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 2021-09-18

00:43:00 Quit dconrad (Remote host closed the connection)
01:23:30 Quit Moriar (Ping timeout: 268 seconds)
01:43:47***Saving seen data "./dancer.seen"
02:17:30 Join ZincAlloy [0] (
02:19:19 Join lebellium [0] (
02:20:45 Quit massiveH (Quit: Leaving)
02:24:16 Quit lebellium (Quit: Leaving)
02:26:14 Join lebellium [0] (
02:28:23 Join lebellium_ [0] (
02:31:10 Quit lebellium (Ping timeout: 240 seconds)
03:43:49***Saving seen data "./dancer.seen"
05:02:24 Quit S|h|a|w|n (Quit: Leaving)
05:43:51***Saving seen data "./dancer.seen"
06:28:59 Quit pixelma (Quit: .)
06:28:59 Quit amiconn (Quit: - Chat comfortably. Anywhere.)
06:31:53 Join pixelma [0] (
06:31:53 Join amiconn [0] (
07:28:34 Quit Piece_Maker (Ping timeout: 268 seconds)
07:43:54***Saving seen data "./dancer.seen"
08:06:18 Join Piece_Maker [0] (
09:43:58***No seen item changed, no save performed.
11:43:59***No seen item changed, no save performed.
11:45:44 Join Moriar [0] (
12:45:00 Join Strife89 [0] (
13:17:06 Join dconrad [0] (~dconrad@
13:44:02***No seen item changed, no save performed.
13:44:28 Join amachronic [0] (~amachroni@user/amachronic)
13:45:57amachronicquick question, is there anyone still active familiar with RB's usb stack?
13:48:05speachyI've gone spelunking into it far, far more than I'd like (when sorting out the jz47xx USB issues)
13:48:50speachybut I don't think any of the folks directly responsible for it are still active.
13:49:20amachronicokay. I guessed so.
13:50:17amachronicI've realized a lot of control transfer related stuff is broken, both in drivers and in the common usb code.
13:51:20speachyThat doesn't surprise me, given the persistent reports of intermittent flakiness on a wide variety of targets.
13:51:35amachronicfor example, the core code doesn't return STALL properly for all types of control transfers. and control writes pretty much don't work.
13:52:09amachronicone of the USB drivers doesn't even know the difference between stalling the in vs. out endpoint.
13:52:33CtcpIgnored 2 channel CTCP requests in 11 hours and 19 minutes at the last flood
13:52:33*speachy winces.
13:52:40speachyI ran into that in the jz47xx code too
13:53:22speachywhere it gets messy is that there's a ton of variation in the relative capabilities of our USB drivers, some handle more stuff automatically.
13:54:18speachyand we sort of tried to make later drivers map the APIs of the original ones.
13:55:10amachronicthe current API takes a micromangement approach. I think this needs to be changed.
13:55:37amachroniclike, instead of send/recv ZLP packets and other low-level voodoo we should handle some stuff at a higher level.
13:56:19amachronicI'm just not sure how to do this without breaking every other USB driver.
13:57:09speachy... gotta start somewhere
13:57:36speachythere was a big SD API rewrite along those same lines a while back.
13:58:15speachyour dev builds have been remarkably stable for the past couple of years. :D
13:59:53speachy... well, propose a new API, I'll be glad to help review it. I assume you'll want to port the dwc2 driver to the new API; I'll do the same with the jz47xx
14:00:27speachyand we can work back from there
14:01:00amachronicthat'd be great, thanks.
14:01:01speachythe as3525 and pp50xx are probably the most important ones to have covered
14:01:34speachyI wonder if this b0rkage has anything to do with that optmization-related regression on the as3525v1 targets..
14:02:23speachyI have a proper hardware USB sniffer BTW.
14:03:17amachroniccool. I'm stuck guessing with wireshark.
14:07:31speachy(turns out the dwc2 in host mode doesn't quite handle flow control properly, causing a DoS on some devices)
14:15:13speachyoh, one thing I should point out first −− when I hacked in that "prompt the user what to do" when USB was inserted
14:16:03speachyon some controllers we only know USB is connected because a posted URB gets returned.
14:16:14speachyothers we're told and can initiate it at our leisure
14:16:28speachythat sort of crap really complicates clean APIs.
14:17:26amachronicdconrad was fighting with that a while back, for some reason.
14:19:59amachroniccontrol requests seem to be the main buggy thing so that's where I'm focusing for now.
14:22:25speachygotta run and deal with more meatspace stuff. but I'm looking forward to seeing what you come up with
15:10:31 Quit amachronic (Quit: amachronic)
15:33:37 Quit Strife89 (Quit: Connection closed for inactivity)
15:44:03***Saving seen data "./dancer.seen"
17:00:37 Quit Moriar (Ping timeout: 268 seconds)
17:10:00 Quit dconrad ()
17:44:04***Saving seen data "./dancer.seen"
17:55:18 Quit lebellium_ (Quit: Leaving)
18:48:09 Quit ZincAlloy (Quit: Leaving.)
18:49:10 Join Strife89 [0] (
19:08:29 Quit tomato (Ping timeout: 268 seconds)
19:41:04 Join tomato [0] (~tomato@user/tomato)
19:44:08***Saving seen data "./dancer.seen"
21:08:12 Join Moriar [0] (
21:15:19 Quit Piece_Maker (Ping timeout: 252 seconds)
21:25:56 Join Piece_Maker [0] (
21:29:14 Join massiveH [0] (
21:44:11***Saving seen data "./dancer.seen"
22:18:36 Join advcomp2019 [0] (~advcomp20@user/advcomp2019)
22:21:50 Quit advcomp2019_ (Ping timeout: 260 seconds)
22:23:01 Quit advcomp2019 (Ping timeout: 252 seconds)
22:23:21 Join advcomp2019 [0] (~advcomp20@user/advcomp2019)
23:33:37 Quit Strife89 (Quit: Connection closed for inactivity)
23:44:13***Saving seen data "./dancer.seen"

Previous day | Next day