[lkml]   [2011]   [May]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: [15/24] tmpfs: fix spurious ENOSPC when racing with unswap
On Fri, May 20, 2011 at 10:49:59AM -0700, Hugh Dickins wrote:
> On Thu, 19 May 2011, Greg KH wrote:
> > 2.6.33-longterm review patch. If anyone has any objections, please let us know.
> We've got in a muddle on this one: too many tmpfs races and too many trees!

Yeah, it's messy :(

> So far, so good: indeed let's have the spurious ENOSPC fix in
> 33-longterm.
> But here this 15/24 patch veers off into a quite different patch,
> for "tmpfs: fix race between umount and writepage" (46/71 in the
> 38-stable series). I've appended the actual ENOSPC backport at the end.
> Yes, let's have this writepage fix in 33-longterm too (the 38-stable
> patch should apply), but it does need "tmpfs: fix race between swapoff
> and writepage" (47/71 in the 38-stable series) on top to correct it,
> please add that in too.
> For differing reasons, none of these races is as likely in 2.6.33
> as in 2.6.38, but good to include the fixes anyway; whereas 2.6.32
> gets more complicated for some of them, so I haven't bothered there.
> (I think I'm reading the mails right, but of course made a fool of
> myself in the past, because of how gmail "rationalized" my view of
> them: I hope this won't be another such case.)

Ok, as this really isn't a big deal for .33, how about I just drop all
of the tmpfs patches for .33-longterm and we call it a day :)

It's easier that way for me by far.

If you think these really should all go to .33, can you send me the
individual patches so I know exactly what ones, and in what order, to


greg k-h

 \ /
  Last update: 2011-05-21 23:53    [W:0.049 / U:6.372 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site