08:33:24 <bero> #startmeeting
08:33:24 <linarobot> Meeting started Thu Dec 17 08:33:24 2015 UTC.  The chair is bero. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:33:24 <linarobot> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:33:35 <bero> #topic sprint
08:33:57 <bero> Tom says he still hasn't heard back from some people about whether or not they can attend the sprint in the week of January 4th
08:34:22 <bero> please get back to him as soon as possible if you haven't - we need to arrange travel details etc.
08:34:44 <bero> #topic status updates
08:35:06 <bero> So on to the status updates -- yspan: do you want to go first?
08:35:11 <yspan> ok
08:35:24 <yspan> What did you do?
08:35:24 <yspan> Start LMG-941, read the article (https://01.org/android-ia/user-guides/android-memory-tuning-android-5.0-and-5.1)
08:35:24 <yspan> What are you going to do?
08:35:24 <yspan> Try to adjust those parameters with LCR-15.11 on Juno
08:35:24 <yspan> What problems do you have?
08:35:26 <yspan> no
08:35:50 <Qian__> I have sent the email and asked about whether we could send more people, but he has answered to me
08:36:30 <bero> Qian__: do you want to go next?
08:36:48 <Qian__> What did you do?
08:36:48 <Qian__> 1. string_copy.S causes performance regression on PRO 5 and MX5
08:36:48 <Qian__> 2. Investigate about lmg-966, i will drop handling app processes of android activities.
08:36:50 <Qian__> Instead, I will investigate to force swapping out the memory of specific process of native or kernel
08:36:52 <Qian__> What are you going to do?
08:36:54 <Qian__> 1. lmg-966
08:36:56 <Qian__> 2. try the find out un-frequently used processeses
08:36:58 <Qian__> 3. investigate to swap out the memory of specific process
08:37:00 <Qian__> What problems do you have?
08:37:02 <Qian__> 1. Can’t write dailyreport, only can view
08:38:16 <Qian__> bero: I think android has done well with app processes now
08:38:30 <bero> The write problem is a little odd, I just checked and your permissions are set to "Can edit" - by any chance, are you logged in with an account other than qian.chen@linaro.org?
08:39:21 <bero> Qian__: yes, M seems to do a pretty good job
08:39:22 <Qian__> bero: I have not logged in, i can't , don't know why
08:39:53 <bero> that's odd, I'll copy stuff from irc today, hope it'll work again tomorrow
08:40:00 <Qian__> bero: I think we should pay more attention on processes of kernel
08:40:05 <Qian__> or native
08:40:13 <liuyq> bero, it might be caused by the GFW problem. I can not use gmail last night, today is better
08:40:54 <Qian__> bero: liuyq, it can't work since yesterday
08:41:13 <bero> Qian__: yes, that's probably the easier ones to control
08:42:08 <bero> if they don't fix it soon, guess we'll move to etherpad
08:42:14 <Qian__> and i will collect the un-frequently used process as dumpsys batterystats done
08:42:25 <bero> sunao: do you want to go next?
08:42:35 <sunao> What did you do?
08:42:37 <sunao> 1.   porting ptmalloc to bionic c. In progress. init crashed in pthread_mutex_lock, hard to debug. trying.
08:42:38 <sunao> what are you going to do?
08:42:40 <sunao> 1.  continue porting ptmalloc2.
08:42:43 <sunao> what problems do you have?.
08:42:43 <sunao> None.
08:43:36 <bero> xavierhsu: do you want to go next?
08:43:51 <xavierhsu> What did you do?
08:43:51 <xavierhsu> 1. I trace the code for finding the jpeg decoder's error.
08:43:51 <xavierhsu> What are you going to do?
08:43:51 <xavierhsu> 1. I try to find which programs and functions will be called by jpeg_build_huffman_index and then builds these program into libjpeg.so.
08:43:51 <xavierhsu> What problems do you have?
08:43:51 <xavierhsu> 1. The build-in app still can not be used.
08:46:12 <Qian__> xavierhsu: seems hard work, maybe you can see the git log of android
08:46:37 <xavierhsu> ok
08:46:48 <xavierhsu> Qian, Thanks.
08:46:50 <liuyq> Qian__, bero, might be better to use dumpsys procstats to check the rarely used processes, https://source.android.com/devices/tech/debug/procstats.html
08:47:37 <Qian__> liuyq: Thanks, seems better
08:47:39 <bero> liuyq: yes, that looks interesting...
08:47:52 <bero> liuyq: do you want to go next?
08:48:19 <liuyq> What did you do?
08:48:19 <liuyq> Finished the cts part2 test on member hikey RC
08:48:19 <liuyq> Rebuild the Lollipop images with bootchart enabled for first time(the old version mali binary can not be used with the latest kernel version)
08:48:19 <liuyq> Finished the draft version release notes for 15.12
08:48:19 <liuyq> What are you going to do?
08:48:22 <liuyq> Continue to run the cts part1 and part2 test on reference HiKey RC
08:48:26 <liuyq> Finish to comparison on bootchart between Lollipop and Marshmallow
08:48:28 <liuyq> Check bugs on Member builds
08:48:30 <liuyq> What problems do you have?
08:48:32 <liuyq> None
08:48:40 <bero> eric____: do you want to go next?
08:48:48 <xavierhsu> liuyq, Thanks. I will use it to analysis the memory usage.
08:49:42 <liuyq> bero, do you know if we have any bionic optimizations for KitKat?
08:50:55 <liuyq> bero, I have no idea on how to find such optimizations, it's long time ago:(
08:50:57 <bero> liuyq: we sure used to, but that's centuries ago. Is this for the ZTE car thing?
08:51:48 <bero> Pretty sure they're on the linaro_android_4.4.4 branch in git
08:53:09 <sunao> liuyq: hi, i need to use the hackbox during sprint, whom I can turn to for apply?
08:53:28 <liuyq> bero, might be, I am not sure
08:53:48 <liuyq> sunao, send mail to its@ and have bero, xruxa cced
08:54:02 <sunao> liuyq: thanks
08:54:24 <liuyq> sunao, please send the mail as soon as possible, its guy might go on vacation sooner
08:54:53 <sunao> liuyq; I will do it right now
08:55:17 <bero> eric____: do you want to go next?
08:55:34 <eric____> fine
08:55:57 <eric____> I just had another meeting, and it's available now
08:56:41 <eric____> I found it is more files refer to zlib.gyp than expected
08:57:15 <bero> yes, gyp is a real pain
08:57:39 <eric____> so I have to check it one by one, just finished ,but there's still a litte problem
08:58:15 <eric____> when gclient runhooks
08:58:26 <eric____> try to fix it now
08:58:51 <eric____> bero: any suggestion?
08:59:21 <bero> what's the problem?
09:00:00 <eric____> there's two levels zlip.gyp
09:01:47 <eric____> not only in zlib but also in other category
09:02:08 <eric____> so I need distinguish them
09:03:20 <bero> hmm, I don't really know much about gyp, but I guess you can refer to them with path names
09:03:26 <bero> does anyone have more experience with gyp?
09:03:59 <eric____> also found other files who refer to zlib.gyp sometimes call zlib, come time call minizip, which both generate from zlib.gpy
09:05:39 <eric____> I think it's OK, for this is the right way to do, and it could be done at the end
09:06:00 <eric____> even if there's no one can help
09:06:22 <bero> ok thanks -- if anyone does know about gyp please speak up ;)
09:06:41 <bero> wuhai: do you want to go next?
09:06:48 <wuhai> OK
09:06:50 <eric____> yes, with 'YOUR' help, maybe it could move more faster
09:07:01 <wuhai> What did you do?
09:07:01 <wuhai> 1.Moving source code and document to local computer and compiled T-shark baseline;
09:07:01 <wuhai> 2.Analyzes the structure of code LocationManagerService,and modify source code.
09:07:01 <wuhai> What are you going to do?
09:07:01 <wuhai> 1.Continue to analyze the structure of code LocationManagerService,try to find the code that can be converted to C/C++ through JNI
09:07:01 <wuhai> What problems do you have?
09:07:02 <wuhai> None
09:07:10 <Qian__> eric____: maybe you can see the git log of zlib
09:07:27 <Qian__> Add minizip to the exported include paths
09:07:38 <bero> wangjian: do you want to go next?
09:07:57 <wangjian> What did you do?
09:07:57 <wangjian> I listed all Android.mk under extern/ and framework/ which have the hardcode specific -O flag.
09:07:57 <wangjian> What are you going to do?
09:07:57 <wangjian> Try to remove all hardcode specific -O flag and complie the T-shark code to test.
09:07:57 <wangjian> What problems do you have?
09:07:57 <wangjian> none.
09:08:30 <bero> yuxing__: do you want to go next?
09:09:27 <yuxing__> What did you do?
09:09:27 <yuxing__> 1.  Investigate whether or not the lib.so file can be mark MADV_MERGEABLE
09:09:27 <yuxing__> 2. capture proc/pid/smaps/ data from a java app and native program. find out the address which be marked MADV_MERGEABLE
09:09:27 <yuxing__> and find out how many address not be  marked MADV_MERGEABLE,  investigate whether or not these address can be marked ?
09:09:27 <yuxing__> What are you going to do?
09:09:27 <yuxing__> 1. I have found the linker mark the address of the lib.so for MADV_MERGEABLE, the linker through call mmap() to map the lib.so to address space of current program
09:09:27 <yuxing__> so the address space which the lib.so file occupy are be marked the MADV_MERGEABLE.
09:09:28 <yuxing__> 2.  From the java app's(com.android.email) smaps file,  I found the following type address space are not be marked  MADV_MERGEABLE:
09:09:28 <yuxing__> 1、current progress's stack
09:09:29 <yuxing__> 2、the address which maped  executable file occupy
09:09:30 <yuxing__> 3.  the address which maped  /dev/__properties__
09:09:30 <liuyq> wangjian, please click the "Assign to me"  for this https://projects.linaro.org/browse/LMG-924
09:09:30 <yuxing__> 3. From the native progress's  I found the following type address space are not be  marked the MADV_MERGEABLE.
09:09:30 <yuxing__> 1、current progress's stack
09:09:31 <yuxing__> 2、the address which maped  executable file occupy
09:10:39 <bero> thanks
09:10:46 <bero> I think that leaves me:
09:10:48 <bero> Bernhard Rosenkränzer <bernhard.rosenkranzer@linaro.org>
09:10:49 <bero> What did you do?
09:10:49 <bero> Fixed some clang 3.7/3.8 related bits
09:10:49 <bero> What are you going to do?
09:10:49 <bero> Investigate and fix more clang 3.7/3.8 bits
09:10:49 <bero> What problems do you have?
09:10:49 <bero> way more dependencies on a particular version of llvm than expected
09:10:57 <bero> #topic AOB
09:11:06 <bero> Does anyone have anything else they want to bring up?
09:11:11 <bero> (Or did I forget anyone?)
09:11:37 <wangjian> liuyq: I has assign it to me
09:12:07 <liuyq> wangjian, ok, thanks
09:13:42 <bero> ok, guess we're done - have a great rest of the day!
09:13:44 <bero> #endmeeting