linaro-dev

linaro-dev@lists.linaro.org

4332 discussions

by Zach Pfeffer

I have some exciting news. We have our first working Android Snowball build! By taking AOSP 2.3.3 and adding a new board config based on Panda support the STE landing team with help from the Android platform team and our valued partners at STE were able to pull together a non-accelerated LEB for Snowball. This build will allow us a great base to enable acceleration, bring up John Stultz's common kernel and enable Android development on this fantastic platform. Many thanks to Mathieu, Frans, Patrik, Lee, Anmar, Scott and others who made this happen. -Zach Note: We're still working through some display and USB issues. Once those have been worked out we'll make an official release.

13 years, 10 months

[NOTES] Dev Platforms Status Meeting Notes and Actions 7/07/2011

by Tom Gall

Enclosed you'll find a link to the agenda, notes and actions from the Linaro Developer Platforms Weekly Status meeting held on July 07th in #linaro-meeting on irc.freenode.net at 14:00 UTC. http://wiki.linaro.org/Platform/DevPlatform/Meetings/2011-07-07 New Actions: * hrw to document how to update the cross packages for oneiric and also how to make that available at backports * hrw to document how he generated the first star rating test matrix, describing what tests were done * rsalveti to ping the whole team about vacation plans * rsalveti to set up a fastboot sync up meeting with jcrigby asac and pfefferz * rsalveti to ping asac about getting http://www.amazon.com/Adesso-Black-Keyboard-Glidepoint-Touchpad/dp/B00008XO… for some of the team -- Regards, Tom "We want great men who, when fortune frowns will not be discouraged." - Colonel Henry Knox Linaro.org │ Open source software for ARM SoCs w) tom.gall att linaro.org w) tom_gall att vnet.ibm.com h) tom_gall att mac.com

13 years, 10 months

[PATCH v6] mmc: documentation of mmc non-blocking request usage and design.

by Per Forlin

Documentation about the background and the design of mmc non-blocking. Host driver guidelines to minimize request preparation overhead. Signed-off-by: Per Forlin <per.forlin(a)linaro.org> Acked-by: Randy Dunlap <rdunlap(a)xenotime.net> --- ChangeLog: v2: - Minor updates after proofreading comments from Chris v3: - Minor updates after more comments from Chris v4: - Minor updates after comments from Randy v5: - Fixed one more comment and Acked-by from Randy v6: - Write out full function names and use () for all functions, feedback from James. Documentation/mmc/00-INDEX | 2 + Documentation/mmc/mmc-async-req.txt | 88 +++++++++++++++++++++++++++++++++++ 2 files changed, 90 insertions(+), 0 deletions(-) create mode 100644 Documentation/mmc/mmc-async-req.txt diff --git a/Documentation/mmc/00-INDEX b/Documentation/mmc/00-INDEX index 93dd7a7..a9ba672 100644 --- a/Documentation/mmc/00-INDEX +++ b/Documentation/mmc/00-INDEX @@ -4,3 +4,5 @@ mmc-dev-attrs.txt - info on SD and MMC device attributes mmc-dev-parts.txt - info on SD and MMC device partitions +mmc-async-req.txt + - info on mmc asynchronous requests diff --git a/Documentation/mmc/mmc-async-req.txt b/Documentation/mmc/mmc-async-req.txt new file mode 100644 index 0000000..aac5634 --- /dev/null +++ b/Documentation/mmc/mmc-async-req.txt @@ -0,0 +1,88 @@ +Rationale +========= + +How significant is the cache maintenance overhead? +It depends. Fast eMMC and multiple cache levels with speculative cache +pre-fetch makes the cache overhead relatively significant. If the DMA +preparations for the next request are done in parallel with the current +transfer, the DMA preparation overhead would not affect the MMC performance. +The intention of non-blocking (asynchronous) MMC requests is to minimize the +time between when an MMC request ends and another MMC request begins. +Using mmc_wait_for_req(), the MMC controller is idle while dma_map_sg and +dma_unmap_sg are processing. Using non-blocking MMC requests makes it +possible to prepare the caches for next job in parallel with an active +MMC request. + +MMC block driver +================ + +The mmc_blk_issue_rw_rq() in the MMC block driver is made non-blocking. +The increase in throughput is proportional to the time it takes to +prepare (major part of preparations are dma_map_sg() and dma_unmap_sg()) +a request and how fast the memory is. The faster the MMC/SD is +the more significant the prepare request time becomes. Roughly the expected +performance gain is 5% for large writes and 10% on large reads on a L2 cache +platform. In power save mode, when clocks run on a lower frequency, the DMA +preparation may cost even more. As long as these slower preparations are run +in parallel with the transfer performance won't be affected. + +Details on measurements from IOZone and mmc_test +================================================ + +https://wiki.linaro.org/WorkingGroups/Kernel/Specs/StoragePerfMMC-async-req + +MMC core API extension +====================== + +There is one new public function mmc_start_req(). +It starts a new MMC command request for a host. The function isn't +truly non-blocking. If there is on ongoing async request it waits +for completion of that request and starts the new one and returns. It +doesn't wait for the new request to complete. If there is no ongoing +request it starts the new request and returns immediately. + +MMC host extensions +=================== + +There are two optional members in the +mmc_host_ops -- pre_req() and post_req() -- that the host +driver may implement in order to move work to before and after the actual +mmc_host_ops.request() function is called. In the DMA case pre_req() may do +dma_map_sg() and prepare the DMA descriptor, and post_req() runs +the dma_unmap_sg(). + +Optimize for the first request +============================== + +The first request in a series of requests can't be prepared in parallel with +the previous transfer, since there is no previous request. +The argument is_first_req in pre_req() indicates that there is no previous +request. The host driver may optimize for this scenario to minimize +the performance loss. A way to optimize for this is to split the current +request in two chunks, prepare the first chunk and start the request, +and finally prepare the second chunk and start the transfer. + +Pseudocode to handle is_first_req scenario with minimal prepare overhead: + +if (is_first_req && req->size > threshold) + /* start MMC transfer for the complete transfer size */ + mmc_start_command(MMC_CMD_TRANSFER_FULL_SIZE); + + /* + * Begin to prepare DMA while cmd is being processed by MMC. + * The first chunk of the request should take the same time + * to prepare as the "MMC process command time". + * If prepare time exceeds MMC cmd time + * the transfer is delayed, guesstimate max 4k as first chunk size. + */ + prepare_1st_chunk_for_dma(req); + /* flush pending desc to the DMAC (dmaengine.h) */ + dma_issue_pending(req->dma_desc); + + prepare_2nd_chunk_for_dma(req); + /* + * The second issue_pending should be called before MMC runs out + * of the first chunk. If the MMC runs out of the first data chunk + * before this call, the transfer is delayed. + */ + dma_issue_pending(req->dma_desc); -- 1.7.4.1

13 years, 10 months

[ACTIVITY] Bernhard Rosenkraenzer 2011/07/03 to 2011/07/09

by Bernhard Rosenkranzer

=== Highlights === * Multimedia test extensions (esp. ffmpeg builder) * Built toolchains * Got i.MX53 related files into the tree === Plans === * Check if i.MX53 build works; fix if it doesn't * Sort out test content hosting for multimedia test === Issues === * Atrocious connectivity while doing on-site work for the day job (partially fixed) * i.MX53 build fails on android-build.linaro.org (but succeeds on the hack box)

13 years, 10 months

[ACTIVITY] Chao Yang 2011-07-03 to 2011-07-09

by Chao Yang

Chao Yang <cyang> Highlights - Working environment setup - Successfully build linaro toolchain-4.5-2011.06-0, toolchain-4.6-2011.05-0, toolchain-4.6-2011.06-0. - Successfully build Linaro Android with toolchain-4.5-2011.06-0 and toolchain-4.6-2011.05-0 - gcc segmentation fault found when building Linaro Android with toolchain-4.6-2011.06-0 - Investigation is ongoing tracked by https://bugs.launchpad.net/gcc-linaro/+bug/809768 Plans - Working with toolchain team to fix gcc segmentation fault issue - Benchmark 11.07 Toolchains Issues - Pandaboard HW not received yet. -- Chao Yang Android Platform Team Linaro.org │ Open source software for ARM SoCs Follow Linaro: http://www.facebook.com/pages/Linaro/155974581091106 http://twitter.com/#!/linaroorg http://www.linaro.org/linaro-blog/

13 years, 10 months

[ACTIVITY] Botao Sun 2011/07/06 to 2011/07/13.

by Botao Sun

Botao Sun <botao_sun> Highlights: 1. UK Business VISA application is under processing at Beijing UK embassy. 2. Made builds from source code for Panda, leb-Panda, Beagle, Linaro Android Toolchain successfully on "Hack Box" machine. 3. Busybox source code has been compiled successfully by "gcc-4.5-arm-linux-gnueabi" compiler, but failed with Linaro's toolchain ("android-toolchain-eabi-linaro-4.5-2011.06-0-2-2011-06-21_00-59-57-linux-x86.tar.bz2"). I'm discussing this issue with the technical leader of Linaro toolchain team - Mr. Michael Hope. 4. Got a general image about Samsung Origen board from Mr. Angus Ainslie, but haven't started the relevant work items. Plans: 1. Try to figure out & solve the Busybox compiling issue with our own toolchain. 2. I will get the Panda board & Samsung Origen board in next few days. 3. Discuss with Mr. Angus Ainslie for more details on Samsung Origen board & build. Issues: 1. Haven't find the root cause of Busybox compiling issue with our own toolchain. I have got the reply from Mr. Michael Hope, but he didn't provide a solution. 2. Risk for my UK business VISA application: Although my application has been submitted to Beijing UK embassy VISA application office, the staff told me that they will finish it around 2 weeks, without guarantee. 3. No Panda & Samsung Origen board (They are on the way). BR Botao Sun

13 years, 10 months

[ACTIVITY] Tony Mansson 2011/07/03 to 2011/07/09

by Tony Mansson

Tony Mansson <mansson> === Highlights === Gearing up for PM tasks. Support by asac and Mounir. Participated in 11.07 standup release meetings. Installed a daily build on Beagleboard XM and fired it up. Helped Zack with some PM tasks. Established how to work with Mumble (Must leave the office). Verifying build environment is ongoing. === Plans === Verify build environment by using the 11.06 release Tackle the "Enable Launcher2 in Android Builds" BP === Issues === Manufacturing problems with Snowballs. Pulled to help with Snowball verification & release of new firmware. The current 2.6.38 BSP Kernel git does not run. *No* Android-on-Snowball resources available at ST-Ericsson for another three weeks.

13 years, 10 months

[ACTIVITY] Frans Gifford 2011/07/03 to 2011/07/09

by Frans Gifford

== Frans Gifford <fgiff> == === Highlights === * Continued weekly hacking sessions with LT. Android builds now available for Snowball. * Got prototype build job working on android-build.l.o sandbox and generating JSON for validation.l.o. === Plans === * Work with LAVA team to get tests started securely from android-build.l.o. * Continue work with LT to improve android on Snowball. * Start work on Android demo, see what does and doesn't work on Snowball. === Issues === * Need to source a MiniUSB-A to USB-A adapter so I can connect USB hub and then keyboard to Snowball. -- Linaro.org │ Open source software for ARM SoCs Follow Linaro: Facebook | Twitter | Blog

13 years, 10 months

Re: [ACTIVITY] Zach Pfeffer 2011-7-3 to 2011-7-10

by Zach Pfeffer

Original at https://wiki.linaro.org/Platform/Android/Meetings/2011-07-13 === Highlights === * Supported unaccelerated Snowball Android bringup and got the first leb-snowball build going in https://android-build.linaro.org/ * Drove android-build to LAVA integration * Onboarded Chao Yang, Tony Mansson and Botao Sun * Started new MLO integration into TI 3.0 LEB * Reworked HDMI parsing patch to unblock 2.6.39 and TI LT kernel === Plans === * Make the first 11.07 release * Finish MLO integration * Work on libjpeg-turbo Android integration * Switch upstream focused builds to jstultz 3.0 tree * Keep supporting all LTs === Issues === * Need to get WG's work landed early and often so their work will ship

13 years, 10 months

[ACTIVITY] Android Platform Team 2011-06-26 to 2011-07-02

by Tony Mansson

Key Points for wider discussion: 8 engineers are now on board and operative in the Android Platform Team. LT/Android plugfests are all running and seem to be speeding up integration We'd like all teams to ensure that what their doing is always included in our daily Android builds in support of our switch to a continuously integrated flow. Team Highlights: "Include Busybox in Android Builds" for Panda and Beagle boards is ongoing. Team members are setting up build and work environments. The Quickstart guide ( https://wiki.linaro.org/Platform/Android/ImageInstallation) has been updated. A multimedia test script with CPU usage coverage is ongoing. A HDMI issue in TI 2.6.39 Kernel has been fixed. Cooperation with Landing Teams are working well. 11.07 Android Blueprints can be found at https://launchpad.net/linaro-android/+milestone/11.07 Miscellaneous: Please pull and test our daily builds: https://android-build.linaro.org/builds/~linaro-android/panda/ https://android-build.linaro.org/builds/~linaro-android/leb-panda/ https://android-build.linaro.org/builds/~linaro-android/beagle/ https://android-build.linaro.org/builds/~linaro-android/leb-snowball/ (note: if your contribution isn't in these builds it probably won't get in for 11.07 unless you talk to me ASAP) and our first 11.07 releases: https://android-build.linaro.org/builds/~linaro-android/beagle-11.07-releas… https://android-build.linaro.org/builds/~linaro-android/panda-11.07-release/

13 years, 10 months

Jump to page:

2025

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

linaro-dev