#qi-hardware IRC log for Wednesday, 2013-08-21

apeletelarsc: are you there ?19:40
apeleteI'm trying to test the musb glue layer skeleton I' ve written: http://seketeli.fr/git/~apelete/qi-kernel.git/commit/?h=jz4740-udc19:41
apeletewhen I load it nothing happens19:41
larscdid you create a platform device for it in qi_lb60.c?19:42
apeleteno, should I ?19:43
larscyes, or better just change the name of the jzudc device19:44
larscin arch/mips/jz4740/platform.c19:45
apeleteI don't get it. qi_lb60.c seems to be asoc related, why should I create a platform device there ?19:46
larscboard-qi_lb60.c, sorry19:47
apeletein board-qi_lb60.c, I found:19:58
apeletestatic struct platform_device *jz_platform_devices[] __initdata = {19:58
apeletewhich correspond to the following in arch/mips/jz4740/platform.c:19:58
apeletestruct platform_device jz4740_udc_device = {19:58
apelete.name= "jz-udc",19:58
apelete 19:58
larscyes, replace "jz-udc" with "musb-jz4740"20:00
apeleteokay, so I don't need to modifiy drivers/usb/musb/Makefile and drivers/usb/musb/Kconfig to add rules for musb-jz4740 ?20:02
apeletebecause all the musb glue layers seem to have some sort of configuration options in those latter files20:03
apeleteand they (the other musb glue layers) all depend on the USB_MUSB_HDRC kernel config parameter20:05
apeleteI compiled musb-jz4740 out of tree and loaded it, but nothing happened. so I think I'm missing the corresponding USB_MUSB_HDRC driver in the kernel: I think musb-jz4740 should be used by USB_MUSB_HDRC driver20:10
apeletelarsc: am I mistaking here ?20:10
larscyes, you need to load the musb module as well20:16
apeleteok. so I need to modify drivers/usb/musb/Makefile and drivers/usb/musb/Kconfig as well, because the musb module cannot be activated in the menuconfig without a platform glue layer20:19
apeletelarsc: I do not know what to put in the drivers/usb/musb/Kconfig for musb-jz4740 exactly, I explain:20:23
apeletelet's take for instance the platform glue layer config section of a TI controller in drivers/usb/musb/Makefile and drivers/usb/musb/Kconfig:20:24
apeleteprompt "Platform Glue Layer"20:24
apelete 20:24
apeleteconfig USB_MUSB_DAVINCI20:24
apeletetristate "DaVinci"20:24
apeletedepends on ARCH_DAVINCI_DMx20:24
apelete 20:24
Last message repeated 1 time(s).20:25
apeletelarsc: it depends on ARCH_DAVINCI_DMx. what should musb-jz4740 depends on ?20:25
apeleteI'm thinking of something like ARCH_MIPS maybe, but not sure about that20:26
apeletes/ config section of a TI controller in drivers/usb/musb/Makefile and drivers/usb/musb/Kconfig / config section of a TI controller indrivers/usb/musb/Kconfig /20:28
larscdepends on MACH_JZ4740 should work20:53
apeleteok thanks21:01
apeletelarsc: I won't bother you longer, will try those modifications first and see what happens :)21:02
apeletethanks a lot for your time21:02
larscno problem21:05
DocScrutinizer05MUSB WAAAAAH!!!21:53
DocScrutinizer05don't tell me JZ4740 is using the fsckdup mentalgrafucks musb core21:55
DocScrutinizer05had my year of torture trying to fix stuff in drivers for that crap21:56
DocScrutinizer05ask PaulFertser what he thinks about musb-hdrc ;-P21:56
DocScrutinizer05ARCH_DAVINCI is OMAP36xx, no?21:58
DocScrutinizer05actually Nokia's idea of a OMAP36xx platform21:59
larscdavinci is davinci22:00
DocScrutinizer05anyway DAVINCI been the codename for the N9(50) prototypes at Nokia22:00
larscthey sometimes share IP cores between davinci and omap though22:01
DocScrutinizer05so what's the "top" meaning of davinci term? A SoC, a platform, a project, a son of flying spaghetti monster?22:04
larscfamily of SoCs22:05
DocScrutinizer05right, I seem to recall something at TI's website22:06
DocScrutinizer05might be wrong though22:06
DocScrutinizer05thought as much: http://www.ti.com/sitesearch/docs/universalsearch.tsp?searchTerm=DAVINCI%20PROCESSORS&linkId=122:10
DocScrutinizer05http://processors.wiki.ti.com/index.php/Category:DaVinci_Linux looks actually not too bad22:16
mthapelete: we've got a JZ4770 kernel here that uses the musb driver: https://github.com/gcwnow/linux23:16
mththe 4770 version can do OTG, which I guess the 4740 version cannot, so it's probably not exactly the same23:16
mththe glue is slightly cleaned up compared to Ingenic's code, but it's not great yet23:17
mthit does work reliably though23:17
--- Thu Aug 22 201300:00

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