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 2024-03-28

00:03:44 Quit q3k (Ping timeout: 268 seconds)
00:05:23 Join q3k [0] (
00:07:04_bilgusamachronic g#5610 should fix that issue as well NO TOO.
00:07:07rb-bluebotGerrit review #5610 at : [Bugfix] ft_assemble_path extra slashes by William Wilgus
00:27:01 Join jacobk [0] (~quassel@
00:28:24 Quit othello7 (Ping timeout: 268 seconds)
00:59:09 Join PheralSparky [0] (~Shawn@user/shawn/x-4432647)
01:38:58***Saving seen data "./dancer.seen"
01:53:01 Quit l0x0l (Quit: WeeChat 4.2.1)
02:07:05 Quit yang (Remote host closed the connection)
02:07:13 Join yang [0] (~yang@fsf/member/yang)
02:07:13 Quit rb-bluebot (Quit: connection lost?)
02:07:21 Join rb-bluebot [0] (~rb-bluebo@rockbox/bot/utility)
02:07:33 Quit bleb (Ping timeout: 256 seconds)
02:07:33 Nick bleb_ is now known as bleb (~cm@user/bleb)
02:08:26 Join bleb_ [0] (~cm@user/bleb)
02:20:39 Quit Bubblegumdrop (Ping timeout: 260 seconds)
02:29:58 Join Bubblegumdrop [0] (~znc@user/Bubblegumdrop)
02:31:55 Quit JanC (Ping timeout: 256 seconds)
02:32:32 Join JanC [0] (~janc@user/janc)
02:57:25 Quit JanC (Ping timeout: 256 seconds)
02:58:15 Join JanC [0] (~janc@user/janc)
03:02:27 Join JanC_ [0] (~janc@user/janc)
03:02:39 Nick JanC is now known as Guest8427 (~janc@user/janc)
03:02:39 Nick JanC_ is now known as JanC (~janc@user/janc)
03:03:55 Quit Guest8427 (Ping timeout: 255 seconds)
03:05:39 Quit jacobk (Ping timeout: 268 seconds)
03:39:02***Saving seen data "./dancer.seen"
04:50:58 Quit PheralSparky (Read error: Connection reset by peer)
05:27:01 Quit JanC (Ping timeout: 255 seconds)
05:39:06***Saving seen data "./dancer.seen"
06:55:13 Join JanC [0] (~janc@user/janc)
07:39:10***No seen item changed, no save performed.
08:25:05rb-bluebotBuild Server message: New build round started. Revision 19f21a2b3f, 304 builds, 9 clients.
08:25:05rb-bluebotCodecs: mp4: Enable FAAD_STATIC_ALLOC by Roman Artiukhin
08:34:46rb-bluebotBuild Server message: Build round completed after 582 seconds.
08:34:47rb-bluebotBuild Server message: Revision 19f21a2b3f result: All green
08:39:38 Join jacobk [0] (~quassel@
09:39:13***No seen item changed, no save performed.
09:41:46 Join mmlant [0] (
09:48:58 Quit mmlant (Remote host closed the connection)
09:49:16 Quit q3k (Changing host)
09:49:16 Join q3k [0] (q3k@hswaw/infra/q3k)
09:49:42 Join mmlant [0] (
09:51:22 Quit mmlant (Remote host closed the connection)
09:55:23 Join mmlant [0] (
09:55:55 Join othello7 [0] (
09:59:08 Quit mmlant (Remote host closed the connection)
10:01:03 Join mmlant [0] (
10:02:18 Quit mmlant (Remote host closed the connection)
10:30:40 Quit JanC (Ping timeout: 256 seconds)
10:32:31 Join JanC [0] (~janc@user/janc)
10:43:45 Quit JanC (Read error: Connection reset by peer)
10:43:52 Join JanC_ [0] (~janc@user/janc)
10:43:52 Nick JanC_ is now known as JanC (~janc@user/janc)
11:09:06 Join JanC_ [0] (~janc@user/janc)
11:09:18 Nick JanC is now known as Guest3663 (~janc@user/janc)
11:09:18 Quit Guest3663 (Killed ( (Nickname regained by services)))
11:09:18 Nick JanC_ is now known as JanC (~janc@user/janc)
11:14:16 Quit CH23 (Ping timeout: 260 seconds)
11:14:48 Join davisr [0] (~davisr@fsf/emeritus/davisr)
11:28:51 Quit othello7 (Ping timeout: 268 seconds)
11:39:16***Saving seen data "./dancer.seen"
11:44:53 Quit CH23_M (Ping timeout: 268 seconds)
11:45:24 Join othello7 [0] (
11:46:30 Join CH23_M [0] (~CH23@revspace/participant/ch23)
12:17:21 Quit CH23_M (Read error: Connection reset by peer)
12:17:47 Join CH23_M [0] (~CH23@revspace/participant/ch23)
12:22:01 Join CH23 [0] (~CH23@revspace/participant/ch23)
12:24:31 Quit jacobk (Ping timeout: 260 seconds)
13:39:18***Saving seen data "./dancer.seen"
14:06:39 Quit JanC (Ping timeout: 272 seconds)
14:10:20 Join JanC [0] (~janc@user/janc)
14:13:56 Join decky [0] (~decky_@
14:16:51 Quit decky_e (Ping timeout: 256 seconds)
14:19:37 Join jacobk [0] (~quassel@
14:19:59 Quit JanC (Read error: Connection reset by peer)
14:20:28 Join JanC [0] (~janc@user/janc)
14:41:37 Join lebellium [0] (
15:39:20***Saving seen data "./dancer.seen"
16:00:11 Quit jacobk (Ping timeout: 268 seconds)
16:05:05 Quit JanC (Ping timeout: 256 seconds)
16:05:50 Join JanC [0] (~janc@user/janc)
16:27:15 Quit JanC (Ping timeout: 272 seconds)
16:37:33 Join JanC [0] (~janc@user/janc)
17:07:23 Join amachronic [0] (~amachroni@user/amachronic)
17:15:33amachronic_bilgus let me see if we can just get rid of buflib in the bootloaders, IIRC there was an ifdef to use static buffers.
17:19:12amachronichmm we're already doing that except for pp502x
17:19:48_bilgusits the as3525 e200v2
17:20:30_bilgusits calling core alloc in usb_storage give me a moment to look at the exact fn
17:22:35_bilgus usb_handle = core_alloc_ex(ALLOCATE_BUFFER_SIZE + MAX_CBW_SIZE + 31,
17:22:35_bilgus &buflib_ops_locked);
17:22:46_bilgusI see nothing around it
17:22:47amachronicyeah that is covered by the big #ifdef above, if you check objdump it isn't getting called.
17:23:09amachronicthe #else
17:23:28_bilgusok let me double check that its not just the code page stuff calling this
17:23:42amachronici think it might be, that was what grepping turned up
17:23:43_bilgusthats the other emitted core_alloc
17:23:58 Join JanC_ [0] (~janc@user/janc)
17:24:10 Nick JanC is now known as Guest2005 (~janc@user/janc)
17:24:10 Quit Guest2005 (Killed ( (Nickname regained by services)))
17:24:10 Nick JanC_ is now known as JanC (~janc@user/janc)
17:24:12amachronicyep it's there in objdump
17:26:45amachronicremoving that shrinks it ~2.5k
17:27:58_bilgussure enough, even better thank you
17:28:35_bilgusI'm all for bahus kicking out bmps and the logo but that 2.5 k bought backtraces
17:31:13_bilgusnow that still leaves that commit they identified I still think its a coincidental kinda thing due to the alignment changing maybe unless you can think of something in the name removal
17:32:04 Quit davisr (Remote host closed the connection)
17:32:15_bilgusyeah the bootloader is still 10k over or so including that
17:34:31amachronici'm at 98532, not sure what our limit is
17:34:39amachronic98532 bytes
17:35:18 Join IPG [0] (
17:35:38 Join jacobk [0] (
17:36:29_bilguswell it all depends on what UCL buys you but I didn't get it to fit till my binary was 73 k IIRC
17:38:32amachronicok yeah i was confused by the different sizes
17:39:23***Saving seen data "./dancer.seen"
17:44:11 Quit f_ (Remote host closed the connection)
17:46:22_bilgus77.8 is still too big that no logo or alpha bitmap, and your patch
17:46:55_bilguswonder what enabling thumb does
17:46:57amachroniccommenting out show_logo in the bootloader gets the delta down to 4k
17:49:07_bilgusstill too big (123187 bytes) doesn't fit in the firmware (120860 bytes)
17:49:34_bilgusi think flip and invert only buys like 150 bytes
17:50:08amachronicknocking out alpha blending saves ~700 bytes
17:50:27_bilgusif I knew we could get rid of fat16 we'd probably have it
17:52:05amachronicthere are still bits of buflib being pulled in and if I can find out why that can save quite a bit, at least 760 uncompressed bytes
17:52:13_bilgusmarking it as having 2 mb knocks off 100 bytes so thats something
17:53:24_bilgustheres not much left thats not needed I can likely dump more of the name space stuff
17:55:30amachronicah it's usb_storage_disconnect() calling core_free()
17:57:50amachronic3.4k to go after that
18:01:26 Join Saijin_Naib [0] (
18:02:00 Join PheralSparky [0] (~Shawn@user/shawn/x-4432647)
18:02:34 Join f_ [0] (~AUGESOUND@fases/developer/funderscore)
18:05:19_bilgusI have a patch that turns memcpy into memmov calls I think it gains like 2-300 bytes
18:05:50_bilgus g#4739
18:05:53rb-bluebotGerrit review #4739 at : Bootloader merge memcpy and memmove by William Wilgus
18:07:10amachronicdefinitely worth it, memcpy is 680 bytes
18:09:44amachronichmm apparently that saved nothing since memmove isn't used
18:10:33_bilgusI was just coming to say the same :/
18:11:03amachronicwe can try to nuke scrolling, that seems to be in there
18:13:10_bilgusits just a dummy fn
18:13:24_bilgusunless itsgotten back in there again somehow
18:14:04_bilgusfat16 eats 540 bytes but I still don't have confirmation that there haven't been v2's seen with it
18:15:51amachronicit's lcd_scroll_stop, cutting it saves 268 bytes
18:17:45_bilguswell thats half of whats left then
18:18:09_bilgusneed to make sure its not doing something we need after init though
18:20:04_bilgusyeah its not
18:25:30 Quit JanC (Ping timeout: 255 seconds)
18:29:22_bilgusdown to about 1k left
18:29:40_bilgusi guess flip and invert will count
18:32:15 Quit f_ (Remote host closed the connection)
18:32:54 Quit lebellium (Quit: Leaving)
18:36:11_bilgusoh its only like 20 bytes : (
18:37:39amachronicooh turning off priority scheduling gives 2k
18:38:02_bilgusyeah that was the other thing I wanted to ask you about
18:39:59_bilgusIIRC it mentioned it was wanted in the bootloader
18:40:12amachronicso just that with bahus's patches (including SD boot) gives a binary that fits
18:40:17amachronicno thumb needed either
18:40:58_bilgusawesome and will it work without priority scheduling then?
18:41:48amachronici'm *fairly* sure that prio scheduling only affects how much a thread gets to run / which gets woken first when there are several running threads.
18:42:19 Join JanC [0] (~janc@user/janc)
18:42:26amachronicthere is a realtime level used in the USB stack and that may be why it was enabled.
18:42:35amachroniciirc the realtime one is for handling control requests and whatnot
18:43:03amachronicunless the bootloader has lots of threads running I don't think it should matter though...
18:46:02_bilgusyou want to build your changes in that patch set and we can have him try it and see if that particular commit still crashes it
18:46:35_bilgusrather place your changes in there or another and have him, build it
18:47:00_bilgusmy bet is if it crashes it will be on a different one
18:47:47 Join Moriar [0] (
18:51:41 Quit JanC (Ping timeout: 256 seconds)
18:52:26 Join JanC [0] (~janc@user/janc)
18:58:39_bilgusI mentioned it in the forum post
19:01:14 Quit sam_d (Read error: Connection reset by peer)
19:05:01amachronicactually it still doesn't fit with backtraces enabled
19:05:07 Join sam_d [0] (~sam@user/sam-d/x-8933526)
19:05:17 Quit Saijin_Naib (Ping timeout: 240 seconds)
19:08:29_bilgusoh its fitting with changes I made
19:09:38_bilgusah thumb mode
19:10:00_bilgusthat buys you 2k I think
19:10:16_bilguswe have thumb enabled for the clipv1
19:10:57_bilgusdo you think its important not to enable it I think we can probably make 1-2 k with some creative trimming
19:13:15_bilgusyeah back to 1.5k
19:13:41amachronicif thumb works I don't see why we can't enable it
19:14:21amachroniccurrently I am wondering if the bootloader needs to be able to write to the filesystem, if not we could trim out _lots_ of code.
19:16:26_bilgusI think I have a patch about removing parts of file open in the bootloader
19:17:19_bilgus g#1562
19:17:21rb-bluebotGerrit review #1562 at : Add DISABLE_O_CREAT define to allow removal of file creation codepath by William Wilgus
19:18:44_bilgusbut that seems like it could be a bad idea
19:21:06amachronicyeah it is definitely more involved
19:23:27 Join massiveH [0] (~massiveH@2600:4040:a982:c800:ed41:eab7:e78f:36e1)
19:25:17_bilgusHAVE_BUTTON_LIGHT isn't needed thats 120 bytes
19:31:55 Quit JanC (Ping timeout: 256 seconds)
19:33:06amachronicremoving reset_poweroff_timer() in usb.c gives 80 bytes (safe since nothing else in powermgmt is used)
19:33:24 Join JanC [0] (~janc@user/janc)
19:36:24_bilgusfat_time might be able to be shrunk
19:36:32 Join Saijin_Naib [0] (
19:37:44amachronicyeah, that one is quite big
19:38:32amachronicif we can get rid of lseek from load_firmware_filename that gives 190 bytes
19:39:27***Saving seen data "./dancer.seen"
19:40:40amachronicbasically everyone is putting CRC_OFFSET = 0 and DATA_OFFSET = 8, so I think that's feasible
19:47:41 Quit jacobk (Ping timeout: 240 seconds)
19:51:23_bilguswell you could use a macro right? #if FIRMWARE_OFFSET_FILE_CRC == 0
19:53:27_bilguslooks like it only read 4 though so I guess you'll have to read 4 more bytes
19:54:53amachronicthat function only works with the "scramble -add" format so the defines are redundant
19:55:52amachronicthe few that aren't 0/8 seem wrong, like CRC_OFFSET = DATA_OFFSET = 0 in samsungyh820.h
19:57:12_bilgusoh nm I read that worng its absolute then
19:57:34amachronicwith these various hacks I got it to fit with backtraces:
19:58:19amachronicthe currently broken part is the file mrsw lock, I suspect if we make everything read-only it can be a normal mutex
19:59:30_bilguswell given a few options maybe that will help narrow down the crash incase some thumb bug is happening
19:59:57_bilgusI really wouldn't want USB without being able to write to it TBH
20:02:04amachronicUSB works on a lower layer, it can still have write access even if the filesystem doesn't
20:02:53amachronicknocking out mkdir and O_CREAT saves 2.2k
20:02:54_bilgusIn that case hell yeah
20:03:37_bilgusin theory it'll make the bootloader that much faster
20:05:49 Join jacobk [0] (
20:19:08_bilguslcd_clear_viewport can call Clear_display instead
20:25:39_bilgus470 bytes
20:41:28amachronicbut clear_display calls clear_viewport
20:49:11_bilgusok I guess I better not bother testing that
20:50:14 Join othello8 [0] (
20:51:10_bilgusbitmap common calls clear display but only if DEPTH is < 8 but that clear display is in 16 bit
20:53:06 Quit Saijin_Naib (Ping timeout: 268 seconds)
20:59:50 Quit othello8 (Quit: othello8)
21:12:32 Join JanC_ [0] (~janc@user/janc)
21:12:44 Nick JanC is now known as Guest9843 (~janc@user/janc)
21:12:44 Nick JanC_ is now known as JanC (~janc@user/janc)
21:12:47 Quit Guest9843 (Ping timeout: 256 seconds)
21:39:28***Saving seen data "./dancer.seen"
21:51:23 Quit sam_d (Read error: Connection reset by peer)
21:53:59 Join sam_d [0] (~sam@user/sam-d/x-8933526)
22:52:48 Join Saijin_Naib [0] (
23:04:59 Quit JanC (Ping timeout: 256 seconds)
23:05:54 Join JanC [0] (~janc@user/janc)
23:14:07 Quit jacobk (Ping timeout: 260 seconds)
23:20:09 Join parahola [0] (~a@
23:39:29***Saving seen data "./dancer.seen"

Previous day | Next day