SDIO stability [was Re: SDIO Performance once again]

[Sven Neumann]

Hi,

On Mon, 2009-02-09 at 10:59 -0500, Dan Williams wrote:

> You might try using the MMC debug options to see what's actually
> happening at the SDIO level, and see if there are errors even below the
> libertas code.  If the MMC/SDIO stack and the card are having problems
> talking, obviously the libertas driver is going to have problems too.

OK, I've enabled CONFIG_MMC_DEBUG in my kernel build. But I don't quite
know how to interpret the debug output. What I get is something along
these lines:

mmc1: starting CMD53 arg 9a000007 flags 000001b5
mmc1:     blksz 256 blocks 7 flags 00000100 tsac 1000 ms nsac 0
PXAMCI: irq 00000004 stat 00002040
PXAMCI: irq 00000001 stat 00003940
mmc1: req done (CMD53): 0: 00002000 00000000 00000000 00000035
mmc1:     1792 bytes transferred: 0
PXAMCI: irq 00000800 stat 0000b940
mmc1: starting CMD52 arg 00000a00 flags 00000195
PXAMCI: irq 00000004 stat 0000a140
mmc1: req done (CMD52): 0: 00001002 00000000 00000000 00000034
mmc1: starting CMD52 arg 10000a00 flags 00000195
PXAMCI: irq 00000004 stat 0000a140
mmc1: req done (CMD52): 0: 00001002 00000000 00000000 00000034
mmc1: starting CMD52 arg 90000afd flags 00000195
PXAMCI: irq 00000004 stat 00002140
mmc1: req done (CMD52): 0: 000010fd 00000000 00000000 00000034
mmc1: starting CMD52 arg 10004000 flags 00000195
PXAMCI: irq 00000004 stat 00002140
mmc1: req done (CMD52): 0: 00001009 00000000 00000000 00000034

Sorry if that is a stupid question. But what should I be looking out for
to determine if the MMC/SDIO stack and the card are having problems
talking to each other?


Thanks for your help,
Sven