Re: [RFC] Imminent death of /proc/locks predicted; film at 11

[Matthew Wilcox]

On Sat, Sep 30, 2000 at 08:23:37PM -0500, Peter Samuelson wrote:
> 
> [Matthew Wilcox]
> > if fcntl took a 4th argument specifying the length of the buffer, i'd
> > recommend a F_GETLKS fcntl.  a horrid work around for this would be
> > that the first 4 bytes of the buffer pointed to by the third argument
> > of the fcntl is the length of the buffer.
> 
> Ewwww!  You're right, it's horrid.  Anyway, I think this one can be
> solved in userspace -- as long as you don't need atomicity.  Untested
> code with no error checking:

thanks for snipping the part of my email where i explain this won't work.
examples:

process 1 locks bytes 1 to 7 nonexclusively
process 2 locks bytes 2 to 5 nonexclusively

you now can't see the second lock.

and there's no way of seeing the blocked lock.  This needs kernel support
some how.

-- 
Revolutions do not require corporate support.