September 2023 - Alsa-devel - mailman.alsa-project.org

[PATCH v3 0/3] ASoC: wm8782: Allow higher audio rates
by John Watts 18 Sep '23

18 Sep '23

The wm8782 supports higher audio rates than just 48kHz. This is configured by setting the FSAMPEN pin on the codec chip. This patch series introduces the 'wlf,fsampen' device tree property to indicate the pin status and control the maximum rate available when using the codec. v2 -> v3: - Rate is now properly constrained using ALSA constraints v1 -> v2: - Switched from max-rate property to wlf,fsampen property - Clarified property is optional, not required John Watts (3): ASoC: wm8782: Constrain maximum audio rate at runtime ASoC: wm8782: Use wlf,fsampen device tree property ASoC: dt-bindings: wlf,wm8782: Add wlf,fsampen property .../devicetree/bindings/sound/wm8782.txt | 5 ++ sound/soc/codecs/wm8782.c | 63 +++++++++++++++---- 2 files changed, 55 insertions(+), 13 deletions(-) -- 2.42.0

2 7

[PATCH 5.15 312/511] um: Fix hostaudio build errors
by Greg Kroah-Hartman 17 Sep '23

17 Sep '23

5.15-stable review patch. If anyone has any objections, please let me know. ------------------ From: Randy Dunlap <rdunlap(a)infradead.org> [ Upstream commit db4bfcba7bb8d10f00bba2a3da6b9a9c2a1d7b71 ] Use "select" to ensure that the required kconfig symbols are set as expected. Drop HOSTAUDIO since it is now equivalent to UML_SOUND. Set CONFIG_SOUND=m in ARCH=um defconfig files to maintain the status quo of the default configs. Allow SOUND with UML regardless of HAS_IOMEM. Otherwise there is a kconfig warning for unmet dependencies. (This was not an issue when SOUND was defined in arch/um/drivers/Kconfig. I have done 50 randconfig builds and didn't find any issues.) This fixes build errors when CONFIG_SOUND is not set: ld: arch/um/drivers/hostaudio_kern.o: in function `hostaudio_cleanup_module': hostaudio_kern.c:(.exit.text+0xa): undefined reference to `unregister_sound_mixer' ld: hostaudio_kern.c:(.exit.text+0x15): undefined reference to `unregister_sound_dsp' ld: arch/um/drivers/hostaudio_kern.o: in function `hostaudio_init_module': hostaudio_kern.c:(.init.text+0x19): undefined reference to `register_sound_dsp' ld: hostaudio_kern.c:(.init.text+0x31): undefined reference to `register_sound_mixer' ld: hostaudio_kern.c:(.init.text+0x49): undefined reference to `unregister_sound_dsp' and this kconfig warning: WARNING: unmet direct dependencies detected for SOUND Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Fixes: d886e87cb82b ("sound: make OSS sound core optional") Signed-off-by: Randy Dunlap <rdunlap(a)infradead.org> Reported-by: kernel test robot <lkp(a)intel.com> Closes: lore.kernel.org/r/202307141416.vxuRVpFv-lkp@intel.com Cc: Richard Weinberger <richard(a)nod.at> Cc: Anton Ivanov <anton.ivanov(a)cambridgegreys.com> Cc: Johannes Berg <johannes(a)sipsolutions.net> Cc: linux-um(a)lists.infradead.org Cc: Tejun Heo <tj(a)kernel.org> Cc: Takashi Iwai <tiwai(a)suse.de> Cc: Jaroslav Kysela <perex(a)perex.cz> Cc: Masahiro Yamada <masahiroy(a)kernel.org> Cc: Nathan Chancellor <nathan(a)kernel.org> Cc: Nick Desaulniers <ndesaulniers(a)google.com> Cc: Nicolas Schier <nicolas(a)fjasle.eu> Cc: linux-kbuild(a)vger.kernel.org Cc: alsa-devel(a)alsa-project.org Reviewed-by: Masahiro Yamada <masahiroy(a)kernel.org> Signed-off-by: Richard Weinberger <richard(a)nod.at> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- arch/um/configs/i386_defconfig | 1 + arch/um/configs/x86_64_defconfig | 1 + arch/um/drivers/Kconfig | 16 +++------------- arch/um/drivers/Makefile | 2 +- sound/Kconfig | 2 +- 5 files changed, 7 insertions(+), 15 deletions(-) diff --git a/arch/um/configs/i386_defconfig b/arch/um/configs/i386_defconfig index fb51bd206dbed..4d7f99a02c1eb 100644 --- a/arch/um/configs/i386_defconfig +++ b/arch/um/configs/i386_defconfig @@ -35,6 +35,7 @@ CONFIG_TTY_CHAN=y CONFIG_XTERM_CHAN=y CONFIG_CON_CHAN="pts" CONFIG_SSL_CHAN="pts" +CONFIG_SOUND=m CONFIG_UML_SOUND=m CONFIG_DEVTMPFS=y CONFIG_DEVTMPFS_MOUNT=y diff --git a/arch/um/configs/x86_64_defconfig b/arch/um/configs/x86_64_defconfig index 477b873174243..4bdd83008f623 100644 --- a/arch/um/configs/x86_64_defconfig +++ b/arch/um/configs/x86_64_defconfig @@ -33,6 +33,7 @@ CONFIG_TTY_CHAN=y CONFIG_XTERM_CHAN=y CONFIG_CON_CHAN="pts" CONFIG_SSL_CHAN="pts" +CONFIG_SOUND=m CONFIG_UML_SOUND=m CONFIG_DEVTMPFS=y CONFIG_DEVTMPFS_MOUNT=y diff --git a/arch/um/drivers/Kconfig b/arch/um/drivers/Kconfig index f145842c40b94..3dd74d369f995 100644 --- a/arch/um/drivers/Kconfig +++ b/arch/um/drivers/Kconfig @@ -104,24 +104,14 @@ config SSL_CHAN config UML_SOUND tristate "Sound support" + depends on SOUND + select SOUND_OSS_CORE help This option enables UML sound support. If enabled, it will pull in - soundcore and the UML hostaudio relay, which acts as a intermediary + the UML hostaudio relay, which acts as a intermediary between the host's dsp and mixer devices and the UML sound system. It is safe to say 'Y' here. -config SOUND - tristate - default UML_SOUND - -config SOUND_OSS_CORE - bool - default UML_SOUND - -config HOSTAUDIO - tristate - default UML_SOUND - endmenu menu "UML Network Devices" diff --git a/arch/um/drivers/Makefile b/arch/um/drivers/Makefile index cd1a58bb3e9cd..dc0e6fe77de10 100644 --- a/arch/um/drivers/Makefile +++ b/arch/um/drivers/Makefile @@ -54,7 +54,7 @@ obj-$(CONFIG_UML_NET) += net.o obj-$(CONFIG_MCONSOLE) += mconsole.o obj-$(CONFIG_MMAPPER) += mmapper_kern.o obj-$(CONFIG_BLK_DEV_UBD) += ubd.o -obj-$(CONFIG_HOSTAUDIO) += hostaudio.o +obj-$(CONFIG_UML_SOUND) += hostaudio.o obj-$(CONFIG_NULL_CHAN) += null.o obj-$(CONFIG_PORT_CHAN) += port.o obj-$(CONFIG_PTY_CHAN) += pty.o diff --git a/sound/Kconfig b/sound/Kconfig index e56d96d2b11ca..1903c35d799e1 100644 --- a/sound/Kconfig +++ b/sound/Kconfig @@ -1,7 +1,7 @@ # SPDX-License-Identifier: GPL-2.0-only menuconfig SOUND tristate "Sound card support" - depends on HAS_IOMEM + depends on HAS_IOMEM || UML help If you have a sound card in your computer, i.e. if it can say more than an occasional beep, say Y. -- 2.40.1

1 0

[PATCH 5.10 264/406] um: Fix hostaudio build errors
by Greg Kroah-Hartman 17 Sep '23

17 Sep '23

5.10-stable review patch. If anyone has any objections, please let me know. ------------------ From: Randy Dunlap <rdunlap(a)infradead.org> [ Upstream commit db4bfcba7bb8d10f00bba2a3da6b9a9c2a1d7b71 ] Use "select" to ensure that the required kconfig symbols are set as expected. Drop HOSTAUDIO since it is now equivalent to UML_SOUND. Set CONFIG_SOUND=m in ARCH=um defconfig files to maintain the status quo of the default configs. Allow SOUND with UML regardless of HAS_IOMEM. Otherwise there is a kconfig warning for unmet dependencies. (This was not an issue when SOUND was defined in arch/um/drivers/Kconfig. I have done 50 randconfig builds and didn't find any issues.) This fixes build errors when CONFIG_SOUND is not set: ld: arch/um/drivers/hostaudio_kern.o: in function `hostaudio_cleanup_module': hostaudio_kern.c:(.exit.text+0xa): undefined reference to `unregister_sound_mixer' ld: hostaudio_kern.c:(.exit.text+0x15): undefined reference to `unregister_sound_dsp' ld: arch/um/drivers/hostaudio_kern.o: in function `hostaudio_init_module': hostaudio_kern.c:(.init.text+0x19): undefined reference to `register_sound_dsp' ld: hostaudio_kern.c:(.init.text+0x31): undefined reference to `register_sound_mixer' ld: hostaudio_kern.c:(.init.text+0x49): undefined reference to `unregister_sound_dsp' and this kconfig warning: WARNING: unmet direct dependencies detected for SOUND Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Fixes: d886e87cb82b ("sound: make OSS sound core optional") Signed-off-by: Randy Dunlap <rdunlap(a)infradead.org> Reported-by: kernel test robot <lkp(a)intel.com> Closes: lore.kernel.org/r/202307141416.vxuRVpFv-lkp@intel.com Cc: Richard Weinberger <richard(a)nod.at> Cc: Anton Ivanov <anton.ivanov(a)cambridgegreys.com> Cc: Johannes Berg <johannes(a)sipsolutions.net> Cc: linux-um(a)lists.infradead.org Cc: Tejun Heo <tj(a)kernel.org> Cc: Takashi Iwai <tiwai(a)suse.de> Cc: Jaroslav Kysela <perex(a)perex.cz> Cc: Masahiro Yamada <masahiroy(a)kernel.org> Cc: Nathan Chancellor <nathan(a)kernel.org> Cc: Nick Desaulniers <ndesaulniers(a)google.com> Cc: Nicolas Schier <nicolas(a)fjasle.eu> Cc: linux-kbuild(a)vger.kernel.org Cc: alsa-devel(a)alsa-project.org Reviewed-by: Masahiro Yamada <masahiroy(a)kernel.org> Signed-off-by: Richard Weinberger <richard(a)nod.at> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- arch/um/configs/i386_defconfig | 1 + arch/um/configs/x86_64_defconfig | 1 + arch/um/drivers/Kconfig | 16 +++------------- arch/um/drivers/Makefile | 2 +- sound/Kconfig | 2 +- 5 files changed, 7 insertions(+), 15 deletions(-) diff --git a/arch/um/configs/i386_defconfig b/arch/um/configs/i386_defconfig index fb51bd206dbed..4d7f99a02c1eb 100644 --- a/arch/um/configs/i386_defconfig +++ b/arch/um/configs/i386_defconfig @@ -35,6 +35,7 @@ CONFIG_TTY_CHAN=y CONFIG_XTERM_CHAN=y CONFIG_CON_CHAN="pts" CONFIG_SSL_CHAN="pts" +CONFIG_SOUND=m CONFIG_UML_SOUND=m CONFIG_DEVTMPFS=y CONFIG_DEVTMPFS_MOUNT=y diff --git a/arch/um/configs/x86_64_defconfig b/arch/um/configs/x86_64_defconfig index 477b873174243..4bdd83008f623 100644 --- a/arch/um/configs/x86_64_defconfig +++ b/arch/um/configs/x86_64_defconfig @@ -33,6 +33,7 @@ CONFIG_TTY_CHAN=y CONFIG_XTERM_CHAN=y CONFIG_CON_CHAN="pts" CONFIG_SSL_CHAN="pts" +CONFIG_SOUND=m CONFIG_UML_SOUND=m CONFIG_DEVTMPFS=y CONFIG_DEVTMPFS_MOUNT=y diff --git a/arch/um/drivers/Kconfig b/arch/um/drivers/Kconfig index 2e7b8e0e7194b..01dfbd57e29d7 100644 --- a/arch/um/drivers/Kconfig +++ b/arch/um/drivers/Kconfig @@ -104,24 +104,14 @@ config SSL_CHAN config UML_SOUND tristate "Sound support" + depends on SOUND + select SOUND_OSS_CORE help This option enables UML sound support. If enabled, it will pull in - soundcore and the UML hostaudio relay, which acts as a intermediary + the UML hostaudio relay, which acts as a intermediary between the host's dsp and mixer devices and the UML sound system. It is safe to say 'Y' here. -config SOUND - tristate - default UML_SOUND - -config SOUND_OSS_CORE - bool - default UML_SOUND - -config HOSTAUDIO - tristate - default UML_SOUND - endmenu menu "UML Network Devices" diff --git a/arch/um/drivers/Makefile b/arch/um/drivers/Makefile index 2a249f6194671..207d62ab519df 100644 --- a/arch/um/drivers/Makefile +++ b/arch/um/drivers/Makefile @@ -52,7 +52,7 @@ obj-$(CONFIG_UML_NET) += net.o obj-$(CONFIG_MCONSOLE) += mconsole.o obj-$(CONFIG_MMAPPER) += mmapper_kern.o obj-$(CONFIG_BLK_DEV_UBD) += ubd.o -obj-$(CONFIG_HOSTAUDIO) += hostaudio.o +obj-$(CONFIG_UML_SOUND) += hostaudio.o obj-$(CONFIG_NULL_CHAN) += null.o obj-$(CONFIG_PORT_CHAN) += port.o obj-$(CONFIG_PTY_CHAN) += pty.o diff --git a/sound/Kconfig b/sound/Kconfig index 36785410fbe15..aaf2022ffc57d 100644 --- a/sound/Kconfig +++ b/sound/Kconfig @@ -1,7 +1,7 @@ # SPDX-License-Identifier: GPL-2.0-only menuconfig SOUND tristate "Sound card support" - depends on HAS_IOMEM + depends on HAS_IOMEM || UML help If you have a sound card in your computer, i.e. if it can say more than an occasional beep, say Y. -- 2.40.1

1 0

[PATCH 5.10 127/406] ASoC: stac9766: fix build errors with REGMAP_AC97
by Greg Kroah-Hartman 17 Sep '23

17 Sep '23

5.10-stable review patch. If anyone has any objections, please let me know. ------------------ From: Randy Dunlap <rdunlap(a)infradead.org> [ Upstream commit c70064b96f509daa78f57992aeabcf274fb2fed4 ] Select REGMAP_AC97 to fix these build errors: ERROR: modpost: "regmap_ac97_default_volatile" [sound/soc/codecs/snd-soc-stac9766.ko] undefined! ERROR: modpost: "__regmap_init_ac97" [sound/soc/codecs/snd-soc-stac9766.ko] undefined! Fixes: 6bbf787bb70c ("ASoC: stac9766: Convert to regmap") Signed-off-by: Randy Dunlap <rdunlap(a)infradead.org> Cc: Lars-Peter Clausen <lars(a)metafoo.de> Cc: Mark Brown <broonie(a)kernel.org> Cc: Liam Girdwood <lgirdwood(a)gmail.com> Cc: alsa-devel(a)alsa-project.org Link: https://lore.kernel.org/r/20230701044836.18789-1-rdunlap@infradead.org Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/soc/codecs/Kconfig b/sound/soc/codecs/Kconfig index f1c9e563994b2..04a7070c78e28 100644 --- a/sound/soc/codecs/Kconfig +++ b/sound/soc/codecs/Kconfig @@ -1295,6 +1295,7 @@ config SND_SOC_STA529 config SND_SOC_STAC9766 tristate depends on SND_SOC_AC97_BUS + select REGMAP_AC97 config SND_SOC_STI_SAS tristate "codec Audio support for STI SAS codec" -- 2.40.1

1 0

[PATCH 5.15 128/511] ASoC: stac9766: fix build errors with REGMAP_AC97
by Greg Kroah-Hartman 17 Sep '23

17 Sep '23

5.15-stable review patch. If anyone has any objections, please let me know. ------------------ From: Randy Dunlap <rdunlap(a)infradead.org> [ Upstream commit c70064b96f509daa78f57992aeabcf274fb2fed4 ] Select REGMAP_AC97 to fix these build errors: ERROR: modpost: "regmap_ac97_default_volatile" [sound/soc/codecs/snd-soc-stac9766.ko] undefined! ERROR: modpost: "__regmap_init_ac97" [sound/soc/codecs/snd-soc-stac9766.ko] undefined! Fixes: 6bbf787bb70c ("ASoC: stac9766: Convert to regmap") Signed-off-by: Randy Dunlap <rdunlap(a)infradead.org> Cc: Lars-Peter Clausen <lars(a)metafoo.de> Cc: Mark Brown <broonie(a)kernel.org> Cc: Liam Girdwood <lgirdwood(a)gmail.com> Cc: alsa-devel(a)alsa-project.org Link: https://lore.kernel.org/r/20230701044836.18789-1-rdunlap@infradead.org Signed-off-by: Mark Brown <broonie(a)kernel.org> Signed-off-by: Sasha Levin <sashal(a)kernel.org> --- sound/soc/codecs/Kconfig | 1 + 1 file changed, 1 insertion(+) diff --git a/sound/soc/codecs/Kconfig b/sound/soc/codecs/Kconfig index c3deb82c5da3a..1750cc888bbe8 100644 --- a/sound/soc/codecs/Kconfig +++ b/sound/soc/codecs/Kconfig @@ -1383,6 +1383,7 @@ config SND_SOC_STA529 config SND_SOC_STAC9766 tristate depends on SND_SOC_AC97_BUS + select REGMAP_AC97 config SND_SOC_STI_SAS tristate "codec Audio support for STI SAS codec" -- 2.40.1

1 0

[PATCH] selftests: ALSA: fix warnings in 'test-pcmtest-driver'
by Javier Carrasco 16 Sep '23

16 Sep '23

Defining the 'len' variable inside the 'patten_buf' as unsigned makes it more consistent with its actual meaning and the rest of the size variables in the test. Moreover, this removes an implicit conversion in the fscanf function call. Additionally, remove the unused variable 'it' from the reset_ioctl test. Signed-off-by: Javier Carrasco <javier.carrasco.cruz(a)gmail.com> --- Defining the 'len' variable inside the 'patten_buf' as unsigned makes it more consistent with its actual meaning and the rest of the size variables in the test. Moreover, this removes an implicit conversion in the fscanf function call. Additionally, remove the unused variable 'it' from the reset_ioctl test. --- tools/testing/selftests/alsa/test-pcmtest-driver.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/tools/testing/selftests/alsa/test-pcmtest-driver.c b/tools/testing/selftests/alsa/test-pcmtest-driver.c index 357adc722cba..f0dae651e495 100644 --- a/tools/testing/selftests/alsa/test-pcmtest-driver.c +++ b/tools/testing/selftests/alsa/test-pcmtest-driver.c @@ -13,7 +13,7 @@ struct pattern_buf { char buf[1024]; - int len; + unsigned int len; }; struct pattern_buf patterns[CH_NUM]; @@ -313,7 +313,6 @@ TEST_F(pcmtest, ni_playback) { */ TEST_F(pcmtest, reset_ioctl) { snd_pcm_t *handle; - unsigned char *it; int test_res; struct pcmtest_test_params *params = &self->params; --- base-commit: 0bb80ecc33a8fb5a682236443c1e740d5c917d1d change-id: 20230916-topic-pcmtest_warnings-ed074edee338 Best regards, -- Javier Carrasco <javier.carrasco.cruz(a)gmail.com>

2 3

AIF2 ADC Stereo Capture Route is uninitialised in alsa-ucm-conf
by GitHub issues - edited 16 Sep '23

16 Sep '23

alsa-project/alsa-ucm-conf issue #351 was edited from broukema: DESCRIPTION: `AIF2 ADC Stereo Capture Route` is not initialised in `alsa-ucm-conf`. This implies that if its setting is significant, any accidental change by the user (using any of the various `alsa` tools) to a wrong setting will not be corrected by any of the three `PinePhone/*.conf` files, leading, in particular, to audio errors during a phone call. This was discovered by testing variations on the Mobian version of the PinePhone alsa-ucm-conf files. CODING ARGUMENT: All parameters that can significantly affect the audio system should be either initialised or overridden in configuration files that the user expects to necessarily work, such as those for phone calls. In `alsa-ucm-conf` commit b68aa52 (one beyond v1.2.10): ```` $ grep -ri "AIF2 ADC Stereo Capture" . $ ```` shows that `AIF2 ADC Stereo Capture Route` is not set for any device at all (PinePhone or other). HARDWARE ARGUMENT: For the PinePhone (PP), Megi recommends in https://xnux.eu/devices/feature/audio-pp.html that `AIF2 ADC Stereo Capture Route` is set to `Mix Mono`. TESTING ARGUMENT: Experimentation with different stored alsa settings (from ```` sudo alsactl store -f - > YYYYMMDD_example_alsa_settings ```` ) showed that settings files whose only substantial difference was in the setting of `AIF2 ADC Stereo Capture Route` had the effect that audio from the PP was heard on the remote phone when the value was `Mix Mono` and was not heard with the value `Stereo`. Testing was done with the equivalent of https://codeberg.org/boud/pinephone_hacks/src/commit/7c7b6660a4446a24747f2c… with the settings file https://codeberg.org/boud/pinephone_hacks/src/commit/7c7b6660a4446a24747f2c… that has the `Mix Mono` value selected. RECOMMENDATIONS: * `AIF2 ADC Stereo Capture Route` should be initialised or set to *something* in one of the .conf files (in the PP case, most likely in `VoiceCall.conf`), at least for the PinePhone, to prevent users who experiment and have difficulty remembering which of the 50 PinePhone alsa parameters should be set in which way from accidentally sabotaging their otherwise working audio calls; * Based on Megi's 2020 page and my testing, it seems that the correct value should be `Mix Mono`. COMMENT: While setting all 50 of the alsa settings for the PinePHone ```` $ sudo alsactl store -f - |grep "name " |wc 50 258 1788 ```` in the .conf files is most likely unnecessary, there could be others that are necessary too. Ideally, someone who really understands how the PP audio system works (including Megi's two diagrams at https://xnux.eu/devices/feature/audio-pp.html) should be able to make these recommendations. Some other phones may need this or other parameters initialised too. Issue URL : https://github.com/alsa-project/alsa-ucm-conf/issues/351 Repository URL: https://github.com/alsa-project/alsa-ucm-conf

1 0

[PATCH] ALSA: rawmidi: Fix NULL dereference at proc read
by Takashi Iwai 16 Sep '23

16 Sep '23

At the implementation of the optional proc fs in rawmidi, I forgot that rmidi->ops itself is optional and can be NULL. Add the proper NULL check for avoiding the Oops. Fixes: fa030f666d24 ("ALSA: ump: Additional proc output") Reported-and-tested-by: Mark Hills <mark(a)xwax.org> Closes: https://lore.kernel.org/r/ef9118c3-a2eb-d0ff-1efa-cc5fb6416bde@xwax.org Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/core/rawmidi.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/sound/core/rawmidi.c b/sound/core/rawmidi.c index ba06484ac4aa..1431cb997808 100644 --- a/sound/core/rawmidi.c +++ b/sound/core/rawmidi.c @@ -1770,7 +1770,7 @@ static void snd_rawmidi_proc_info_read(struct snd_info_entry *entry, if (IS_ENABLED(CONFIG_SND_UMP)) snd_iprintf(buffer, "Type: %s\n", rawmidi_is_ump(rmidi) ? "UMP" : "Legacy"); - if (rmidi->ops->proc_read) + if (rmidi->ops && rmidi->ops->proc_read) rmidi->ops->proc_read(entry, buffer); mutex_lock(&rmidi->open_mutex); if (rmidi->info_flags & SNDRV_RAWMIDI_INFO_OUTPUT) { -- 2.35.3

1 0

[REGRESSION] rust midir MIDI library causes kernel oops
by Ash Holland 16 Sep '23

16 Sep '23

Hello, I upgraded to Linux 6.5 and found that my MIDI-input application no longer works, and causes an oops when I launch it. The application can be found at https://github.com/sersorrel/lp; `cargo run` is enough to cause the oops, though it has many undocumented dependencies, sorry (including a Novation Launchpad Mini Mk3). Once the oops occurs, it seems like it can still send MIDI to the Launchpad (i.e. display things on it), but input from the Launchpad doesn't work. I use NixOS with minimally-altered kernel configuration (blacklisted r8152 module and `amdgpu.reset_method=4` parameter), and was happily using kernel 6.4.9 or so before upgrading to 6.5. I bisected this to: commit f80e6d60d677be1d4dbbcdbf97379b8fbcf97ff0 Author: Takashi Iwai <tiwai(a)suse.de> Date: 2023-05-23 09:53:38 +0200 ALSA: seq: Clear padded bytes at expanding events There can be a small memory hole that may not be cleared at expanding an event with the variable length type. Make sure to clear it. Reviewed-by: Jaroslav Kysela <perex(a)perex.cz> Link: https://lore.kernel.org/r/20230523075358.9672-18-tiwai@suse.de Signed-off-by: Takashi Iwai <tiwai(a)suse.de> #regzbot introduced: f80e6d60d677be1d4dbbcdbf97379b8fbcf97ff0 I guess the problematic part is the `memset(buf + len, 0, newlen - len)`, which tries to memset a buffer that can be allocated in userspace. The oops: Sep 02 13:40:35 kernel: BUG: unable to handle page fault for address: 000055efb39dffb5 Sep 02 13:40:35 kernel: #PF: supervisor write access in kernel mode Sep 02 13:40:36 kernel: #PF: error_code(0x0003) - permissions violation Sep 02 13:40:36 kernel: PGD 1aff38067 P4D 1aff38067 PUD 1aff37067 PMD 1a975f067 PTE 80000001b315b067 Sep 02 13:40:36 kernel: Oops: 0003 [#1] PREEMPT SMP NOPTI Sep 02 13:40:36 kernel: CPU: 3 PID: 4441 Comm: midir ALSA inpu Not tainted 6.5.0 #1-NixOS Sep 02 13:40:36 kernel: Hardware name: To Be Filled By O.E.M. X570S PG Riptide/X570S PG Riptide, BIOS P5.01 01/17/2023 Sep 02 13:40:36 kernel: RIP: 0010:memset+0xf/0x20 Sep 02 13:40:36 kernel: Code: 44 88 1f e9 83 69 01 00 0f 1f 00 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 90 f3 0f 1e fa 66 90 49 89 f9 40 88 f0 48 89 d1 <f3> aa 4c 89 c8 e9 57 69 01 00 0f 1f 80 00 00 00 00 90 90 90 90 90 Sep 02 13:40:36 kernel: RSP: 0018:ffffa699c3607dd0 EFLAGS: 00010202 Sep 02 13:40:36 kernel: RAX: 0000000000000000 RBX: 0000000000000009 RCX: 0000000000000013 Sep 02 13:40:36 kernel: RDX: 0000000000000013 RSI: 0000000000000000 RDI: 000055efb39dffb5 Sep 02 13:40:36 kernel: RBP: 000000000000001c R08: 0000000000000009 R09: 000055efb39dffb5 Sep 02 13:40:36 kernel: R10: ffffa699c3607e68 R11: 0000000000000000 R12: 000055efb39dffac Sep 02 13:40:36 kernel: R13: 000055efb39dff00 R14: ffff9d2a2bd671e0 R15: ffff9d2aa2b36d00 Sep 02 13:40:36 kernel: FS: 00007fe33e2d86c0(0000) GS:ffff9d38feac0000(0000) knlGS:0000000000000000 Sep 02 13:40:36 kernel: CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 Sep 02 13:40:36 kernel: CR2: 000055efb39dffb5 CR3: 000000018e25c000 CR4: 0000000000750ee0 Sep 02 13:40:36 kernel: PKRU: 55555554 Sep 02 13:40:36 kernel: Call Trace: Sep 02 13:40:36 kernel: <TASK> Sep 02 13:40:36 kernel: ? __die+0x23/0x70 Sep 02 13:40:36 kernel: ? page_fault_oops+0x17d/0x4b0 Sep 02 13:40:36 kernel: ? exc_page_fault+0x6d/0x150 Sep 02 13:40:36 kernel: ? asm_exc_page_fault+0x26/0x30 Sep 02 13:40:36 kernel: ? memset+0xf/0x20 Sep 02 13:40:36 kernel: snd_seq_expand_var_event+0x6b/0xa0 [snd_seq] Sep 02 13:40:36 kernel: snd_seq_read+0x1b5/0x270 [snd_seq] Sep 02 13:40:36 kernel: vfs_read+0xaf/0x350 Sep 02 13:40:36 kernel: ? srso_alias_return_thunk+0x5/0x7f Sep 02 13:40:36 kernel: ? __fget_light+0x9d/0x100 Sep 02 13:40:36 kernel: ksys_read+0xbb/0xf0 Sep 02 13:40:36 kernel: do_syscall_64+0x3e/0x90 Sep 02 13:40:36 kernel: entry_SYSCALL_64_after_hwframe+0x6e/0xd8 Sep 02 13:40:36 kernel: RIP: 0033:0x7fe33ebf074c Sep 02 13:40:36 kernel: Code: ec 28 48 89 54 24 18 48 89 74 24 10 89 7c 24 08 e8 59 bb f8 ff 48 8b 54 24 18 48 8b 74 24 10 41 89 c0 8b 7c 24 08 31 c0 0f 05 <48> 3d 00 f0 ff ff 77 34 44 89 c7 48 89 44 24 08 e8 af bb f8 ff 48 Sep 02 13:40:36 kernel: RSP: 002b:00007fe33e2d7a70 EFLAGS: 00000246 ORIG_RAX: 0000000000000000 Sep 02 13:40:36 kernel: RAX: ffffffffffffffda RBX: 000055efb39d9440 RCX: 00007fe33ebf074c Sep 02 13:40:36 kernel: RDX: 00000000000036b0 RSI: 000055efb39dff90 RDI: 0000000000000005 Sep 02 13:40:36 kernel: RBP: 0000000000000002 R08: 0000000000000000 R09: 0000000000000020 Sep 02 13:40:36 kernel: R10: 2b7b3401fd8f65e7 R11: 0000000000000246 R12: 00007fe33e2d7d01 Sep 02 13:40:36 kernel: R13: 00007fe33e2d7b80 R14: 00007fe32c000cd0 R15: 0000000000000001 Sep 02 13:40:36 kernel: </TASK> Sep 02 13:40:36 kernel: Modules linked in: snd_hrtimer snd_seq_midi snd_seq_dummy snd_seq_midi_event snd_seq xt_MASQUERADE xt_mark nft_chain_nat nf_nat af_packet rfkill amdgpu snd_hda_codec_realtek snd_hda_codec_generic nls_iso8859_1 nls_cp437 ledtrig_audio vfat fat amdxcp snd_hda_codec_hdmi iommu_v2 drm_buddy gpu_sc> Sep 02 13:40:36 kernel: xt_tcpudp nft_compat nf_tables sch_fq_codel nfnetlink uinput i2c_piix4 i2c_dev ctr atkbd libps2 serio vivaldi_fmap loop tun tap macvlan bridge stp llc kvm_amd ccp kvm drm irqbypass fuse deflate efi_pstore backlight configfs efivarfs dmi_sysfs ip_tables x_tables autofs4 dm_crypt cbc encrypted_> Sep 02 13:40:36 kernel: CR2: 000055efb39dffb5 Sep 02 13:40:36 kernel: ---[ end trace 0000000000000000 ]--- thanks, Ash -- Ash Holland, she/her

3 8

[PATCH] ASoC: cs42l43: Add shared IRQ flag for shutters
by Charles Keepax 16 Sep '23

16 Sep '23

The microphone and speaker shutters on cs42l43 can be configured to trigger from the same GPIO, in this case the current code returns an error as we attempt to request two IRQ handlers for the same IRQ. Fix this by always requesting the shutter IRQs with the IRQF_SHARED flag. Signed-off-by: Charles Keepax <ckeepax(a)opensource.cirrus.com> --- sound/soc/codecs/cs42l43.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-) diff --git a/sound/soc/codecs/cs42l43.c b/sound/soc/codecs/cs42l43.c index 1a95c370fc4c9..5643c666d7d04 100644 --- a/sound/soc/codecs/cs42l43.c +++ b/sound/soc/codecs/cs42l43.c @@ -2077,7 +2077,8 @@ static const struct cs42l43_irq cs42l43_irqs[] = { static int cs42l43_request_irq(struct cs42l43_codec *priv, struct irq_domain *dom, const char * const name, - unsigned int irq, irq_handler_t handler) + unsigned int irq, irq_handler_t handler, + unsigned long flags) { int ret; @@ -2087,8 +2088,8 @@ static int cs42l43_request_irq(struct cs42l43_codec *priv, dev_dbg(priv->dev, "Request IRQ %d for %s\n", ret, name); - ret = devm_request_threaded_irq(priv->dev, ret, NULL, handler, IRQF_ONESHOT, - name, priv); + ret = devm_request_threaded_irq(priv->dev, ret, NULL, handler, + IRQF_ONESHOT | flags, name, priv); if (ret) return dev_err_probe(priv->dev, ret, "Failed to request IRQ %s\n", name); @@ -2124,11 +2125,11 @@ static int cs42l43_shutter_irq(struct cs42l43_codec *priv, return 0; } - ret = cs42l43_request_irq(priv, dom, close_name, close_irq, handler); + ret = cs42l43_request_irq(priv, dom, close_name, close_irq, handler, IRQF_SHARED); if (ret) return ret; - return cs42l43_request_irq(priv, dom, open_name, open_irq, handler); + return cs42l43_request_irq(priv, dom, open_name, open_irq, handler, IRQF_SHARED); } static int cs42l43_codec_probe(struct platform_device *pdev) @@ -2178,7 +2179,8 @@ static int cs42l43_codec_probe(struct platform_device *pdev) for (i = 0; i < ARRAY_SIZE(cs42l43_irqs); i++) { ret = cs42l43_request_irq(priv, dom, cs42l43_irqs[i].name, - cs42l43_irqs[i].irq, cs42l43_irqs[i].handler); + cs42l43_irqs[i].irq, + cs42l43_irqs[i].handler, 0); if (ret) goto err_pm; } -- 2.30.2

2 1