Re: SD card on Dell D420 (Ricoh R5C822)

[Maxim Levitsky <maximlevitsky@gmail.com>]

On Thu, 2010-06-10 at 22:54 +0100, Mark Hills wrote: 
> On Thu, 10 Jun 2010, Maxim Levitsky wrote:
> 
> > On Thu, 2010-06-10 at 09:08 +0100, Mark Hills wrote:
> > > On Wed, 9 Jun 2010, Chris Ball wrote:
> > > 
> > > > Hi Mark,
> > > > 
> > > >    >   mmc0: error -84 whilst initialising SD card
> > > > 
> > > > Not sure how much this'll help, but:
> > > > 
> > > > errno -84 is defined in include/asm-generic/errno.h as EILSEQ, and
> > > > there are two occurences of EILSEQ in sdhci.c, one of which has a
> > > > printk next to it that you aren't seeing.  So, it's likely that you're
> > > > seeing:
> > > > 
> > > >     else if (intmask & (SDHCI_INT_CRC | SDHCI_INT_END_BIT | SDHCI_INT_INDEX))
> > > >         host->cmd->error = -EILSEQ;
> > > > 
> > > > So, it seems like your controller is returning a successful command
> > > > response, but with some sort of data error attached.  Have you tried
> > > > multiple SD cards, and are they known-working?
> > > 
> > > I have two SD cards -- a 2GB Crucial, and a 32MB Sandisk (actually a 
> > > micro-SD in an adaptor). Both cards work in a USB card reader with the 
> > > same kernel, and in my digital radio.
> > >  
> > > > If you're sure it's not the card, try passing debug_quirks=1 to the
> > > > sdhci module.  (Let me know if you're not sure how to do that.)
> > > 
> > > I've tried this with 2.6.35-rc2, and the result is the same. I also diff'd 
> > > the dmesg, with no relevant differences.
> > 
> > Could you disable the CONFIG_MMC_RICOH_MMC and try again?
> > This setting is only relevant to MMC cards, and I have seen some
> > problems with it.
> 
> Okay I disabled it, unfortunately with the same result.
> 
> > Also please post lsmod, dmesg, lspci output.
> 
> Sure, they are attached.
> 
> > Did you do a suspend/resume cycle before this test?
> 
> No, I didn't suspect or resume the laptop. The dmesg is from a cold 
> reboot, where I let the machine boot, then inserted the card (and the same 
> in the previous email)

In fact, you need to undo the work of CONFIG_MMC_RICOH_MMC you need to
shutdown the system and then boot into kernel that hasn't this option.
But anyway, I don't think its to blame.

Thanks for kernel log I hope I get enough clues from it.
I am interested in this problem because this isn't first time I seen it.


It might be related to too high clock or hispeed mode that card doesn't
support.

As suggested, try 'debug_quirks=1' too.

Best regards,
Maxim Levitsky