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 2020-04-04

00:13:16 Quit lebellium (Quit: Leaving)
00:14:47cockroachI'm trying to setup a new sd card for my ipod but Rockbox refuses to boot from it. Apparently, because the cluster size is too large. Since I really don't know much (anything) about FAT, can anyone tell me why that would be a problem?
00:20:36 Quit prof_wolfff (Ping timeout: 264 seconds)
00:34:58 Join prof_wolfff [0] (~prof_wolf@
01:04:33cockroachhmm, is this really the right place for the secmult?
01:05:08cockroachthis makes the bpb_secperclus value needlessly large for file systems with 2048 bytes per sector
01:05:44cockroachwhich is why says that the file system isn't sane
01:10:15cockroachor;a=blob;f=firmware/drivers/fat.c#l1129 and;a=blob;f=firmware/drivers/fat.c#l394 for more official (and less outdated) links
01:18:46 Quit krabador (Remote host closed the connection)
01:21:36 Join ac_laptop [0] (
01:32:43 Quit ac_laptop (Ping timeout: 265 seconds)
01:55:28***Saving seen data "./dancer.seen"
02:02:52 Quit GeekShadow (Ping timeout: 265 seconds)
02:03:30 Join GeekShadow [0] (
02:03:30 Quit GeekShadow (Changing host)
02:03:30 Join GeekShadow [0] (~antoine@reactos/tester/GeekShadow)
02:38:27speachyat first glance that does seem wrong
02:40:43cockroachyay, so I'm not entirely wrong! :)
02:43:07 Join reductum [0] (
02:43:56 Quit PimpiN8 (Quit: My MacBook has gone to sleep. ZZZzzz…)
02:45:28speachylooks like all uses of that variable treat it as sectors per cluster too.
02:46:18speachyso get rid of that secmult and see what happens..
02:46:31cockroachif I simply remove it it won't work
02:46:49cockroachrockbox doesn't find any files
02:47:20cockroachbut if you agree that it looks wrong, I'll see if I can come up with something that *does* work
02:49:09speachySECTOR_SIZE seems hardcoded to 512
02:49:57cockroachso we'll probably have to multiply it somewhere else then to get to the right address
02:51:22speachythat has a hardcoded SECTOR_SIZE of 4K
02:51:47speachyas do many of the ipods
02:51:52speachyipodvideo is 2K.
02:52:18cockroachisn't the 5g an ipod video?
02:52:39speachyI'm very unfamiliar with this code
02:52:52speachyand mostly clueless on the ipod ecosystem
02:53:26speachywhat I think is happening is that secmult usually == 1
02:54:11cockroachmakes sense. that's why it usually works.
02:54:14speachybut on targets that have funky sector sizes, that calculation ends up wrong
02:54:33speachyand usually ends up in a situation where two bugs cancel each other out. :)
02:54:40speachy... but perhaps not always.
02:54:47cockroachheh yep
02:55:25cockroachI'll probably need to go over this with a piece of paper, it too many things-per-other-thing to keep in mind.
02:55:50cockroachthe good thing is that I probably have a particularly bad setup with this sd card, so if that one works, it should be reasonably safe
02:59:36 Quit koniu (Quit: WeeChat 2.3)
03:00:25 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
03:03:12 Quit _Bilgus (Ping timeout: 240 seconds)
03:08:30 Quit Bilgus (Ping timeout: 260 seconds)
03:21:10 Quit prof_wolfff (Ping timeout: 265 seconds)
03:46:37speachysuccess! I _finally_ was able to build a working AGPTek Rocker bootloader + rockbox
03:46:47speachyincluding patching the flash binary
03:48:42speachyin the process I discovered my Rocker has had a loose screw inside it since I bought it. now the keypad feels much nicer..
03:52:49speachyunfortunately the image has to be extraced as root to maintain permissions/attributes. Otherwise when re-assembled it won't boot.
03:53:40CtcpIgnored 1 channel CTCP requests in 0 seconds at the last flood
03:53:40*__builtin needs to get a Rocker
03:54:56speachythe sound quality is only so-so and the battery life kinda sucks..
03:55:20speachythough it doesn't feel as flimsy with all the internal screws reattached.
03:55:29***Saving seen data "./dancer.seen"
03:57:05speachynow to see if I can repeat this with the X3.
03:57:57speachysweet, the clip+ with a gcc4.9.4 build is working too. guess the flash yet lives.
03:59:26 Quit cockroach (Quit: leaving)
04:29:37 Quit MrZeus_ (Ping timeout: 256 seconds)
04:30:08speachyX3II image patched right up, yay.
04:41:07 Quit reductum (Quit: WeeChat 2.8)
05:15:45 Quit TheSeven (Ping timeout: 240 seconds)
05:16:14 Join TheSeven [0] (~quassel@rockbox/developer/TheSeven)
05:55:33***Saving seen data "./dancer.seen"
06:06:48 Quit dys (Ping timeout: 264 seconds)
07:55:38***Saving seen data "./dancer.seen"
08:08:18 Join reductum [0] (
09:26:23 Quit reductum (Quit: WeeChat 2.8)
09:55:40***Saving seen data "./dancer.seen"
10:19:16 Join lebellium [0] (
10:20:46 Join ZincAlloy [0] (~Adium@2a02:8108:943f:d824:b008:3a9c:afe8:c3c6)
11:03:17 Join TheLemonMan [0] (~lemonboy@irssi/staff/TheLemonMan)
11:28:24 Join MrZeus_ [0] (
11:37:58gevaertsI'd recommend trying to find someone who knows about it before messing with large sectors and FAT
11:55:44***No seen item changed, no save performed.
12:10:40 Join petur [0] (~petur@rockbox/developer/petur)
13:29:47 Join dys [0] (
13:31:13 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
13:33:47 Join prof_wolfff [0] (~prof_wolf@
13:55:49***No seen item changed, no save performed.
14:04:24 Quit prof_wolfff (Ping timeout: 264 seconds)
14:19:11 Join prof_wolfff [0] (
14:24:56 Quit prof_wolfff (Ping timeout: 258 seconds)
14:29:27 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
15:30:16 Join cockroach [0] (~blattodea@pdpc/supporter/active/cockroach)
15:55:50***Saving seen data "./dancer.seen"
15:57:28 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
16:01:52 Quit pamaury (Ping timeout: 240 seconds)
16:12:07cockroachI've been looking into ways to really fix but from what I understand, the FAT code would need quite an overhaul to deal with different sector sizes. Which is probably why the current way of dealing with it (that secmult hack) was made in the first place.
16:30:11 Join johnb2 [0] (
17:08:30 Quit johnb2 (Ping timeout: 260 seconds)
17:12:38 Join pamaury [0] (~pamaury@rockbox/developer/pamaury)
17:33:33 Join krabador [0] (~krabador@unaffiliated/krabador)
17:35:09 Join PimpiN8 [0] (~PimpiN8@2001:1c04:3309:f200:2896:1ccb:b875:79bd)
17:37:47 Quit XDjackieXD (Read error: Connection reset by peer)
17:39:13 Join johnb2 [0] (
17:40:04 Join XDjackieXD [0] (
17:41:32 Join prof_wolfff [0] (~prof_wolf@
17:55:52***Saving seen data "./dancer.seen"
18:07:50 Quit krabador (Remote host closed the connection)
18:18:04 Quit johnb2 (Quit: Quit)
18:18:04 Quit Loeb (Ping timeout: 256 seconds)
18:25:06 Quit beencubed (Remote host closed the connection)
18:50:11 Quit pamaury (Ping timeout: 258 seconds)
18:56:27 Join krabador [0] (~krabador@unaffiliated/krabador)
19:04:42 Join beencubed [0] (~beencubed@
19:19:50 Quit _3dsv (Ping timeout: 256 seconds)
19:24:32 Join _3dsv [0] (
19:26:12 Quit prof_wolfff (Ping timeout: 240 seconds)
19:42:00 Join prof_wolfff [0] (~prof_wolf@
19:43:25 Quit Aldem (Read error: Connection reset by peer)
19:47:00 Quit prof_wolfff (Ping timeout: 264 seconds)
19:54:54 Quit krabador (Remote host closed the connection)
19:55:56***Saving seen data "./dancer.seen"
20:02:39 Join Aldem [0] (~Aldem@unaffiliated/aldem)
20:20:58 Quit sakax (Remote host closed the connection)
21:06:44bluebrotherspeachy: ok, that new mini2g version crashes on startup.
21:07:07bluebrotherundefined instruction at 0006fc18
21:07:36bluebrotherand something is off with the display, the text is cut off at both left and right border.
21:09:16speachyso.. it's worse. Heh.
21:09:43bluebrotherwell, at least it shows something other than the Rockbox logo at startup now :)
21:13:41bluebrotherIMO the patches to the compilers would be better somewhere like in tools/patches instead of downloading them from the website.
21:13:52bluebrotherso everything is in the same place
21:15:48speachyI just updated the mini2g build. Don't expect any functional change but I'd accidentally nuked the build dir with the debugging symbols
21:16:10speachyyeah, having the patches in the main repo makes sense.
21:18:33bluebrothersame URL as before? Then I'll try that once more.
21:18:41speachyyes please
21:23:53bluebrotherundefined instruction at 0006fc18
21:26:07speachythank you.
21:27:31speachywill look into that once I finish reworking to pull patches locally
21:27:59bluebrothernice, thanks.
21:41:38speachytake five, and it's good.
21:45:05speachyI need to clean this up for commit, and then rebase the 4.9.4 patch on top.
21:52:57 Quit TheLemonMan (Quit: "It's now safe to turn off your computer.")
21:54:09 Join sakax [0] (~r0b0t@unaffiliated/r0b0t)
21:55:58***Saving seen data "./dancer.seen"
21:57:23speachyokay, merged the "store toolchain patches locally" patch.
21:57:48speachyand g#2305 is rebased/updated.
21:57:50fs-bluebotGerrit review #2305 at : Build: Bump all compilers to GCC 4.9.4 (WIP, do not merge) by Solomon Peachy
21:59:44 Quit Bilgus (Read error: Connection reset by peer)
22:00:35speachyokay, that crash is in system_init
22:00:55speachy6fc18 udf #0
22:04:06bluebrothertime to build some toolchains now :)
22:04:36speachythis doesn't seem to be a toolchain "bug" so much as the compiler getting stricter about undefined behavior
22:05:27bluebrotherwouldn't be surprised about that :)
22:06:13speachythat UDF insertion (literally, "permanently undefined instruction") is deliberate and marked as WONTFIX by GCC.
22:10:28speachyyep, there it is, the end of system-pp502x.c:280
22:11:33speachywe're deliberately deferencing a null pointer
22:11:48bluebrothersounds like a good idea :)
22:12:36speachynow address 0 is valid, and we're just using it to fill the cache with valid data
22:14:45speachyokay, updated the zip with this fix.
22:15:00speachynow insted of reading it from DRAM we read it from IRAM.
22:15:47speachyactually, hang on a sec, let me change that
22:17:37speachyokay, now it's DRAM but starting at 1K instead of 0K
22:20:53 Join Bilgus [0] (~Bilgus@unaffiliated/bilgus)
22:21:56speachyd'oh, screwed up again.
22:22:44speachyyet another version uploaded.
22:37:39speachythis particular bug was brought out due to using -Os
22:50:44bluebrotherok, that's the first version that successfully rolo's. Still doesn't boot.
22:51:01bluebrotherdata abort at 000756bc
22:52:15speachywhat's the joke, 99 bugs in the code, 99 bugs in the code, fix one down, recompile it around, 101 bugs in the code...
22:52:33bluebrothernah, that's 127 after fixing the 1st one ;-)
22:57:23speachythat's in lock_blocker_thread(), looks like the data abort is due to a bad 'struct blocker *' argument
22:57:43speachycould be something unaligned or a completely trashed pointer.
23:03:49speachyoh, for a proper stack trace..
23:29:35 Quit dys (Ping timeout: 264 seconds)
23:31:01Bilgusit was 153 errors now its 152
23:31:30Bilgusforgot to convert to dec
23:38:17 Join krabador [0] (~krabador@unaffiliated/krabador)
23:56:01***Saving seen data "./dancer.seen"

Previous day | Next day