Here are the build results from automated periodic testing.
The tree being built was mainline, found at:
http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git (master branch)
Topmost commit:
65ea11e x86/hweight: Don't clobber %rdi
Build logs (stderr only) can be found at the following link (experimental):
http://arm-soc.lixom.net/buildlogs/mainline/v4.8-rc1-1-g65ea11e/
Runtime: 56m 35s
Passed: 122 Failed: 1
Warnings: 1
Section mismatches: 0
-------------------------------------------------------------------------------
Failed defconfigs: arm.multi_v7_nosmp_defconfig
-------------------------------------------------------------------------------
Errors:
arm.multi_v7_nosmp_defconfig: pm-rcar-gen2.c:(.init.text+0x740): undefined reference to `platform_can_secondary_boot'
Warnings: 1 drivers/tty/serial/8250/8250_fintek.c:34:0: warning: "IRQ_MODE" redefined
Section mismatches:
On Mon, Aug 08, 2016 at 12:10:37PM -0700, Olof's autobuilder wrote:
Errors:
arm.multi_v7_nosmp_defconfig: pm-rcar-gen2.c:(.init.text+0x740): undefined reference to `platform_can_secondary_boot'
This is a real pain, because it prevents boot testing on many platforms - who's looking at fixing this?
Also, how did the problem creep in? Surely it should've been found prior to the merge window... Looks like it first appeared 1st August.
On Wednesday, August 10, 2016 5:16:15 PM CEST Russell King - ARM Linux wrote:
On Mon, Aug 08, 2016 at 12:10:37PM -0700, Olof's autobuilder wrote:
Errors:
arm.multi_v7_nosmp_defconfig:pm-rcar-gen2.c:(.init.text+0x740): undefined reference to `platform_can_secondary_boot'
This is a real pain, because it prevents boot testing on many platforms - who's looking at fixing this?
Also, how did the problem creep in? Surely it should've been found prior to the merge window... Looks like it first appeared 1st August.
Basically a series of vacation-related delays for various people in series.
I found it earlier than that (when the broken patch made it into linux-next through the renesas tree), but the shmobile maintainers at first didn't pick it up as I had expected, then I didn't realize that the bug had landed in arm-soc, and in the end Olof applied my fix manually, but we haven't sent our fixes branch since then.
It should be good as soon as Linus picks up arm-soc/fixes now.
Arnd
On Wed, Aug 10, 2016 at 09:28:11PM +0200, Arnd Bergmann wrote:
On Wednesday, August 10, 2016 5:16:15 PM CEST Russell King - ARM Linux wrote:
On Mon, Aug 08, 2016 at 12:10:37PM -0700, Olof's autobuilder wrote:
Errors:
arm.multi_v7_nosmp_defconfig:pm-rcar-gen2.c:(.init.text+0x740): undefined reference to `platform_can_secondary_boot'
This is a real pain, because it prevents boot testing on many platforms - who's looking at fixing this?
Also, how did the problem creep in? Surely it should've been found prior to the merge window... Looks like it first appeared 1st August.
Basically a series of vacation-related delays for various people in series.
I found it earlier than that (when the broken patch made it into linux-next through the renesas tree), but the shmobile maintainers at first didn't pick it up as I had expected, then I didn't realize that the bug had landed in arm-soc, and in the end Olof applied my fix manually, but we haven't sent our fixes branch since then.
Given it's impact (on one of our major defconfigs), I'll have to wait until this gets merged and base my tree off a later point (rc2?), or I just ignore the build results for this cycle, as they'll keep telling me each time I update my for-next branch that all these platforms failed because this defconfig failed.
These failures are rather annoying and inconvenient. Can we please do better to avoid them?
On Thursday, August 11, 2016 12:36:01 AM CEST Russell King - ARM Linux wrote:
On Wed, Aug 10, 2016 at 09:28:11PM +0200, Arnd Bergmann wrote:
On Wednesday, August 10, 2016 5:16:15 PM CEST Russell King - ARM Linux wrote:
On Mon, Aug 08, 2016 at 12:10:37PM -0700, Olof's autobuilder wrote:
Errors:
arm.multi_v7_nosmp_defconfig:pm-rcar-gen2.c:(.init.text+0x740): undefined reference to `platform_can_secondary_boot'
This is a real pain, because it prevents boot testing on many platforms - who's looking at fixing this?
Also, how did the problem creep in? Surely it should've been found prior to the merge window... Looks like it first appeared 1st August.
Basically a series of vacation-related delays for various people in series.
I found it earlier than that (when the broken patch made it into linux-next through the renesas tree), but the shmobile maintainers at first didn't pick it up as I had expected, then I didn't realize that the bug had landed in arm-soc, and in the end Olof applied my fix manually, but we haven't sent our fixes branch since then.
Given it's impact (on one of our major defconfigs), I'll have to wait until this gets merged and base my tree off a later point (rc2?), or I just ignore the build results for this cycle, as they'll keep telling me each time I update my for-next branch that all these platforms failed because this defconfig failed.
Maybe you could merge the current arm-soc/fixes branch into your for-next branch to avoid the warning.
I agree it's very annoying to see the warning all the time
These failures are rather annoying and inconvenient. Can we please do better to avoid them?
I think we are already doing much better than in the past. There are only a handful of patches missing to get zero warnings and failures in kernelci, and this one (multi-v7 without SMP) is not a configuration that anyone uses in practice or that is part of upstream arch/arm/configs/.
Of course we should always try to improve.
Arnd
kernel-build-reports@lists.linaro.org
-
Arnd Bergmann -
Olof's autobuilder -
Russell King - ARM Linux