Previous day | Jump to hour: 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 | Next day

Seconds: Show Hide | Joins: Show Hide | View raw
Font: Serif Sans-Serif Monospace | Size: Small Medium Large

Click in the nick column to highlight everything a person has said.
The Logo icon identifies that the person is a core developer (has commit access).

#rockbox log for 2017-03-26

00:01:04 Quit Moarc (Ping timeout: 240 seconds)
00:01:11 Quit paulk-collins (Read error: Connection reset by peer)
00:02:20 Join Moarc [0] (~chujko@a105.net128.okay.pl)
00:26:45 Join PurlingNayuki [0] (~Thunderbi@61.148.242.216)
00:31:11 Quit PurlingNayuki (Ping timeout: 260 seconds)
00:44:10 Quit Moarc (Ping timeout: 260 seconds)
00:47:27 Join Moarc [0] (~chujko@a105.net128.okay.pl)
01:00
01:00:12 Join MrZeus1 [0] (~MrZeus@2a02:c7f:7066:fb00:910:63f9:9b66:a631)
01:14:16 Quit ZincAlloy (Quit: Leaving.)
01:22:06***Saving seen data "./dancer.seen"
01:22:31 Quit pamaury (Ping timeout: 260 seconds)
03:00
03:01:26 Quit MrZeus1 (Ping timeout: 240 seconds)
03:58:02 Join PurlingNayuki [0] (~Thunderbi@61.148.242.216)
04:00
04:02:24 Quit PurlingNayuki (Ping timeout: 260 seconds)
04:15:54 Quit duo8 (Ping timeout: 246 seconds)
04:22:07***Saving seen data "./dancer.seen"
04:29:24 Join PurlingNayuki [0] (~Thunderbi@61.148.242.216)
04:35:32 Quit PurlingNayuki (Ping timeout: 268 seconds)
04:37:10 Quit smoke_fumus (Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/)
05:00
05:00:50 Quit alexweissman (Remote host closed the connection)
05:21:16 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
05:37:34 Quit alexweissman (Read error: Connection reset by peer)
05:38:07 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
06:00
06:06:48 Quit robertd1 (Ping timeout: 268 seconds)
06:15:30 Quit alexweissman (Remote host closed the connection)
06:15:45 Join duo8 [0] (~ZNC-SRV-H@117.7.201.231)
06:22:10***Saving seen data "./dancer.seen"
06:32:17 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
06:36:31 Quit alexweissman (Remote host closed the connection)
06:38:01 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
06:49:20 Quit TheSeven (Disconnected by services)
06:49:28 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
07:00
07:27:58 Quit alexweissman (Remote host closed the connection)
07:34:47 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
07:38:49 Quit alexweissman (Remote host closed the connection)
08:00
08:22:11***Saving seen data "./dancer.seen"
08:49:24 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
08:53:47 Quit alexweissman (Ping timeout: 240 seconds)
09:00
09:23:39 Quit Jinx (Ping timeout: 240 seconds)
09:48:39 Join PurlingNayuki [0] (~Thunderbi@61.148.242.216)
10:00
10:22:15***Saving seen data "./dancer.seen"
10:32:35 Join lebellium [0] (~chatzilla@89-93-177-91.hfc.dyn.abo.bbox.fr)
10:40:42 Join petur [0] (~petur@rockbox/developer/petur)
10:49:45 Join paulk-collins [0] (~paulk@gagarine.paulk.fr)
11:00
11:10:13 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
11:47:16 Join Guest24460 [0] (~gitterast@gsr198.internetdsl.tpnet.pl)
11:56:44 Quit Guest24460 (Quit: My laptop has gone to sleep. ZZZzzz�)
12:00
12:22:17***Saving seen data "./dancer.seen"
12:32:35 Quit PurlingNayuki (Ping timeout: 256 seconds)
13:00
13:11:01 Quit paulk-collins (Quit: Leaving)
13:11:20 Join paulk-collins [0] (~paulk@gagarine.paulk.fr)
13:53:45 Quit diox (Ping timeout: 264 seconds)
13:56:47 Quit rela (Ping timeout: 240 seconds)
13:57:11 Join diox [0] (~u@c80-216-194-28.bredband.comhem.se)
14:00
14:00:21 Join rela [0] (~x@pdpc/supporter/active/rela)
14:06:25 Join MrZeus1 [0] (~MrZeus@2a02:c7f:7066:fb00:e4e3:80e5:2592:a214)
14:10:46lebelliumpamaury: look like you have been much busy lately :)
14:14:52pamaurylebellium: yes, I have been away several times for work, I have had interviews for job and deadlines.
14:15:04pamauryAnd April will be similar
14:15:34 Join wodz [0] (~wodz@89-74-169-198.dynamic.chello.pl)
14:16:14wodzpamaury: Do you have an idea how to trace why rb dies at startup on agptek?
14:16:53lebelliumLeo told me that based on my suggestion they are working on a new Rocker with an "HD" screen :)
14:17:47pamaurywodz: does it die before main ? Did you try simpler programs than rockbox ?
14:17:54pamaurytry with strace
14:19:25wodzpamaury: It dies before main, yes. I tried simple programs and this works (and I can gdb them).
14:19:34wodzpamaury: strace doesn't show up anything
14:21:45pamaurywodz: and it doesn't print any error message in the console?
14:21:50pamaurydid you look in dmesg?
14:22:18***Saving seen data "./dancer.seen"
14:23:17wodzpamaury: console says simply 'Killed.'
14:24:12wodzpamaury: BTW the problem with setrlimit64/getrlimit64 is related with 'our' glibc compiled with large files support.
14:31:47 Quit michaelni (Ping timeout: 240 seconds)
14:32:57pamauryI don't see many possibilities for a kill before main(). Either a missing symbol (but I would expect an error message) or too much memory allocation in bss for example
14:35:58wodzI was thinking about ulimit thing but I'd expect some error message to give a clue about exhausted resource
14:36:52 Quit alucryd (Remote host closed the connection)
14:37:58 Join alucryd [0] (~quassel@archlinux/developer/alucryd)
14:38:27pamaurywodz: what is the size of the bss section ? I don't remember how memory is handled on hosted
14:39:26wodzpamaury: according to readelf it is 211d844 which is enormous
14:41:37pamauryhum
14:41:49pamaurywhat is the memory size configured in configure ?
14:42:25wodzmemory=32
14:42:42pamauryhow much memory does the device have ?
14:42:46wodz32
14:42:51pamaurythen it's way to much
14:43:12pamaurythe kernel takes several megabytes already, not even counting the other programs
14:43:14pamaurytry with 16
14:44:36wodzstill 211d844 is waaaaay over 32MB
14:45:21pamauryisn't it like 33MB ?
14:45:24wodzwith memory=16 readelf reports 111d254 for bss
14:45:54wodzhmm indeed
14:50:18wodzHa, now error is different
14:50:39wodz/mnt/sd_0/rockbox: /usr/lib/libasound.so.2: no version information available (required by /mnt/sd_0/rockbox)
14:50:49wodzpamaury: ^
14:51:52pamauryweird
14:52:21pamaurygive me a minute
14:52:33 Quit cc___ (Quit: WeeChat 1.6)
14:53:39 Join cc___ [0] (~ac@2001:910:113f:1:6a05:caff:fe1c:1627)
14:54:37wodzhmm, looking at strace it looks like it dies on some mmap
14:54:38pamaurywodz: it's a bit unusual
14:54:50pamaurylibasound on agptek doesn't have symbol versions
14:54:50 Join michaelni [0] (~michael@213-47-41-20.cable.dynamic.surfer.at)
14:56:22pamaurymaybe there is an option to disable symbol versioning, although I am not sure why one would do that
14:57:59pamaurywodz: apparently one can add −−with-versioned=no to alsa-lib configure
14:59:08wodzIs missing symbol versioning error or warrning?
14:59:10pamauryone annoying thing is that I don't know if an unversioned works on versioned. If not, that would mean we need two different alsa-lib versions
15:00
15:00:47pamaurywodz: I updated g#1570
15:00:49fs-bluebotGerrit review #1570 at http://gerrit.rockbox.org/r/1570 : Modernize toolchain script and add generic arm toolchain by Amaury Pouly
15:00:54pamauryI haven't tried it
15:01:45pamauryyou don't have to rebuild everything
15:01:50wodzpamaury: I reduced memory further to 8 and now I can run binary in gdb and get backtrace. http://paste.debian.net/924411/
15:02:40pamaurywodz: ah so it's actually just a warning, good
15:03:00wodzI know why it tries to throw panicf() at least
15:03:51pamauryah?
15:05:30wodzI was opening battery status file with O_RDWR, but this file is read only
15:06:05pamauryyou will probably have to play with memory size to find a good value
15:06:16pamaurymaybe 12
15:06:55 Join JanC_ [0] (~janc@lugwv/member/JanC)
15:07:47wodzI'll stick with 8 for now as having ability to run in gdb is really useful
15:08:10 Quit JanC (Killed (karatkievich.freenode.net (Nickname regained by services)))
15:08:10 Nick JanC_ is now known as JanC (~janc@lugwv/member/JanC)
15:09:41wodzNow it dies on switch_thread()
15:10:02wodzbut I have rockbox logo on screen!
15:10:14pamaurywodz: gdb might not like switch_thread
15:10:23wodzgood point
15:10:35pamaurybecause we use signals to switch threads I think
15:10:57wodzwithout gdb rockbox segfaults as well
15:17:34pamauryI don't know much about threading
15:17:39pamaurymaybe jhMikeS knows
15:17:51 Quit jhMikeS (Ping timeout: 258 seconds)
15:21:22wodzhe was apparently scared by the question :P
15:37:19wodzpamaury: ok, I know what is direct cause of crash in switch_thread() BUT I have no clue why it is like that. thread->stack is NULL and hence stack overflow check makes NULL ptr dereference. http://paste.debian.net/924415/
15:41:14pamauryI know nothing about the threading code, and it's quite complicated :-/ One possibility I see is that you are creating/using threads before the threading code is initialised?
15:43:47wodzThat was my thought as well. Need to step through main() to see what it calls actually
15:46:52 Quit preglow (Read error: Connection reset by peer)
15:47:08 Join preglow [0] (~thomj@2001:840:4243:3::101)
15:50:00wodzpamaury: What sets stackbegin and stackend on hosted?
15:52:26pamaurywodz: I think it's set in ../firmware/target/hosted/samsungypr/<device>/system-<device>.c:
15:52:41pamaury../firmware/target/hosted/samsungypr/ypr1/system-ypr1.c for example
15:52:54wodzI see. How does it actually work?
15:52:54pamauryit's supposed to be completely irrelevant for hosted
15:53:46wodzpamaury: But __get_main_stack() does *stacksize = (uintptr_t)stackend - (uintptr_t)stackbegin;
15:54:21wodzhmm, maybe stacksize can be 0 but actual pointer must be valid
15:54:23pamaurybut is the result actually used on hosted? I know nothing about those. Just that some comments says it's not used I think
15:56:42 Quit petur (Quit: Leaving)
16:00
16:00:02wodznow it crashes in some speex code
16:00:20 Quit zoktar (Ping timeout: 240 seconds)
16:00:40 Join Strife1989 [0] (~quassel@adsl-98-80-187-37.mcn.bellsouth.net)
16:02:29 Quit Strife89 (Ping timeout: 268 seconds)
16:02:39 Join zoktar [0] (~zoktar@78-72-39-137-no186.tbcn.telia.com)
16:02:39 Quit zoktar (Changing host)
16:02:40 Join zoktar [0] (~zoktar@unaffiliated/zoktar)
16:07:18 Quit zoktar (Ping timeout: 258 seconds)
16:09:05 Join zoktar [0] (~zoktar@78-72-39-137-no186.tbcn.telia.com)
16:09:05 Quit zoktar (Changing host)
16:09:05 Join zoktar [0] (~zoktar@unaffiliated/zoktar)
16:19:33 Join PurlingNayuki [0] (~Thunderbi@61.148.242.216)
16:22:21***Saving seen data "./dancer.seen"
16:23:00 Part Asara
16:30:18wodzpamaury: After commenting out HAVE_TAGCACHE so tagcache_init() is skipped I have crash in lcd_alpha_bitmap_part_mix() which looks like some memory corruption so something is fundamentally broken
16:45:34 Join ender [0] (krneki@foo.eternallybored.org)
16:48:12 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
16:48:21 Quit MrZeus1 (Ping timeout: 264 seconds)
16:56:08 Quit PurlingNayuki (Ping timeout: 264 seconds)
17:00
17:05:58pamaurywodz: I don't know, it works like a charm on nwz...
17:10:25pamauryth only obvious difference is arm vs mips
17:11:07 Quit wodz (Ping timeout: 240 seconds)
17:17:14duo8tfw x1000 going faster than jz
17:17:40 Quit zoktar (Ping timeout: 240 seconds)
17:18:10 Join smoke_fumus [0] (~smoke_fum@dynamic-vpdn-46-53-165-217.telecom.by)
17:18:16 Join zoktar [0] (~zoktar@78-72-39-137-no186.tbcn.telia.com)
17:18:16 Quit zoktar (Changing host)
17:18:16 Join zoktar [0] (~zoktar@unaffiliated/zoktar)
17:29:47 Join wodz [0] (~wodz@89-74-169-198.dynamic.chello.pl)
17:47:42wodzpamaury: In order to exclude 32bit lcd driver issue, I set it up as 24bit driver (knowing lcd will be garbage) but rb still crashes. It crashes on dereferencing pointer which is crap (like NULL or 0x1) in speex code but I think the corruption must be much earlier and speex just happen to trigger that
17:51:10pamaurywodz: if the corruption is visible with 32bit lcd, that means it happens early on boot, you should investigate with this, take a pointer that becomes corrupted and narrow down the last places where is was still ok and then bad
17:52:27wodzpamaury: with 32bit driver rockbox logo is displayed correctly and then it goes mad somewhere in tagcache_init() afaik
17:56:34wodzpamaury: I don't really understand the thing with stackbegin and stackend. IMO it works purely by luck. If you define stackbegin to some variable allocated on the stack in system_init() it becomes invalid just after returning from this function.
18:00
18:00:32pamaurywodz: my understanding is that stackbegin and stackend are only used to get the stack size on native
18:00:51pamauryand then check for stack overflow I guess
18:01:05pamauryI just copied the code from other hosted targets
18:01:10wodzpamaury: yes, stackoverflow is checked on hosted as well.
18:01:35wodzpamaury: I know but IMO this is simply wrong
18:01:48wodzand misleading
18:02:22pamaurywodz: hosted doesn't check for stack overflow on main thread because if you set stackbegin=stackend, then stacksize=0 and the stkov code ignores 0 size stacks
18:02:36pamauryon hosted you cannot get the pointer and size of the main stack anyway
18:03:03pamaurythough I agree that the proper approach would be to handle this in thread.c and not in every target
18:04:26pamauryI might be wrong of course, I don't understand much of thread.c
18:04:46pamaurybut since all hosted targets do that and all work except x1000, I assume that the problem is not there
18:04:50wodzgdb disagrees with you. In switch_thread() there is line if (UNLIKELY(thread->stack[0] != DEADBEEF) && thread->stack_size > 0) which means that thread->stack[0] will be checked even when thread->stack_size == 0
18:06:12wodzI agree that problem is elsewhere but It took me fair bit of time to understand why it crashed in switch_thread()
18:06:32pamauryah you are right, this condition seems like a stupid overlook, you should exchange the two conditions order
18:06:42wodzyep
18:07:05pamauryit's a bit strange though
18:07:23pamaurybecause on nwz, it returns an address which is on the stack, so this address will always be valid
18:07:34pamaury(even if it's content is garbage)
18:09:32wodzit is valid in the sense it is on the stack but don't you think this is stupid to return pointer to random stack part and check if this is not DEADBEEF?
18:09:41 Join ZincAlloy [0] (~Adium@p57A72E5A.dip0.t-ipconnect.de)
18:09:52pamaurywodz: I think the check in thread.c is stupid
18:10:12pamauryit's the classic if(array[0] == .... && array_size > 0)
18:10:23pamauryie checking content before checking size
18:12:44pamaurybut actually, it seems to me now that we should just set stackbegin = stackend = 0 in this case
18:12:57pamaurysince it's supposed to be useless
18:22:25***Saving seen data "./dancer.seen"
18:26:01*__builtin gets Ballerburg up and running on SDL
18:27:32__builtinthe big issue is that everything is hardcoded for 640x480 screen res
18:28:09__builtinso for now I can only see 25% of the screen :D
18:37:53wodzYou are insane with multiplying number of plugins. You should be responsible for defining keymaps for plugins on each new port :P
18:38:43__builtinI feel the ad-hoc keymaps need to be centralized
18:39:12wodz__builtin: feel free to propose something
18:39:58__builtinPLA is great and all for simple stuff, but when it gets more involved (i.e. simultaneous presses), keymaps are the only way to go
18:41:02wodzI know. I know also that defining all keymaps for new port is nightmare
18:41:12__builtinI think there's a lot of duplication of essentially the same keymap between plugins though
18:41:35__builtini.e. 4 directions + action + menu
18:42:32wodz+ quit
18:42:59__builtinit would make life easier for everyone to have these definitions in a centralized header somewhere
18:48:45__builtinthis probably wouldn't replace PLA, but instead provide a more flexible version of it
18:50:25wodz__builtin: BTW maybe you are interested in finishing g#1389 ? I think working tts will be highly appreciated feature.
18:50:27fs-bluebotGerrit review #1389 at http://gerrit.rockbox.org/r/1389 : Port of picoTTS by Marcin Bukat
18:53:43__builtinI'll take a look
18:54:57wodz__builtin: tts engine basically works, audio glue code is f*** up. I had simpler version which worked but I lost it somehow. The code is borrowed from mpgplayer actually.
18:56:26pamaury__builtin: I'm all for simplifying keymaps for plugins, but if you do, please please please, look at all targets before doing that. Some of them have crazy keys so we need a flexible system
18:56:39pamauryalso some plugins have ways to many keys required, that's a nightmare
18:57:56__builtinI'll probably just leave those that require a bunch of keys alone for the time being
18:58:34wodzIMO the most desired functionality would be per plugin enable for each target. So you could define keymap and enable plugin one by one and not all or nothing
18:59:23__builtinand they would be enabled in the target header?
18:59:36__builtinsuch as PLUGIN_ENABLE_DOOM
18:59:38__builtin?
19:00
19:01:33pamauryno please not the target header, and possibly not new defines
19:01:51wodzI don't know how to do it cleanly
19:02:05pamauryI don't know how to do it cleanly but it has to be clean
19:02:47pamauryI would actually suggest to come up with a header that somehow lists plugins to NOT compile
19:03:43pamaurybecause the goal at the end of the day is to compile everything, you only want to do work if you want to disable something imo
19:14:45wodzpamaury: It looks like threading code for mips doesn't work as it should on hosted
19:18:19pamaurynot sure what you mean
19:19:16wodzpamaury: the corruption happens on load_context() in rockbox/firmware/asm/mips/thread-mips32.c
19:21:29pamaurywodz: I'm a bit confused, I thought on hosted it's not using this code
19:21:51pamauryit's supposed to use thread-unix.c
19:22:08wodzThat would explain a lot.
19:22:25wodzThere is some magic define needed somewhere
19:22:47pamaurywodz: probably asm.make
19:23:36wodzpamaury: I don't see anything hosted specific there
19:23:42pamaurylet me see
19:25:18pamaurywodz: it's a simple fix actually
19:25:21wodzHAVE_SIGALTSTACK_THREADS should be defined somewhere
19:25:24pamaurylook at firmware/asm/arm/thread.h
19:25:31pamauryit takes into account hosted
19:25:38pamaurythe asm/mips/thread.h does not
19:25:46pamauryI think
19:26:14pamauryhum, I may have spoken too quickly
19:26:14wodzit only include <errno.h> nothing more. It is needed on hosted
19:26:41pamauryso thread.c -> thread-internal.h -> asm/thread.h
19:27:12pamauryand thread.c -> asm/thread.c
19:27:48pamauryand asm/thread.c uses thread-unix.c if HAVE_SIGALTSTACK_THREADS is defined
19:28:07pamaurybut I can't find where HAVE_SIGALTSTACK_THREADS is defined, maybe in configure
19:28:10wodzaaaah, it should be defined in tools/configure
19:30:30pamauryI don't really understand in configure though
19:32:23pamaurywodz: actually on the nwz port, autoconfig.h defines ASSEMBLER_THREADS
19:32:42pamauryand the ypr0 code doesn't seem to do anything special
19:33:33pamauryso, either 1) the arm code somehow works on both native and hosted, but not the mips one, or 2) it's an overlook and it works on arm purely by chance
19:33:50wodzautoconfig.h? I don't have such file here (in build dir)
19:34:17pamauryautoconf.h
19:34:23pamaurythat's the file generated by configure
19:34:37wodzah, no I have it
19:34:56wodz#define ASSEMBLER_THREADS
19:35:23wodzso it looks like on arm it is working on hosted as well
19:35:41wodzI am wondering how threading worked on android mips
19:35:43pamauryit might be that the mips code is making some assumption not true on hosted
19:35:57wodzgevaerts: ping
19:36:48*gevaerts fears a question he doesn
19:36:51*pamaury is looking at thread-mips32.c but fails to understand the code
19:36:55gevaertst know the answer to!
19:37:55wodzgevaerts: the question is how autoconf.h looks like on android mips. Which threading model is used?
19:38:20 Join TheLemonMan [0] (~root@irssi/staff/TheLemonMan)
19:38:32wodzI don't have Android SDK and NDK installed to check
19:38:48gevaerts#define HAVE_SIGALTSTACK_THREADS
19:39:05wodzaha
19:39:21pamaurymaybe we should understand why the assembler thread code is not working on hosted ?
19:39:34pamaury(whereas it does for arm)
19:40:17wodzpamaury: I guess we under store registers since rockbox native assumes cooperative threading so it might not save some registers
19:41:00pamaurythat's also the case on arm
19:41:16pamaurywe do cooperative threading on hosted arm
19:44:00wodzlooks like it is not storing s8 register
19:44:34pamaurys8?
19:44:36wodzhmm, maybe not, s8 is aliased as fp
19:44:37pamauryyou mean fp
19:45:06wodzaccording to this: https://www.student.cs.uwaterloo.ca/~cs350/W11/notes/threads.pdf the store part is correct
19:46:53 Join MrZeus1 [0] (~MrZeus@2a02:c7f:7066:fb00:555d:2828:5e03:593e)
19:48:29pamauryI am a bit confused about start_thread
19:49:13pamauryI don't understand why it loads from $8 and $9
19:49:36pamauryah ok because of load_context
19:52:02wodzI set breakepoint in load_context() and start param points in the middle of sb_decode() while it is supposed to be "scroll" thread. something is really fishy
20:00
20:01:25pamaurywodz: you mean context->start is completely wrong ?
20:05:23wodzyes
20:06:07pamaurywodz: what about the context pointer itself ?
20:06:30wodzlooks valid
20:07:23pamaurywodz: what is value of context->r[0] and context->r[1] ?
20:16:34wodzcontext->r[0] = 0xe2712c, context->r[1] = 0x4a9ea8
20:16:52pamaurywodz: r[0] should be a pointer to the context itself
20:16:59pamauryand r[1] to the main thread function
20:17:09pamauryat least on thread startup
20:17:29wodz(struct thread_entry *) 0xe2712c <__thread_entries+156>
20:17:43wodzso context->r[0] indeed points to self
20:17:54pamauryyeah, good. And r[1] ?
20:18:06pamauryand context->start
20:18:18wodzyeah, it points to scroll_thread()
20:18:43wodzcontext->start is bogus
20:19:03pamaurywodz: can you dump the code of the function start_thread() ?
20:19:15pamauryand _start_thread()
20:19:24pamaurythe code is a bit weird, it seems to be a huge hack
20:19:34wodzyou mean objdump rockbox.elf?
20:20:19pamauryyeah
20:20:27wodzhttp://paste.debian.net/924462/
20:20:40pamauryI don't understand why the code doesn't simply put naked attribute
20:21:13pamauryok looks correct
20:21:24pamauryso context->start doesn't point to 004b959c ?
20:21:55wodzMaybe naked was broken at that time.
20:22:04wodzno, context->start = 0x4c0000
20:22:26pamaurywodz: can you try to rename _start_thread to start_thread, add naked attribute
20:22:30***No seen item changed, no save performed.
20:22:32pamauryie like asm/arm/thread.c
20:23:56wodzsure, its just I need to get kids to bed so it will take some time
20:25:46 Quit MrZeus1 (Ping timeout: 240 seconds)
20:27:31wodzpamaury: warning: ‘naked’ attribute directive ignored
20:27:41wodzpamaury: So I guess thats the answer
20:28:45pamauryinteresting
20:29:34pamauryand what does 0x4c0000 correspond to ? (in rockbox.map)
20:31:30wodzpamaury: it is in the middle of some speex function
20:32:05TheLemonManlooks like the bottom 16 bits are missing, since this is MIPS I'd suspect there's a missing ori/addi following a lui heh
20:32:33wodzpretty possible
20:32:33pamaurywodz: maybe the compiler emits the wrong type of relocation for some reason
20:33:12pamaurywodz: mybe you could try to put this start_thread function in .S file with proper .global start_thread and all
20:33:24wodzbingo! Can't find matching LO16 reloc against `.text' for R_MIPS_GOT16 at 0x132c in section `.text'
20:33:50pamaurywodz: maybe try to add .global start_thread in the asm block ?
20:34:02*pamaury thinks defining this function in C code is a hack
20:39:35wodznop, that doesn't work. IMO this should be moved to .S as it is pure asm anyway
20:40:43pamauryagreed
20:41:00pamaurywodz: alternatively, you could turn start_thread into a proper function
20:41:10pamauryand make load_context load the context into a0
20:41:19wodzthat too
20:41:39wodzactually why it doesn't follow calling convention actually?
20:42:06pamauryno idea, seems a bit unatural
20:43:42pamaurywodz: something like this pamaury/48e17fbd5ee7c1078d1eea80e5d98c2d">https://gist.github.com/pamaury/48e17fbd5ee7c1078d1eea80e5d98c2d ?
20:48:55wodzcrashes in switch_thread()
20:49:34pamaurymaybe I got the code wrong
20:49:45pamauryis the address in context->start correct now ?
20:50:58wodzlet me see
20:55:01wodzyes, context->r[0], context->r[1] and context->start seems valid
20:56:35wodzhmm, ra is 0
20:56:36pamauryI must have screwed up something
20:57:35pamaurywodz: can you objdump start_thread ?
20:58:14wodzhttp://paste.debian.net/924467
20:59:46pamaurythis code is crazy stupid
20:59:56wodzit compiled with -O0
21:00
21:01:38pamauryI don't see what is wrong
21:01:46pamauryyou might have to single step from load_context to catch the problem
21:09:22 Quit wodz (Ping timeout: 258 seconds)
21:10:46 Quit ZincAlloy (Quit: Leaving.)
21:12:14 Join wodz [0] (~wodz@89-74-169-198.dynamic.chello.pl)
21:18:52wodzpamaury: After full recompile it seems to be working (i.e I get proper panicf() from my code)
21:21:26 Join ZincAlloy [0] (~Adium@p57A72E5A.dip0.t-ipconnect.de)
21:21:35pamaurywodz: with old code or new code ?
21:59:36wodzwith your patch in start_thread
22:00
22:17:15 Join webguest194 [0] (~5d41e969@www.haxx.se)
22:20:46 Quit wodz (Ping timeout: 256 seconds)
22:22:10 Quit webguest194 (Quit: CGI:IRC (Ping timeout))
22:22:34***Saving seen data "./dancer.seen"
22:23:54 Join wodz [0] (~wodz@89-74-169-198.dynamic.chello.pl)
22:27:22*wodz \o/ main menu reached
22:28:20lebellium:)
22:36:43pamaurywodz: great :)
22:37:09wodzpamaury: now I get crash when entering 'Files' from main menu :P
22:40:13pamaurywhere is the crash ?
22:42:10wodzpamaury: http://paste.debian.net/924492
22:42:54wodzbut this is strange since the offending line is dptr->attr = info.attribute and in gdb I can do set dptr->attr = info.attribute just fine
22:45:23__builtinany german speakers here?
22:45:43 Quit preglow (Ping timeout: 252 seconds)
22:47:22*[Saint] summons pixelma
22:47:40__builtinmmh, google translate barely works
22:48:07[Saint]Yeah, Google translate for non-germanic language -> germanic language is hilarious.
22:48:18[Saint]For bonus points, translate it back and forth a few times.
22:48:29[Saint]It usually becomes garbage.
22:48:54wodz[Saint]: you could try this trick with slavic language too :P
22:55:55DEBUGEOF from server (Connection reset by peer) (snapshot: netstuff.c line 545)
22:55:55***Cleanup
22:55:55***Cleanup
22:55:55***Saving seen data "./dancer.seen"
22:55:55***Exit
22:55:57***Started Dancer V4.16
22:55:57***Connected to irc.freenode.net on port 6667
22:55:57***Logfile for #rockbox started
22:55:57Mode"logbot :+i" by logbot
22:56:04***Server message 501: 'logbot :Unknown MODE flag'
22:56:04 Join logbot [0] (~rockbox@giant.haxx.se)
22:56:04 Join aevin [0] (eivindsy@unaffiliated/aevin)
22:56:04 Join froggyman [0] (~frogs@unaffiliated/froggyman)
22:56:04 Join wodz [0] (~wodz@89-74-169-198.dynamic.chello.pl)
22:56:04 Join ZincAlloy [0] (~Adium@p57A72E5A.dip0.t-ipconnect.de)
22:56:04 Join TheLemonMan [0] (~root@irssi/staff/TheLemonMan)
22:56:04 Join zoktar [0] (~zoktar@unaffiliated/zoktar)
22:56:04 Join smoke_fumus [0] (~smoke_fum@dynamic-vpdn-46-53-165-217.telecom.by)
22:56:04 Join alexweissman [0] (~alexweiss@c-68-51-123-75.hsd1.in.comcast.net)
22:56:04 Join ender [0] (krneki@foo.eternallybored.org)
22:56:04 Join Strife1989 [0] (~quassel@adsl-98-80-187-37.mcn.bellsouth.net)
22:56:04 Join JanC [0] (~janc@lugwv/member/JanC)
22:56:04 Join michaelni [0] (~michael@213-47-41-20.cable.dynamic.surfer.at)
22:56:04 Join cc___ [0] (~ac@2001:910:113f:1:6a05:caff:fe1c:1627)
22:56:04 Join alucryd [0] (~quassel@archlinux/developer/alucryd)
22:56:04 Join rela [0] (~x@pdpc/supporter/active/rela)
22:56:04 Join diox [0] (~u@c80-216-194-28.bredband.comhem.se)
22:56:04 Join paulk-collins [0] (~paulk@gagarine.paulk.fr)
22:56:04 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
22:56:04 Join [7] [0] (~quassel@rockbox/developer/TheSeven)
22:56:04 Join duo8 [0] (~ZNC-SRV-H@117.7.201.231)
22:56:04 Join Moarc [0] (~chujko@a105.net128.okay.pl)
22:56:04 Join Bilgus [0] (~Bilgus@gateway/tor-sasl/bilgus)
22:56:04 Join sparetire [0] (~sparetire@unaffiliated/sparetire)
22:56:04 Join fs-bluebot [0] (~fs-bluebo@xd9baf2b7.dyn.telefonica.de)
22:56:04 Join bluebrother [0] (~dom@rockbox/developer/bluebrother)
22:56:04 Join dys [0] (~dys@109.40.0.59)
22:56:04 Join ender` [0] (krneki@2a01:260:4094:1:42:42:42:42)
22:56:04 Join ParkerR [0] (ParkerR@unaffiliated/parkerr)
22:56:04 Join zoid [0] (~zoid@unaffiliated/taxationistheft)
22:56:04 Join Rower [0] (husvagn@d83-183-134-99.cust.tele2.se)
22:56:04 Join [Saint] [0] (~sinner@rockbox/staff/saint)
22:56:04 Join tchan [0] (~tchan@lunar-linux/developer/tchan)
22:56:04 Join amiconn [0] (~amiconn@rockbox/developer/amiconn)
22:56:04 Join pixelma [0] (~pixelma@rockbox/staff/pixelma)
22:56:04 Join prg318 [0] (~prg318@deadcodersociety/prg318)
22:56:04 Join idonob [0] (~Owner@S010600259c3e7d7b.vs.shawcable.net)
22:56:04 Join Jack87 [0] (~Jack87@nasadmin/admin/jack87)
22:56:04 Join Petri152 [0] (~Petri@petritrebs.ca)
22:56:04 Join The_Prospector [0] (~The_Prosp@unaffiliated/cornman)
22:56:04 Join puckipedia [0] (puck@puckipedia.com)
22:56:04 Join vifino- [0] (~vifino@tty.sh)
22:56:04 Join utrack_ [0] (~utrack@21422.s.t4vps.eu)
22:56:04 Join funman [0] (~fun@rockbox/developer/funman)
22:56:04 Join munch [0] (pls@gateway/shell/elitebnc/x-elfucwuiphdaphfn)
22:56:04 Join x56 [0] (~0x56@unaffiliated/x56)
22:56:04 Join Horrorcat [0] (129994c4c6@unaffiliated/horrorcat)
22:56:04 Join dfkt [0] (~dfkt@unaffiliated/dfkt)
22:56:04 Join quaz0r [0] (quaz@c-67-183-243-24.hsd1.wa.comcast.net)
22:56:04 Join Amboyna [0] (~TorC@fsf/member/TorC)
22:56:04 Join kugel [0] (~kugel@rockbox/developer/kugel)
22:56:04 Join rasher [0] (~rasher@rockbox/developer/rasher)
22:56:04 Join foolsh [0] (~starchase@162-204-199-234.lightspeed.sbndin.sbcglobal.net)
22:56:04 Join St0neHead [0] (stonehead@2a01:7e00:e001:3700:6667::2)
22:56:04 Join mercutio [0] (ben@pearl.meh.net.nz)
22:56:04 Join gevaerts [0] (~fg@rockbox/developer/gevaerts)
22:56:04 Join madk [0] (~noneofyou@107.170.140.154)
22:56:04 Join igitoor [0] (igitur@2a00:d880:3:1::c1ca:a648)
22:56:04 Join akaWolf [0] (~akaWolf@akawolf.org)
22:56:04 Join Ruhan [0] (uid76353@gateway/web/irccloud.com/x-dxgqqpstcysrghut)
22:56:04 Join MrZeus [0] (~MrZeus@81.144.218.162)
22:56:04 Join Galois [0] (djao@efnet.math.uwaterloo.ca)
22:56:04 Join Sockbat [0] (sockbat@gateway/shell/fnordserver.eu/x-zieimjmizsfjlhat)
22:56:04 Join GeekShadow [0] (~antoine@reactos/tester/GeekShadow)
22:56:04 Join JdGordon [0] (~jonno@rockbox/developer/JdGordon)
22:56:04 Join user890104 [0] (Venci@unaffiliated/user890104)
22:56:04 Join olspookishmagus [0] (~pookie@snf-137798.vm.okeanos.grnet.gr)
22:56:04 Join prof_wolfff [0] (~prof_wolf@62.83.49.172.dyn.user.ono.com)
22:56:04 Join dan- [0] (~d@freenode/corporate-sponsor/privateinternetaccess.com/doaks)
22:56:04 Join mc2739 [0] (~mc2739@rockbox/developer/mc2739)
22:56:04 Join shmibs [0] (~shmibs@shmibbles.me)
22:56:04 Join advcomp2019 [0] (~advcomp20@unaffiliated/advcomp2019)
22:56:04 Join scorche [0] (~scorche@rockbox/administrator/scorche)
22:56:04 Join tomflint [0] (~tomflint@unaffiliated/tomflint)
22:56:04 Join derf [0] (~derf@static-108-18-126-14.washdc.fios.verizon.net)
22:56:04 Join neutric [0] (~ah@unaffiliated/neutric)
22:56:04 Join hiro [0] (~hiro@2001:41d0:8:e46:5054:ff:feac:f713)
22:56:04 Join naraic [0] (naraic@spoon.netsoc.tcd.ie)
22:56:04 Join bzed [0] (~bzed@shell.bzed.at)
22:56:04 Join GodEater [0] (~whoknows@rockbox/staff/GodEater)
22:56:04 Join @ChanServ [0] (ChanServ@services.)
22:56:04 Join CustosL1men [0] (~CustosLim@unaffiliated/cust0slim3n)
22:56:04 Join knittl [0] (~knittl@unaffiliated/knittl)
22:56:04 Join TD-Linux [0] (~Thomas@about/essy/indecisive/TD-Linux)
22:56:04 Join andrewthecoder [0] (~beveradb@irc.beveridge.uk)
22:56:04 Join zu [0] (~zu@ks387228.kimsufi.com)
22:56:04 Join ruskie [0] (ruskie@sourcemage/mage/ruskie)
22:56:04 Join jvoisin [0] (~jvoisin@dustri.org)
22:56:04 Join cttttt [0] (sid135570@gateway/web/irccloud.com/x-inevvxllsjjeegeo)
22:56:04 Join Cu5tosLimen [0] (~CustosLim@unaffiliated/cust0slim3n)
22:56:04 Join anonus [0] (~anonymous@citadel.niflheim.info)
22:56:04 Join APLU [0] (~mulx@eva.aplu.fr)
22:56:04 Join ved [0] (ved@ddsbox.tk)
22:56:04 Join maraz [0] (maraz@kapsi.fi)
22:56:04 Join shrizza_ [0] (~shrizza@oki-dc-urasoe01-187.glbb.ne.jp)
22:56:04 Join evilnick [0] (~evilnick@rockbox/staff/evilnick)
22:56:04 Join rogeliodh [0] (~rogeliodh@ec2-52-90-241-120.compute-1.amazonaws.com)
22:56:04 Join vflyson [0] (~vflyson@cupcake.uberspace.net)
22:56:04 Join rudi_s [0] (~simon@kraftwerk.ruderich.eu)
22:56:04 Join yosafbridge [0] (~yosafbrid@68.ip-149-56-14.net)
22:56:04 Join alexbobp [0] (~alex@testificate.xen.prgmr.com)
22:56:04 Join gluytium [0] (U2FsdGVkX1@ma.sdf.org)
22:56:04 Join __builtin [0] (~xray@rockbox/developer/builtin)
22:56:04 Join Elfish [0] (amba@84.201.30.174)
22:56:04 Join uwe_android [0] (~uwe_andro@static.173.76.9.176.clients.your-server.de)
22:56:04 Join uwe_ [0] (~uwe_@dslb-084-056-050-120.084.056.pools.vodafone-ip.de)
22:56:04 Join mikroflops [0] (~yogurt@178.174.137.46)
22:56:04 Join ranmachan [0] (~ranma@yumi.uguu.de)
22:56:04 Join Rondom [0] (~rondom@modo.nonmodosedetiam.net)
22:56:04 Join n17ikh [0] (~n17ikh@unaffiliated/n17ikh)
22:56:04 Join fIorz [0] (nobody@rain.florz.de)
22:56:04 Join Kohlrabi [0] (~kohlrabi@kohlio.de)
22:56:04 Join dongs [0] (~dongs@bcas.tv)
22:56:04 Join CommunistWitchDr [0] (quasselcor@97-87-177-85.dhcp.stls.mo.charter.com)
22:56:04 Join marex-cloud [0] (sid137234@gateway/web/irccloud.com/x-wkxgsoyjpzyvcvzf)
22:56:04 Join ps-auxw [0] (~arneb@p578F7FEC.dip0.t-ipconnect.de)
22:56:04 Join Marex [0] (~Marex@195.140.253.167)
22:56:04 Join benedikt93 [0] (~quassel@unaffiliated/benedikt93)
22:56:04 Join atsampso1 [0] (~ats@cartman.offog.org)
22:56:04 Join K1773R [0] (~K1773R@unaffiliated/k1773r)
22:56:04 Join WakiMiko [0] (~WakiMiko@unaffiliated/wakimiko)
22:56:04 Join snw [0] (~snow_bcks@ganon.dot-server.net)
22:56:04 Join Slasheri [0] (~miipekk@rockbox/developer/Slasheri)
22:56:35__builtinwhat does "Durchführen eines Zuges" mean when in the context of code?
22:58:00[Saint]something a train?
22:58:10[Saint]oops, thought this was *-community
22:59:07dysunlikely. more like "performing a move" in case the code is from a game :-)
23:00
23:00:40 Join webguest46 [0] (~5d41e969@www.haxx.se)
23:02:19 Quit ZincAlloy (Quit: Leaving.)
23:03:25[Saint]piece-meal translate of the subset of words I know in German foils me again.
23:03:40 Join MrZeus1 [0] (~MrZeus@2a02:c7f:7066:fb00:f911:2965:566b:456a)
23:05:38 Quit webguest46 (Quit: CGI:IRC (Ping timeout))
23:19:28 Join arandomone [0] (~Creator@176-12-25-113.pon.spectrumnet.bg)
23:21:21 Quit wodz (Quit: Leaving)
23:27:52 Join Jinx [0] (Dojo@unaffiliated/jinx)
23:35:25 Join jhMikeS [0] (~jethead71@24.192.177.173)
23:35:50 Quit paulk-collins (Quit: Leaving)
23:41:32 Join roxfan [0] (~roxfan@43.234-245-81.adsl-dyn.isp.belgacom.be)
23:51:26 Join krabador [0] (~krabador@unaffiliated/krabador)
23:52:27 Join preglow [0] (~thomj@2001:840:4243:3::101)
23:58:36 Quit pamaury (Ping timeout: 240 seconds)

Previous day | Next day