[PATCH AUTOSEL 6.0 07/44] ALSA: hda: Fix page fault in snd_hda_codec_shutdown()
From: Cezary Rojewski cezary.rojewski@intel.com
[ Upstream commit f2bd1c5ae2cb0cf9525c9bffc0038c12dd7e1338 ]
If early probe of HDAudio bus driver fails e.g.: due to missing firmware file, snd_hda_codec_shutdown() ends in manipulating uninitialized codec->pcm_list_head causing page fault.
Initialization of HDAudio codec in ASoC is split in two: - snd_hda_codec_device_init() - snd_hda_codec_device_new()
snd_hda_codec_device_init() is called during probe_codecs() by HDAudio bus driver while snd_hda_codec_device_new() is called by codec-component's ->probe(). The second call will not happen until all components required by related sound card are present within the ASoC framework. With firmware failing to load during the PCI's deferred initialization i.e.: probe_work(), no platform components are ever registered. HDAudio codec enumeration is done at that point though, so the codec components became registered to ASoC framework, calling snd_hda_codec_device_init() in the process.
Now, during platform reboot snd_hda_codec_shutdown() is called for every codec found on the HDAudio bus causing oops if any of them has not completed both of their initialization steps. Relocating field initialization fixes the issue.
Reviewed-by: Kai Vehmanen kai.vehmanen@linux.intel.com Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com Signed-off-by: Cezary Rojewski cezary.rojewski@intel.com Link: https://lore.kernel.org/r/20220816111727.3218543-7-cezary.rojewski@intel.com Signed-off-by: Takashi Iwai tiwai@suse.de Signed-off-by: Sasha Levin sashal@kernel.org --- sound/pci/hda/hda_codec.c | 41 +++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 21 deletions(-)
diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 384426d7e9dd..4ae8b9574778 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -931,8 +931,28 @@ snd_hda_codec_device_init(struct hda_bus *bus, unsigned int codec_addr, }
codec->bus = bus; + codec->depop_delay = -1; + codec->fixup_id = HDA_FIXUP_ID_NOT_SET; + codec->core.dev.release = snd_hda_codec_dev_release; + codec->core.exec_verb = codec_exec_verb; codec->core.type = HDA_DEV_LEGACY;
+ mutex_init(&codec->spdif_mutex); + mutex_init(&codec->control_mutex); + snd_array_init(&codec->mixers, sizeof(struct hda_nid_item), 32); + snd_array_init(&codec->nids, sizeof(struct hda_nid_item), 32); + snd_array_init(&codec->init_pins, sizeof(struct hda_pincfg), 16); + snd_array_init(&codec->driver_pins, sizeof(struct hda_pincfg), 16); + snd_array_init(&codec->cvt_setups, sizeof(struct hda_cvt_setup), 8); + snd_array_init(&codec->spdif_out, sizeof(struct hda_spdif_out), 16); + snd_array_init(&codec->jacktbl, sizeof(struct hda_jack_tbl), 16); + snd_array_init(&codec->verbs, sizeof(struct hda_verb *), 8); + INIT_LIST_HEAD(&codec->conn_list); + INIT_LIST_HEAD(&codec->pcm_list_head); + INIT_DELAYED_WORK(&codec->jackpoll_work, hda_jackpoll_work); + refcount_set(&codec->pcm_ref, 1); + init_waitqueue_head(&codec->remove_sleep); + return codec; } EXPORT_SYMBOL_GPL(snd_hda_codec_device_init); @@ -985,29 +1005,8 @@ int snd_hda_codec_device_new(struct hda_bus *bus, struct snd_card *card, if (snd_BUG_ON(codec_addr > HDA_MAX_CODEC_ADDRESS)) return -EINVAL;
- codec->core.dev.release = snd_hda_codec_dev_release; - codec->core.exec_verb = codec_exec_verb; - codec->card = card; codec->addr = codec_addr; - mutex_init(&codec->spdif_mutex); - mutex_init(&codec->control_mutex); - snd_array_init(&codec->mixers, sizeof(struct hda_nid_item), 32); - snd_array_init(&codec->nids, sizeof(struct hda_nid_item), 32); - snd_array_init(&codec->init_pins, sizeof(struct hda_pincfg), 16); - snd_array_init(&codec->driver_pins, sizeof(struct hda_pincfg), 16); - snd_array_init(&codec->cvt_setups, sizeof(struct hda_cvt_setup), 8); - snd_array_init(&codec->spdif_out, sizeof(struct hda_spdif_out), 16); - snd_array_init(&codec->jacktbl, sizeof(struct hda_jack_tbl), 16); - snd_array_init(&codec->verbs, sizeof(struct hda_verb *), 8); - INIT_LIST_HEAD(&codec->conn_list); - INIT_LIST_HEAD(&codec->pcm_list_head); - refcount_set(&codec->pcm_ref, 1); - init_waitqueue_head(&codec->remove_sleep); - - INIT_DELAYED_WORK(&codec->jackpoll_work, hda_jackpoll_work); - codec->depop_delay = -1; - codec->fixup_id = HDA_FIXUP_ID_NOT_SET;
#ifdef CONFIG_PM codec->power_jiffies = jiffies;
On 10/9/22 18:48, Sasha Levin wrote:
From: Cezary Rojewski cezary.rojewski@intel.com
[ Upstream commit f2bd1c5ae2cb0cf9525c9bffc0038c12dd7e1338 ]
This commit on linux-stable seems to have broken a number of platforms.
6.0.2 worked fine. 6.0.3 does not
reverting this commit solves the problem, see https://github.com/thesofproject/linux/issues/3960 for details.
Are we missing a prerequisite patch for this commit?
If early probe of HDAudio bus driver fails e.g.: due to missing firmware file, snd_hda_codec_shutdown() ends in manipulating uninitialized codec->pcm_list_head causing page fault.
Initialization of HDAudio codec in ASoC is split in two:
- snd_hda_codec_device_init()
- snd_hda_codec_device_new()
snd_hda_codec_device_init() is called during probe_codecs() by HDAudio bus driver while snd_hda_codec_device_new() is called by codec-component's ->probe(). The second call will not happen until all components required by related sound card are present within the ASoC framework. With firmware failing to load during the PCI's deferred initialization i.e.: probe_work(), no platform components are ever registered. HDAudio codec enumeration is done at that point though, so the codec components became registered to ASoC framework, calling snd_hda_codec_device_init() in the process.
Now, during platform reboot snd_hda_codec_shutdown() is called for every codec found on the HDAudio bus causing oops if any of them has not completed both of their initialization steps. Relocating field initialization fixes the issue.
Reviewed-by: Kai Vehmanen kai.vehmanen@linux.intel.com Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com Signed-off-by: Cezary Rojewski cezary.rojewski@intel.com Link: https://lore.kernel.org/r/20220816111727.3218543-7-cezary.rojewski@intel.com Signed-off-by: Takashi Iwai tiwai@suse.de Signed-off-by: Sasha Levin sashal@kernel.org
sound/pci/hda/hda_codec.c | 41 +++++++++++++++++++-------------------- 1 file changed, 20 insertions(+), 21 deletions(-)
diff --git a/sound/pci/hda/hda_codec.c b/sound/pci/hda/hda_codec.c index 384426d7e9dd..4ae8b9574778 100644 --- a/sound/pci/hda/hda_codec.c +++ b/sound/pci/hda/hda_codec.c @@ -931,8 +931,28 @@ snd_hda_codec_device_init(struct hda_bus *bus, unsigned int codec_addr, }
codec->bus = bus;
codec->depop_delay = -1;
codec->fixup_id = HDA_FIXUP_ID_NOT_SET;
codec->core.dev.release = snd_hda_codec_dev_release;
codec->core.exec_verb = codec_exec_verb; codec->core.type = HDA_DEV_LEGACY;
mutex_init(&codec->spdif_mutex);
mutex_init(&codec->control_mutex);
snd_array_init(&codec->mixers, sizeof(struct hda_nid_item), 32);
snd_array_init(&codec->nids, sizeof(struct hda_nid_item), 32);
snd_array_init(&codec->init_pins, sizeof(struct hda_pincfg), 16);
snd_array_init(&codec->driver_pins, sizeof(struct hda_pincfg), 16);
snd_array_init(&codec->cvt_setups, sizeof(struct hda_cvt_setup), 8);
snd_array_init(&codec->spdif_out, sizeof(struct hda_spdif_out), 16);
snd_array_init(&codec->jacktbl, sizeof(struct hda_jack_tbl), 16);
snd_array_init(&codec->verbs, sizeof(struct hda_verb *), 8);
INIT_LIST_HEAD(&codec->conn_list);
INIT_LIST_HEAD(&codec->pcm_list_head);
INIT_DELAYED_WORK(&codec->jackpoll_work, hda_jackpoll_work);
refcount_set(&codec->pcm_ref, 1);
init_waitqueue_head(&codec->remove_sleep);
return codec;
} EXPORT_SYMBOL_GPL(snd_hda_codec_device_init); @@ -985,29 +1005,8 @@ int snd_hda_codec_device_new(struct hda_bus *bus, struct snd_card *card, if (snd_BUG_ON(codec_addr > HDA_MAX_CODEC_ADDRESS)) return -EINVAL;
- codec->core.dev.release = snd_hda_codec_dev_release;
- codec->core.exec_verb = codec_exec_verb;
- codec->card = card; codec->addr = codec_addr;
- mutex_init(&codec->spdif_mutex);
- mutex_init(&codec->control_mutex);
- snd_array_init(&codec->mixers, sizeof(struct hda_nid_item), 32);
- snd_array_init(&codec->nids, sizeof(struct hda_nid_item), 32);
- snd_array_init(&codec->init_pins, sizeof(struct hda_pincfg), 16);
- snd_array_init(&codec->driver_pins, sizeof(struct hda_pincfg), 16);
- snd_array_init(&codec->cvt_setups, sizeof(struct hda_cvt_setup), 8);
- snd_array_init(&codec->spdif_out, sizeof(struct hda_spdif_out), 16);
- snd_array_init(&codec->jacktbl, sizeof(struct hda_jack_tbl), 16);
- snd_array_init(&codec->verbs, sizeof(struct hda_verb *), 8);
- INIT_LIST_HEAD(&codec->conn_list);
- INIT_LIST_HEAD(&codec->pcm_list_head);
- refcount_set(&codec->pcm_ref, 1);
- init_waitqueue_head(&codec->remove_sleep);
- INIT_DELAYED_WORK(&codec->jackpoll_work, hda_jackpoll_work);
- codec->depop_delay = -1;
- codec->fixup_id = HDA_FIXUP_ID_NOT_SET;
#ifdef CONFIG_PM codec->power_jiffies = jiffies;
On Tue, Oct 25, 2022 at 09:27:32AM -0500, Pierre-Louis Bossart wrote:
On 10/9/22 18:48, Sasha Levin wrote:
From: Cezary Rojewski cezary.rojewski@intel.com
[ Upstream commit f2bd1c5ae2cb0cf9525c9bffc0038c12dd7e1338 ]
This commit on linux-stable seems to have broken a number of platforms.
6.0.2 worked fine. 6.0.3 does not
reverting this commit solves the problem, see https://github.com/thesofproject/linux/issues/3960 for details.
Are we missing a prerequisite patch for this commit?
Please see https://lore.kernel.org/r/20221024143931.15722-1-tiwai@suse.de
Does that solve it for you?
thanks,
greg k-h
On 10/25/22 09:50, Greg KH wrote:
On Tue, Oct 25, 2022 at 09:27:32AM -0500, Pierre-Louis Bossart wrote:
On 10/9/22 18:48, Sasha Levin wrote:
From: Cezary Rojewski cezary.rojewski@intel.com
[ Upstream commit f2bd1c5ae2cb0cf9525c9bffc0038c12dd7e1338 ]
This commit on linux-stable seems to have broken a number of platforms.
6.0.2 worked fine. 6.0.3 does not
reverting this commit solves the problem, see https://github.com/thesofproject/linux/issues/3960 for details.
Are we missing a prerequisite patch for this commit?
Please see https://lore.kernel.org/r/20221024143931.15722-1-tiwai@suse.de
Does that solve it for you?
Yep, that's the revert I tested.
participants (3)
-
Greg KH
-
Pierre-Louis Bossart
-
Sasha Levin