#milkymist IRC log for Tuesday, 2011-07-26

wolfspraula short update from Adam: he has now applied fix2 to 70 boards already, and will finish all of them, probably by the end of today04:01
wolfspraulthen a batch of impedance/power-on/reflash testing, then we see04:01
wolfspraulmost likely that'll be tomorrow04:02
GitHub116[linux-milkymist] larsclausen pushed 1 new commit to master: https://github.com/milkymist/linux-milkymist/commit/54b8f87dadb0cf99ccacfedbc5212ebb747f969807:18
GitHub116[linux-milkymist/master] lm32: Disable interrupts while restoring irq and syscall stackframe - Lars-Peter Clausen07:18
larsclooks as if it is finanly stable, at least my stress tests don't crash anymore07:19
wolfspraulnice, I should mention that in the news I guess07:25
wolfspraulwhich linux version is this actually?07:29
larsc3.007:32
Alarmwhen I load a file "patch" with the command 'first patch', I always get the message 'No patches found' when click on Go?09:10
larsclekernel: uart "thru" mode is loopback?09:12
lekernelyes09:12
lekernelAlarm, use "configured" mode09:13
lekernelor use the patch editor09:13
awrc3 status: all reworks done, next to clean all.09:42
Alarmk thank you. It works.09:44
AlarmFor FTP, DHCP is not working? I set to manual IP address09:44
lekerneldhcp is working, or should be ...09:45
wolfspraulaw: phew, great!09:54
rohuh. aw. you are fast10:12
roh:)10:12
awroh, :) try to test them tomorrow. my soldering status will be approved then. before each board gets test surely I need to visually inspect and measured myself. :)10:15
rohi should hurry up completing the glueing today10:16
awroh, sounds good news from you. :)10:18
wpwrakroh: new job title: "glue sniffer" ;-)10:25
Action: lekernel found a wiki engine that sucks more than mediawiki: the redmine wiki10:52
lekernelin fact, I wonder what is worse, spam dousing or the most obscure syntax ever10:54
Action: methril_work nods about redminke wiki engine11:17
methril_workmorning11:17
GitHub11[uclibc-lm32] larsclausen pushed 4 new commits to master: https://github.com/milkymist/uclibc-lm32/compare/679f42b...71624d511:46
GitHub11[uclibc-lm32/master] Merge branch 'master' of github.com:milkymist/uclibc-lm32 - Lars-Peter Clausen11:46
GitHub11[uclibc-lm32/master] lm32: Drop unused __syscall_return - Lars-Peter Clausen11:46
GitHub11[uclibc-lm32/master] lm32: Cleanup clone - Lars-Peter Clausen11:46
GitHub141[linux-milkymist] larsclausen pushed 3 new commits to master: https://github.com/milkymist/linux-milkymist/compare/54b8f87...989f40511:55
GitHub141[linux-milkymist/master] lm32: Remove incomplete dts - Lars-Peter Clausen11:55
GitHub141[linux-milkymist/master] lm32: Speed up aligned memcpy - Lars-Peter Clausen11:55
GitHub141[linux-milkymist/master] ASoC: Update milkymist platform to the current API - Lars-Peter Clausen11:55
AlarmWhat procedure must be followed to program the keys of RC5 remote control?12:45
wolfspraulAlarm: do you have an RC5 control that works?12:49
Alarmyes12:49
Alarmphilips12:49
wolfsprauldid you test it with m1?12:49
Alarmm1?12:50
wolfspraulyou go to the IR remote dialog, and press a number on your remote, say '5', you should see the value 0x0512:50
wolfspraulMilkymist One12:50
Alarm:)12:50
wolfsprauldid you try that?12:50
wolfspraulyou see the correct number?12:50
Alarmno12:51
wolfspraulI'm asking because the rc5 implementation in the Milkymist SoC right now is very timing sensitive, and will not be as robust/flexible as other rc5 implementations12:51
wolfspraulso it's a bit hyperbole to call it 'rc-5' actually :-)12:51
wolfspraulit's most like "some rc-5" right now12:51
wolfspraulwe changed the timing to match the remote we are including with future units, but now some other remotes don't work anymore12:52
wolfspraulthe best would be a more robust rc-5 implementation but that's a lot of work and nobody sees it as enough of a priority...12:52
Alarmc'est  aleatoire12:52
lekernelon the other hand, it could work with non-RC5 remotes as well by changing the code12:52
wolfspraulAlarm: if you have a philips, maybe if you go back to the old value your remote will work again12:53
Alarmit is an old remote a philips DVD player12:54
wolfsprauldo you know how to build the SoC from source?12:54
wolfspraulyou can try changing that one value back and maybe you are lucky :-)12:54
wpwrakwolfspraul: no, don't make a better rc5 implementation. make a universal detector that doesn't even try to "understand" the code12:55
wolfspraulwe tested a lot of remotes, like 10-15 (all rc5), and with old or new setting _some_ will work, some don't12:55
wpwrakwolfspraul: lekernel promised me one of his old M1 boards if i hack that part. so maybe it'll happen one day ;-)12:56
wolfspraulok the more flexible you want to make it the more work it will be12:56
lekernelnot necessarily12:56
lekernelif it's generic, it can overcome pesky problems like timing variations between remotes12:57
wpwrakmore flexible would actually be much simpler :)12:57
wolfspraulalright, I rest my case12:57
lekernelit's less annoying testing12:57
wolfspraulI'll write it on my Santa Claus wishlist this year12:57
Action: wpwrak dons the red pointy hat13:08
roh   ~.13:08
rohre13:11
awrc3:  clean done. Let's test impedance tomorrow. :-)13:43
larscwhoa, linux has survived 8h of constantly spawning and exiting processes on the mm one board. that's like the first time ever15:16
lekernellarsc, shall I create a repository for openwrt?16:10
lekernellarsc, this is cool, congratulations :)16:10
larscrepo for openwrt would be nice16:15
lekernelhttps://github.com/milkymist/openwrt-milkymist16:20
lekernelyou should have access already16:20
larscthanks16:20
GitHub18[extras-m1] yizhangsh pushed 1 new commit to master: https://github.com/milkymist/extras-m1/commit/5aee94251285c432afdb59ba3b9e35caa3f5f0c516:41
GitHub18[extras-m1/master] add size 285x210 mm brochure artwork. - Yi Zhang16:41
GitHub110[milkymist] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/7abd367bf4baead9a7c7473fa2a2f4efec0577db17:13
GitHub110[milkymist/master] TMU prefetch: flush cache - Sebastien Bourdeauducq17:13
GitHub24[milkymist] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/df573f94f9875cf4a2dcee409a6e0f6b83c644be17:29
GitHub24[milkymist/master] TMU prefetch: fix commit FIFO write indexing - Sebastien Bourdeauducq17:29
GitHub23[extras-m1] yizhangsh pushed 1 new commit to master: https://github.com/milkymist/extras-m1/commit/d54abfc33748cae0e13f5125f4206dd40cd2443417:51
GitHub23[extras-m1/master] added A3 brochure - Yi Zhang17:51
kristianpaullarsc: how wwas the overall performance of your test?18:35
kristianpauli mean did you check memory and cpu stats as well?18:37
larsci did not test for performance18:37
GitHub92[milkymist] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/c219760cbb34842d8790048fba5b9c61ea664a3519:16
GitHub92[milkymist/master] TMU prefetch: faster commit - Sebastien Bourdeauducq19:16
mwallebtw qemu 0.15 is forked19:24
lekerneland does it work?19:51
GitHub84[milkymist] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/639f1242e92a976fb33a7c446eb8fa75f8b9ecaf19:51
GitHub84[milkymist/master] TMU prefetch: even faster commit - Sebastien Bourdeauducq19:51
mwallelekernel: nope, but the patch was sent last week, i'll request a pull for it19:55
GitHub74[milkymist] sbourdeauducq pushed 1 new commit to master: https://github.com/milkymist/milkymist/commit/2c430c95c8a06edd24315943e6974d705c8b10dc20:16
GitHub74[milkymist/master] TMU prefetch: fix qpram read-after-write - Sebastien Bourdeauducq20:16
wpwrakfun idea: if you have a display wall made up of several LCD/plasma screens, perhaps a set of M1 could be used to display things across these screens. would need the ability to pretend do render things off-screen, e.g., if a particle crosses from one screen to the next, M1 #1 would first render it in its visible area, while M1 #2 would pretend to render (tracking the motion) but doesn't show anything. then, when the particle crosses, the20:20
wpwrakroles would be reversed. also needs the ability to synchronize rendering clocks with reasonable accuracy (i guess you could be a few pixels off without anyone noticing, though)20:20
kristianpaulis not that something very common in concerts?20:43
wpwrakshould be, yes. and the equipment is probably quite $$$ :)20:47
wpwrakM1 could be a super low-cost alternative if you can find some suitable panels :)20:47
GitHub167[milkymist] sbourdeauducq pushed 2 new commits to master: https://github.com/milkymist/milkymist/compare/2c430c9...c43799e20:51
GitHub167[milkymist/master] TMU prefetch: fix pipe ack generation in tagmem - Sebastien Bourdeauducq20:51
GitHub167[milkymist/master] TMU: testbench: display mask on bus writes - Sebastien Bourdeauducq20:51
GitHub47[milkymist] sbourdeauducq pushed 1 new commit to hires: https://github.com/milkymist/milkymist/commit/af1f9318b0f7078024d4687a04ceb2d91c6658cf20:59
GitHub47[milkymist/hires] Merge branch 'master' into hires - Sebastien Bourdeauducq20:59
GitHub106[linux-milkymist] mwalle pushed 3 new commits to master: https://github.com/milkymist/linux-milkymist/compare/989f405...9382ce721:51
GitHub106[linux-milkymist/master] lm32: remove debug output in milkymistfb - Michael Walle21:51
GitHub106[linux-milkymist/master] lm32: refactor lm32 architecture - Michael Walle21:51
GitHub106[linux-milkymist/master] lm32: simplify architecture setup - Michael Walle21:51
mwallelarsc: i'm not sure if the memory initialization is correct, see arm/uc32_bootmem_{init,free} for a pretty generic one22:05
GitHub3[linux-milkymist] mwalle pushed 1 new commit to master: https://github.com/milkymist/linux-milkymist/commit/fa90900f345610929ae415ea5e0a5bdb0174384522:22
GitHub3[linux-milkymist/master] lm32: make milkymist gpio driver OF compatible - Michael Walle22:22
--- Wed Jul 27 201100:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!