kernel/sound/soc/sof, branch linux-6.2.y

ASoC: SOF: pm: Tear down pipelines only if DSP was active

2023-04-26T12:30:09Z

commit 0b186bb06198653d74a141902a7739e0bde20cf4 upstream. With PCI if the device was suspended it is brought back to full power and then suspended again. This doesn't happen when device is described via DT. We need to make sure that we tear down pipelines only if the device was previously active (thus the pipelines were setup). Otherwise, we can break the use_count: [ 219.009743] sof-audio-of-imx8m 3b6e8000.dsp: sof_ipc3_tear_down_all_pipelines: widget PIPELINE.2.SAI3.IN is still in use: count -1 and after this everything stops working. Fixes: d185e0689abc ("ASoC: SOF: pm: Always tear down pipelines before DSP suspend") Reviewed-by: Pierre-Louis Bossart Reviewed-by: Ranjani Sridharan Signed-off-by: Daniel Baluta Link: https://lore.kernel.org/r/20230405092655.19587-1-daniel.baluta@oss.nxp.com Signed-off-by: Mark Brown Signed-off-by: Greg Kroah-Hartman

ASoC: SOF: ipc4-topology: Clarify bind failure caused by missing fw_module

2023-04-26T12:30:04Z

commit de6aa72b265b72bca2b1897d5000c8f0147d3157 upstream. The original patch uses a feature in lib/vsprintf.c to handle the invalid address when tring to print *_fw_module->man4_module_entry.name when the *rc_fw_module is NULL. This case is handled by check_pointer_msg() internally and turns the invalid pointer to '(efault)' for printing but it is hiding useful information about the circumstances. Change the print to emmit the name of the widget and a note on which side's fw_module is missing. Fixes: e3720f92e023 ("ASoC: SOF: avoid a NULL dereference with unsupported widgets") Reported-by: Dan Carpenter Link: https://lore.kernel.org/alsa-devel/4826f662-42f0-4a82-ba32-8bf5f8a03256@kili.mountain/ Signed-off-by: Peter Ujfalusi Rule: 'Cc: stable@vger.kernel.org' or 'commit upstream.' Link: https://lore.kernel.org/r/20230403090909.18233-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Greg Kroah-Hartman

ASoC: SOF: avoid a NULL dereference with unsupported widgets

2023-04-13T15:02:52Z

commit e3720f92e0237921da537e47a0b24e27899203f8 upstream. If an IPC4 topology contains an unsupported widget, its .module_info field won't be set, then sof_ipc4_route_setup() will cause a kernel Oops trying to dereference it. Add a check for such cases. Cc: stable@vger.kernel.org # 6.2 Signed-off-by: Guennadi Liakhovetski Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230329113828.28562-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Greg Kroah-Hartman

ASoC: SOF: ipc4: Ensure DSP is in D0I0 during sof_ipc4_set_get_data()

2023-04-13T15:02:35Z

[ Upstream commit e51f49512d98783b90799c9cc2002895ec3aa0eb ] The set_get_data() IPC op bypasses the check for the no_pm flag as done with the regular IPC tx_msg op. Since set_get_data should be performed when the DSP is in D0I0, set the DSP power state to D0I0 before sending the IPC's in sof_ipc4_set_get_data(). Fixes: ceb89acc4dc8 ("ASoC: SOF: ipc4: Add support for mandatory message handling functionality") Signed-off-by: Ranjani Sridharan Reviewed-by: Bard Liao Reviewed-by: Péter Ujfalusi Reviewed-by: Pierre-Louis Bossart Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230322085538.10214-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: SOF: IPC4: update gain ipc msg definition to align with fw

2023-04-06T10:12:27Z

[ Upstream commit e45cd86c3a78bfb9875a5eb8ab5dab459b59bbe2 ] Recent firmware changes modified the curve duration from 32 to 64 bits, which breaks volume ramps. A simple solution would be to change the definition, but unfortunately the ASoC topology framework only supports up to 32 bit tokens. This patch suggests breaking the 64 bit value in low and high parts, with only the low-part extracted from topology and high-part only zeroes. Since the curve duration is represented in hundred of nanoseconds, we can still represent a 400s ramp, which is just fine. The defacto ABI change has no effect on existing users since the IPC4 firmware has not been released just yet. Link: https://github.com/thesofproject/linux/issues/4026 Signed-off-by: Rander Wang Reviewed-by: Ranjani Sridharan Reviewed-by: Pierre-Louis Bossart Reviewed-by: Bard Liao Reviewed-by: Péter Ujfalusi Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230307110656.1816-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: SOF: Intel: hda-ctrl: re-add sleep after entering and exiting reset

2023-04-06T10:12:27Z

[ Upstream commit 8bac40b8ed17ab1be9133e9620f65fae80262b7e ] This reverts commit a09d82ce0a867 ("ASoC: SOF: Intel: hda-ctrl: remove useless sleep") It was a mistake to remove those delays, in light of comments in the HDaudio spec captured in snd_hdac_bus_reset_link() that the codec needs time for its initialization and PLL lock. Signed-off-by: Pierre-Louis Bossart Reviewed-by: Ranjani Sridharan Reviewed-by: Péter Ujfalusi Reviewed-by: Rander Wang Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230307095412.3416-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: SOF: Intel: hda-dsp: harden D0i3 programming sequence

2023-04-06T10:12:26Z

[ Upstream commit 52a55779ed14792a150421339664193d6eb8e036 ] Add delay between set and wait command according to hardware programming sequence. Also add debug log to detect error. Signed-off-by: Rander Wang Reviewed-by: Péter Ujfalusi Reviewed-by: Pierre-Louis Bossart Reviewed-by: Péter Ujfalusi Reviewed-by: Ranjani Sridharan Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230307095453.3719-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: SOF: Intel: pci-tng: revert invalid bar size setting

2023-04-06T10:12:26Z

[ Upstream commit ca09e2a351fbc7836ba9418304ff0c3e72addfe0 ] The logic for the ioremap is to find the resource index 3 (IRAM) and infer the BAR address by subtracting the IRAM offset. The BAR size defined in hardware specifications is 2MB. The commit 5947b2726beb6 ("ASoC: SOF: Intel: Check the bar size before remapping") tried to find the BAR size by querying the resource length instead of a pre-canned value, but by requesting the size for index 3 it only gets the size of the IRAM. That's obviously wrong and prevents the probe from proceeding. This commit attempted to fix an issue in a fuzzing/simulated environment but created another on actual devices, so the best course of action is to revert that change. Reported-by: Ferry Toth Tested-by: Ferry Toth (Intel Edison-Arduino) Link: https://github.com/thesofproject/linux/issues/3901 Signed-off-by: Pierre-Louis Bossart Reviewed-by: Péter Ujfalusi Reviewed-by: Ranjani Sridharan Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230307095341.3222-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: SOF: ipc4-topology: Fix incorrect sample rate print unit

2023-04-06T10:12:26Z

[ Upstream commit 9e269e3aa9006440de639597079ee7140ef5b5f3 ] This patch fixes the sample rate print unit from KHz to Hz. E.g. 48000KHz becomes 48000Hz. Signed-off-by: Seppo Ingalsuo Reviewed-by: Pierre-Louis Bossart Reviewed-by: Ranjani Sridharan Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230307110751.2053-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin

ASoC: SOF: ipc3: Check for upper size limit for the received message

2023-04-06T10:12:26Z

[ Upstream commit 989a3e4479177d0f4afab8be1960731bc0ffbbd0 ] The sof_ipc3_rx_msg() checks for minimum size of a new rx message but it is missing the check for upper limit. Corrupted or compromised firmware might be able to take advantage of this to cause out of bounds reads outside of the message area. Reported-by: Curtis Malainey Signed-off-by: Peter Ujfalusi Reviewed-by: Pierre-Louis Bossart Reviewed-by: Curtis Malainey Signed-off-by: Peter Ujfalusi Link: https://lore.kernel.org/r/20230307114917.5124-1-peter.ujfalusi@linux.intel.com Signed-off-by: Mark Brown Signed-off-by: Sasha Levin