Pádraig Brady <[email protected]> wrote:
> On 01/28/2013 01:14 PM, Pádraig Brady wrote:
> >On 12/14/2012 04:17 AM, Jim Meyering wrote:
> >>If we're omitting "devfs", then should "devtmpfs" also be omitted?
> >
> >I don't think "devtmpfs" should be marked as dummy
> >as there is storage associated with it.
> >I.E. you can write normal files to /dev and space will be consumed.
> >
> >Note df already excludes most dummy file systems
> >by default by checking the size like:
> >
> > if (fsu.fsu_blocks == 0 && !show_all_fs && !show_listed_fs)
> > return;
Thanks for that hint[1].
> I'll apply the attached in a while unless there are objections.
No objection here.
[1] - However, I still hit the stat() stall from get_fs_usage() on
my system :<
Checking my eglibc (2.13-37 on Debian testing), it is ST_VALID-capable
(as glibc 2.13 is). However, I think it was built by the Debian
maintainer on an older kernel so it's not using ST_VALID (I'm on a
lightly-patched 3.7.4 kernel).
So I think the fsusage gnulib module needs a stricter check for
statvfs() + ST_VALID use along with the 2.6.36 kernel check...
Rebuilding eglibc now to confirm...
