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 2020-08-15

00:12:02 Join lemon_jesus6 [0] (
00:13:47 Quit lemon_jesus (Ping timeout: 246 seconds)
00:13:48 Nick lemon_jesus6 is now known as lemon_jesus (
00:31:12***Saving seen data "./dancer.seen"
00:52:05 Quit __Bilgus_ (Remote host closed the connection)
00:54:00 Join caviar_hands [0] (
00:57:25 Nick caviar_hands is now known as stumper (
00:58:00 Part stumper
01:11:59 Quit Soap (Read error: Connection reset by peer)
01:29:42 Join Rower [0] (
01:40:00 Quit massiveH (Quit: Leaving)
02:31:15***Saving seen data "./dancer.seen"
02:38:55 Quit TheEaterOfSouls (Remote host closed the connection)
02:41:26 Join blbro[m]1 [0] (blbromatri@gateway/shell/
03:57:04 Join lebellium [0] (
04:31:17***Saving seen data "./dancer.seen"
04:41:59 Join lemon_jesus0 [0] (
04:43:45 Quit lemon_jesus (Ping timeout: 240 seconds)
04:43:45 Nick lemon_jesus0 is now known as lemon_jesus (
05:39:12 Join __Bilgus_ [0] (41ba23be@
05:39:28 Join johnb5 [0] (
05:41:57johnb5__Bilgus_ : with that pastebin, the batterybench works fine (tested on two devices).
05:43:13__Bilgus_that stinks I guess Ill try to get to the bottom of this before I just revert to a static stack
05:44:00__Bilgus_I just wonder why the announce plugin doesn't exhibit the same
05:46:06johnb5Just to make sure: the ask for the second trial was
05:46:08johnb5static struct event_queue thread_q SHAREDBSS_ATTR;
05:46:14johnb5static struct event_queue thread_q;
05:47:14johnb5yep, that's what I had done.
05:50:40__Bilgus_Ah I found where builtin did something similar for a plugin lets try copying in his implementation
05:51:03__Bilgus_probably just a weird arm alignment thing I hope
06:16:29__Bilgus_In __builtin's code he initializes the buffer with tlsf first so maybe it handles the alignment better
06:18:06 Nick johnb5 is now known as johnb2 (
06:31:18***Saving seen data "./dancer.seen"
08:04:32 Quit johnb2 (Ping timeout: 256 seconds)
08:06:32 Join ac_laptop [0] (~ac_laptop@
08:12:52 Join MrZeus [0] (
08:26:48 Quit jdarnley (Read error: Connection reset by peer)
08:31:22***Saving seen data "./dancer.seen"
08:37:30 Join J_Darnley [0] (
08:39:19__Bilgus_johnb when you get a chance couild you try g#2661 ?
08:39:21fs-bluebotGerrit review #2661 at : Battery_bench stack alignment fix by William Wilgus
08:56:55__Bilgus_oh you know what I bet I'm wrong on the stack growing down in that original code
09:13:39 Join johnb2 [0] (
09:15:01johnb2__Bilgus_ : Do you still want me to test your current g#2661 or should I wait?
09:15:03fs-bluebotGerrit review #2661 at : Battery_bench stack alignment fix by William Wilgus
09:17:34 Quit MrZeus (Ping timeout: 260 seconds)
09:17:57__Bilgus_i'm compiling now to check if I have the stack backwards
09:18:31__Bilgus_I just happen to be on my laptop so it takes forever
09:23:51__builtin__Bilgus_: why are you bothering to align the result of malloc?
09:24:18__builtinit should already be aligned, otherwise something is very broken
09:25:13__Bilgus_get_plugin_buffer isn't aligned for you
09:25:36__Bilgus_any malloc code is from your puzzles plugin
09:26:00__Bilgus_I think my issue is that I thought the stack grows up but it apparently doesnt
09:26:32__builtinyeah, stacks on arm are descending in numerical address
09:27:01__builtinI still don't think you need to manually align though
09:27:56__builtinhave you tested to verify that malloc is actually returning an unaligned result?
09:28:29__Bilgus_johnb try this one
09:28:49__Bilgus_funny thing is that it works either way on my player
09:28:59__builtinalso - why do you even need tlsf for a single allocation?
09:29:25__Bilgus___builtin malloc has nothing to do with this
09:29:51__Bilgus_I moved battery bench to using the plugin buffer for stack
09:30:23__Bilgus_worked fine here and I assume for mendel but johnb and strife both had issues
09:30:56__Bilgus_the other plugin where I used a 'dynamic' stack like that worked fine as well
09:31:07__builtinwhat exactly is going wrong?
09:31:10__Bilgus_including johnb
09:31:31__Bilgus_unchecked stack overflow
09:32:13__builtinwhat do you mean by "unchecked?"
09:32:15__Bilgus_I assume its hangs and crashes so I figured we'd try your implementation and see if it fixed it for john
09:32:49__builtinbut you aren't sure it's actually a stkov?
09:32:53__Bilgus_as in it doesn't catch the stack ovfl it runs some invalid memory address
09:33:12johnb2yes, this one works
09:33:29__Bilgus_it was that I had the stack growing the wrong way
09:33:40__builtindo you get a crash dump?
09:33:41__Bilgus_which I realized when I saw your code
09:34:08__builtinoh wait - what?
09:34:29__Bilgus_I was looking at your implementation
09:34:38__Bilgus_and realized I flipped the stack
09:34:44__builtinI'm almost certain you don't pass the high limit of the stack to create_thread
09:34:55__Bilgus_because your code did not flip the stack
09:34:59__builtinthat's almost definitely wrong
09:36:28__builtinyou pass create_thread the starting, lower address and it takes care of the platform quirks for you
09:36:44__Bilgus_what I did is most definitely wrong per johnb 's experience
09:37:07__Bilgus_but its been working for 2+ years elsewhere for me
09:37:14__builtinin no case should you have to account for the stack growth direction
09:37:26__Bilgus_so good lesson to remember
09:37:46__builtinjeez, you've done this elsewhere too?
09:38:24__Bilgus_not in anything in the codebase except the announce plugin
09:38:39__builtinyeah I'd fix that haha
09:39:59__builtinwhat I'd do if a thread needs a bigger stack is just declare a global long[] array
09:40:25__Bilgus_it ends up in the same region either way
09:40:27__builtinthat'll take care of alignment and avoids tlsf
09:40:45__builtinyeah, but with less opportunities to shoot yourself in the foot
09:40:54__Bilgus_tlsf isn't going to be used
09:41:11__Bilgus_look at that last pb
09:41:33__Bilgus_I just take the rest of the buffer
09:42:41fs-bluebotGerrit review #2661 at : Battery_bench stack alignment fix by William Wilgus
09:43:20__builtinoh pastebin
09:43:22__Bilgus_ @654
09:46:38__builtinwhere is ALIGN_BUFFER coming from?
10:02:52fs-bluebotBuild Server message: New build round started. Revision 84bfd68, 280 builds, 9 clients.
10:03:40__Bilgus_johnb thanks for your help as always should be fixed in head now
10:06:41johnb2__Bilgus_ : for something completely different: I was thinking of enabling multiboot for the e200 and had started mimicking what you had done for the AMS based devices, until I hit the CRT / assembler related changes, which is beyond my knowledge.
10:07:02__Bilgus_pamaury helped me with that originally
10:07:08johnb2bootdata that is.
10:09:57__Bilgus_johnb2 where is the ctr file for uhe e200?
10:10:48__Bilgus_crt lol
10:12:36__Bilgus_if its an arm device you shouldn't need to do anything
10:18:44johnb2I couldn't even determine the location pp / arm.
10:18:52fs-bluebotBuild Server message: Build round completed after 960 seconds.
10:18:54fs-bluebotBuild Server message: Revision 84bfd68 result: All green
10:20:57johnb2I if recall correctly, while compiling I got an error that the structure "bootdata" was missing, so I assumed it was related to your changes into crt.
10:22:25__Bilgus_very well possible
10:22:41__Bilgus_if you look in the map file you should be able to tell which
10:23:08__Bilgus_it could also be that you don't have it defined properly
10:23:25__Bilgus_there were like 2 or 3 different commits for that
10:25:58__Bilgus_I think maybe there wasn't enough room but it might have been one of the other players
10:29:16johnb2You might have misunderstood: I was referring to e200 v1 which is portalplayer based. I had tried the v2 before (and we chatted about it), here the BL didn't fit into the firmware image (amspatcher complained).
10:29:58johnb2For v2 I didn't succeed in reducing the BL size enough.
10:31:25***Saving seen data "./dancer.seen"
10:36:37__Bilgus_does this naive try work?
10:37:11__Bilgus_I'm really not familar with the quirks for the crt files that why pamaury did that part
10:40:40 Join MrZeus [0] (
10:49:53 Join blbro[m] [0] (blbrostrat@gateway/shell/
11:50:20 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
12:14:10 Quit johnb2 (Quit: Nettalk6 -
12:31:28***Saving seen data "./dancer.seen"
12:34:30 Join Bilgus [0] (
12:35:10Bilgusjohnb you added the definition 2x in rolo.c
12:36:24 Quit Bilgus (Client Quit)
12:44:59 Join TheEaterOfSouls [0] (~souls@unaffiliated/theeaterofsouls)
13:23:43 Quit livvy (Ping timeout: 240 seconds)
13:24:25 Join livvy [0] (~livvy@gateway/tor-sasl/livvy)
14:31:32***Saving seen data "./dancer.seen"
15:15:36 Quit Rower (Ping timeout: 256 seconds)
15:23:45 Quit kugel (Quit: Lost terminal)
15:26:36 Join kugel [0] (
15:26:37 Quit kugel (Changing host)
15:26:37 Join kugel [0] (~kugel@rockbox/developer/kugel)
15:33:40 Quit kugel (Quit: Lost terminal)
15:40:42 Join johnb4 [0] (
15:40:53 Join kugel_ [0] (
15:40:53 Quit kugel_ (Changing host)
15:40:53 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
15:44:44johnb4__Bilgus_ : yes, that was a desperate trial, but it is still missing (see the error message). Ain't I ending up in the MI4_FORMAT clause?
15:59:37johnb4Error : rolo.c:(.text+0xf4): undefined reference to `write_bootdata'
15:59:51 Quit kugel_ (Quit: Lost terminal)
16:00:14johnb4 g#2662
16:00:16fs-bluebotGerrit review #2662 at : Enable boot from SD for Sansa e200 (WIP) by Johannes Rauh
16:02:18 Quit johnb4 (Quit: Nettalk6 -
16:31:35***Saving seen data "./dancer.seen"
16:41:22 Quit ac_laptop (Ping timeout: 260 seconds)
16:45:42 Join kugel [0] (
16:45:42 Quit kugel (Changing host)
16:45:42 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:49:14 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
16:52:30 Quit kugel (Quit: Lost terminal)
16:53:41 Join kugel [0] (~kugel@rockbox/developer/kugel)
16:53:53 Quit kugel (Client Quit)
16:55:33 Join kugel_ [0] (
16:55:33 Quit kugel_ (Changing host)
16:55:33 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
17:08:19 Quit kugel_ (Quit: Lost terminal)
17:08:53 Join kugel_ [0] (
17:08:54 Quit kugel_ (Changing host)
17:08:54 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
17:09:21 Quit kugel_ (Client Quit)
17:22:37 Join kugel_ [0] (
17:22:38 Quit kugel_ (Changing host)
17:22:38 Join kugel_ [0] (~kugel@rockbox/developer/kugel)
17:23:14 Quit kugel_ (Client Quit)
17:25:11 Join kugel [0] (
17:25:11 Quit kugel (Changing host)
17:25:12 Join kugel [0] (~kugel@rockbox/developer/kugel)
18:03:30 Quit pamaury (Ping timeout: 256 seconds)
18:24:43 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
18:31:39***Saving seen data "./dancer.seen"
18:37:16 Join ac_laptop [0] (~ac_laptop@
18:56:22 Quit lebellium (Quit: Leaving)
19:37:23 Quit sakax (Ping timeout: 240 seconds)
20:22:02__Bilgus_johnb g#2662 compiles now it needs tested though and there might be other gotchas I've missed
20:22:04fs-bluebotGerrit review #2662 at : Enable boot from SD for Sansa e200 (WIP) by Johannes Rauh
20:31:41***Saving seen data "./dancer.seen"
21:00:36 Join massiveH [0] (
21:29:46 Quit akaWolf (Ping timeout: 256 seconds)
21:46:59 Quit MrZeus (Ping timeout: 240 seconds)
22:31:42***Saving seen data "./dancer.seen"
23:34:16 Quit [7] (Disconnected by services)
23:34:22 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)

Previous day | Next day