Olof, Kevin,
I was wondering if you had thought of adding the testing of maintained stable kernel versions to the boot farm? I've been testing 3.10-stable and 3.14-stable yesterday/today on Marvell Armada XP, and none of them boot. I don't know (yet) if it's a regression introduced by the 3.10-stable and 3.14-stable branches, or some original problem in 3.10/3.14, but it would be very useful to have a monitoring on the bootability of the stable kernels.
As of now, the maintained kernels are 2.6.32, 3.2, 3.4, 3.10, 3.12, 3.14, 3.17 and 3.18. I personally don't care about anything before 3.10. Don't know about others.
Thoughts?
Thomas
On Wed, Jan 07, 2015 at 10:47:36AM +0100, Thomas Petazzoni wrote:
I was wondering if you had thought of adding the testing of maintained stable kernel versions to the boot farm? I've been testing 3.10-stable and 3.14-stable yesterday/today on Marvell Armada XP, and none of them boot. I don't know (yet) if it's a regression introduced by the 3.10-stable and 3.14-stable branches, or some original problem in 3.10/3.14, but it would be very useful to have a monitoring on the bootability of the stable kernels.
Both have been booting for me on Panda FWIW, and v3.14 seems to boot on BBB just fine too.
As of now, the maintained kernels are 2.6.32, 3.2, 3.4, 3.10, 3.12, 3.14, 3.17 and 3.18. I personally don't care about anything before 3.10. Don't know about others.
Thoughts?
FWIW my own scripts are just set up to look at all the branches in the stable tree - that seemed easier than worrying about which ones to look at, if something goes out of support there won't be any updates, the cost of looking is virtually zero. If nothing else looking at the older trees is a reminder of how much good paying attention to buildability has been doing.
Hi Thomas,
On Wed, Jan 7, 2015 at 1:47 AM, Thomas Petazzoni thomas.petazzoni@free-electrons.com wrote:
I was wondering if you had thought of adding the testing of maintained stable kernel versions to the boot farm?
Yes, we've been doing stable build/boot testing for awhile, here are a few examples from my boot farm:
latest 3.10.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006614.... latest 3.12.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006586.... latest 3.14.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006593.... lastet 3.17.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006595.... latest 3.18.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006810....
I've been testing 3.10-stable and 3.14-stable yesterday/today on Marvell Armada XP, and none of them boot.
Based on the boot reports above, looks like openblocks was booting on v3.10.y and v3.14.y just fine for me. BTW, I have the big-endian boot testing blacklisted on v3.10 since it wasn't known working then.
I don't know (yet) if it's a regression introduced by the 3.10-stable and 3.14-stable branches, or some original problem in 3.10/3.14, but it would be very useful to have a monitoring on the bootability of the stable kernels.
You do. :)
As of now, the maintained kernels are 2.6.32, 3.2, 3.4, 3.10, 3.12, 3.14, 3.17 and 3.18. I personally don't care about anything before 3.10. Don't know about others.
Thoughts?
I only test v3.10 and later.
Kevin
Hello Kevin,
On Wed, 7 Jan 2015 08:16:02 -0800, Kevin Hilman wrote:
Yes, we've been doing stable build/boot testing for awhile, here are a few examples from my boot farm:
latest 3.10.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006614.... latest 3.12.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006586.... latest 3.14.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006593.... lastet 3.17.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006595.... latest 3.18.y: http://lists.linaro.org/pipermail/kernel-build-reports/2014-December/006810....
Ah, great. Maxime indeed point me to these tests after seeing my e-mail. I was just surprised that 3.10 and 3.14 were not working.
After some investigation:
* 3.14 was not booting due to a DT issue around the internal register base address (depending on which bootloader version is used, the DT needs to be adjusted)
* 3.10 boots fine on the AX3 you use for testing, but not on some other Armada XP development platform, for a reason that I haven't found yet (latest mainline boots fine).
So none of these problems affect the AX3 used in the boot farm.
Thanks for your feedback, and for all the testing work!
Thomas
On Wed, Jan 7, 2015 at 1:47 AM, Thomas Petazzoni thomas.petazzoni@free-electrons.com wrote:
Olof, Kevin,
I was wondering if you had thought of adding the testing of maintained stable kernel versions to the boot farm? I've been testing 3.10-stable and 3.14-stable yesterday/today on Marvell Armada XP, and none of them boot. I don't know (yet) if it's a regression introduced by the 3.10-stable and 3.14-stable branches, or some original problem in 3.10/3.14, but it would be very useful to have a monitoring on the bootability of the stable kernels.
As of now, the maintained kernels are 2.6.32, 3.2, 3.4, 3.10, 3.12, 3.14, 3.17 and 3.18. I personally don't care about anything before 3.10. Don't know about others.
Thoughts?
The only Marvell hardware I have is a Cubox with Dove, so I don't do any Armada testing at all.
I also stopped testing stable releases -- I was trying to test them at -rc time but Greg doesn't publish them in a format that makes it easy and I lost interest. I've turned off all stable builds.
-Olof
Dear Olof Johansson,
On Wed, 7 Jan 2015 10:07:14 -0800, Olof Johansson wrote:
Thoughts?
The only Marvell hardware I have is a Cubox with Dove, so I don't do any Armada testing at all.
Ok. I was actually more thinking about Kevin's farm, since he does have some platforms of interest to us (Armada 370 and Armada XP).
I also stopped testing stable releases -- I was trying to test them at -rc time but Greg doesn't publish them in a format that makes it easy and I lost interest. I've turned off all stable builds.
Ah too bad :-/
Thomas
kernel-build-reports@lists.linaro.org
-
Kevin Hilman -
Mark Brown -
Olof Johansson -
Thomas Petazzoni