> Eugene Crosser wrote:
>
> > Is anyone working on the kernel nfs server? is linux-nfs-0.4.21.tar.gz
> > the most recent version? I tried to compile it with the kernel 2.1.112
> > and 2.1.115 and the tools did not compile because they indirectly include
> > asm/system.h which has in like 40:
> >
> > extern void FASTCALL(__switch_to(struct task_struct *prev, ....
> >
> > FASTCALL is defined in linux/kernel.h, but inside #ifdef __KERNEL__.
> > NFS tools are (naturally) compiled without this #define. As a result,
> > there is a syntax error. Moving #define FASTCALL outside of #ifdef
> > __KERNEL__ helps. I am attaching the diff but I am not sure if it
> > is right.
>
> You don't need to build anything in the kernel directory of linux-nfs_xcxx,
> as all of the code is already in the kernel. Just modify the makefile to
> comment that part out.
>
> > knfsd (or kmountd?) works in a weird way anyway... Sometimes it refuses
> > to satisfy mount request without obvious reason. Locking seems to always
> > fail...
>
> knfsd works differently than unfsd in the mounts it will accept. Locking
> should be working, but you'll need to run statd, IIRC.
>
What happens if I run NFS on 2.1 kernel, without linux-nfs-0.4.21
package?
Or to rephrase the question. Do I need anything from that package if
I'm only 2.1 NFS client? Can utilities made for 2.0 kernels do their
job correctly (rpc.portmap...)?
I'm asking because I was running setup like that all the time, but
performance sometimes really sucked. Solaris 2.6 is on the other side,
as a server.
On the other side, I heard some rumours that Solaris NFS implemen-
tation won't do write behinds on the server. Server won't acknowledge
NFS write operation until blocks are physically on disk. So Sun's
cache data on client side and that speeds things up for them.
Does Linux have client-side caching?
Sorry, I'm not that familiar with NFS code, nor implementation
details.
Later,
-- Posted by Zlatko Calusic E-mail: <Zlatko.Calusic@CARNet.hr> --------------------------------------------------------------------- Oops. My brain just hit a bad sector.- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.altern.org/andrebalsa/doc/lkml-faq.html