Hi,
I had a really good meeting with Paul Larson this week to discuss
the automated test framework that Linaro is developing and wanted
to follow up with the team on brainstorming around what we want
to be running on a regular basis to test the kernels. The current testing
includes:
"x11perf",
"gtkperf",
"glmemperf",
"posixtestsuite",
"gmpbench",
"pybench",
"stream"
I think the obvious missing one is LTP which Shawn is working on
getting up and running on ARM for the 11.11 cycle. Arnd's flash
testing tool might also be a good fit, but I am not sure of the side
effects it might have to the underlying device after extended use.
I also think that adding board specific tests such as "run i2cdetect
and validate that we see all the devices we expect" is something
useful to do in the long-term to validate that driver changes, DT
updates, etc do not break support for a given board.
I'm interested in hearing ideas that others have on what kinds of
tests we should be running.
Thanks,
~Deepak
Hi,
after Nicolas' announcement of linaro-2.6.39 [1] I had a look to the
bsp/freescale/linux-linaro-natty.git lt-2.6.38 kernel [2] what has to
be done to port that to 2.6.39.
Looking at that lt-2.6.38 branch, I exported 184 patches [3] which
seem to be specific for this lt kernel (?).
Then, I rebase them to Nicolas' linaro-2.6.39 using quilt. See the
attached quilt series file for which patches were used and which were
dropped (because they seem to be already in 2.6.39 mainline).
The result does compile with the imx5_defconfig. I've done no boot
tests, yet. So this is only compile tested.
The resulting quilt patch series is ~2,5MB bz2 compressed. I don't
want to send it to the mailing list. I will send it as a reply to this
mail to Eric and Nicolas. Let me know if you are interested in it, too
(or where to put it so that everybody can access it).
Note: I couldn't figure how to put binary firmware files into a patch.
This does mean that the patch series misses the firmware/imx/sdma/
files. You have to download them from [4] and put them manually to
firmware/imx/sdma/ to be able to compile imx5_defconfig successfully.
Opinions?
Best regards
Dirk
[1] http://lists.linaro.org/pipermail/linaro-kernel/2011-June/000461.html
[2]
http://git.linaro.org/gitweb?p=bsp/freescale/linux-linaro-natty.git;a=short…
[3] Starting with
http://git.linaro.org/gitweb?p=bsp/freescale/linux-linaro-natty.git;a=commi…
i.e.
4f2a27a2bf1185e553c85a957102022db67773d6...adcf32a87785f44c19189996a23ecfc09cd0ccef
(adcf32a87785f44c19189996a23ecfc09cd0ccef == recent head)
[4]
http://git.linaro.org/gitweb?p=bsp/freescale/linux-linaro-natty.git;a=tree;…
== Highlights ==
* Fill in the blueprint of 11.11. cycle, continue study some information
* Â MX53 u-boot upstream, loco board support patch serials already in
u-boot master branch
and will be included into the v2011.06 release.
== Plans ==
* Fill the blueprint spec.
* Some u-boot patches upstream review and revision.
== Absence ==
* vacation from 12 June to 29 June,
== Per Forlin <perfor> ==
=== Highlights ===
* Reviewing dependencies of the storage blueprints,
* investigated sdio single irq issue in libertas.
* Setting up wl12xx_sdio_test on pandaboard (ongoing)
== Thomas Abraham <thomas-ab> ==
=== Highlights ===
* Discussed with IBM Engineer, Ankita Garg, about memory regions patchset.
* Study of DDR3 specification and a Samsung DDR3 part datasheet to
understand power consumed during various bank states.
* Measured power consumed by DDR3 on smdkv310 board.
* Discussed with Samsung Landing team about ARM consolidation work.
=== Plans ===
* Complete all the power measurements on smdkv310 board as suggested by Ankita.
* Test memory regions patches on Exynos4 platform.
== Shawn Guo (shawnguo) ==
=== Device Tree ===
* It exposed build error when sdhci-pltfm&OF-consolidation patch set
hits linux-next tree. It turns out there is problem with multiple
sdhci-pltfm based drivers compiling and module build. Sent a
follow-up patch to fix these problems. Chris had picked it up on
mmc-next tree.
=== ARM/imx consolidation ===
* The patches that move imx gpio drivers under subarch into
drivers/gpio have hit Grant's tree.
* The patches converting gpio-mxc and gpio-mxs to use basic_mmio_gpio
library are on the way to be accepted. Tens of LOC will be saved
for each driver.
=== Misc ===
* Sent the v2 of dma-set/get-max_segment_number patch to explain why
the patch is needed.
* Added leds-gpio support for mx28evk board.
=== Plan ===
* Keep looking at if the imx irq_chip and gpio_irq_chip can be migrated
to generic_irq_chip
--
Regards,
Shawn
Just a heads up, I'm out Weds-Friday next week.
=== Highlights ===
* Rebased linaro+android tree to Nico's final 2.6.38 kernel for the
11.05 release
* Sent out first pass at suspend/resume drift improvement (Similar
method to what Android uses)
* Had an initial meeting with AndyG and Zach
* Sent out a number of clocksource_register_khz/hz cleanups
* Queued an prepped a bunch of community RTC, and PTP fixups for 3.0 and
sent Thomas pull requests.
* Queued a few RTC fixups and TSC improvements for 3.1
* Started work on 2.6.39 based Linaro+Android tree on top of Nico's
2.6.39 tree.
* BeagleBoard seems to be working fine
* Panda HDMI output support is gone (need to be rebased and merged
into Nico's tree)
* Started working on an internal presentation.
=== Plans ===
* Short week: I'm on vacation Weds-Friday.
* Try to queue suspend/resume drift correction for 3.1
* Continue working bug #789089
* Continue to work RTC issue from lkml.
* More internal presentation prep
=== Issues ===
* NA
== Paul E. McKenney <paulmck> ==
=== Highlights ==
* Continued tracking down issues relating to RCU priority boosting in mainline,mostly boot-time slowdowns that I cannot reproduce.
* Worked on ARM server whitepaper.
* Attended public plan reviews.
== Miscellaneous ==
* Got out third release of "Is Parallel Programming Hard, And, If So, What Can You Do About It?" (http://paulmck.livejournal.com/26906.html).
* Linux Plumbers Conference refereed-track submissions reviewed. Jesse Barker's "Unified Memory Management" microconference proposal accepted.
Sorry for this being late! Took most of Friday off and forgot to send it
before heading to the coast!
=== Highlights ===
* "Facilitated discussion" (AKA: Nagged enough developers) on android
scheduler trivial tree patch I sent out earlier, that it got queued and
merged.
* Android trivial mm patch was merged upstream
* Spent a few days chasing down a HDMI regression on beagle XM (thanks
to Ricardo Salveti for helping!). Finally got it resolved and pushed the
fix.
* Merged and made a git pull request for Richard Cochran's PTP patches
that have been in the works for awhile. Patches were able to land for
3.0-rc1.
* Merged RTC fixups for stmp3xxx from Wolfram, but they will need
further work to get upstream.
* Took an initial stab at the RTC suspend/resume drift correction
suggested by Arve.
* Starting working on the repeated console suspend failure notification
issue ( https://bugs.launchpad.net/linaro-android/+bug/789089 ).
=== Plans ===
* Refresh Linaro Android tree with latest bits from Nico
* Continue working bug #789089
* Try to get a patch to lkml for suspend/resume drift correction
* Rework Wolfram's RTC patches
* Try to diagnose RTC issue from lkml.
=== Issues ===
* NA
== Dave Martin <dmart> ==
=== General Activity ===
* Discussed work items and assignments with Deepak Saxena.
* kernel snapshot testing
=== linaro-kernel-o-standard-arch ===
* {set,get}_fiq_regs patches for Thumb-2 compatibility now
merged upstream.
=== Plans ===
* Test Thumb-2 kernel on vexpress. If it works (likely)
follow up with John Rigby so we can use Thumb-2 on vexpress
by default.
* Sync up with freescale landing team regarding Thumb-2 kernel
testing and reployment.
