|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
Re: mean vfs bug in *bsd
Brian Feldman (green
FELDMAN.DYN.ML.ORG)Sun, 31 May 1998 00:44:50 -0400
- Messages sorted by: [ date ][ thread ][ subject ][ author ]
- Next message: Jim Dennis: "Re: Linux auto idle logout & vlock possible security problem"
- Previous message: aleph1
NATIONWIDE.NET: "Patch to prevent setuid bash shells"
- Maybe in reply to: Arend-Jan Wijtzes: "mean vfs bug in *bsd"
- Next in thread: matthew green: "Re: mean vfs bug in *bsd"
> As we upgraded to freeBSD 2.2.6-STABLE the following bug came to my attention: > when linking a device like /dev/null to a file on a NFS mounted fs > the machien freeses. A hard link links via a true link, as follows: a symlink is just a specially marked file with the path and name of another file which it refers to. A _hard_ link creates a new fs entry and sets its data inode number to point to that of the other file, in which case it becomes the file. In this way a hard link does NOT traverse different file systems/mount points whatsoever. I'm pretty certain your /dev is locally mounted, not by NFS, therefore I tend to believe that you shouldn't even be attempting a hard link. If in actuality, this does occur linking two files, one a device, both on the same NFS filesystem, then this is significant, otherwise it just means that ln(1) needs to have some better error-checking so it doesn't make hard links span drives. -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= my->name = "Brian Feldman"; my->email[0] = "brianfeldmanhotmail.com"; my->email[1] = "green
feldman.dyn.ml.org"; -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
- Next message: Jim Dennis: "Re: Linux auto idle logout & vlock possible security problem"
- Previous message: aleph1
NATIONWIDE.NET: "Patch to prevent setuid bash shells"
- Maybe in reply to: Arend-Jan Wijtzes: "mean vfs bug in *bsd"
- Next in thread: matthew green: "Re: mean vfs bug in *bsd"