#qi-hardware IRC log for Wednesday, 2013-11-13

wpwrakhmm, i think the mantle (née maemo) guys should add an entry to wikipedia. now that AMD are using "mantle" as well, there'll be a heck of a confusion09:37
apeletemorning everyone10:38
apeletelarsc mth: just found out why the musb core couldn't discover the udc config by reading from hw10:40
apeleteit seems there is a register called configdata that usually holds usb config data, but it does not exists in the jz474010:41
apeletethat means there is no way to autodetect config data nor endpoints...10:42
apeleteI'm angry with myself for having wasted some much time on this :-(10:43
apeleteso, next step would be to find out how to trick musb core into calling ep_config_from_table() instead of ep_config_from_hw() to setup endpoints10:47
DocScrutinizer05wpwrak: fremantle14:38
DocScrutinizer05apelete: :EEEEK:14:39
DocScrutinizer05apelete: this all sounds so terribly similar to H-E-N hostmode14:40
DocScrutinizer05apelete: temporarily we had a patch that allowed us to step maually thru every possible state of musbcore state machine, by sending commands to err /sys/musb/state or whatever (I forgot the name of sysnode, or was it even in /proc ?)14:41
apeleteDocScrutinizer05: let's hope the rest of the udc core in jz4740 is more "usb spec compliant"15:03
apeletethere seems to be some sort of support for devices that does not hold config data in the hardware, but I don't know how well it's going to work with jz474015:04
apeletes/support/support in the kernel/15:05
kyakviric: why you're making it so hard to package? :) The download URL is http://vicerveza.homeunix.net/~viric/cgi-bin/filegive/tarball/filegive-0.7.1.tar.gz?uuid=v0.7, that means i need to hardcode 0.7 or use two separate variables15:06
kyakviric: also, do you publish gpg key or md5sum somewhere? Just to be sure i downloaded what you actually packaged15:07
apeleterelying on ep_config_from_table() does not make it any better -> http://paste.debian.net/65724/:23:09
apelete[    1.230000] musb-jz4740: glue kzalloc OK!23:09
apelete[    1.240000] musb-jz4740: platform_device_alloc OK!23:09
apelete[    1.250000] musb-jz4740: devm_clk_get OK!23:09
apelete[    1.260000] musb-jz4740: clk_prepare_enable OK!23:09
apelete[    1.270000] musb-jz4740: platform_set_drvdata OK!23:09
apelete[    1.280000] musb-jz4740: platform_device_add_resources OK!23:09
apelete[    1.290000] musb-jz4740: platform_device_add_data OK!23:09
apelete[    1.300000] musb-jz4740: musb->dyn_fifo = 123:09
apelete[    1.310000] musb-jz4740: init OK!23:09
apelete[    1.320000] musb-hdrc: ConfigData=0x00 (UTMI-8)23:09
apelete[    1.330000] musb-hdrc: kernel must blacklist external hubs23:09
apelete[    1.340000] musb-hdrc: HDRC RTL version 2.30023:09
apelete[    1.350000] musb-hdrc: setup fifo_mode 223:09
apelete[    1.360000] musb-hdrc: fifo_setup failed with -EMSGSIZE = -9723:09
wpwrakhow about using pastebins ? :)23:10
wpwrakprobably won't make the code work any better, though23:11
apeletewpwrak: just did above, but thought pasting the relevant part would help, sorry about that :-)23:12
apelete 23:12
Last message repeated 1 time(s).23:13
apeleteanyway, the -EMSGSIZE return code is from fifo_setup() in musb_core.c, it seems ram address size is not read from raminfo register here...23:14
apeletethis is getting tedious, going to give up for tonight :-)23:14
wpwrakdon't you have a working driver you can use for comparison ? that's normally the quickest way to figure out where the problems are23:36
--- Thu Nov 14 201300:00

Generated by irclog2html.py 2.9.2 by Marius Gedminas - find it at mg.pov.lt!