Re: [Nbd] nbd-server segfault on x86_64

[Wouter Verhelst <wouter@...3...>]

On Mon, Feb 13, 2006 at 10:53:32PM +0100, JaniD++ wrote:
[...]
> > > The system:
> > > P4 Cual core(64 bit), Fedora Core 4 X86_64, Kernel 2.6.16-rc1-git4, nbd
> > > 2.8.2, compiled on this system.
> >
> > I believe these problems have been fixed in 2.8.3, though I'm not
> > entirely sure.  Could you try with 2.8.3? If that does not work, we'll
> > need to debug a bit more.
> 
> Not needed.
> It looks like fixed on 2.8.3.

Right, I thought as much.

> > > And another one issue:
> > >
> > > My client machine is i386 system, and i get this, when i try to
> > > connect to 3.3TB blockdev: "Exported device is too big for me. Get
> > > 64-bit machine :-("
> > >
> > > This is really need 64 bit machine?
> > > This problem is cannot be solve with bit-shifting?
> >
> > Maybe, but that would require changes to the kernel driver.
> >
> > > This needs the >4GB memory access?
> >
> > No, it needs a useful unsigned long data type to be able to store
> > offsets.
> >
> > I'm not 100% sure this check is still required, though. I'll try to find
> > out.
> 
> Can i help to try / test?

No, I can test it using sparse files. No need :-)

> (I have 4 nodes, all 3.3TB, but the NBD looks only 4x 2.0TB, and inside one
> 8TB XFS, with valuable data.
> But the blockdevice is free from 8TB to 13.2TB.)
> 
> only note:
> The NBD write-deadlock problem is finally fixed in kernel 2.6.16!

Whee!

> (thanks, to Herbert Xu, and me :-)

That's _really_ cool. Can you point me to the details of that one? I'd
be interested in finding out what the heck was wrong; I could then
perhaps get a workaround in nbd-server for older kernels as well.

-- 
Fun will now commence
  -- Seven Of Nine, "Ashes to Ashes", stardate 53679.4