IRC log for #openezx on 20090219

00:48.06WyrMFound it!!! :)
00:48.15WyrMbut have no idea on how to fix ;)
01:17.35*** join/#openezx sn9_ (n=danielg4@gimpelevich.san-francisco.ca.us)
01:48.54*** join/#openezx sn9_ (n=danielg4@gimpelevich.san-francisco.ca.us)
02:00.47*** join/#openezx Fosforo (n=Fosforo@20151053253.user.veloxzone.com.br)
02:30.08*** join/#openezx purl_ (i=ibot@rikers.org)
02:30.08*** topic/#openezx is OpenEZX Developer Lounge -- new gen-blob at http://people.openezx.org/wyrm/gen-blob/gen-blob -- http://openezx.org
02:47.57*** join/#openezx sn9_ (n=danielg4@gimpelevich.san-francisco.ca.us)
06:42.36*** join/#openezx pleemans (n=toi@116.54-246-81.adsl-static.isp.belgacom.be)
07:20.27*** join/#openezx agd5f (n=alex@c-69-143-4-28.hsd1.va.comcast.net)
07:29.58*** join/#openezx pleemans (n=toi@116.54-246-81.adsl-static.isp.belgacom.be)
07:36.59*** join/#openezx gmzhuo (i=gmzhuo@125.70.172.86)
08:39.20Hopscotchgood morning
08:54.00*** join/#openezx gmzhuo (i=gmzhuo@125.69.11.56)
09:13.58*** join/#openezx methril (n=Methril@213.27.233.98)
09:29.07*** join/#openezx gmzhuo (n=gmzhuo@125.69.132.7)
09:31.11*** join/#openezx gmzhuo- (i=gmzhuo@125.69.11.56)
09:41.13*** join/#openezx ao2 (n=u@2001:1418:117:0:0:0:0:1)
10:16.04*** join/#openezx jonwil (n=jonwil@123-243-219-11.tpgi.com.au)
10:18.36jonwilhi
10:20.57*** join/#openezx l403 (n=l@85.132.159.239)
10:59.34*** join/#openezx flyhorse|linux (n=flyhorse@116.4.155.80)
11:07.00*** join/#openezx stefan_schmidt (n=stefan@p5B03621C.dip.t-dialin.net)
13:24.17*** join/#openezx flo_lap (n=fuchs@f048080072.adsl.alicedsl.de)
14:18.11*** join/#openezx johanbr (n=j@blk-7-131-230.eastlink.ca)
15:02.41WyrMI found exactly where the kernel is freezing at resume.
15:02.51WyrMbut I have no idea on how to solve it.
15:02.58WyrMas it is a resume ordering issue.
15:03.09WyrMbetween PCAP/SPI, UDC and MMC.
15:03.30WyrMthe kernel is freezing at the mmc powerup.
15:03.54WyrMwhen it calls ezx_pcap_read/write to set the mmc voltage.
15:04.21WyrMlooks like the SPI is not ready at resume, and that's why it freezes.
15:20.54sabrodHi Wyrm!
15:21.09WyrMhi sabrod.
15:21.11sabrodCongrats!
15:21.20WyrMheh, its not fixed.
15:21.23WyrMbut thanks :)
15:22.59sabrodWell , this already needed enough peregrinations for a congrat!
15:24.36WyrMindeed, this kind of debug is really boring to do.
15:29.09sabrodYes, yesterday i spoke about jtag, you did without it... I would risk any other advise... Too much danger to damage your brain! :)
15:29.37sabrodhumm, WOULDN'T risk! :p
15:30.59WyrMits not a "brain intensive" activity ;)
15:31.09WyrMits more like monkey work. lol.
15:31.49sabrodI as a monkey wouldn't say so... :)
15:37.10WyrMseriously.
15:37.44WyrMthis debug work consists on a lot of cut/paste work to put a lot of
15:38.03WyrM*(unsigned long *)(phys_to_virt(0xa0000000)) = 0x??????????;
15:38.44WyrMon the code, then try to resume wait some seconds, remove the battery and see at the bootloader what value it reports as FLAG.
15:39.07WyrMthis way I can see how far the resume proccess went.
15:39.18WyrMnot hard at all, very repetitive.
15:44.56*** join/#openezx p3t3r__ (n=peter@134.245.164.105)
16:03.15WyrMhah
16:03.22WyrMlooks like this problem is already known.
16:03.34WyrMhttp://osdir.com/ml/linux.kernel.spi.devel/2008-02/msg00123.html
16:03.53WyrMbut it's a year ago, should already be fixed.
16:31.40*** join/#openezx darion76 (i=darion@94-248-121-132.dynamic.peoplenet.ua)
16:37.03*** join/#openezx johanbr (n=j@JBrannlund2.MathStat.Dal.Ca)
17:21.50*** join/#openezx sn9_ (n=danielg4@gimpelevich.san-francisco.ca.us)
18:58.12tmztWyrM: would the new sequencing help here, it looks like the kernel should be able to suspend resume drivers in a proper order but I don't know if it works yet
18:59.19WyrMtmzt: it works, the problem is that there is too much abstraction being done by arch/arm/mach-pxa/devices.c
18:59.51*** join/#openezx pleemans (n=toi@d54C2AAB7.access.telenet.be)
18:59.52WyrMand it doesn't allow me to set the platform_device's parent.
19:00.12WyrMbut im not sure if it is ordering.
19:00.29WyrMlooks more like a bug on pxa2xx-spi.c
19:01.38WyrMI added a pxa_set_mci_parent and pxa_set_udc_parent on devices.c, and set the parents accordingly, but it still freezes on the pxa2xx-spi driver.
19:02.11tmzthow do you know what the parents are?
19:03.50WyrMthe parent is the ezx-pcap spi driver.
19:04.37WyrMI can see the order at which the suspend/resume functions are called by setting pm_test
19:05.29WyrMwith "echo core > /sys/power/pm_test" I can suspend/resume sucessfully, and see the resume order.
19:06.14WyrMbut with pm_test set to "none", it freezes on pxa2xx-spi code.
19:11.30WyrMI found: http://www.mail-archive.com/spi-devel-general@lists.sourceforge.net/msg00779.html and a git-log shows no changes on resume behaviour since this email other than a fix for a typo introduced by a later commit
19:54.07*** join/#openezx sabrod-laptop (n=sab@lns-bzn-45-82-65-152-113.adsl.proxad.net)
20:02.01*** join/#openezx darion76 (i=darion@94-248-7-69.dynamic.peoplenet.ua)
21:25.35*** join/#openezx jonwil (n=jonwil@123-243-219-11.tpgi.com.au)
21:25.58jonwilhi
21:26.45ao2WyrM, have you already tried the quick "not quite right" fix the poster proposed?
21:28.41WyrMao2: no.
21:28.54WyrMao2: but depending on my patience, I will ;)
21:29.39WyrMbut first I will try dumping the registers on suspend and then on resume to see if I can propose a proper fix.
21:31.00ao2we can cheer a bit if you need it
22:09.51*** join/#openezx ecloud (n=srutledg@gw.autoxray.com)
22:20.19*** join/#openezx sn9 (n=danielg4@gimpelevich.san-francisco.ca.us)
23:00.15sabrod-laptopQtopia fw, has a different kct table than your a1200, wyrm.
23:00.31sabrod-laptopDoes the kernel "build" it?
23:00.59sabrod-laptopAccording to it's mtd map? I don't see why...
23:01.47sabrod-laptophttp://pastebin.com/m3eefc3a8
23:01.55WyrMno, who made it probably hex edited the firmware.
23:02.06sabrod-laptopincomplete, but relevent.
23:02.30WyrMthis was generated by your script?
23:02.31sabrod-laptopWell, I'll see the flash file...
23:05.53sabrod-laptopOh, right, that is in the flash files...
23:07.48sabrod-laptopIt seems needed by the rootfs... Or maybe they only did this to change some command line params...
23:22.51*** join/#openezx l403 (n=l@85.132.159.239)
23:30.43WyrMNo, I think that its read by MBM and by RDL.
23:31.05jonwilthat table is used by mbm to tell mbm where all the codegroups are
23:31.11jonwilRDL doesnt use it
23:31.13jonwilAFAIK
23:31.31jonwilIts used for sign checks on magx platform for one
23:32.18WyrMon MAGX it is read by RDL too.
23:32.26jonwilfor what?
23:32.28jonwilmore sig checks?
23:33.05sabrod-laptopHeh, I though rootfs was so a mess, that it prefered to read mtd layout from this KCT table... :)
23:34.36WyrMjonwil: no idea.
23:34.45jonwilok
23:34.57jonwilanyhow, with NoRSA we can ignore it really
23:35.11WyrMthat's what the russian that come up with the norsa patch said.
23:35.21WyrMalso, you can't really ignore this table..
23:35.46jonwilI mean we can ignore it as in we dont need to change it
23:35.49jonwilfor magx anyway
23:35.54WyrMok.
23:36.09jonwilalthough if we need to make some CGs bigger or smaller we may want to edit it
23:36.21WyrMbut very likely, if you change blob's START, or the kernel's START the phone will not boot.
23:36.40jonwilmagx doesnt have blob
23:36.41jonwilit has lbl
23:36.50jonwil:)
23:37.03jonwilbut yeah kernel and lbl wont need to be moved I dont think
23:37.15WyrMif you want to change partitions layout, this table will not help.
23:37.16jonwilsince they wont be getting significantly bigger on magx
23:37.27WyrMand blob or lbl, its just a name.
23:37.31jonwilyeah
23:37.34WyrMits LBL on a1200 too.
23:37.56WyrMand yet, motorola calls it blob on a1200's KCT table.
23:38.20jonwilok, well its called lbl on motomagx
23:38.39WyrMthat's what I saying.
23:38.46WyrMits LBL on ezx gen2 too.
23:38.48WyrMlol.
23:38.51jonwilI mean its called lbl in the cdt table
23:38.58jonwiland not blob
23:39.23jonwileither way it seems to be a simple bootloader derived from full blob of gen1 ezx designed to do stuff previously done by gpl code of blob
23:39.37jonwili.e. a place to put those pieces of bootloader code that are derived from GPL code
23:46.33*** join/#openezx sn9_ (n=danielg4@gimpelevich.san-francisco.ca.us)
23:50.37*** join/#openezx p3t3r__ (n=peter@wlan1.net-you.de)

Generated by irclog2html.pl Modified by Tim Riker to work with infobot.