08:33:08 <bero|2> #startmeeting
08:33:08 <linarobot> Meeting started Tue Dec 22 08:33:08 2015 UTC.  The chair is bero|2. Information about MeetBot at http://wiki.debian.org/MeetBot.
08:33:08 <linarobot> Useful Commands: #action #agreed #help #info #idea #link #topic.
08:33:16 <bero|2> hi all
08:33:20 <bero|2> no big news today...
08:33:40 <zhizhoutian__> hi bero|2
08:33:44 <zhizhoutian__> hi all
08:33:50 <bero|2> if anyone who can attend the sprint hasn't signed up yet, please do today
08:34:18 <bero|2> that's already all updates from my side...
08:34:24 <bero|2> let's go to status updates
08:34:25 <zhizhoutian__> liuyq: can you tell me where is the online document for edit what have done/what to do...
08:34:27 <Qian__> bero|2: Do you mean ShenZhou or BKK?
08:34:50 <Qian__> sorry, ShenZhen
08:34:50 <sunao> hi all
08:34:59 <bero|2> Qian__: Shenzhen. But signing up for BKK in time is important too
08:35:02 <bero|2> so both ;)
08:35:23 <zhizhoutian__> bero|2: where to signed up to attend sprint in ShenZhen?
08:36:03 <bero|2> simply tell Tom you're attending
08:36:18 <Qian__> bero|2, zhizhoutian__: We should sing up?That's what i wants to ask too
08:36:20 <liuyq> zhizhoutian__, https://docs.google.com/document/d/1CgwSksOXW5r6V3M7EPxv5ot-IZNBuTo_h2_OHfOi3aA/edit#
08:36:35 <zhizhoutian__> bero|2: ok, thanks
08:36:40 <zhizhoutian__> liuyq: ok thanks
08:37:26 <zhizhoutian__> Qian__: I have no idea, maybe you should ask your boss;)
08:37:57 <bero|2> sunao: do you want to go first with status updates?
08:38:02 <Qian__> bero|2: OK
08:38:07 <sunao> What did you do?
08:38:09 <sunao> 1.   turn to read the code of glibc to see how the ptmalloc is ported.
08:38:10 <sunao> what are you going to do?
08:38:12 <sunao> 1.  continue porting ptmalloc2.
08:38:13 <sunao> what problems do you have?.
08:38:15 <sunao> None.
08:39:08 <bero|2> good, but please keep in mind that glibc uses a different license from Bionic - so copying code from glibc can be problematic
08:39:13 <Qian__> zhizhoutian__: :)
08:39:23 <bero|2> looking at it for ideas shouldn't be a problem though
08:39:41 <bero|2> Qian__: do you want to go next?
08:39:50 <Qian__> ok
08:40:07 <Qian__> What did you do?
08:40:07 <Qian__> Investigate lmg-966, there are few processes can be swapped their memories out, like uevent, vold, installd, drmserver, healthd
08:40:07 <Qian__> Investigate batterystatsImpl to collect more infos
08:40:09 <Qian__> What are you going to do?
08:40:11 <Qian__> Continue investigate batterystatsImpl to get more info
08:40:12 <sunao> bero|2, thanks for you reminder
08:40:13 <Qian__> Try to swapped out memory of specific process
08:40:15 <Qian__> Look into new mechanism about  swapping activities
08:40:17 <Qian__> What problems do you have?
08:40:19 <Qian__> Does anybody have tried swapped out memory of specific process?
08:41:14 <bero|2> the only way I'm aware of to force swapping a specific process is putting it into its own cgroup and forcing that cgroup's memory limit to something really small
08:41:39 <bero|2> https://wiki.archlinux.org/index.php/Cgroups#Ad-hoc_groups <--- like that
08:42:01 <bero|2> (this is on "normal" Linux, but the principle is the same on Android)
08:42:19 <Qian__> bero|2: But i think we can force swapping out its memories
08:42:31 <bero|2> There might be another way to do it, this just happens to be the only one I know off the top of my head
08:42:57 <Qian__> as we know the information of each process
08:43:59 <Qian__> bero|2: Thanks, it's better to have a second choise
08:44:50 <bero|2> yspan: do you want to go next?
08:45:11 <yspan> What did you do?
08:45:12 <yspan> I asked a leave yesterday, because my daughter got a cold.
08:45:12 <yspan> What are you going to do?
08:45:12 <yspan> Will try to adjust those parameters with LCR-15.12 on Juno
08:45:12 <yspan> What problems do you have?
08:45:12 <yspan> none.
08:45:30 <bero|2> hope your daughter will get well soon!
08:45:39 <yspan> thanks
08:45:56 <bero|2> liuyq: do you want to go next?
08:46:09 <liuyq> What did you do?
08:46:10 <liuyq> Compared the bootchart between Lollipop and Marshmallow.(Lollipop use less boottime and the maxium disk throughput on Lollipop(around 20MB/s) is much more than Marshmallow(more than 100MB): )
08:46:10 <liuyq> Updated CTS package to 6.0_r2.
08:46:10 <liuyq> Checked about problem on Andebenchpro2015
08:46:10 <liuyq> What are you going to do?
08:46:10 <liuyq> Merge the change on cts job templates and follow up the job status
08:46:13 <liuyq> Continue on the automaton of Andebenchpro2015
08:46:14 <liuyq> Submit changes on ARM and Thumb mode switch for pdfium project.
08:46:21 <liuyq> What problems do you have?
08:46:22 <liuyq> Problem to submit changes for review to android-review, also problem to apply patches from android-review.
08:47:06 <bero|2> what's going wrong with android-review? Haven't tried submitting anything the last couple of days...
08:47:21 <bero|2> is it a problem to bug its with or is it more likely a firewall issue?
08:47:33 <liuyq> Qian__, zhizhoutian__, sunao, if you have problem with submitting changes to android-review or problem to apply the android patch, please update this bug: https://bugs.linaro.org/show_bug.cgi?id=1953
08:48:04 <liuyq> bero|2, I think it's only a problem for us in China
08:48:18 <zhizhoutian__> liuyq: ok
08:49:52 <bero|2> hmm, definitely needs fixing
08:50:09 <bero|2> does ssh to hackbox and pushing from there work?
08:50:24 <Qian__> liuyq: Thanks
08:50:54 <liuyq> bero|2, yes, from hackbox works
08:51:25 <bero|2> ok, that's a pretty crappy workaround, but at least a workaround
08:51:49 <bero|2> zhizhoutian__: do you want to go next?
08:52:03 <zhizhoutian__> What did you do?
08:52:03 <zhizhoutian__> * investigate difference between linux4.3 and linux-3.10 in swap module
08:52:03 <zhizhoutian__> - swap out pages without queuing bio request(use bdev_write_page---> zram_rw_page)
08:52:03 <zhizhoutian__> - this should increase execute effect and speed
08:52:03 <zhizhoutian__> * investigate difference between swap to zram and swap to swapfile
08:52:04 <zhizhoutian__> - zram use zram_make_request to handle bio request
08:52:04 <zhizhoutian__> - swapfile use blk_queue_bio to handle bio request
08:52:05 <zhizhoutian__> * see this: https://cacoo.com/diagrams/OGk6QP0jLSNX42EW-ED6EC.png
08:52:05 <zhizhoutian__> * we have got a tshark phone with stable usb connect
08:52:06 <zhizhoutian__> What are you going to do?
08:52:06 <zhizhoutian__> * investigate why swapon use so much memory
08:52:07 <zhizhoutian__> * try to backport linux-4.3 to linux3.10
08:52:19 <zhizhoutian__> none yet
08:53:45 <bero|2> I don't seem to have access to that cacoo URL
08:54:03 <zhizhoutian__> bero|2: let me see...
08:54:04 <Qian__> zhizhoutian__: my company colleage said using LZ4 ot instead of LZO, it may be more quickly
08:54:25 <zhizhoutian__> bero|2: ok, how can i transport a png image to webside?
08:54:54 <bero|2> you can download it to your local box and scp it to people.linaro.org:public_html/
08:55:16 <bero|2> then it will show up as http://people.linaro.org/~zhizhou.tian/whatever.png
08:55:58 <zhizhoutian__> bero|2: ok
08:56:03 <bero|2> eric____: do you want to go next?
08:56:12 <eric____> fine
08:56:35 <zhizhoutian__> Qian__: lz4 is faster than lzo? ok
08:57:07 <bero|2> I think the difference between lz4 and lzo is that lz4 is slightly faster, but slightly less effective in terms of memory saved
08:57:09 <eric____> I've replace all zlib.gyp files with system libs, but there're some errors when running gclient runhooks
08:57:21 <eric____> I
08:57:48 <bero|2> So we should check both of them to see which saves enough memory with decent performance
08:57:49 <eric____> I'm checking all these places, to find if there's some issue when replace it
08:59:33 <Qian__> zhizhoutian__: they said so
09:01:33 <bero|2> wangjian: do you want to go next?
09:02:09 <wangjian> What did you do?
09:02:09 <wangjian> Downloaded AOSP, merged linaro gcc patch on it, removed all hardcode specific -O flag in Android.mk under external/ and framework/ and building.
09:02:09 <wangjian> No more updates, because I was doing other work for our company these days.
09:02:09 <wangjian> What are you going to do?
09:02:09 <wangjian> Test both T-shark and AOSP image size and memory usage after remove the -O flag.
09:02:09 <wangjian> I think I will work on LMG-924 the day after tomorrow because I have other company works to do.
09:02:10 <wangjian> What problems do you have?
09:02:10 <wangjian> None
09:03:09 <bero|2> ok, I think that was everyone who is here...
09:03:12 <bero|2> #topic AOB
09:03:22 <bero|2> does anyone have anything else they'd like to talk about?
09:04:42 <zhizhoutian__> bero|2: this should be ok - http://people.linaro.org/~zhizhou.tian/SwapToZramOrFile.png
09:06:06 <bero|2> zhizhouzian__: yes, thanks. Can't read the comments, but I think it's pretty clear what they're about from the context
09:06:33 <zhizhoutian__> bero|2: oh, yes, sorry..
09:06:51 <bero|2> zhizhoutian__: no problem
09:07:34 <bero|2> I'll figure out some way to OCR  them for translate.google.com or the likes if I can't figure something out
09:08:28 <zhizhoutian__> bero|2: powerful;)
09:08:36 <bero|2> looks like we don't have anything else - so have a great rest of the week and happy holidays for those who have them
09:08:40 <bero|2> #endmeeting