| Summary: | mounting on top of a mounted file system | ||
|---|---|---|---|
| Product: | Base System | Reporter: | Randall Hopper <aa8vb> |
| Component: | kern | Assignee: | Alfred Perlstein <alfred> |
| Status: | Closed FIXED | ||
| Severity: | Affects Only Me | CC: | aa8vb |
| Priority: | Normal | ||
| Version: | 3.2-RELEASE | ||
| Hardware: | Any | ||
| OS: | Any | ||
Responsible Changed From-To: freebsd-bugs->billf I have a number of patches from Martin Blapp <blapp@attic.ch> one of which addresses this problem. <<On Sun, 15 Aug 1999 13:34:01 -0700 (PDT), <billf@FreeBSD.ORG> said: > I have a number of patches from Martin Blapp <blapp@attic.ch> > one of which addresses this problem. I disagree that it's a problem. Unexpected to users of other operating systems, yes, but I'd still consider it a feature (and the behavior of those other systems a bug). AIX has similar flexible mount semantics (in fact, more so -- AIX can mount files as well as directories). -GAWollman Solaris 7 does know the option '-O' to allow overlay mounts. I've patched mount and all mount_xxx to recognize and controll overlay mounts. With my fix, mounting on top will be still possible. It's the best for both sides, I think. We don't loose functionality, and it is more userfriendly. The other thing is, if we allow overlay mounts, umount should know to handle them properly. I've patched umount too, to always unmount overlay mounts in right order. Martin Responsible Changed From-To: billf->alfred Alfred has been working with Martin Blapp who has patches to address this. State Changed From-To: open->closed Martin says that the "overlay" mount stuff fixes this. |
I had: > df | grep 32_share /dev/wd1s3e 2031953 1353017 516380 72% /32_share Then I mistakenly did a: > mount /dev/wd0s3e /32_share ^ And it didn't generate an error (!?). I then had: > df | grep 32_share /dev/wd1s3e 2031953 1353017 516380 72% /32_share /dev/wd0s3e 2032623 1169007 701007 63% /32_share Odd things started happening (vi failed, missing directories, etc.), so I rebooted. All is normal now, and nothing lost AFAIK, but that sure was strange. This looks like a bug. Fix: Unknown. I think that mount should issue an error when this is attempted. How-To-Repeat: > mount /dev/wd1s1 /abc > mount /dev/wd1s2 /abc In this case, both were FreeBSD UFSs.