On Tue, Dec 13, 2022, at 08:48, Naresh Kamboju wrote:
On Mon, 12 Dec 2022 at 18:43, Greg Kroah-Hartman gregkh@linuxfoundation.org wrote:
Regression detected on arm64 Raspberry Pi 4 Model B the NFS mount failed.
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
Following changes have been noticed in the Kconfig file between good and bad. The config files attached to this email.
-CONFIG_BCMGENET=y -CONFIG_BROADCOM_PHY=y +# CONFIG_BROADCOM_PHY is not set -CONFIG_BCM7XXX_PHY=y +# CONFIG_BCM7XXX_PHY is not set -CONFIG_BCM_NET_PHYLIB=y
Full test log details,
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
Where does the kernel configuration come from? Is this a plain defconfig that used to work, or do you have a board specific config file?
This is most likely caused by the added dependency on CONFIG_PTP_1588_CLOCK that would lead to the BCMGENET driver not being built-in if PTP support is in a module.
Arnd
On Tue, 13 Dec 2022 at 14:50, Arnd Bergmann arnd@arndb.de wrote:
On Tue, Dec 13, 2022, at 08:48, Naresh Kamboju wrote:
On Mon, 12 Dec 2022 at 18:43, Greg Kroah-Hartman gregkh@linuxfoundation.org wrote:
Regression detected on arm64 Raspberry Pi 4 Model B the NFS mount failed.
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
Following changes have been noticed in the Kconfig file between good and bad. The config files attached to this email.
-CONFIG_BCMGENET=y -CONFIG_BROADCOM_PHY=y +# CONFIG_BROADCOM_PHY is not set -CONFIG_BCM7XXX_PHY=y +# CONFIG_BCM7XXX_PHY is not set -CONFIG_BCM_NET_PHYLIB=y
Full test log details,
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
Where does the kernel configuration come from? Is this a plain defconfig that used to work, or do you have a board specific config file?
This is most likely caused by the added dependency on CONFIG_PTP_1588_CLOCK that would lead to the BCMGENET driver not being built-in if PTP support is in a module.
Here is the build command which is the same for working and not working kernels.
# tuxmake --runtime podman --target-arch arm64 --toolchain gcc-11 --kconfig defconfig --kconfig-add https://raw.githubusercontent.com/Linaro/meta-lkft/kirkstone/meta/recipes-ke... --kconfig-add https://raw.githubusercontent.com/Linaro/meta-lkft/kirkstone/meta/recipes-ke... --kconfig-add https://raw.githubusercontent.com/Linaro/meta-lkft/kirkstone/meta/recipes-ke... --kconfig-add https://raw.githubusercontent.com/Linaro/meta-lkft/kirkstone/meta/recipes-ke... --kconfig-add https://raw.githubusercontent.com/Linaro/meta-lkft/kirkstone/meta/recipes-ke... --kconfig-add CONFIG_ARM64_MODULE_PLTS=y --kconfig-add CONFIG_SYN_COOKIES=y --kconfig-add CONFIG_SCHEDSTATS=y CROSS_COMPILE_COMPAT=arm-linux-gnueabihf-
- Naresh
Arnd
On Tue 2022-12-13 10:20:30, Arnd Bergmann wrote:
On Tue, Dec 13, 2022, at 08:48, Naresh Kamboju wrote:
On Mon, 12 Dec 2022 at 18:43, Greg Kroah-Hartman gregkh@linuxfoundation.org wrote:
Regression detected on arm64 Raspberry Pi 4 Model B the NFS mount failed.
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
Following changes have been noticed in the Kconfig file between good and bad. The config files attached to this email.
-CONFIG_BCMGENET=y -CONFIG_BROADCOM_PHY=y +# CONFIG_BROADCOM_PHY is not set -CONFIG_BCM7XXX_PHY=y +# CONFIG_BCM7XXX_PHY is not set -CONFIG_BCM_NET_PHYLIB=y
Full test log details,
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
Where does the kernel configuration come from? Is this a plain defconfig that used to work, or do you have a board specific config file?
This is most likely caused by the added dependency on CONFIG_PTP_1588_CLOCK that would lead to the BCMGENET driver not being built-in if PTP support is in a module.
There's no PTP_1588_CLOCK_OPTIONAL in 5.10. This needs to be dropped:
|2a4912705 421f86 o: 5.10| net: broadcom: Add PTP_1588_CLOCK_OPTIONAL dependency for BCMGENET under ARCH_BCM2835
Best regards, Pavel
On Tue, Dec 13, 2022 at 10:20:30AM +0100, Arnd Bergmann wrote:
On Tue, Dec 13, 2022, at 08:48, Naresh Kamboju wrote:
On Mon, 12 Dec 2022 at 18:43, Greg Kroah-Hartman gregkh@linuxfoundation.org wrote:
Regression detected on arm64 Raspberry Pi 4 Model B the NFS mount failed.
Reported-by: Linux Kernel Functional Testing lkft@linaro.org
Following changes have been noticed in the Kconfig file between good and bad. The config files attached to this email.
-CONFIG_BCMGENET=y -CONFIG_BROADCOM_PHY=y +# CONFIG_BROADCOM_PHY is not set -CONFIG_BCM7XXX_PHY=y +# CONFIG_BCM7XXX_PHY is not set -CONFIG_BCM_NET_PHYLIB=y
Full test log details,
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
https://qa-reports.linaro.org/lkft/linux-stable-rc-linux-5.10.y/build/v5.10....
Where does the kernel configuration come from? Is this a plain defconfig that used to work, or do you have a board specific config file?
This is most likely caused by the added dependency on CONFIG_PTP_1588_CLOCK that would lead to the BCMGENET driver not being built-in if PTP support is in a module.
I've dropped the patch that caused this and will push out a -rc2 in a bit.
thanks all!
greg k-h
-
Arnd Bergmann -
Greg Kroah-Hartman -
Naresh Kamboju -
Pavel Machek