Messages in this thread | | | Date | Fri, 9 Jul 1999 00:31:35 -0400 (EDT) | From | Chris Wing <> | Subject | Re: ANNOUNCE: 32-bit UID patch for Linux 2.0, 2.2 |
| |
Jordan:
Thanks for writing back.
> The quota code doesn't appear to be updated. I know this will break backwards > compatibility and the quota code in general needs to be redone, but it should be > done or at least be put on the to do list.
I've already corresponded with Jan Kara <jack@atrey.karlin.mff.cuni.cz>, who has been working on the quota code, it turns out that quota works properly more or less without any changes-- one thing, though, is that due to the structure of the quota file, you can't set quotas for the full 32-bit UID range (it works only up to about 26 bits, or 64 million users). Beyond that, you exceed the maximum filesize (2GB) for the quota file on 32-bit systems. I bet that on systems that support 64-bit file sizes (Alpha and Ultrasparc for now) there wouldn't be any problems.
In any case, the user quota utilities bog down to useless speed when dealing with a quota file that big. (the kernel is fine)
This doesn't break backwards compatibility, in any case, but other changes to the quota system in the future likely will.
> The /proc filesystem needs to have a few things updated as well, things like > /proc/<pid>/status list the uid & gid of the owning process, otherwise ps is > going to break (though it will need updating anyway).
From the code, this shouldn't be a problem:
sprintf( ... "Uid:\t%d\t%d\t%d\t%d\n", "Gid:\t%d\t%d\t%d\t%d\n", ... p->uid, p->euid, p->suid, p->fsuid, .... etc)
Likewise, the ps program seems to be using type 'int' to parse /proc, as it doesn't have a problem displaying the information.
Thanks,
Chris Wing wingc@engin.umich.edu
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |