| Message ID | 20221110012532.8796-1-hui.wang@canonical.com |
|---|---|
| Headers | show |
| Series | alsa: soc: the kernel print UBSAN calltrace on the machine with cs35l41 codec | expand |
On 11/9/22 6:25 PM, Hui Wang wrote: > BugLink: https://bugs.launchpad.net/bugs/1996121 > > The cs35l41 codec driver was backported to Jammy kernel a couple of > months ago, recently the Jammy kernel was adopted as the hwe kernel > for OEM machines and with this kernel the UBSAN calltrace was seen, > need to backport these 2 patches. Kinetic kernel already has the > patches, so this SRU is only for Jammy kernel. > > [Impact] > Install the 5.15 hwe kernel on the machines with cs35l41 > codec, after booting and check the dmesg, there is a UBSAN calltrace. > > [Fix] > Backport 2 patches from mainline kernel to fix this problem. > > [Test] > boot the patched kernel on the machine, check the dmesg, there > is no UBSAN calltrace, and test the audio functions, all input > and output devices coudl work as well as before. > > > [Where problems could occur] > The patch only affects the cirruslogic asoc codec driver, if it > could introduce the regression, it will be on the machines with > cs codec, and it will make the audio driver initialization fail > and make the system lose the audio function. But this possibility > is very low, we tested the patches on many oem machines, all > worked well. > > > Hui Wang (2): > ASoC: cs35l41: Add one more variable in the debug log > ASoC: cs35l41: Fix an out-of-bounds access in otp_packed_element_t > > include/sound/cs35l41.h | 1 - > sound/soc/codecs/cs35l41-lib.c | 18 +++++++++--------- > 2 files changed, 9 insertions(+), 10 deletions(-) > Acked-by: Tim Gardner <tim.gardner@canonical.com>
On 10.11.22 02:25, Hui Wang wrote: > BugLink: https://bugs.launchpad.net/bugs/1996121 > > The cs35l41 codec driver was backported to Jammy kernel a couple of > months ago, recently the Jammy kernel was adopted as the hwe kernel > for OEM machines and with this kernel the UBSAN calltrace was seen, > need to backport these 2 patches. Kinetic kernel already has the > patches, so this SRU is only for Jammy kernel. > > [Impact] > Install the 5.15 hwe kernel on the machines with cs35l41 > codec, after booting and check the dmesg, there is a UBSAN calltrace. > > [Fix] > Backport 2 patches from mainline kernel to fix this problem. > > [Test] > boot the patched kernel on the machine, check the dmesg, there > is no UBSAN calltrace, and test the audio functions, all input > and output devices coudl work as well as before. > > > [Where problems could occur] > The patch only affects the cirruslogic asoc codec driver, if it > could introduce the regression, it will be on the machines with > cs codec, and it will make the audio driver initialization fail > and make the system lose the audio function. But this possibility > is very low, we tested the patches on many oem machines, all > worked well. > > > Hui Wang (2): > ASoC: cs35l41: Add one more variable in the debug log > ASoC: cs35l41: Fix an out-of-bounds access in otp_packed_element_t > > include/sound/cs35l41.h | 1 - > sound/soc/codecs/cs35l41-lib.c | 18 +++++++++--------- > 2 files changed, 9 insertions(+), 10 deletions(-) > Acked-by: Stefan Bader <stefan.bader@canonical.com>
On 10.11.22 02:25, Hui Wang wrote: > BugLink: https://bugs.launchpad.net/bugs/1996121 > > The cs35l41 codec driver was backported to Jammy kernel a couple of > months ago, recently the Jammy kernel was adopted as the hwe kernel > for OEM machines and with this kernel the UBSAN calltrace was seen, > need to backport these 2 patches. Kinetic kernel already has the > patches, so this SRU is only for Jammy kernel. > > [Impact] > Install the 5.15 hwe kernel on the machines with cs35l41 > codec, after booting and check the dmesg, there is a UBSAN calltrace. > > [Fix] > Backport 2 patches from mainline kernel to fix this problem. > > [Test] > boot the patched kernel on the machine, check the dmesg, there > is no UBSAN calltrace, and test the audio functions, all input > and output devices coudl work as well as before. > > > [Where problems could occur] > The patch only affects the cirruslogic asoc codec driver, if it > could introduce the regression, it will be on the machines with > cs codec, and it will make the audio driver initialization fail > and make the system lose the audio function. But this possibility > is very low, we tested the patches on many oem machines, all > worked well. > > > Hui Wang (2): > ASoC: cs35l41: Add one more variable in the debug log > ASoC: cs35l41: Fix an out-of-bounds access in otp_packed_element_t > > include/sound/cs35l41.h | 1 - > sound/soc/codecs/cs35l41-lib.c | 18 +++++++++--------- > 2 files changed, 9 insertions(+), 10 deletions(-) > Applied to jammy:linux/master-next. Thanks. -Stefan