Messages in this thread | | | Date | Tue, 24 Jun 2008 11:34:53 +0200 | From | Nicolas Ferre <> | Subject | Re: sdio: bad CISTPL_FUNCE error |
| |
Nicolas Pitre : > On Mon, 23 Jun 2008, Nicolas Ferre wrote: > >> Here is the log : >> mmc0: bad function CISTPL_FUNCE size 28 type 1 > > If you look in cistpl_funce_func() you'll find this code: > > vsn = func->card->cccr.sdio_vsn; > min_size = (vsn == SDIO_SDIO_REV_1_00) ? 28 : 42; > if (size < min_size || buf[0] != 1) > return -EINVAL; > > I suspect your card might be buggy wrt the SDIO standard. > > You could try printing out the value of vsn here, and then hardcode it > to SDIO_SDIO_REV_1_00 to see if that helps.
Indeed, I hardcode the value with SDIO_SDIO_REV_1_00 and the card seems to identify itself better :
vsn = 0x1, buf[0] = 0x1 mmc0: starting CMD52 arg 0020e400 flags 00000195 Sending command 52 as 00001074, arg = 0020E400, blocks = 0, length = 0 (MR = 00001B01) MCI irq: status = 0000C1E5, C07F0001, 00000001 Command ready Completed command Status = 00000001/0000c1e5 [000010FF 00000000 00000000 00000000] mmc0: req done (CMD52): 0: 000010ff 00000000 00000000 00000000 mmc0: read tpl 0xff mmc0: new SDIO card at address 0001
So far, so good...
But when going to the next step, it hangs in sdio_uart probing : mmc0: starting CMD52 arg 800220c0 flags 00000195 Sending command 52 as 00001074, arg = 800220C0, blocks = 0, length = 0 (MR = 00001B01) MCI irq: status = 0000C1E5, C07F0001, 00000001 Command ready Completed command Status = 00000001/0000c1e5 [000010C0 00000000 00000000 00000000] mmc0: req done (CMD52): 0: 000010c0 00000000 00000000 00000000 mmc0: starting CMD52 arg 80022201 flags 00000195 Sending command 52 as 00001074, arg = 80022201, blocks = 0, length = 0 (MR = 00001B01) MCI irq: status = 0000C1E5, C07F0001, 00000001 Command ready Completed command Status = 00000001/0000c1e5 [00001101 00000000 00000000 00000000] mmc0: req done (CMD52): 0: 00001101 00000000 00000000 00000000 sdio_uart: probe of mmc0:0001:1 failed with error -34
It seems that I have a out of range error in R5 now...
Regards, -- Nicolas Ferre
| |