July 2019 - Alsa-devel - mailman.alsa-project.org

[alsa-devel] [PATCH 0/2] ASoC: samsung: odroid: fix err handling of odroid_audio_probe
by Wen Yang 16 Jul '19

16 Jul '19

We developed a coccinelle SmPL to detect sound/soc/samsung/odroid.c and found some use-after-free problems. This patch series fixes those problems. Wen Yang (2): ASoC: samsung: odroid: fix an use-after-free issue for codec ASoC: samsung: odroid: fix a double-free issue for cpu_dai sound/soc/samsung/odroid.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) Cc: Krzysztof Kozlowski <krzk(a)kernel.org> Cc: Sangbeom Kim <sbkim73(a)samsung.com> Cc: Sylwester Nawrocki <s.nawrocki(a)samsung.com> Cc: Liam Girdwood <lgirdwood(a)gmail.com> Cc: Mark Brown <broonie(a)kernel.org> Cc: Jaroslav Kysela <perex(a)perex.cz> Cc: Takashi Iwai <tiwai(a)suse.com> Cc: alsa-devel(a)alsa-project.org Cc: linux-kernel(a)vger.kernel.org -- 2.9.5

5 12

[alsa-devel] [PATCH] ASoC: Intel: bytcht_es8316: Add quirk for Irbis NB41 netbook
by Hans de Goede 16 Jul '19

16 Jul '19

The Irbis NB41 netbook has its internal mic on IN2, inverted jack-detect and stereo speakers, add a quirk for this. Cc: russianneuromancer(a)ya.ru Reported-and-tested-by: russianneuromancer(a)ya.ru Signed-off-by: Hans de Goede <hdegoede(a)redhat.com> --- sound/soc/intel/boards/bytcht_es8316.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/sound/soc/intel/boards/bytcht_es8316.c b/sound/soc/intel/boards/bytcht_es8316.c index 2fe1ce879123..c360ebc3ccc7 100644 --- a/sound/soc/intel/boards/bytcht_es8316.c +++ b/sound/soc/intel/boards/bytcht_es8316.c @@ -436,6 +436,14 @@ static const struct acpi_gpio_mapping byt_cht_es8316_gpios[] = { /* Please keep this list alphabetically sorted */ static const struct dmi_system_id byt_cht_es8316_quirk_table[] = { + { /* Irbis NB41 */ + .matches = { + DMI_MATCH(DMI_SYS_VENDOR, "IRBIS"), + DMI_MATCH(DMI_PRODUCT_NAME, "NB41"), + }, + .driver_data = (void *)(BYT_CHT_ES8316_INTMIC_IN2_MAP + | BYT_CHT_ES8316_JD_INVERTED), + }, { /* Teclast X98 Plus II */ .matches = { DMI_MATCH(DMI_SYS_VENDOR, "TECLAST"), -- 2.21.0

2 1

[alsa-devel] [PATCH 1/2] ALSA: hda/hdmi - Remove duplicated define
by Takashi Iwai 16 Jul '19

16 Jul '19

INTEL_GET_VENDOR_VERB is defined twice identically. Let's remove a superfluous line. Fixes: b0d8bc50b9f2 ("ALSA: hda: hdmi - add Icelake support") Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/pci/hda/patch_hdmi.c | 1 - 1 file changed, 1 deletion(-) diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 40323d91f9e4..1e6c489bca15 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -2416,7 +2416,6 @@ static void intel_haswell_fixup_connect_list(struct hda_codec *codec, snd_hda_override_conn_list(codec, nid, spec->num_cvts, spec->cvt_nids); } -#define INTEL_GET_VENDOR_VERB 0xf81 #define INTEL_GET_VENDOR_VERB 0xf81 #define INTEL_SET_VENDOR_VERB 0x781 #define INTEL_EN_DP12 0x02 /* enable DP 1.2 features */ -- 2.16.4

2 4

[alsa-devel] [PATCH v2] ALSA: hda/hdmi - Fix i915 reverse port/pin mapping
by Takashi Iwai 16 Jul '19

16 Jul '19

The recent fix for Icelake HDMI codec introduced the mapping from pin NID to the i915 gfx port number. However, it forgot the reverse mapping from the port number to the pin NID that is used in the ELD notifier callback. As a result, it's processed to a wrong widget and gives a warning like snd_hda_codec_hdmi hdaudioC0D2: HDMI: pin nid 5 not registered This patch corrects it with a proper reverse mapping function. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=204133 Fixes: b0d8bc50b9f2 ("ALSA: hda: hdmi - add Icelake support") Reviewed-by: Kai Vehmanen <kai.vehmanen(a)linux.intel.com> Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- v1->v2: Add missing error check from intel_port2pin() sound/pci/hda/patch_hdmi.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-) diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 1e6c489bca15..0b2a26e2c5f1 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -2524,18 +2524,32 @@ static int intel_pin2port(void *audio_ptr, int pin_nid) return -1; } +static int intel_port2pin(struct hda_codec *codec, int port) +{ + struct hdmi_spec *spec = codec->spec; + + if (!spec->port_num) { + /* we assume only from port-B to port-D */ + if (port < 1 || port > 3) + return 0; + /* intel port is 1-based */ + return port + intel_base_nid(codec) - 1; + } + + if (port < 1 || port > spec->port_num) + return 0; + return spec->port_map[port - 1]; +} + static void intel_pin_eld_notify(void *audio_ptr, int port, int pipe) { struct hda_codec *codec = audio_ptr; int pin_nid; int dev_id = pipe; - /* we assume only from port-B to port-D */ - if (port < 1 || port > 3) + pin_nid = intel_port2pin(codec, port); + if (!pin_nid) return; - - pin_nid = port + intel_base_nid(codec) - 1; /* intel port is 1-based */ - /* skip notification during system suspend (but not in runtime PM); * the state will be updated at resume */ -- 2.16.4

1 0

[alsa-devel] [PATCH 00/14] pending doc patches for 5.3-rc
by Mauro Carvalho Chehab 16 Jul '19

16 Jul '19

Those are the pending documentation patches after my pull request for this branch: git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media.git tags/docs/v5.3-1 Patches 1 to 13 were already submitted, but got rebased. Patch 14 is a new fixup one. Patches 1 and 2 weren't submitted before due to merge conflicts that are now solved upstream; Patch 3 fixes a series of random Documentation/* references that are pointing to the wrong places. Patch 4 fix a longstanding issue: every time a new book is added, conf.py need changes, in order to allow generating a PDF file. After the patch, conf.py will automatically recognize new books, saving the trouble of keeping adding documents to it. Patches 5 to 11 are due to fonts support when building translations.pdf. The main focus is to add xeCJK support. While doing it, I discovered some bugs at sphinx-pre-install script after running it with 7 different distributions. Patch 12 improves support for partial doc building. Currently, each subdir needs to have its own conf.py, in order to support partial doc build. After it, any Documentation subdir can be used to roduce html/pdf docs with: make SPHINXDIRS="foo bar" htmldocs (or pdfdocs, latexdocs, epubdocs, ...) Patch 13 is a cleanup patch: it simply get rid of all those extra conf.py files that aren't needed anymore. The only extra config file after it is this one: Documentation/media/conf_nitpick.py With enables some extra optional Sphinx features. Patch 14 adds Documentation/virtual to the main index.rst file and add a new *.rst file that was orphaned there. - After this series, there's just one more patch meant to be applied for 5.3, with is still waiting for some patches to be merged from linux-next: https://git.linuxtv.org/mchehab/experimental.git/commit/?id=b1b5dc7d7bbfbbf… Mauro Carvalho Chehab (14): docs: powerpc: convert docs to ReST and rename to *.rst docs: power: add it to to the main documentation index docs: fix broken doc references due to renames docs: pdf: add all Documentation/*/index.rst to PDF output docs: conf.py: add CJK package needed by translations docs: conf.py: only use CJK if the font is available scripts/sphinx-pre-install: fix script for RHEL/CentOS scripts/sphinx-pre-install: don't use LaTeX with CentOS 7 scripts/sphinx-pre-install: fix latexmk dependencies scripts/sphinx-pre-install: cleanup Gentoo checks scripts/sphinx-pre-install: seek for Noto CJK fonts for pdf output docs: load_config.py: avoid needing a conf.py just due to LaTeX docs docs: remove extra conf.py files docs: virtual: add it to the documentation body Documentation/PCI/pci-error-recovery.rst | 5 +- Documentation/RCU/rculist_nulls.txt | 2 +- Documentation/admin-guide/conf.py | 10 -- Documentation/conf.py | 30 +++- Documentation/core-api/conf.py | 10 -- Documentation/crypto/conf.py | 10 -- Documentation/dev-tools/conf.py | 10 -- .../devicetree/bindings/arm/idle-states.txt | 2 +- Documentation/doc-guide/conf.py | 10 -- Documentation/driver-api/80211/conf.py | 10 -- Documentation/driver-api/conf.py | 10 -- Documentation/driver-api/pm/conf.py | 10 -- Documentation/filesystems/conf.py | 10 -- Documentation/gpu/conf.py | 10 -- Documentation/index.rst | 3 + Documentation/input/conf.py | 10 -- Documentation/kernel-hacking/conf.py | 10 -- Documentation/locking/spinlocks.rst | 4 +- Documentation/maintainer/conf.py | 10 -- Documentation/media/conf.py | 12 -- Documentation/memory-barriers.txt | 2 +- Documentation/networking/conf.py | 10 -- Documentation/power/index.rst | 2 +- .../{bootwrapper.txt => bootwrapper.rst} | 28 +++- .../{cpu_families.txt => cpu_families.rst} | 23 +-- .../{cpu_features.txt => cpu_features.rst} | 6 +- Documentation/powerpc/{cxl.txt => cxl.rst} | 46 ++++-- .../powerpc/{cxlflash.txt => cxlflash.rst} | 10 +- .../{DAWR-POWER9.txt => dawr-power9.rst} | 15 +- Documentation/powerpc/{dscr.txt => dscr.rst} | 18 +- ...ecovery.txt => eeh-pci-error-recovery.rst} | 108 ++++++------ ...ed-dump.txt => firmware-assisted-dump.rst} | 117 +++++++------ Documentation/powerpc/{hvcs.txt => hvcs.rst} | 108 ++++++------ Documentation/powerpc/index.rst | 34 ++++ Documentation/powerpc/isa-versions.rst | 15 +- .../powerpc/{mpc52xx.txt => mpc52xx.rst} | 12 +- ...nv.txt => pci_iov_resource_on_powernv.rst} | 15 +- .../powerpc/{pmu-ebb.txt => pmu-ebb.rst} | 1 + Documentation/powerpc/ptrace.rst | 156 ++++++++++++++++++ Documentation/powerpc/ptrace.txt | 151 ----------------- .../{qe_firmware.txt => qe_firmware.rst} | 37 +++-- .../{syscall64-abi.txt => syscall64-abi.rst} | 29 ++-- ...al_memory.txt => transactional_memory.rst} | 45 ++--- Documentation/process/conf.py | 10 -- Documentation/sh/conf.py | 10 -- Documentation/sound/conf.py | 10 -- Documentation/sphinx/load_config.py | 27 ++- .../translations/ko_KR/memory-barriers.txt | 2 +- Documentation/userspace-api/conf.py | 10 -- Documentation/virtual/kvm/index.rst | 1 + Documentation/vm/conf.py | 10 -- Documentation/watchdog/hpwdt.rst | 2 +- Documentation/x86/conf.py | 10 -- MAINTAINERS | 14 +- arch/powerpc/kernel/exceptions-64s.S | 2 +- drivers/gpu/drm/drm_modes.c | 2 +- drivers/i2c/busses/i2c-nvidia-gpu.c | 2 +- drivers/scsi/hpsa.c | 4 +- drivers/soc/fsl/qe/qe.c | 2 +- drivers/tty/hvc/hvcs.c | 2 +- include/soc/fsl/qe/qe.h | 2 +- scripts/sphinx-pre-install | 118 ++++++++++--- 62 files changed, 738 insertions(+), 678 deletions(-) delete mode 100644 Documentation/admin-guide/conf.py delete mode 100644 Documentation/core-api/conf.py delete mode 100644 Documentation/crypto/conf.py delete mode 100644 Documentation/dev-tools/conf.py delete mode 100644 Documentation/doc-guide/conf.py delete mode 100644 Documentation/driver-api/80211/conf.py delete mode 100644 Documentation/driver-api/conf.py delete mode 100644 Documentation/driver-api/pm/conf.py delete mode 100644 Documentation/filesystems/conf.py delete mode 100644 Documentation/gpu/conf.py delete mode 100644 Documentation/input/conf.py delete mode 100644 Documentation/kernel-hacking/conf.py delete mode 100644 Documentation/maintainer/conf.py delete mode 100644 Documentation/media/conf.py delete mode 100644 Documentation/networking/conf.py rename Documentation/powerpc/{bootwrapper.txt => bootwrapper.rst} (93%) rename Documentation/powerpc/{cpu_families.txt => cpu_families.rst} (95%) rename Documentation/powerpc/{cpu_features.txt => cpu_features.rst} (97%) rename Documentation/powerpc/{cxl.txt => cxl.rst} (95%) rename Documentation/powerpc/{cxlflash.txt => cxlflash.rst} (98%) rename Documentation/powerpc/{DAWR-POWER9.txt => dawr-power9.rst} (95%) rename Documentation/powerpc/{dscr.txt => dscr.rst} (91%) rename Documentation/powerpc/{eeh-pci-error-recovery.txt => eeh-pci-error-recovery.rst} (82%) rename Documentation/powerpc/{firmware-assisted-dump.txt => firmware-assisted-dump.rst} (80%) rename Documentation/powerpc/{hvcs.txt => hvcs.rst} (91%) create mode 100644 Documentation/powerpc/index.rst rename Documentation/powerpc/{mpc52xx.txt => mpc52xx.rst} (91%) rename Documentation/powerpc/{pci_iov_resource_on_powernv.txt => pci_iov_resource_on_powernv.rst} (97%) rename Documentation/powerpc/{pmu-ebb.txt => pmu-ebb.rst} (99%) create mode 100644 Documentation/powerpc/ptrace.rst delete mode 100644 Documentation/powerpc/ptrace.txt rename Documentation/powerpc/{qe_firmware.txt => qe_firmware.rst} (95%) rename Documentation/powerpc/{syscall64-abi.txt => syscall64-abi.rst} (82%) rename Documentation/powerpc/{transactional_memory.txt => transactional_memory.rst} (93%) delete mode 100644 Documentation/process/conf.py delete mode 100644 Documentation/sh/conf.py delete mode 100644 Documentation/sound/conf.py delete mode 100644 Documentation/userspace-api/conf.py delete mode 100644 Documentation/vm/conf.py delete mode 100644 Documentation/x86/conf.py -- 2.21.0

1 1

[alsa-devel] [PATCH v3 0/5] Add HDMI jack support on RK3288
by Cheng-Yi Chiang 16 Jul '19

16 Jul '19

This patch series supports HDMI jack reporting on RK3288, which uses DRM dw-hdmi driver and hdmi-codec codec driver. The previous discussion about reporting jack status using hdmi-notifier and drm_audio_component is at https://lore.kernel.org/patchwork/patch/1083027/ The new approach is to use a callback mechanism that is specific to hdmi-codec. Changes from v2 to v3: - dw-hdmi-i2s-audio.c: Use fixed ID instead of auto ID. - rk3288_hdmi_analog.c: Use the fixed name hdmi-audio-codec for codec device. - rockchip_max98090: Use the fixed name hdmi-audio-codec for codec device. - rockchip_max98090: Fix the dependency of hdmi-codec in Kconfig. Cheng-Yi Chiang (5): ASoC: hdmi-codec: Add an op to set callback function for plug event drm: bridge: dw-hdmi: Report connector status using callback drm: dw-hdmi-i2s: Use fixed id for codec device ASoC: rockchip_max98090: Add dai_link for HDMI FROMLIST: ASoC: rockchip_max98090: Add HDMI jack support .../gpu/drm/bridge/synopsys/dw-hdmi-audio.h | 3 + .../drm/bridge/synopsys/dw-hdmi-i2s-audio.c | 12 +- drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 55 ++++++++- include/sound/hdmi-codec.h | 16 +++ sound/soc/codecs/hdmi-codec.c | 45 +++++++ sound/soc/rockchip/Kconfig | 3 +- sound/soc/rockchip/rk3288_hdmi_analog.c | 3 +- sound/soc/rockchip/rockchip_max98090.c | 116 ++++++++++++++---- 8 files changed, 226 insertions(+), 27 deletions(-) -- 2.22.0.510.g264f2c817a-goog

3 11

[alsa-devel] [PATCH] ALSA: hda - Don't resume forcibly i915 HDMI/DP codec
by Takashi Iwai 16 Jul '19

16 Jul '19

We apply the codec resume forcibly at system resume callback for updating and syncing the jack detection state that may have changed during sleeping. This is, however, superfluous for the codec like Intel HDMI/DP, where the jack detection is managed via the audio component notification; i.e. the jack state change shall be reported sooner or later from the graphics side at mode change. This patch changes the codec resume callback to avoid the forcible resume conditionally with a new flag, codec->relaxed_resume, for reducing the resume time. The flag is set in the codec probe. Although this doesn't fix the entire bug mentioned in the bugzilla entry below, it's still a good optimization and some improvements are seen. Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=201901 Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- include/sound/hda_codec.h | 2 ++ sound/pci/hda/hda_codec.c | 8 ++++++-- sound/pci/hda/patch_hdmi.c | 6 +++++- 3 files changed, 13 insertions(+), 3 deletions(-) diff --git a/include/sound/hda_codec.h b/include/sound/hda_codec.h index 8f46ff3449d5..871993696c5f 100644 --- a/include/sound/hda_codec.h +++ b/include/sound/hda_codec.h @@ -252,6 +252,8 @@ struct hda_codec { unsigned int auto_runtime_pm:1; /* enable automatic codec runtime pm */ unsigned int force_pin_prefix:1; /* Add location prefix */ unsigned int link_down_at_suspend:1; /* link down at runtime suspend */ + unsigned int relaxed_resume:1; /* don't resume forcibly for jack */ + #ifdef CONFIG_PM unsigned long power_on_acct; unsigned long power_off_acct; diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 5346631df1ec..e30e86ca6b72 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -2941,15 +2941,19 @@ static int hda_codec_runtime_resume(struct device *dev) #ifdef CONFIG_PM_SLEEP static int hda_codec_force_resume(struct device *dev) { + struct hda_codec *codec = dev_to_hda_codec(dev); + bool forced_resume = !codec->relaxed_resume; int ret; /* The get/put pair below enforces the runtime resume even if the * device hasn't been used at suspend time. This trick is needed to * update the jack state change during the sleep. */ - pm_runtime_get_noresume(dev); + if (forced_resume) + pm_runtime_get_noresume(dev); ret = pm_runtime_force_resume(dev); - pm_runtime_put(dev); + if (forced_resume) + pm_runtime_put(dev); return ret; } diff --git a/sound/pci/hda/patch_hdmi.c b/sound/pci/hda/patch_hdmi.c index 40323d91f9e4..72000f6fedca 100644 --- a/sound/pci/hda/patch_hdmi.c +++ b/sound/pci/hda/patch_hdmi.c @@ -2292,8 +2292,10 @@ static void generic_hdmi_free(struct hda_codec *codec) struct hdmi_spec *spec = codec->spec; int pin_idx, pcm_idx; - if (codec_has_acomp(codec)) + if (codec_has_acomp(codec)) { snd_hdac_acomp_register_notifier(&codec->bus->core, NULL); + codec->relaxed_resume = 0; + } for (pin_idx = 0; pin_idx < spec->num_pins; pin_idx++) { struct hdmi_spec_per_pin *per_pin = get_pin(spec, pin_idx); @@ -2566,6 +2568,8 @@ static void register_i915_notifier(struct hda_codec *codec) spec->drm_audio_ops.pin_eld_notify = intel_pin_eld_notify; snd_hdac_acomp_register_notifier(&codec->bus->core, &spec->drm_audio_ops); + /* no need for forcible resume for jack check thanks to notifier */ + codec->relaxed_resume = 1; } /* setup_stream ops override for HSW+ */ -- 2.16.4

1 0

[alsa-devel] [PATCH 0/2] expand dshare to allow audio clock when not streaming data
by Adam Miartus 16 Jul '19

16 Jul '19

The general idea of these changes is to be able to open the device without defined bindings. This would start the audio clock without blocking a channel. This might be useful for hardware that requires running clock early at system start and to be available even without application streaming audio data. Andreas Pape (2): alsa: pcm: add unsupported OPS alsa: dshare: allow missing bindings src/pcm/Makefile.am | 4 +- src/pcm/pcm_dshare.c | 72 +++++++++++-- src/pcm/pcm_unsupported.c | 265 ++++++++++++++++++++++++++++++++++++++++++++++ src/pcm/pcm_unsupported.h | 167 +++++++++++++++++++++++++++++ 4 files changed, 495 insertions(+), 13 deletions(-) create mode 100644 src/pcm/pcm_unsupported.c create mode 100644 src/pcm/pcm_unsupported.h -- 2.7.4

3 7

[alsa-devel] [PATCH v2 0/6]
by vanitha.channaiah＠in.bosch.com 16 Jul '19

16 Jul '19

- Updated the patches by incorporating review comments from Takashi Iwai-san [PATCH 1/5] pcm: direct: Add generic hw_ptr_alignment function for dmix, dshare and dsnoop [Takashi Iwai:] > The patch description needs rephrasing. What actually this does is to move the code from pcm_dmix.c to pcm_direct.c > and its header so that the helper code can be re-used by other direct-PCM plugins. - Commit message is rephrased as suggested. - New commit: [PATCH v2 1/6] pcm: direct: Add generic hw_ptr_alignment function for [PATCH 2/5] pcm: dshare: Added "hw_ptr_alignment" option in configuration for alignment of slave pointers [Takashi Iwai:] > Again, this patch description is too ambiguous. > And, if it enables the hw_ptr_alignment option, update the documentation as well. - Commit message is explained in detail for the changes done. - Documentation updated. - New commit: [PATCH v2 2/6] pcm: dshare: Added "hw_ptr_alignment" option in [PATCH 3/5] pcm: dsnoop: Added "hw_ptr_alignment" option in configuration for slave pointer alignment [Takashi Iwai:] > Ditto as patch 2, the description is too ambiguous, and the update of documentation is missing. > It's not good to change the helper function semantics out of sudden, even without any description. - Commit message is explained in detail for the changes done. - Documentation updated. - Divided the patch with commit ("pcm: dsnoop: Add hw_ptr_alignment option in configuration") into additional patch commit ("pcm: direct: Round up of slave_app_ptr pointer if buffer") - Usecase scenario is described for the changes done in helper function. - New commit: [PATCH v2 3/6] pcm: dsnoop: Added "hw_ptr_alignment" option in [PATCH v2 4/6] pcm: direct: Round up of slave_app_ptr pointer if buffer [PATCH 4/5] pcm: restructuring sw params function [Takashi Iwai:] > I see no reason to do that. Please describe. - Commit message is explained in detail why reformating was done. - New commit: [PATCH v2 5/6] pcm: restructuring sw params function [PATCH 5/5] pcm: Update pcm->avail_min with needed_slave_avail_min, after reading unaligned frames [Takashi Iwai:] > This kind of changes in the core code should be avoided as much as possible, especially if it's only relevant with the specific plugins. > Sorry, this isn't convincing enough. If this is a MUST, please clarify better. - Commit message is explained in detail with the generic usecase and specific use case we came across. - New commit: [PATCH v2 6/6] pcm: Update pcm->avail_min with needed_slave_avail_min,

4 20

[alsa-devel] [PATCH] ALSA: seq: Break too long mutex context in the write loop
by Takashi Iwai 16 Jul '19

16 Jul '19

The fix for the racy writes and ioctls to sequencer widened the application of client->ioctl_mutex to the whole write loop. Although it does unlock/relock for the lengthy operation like the event dup, the loop keeps the ioctl_mutex for the whole time in other situations. This may take quite long time if the user-space would give a huge buffer, and this is a likely cause of some weird behavior spotted by syzcaller fuzzer. This patch puts a simple workaround, just adding a mutex break in the loop when a large number of events have been processed. This shouldn't hit any performance drop because the threshold is set high enough for usual operations. Fixes: 7bd800915677 ("ALSA: seq: More protection for concurrent write and ioctl races") Reported-by: syzbot+97aae04ce27e39cbfca9(a)syzkaller.appspotmail.com Reported-by: syzbot+4c595632b98bb8ffcc66(a)syzkaller.appspotmail.com Cc: <stable(a)vger.kernel.org> Signed-off-by: Takashi Iwai <tiwai(a)suse.de> --- sound/core/seq/seq_clientmgr.c | 11 ++++++++++- 1 file changed, 10 insertions(+), 1 deletion(-) diff --git a/sound/core/seq/seq_clientmgr.c b/sound/core/seq/seq_clientmgr.c index a60e7a17f0b8..7737b2670064 100644 --- a/sound/core/seq/seq_clientmgr.c +++ b/sound/core/seq/seq_clientmgr.c @@ -1021,7 +1021,7 @@ static ssize_t snd_seq_write(struct file *file, const char __user *buf, { struct snd_seq_client *client = file->private_data; int written = 0, len; - int err; + int err, handled; struct snd_seq_event event; if (!(snd_seq_file_flags(file) & SNDRV_SEQ_LFLG_OUTPUT)) @@ -1034,6 +1034,8 @@ static ssize_t snd_seq_write(struct file *file, const char __user *buf, if (!client->accept_output || client->pool == NULL) return -ENXIO; + repeat: + handled = 0; /* allocate the pool now if the pool is not allocated yet */ mutex_lock(&client->ioctl_mutex); if (client->pool->size > 0 && !snd_seq_write_pool_allocated(client)) { @@ -1093,12 +1095,19 @@ static ssize_t snd_seq_write(struct file *file, const char __user *buf, 0, 0, &client->ioctl_mutex); if (err < 0) break; + handled++; __skip_event: /* Update pointers and counts */ count -= len; buf += len; written += len; + + /* let's have a coffee break if too many events are queued */ + if (++handled >= 200) { + mutex_unlock(&client->ioctl_mutex); + goto repeat; + } } out: -- 2.16.4

1 0