACPI tables to sound/hda. > > > > Integration note: > > This patchset touches directories maintained by Vinod, Takashi and > > Mark in separate trees, and will impact additional changes to use the > > auxiliary bus in drivers/soundwire/. > > I can think of two options, both of which are fine: > > 1. Mark merges the patches with Vinod and Takashi Acked-by tags, then > > Mark provides an immutable tag to Vinod. > > 2. Vinod merges the patches with Mark and Takashi Acked-by tags, then > > Vinod provides an immutable tag to Mark > > Acked-by: Takashi Iwai <tiwai@suse.de> Since most of the changes are sound/ it makes sense for this to go thru sound tree. I have acked the sdw part.. Thanks
> OK, I applied those and pushed to topic/sound-sdw-kconfig-fixes > immutable branch. A signed tag, sound-sdw-kconfig-fixes, is found > there, too. > > Mark, Vinod, feel free to pull from: > git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git tags/sound-sdw-kconfig/fixes > Thanks Takashi! > I suppose those are supposed to be merged in 5.12? If so, I'm going > to merge to for-linus branch as well. We finished those changes rather late in the cycle, so I didn't want to send this before 5.12-rc1 and create issues. But we've been testing them for at least a month now, so the risk is low if this series was added in an 5.12-rcX branch.
On Tue, 02 Mar 2021 16:03:37 +0100, Mark Brown wrote: > > On Tue, Mar 02, 2021 at 03:43:23PM +0100, Takashi Iwai wrote: > > > Mark, Vinod, feel free to pull from: > > git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git tags/sound-sdw-kconfig/fixes > > $ git pull git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git tags/sound-sdw-kconfig/fixes Sorry, copy&paste error, it's not with "/fixes" but "-fixes". git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound.git tags/sound-sdw-kconfig-fixes Takashi
On Tue, Mar 2, 2021 at 1:31 AM Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> wrote: > > In January, Intel kbuild bot and Arnd Bergmann reported multiple > issues with randconfig. This patchset builds on Arnd's suggestions to > > a) expose ACPI and PCI devices in separate modules, while sof-acpi-dev > and sof-pci-dev become helpers. This will result in minor changes > required for developers/testers, i.e. modprobe snd-sof-pci will no > longer result in a probe. The SOF CI was already updated to deal with > this module dependency change and introduction of new modules. > > b) Fix SOF/SoundWire/DSP_config dependencies by moving the code > required to detect SoundWire presence in ACPI tables to sound/hda. > > Integration note: > This patchset touches directories maintained by Vinod, Takashi and > Mark in separate trees, and will impact additional changes to use the > auxiliary bus in drivers/soundwire/. > I can think of two options, both of which are fine: > 1. Mark merges the patches with Vinod and Takashi Acked-by tags, then > Mark provides an immutable tag to Vinod. > 2. Vinod merges the patches with Mark and Takashi Acked-by tags, then > Vinod provides an immutable tag to Mark > > Acknowledgements: > Thanks to Arnd for suggesting fixes and testing these patches with > more randconfigs. > Thanks to Vinod Koul for his feedback on the move of this common > helper to sound/hda/ Hi Pierre-Louis, I have not done as much randconfig tested recently as I used to, but since I ramped up the rate again after the merge window, I have unfortunately once more run into a variation of this problem, with CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE=m and CONFIG_SND_SOC_SOF_HDA_COMMON=y: x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_dsp_interrupt_thread': hda.c:(.text+0x739): undefined reference to `sdw_intel_thread' x86_64-linux-ld: hda.c:(.text+0x764): undefined reference to `sdw_intel_process_wakeen_event' x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_sdw_int_enable': hda.c:(.text+0x782): undefined reference to `sdw_intel_enable_irq' x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_sdw_startup': hda.c:(.text+0x7ac): undefined reference to `sdw_intel_startup' x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_sdw_process_wakeen': hda.c:(.text+0x7dc): undefined reference to `sdw_intel_process_wakeen_event' x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_dsp_probe': hda.c:(.text+0xee1): undefined reference to `sdw_intel_probe' x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_dsp_remove': hda.c:(.text+0xf7f): undefined reference to `sdw_intel_enable_irq' x86_64-linux-ld: hda.c:(.text+0xf8e): undefined reference to `sdw_intel_exit' This only seems to appear very rarely, as I still did around 25000 randconfig builds since you merged the series, but this config does show the problem for any kernel version between the fix and the latest linux-next: https://pastebin.com/raw/VCjcB3EU I have only done the quick bisection but no analysis regarding what we may have missed back in March. I hope you can figure this out. Arnd
> I have not done as much randconfig tested recently as I used to, but > since I ramped up the rate again after the merge window, I have unfortunately > once more run into a variation of this problem, with > CONFIG_SND_SOC_SOF_INTEL_SOUNDWIRE=m and > CONFIG_SND_SOC_SOF_HDA_COMMON=y: > > x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function > `hda_dsp_interrupt_thread': > hda.c:(.text+0x739): undefined reference to `sdw_intel_thread' > x86_64-linux-ld: hda.c:(.text+0x764): undefined reference to > `sdw_intel_process_wakeen_event' > x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_sdw_int_enable': > hda.c:(.text+0x782): undefined reference to `sdw_intel_enable_irq' > x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_sdw_startup': > hda.c:(.text+0x7ac): undefined reference to `sdw_intel_startup' > x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function > `hda_sdw_process_wakeen': > hda.c:(.text+0x7dc): undefined reference to `sdw_intel_process_wakeen_event' > x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_dsp_probe': > hda.c:(.text+0xee1): undefined reference to `sdw_intel_probe' > x86_64-linux-ld: sound/soc/sof/intel/hda.o: in function `hda_dsp_remove': > hda.c:(.text+0xf7f): undefined reference to `sdw_intel_enable_irq' > x86_64-linux-ld: hda.c:(.text+0xf8e): undefined reference to `sdw_intel_exit' > > This only seems to appear very rarely, as I still did around 25000 randconfig > builds since you merged the series, but this config does show the problem > for any kernel version between the fix and the latest linux-next: > > https://pastebin.com/raw/VCjcB3EU > > I have only done the quick bisection but no analysis regarding what > we may have missed back in March. I hope you can figure this out. Thanks for this report Arnd. I can reproduce the problem, the changes made 6 months back don't make too much sense to me any longer. I have really no idea why I moved the INTEL_SOUNDWIRE config to a boolean, that was asking for trouble. I have a tentative fix being tested at https://github.com/thesofproject/linux/pull/3066. I just moved the select operations to a different option and reverted the INTEL_SOUNDWIRE config to be an on/off switch from the user perspective, with dependencies on built-in v. module checked inside.
On Mon, Jul 26, 2021 at 11:01 PM Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com> wrote: > > Thanks for this report Arnd. > > I can reproduce the problem, the changes made 6 months back don't make > too much sense to me any longer. I have really no idea why I moved the > INTEL_SOUNDWIRE config to a boolean, that was asking for trouble. > > I have a tentative fix being tested at > https://github.com/thesofproject/linux/pull/3066. I just moved the > select operations to a different option and reverted the INTEL_SOUNDWIRE > config to be an on/off switch from the user perspective, with > dependencies on built-in v. module checked inside. Ah, I'm glad you at least came up with a patch quickly, even if neither of us remembers what we were thinking then ;-) I've added this to my test tree, and will be able to at least verify that this doesn't break anything obvious over night. Arnd