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 2022-12-11

00:42:40 Quit massiveH (Quit: Leaving)
00:54:53_bilgus_dconrad (logs) see thats why we can't have nice things, I really like that patch so lets hope its coincidence
00:57:19_bilgus_its probably more likely Haas Surround has a bug that this is making come out of the wood work
01:40:14 Quit Romster (Quit: Leaving)
01:54:15_bilgus_well its reproducible in the simulator ASAN says AddressSanitizer: SEGV on unknown address 0x000000000000
01:54:43***Saving seen data "./dancer.seen"
02:13:46_bilgus_changing the code back to the original before 4691 the error still occurs
03:06:50 Join Romster [0] (~romster@
03:11:32 Quit Romster (Ping timeout: 248 seconds)
03:45:20bluebrothersomeone familiar with ipod6g bootloader install? I've been looking at the manual and the manual install for that are ... well, outdated.
03:45:57bluebrotherit still links to freemyipod, but we have our official bootloader now. Also, it instructs the user to use Rockbox Utility with the "show disabled targets" option, which is simply wrong these days −− ipod6g isn't a disabled target anymore.
03:46:30bluebrotherI also expect bootloader install to work on Windows if you have mks5lboot for Windows. Though I don't have an ipod6g, so I cannot check.
03:46:43bluebrotherand last, it instructs the user to install the build first. Is that really necessary?
03:47:38bluebrother(plus, using Rockbox Utility, but if you want to install manually you likely want to install from the zip anyway, and it's in the "bootloader install" subsection, so that is also messed up)
03:48:57bluebrotherI'd update the docs, but should have confirmation on how it actually works :)
03:51:24 Join lebellium [0] (
03:54:44***Saving seen data "./dancer.seen"
04:32:03 Join Romster [0] (~romster@
05:32:52 Quit Richard42 (Ping timeout: 268 seconds)
05:54:28rb-bluebotBuild Server message: New build round started. Revision dfd9c10589, 303 builds, 7 clients.
05:54:47***Saving seen data "./dancer.seen"
06:02:09 Quit jackie (Quit: No Ping reply in 180 seconds.)
06:03:17 Join jackie [0] (
06:28:32rb-bluebotBuild Server message: Build round completed after 2044 seconds.
06:28:35rb-bluebotBuild Server message: Revision dfd9c10589 result: All green
06:30:09rb-bluebotBuild Server message: New build round started. Revision ac9066dd44, 303 builds, 7 clients.
06:35:39 Quit CH23_M (Ping timeout: 265 seconds)
06:38:22 Join CH23_M [0] (~CH23@revspace/participant/ch23)
06:55:44 Quit CH23_M (Read error: Connection reset by peer)
06:56:03 Join CH23_M [0] (~CH23@revspace/participant/ch23)
07:03:11rb-bluebotBuild Server message: Build round completed after 1982 seconds.
07:03:12rb-bluebotBuild Server message: Revision ac9066dd44 result: All green
07:54:50***Saving seen data "./dancer.seen"
08:25:45_bilgus_besides td speed the dsp stuff doesn't check for a successful handle from core_alloc
08:26:05_bilgus_but thats still not fixing the crash in the sim
08:26:29_bilgus_and I've followed it to prior to asan being added
08:36:59_bilgus_nm they check just not in the alloc proc
08:56:48_bilgus_starting in dsp/surround.c followed it to surround_new_format> dsp_proc_activate(dsp/dsp_core.c) adding sleep(100) at the end of the function makes it not crash
08:57:05_bilgus_oddly I don't see a rhyme or reason
08:57:37_bilgus_race condition I suppose
09:03:13_bilgus_dsp_process(dsp/dsp_core.c) looks to be our running thread
09:54:54***No seen item changed, no save performed.
10:22:59_bilgus_dconrad (logs) could you try g#4905 and see if the problem persists?
10:23:02rb-bluebotGerrit review #4905 at : haas surround crash diag by William Wilgus
10:35:54_bilgus_from what I know so far its calling a NULL function pointer now why is that process function pointer still NULL I don't know
10:43:43 Join amachronic [0] (~amachroni@user/amachronic)
10:48:29amachronic_bilgus about that INIT_ATTR linker thing: g#4906
10:48:32rb-bluebotGerrit review #4906 at : configure: Detect linker's NO_CROSSREFS_TO() support by Aidan MacDonald
10:48:46_bilgus_that patch fixes it for me and it appears the surround works too but I still don't know what is missing, doing a logf build it appears it calls the NULL function till it gets setup but why it only appears when you se through the QuickScreen Ive no idea, its almost as if it skips something critical where as running it through the settings menu it works fine, WEIRD
10:48:52amachronicsample linker script in g#4909
10:48:54rb-bluebotGerrit review #4909 at : x1000: Enable support for INIT_ATTR by Aidan MacDonald
10:49:45_bilgus_amachronic, nice ill have to go through and see where else INIT_ATTR is waiting to crash us
10:50:42_bilgus_re your playlist loader patch I think its a coincidence I'm waiting to hear back from dconrad if this patch makes his player stop crashing
10:52:17_bilgus_ASAN really doesn't do much for tracking it down probably have to be pulled apart a bit more, either that of we just leave the empty fn there its not adding processing in the hot path that way
10:53:50amachronichuh, well i'll try an asan build now anyway
10:54:11_bilgus_it smells like a race condition being that adding a sleep makes it work too
10:56:19_bilgus_ASAN cathches it its just got no context I knew the pc of 0 pointed to a null deref but I just had to go through and check every pointer till I found it
10:56:47_bilgus_the first ver of that patch has the check if you want to see where it was
11:03:14amachronic_bilgus_ okay I see what you mean with ASAN, when it bails like that you can use gdb to get a backtrace
11:03:21amachronicASAN_OPTIONS="abort_on_error=1" gdb ./rockboxui
11:03:44_bilgus_oh nice I was wondering if I could attach a debugger to it
11:04:19_bilgus_where do I define that? in the call to configure or in the script?
11:05:21amachronicyou just set the env var before you run the sim
11:05:22_bilgus_or on the commend line?
11:05:28_bilgus_ah ok
11:06:28amachronicwell i only found out what you already did −− the process fn pointer is NULL
11:07:51_bilgus_I have no idea why dconrad found 4691 to be the start I still found it all the way back to ASAN being added
11:08:12_bilgus_kinda stopped me in my tracks there
11:09:49amachronici'd like to say it's happening because dsp_proc_enable_enlink() is linking in the dsp before the fn pointer gets set by configure(DSP_PROC_INIT)
11:10:35_bilgus_but whats different between doing it through the menu versus the QS they appear to do the same
11:10:44amachronicthat I don't know :)
11:10:51_bilgus_must be something I'm missing
11:16:02 Join dconrad [0] (~dconrad@
11:21:54dconrad_builgus_ I'm getting that patch built now
11:22:31dconradthanks for looking at it
11:26:35amachronicwell I added sleep(HZ) right before ->configure(DSP_PROC_INIT) and got a crash from the menu eventually
11:26:42amachronicdoesn't happen every time though
11:29:11dconradthat patch does appear to make it work
11:29:24dconradlet me try it a few more times and see if its consistent
11:34:46amachronicif I comment out the pcmbuf_set_low_latency() calls in lowlatency_callback() then the menu triggers it reliably without the sleep
11:35:21amachronicthat must be the difference between the menu and quickscreen.
11:36:10amachronici guess low latency mode makes it harder to trigger the race?
11:37:07amachronic_bilgus_ i think your patch is the best solution
11:38:51dconradso I've tried it 4 times and didn't get the crash once
11:39:19dconradI have noticed that dsp options seem to take a few seconds to "kick in" from the QS, is this related?
11:40:01dconrador, could it be related
11:43:01amachronici think that's because the QS doesn't enable low latency mode
11:43:59dconradthat would make sense, does just going into the settings enable low latency mode normally? or, certain settings I guess
11:44:27amachronicyeah only certain sound settings do it
11:44:54dconradlike, enter the haas surround setting menu and it triggers it then
11:45:05dconradI see
11:45:37dconradwell that answers a long-standing question I've had haha
11:50:34 Quit amachronic (Quit: amachronic)
11:51:32_bilgus_thats strange I saw the low_latency call and tried adding that but still got the crash
11:51:53_bilgus_when I was first trying to pin it down
11:53:25_bilgus_maybe it still happens even with low_latency enabled just less frquent
11:54:57***Saving seen data "./dancer.seen"
12:34:34 Join dys [0] (~dys@user/dys)
13:25:51 Quit dconrad (Remote host closed the connection)
13:35:08 Join mink [0] (~mink@
13:42:09 Join amiconn_ [0] (
13:42:09 Nick amiconn is now known as Guest4891 (
13:42:09 Quit Guest4891 (Killed ( (Nickname regained by services)))
13:42:09 Nick amiconn_ is now known as amiconn (
13:42:46 Join pixelma_ [0] (
13:42:46 Nick pixelma is now known as Guest4402 (
13:42:46 Nick pixelma_ is now known as pixelma (
13:43:16 Quit Guest4402 (Ping timeout: 256 seconds)
13:55:01***Saving seen data "./dancer.seen"
14:46:37 Quit fourHZ (Quit: Client closed)
14:48:01 Join fourHZ [0] (
15:02:04 Join dconrad [0] (~dconrad@
15:06:59 Quit dconrad (Ping timeout: 268 seconds)
15:08:22 Join dconrad [0] (~dconrad@
15:11:11dconradsmall caveat with patchset 4905 _bilgus_, it doesn't seem to save the setting if I change it from the QS and then power off the player
15:11:36dconradthough that's probably a completely separate issue
15:14:34dconradcoincidentally, after months and months of not seeing it, I got the "screen wakeup freeze" like 4 or 5 times today in quick succession... I may try some more to find a repeatable way to trigger it
15:15:47 Quit fourHZ (Quit:
15:35:09 Quit mink (Remote host closed the connection)
15:36:35_bilgus_dconrad, I'll look into it a bit more before I push it I still need to check a few things (in my patch comment) whats this 'screen wakeup freeze'?
15:37:59_bilgus_I think as things change and or get faster we will probably see more of these things pop up I hate heisenbugs
15:38:48dconradthe screen wakeup freeze is something unrelated, but has been with the eros q native since the port started
15:39:49dconradI thought it was taken care of when amachronic found the lcd instructions didn't have the end marker, and I haven't seen it in a long time, only for it to pop up again today
15:40:19dconradbasically the player is doing its thing and you wake up the screen and the whole device freezes, you have to do a force power off
15:43:28dconradnot sure if coincidence or not, but most of my library is flac, but I'm listening to some ancient mp3s today which I don't think I've listened to in a while
15:55:03***Saving seen data "./dancer.seen"
15:59:55_bilgus_more heisenbugs it is then :p\
16:29:40dconradhmm, I guess I can't save the surround enable setting at all, even through the menu
16:36:12dconradeven with a blank settings file
16:53:32 Join bilgus_ph [0] (~bilgus_ph@
16:54:04bilgus_phDconrad probably something with the recent settings rewrites I'll see what I can find tonight or tomorrow
16:58:55 Quit bilgus_ph (Ping timeout: 260 seconds)
17:21:26 Quit dconrad (Remote host closed the connection)
17:26:08 Quit lebellium (Quit: Leaving)
17:28:08 Quit advcomp2019_ (Quit: Leaving)
17:33:52 Join advcomp2019 [0] (~advcomp20@user/advcomp2019)
17:47:33 Join dconrad [0] (~dconrad@
17:50:32 Quit dconrad (Remote host closed the connection)
17:51:02 Join dconrad [0] (~dconrad@
17:51:08 Quit dconrad (Client Quit)
17:55:05***Saving seen data "./dancer.seen"
18:56:24 Quit CH23_M (Read error: Connection reset by peer)
18:56:56 Join CH23_M [0] (~CH23@revspace/participant/ch23)
19:55:08***Saving seen data "./dancer.seen"
21:55:13***No seen item changed, no save performed.
22:17:14rb-bluebotBuild Server message: New build round started. Revision 863538c50a, 303 builds, 6 clients.
22:54:16rb-bluebotBuild Server message: Build round completed after 2222 seconds.
22:54:19rb-bluebotBuild Server message: Revision 863538c50a result: All green
23:37:10 Quit m01 (Quit: Konversation terminated.)
23:39:22 Join m01 [0] (
23:42:57 Join WilliamC [0] (
23:43:25WilliamCI have some Vorbis files, with images, that are causing rockbox to crash. Is this a known issue?
23:43:58WilliamCI removed the images to see if this is the case, otherwise I don't know what the commonality is.
23:45:29 Quit Acou_Bass (Quit: ZNC 1.8.2 -
23:48:12 Join Acou_Bass [0] (
23:55:15***Saving seen data "./dancer.seen"

Previous day | Next day