mount bug [was: Re: tmpfs]

Chey chey.smith at gmail.com
Sun Aug 20 21:50:20 UTC 2006


On 8/20/06, Rob Landley <rob at landley.net> wrote:
> On Sunday 20 August 2006 9:50 am, Bernhard Fischer wrote:
> > On Sun, Aug 20, 2006 at 09:38:54AM -0400, Chey wrote:
> > >Is anyone else having this problem?
>
> I can't reproduce it.
>
> > This is likely a generic bug in the new mount and should go away if you
> > revert to the mount from version 1.0+fixes-but-no-rewrites.
>
> Bernhard: are you implying that you can reproduce his problem?  Because I
> can't.  And I have no idea what version you're referring to here.
>
> > Rob may already have a fix for this pending, though.
>
> I'll happily fix a bug I can reproduce.
>
> > HTH,
> > >
> > >
> > >On 8/18/06, Chey <chey.smith at gmail.com> wrote:
> > >> Hello,
> > >>    I am using busybox 1.2.1 and I'm having trouble mounting tmpfs with
> > >> size constraints:
> > >>
> > >> root at generic:/# mount -t tmpfs -o size=4264k none /tmp
> > >>
> > >> root at generic:/# mount
> > >> ...(truncated)
> > >> none on /tmp type tmpfs (rw)
>
> Note that the kernel doesn't spit back options for tmpfs from /proc/mounts, so
> it won't tell you about the size thing.  This is a kernel limitation.  (It
> doesn't tell you what a loopback mount is associated with, either.)

I didn't realize this. I was assuming that since my main system
(Debian) mount displays "the size thing" I could expect the same
results from BB mount.

>
> > >> root at generic:/# dd if=/dev/zero of=/tmp/foo bs=1000k count=5
> > >> 5+0 records in
> > >> 5+0 records out
>
> I did a busybox mount like your example and then did a cp of a 7 megabyte file
> into the directory with current svn, and it worked.  I grabbed 1.2.1 and
> tried it, and that worked too.  (By "worked" I mean I got no space left on
> device, with 4264k used.)  I'll try it again with dd:
>
> root at driftwood:/home/landley/busybox/release/busybox-1.2.1# ./busybox mount -t
> tmpfs -o size=4264k none testsuite
> root at driftwood:/home/landley/busybox/release/busybox-1.2.1# dd if=/dev/zero
> of=testsuite/walrus bs=1000k count=5
> dd: writing `testsuite/walrus': No space left on device
> 5+0 records in
> 4+0 records out
> 4354048 bytes (4.4 MB) copied, 0.031353 seconds, 139 MB/s
> root at driftwood:/home/landley/busybox/release/busybox-1.2.1#
>
> I cannot reproduce the problem you're seeing.
>
> Rob
> --
> Never bet against the cheap plastic solution.
>

I didn't fully test my theory after Jasons suggestion. I just looked
at the output from mount and didn't see the size in the output. After
fully testing I do get back the error and it appears to be working. So
the problem was the fact that I had CONFIG_TINY_SHMEM enabled in my
kernel.


Thanks everyone. I really appreciate your help!

Chey



More information about the busybox mailing list