Mount and ext3 issues with Busybox 1.2.2.1
Rob Landley
rob at landley.net
Fri Nov 24 03:19:35 UTC 2006
On Thursday 23 November 2006 5:40 pm, Denis Vlasenko wrote:
> On Thursday 23 November 2006 07:56, Jim Gifford wrote:
> > Rob Landley wrote:
> > > On Tuesday 21 November 2006 1:50 am, Jim Gifford wrote:
> > >
> > >> Ok Guys figured out my issue ended up being something really stupid,
> > >> some reason my boot script was failing to create /etc/mtab symlink to
> > >> /proc/mounts. Don't know why but I'm investigating that right now.
> > >>
> > >
> > > Why do you have "CONFIG_FEATURE_MTAB_SUPPORT" switched on?
> > >
> > > Rob
> >
> > I was just using make defconfig, which had that enabled.
>
> Yes, it was, but it is off now in svn.
> Probably my fault...
No, it was on because make defconfig enables all functionality that sane
people might actually want to use.
There are people out there who remove /proc from their kernels for the space
savings (which is actually fairly noticeable in terms of embedded boards
trying to get away with 4 megs of ram), and thus they haven't
got /proc/mounts. (Yeah, it's often the same people who like nommu because
you don't need to waste RAM on page tables.) If you symlink /etc/mtab
to /proc/mounts any attempt to write to mtab (which is all the new code
actually does) is a NOP, so the behavior's the same if the system's set up
right.
Stuff that's off in defconfig essentially never gets tested, because
allyesconfig enables things like SELINUX that nobody anywhere has a build
environment set up for. It doesn't even get compile tested. That's why it
was on by default, so it doesn't bit-rot.
This is a case of "configure what you actually need, understand what you're
actually doing" that shows the project needs more "how to configure the
sucker" documentation somewhere...
Rob
--
"Perfection is reached, not when there is no longer anything to add, but
when there is no longer anything to take away." - Antoine de Saint-Exupery
More information about the busybox
mailing list