Re: [patch 13/13] lseek speedup

[Andrew Morton]

Anton Altaparmakov wrote:
> 
> At 06:31 17/07/02, Andrew Morton wrote:
> >This is a fairly dopey patch to fix up the i_sem contention in lseek.
> >Better ideas are welcome, but I'm offline until Monday so don't think
> >I'm ignoring them...
> 
> I am afraid I don't have any better ideas but I don't think your patch is
> safe. )-:

It wasn't a very good idea in the first place.  Forgot to take
the new lock over in the updaters of f_pos.

And it's attempting to cater for a buggy application on a 32-bit
machine, SMP, where the fd is shared.  It's hard to justify
putting any locking in lseek for that.  (Then again, people
should use pread() more..)

Except for readdir().  Now, why are we taking i_sem for lseek/readdir
exclusion and not a per-file lock?

-