Re: bizarre behavior reading files incorrectly with 4.99.72: FIXED

[Thor Lancelot Simon <tls%rek.tjls.com@localhost>]

On Sat, Aug 09, 2008 at 08:38:15AM -0400, Greg Troxel wrote:
> 
> It turns out the problem was that /tmp (tmpfs) was 512K, not 512M.  So
> there's apparently a bug in patch that fails to error out when /tmp is
> too small.

I've encountered this, too.  The problem is that tmpfs takes its size
(as a mount argument) in bytes, whereas the old MFS took it in blocks.
Naively converting an fstab entry causes this highly unobvious lossage.

I did agitate for changing the tmpfs behavior (or at least modifying
mount_tmpfs to require a scaling factor on the size argument, e.g.
"512k" or "512m") but nobody else seemed to like that idea much.

-- 
Thor Lancelot Simon                                        
tls%rek.tjls.com@localhost
 "My guess is that the minimal training typically provided would only
 have given the party in question multiple new and elaborate ways to do
 something incomprehensibly stupid and dangerous."      -Rich Goldstone