Messages in this thread | | | Date | Fri, 16 Jan 1998 18:34:34 -0500 (EST) | From | "Steven N. Hirsch" <> | Subject | Re: 2.1.79 + latest smbfs patch |
| |
On Fri, 16 Jan 1998, Bill Hawes wrote:
> David Burrows wrote: > > > > Its still a no go on my system I'm afraid. I am scratching my head at the > > moment as to why some people report success and others don't. I run > > RedHat 5 here, which means glibc2, maybe that causes a problem with the > > mount tools.. Well I know it does, because I had to fiddle with the > > includes to get it to compile. Changed a few things like, where files > > caused problems I'd delete the include reference.. and then define what > > was missing manually. > > I thought of some tests you could do to check whether there's a glibc > problem. If you check the definition of the mount structure passed to > smbfs (in smbfs/inode.c, smb_read_super), the values in here must agree > with what's being passed from the smbount program. So if you print them > out on the userland side and again in the kernel (just add some > printks), all the values should agree. > > If something isn't right, it may be a difference in the glibc definition > (e.g. a short vs a long.)
Bill,
If the mount structure has a 'dev_t' field, that is almost certainly the problem. This bit me hard when getting the knfsd support utilities running on the Alpha. Glibc defines dev_t as a 64-bit long, IIRC.
Steve
| |