[PATCH v2 0/5] ASoC: Intel: Balance ACPI device refcount
While looking for the open coded put_device(&adev->dev) cases, where adev stands for ACPI device, I noticed that in a few ASoC Intel driver, among others, the refcount is not balanced properly in some cases.
This series fixes that issue and converts to use acpi_dev_put().
Changelog v2: - split Intel drivers out from others (Pierre) - sent with cover letter (Mark)
Andy Shevchenko (5): ASoC: Intel: bytcht_es8316: Drop reference count of ACPI device after use ASoC: Intel: bytcr_rt5651: Drop reference count of ACPI device after use ASoC: Intel: bytcr_rt5640: Drop reference count of ACPI device after use ASoC: Intel: bytcr_wm5102: Drop reference count of ACPI device after use ASoC: Intel: sof_es8336: Drop reference count of ACPI device after use
sound/soc/intel/boards/bytcht_es8316.c | 20 ++++++++++++-------- sound/soc/intel/boards/bytcr_rt5640.c | 12 ++++++------ sound/soc/intel/boards/bytcr_rt5651.c | 2 +- sound/soc/intel/boards/bytcr_wm5102.c | 2 +- sound/soc/intel/boards/sof_es8336.c | 14 ++++++++------ 5 files changed, 28 insertions(+), 22 deletions(-)
Theoretically the device might gone if its reference count drops to 0. This might be the case when we try to find the first physical node of the ACPI device. We need to keep reference to it until we get a result of the above mentioned call. Refactor the code to drop the reference count at the correct place.
While at it, move to acpi_dev_put() as symmetrical call to the acpi_dev_get_first_match_dev().
Fixes: 3c22a73fb873 ("ASoC: Intel: bytcht_es8316: fix HID handling") Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com --- sound/soc/intel/boards/bytcht_es8316.c | 20 ++++++++++++-------- 1 file changed, 12 insertions(+), 8 deletions(-)
diff --git a/sound/soc/intel/boards/bytcht_es8316.c b/sound/soc/intel/boards/bytcht_es8316.c index 09d1f0f6d686..df157b01df8b 100644 --- a/sound/soc/intel/boards/bytcht_es8316.c +++ b/sound/soc/intel/boards/bytcht_es8316.c @@ -497,21 +497,28 @@ static int snd_byt_cht_es8316_mc_probe(struct platform_device *pdev) if (adev) { snprintf(codec_name, sizeof(codec_name), "i2c-%s", acpi_dev_name(adev)); - put_device(&adev->dev); byt_cht_es8316_dais[dai_index].codecs->name = codec_name; } else { dev_err(dev, "Error cannot find '%s' dev\n", mach->id); return -ENXIO; }
+ codec_dev = acpi_get_first_physical_node(adev); + acpi_dev_put(adev); + if (!codec_dev) + return -EPROBE_DEFER; + priv->codec_dev = get_device(codec_dev); + /* override platform name, if required */ byt_cht_es8316_card.dev = dev; platform_name = mach->mach_params.platform;
ret = snd_soc_fixup_dai_links_platform_name(&byt_cht_es8316_card, platform_name); - if (ret) + if (ret) { + put_device(codec_dev); return ret; + }
/* Check for BYTCR or other platform and setup quirks */ dmi_id = dmi_first_match(byt_cht_es8316_quirk_table); @@ -539,13 +546,10 @@ static int snd_byt_cht_es8316_mc_probe(struct platform_device *pdev)
/* get the clock */ priv->mclk = devm_clk_get(dev, "pmc_plt_clk_3"); - if (IS_ERR(priv->mclk)) + if (IS_ERR(priv->mclk)) { + put_device(codec_dev); return dev_err_probe(dev, PTR_ERR(priv->mclk), "clk_get pmc_plt_clk_3 failed\n"); - - codec_dev = acpi_get_first_physical_node(adev); - if (!codec_dev) - return -EPROBE_DEFER; - priv->codec_dev = get_device(codec_dev); + }
if (quirk & BYT_CHT_ES8316_JD_INVERTED) props[cnt++] = PROPERTY_ENTRY_BOOL("everest,jack-detect-inverted");
Theoretically the device might gone if its reference count drops to 0. This might be the case when we try to find the first physical node of the ACPI device. We need to keep reference to it until we get a result of the above mentioned call. Refactor the code to drop the reference count at the correct place.
While at it, move to acpi_dev_put() as symmetrical call to the acpi_dev_get_first_match_dev().
Fixes: 02c0a3b3047f ("ASoC: Intel: bytcr_rt5651: add MCLK, quirks and cleanups") Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com --- sound/soc/intel/boards/bytcr_rt5651.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/soc/intel/boards/bytcr_rt5651.c b/sound/soc/intel/boards/bytcr_rt5651.c index 81ac6eeda2e6..8fca9b82d4d0 100644 --- a/sound/soc/intel/boards/bytcr_rt5651.c +++ b/sound/soc/intel/boards/bytcr_rt5651.c @@ -922,7 +922,6 @@ static int snd_byt_rt5651_mc_probe(struct platform_device *pdev) if (adev) { snprintf(byt_rt5651_codec_name, sizeof(byt_rt5651_codec_name), "i2c-%s", acpi_dev_name(adev)); - put_device(&adev->dev); byt_rt5651_dais[dai_index].codecs->name = byt_rt5651_codec_name; } else { dev_err(dev, "Error cannot find '%s' dev\n", mach->id); @@ -930,6 +929,7 @@ static int snd_byt_rt5651_mc_probe(struct platform_device *pdev) }
codec_dev = acpi_get_first_physical_node(adev); + acpi_dev_put(adev); if (!codec_dev) return -EPROBE_DEFER; priv->codec_dev = get_device(codec_dev);
Theoretically the device might gone if its reference count drops to 0. This might be the case when we try to find the first physical node of the ACPI device. We need to keep reference to it until we get a result of the above mentioned call. Refactor the code to drop the reference count at the correct place.
While at it, move to acpi_dev_put() as symmetrical call to the acpi_dev_get_first_match_dev().
Fixes: a232b96dcece ("ASoC: Intel: bytcr_rt5640: use HID translation util") Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com --- sound/soc/intel/boards/bytcr_rt5640.c | 12 ++++++------ 1 file changed, 6 insertions(+), 6 deletions(-)
diff --git a/sound/soc/intel/boards/bytcr_rt5640.c b/sound/soc/intel/boards/bytcr_rt5640.c index 4699ca79f3ea..79e0039c79a3 100644 --- a/sound/soc/intel/boards/bytcr_rt5640.c +++ b/sound/soc/intel/boards/bytcr_rt5640.c @@ -1636,13 +1636,18 @@ static int snd_byt_rt5640_mc_probe(struct platform_device *pdev) if (adev) { snprintf(byt_rt5640_codec_name, sizeof(byt_rt5640_codec_name), "i2c-%s", acpi_dev_name(adev)); - put_device(&adev->dev); byt_rt5640_dais[dai_index].codecs->name = byt_rt5640_codec_name; } else { dev_err(dev, "Error cannot find '%s' dev\n", mach->id); return -ENXIO; }
+ codec_dev = acpi_get_first_physical_node(adev); + acpi_dev_put(adev); + if (!codec_dev) + return -EPROBE_DEFER; + priv->codec_dev = get_device(codec_dev); + /* * swap SSP0 if bytcr is detected * (will be overridden if DMI quirk is detected) @@ -1717,11 +1722,6 @@ static int snd_byt_rt5640_mc_probe(struct platform_device *pdev) byt_rt5640_quirk = quirk_override; }
- codec_dev = acpi_get_first_physical_node(adev); - if (!codec_dev) - return -EPROBE_DEFER; - priv->codec_dev = get_device(codec_dev); - if (byt_rt5640_quirk & BYT_RT5640_JD_HP_ELITEP_1000G2) { acpi_dev_add_driver_gpios(ACPI_COMPANION(priv->codec_dev), byt_rt5640_hp_elitepad_1000g2_gpios);
Theoretically the device might gone if its reference count drops to 0. This might be the case when we try to find the first physical node of the ACPI device. We need to keep reference to it until we get a result of the above mentioned call. Refactor the code to drop the reference count at the correct place.
While at it, move to acpi_dev_put() as symmetrical call to the acpi_dev_get_first_match_dev().
Fixes: 9a87fc1e0619 ("ASoC: Intel: bytcr_wm5102: Add machine driver for BYT/WM5102") Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com --- sound/soc/intel/boards/bytcr_wm5102.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/sound/soc/intel/boards/bytcr_wm5102.c b/sound/soc/intel/boards/bytcr_wm5102.c index 1669eb3bd80f..c0706537f673 100644 --- a/sound/soc/intel/boards/bytcr_wm5102.c +++ b/sound/soc/intel/boards/bytcr_wm5102.c @@ -411,9 +411,9 @@ static int snd_byt_wm5102_mc_probe(struct platform_device *pdev) return -ENOENT; } snprintf(codec_name, sizeof(codec_name), "spi-%s", acpi_dev_name(adev)); - put_device(&adev->dev);
codec_dev = bus_find_device_by_name(&spi_bus_type, NULL, codec_name); + acpi_dev_put(adev); if (!codec_dev) return -EPROBE_DEFER;
Theoretically the device might gone if its reference count drops to 0. This might be the case when we try to find the first physical node of the ACPI device. We need to keep reference to it until we get a result of the above mentioned call. Refactor the code to drop the reference count at the correct place.
While at it, move to acpi_dev_put() as symmetrical call to the acpi_dev_get_first_match_dev().
Fixes: a164137ce91a ("ASoC: Intel: add machine driver for SOF+ES8336") Signed-off-by: Andy Shevchenko andriy.shevchenko@linux.intel.com --- sound/soc/intel/boards/sof_es8336.c | 14 ++++++++------ 1 file changed, 8 insertions(+), 6 deletions(-)
diff --git a/sound/soc/intel/boards/sof_es8336.c b/sound/soc/intel/boards/sof_es8336.c index 773e5d1d87d4..894b6610b9e2 100644 --- a/sound/soc/intel/boards/sof_es8336.c +++ b/sound/soc/intel/boards/sof_es8336.c @@ -681,7 +681,6 @@ static int sof_es8336_probe(struct platform_device *pdev) if (adev) { snprintf(codec_name, sizeof(codec_name), "i2c-%s", acpi_dev_name(adev)); - put_device(&adev->dev); dai_links[0].codecs->name = codec_name;
/* also fixup codec dai name if relevant */ @@ -692,16 +691,19 @@ static int sof_es8336_probe(struct platform_device *pdev) return -ENXIO; }
- ret = snd_soc_fixup_dai_links_platform_name(&sof_es8336_card, - mach->mach_params.platform); - if (ret) - return ret; - codec_dev = acpi_get_first_physical_node(adev); + acpi_dev_put(adev); if (!codec_dev) return -EPROBE_DEFER; priv->codec_dev = get_device(codec_dev);
+ ret = snd_soc_fixup_dai_links_platform_name(&sof_es8336_card, + mach->mach_params.platform); + if (ret) { + put_device(codec_dev); + return ret; + } + if (quirk & SOF_ES8336_JD_INVERTED) props[cnt++] = PROPERTY_ENTRY_BOOL("everest,jack-detect-inverted");
On 1/12/23 05:28, Andy Shevchenko wrote:
While looking for the open coded put_device(&adev->dev) cases, where adev stands for ACPI device, I noticed that in a few ASoC Intel driver, among others, the refcount is not balanced properly in some cases.
This series fixes that issue and converts to use acpi_dev_put().
Thanks Andy!
Acked-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com
Changelog v2:
- split Intel drivers out from others (Pierre)
- sent with cover letter (Mark)
Andy Shevchenko (5): ASoC: Intel: bytcht_es8316: Drop reference count of ACPI device after use ASoC: Intel: bytcr_rt5651: Drop reference count of ACPI device after use ASoC: Intel: bytcr_rt5640: Drop reference count of ACPI device after use ASoC: Intel: bytcr_wm5102: Drop reference count of ACPI device after use ASoC: Intel: sof_es8336: Drop reference count of ACPI device after use
sound/soc/intel/boards/bytcht_es8316.c | 20 ++++++++++++-------- sound/soc/intel/boards/bytcr_rt5640.c | 12 ++++++------ sound/soc/intel/boards/bytcr_rt5651.c | 2 +- sound/soc/intel/boards/bytcr_wm5102.c | 2 +- sound/soc/intel/boards/sof_es8336.c | 14 ++++++++------ 5 files changed, 28 insertions(+), 22 deletions(-)
On Thu, 12 Jan 2023 13:28:47 +0200, Andy Shevchenko wrote:
While looking for the open coded put_device(&adev->dev) cases, where adev stands for ACPI device, I noticed that in a few ASoC Intel driver, among others, the refcount is not balanced properly in some cases.
This series fixes that issue and converts to use acpi_dev_put().
Changelog v2:
- split Intel drivers out from others (Pierre)
- sent with cover letter (Mark)
[...]
Applied to
https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next
Thanks!
[1/5] ASoC: Intel: bytcht_es8316: Drop reference count of ACPI device after use commit: 6b1c0bd6fdefbf3c3d75680c2708f5423ef72e46 [2/5] ASoC: Intel: bytcr_rt5651: Drop reference count of ACPI device after use commit: 721858823d7cdc8f2a897579b040e935989f6f02 [3/5] ASoC: Intel: bytcr_rt5640: Drop reference count of ACPI device after use commit: cbf87bcf46e399e9a5288430d940efbad3551c68 [4/5] ASoC: Intel: bytcr_wm5102: Drop reference count of ACPI device after use commit: c8aa49abdeda2ab587aadb083e670f6aa0236f93 [5/5] ASoC: Intel: sof_es8336: Drop reference count of ACPI device after use commit: 64e57b2195725c1ae2246a8a2ce224abb60620ac
All being well this means that it will be integrated into the linux-next tree (usually sometime in the next 24 hours) and sent to Linus during the next merge window (or sooner if it is a bug fix), however if problems are discovered then the patch may be dropped or reverted.
You may get further e-mails resulting from automated or manual testing and review of the tree, please engage with people reporting problems and send followup patches addressing any issues that are reported if needed.
If any updates are required or you are submitting further changes they should be sent as incremental updates against current git, existing patches will not be replaced.
Please add any relevant lists and maintainers to the CCs when replying to this mail.
Thanks, Mark
participants (3)
-
Andy Shevchenko -
Mark Brown -
Pierre-Louis Bossart