Re: flash-kernel puts QNAP on its knees
On Sun, 2015-09-06 at 13:02 +0200, Rob J. Epping wrote:
> Hi,
>
> On 08/21/2015 02:11 PM, Ian Campbell wrote:
> > On Sun, 2015-08-09 at 09:39 +0200, JM wrote:
> >>
> >> This is just a hunch, but perhaps a part of the problem is because
> >> flash-kernel simply cats the image to mtdblock, which is not the
> >> recommended way to deal with flash storage [1].
> >
> > I'm afraid I disagree with this, cat-ing to mtdblock is IMHO
> absolutely
> > fine.
>
> Breaks on my TS-219 and TS-221 too, unfortunately.
>
> > The article you cite is regarding putting normal (i.e non-flash
> -aware)
> > filesystems onto mtdblock devices, but flash-kernel is using the
> raw
> > device which is fine and exactly the sort of thing mtdblock is for.
> >
> >> When I have more time, I'd like to try flashcp from mtd-utils and
> see
> >> if the deadlock still occurs. If it does, it's likely a kernel bug
> >> somewhere in handling flash storage.
> >
> > Despite the above please do let us know if this makes any
> difference?
>
> I hacked flash-kernel to use flashcp from mtd-utils. Where using cat
> triggers ATA timeout and reset, using flashcp worked flawlessly.
AIUI flashcp writes via /dev/mtdN, whereas cat is using mtdblockN. This
suggests the mtdblock driver is at fault.
Ian.
Reply to: