08:34:29 <xruxa> #startmeeting
08:34:29 <linarobot> Meeting started Wed Nov 11 08:34:29 2015 UTC.  The chair is xruxa. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:34:29 <linarobot> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:34:33 <xruxa> #chair bero|2
08:34:33 <linarobot> Current chairs: bero|2 xruxa
08:34:47 <xruxa> bero|2, are you with us?
08:36:08 <lijixin> xruxa: hi
08:36:11 <xruxa> 6 minutes past, lets roll.
08:36:16 <xruxa> sunao, please go first.
08:36:36 <sunao> What did you do?
08:36:37 <sunao> 1. flash the hikey board, got the android run on it. But the HDMI has no output.
08:36:39 <sunao> what are you going to do?
08:36:40 <sunao> 1. try to bring the hikey up, and will test the memory allocator on the board.
08:36:42 <sunao> what problems do you have?
08:36:43 <sunao> 1.  hikey hdmi problem, the android has been booted to home screen, and can see adb on the computer, but  the LCD monitor has no signal.
08:36:45 <sunao> 2. need adb driver on Windows.
08:37:44 <xruxa> #info sunao has issues with HDMI on his HiKey
08:38:23 <xruxa> sunao, please setup also fast model (ARMv8 one) in parallel - in case you get badly stuck on HiKey
08:38:47 <sunao> OK , I will try
08:38:54 <xruxa> bhoj, are you aware of the HDMI being such a common issues on HiKey? IIRC someone else just reported similar problem in this team ...
08:39:18 <eric____> xruxa: hi
08:39:31 <eric____> bero: hi
08:39:39 <bero> it works here, but I've heard the driver has some issues with various displays
08:39:51 <xruxa> #info Xavier has also display problems with HiKey
08:39:55 <bhoj> xruxa, nothing on HDMI could be some other issue. HDMI is flaky but you should get UI on screen for some duration
08:40:19 <bhoj> sunao, did you boot hikey with UEFI ? Did you build it from source ? if so did you include vendor tarball ?
08:40:20 <xruxa> bhoj, thanks.
08:40:51 <sunao> i just download the android release from 96boards, not build from source
08:41:21 <bhoj> sunao, okay. which release did you pick ?
08:41:26 * xruxa hopes HiKey in LAVA will not suffer the same issues
08:41:31 <sunao> the latest
08:41:55 <xruxa> thanks sunao . Feel free to continue the discussion with bhoj or other 96boards guys to get this solved. Check their bugzilla too.
08:41:59 <bhoj> sunao, link please ? did you use UEFI bootloader ?
08:42:12 <xruxa> Qian_, Qian__ - please go next
08:42:15 <sunao> I just follow the get started document
08:42:30 <sunao> http://builds.96boards.org/releases/hikey/linaro/debian/latest/
08:42:30 <Qian_> What did you do?
08:42:30 <Qian_> 1.	   a) have compiled some of SPEC tools successfully.
08:42:30 <Qian_> Suc: xz , rxp, specsum, specinvoke
08:42:32 <Qian_> Failed:  1) perl     [compiled successfully using the master version,  use perl instead of specperl to have try]
08:42:34 <Qian_> 2) other perl reference libs will use X86 version instead
08:42:36 <Qian_> 3)make [using /data/data/com.pdaxrom.cctools/root/cctools/bin/make instead]
08:42:38 <Qian_> 4) tar  (because can’t configure as root, use  /data/data/com.pdaxrom.cctools/root/cctools/bin/make instead)
08:42:40 <Qian_> 2.     compared  spec-perl-5.12.3 with open-perl-5.12.3, SPEC perl has changed some file names, like benchmark. spec-perl-5.12.3 native-compiles failed. i am not sure about some conifguration files.
08:42:42 <Qian_> What are you going to do?
08:42:44 <Qian_> 1.	   a) SPEC tools are almost prepared, i will test runspec now.
08:42:46 <Qian_> 2.	   b) compile and run testcase
08:42:48 <Qian_> What problems do you have?
08:42:50 <Qian_> 1.	none.
08:43:31 <xruxa> Qian_, great! Bit of warning - the whole SPEC seems to take at least 20 hours to run (on PC) - not sure how long that will be on ARM device or model - probably days.
08:43:42 <bhoj> sunao, you have tried debian image ?
08:44:06 <Qian_> xruxa,yes, i will try one by one
08:44:31 <xruxa> Qian_, and please check with bero how you could share the build SPEC binaries with us (in terms of linaro infrastructure) - do not want to have them sent by emails and such ok?
08:44:37 <sunao> sorry,  bhoj, here is the android link: https://builds.96boards.org/releases/hikey/linaro/aosp/latest/
08:44:38 <Qian_> so , if i can runspec,  that seems we have the chance
08:45:00 <xruxa> #action Qian_ share the SPEC ARM binaries with bero and team
08:45:05 <xruxa> Qian_, thanks
08:45:28 <xruxa> luther, can you go next? Do not see anything in the doc ...
08:45:37 <Qian_> firstly, i will check it whether can transplant, and i am writing auto -script nwo
08:45:41 <luther> sorry
08:45:46 * luther What did you do?
08:45:46 <luther> 1. Look at move and forward mechanism
08:45:46 * luther What are you going to do?
08:45:46 <luther> 1. They are very complex, so go on look at them
08:45:47 * luther What problems do you have?
08:45:49 <luther> 1. None till now
08:46:01 <bero> we should put them into private git so lava can pull them
08:46:17 <bhoj> sunao, please use UEFI instead of the proprietary bootloader that you have used .. let me know how it goes.
08:46:25 <xruxa> binaries in GIT - that is only half evil :)
08:46:52 <Qian_> xruxa, i have not tested successfully now, just some SPEC tools available
08:47:14 <xruxa> luther, is it going alright or difficult to progress forward?
08:47:16 <bhoj> sunao, please install by following these instructions: https://github.com/96boards/documentation/wiki/LatestSnapshots
08:47:39 <sunao> Thanks bhoj, I will have a try
08:48:05 <xruxa> is Xavier here?
08:48:33 <xruxa> #info Peter needed to take a day off today
08:48:33 <luther> xruxa, just step by step, no difficult
08:48:42 <xruxa> luther, ok, thanks
08:48:44 <bero> he sent a message on hangouts, he's taking the day off as well
08:48:47 <xruxa> liuyq, please go next
08:49:05 <liuyq> What did you do?
08:49:06 <xruxa> #info Xavier needed to take a day off today
08:49:06 <liuyq> Finished the cts test on Nexus7 with images compiled with thumb mode.(took 24h 43m 44s)
08:49:06 <liuyq> Checked the possible projects in update-externals.xml.
08:49:06 <liuyq> Checked some patches in LOLLIPOP-MLCR-PATCHSET
08:49:06 <liuyq> What are you going to do?
08:49:06 <liuyq> Running cts test on Nexus7 with images compile with arm mode projects
08:49:08 <liuyq> Continue on applying patches in LOLLIPOP-MLCR-PATCHSET  to MARSHMALLOW-MLCR-PATCHSET
08:49:10 <liuyq> What problems do you have?
08:49:12 <liuyq> none
08:49:24 <xruxa> 24 hous for CTS ...
08:49:49 <xruxa> liuyq, so do we have any add-ons in the LOLLIPOP-MLCR-PATCHSET ? What did you cherry pick?
08:50:02 <xruxa> s/LOLLIPOP-MLCR-PATCHSET/MARSHMALLOW-MLCR-PATCHSET
08:50:35 <xruxa> liuyq, also, the CTS passed - so no issues with the new Thumb2 code?
08:50:47 <liuyq> xruxa, there are some patches are already in marshmallow branch, so we don't need them
08:51:05 <xruxa> bero, how do we want to wrap-up the Thumb2 task? Will we take a patch?
08:51:18 <liuyq> xruxa, but there are some in linaro gerrit for bionic or gcc5.x, I will check with bero about them after the meeting
08:52:06 <bero> yes, patch to make use of whatever generates the smallest bits...
08:52:09 <liuyq> xruxa, I just finished the cts test with thumb mode images, need to compare the result running on images with arm mode
08:52:32 <xruxa> liuyq, I see, great, please add comments on the Story in JIRA.
08:52:56 <liuyq> xruxa, ack
08:52:57 <xruxa> liuyq, thanks.
08:53:14 <xruxa> zhizhoutian_, you are next please
08:53:20 <zhizhoutian_> yes
08:53:30 <zhizhoutian_> What did you do?
08:53:30 <zhizhoutian_> LMG-937 Setup builds for low memory work for T-Shark hd
08:53:30 <zhizhoutian_> talk with Lava guys about tshark devices current issue
08:53:30 <zhizhoutian_> fastboot number is not unique
08:53:30 <zhizhoutian_> adb always lost connect
08:53:31 <zhizhoutian_> code base version is too old
08:53:31 <zhizhoutian_> What are you going to do?
08:53:32 <zhizhoutian_> try to solve those issues
08:53:32 <zhizhoutian_> What problems do you have?
08:53:33 <zhizhoutian_> code base issue must wait tshark’s Android6.0 finish updating
08:54:07 <zhizhoutian_> btw, four guys from Tinno and Spreadturm are on board now.
08:54:09 <zhizhoutian_> they are Eric Wang from Spreadtrum, Jixin Li/Hai Wu/Jian Wang from Tinno
08:54:22 <xruxa> zhizhoutian_, I see, lots to do for T-shark. We will be tracking this daily. Thanks.
08:54:35 <xruxa> zhizhoutian_, Ok, lets ping them at the end of status round
08:54:51 <xruxa> bero, your status please?
08:54:52 <zhizhoutian_> xruxa: ok:)
08:55:09 <bero> What did you do?
08:55:19 <bero> none
08:55:38 <xruxa> :)
08:55:40 <zhizhoutian_> :)
08:56:01 <xruxa> that did not sound like bero
08:56:27 <zhizhoutian_> yes, bero 's type speed should not so slow.
08:58:16 <xruxa> bero, you still with us?
08:59:10 <xruxa> lets wait for bero to get back - we need him to talk to newcomers :)
08:59:31 <bero|2> bero is gone ;) but the better version is there
08:59:46 <bero|2> looks like one line was cut off earlier ;)
08:59:49 <Qian_> :)
08:59:53 <xruxa> bero++
09:00:03 <bero|2> the missing bit is:
09:00:04 <bero|2> What did you do?
09:00:04 <bero|2> Started porting the -fstrict-aliasing patchset to M
09:00:04 <bero|2> What are you going to do?
09:00:04 <bero|2> Finish porting the -fstrict-aliasing patchset to M
09:00:04 <bero|2> Help new assignees get started
09:00:04 <bero|2> If time, look into fixing LTO
09:00:04 <bero|2> What problems do you have?
09:00:44 <xruxa> thanks bero|2 , no need to list all your problems :)
09:00:52 <zhizhoutian_> bero##
09:01:04 <xruxa> #topic Newcomers
09:01:05 <bero|2> that line actually made it through ;)
09:01:53 <xruxa> I'd expect Eric Wang, Jixin Li, Hai Wu and also (not so much newcomer) Yu Xing.
09:02:03 <xruxa> zhizhoutian_, are they here?
09:02:11 <zhizhoutian_> yes, i have call them
09:02:24 <eric____> I'm here, :)
09:02:38 <zhizhoutian_> bero|2, xruxa i have send you an email with their resmue.
09:02:44 <xruxa> zhizhoutian_, ack
09:02:47 <yuxing_> I'm here
09:02:58 <bero|2> zhizhoutian_: yes, thanks -- makes it a lot easier to see what everyone can do
09:03:04 <xruxa> Hi eric____ , hi yuxing_
09:03:05 <lijixin> xruxa   I'm here
09:03:11 <xruxa> hi lijixin
09:03:33 <wangjian> hi
09:03:36 <lijixin> hi  xruxa.
09:03:39 <xruxa> hi wangjian
09:03:46 <xruxa> good to see you all on IRC
09:03:57 <wuhai> HI
09:04:10 <bero|2> welcome all - good to see you!
09:04:14 <eric____> good to see you and bero, :)
09:04:24 <xruxa> I understand yuxing_ has aready been assigned a task - https://projects.linaro.org/browse/LMG-939 re-investigate zramfs
09:05:42 <bero|2> I've assigned a random task to everyone given I had nothing but a list of names
09:05:50 <yuxing_> xruxa: yes, I am doing this thing
09:05:56 <bero|2> so we may need to shift some stuff around based on who can do something best
09:06:35 <xruxa> yuxing_, can you please come here to report on progress? Today or tomorrow, like everyone.
09:06:45 <zhizhoutian_> bero|2: Maybe you can assigned me a work relate low memory optimazation;)
09:06:59 <zhizhoutian_> bero|2: i think it is time for me to take part in this work.
09:07:05 <bero|2> zhizhoutian_: sure
09:07:16 <zhizhoutian_> bero|2: thanks:)
09:07:22 <yuxing_> What did you do?
09:07:30 <xruxa> lijixin, wuhai, wangjian, yuxing_ and eric____ - please update your status daily before the meeting at https://docs.google.com/document/d/1CgwSksOXW5r6V3M7EPxv5ot-IZNBuTo_h2_OHfOi3aA/edit#
09:07:40 <yuxing_> 1、enable the zramfs on 1G memory mobile and do some test make the system produce swap
09:07:52 <yuxing_> 2、for the 512MB memory mobile , I set the zram disk size to 300MB and set the proc/sys/vm/swappiness = 100
09:08:07 <yuxing_> What are you going to do? 1、 continuing  do some test , through the test understand the zramfs driver
09:08:16 <yuxing_> what problems do you have?
09:08:34 <yuxing_> 1、 At present , I have one methods for zramfs to reduce memory usage     set the  proc/sys/vm/swappiness=100 , Bigger swappiness implies more preference to swap dirty pages out to ZRAM
09:08:51 <yuxing_> 2、 I don't know much about zram , Whether there is any more optimization strategy
09:09:16 <bero|2> One thing I've been thinking about is checking how long stuff tends to stay in zram
09:09:16 <yuxing_> xruxa: this is my report
09:09:37 <bero|2> if it stays there for a long time, it may make sense to connect an lzma compressor into zramfs instead of using lzo/lz4
09:09:50 <bero|2> but if it's taken out quickly, speed is more important than compression efficiency
09:10:12 <bero|2> and we may have to experiment with the best values for disk size and swappiness
09:11:10 <zhizhoutian_> bero|2: we do care about disk size?
09:11:27 <bero|2> zhizhoutian_: yes, but it's a secondary concern
09:11:35 <zhizhoutian_> bero|2: i see.
09:12:28 <zhizhoutian_> bero|2: ksm also will help
09:12:35 <yuxing_> bero: how to detect the best value  disk size and swappness?
09:12:57 <zhizhoutian_> bero|2: i have do a work before, add ksm feature and saved ~4M memory
09:13:17 <zhizhoutian_> bero|2: maybe i can start from here.
09:13:59 <bero|2> zhizhoutian_: yes, that's currently in the queue and unassigned, let's put that on you
09:14:21 <bero|2> yuxing_: I think it's a matter of experimentation - try a couple of values and see how it affects memory use and performance
09:14:24 <zhizhoutian_> bero|2: fine, thanks:)
09:14:52 <bero|2> essentially, the bigger the zramfs size is, the more memory will be available - but also, things will get slower because of the compression going on
09:15:04 <bero|2> so we need to identify the point where we get the most memory with an acceptable performance
09:15:17 <yuxing_> bero: ok, thanks
09:15:24 <xruxa> bero|2, can we quickly run through the newcomers and their tasks now?
09:15:58 <xruxa> wangjian, I see bero|2 assigned you https://projects.linaro.org/browse/LMG-417 - please read it and ask if you have questions
09:16:31 <wangjian> 1.I don't know how to analyze them. Is there any doc and example?
09:16:39 <wangjian> 2.Which tool or command can I used to analyze the libc++ memory use and performance?
09:17:38 <bero|2> That is one of the more complicated tasks... Let me see if I can find something easier for a first task
09:17:50 <xruxa> wuhai, and you got https://projects.linaro.org/browse/LMG-418
09:17:56 <bero|2> essentially the tool to use here is looking at the code to see if there's anything wasteful
09:18:05 <xruxa> nothing for eric____  ? bero|2 ?
09:18:21 <xruxa> lijixin, you got https://projects.linaro.org/browse/LMG-420
09:18:28 <eric____> xruxa: :)
09:19:19 <xruxa> wangjian, lijixin wuhai - read those, tell us if it is bad match for your stills, ask questions here on IRC from bero or other team members.
09:19:29 <xruxa> we still want some task for eric____  :)
09:19:36 <wuhai> i don't kown how to get start for it too, Is there any documents or related information for it
09:20:04 <lijixin> xruxa, LMG-420 compiler-rt memory related operations in sync with bionic/cortex-strings, I don't know how to start my task. are there some documents or guide for our tast?
09:20:38 <xruxa> I guess we will need bero|2 to sit down with each of you for 5-10 minutes one after another and go through it. Can you do it bero|2 ?
09:20:39 <bero|2> initially there were some more people on the list - e.g. Mingji Lai - does anyone know if they're still coming or if something has changed there?
09:20:50 <bero|2> If they aren't coming, we can reassign their tasks...
09:20:55 <bero|2> xruxa: sure
09:21:20 <zhizhoutian_> bero|2: I am not sure who is Mingji Lai
09:21:38 <wangjian> bero|2: How to see how much memory to use by libc ++? Use "dumpsys meminfo" can not list libc ++ memory usage
09:21:39 <bero|2> ok, will check with Tom
09:21:39 <xruxa> zhizhoutian_, liuyq - did you hear from Ming Ling? He was supposed to work on kernel memory reductions
09:21:48 <lijixin> mingji lai can't access the net.
09:22:12 <bero|2> we'll need to fix that...
09:22:13 <zhizhoutian_> xruxa:  yes, i know Ming Ling, if you mean Ming Ling from Spreadtrum
09:22:31 <xruxa> #info mingji lai can't access the net.
09:22:34 <zhizhoutian_> xruxa: he take an eye on this project
09:22:50 <xruxa> zhizhoutian_, I guess that is the one. Is he doing that work or not? We did not hear from him AFAIK
09:22:58 <bero|2> yes, Ming Ling is from Spreadtrum, Tom said he'll work mostly with sumit
09:23:25 <bero|2> so we need to check if sumit heard from him
09:23:26 <zhizhoutian_> bero|2: I call him now, he is not far from myside
09:24:07 <xruxa> bero|2, I see. He has https://projects.linaro.org/browse/LMG-940 assigned to him ... We need to check with sumits - I did not see him in the kernel meeting either
09:24:29 <zhizhoutian_> xruxa: he is not on his seat...
09:24:58 <zhizhoutian_> xruxa: his phone is closed, i will tell him this after this meeting.
09:25:06 <xruxa> bero|2, ok, lets end the official meeting, so that some people can get back to work
09:25:21 <xruxa> did we forget someone with his status/problems?
09:26:06 <bero|2> I don't think so
09:26:26 <xruxa> #action bero|2 to sit down with newcomers and check/explain their newly assigned tasks
09:26:29 <xruxa> #endmeeting