#qi-hardware IRC log for Thursday, 2013-08-22

DocScrutinizer05mth: OTG in driver *usually* shouldn't introduce any problems, even when the musb-core can't do it00:36
DocScrutinizer05it's just "dead code" since the core never sends the messages that would signal that its state machine gone to OTG mode00:37
apeletemth: thanks for the tip, having the jz4770 example should be very helpful07:18
apeletelarsc: Hi21:16
apeleteI made the changes to integrate musb-jz4740 to the kernel build system and built it successfully: http://seketeli.fr/git/~apelete/qi-kernel.git/commit/?h=jz4740-udc&id=7dedb95c72eb73a0dd4f15fd9bd650872351040621:18
apeleteI set it to use Inventra dma engine from Mentor Graphics just because I don't know better, will see when I get to dma configuration hacking21:20
apeletethe kernel is booting just fine, but the driver is not loading for nom: http://paste.debian.net/28702/21:21
apelete(see "musb-jz4740: failed to get clock" error message)21:21
mthapelete: the clock is named "udc" instead of "usb"21:22
mthsince there is both a USB host (uhc) and device (udc) controller21:23
apeletemth: ah. you mean I should write: clk = clk_get(&pdev->dev, "udc");21:25
apeleteinstead of: clk = clk_get(&pdev->dev, "usb"); ?21:25
apeleteI am trying to guess a lot of things by looking at other drivers actually (thanks mth for pointing me to the musb code for jz4770 btw)21:31
mthyou can simplify the probe and remove code by using the devm_ functions21:32
mththose automatically release resources if the probe fails or when the driver is unloaded21:33
apeleteI saw that in the jz4770 musb code, I was wondering about it :)21:40
apeleteseems convenient, I will try to use the devm_ functions indeed21:41
apeletehmmm, naming the clock "udc" instead of "usb" seems to cause a panic: http://paste.debian.net/28708/21:49
apeletetime to think about adding some deubg parameters into the kernel configuration21:54
--- Fri Aug 23 201300:00

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