gvirstor & UFS

[Ivan Voras]

I'm having trouble recovering from "ENOSPC" situation in gvirstor - when
there's enough space on the virtual device, but not enough physical
space. No matter what I return to the upper layers (UFS), including EIO,
it seems to keep on retrying, spitting enormous amounts of messages to
the kernel log in g_vfs_done (and during this the console is stuck).
This is the same problems reported by testers some time ago.

Maybe the solution is as simple as sticking a check for ENOSPC somewhere
in UFS code to make it a special case, but I doubt it's that simple. The
required behaviour is to, if this condition is reached, drop the current
IO request (returning appropriate errno to the application) and stop
retrying (or at least ignore further error messages) until the condition
is cleared.

Any suggestions?

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 250 bytes
Desc: OpenPGP digital signature
Url : http://lists.freebsd.org/pipermail/freebsd-fs/attachments/20070327/c2305082/signature.pgp