=== Highlights ===
* Worked to get ADB over USB working for both Beagle and Panda boards
for 2.6.39
* Updated linaro+android tree to linaro 11.06 final kernel
* Worked on testing Android Alarm Timers integration with Posix Alarm
Timers.
* Sent pull requests to Tglx for patches I've queued for 3.1
* Started review of Kconfig code, had some discussions with folks from
the Yocto project on how Kconfig fragments might help. Reached out to
the author of Kconfig SAT work from a year ago to learn more.
* Helped narrow down last minute Android Image issue with u-boot on
beagleboard.
* Working with TI patch author that caused the linux-3.0 panda ehci boot
crash regression. Hoping to get a fix or a revert applied soon.
* Sent in request for AOSP license grant, Patrik added me to the group,
but unfortunately my account status hasn't seemed to be updated yet.
=== Plans ===
* Continue working to get panda ehci issue resolved before 3.0-final
* Continue working on Android Alarm Timer rework to utilize upstreamed
code.
* Look at the ADB story on the android-3.0 branch (the ADB code is
heavily reworked there).
* Get back to Kconfig review/research.
=== Issues ==
* Too much breakage & regressions around Panda!
* AOSP license grant group issues still remain before I can submit
patches to AOSP
== Linus Walleij linusw ==
=== Highlights ===
* Andrew merged the lm3530 patches to the mm tree.
* Arnd pulled in the ux500 stuff to arm-soc as proof of concept,
hey this seems to work!
* Played with a RealView PB11MPCore, fixed three TCM
patches when learning that I was mistaken thinking the
PB11MPCore has TCM, actually nothing ARM multicore
seems to have TCM, which simplifies that stuff.
http://marc.info/?l=linux-arm-kernel&m=130944303522448&w=2http://marc.info/?l=linux-arm-kernel&m=130944291922298&w=2http://marc.info/?l=linux-arm-kernel&m=130944294622342&w=2
* Secured the stuff necessary to boot the ARM Integrator
board, meaning I will have access to PB1176, PB11MPCore
and Integrator, if I get it to work.
* Lots of talk about pinctrl/pinmux subsystem with nVidia.
* Submitted some internal DMA40 patches upstream, also
added a patch of my own on top.
http://marc.info/?l=linux-kernel&m=130916853306730&w=2
* Fixed a pesky staging driver for Synaptics RMI4
http://marc.info/?l=linux-kernel&m=130941558024742&w=2
* Talked to people from Lauterbach and ST Micro about their
PID for HW-trace patch, reviewed code and gave some
advice.
* Helped Per Forlin test some generic MMC speedup
patches.
=== Plans ===
* New iteration of the pinctrl + pinmux subsystem
* Drive generalization of U300 and Nomadik GPIO
* Get into struct clk generalization and movement of clk
drivers into drivers/clk - handover from Jeremy Kerr.
We will do this in Cambridge or something like that.
* Vacation 2011-07-02 thru 2011-07-31
I will only do FUN hacking during vacation!
See you in Cambridge.
* Vacation after sprint meeting 2011-08-07 thru 2011-08-12
I will only do FUN hacking during vacation!
=== Issues ===
* Generalizing/consolidating GPIO control is on hold until
the subsystem maintainer (i.e Grant) has time to look into it.
Not much to worry about, we all have shortage of time.
* Sync with Grant & Jeremy overdue, but I wouldn't be of
any big use before the vacation anyway. Let's sync this
in Cambourne.
* Shanghaiied for some internal conflict resolution and patent
advisory at ST-Ericsson. Not too much though.
Thanks for reading,
Linus Walleij
== Kernel maintainance ==
* Merged the first few trees (u8500, zynq) into the arm-soc tree
* reviewed new PWM framework submission
* some more cleanups for randconfig build errors
* cleanup of the 8250 driver as preparation for getting rid of
inb/outb on most platforms.
* Review of highbank and sirfprima platforms
* discussions of DMA mapping API patches from Marek Szyprowski
== Plans ==
* Ask for more subarch pull requests, and merge the trees
* Submit patches from randconfig errors
* finalize 8250 cleanup
Following a suggestion from Nicolas I have implemented table driven
decoding of CPU instructions in kprobes. This is an alternative to the
mass of if/else code in the original kprobes implementation. Also, by
making the register usage in emulation code completely uniform,
emulation functions can be more widely reused and all the separate
prep_emulate_xxx functions unnecessary.
All of infrastructure for this is in patch 1, and I am looking for
feedback on this.
I've also added a second patch which shows how this infrastructure
is used to implement decoding and emulation of all the 32-bit Thumb
instructions.
== Dave Martin <dmart> ==
(Apologies for the late report-- I have been having some
power/connectivity problems)
=== General Activity ===
* CPU info declaration macros -- posted patches; few
comments so far (positive or negative). These may
get merged at the next merge window, depending on
feedback.
* Took a look at a firefox Thumb-2 related problem which
breaks webm video:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/789198 video
It turns out that a "weirdness" in gas is causing a load of ARM code
connected with the webm code to be linked on an odd halfword
boundary. This might be an assembler bug, but it's not 100% clear
what the correct behaviour would be.
=== linaro-kernel-o-standard-arch ===
* thumb-2 now appears to be turned on globally for all boards in
the linux-linaro-oneirii tree. "works" at a basic level for vexpress
and mx51evk. I need to follow up with jcrigby on the exact status
of this.
=== Plans ===
* Check whether the new kernel trees work in Thumb-2 on current
Freescale landing team trees.
* Follow up on Cortex-A15 model/platform availability.
== Dave Martin <dmart> ==
=== General Activity ===
* CPU info declaration macros -- posted patches; few
comments so far (positive or negative). These may
get merged at the next merge window, depending on
feedback.
* Took a look at a firefox Thumb-2 related problem which
breaks webm video:
https://bugs.launchpad.net/ubuntu/+source/firefox/+bug/789198 video
It turns out that a "weirdness" in gas is causing a load of ARM code
connected with the webm code to be linked on an odd halfword
boundary. This might be an assembler bug, but it's not 100% clear
what the correct behaviour would be.
=== linaro-kernel-o-standard-arch ===
* thumb-2 now appears to be turned on globally for all boards in
the linux-linaro-oneirii tree. "works" at a basic level for vexpress
and mx51evk. I need to follow up with jcrigby on the exact status
of this.
=== Plans ===
* Check whether the new kernel trees work in Thumb-2 on current
Freescale landing team trees.
* Follow up on Cortex-A15 model/platform availability.
== Per Forlin <perfor> ==
=== Highlights ===
* Run mmc test to measure how bandwidth is affected by number of sg
elements in an sg list.
* two bug fixes in the mmc non-blocking code, compilation issue and
mmc_test error.
* adjusted ste_dma40 to comply with dmaengine documentation for issue_pending.
* booted snowball on linux 3.0-rc4
=== Completed WIs===
* https://blueprints.edge.launchpad.net/linux-linaro/+spec/dmaengine-improve-…
* Update ste_dma40 to comply with documentation, to be use as
reference for others: DONE
* https://blueprints.edge.launchpad.net/linux-linaro/+spec/mmc-async-request
* linux 3.0 mmc async req, update after review: DONE
=== Plans ===
* rebase and test mmc non-blocking for linux 3.1
* wrap up minor things before going on 4 weeks vacation in July
=== Device Tree ===
* Sent v2 of imx51-dt patch set that addresses the comments on v1
from Grant and Arnd
* Added one patch into the series to get rid of the cpu_is_mx1()
uses in imx serial driver, so that the DT binding can fit in
the driver well.
* Per Grant's request, added a couple of functions into DT
infrastructure code for drivers to identify device id, which
will be widely needed by drivers using pdev->id when migrating
to DT. The patch is under Grant's review.
=== ARM/imx consolidation ===
* Added one more patch into imx-dma device cleanup series to pass
the whole firmware name via platform_data without the churn of
cpu name and TO version being variables. It's been acked by Vinod
and the whole series will go through Sascha's tree.
=== Misc ===
* [Bug 754254] imx51 randomly truncates serial input at 31
characters. Sent a follow-up patch to Nicolas changing CD/WP
to gpio mode on babbage for by-passing issue [1]. The fixing for
the issue has been incorporated into v4 of the series, which gets
no more comments and is waiting for Chris to merge.
[1] http://article.gmane.org/gmane.linux.ports.arm.kernel/120790
=== Plan ===
* DT on imx51
* imx cleanup and consolidation
--
Regards,
Shawn
