lkml.org 
[lkml]   [2013]   [Dec]   [10]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: process 'stuck' at exit.
On Tue, 10 Dec 2013, Linus Torvalds wrote:

> On Tue, Dec 10, 2013 at 2:57 PM, Thomas Gleixner <tglx@linutronix.de> wrote:
> >
> > But how does the access_ok() move do anything helpful here?
>
> Just making it all more obvious.
>
> > We really need it for the fastpath !fshared case, but for the fshared
> > case you actively break working code, because you force a VERIFY_WRITE
> > check into it. The VERIFY_WRITE is necessary for !fshared, because
> > there is no way that one thread can map the futex RO and the other RW,
> > right?
>
> Nobody actually uses that argument any more (it goes back to the old
> i386 "let's manually verify that we have write permissions, because
> the CPU doesn't do it for us in the trap handling"), and it should
> probably be removed.

Fair enough.

> But you're right that it's at least misleading. I'd love to remove it
> entirely, because it's not even syntax-checked, and it's confusing.
> But that would be a humongous patch.

Well, we should ask Julia for a coccinelle patch to limit the
wreckage. :)

Seriously, if that VERIFY_WRITE is completely useless we really want
to get rid of it.

> So these days, "access_ok()" literally just checks that the address is
> in the user address space range. And that would seem to always be
> appropriate for futexes, so why not just do it in the generic code?

Agreed, but as long as the VERIFY_WRITE argument is there it needs at
least a big fat comment :)

Thanks,

tglx


\
 
 \ /
  Last update: 2013-12-11 00:41    [W:0.324 / U:0.040 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site