08:31:22 <xruxa> #startmeeting
08:31:22 <linarobot> Meeting started Wed Dec 16 08:31:22 2015 UTC.  The chair is xruxa. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:31:22 <linarobot> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:31:43 <xruxa> Hello all, lets get started with a status round
08:31:49 <xruxa> #topic Status round
08:31:53 <xruxa> #chair bero
08:31:53 <linarobot> Current chairs: bero xruxa
08:32:25 <xruxa> lets start with someone we missed a lot (not bero2  ) - Peter - yspan please go first
08:32:52 <yspan> What did you do?
08:32:52 <yspan> I made gcov work with shared lib in Android. I ran pngtest (disable write support to simulate normal APPs uses only decoding), the results listed [URL only in shared doc]
08:32:52 <yspan> What are you going to do?
08:32:52 <yspan> Write a wiki page to describe how to use gcov in Android, I think it is very useful if we want to separate some functions out. Tried 3-party APPs that uses png decode to see what functions are less used.
08:32:53 <yspan> What problems do you have?
08:32:55 <yspan> Can someone recommend this kind of APPs
08:33:29 <xruxa> yspan, so is your work on libpng done (patches merged) and you are looking for some additionalunused parts?
08:33:46 <yspan> yes, that's right
08:34:16 <xruxa> yspan, bero2  - I will want to resolve LMG-979 as done (we got libpng patches in) and make another item for the unused bits work
08:34:35 <xruxa> yspan, how agout LMG-941 - ART parameters? Will you still look into that in DEcember?
08:35:12 <yspan> xruxa, no. Which one has higher priority?
08:35:52 <xruxa> yspan - I do not care really, unless bero2  has opinion.
08:36:54 <xruxa> yspan, lets decide tomorrow - for now continue as you were, ok?
08:37:03 <bero2> Let's look at ART parameters first...
08:37:04 <yspan> OK, no problem
08:37:20 <bero2> Just so we know how much we can achieve with it
08:37:21 <yspan> bero2, OK. Then, I will switch to ART
08:37:36 <bero2> We know pretty much how much libpng can save
08:38:07 <xruxa> thanks yspan
08:38:12 <xruxa> liuyq1, please go next?
08:39:45 <xruxa> ok, eric____ please go next while we wait for liuyq1
08:40:00 <eric____> What did you do?
08:40:00 <eric____> Add zlib.gyp file for system zlib, and modify it for build, for there's many different with zlib from chromium;
08:40:00 <eric____> Modify all.gyp make it point to system lib;
08:40:00 <eric____> What are you going to do?
08:40:00 <eric____> make it work with system lib;
08:40:00 <eric____> What problems do you have?
08:40:00 <eric____> None
08:40:34 <xruxa> eric____, thanks.
08:40:36 <eric____> bero2: I think it's a way to do this, and can be done quickly
08:40:38 * liuyq1 is ready
08:40:53 <xruxa> liuyq1, please go ahead
08:41:00 <liuyq1> What did you do?
08:41:00 <liuyq1> Finished the draft version for 15.12 release notes
08:41:00 <liuyq1> Investigated about the mali problem on Lollipop builds(might be kernel problem)
08:41:00 <liuyq1> Finished the cts part1 test for reference Hikey Board
08:41:00 <liuyq1> What are you going to do?
08:41:04 <liuyq1> Continue on the investigation of lollipop images
08:41:06 <liuyq1> Finish the bootchart comparison between lollipop and marshmallow
08:41:08 <eric____> bero2: Do you think so?
08:41:08 <liuyq1> Continue the cts part2 test for member Hikey build
08:41:10 <liuyq1> What problems do you have?
08:41:12 <liuyq1> None
08:41:26 <xruxa> liuyq1, thanks.
08:41:27 <liuyq1> xruxa, pleae help to check the release notes though I merged it already
08:41:47 <xruxa> #info bero2 is on low bandwidth but watching us - do not expect fast replies
08:41:57 <xruxa> liuyq1, ack, right after the meeting :)
08:42:19 <xruxa> do not see Sun Ao here ...
08:42:28 <xruxa> Qian__, please go next?
08:42:34 <Qian___> What did you do?
08:42:35 <Qian___> 1. merge and test bionic on PRO 5, testing now
08:42:35 <Qian___> What are you going to do?
08:42:36 <Qian___> 2. lmg-966
08:42:38 <Qian___> What problems do you have?
08:42:40 <Qian___> 1. libjpeg turbo has been reverted from meizu master, it can’t show pictures acationally
08:42:42 <Qian___> 2. my company colleage wants to ask whether libjpeg-trubo only support 1x, 2x, 4x, 8x  scale times,
08:42:44 <Qian___> Does it will meet problem when scale such as 3x, 5x, 6x, 7x times?
08:43:57 <xruxa> Qian___, you are on LMG-966? It is not in the sprint - so should I bring it in? You have LMG-914 in the sprint as well :)
08:44:55 <xruxa> Qian___, on the LMG-914 (static linkage/copied code) bero says the next step is to identify where switching to dynamic linking helps. Ok?
08:44:58 <Qian___> xruxa: I think lmg-914 has almost finished, left bero2 to check and create some issues
08:45:41 <Qian___> xruxa, bero2, does i misunderstand?
08:46:10 <Qian___> xruxa: bero2 said looks good yesterday
08:46:35 <xruxa> Qian___, ok, if he confirms we can resolve it as done. Thanks. I will bring the LMG-966 in for you as well.
08:46:58 <Qian___> can anyone give me some suggestion about libjpeg trubo
08:47:24 <Qian___> xruxa: OK, that's fine
08:48:08 <Qian___> what a pity it can't be used in product
08:48:35 <xruxa> zhizhoutian__, please go next
08:48:53 <Qian___> the performance of libjpeg trubo has improved about 40%
08:48:54 <zhizhoutian__> xruxa: ok
08:48:58 <liuyq1> Qian__, for the libjpeg-turbo, could you please file bugs on bugzilla
08:49:24 <liuyq1> Qian__, then we can reproduce your problems and check.
08:49:31 <Qian___> liuyq1: i will, i am collecting messages now
08:49:57 <xruxa> #action Qian___ to file bugs for libjpeg-turbo in https://bugs.linaro.org/  product is "Linaro Android" and component "Member confectionery release"
08:50:14 <liuyq1> Qian__, and for the scale problem, I think that needs more investigation, so please file a bug as well on bugzilla
08:50:15 <Qian___> liuyq1: They said it can be reproduced by running monkey
08:50:26 <xruxa> zhizhoutian__, status please
08:50:35 <Qian___> liuyq1: OK,i  will
08:50:51 <zhizhoutian__> What did you do?
08:50:52 <zhizhoutian__> 1. analysis swap code and try to got code work flow
08:50:52 <zhizhoutian__> 2. investigate swap to emmc
08:50:52 <zhizhoutian__> What problems do you have?
08:50:52 <zhizhoutian__> 1. how often swapping happened?
08:51:05 <zhizhoutian__> xruxa: ok, ok
08:51:22 <zhizhoutian__> xruxa: that is it. yesterday i have got a cold.
08:51:37 <xruxa> zhizhoutian__, oh get better soon. Thanks for the report.
08:52:00 <zhizhoutian__> xruxa: thanks
08:52:38 <xruxa> yuxing__, please go next
08:52:57 <yuxing__> What did you do?
08:52:57 <yuxing__> 1.  Investigate whether or not the lib.so file can be mark MADV_MERGEABLE
08:52:57 <yuxing__> 2. Investigate which program is responsible for loading the *.so  library,
08:52:57 <yuxing__> I have found that the .so file  was loaded in the android linker program
08:52:58 <yuxing__> so next ,  find the linker whether or not mark MADV_MERGEABLE for *.so
08:52:58 <yuxing__> 3.  Investigate whether or not the MADV_MERGEABLE page can be swaped out?
08:52:58 <yuxing__> About this problem , I think that I should see the swap code, may be need some times.
08:52:59 <yuxing__> 4. capture proc/pid/smaps/ data from a java app and native program. find out the address which be marked MADV_MERGEABLE
08:53:00 <yuxing__> and find out how many address not be  marked MADV_MERGEABLE,  investigate whether or not these address can be marked ?
08:53:00 <yuxing__> What are you going to do?
08:53:01 <yuxing__> continue above work
08:53:01 <yuxing__> what problems do you have?
08:53:02 <yuxing__> none
08:53:25 <liuyq1> Qian__, the monkey will have different paths with different parameters on different platform, so it's better to have the steps you reproduced the problem
08:53:27 <xruxa> yuxing__, nice. And interesting. Thanks for the report.
08:53:40 <zhizhoutian__> 3.  Investigate whether or not the MADV_MERGEABLE page can be swaped out?
08:53:59 <zhizhoutian__> a merged page can not mirgrated.
08:54:49 <zhizhoutian__> before swapping out, shrink_inactive_list() will check if it is a ksmd page(have merged)
08:54:56 <yuxing__> zhizhoutian__, you mean it is not be swaped out?
08:55:13 <zhizhoutian__> yes, a merged page will never be swapped out.
08:55:18 <xruxa> xavierhsu1, please go next
08:55:24 <xavierhsu1> What did you do?
08:55:25 <xavierhsu1> 1. I modify the source code.
08:55:25 <xavierhsu1> What are you going to do?
08:55:25 <xavierhsu1> 1. I use dlopen to loading the lib (libjpeg-enc.so) when first time demand loading and then did not close the lib (libjpegenc.so).
08:55:25 <xavierhsu1> What problems do you have?
08:55:25 <xavierhsu1> 1. The build-in app (Gallery) can open the picture, but only few second, and then build-in app happen error and close.
08:55:42 <yuxing__> zhizhoutian__, oh, thanks a lot
08:56:15 <xruxa> xavierhsu1, thanks. And thanks for the code. I leave it to bero2 to decide if we should have it in GIT repo or not
08:56:33 <xruxa> Ok, that looks like it. Did we forget someone (other than bero2 )?
08:56:39 <xavierhsu1> xruxa, ok . thanks.
08:56:58 <xruxa> bero2, you want to go?
08:57:07 <bero2> We should always have everything in git (just not on the main branch if it isn't working yet)
08:57:15 <bero2> Sure...
08:58:07 <bero2> What did you do?
08:58:16 <Qian___> liuyq1: I will try, but i am not sure they will continue feeding back as it has been reverted, i can try it my sparetime
08:58:28 <bero2> Fixed some clang 3.7/3.8 related bits
08:58:37 <bero2> What are you going good to do?
08:59:07 <bero2> Investigate and fix some more clang 3.7/3.8 bits
08:59:23 <bero2> What problems do you have?
08:59:45 <bero2> Way more dependencies on a specific version of llvm than expected
09:00:59 <xruxa> bero2, thanks.
09:01:33 <xruxa> it seems https://projects.linaro.org/browse/LMG-924 is without assignee - anyone interesting in getting id of hardcoded -O flash in Android.mk files?
09:01:53 <liuyq1> Qian__, that's OK, but it's still better to have the problems recorded in somewhere like bugzilla, so that we will know that and can working on that in the feature
09:03:02 <xruxa> #info it seems https://projects.linaro.org/browse/LMG-924 is without assignee - anyone interesting in getting id of hardcoded -O flash in Android.mk files can grab it
09:03:10 <xruxa> #topic AOB
09:03:20 <xruxa> ok, status round done, anything else anyone? bero2 ?
09:04:51 <xruxa> hearing none
09:05:03 <xruxa> lets close the meeting. Thank you all for attending.
09:05:09 <xruxa> #endmeeting