>
> Hi
>
> The following is a tiny patch to libxfs/init.c
> to use `unsigned long' as the recipient in the
> BLKGETSIZE ioctl in findsize(). This ioctl can
> return a valid unsigned 32-bit quantity, and it
> does seem to be making mkfs.xfs work on a 1.2TB
> raid setup we have here, where it was failing
> with a negative number of blocks before.
Thanks for the patch.
Hmm, I see you just joined the mailing list, so you probably missed the
message I sent out to someone with a 1.1Tbyte raid earlier today. Take
a look at this message or your inode numbers may end up overflowing 32
bits.
http://oss.sgi.com/projects/xfs/mail_archive/0107/msg00727.html
Steve
>
>
> ----
> Chris J. Bednar <http://optics.tamu.edu/~bednar/>
> Director, Distributed Computing Product Group
> http://AdvancedDataSolutions.com/
> -------------------------------------------------------
>
> --- xfsprogs-1.3.1/libxfs/init.c.findsize Thu Jul 12 13:54:41 2001
> +++ xfsprogs-1.3.1/libxfs/init.c Tue Jul 17 13:25:36 2001
> @@ -132,7 +132,7 @@
> {
> int fd;
> int error;
> - long size;
> + unsigned long size;
> struct stat64 st;
>
> /* Test to see if we are dealing with a regular file rather than a
|