Re: Compactflash (Sandisk 512) hangs on access

[Prashant Viswanathan]

On Fri, 28 Jan 2005 22:12:04 -0800, Prashant Viswanathan
<vprashant@xxxxxxxxx> wrote:
> > > I have been trying unsuccessfully over the last 2 weeks to get
> > > compactflash working on my Linux system based on mini-ITX (Via CL
> > > motherboard, pentium compatible).
> > >
> > > I use a CF->IDE adapter to access it just like a IDE hard disk. My
> > > compactflash is Sandisk SDCFH-512. Linux can detect it. I can even
> > > mount it and do a fdisk on it. However, the moment I try to do
> > > anything substantial like copy multiple files or copy 1000 blocks
> > > using dd, I lose access to it. Linux loses access to it totally. I
> > > can't even do a fdisk on it. I get an error like "Unable to open
> > > /dev/hdc".
> 
> On Thu, 27 Jan 2005 22:07:35 +0100, Willy Tarreau <willy@xxxxxxxxx> wrote:
> > Have you checked that the power connector really provides 5V to the
> > IDE-CF adapter ? I had the exact same behaviour 5 years ago with a power
> > wire cut. Signal lines were powerful enough to bring power to the cheap
> > flash (16 MB), I could even read it, most times. The kernel almost always
> > booted from it, and when it turned to mount the ext2 fs R/W, it hanged. I
> > finally partially destroyed it this way, and it got several defects which
> > could not be cleaned with a simple write or format.
> >
> > Other than that, I have lots of CF cards on IDE adapters (some on motherboard,
> > some hand-made, some bought to serious makers), and never ran into such
> > problems since.
> >
> > Willy
> 
> The power connector is fine.
> 
> I also disabled DMA (some suggestion on this newsgroup to a similar
> error) and now I can't turn it back on.
> 
> <snip>
> everest root # hdparm -d1 /dev/hdc
> 
> /dev/hdc:
> setting using_dma to 1 (on)
> 


Finally figured out the problem (or at least worked around it).
Everything seems to work perfectly after I switched to an older 40
conductor IDE cable!  Before that I was using the newer 80 conductor
IDE cable. I sitll dont understand why the newer ( and apparently
better) cables are problematic. Higher speeds causing race conditions perhaps?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/