#qi-hardware IRC log for Monday, 2013-10-28

qi-bot[commit] Werner Almesberger: atusb/fw/: let include path find board.h instead of using ../board.h (master) http://qi-hw.com/p/ben-wpan/c5f9cc114:24
larscpcercuei: btw. if you don't want to fly back the same day but rather grab a beer (or two) in the evening, I think it shouldn't be a problem to find a couch for you14:59
pcercueisounds good :)14:59
pcercueiit'd already take me 3h to go to Paris by train so I don't think I can do it in just one day anyway15:04
pcercuei(I'm from Normandy)15:05
lekernelpcercuei, where are you going?15:49
pcercueiMunich15:50
pcercueifor a job interview15:50
rohheeeh.20:32
roh1°) find 10 persons who want to buy a M1 R3.5 board at approx 700-800 EUR 20:33
rohsomehow thats sad and funny at the same time... last i was asked if i'd want to sell my mm1 i was proposed to get 150-200E for it. which wasnt enough by far to bring me to sell it ;)20:34
cdethat's not even the whole BOM21:01
rohyup. atleast with all the analog stuff.21:02
rohthe digitzer alone is pricy..21:02
cdeI mean even an basic LX75 eval board is usually more than twice that price21:04
wpwrakwelcome to the free market. anyone is free to try to buy something way below its real value ;-)21:15
larscor above21:17
wpwrakwhere's the challenge in that ? :)21:18
larscthe challange is for the seller, I guess21:19
Action: roh is a techie not a businessman :)21:50
rohmeans i dont talk a lot, i just sell at $price. if somebody wants is too much lower.. i just dont sell.21:59
rohhttp://blog.ioactive.com/2013/10/hacking-counterfeit-money-detector-for.html22:00
apeletelarsc mth: Hi there. ran gdb on the kernel to understand why "musb-hdrc: missing bulk" was appearing as an error message during endpoints configuration -> http://paste.debian.net/62655/22:15
apeleteby putting a breakpoint in ep_config_from_hw() where the endpoints are scanned, it appears struct musb *musb->endpoints contains bogus data22:17
apeleteso that's why the musb_core cannot find a valid bulk endpoint (it seems ep0 data is also bogus btw)22:18
apeleteI know the udc reg address where endpoints info is stored, that's 0x13040078 22:21
apeletebut udc base reg address in platform data is 0x13040000 22:22
apeletelarsc mth: how is musb_core supposed to find out about the udc endpoints information that is stored in musb->endpoints ?22:27
apelete(forget about that epinfo reg address 0x13040078 I was talking about, it seems it only contains the total number of endpoints included in the design, according to programming manual)22:29
mthapelete: base reg is typically the address of the lowest register in a register bank; register access is done using base + offset22:53
mthnote that 0x13040000 is the physical address; it has to be ioremapped or otherwise converted to a kseg1 virtual address to get a valid pointer22:53
apeletegot that, but my question was more about why musb->endpoints data was wrong23:03
apeletemth: looking into the code right now to see where it should be initialised23:03
apeleteanyway, can't get my thinking straight right now, wil look more into it maybe tomorrow if I get some time23:11
--- Tue Oct 29 201300:00

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