sdhc trouble

[Patrick Welche <prlw1%cam.ac.uk@localhost>]

After seeing all those lovely sdhc.c commits, I thought I would try writing
to a card via a

sdhc0 at pci1 dev 0 function 0: Ricoh 5U823 SD/MMC Controller (rev. 0x04)
sdhc0: interrupting at ioapic0 pin 16
sdhc0: SD Host Specification 3.0, rev.5
sdhc0: software reset reg=00000001
sdhc0: using DMA transfer 
sdhc0: base clock frequency 50 MHz
sdhc0: max block length 512 bytes
sdmmc0 at sdhc0 slot 0

again. Here is the end of with debug=1

  mount /dev/ld0e /mnt
  cd /mnt
  ls

as seen by syslog:

sdhc0: writing cmd: blksize=0 blkcnt=0 mode=0002 cmd=0d1a
sdhc0: resp = 00000900
sdhc0: cmd 13 done (flags=00000c33 error=0)
sdmmc0: cmd 13 arg=0x70000 data=0x0 dlen=0 flags=0xc33 (error 0)
sdmmc0: resp=00 09 00 00 
sdmmc_mmc_command: error=0
ld0: I/O operation (dir=IN, blkno=0x4000, bcnt=0x8000)
ld0: I/O operati6e000, size=0x8000
ld0: I/O operati6e000, size=0x8000sdhc0: start cmd 18 arg=00004000 
data=0xffff800046438000 dlen=32768 flags=00000572, status=0
sdhc0: writing cmd: blksize=512 blkcnt=64 mode=0037 cmd=123a
sdhc0: resp = 00000900
sdhc0: resa transfer: resp=00000900 datalen=32768
sdhc0: software reset reg=00000006
sdhc0: softwaransfer done (error=60)
sdhc0: softwaransfer done =00000573 error=60)
sdhc0: somd 18 arg=0x4000 data=0xffff800046438000 dlen=32768 flags=0x573 (error 
60)
sdmmc_mmc_command: error=60
sdmmc_mmc_ 60
sdmmc_mmc_ 608192-8255 (ld0 bn 16384; cn 4 tn 4 sn 4)
ld0: I/O operation (dir=IN, blkno=0x4000, bcnt=0x8000)
ld0:0: addr=0x116e000, size=0x8000
ld0:0: addr=0x116e0cmd=18, arg=0x4000, flags=0x572
ld0:0: addrt cmd 18 arg=00004000 data=0xffff800046438000 dlen=32768 
flags=00000572, status=0x1008000
sdhc0: writing cmd: blksize=512 blkcnt=64 mode=0037 cmd=123a
sdhc0: software reset reg=00000006
sdhc0: cmd 18 abort (flags=00000573 error=60)
sdmmc0: cmd 18 arg=0x4000 data=0xffff800046438000 dlen=32768 flags=0x573 (error 
60)
sdmmc_mmc_command: error=60
ld0: error 60
ld0e: error reading fsbn 8192 of 8192-8255 (ld0 bn 16384; cn 4 tn 4 sn 4)

(error 60 is an ETIMEDOUT, I tried things like hz -> 100*hz for the
*_TIMEOUT defines, but not much changed - that was just guessing...)

Cheers,

Patrick