Re: [PATCH 2/2] add f_flags to struct statfs(64)

[Christoph Hellwig <hch@lst.de>]

On Wed, Jul 07, 2010 at 12:30:50PM -0700, Linus Torvalds wrote:
> On Wed, Jul 7, 2010 at 11:50 AM, Nick Piggin <npiggin@suse.de> wrote:
> >
> > Well we do want to be simple I think, but it's not *totally* just
> > for dbench performance :) Actually dbench is supposed to be an fs
> > syscall reply of samba running a file server workload. For one
> > reason or another, it calls statvfs a lot.
> 
> It seems to call statvfs() mainly because it wants the disk free block
> information, afaik. I can't find anything that cares about f_flag.

People calls statvfs instead of statfs because it's the version that
got standardized in Posix.

And with the glibc emulation of the flags damage is done now, and given
how easy it is to fix (see my patches) there is no reason not to fill
out the flags.  We just need to figure out how.  I see three theoretical
variants:

 (1) add new system calls
 (2) drop the ST_VALID flag and only look at the flags field for
     2.6.36+ kernels.
 (3) keep the ST_VALID flag in my patches and only look at the flags
     field in glibc for 2.6 kernels

(1) is pretty ugly, and Linus already vetoed it, (2) is okayish, but
prevents distros from backporting this easy speedup in a useful way,
so I'd prefer variant (3).  If Uli is okay with that we should put
my second version of the patches into the vfs tree and pull them
for 2.6.36.  If version (2) is preferred I can update the patches
again.