Re: [RFC][PATCH] /proc/pid/maps doesn't match "ipcs -m" shmid

["Serge E. Hallyn" <serge@hallyn.com>]

Quoting Serge E. Hallyn (serge@hallyn.com):
> Quoting Badari Pulavarty (pbadari@us.ibm.com):
> > On Thu, 2007-06-07 at 15:37 -0500, Serge E. Hallyn wrote:
> > > Quoting Badari Pulavarty (pbadari@us.ibm.com):
> > > > On Thu, 2007-06-07 at 12:48 -0700, Andrew Morton wrote:
> > > > > On Thu, 07 Jun 2007 10:06:37 -0700
> > > > > Badari Pulavarty <pbadari@us.ibm.com> wrote:
> > > > > 
> > > > > > On Thu, 2007-06-07 at 12:43 -0400, Albert Cahalan wrote:
> > > > > > > On 6/7/07, Badari Pulavarty <pbadari@us.ibm.com> wrote:
> > > > > > > 
> > > > > > > > BTW, I agree with Eric that its would be nice to use shmid as part
> > > > > > > > of name instead of forcing to be as inode number. It should be
> > > > > > > > possible for pmap to workout shmid from "key" or name. Isn't it ?
> > > > > > > 
> > > > > > > It is not at all nice.
> > > > > > > 
> > > > > > > 1. it's incompatible ABI breakage
> > > > > > > 2. where will you put the key then, in the inode? :-)
> > > > > > 
> > > > > > Nope. Currently "key" is part of the name (but its not unique).
> > > > > > 
> > > > > > > 
> > > > > > > Changing to "SYSVID%d" is no good either. Look, people
> > > > > > > are ***parsing*** this stuff in /proc. The /proc filesystem
> > > > > > > is not some random sandbox to be playing in.
> > > > > > > 
> > > > > > > Before you go messing with it, note that the device number
> > > > > > > also matters. (it's per-boot dynamic, but that's OK)
> > > > > > > That's how one knows that /SYSV00000000 is not just
> > > > > > > a regular file; sadly these didn't get a non-/ prefix.
> > > > > > > (and no you can't fix that now; it's way too late)
> > > > > > > 
> > > > > > > Next time you feel like breaking an ABI, mind putting
> > > > > > > "LET'S BREAK AN ABI!" in the subject of your email?
> > > > > > 
> > > > > > I am not breaking ABI. Its already broken in the current
> > > > > > mainline. I am trying to fix it by putting back the ino#
> > > > > > as shmid. Eric had a suggestion that, instead of depending
> > > > > > on the inode# to be shmid, we could embed shmid into name
> > > > > > (instead of "key" which is currently not unique).
> > > > > > 
> > > > > > > BTW, I suspect this kind of thing also breaks:
> > > > > > > a. fuser, lsof, and other resource usage display tools
> > > > > > > b. various obscure emulators (similar to valgrind)
> > > > > > 
> > > > > > If you strongly feel that "old" behaviour needs to be retained, 
> > > > > 
> > > > > yup, we should put it back.  The change was, afaik, accidental.
> > > > > 
> > > > > > here is the patch I originally suggested.
> > > > > 
> > > > > Confused.  Will this one-liner fix all the userspace breakage to which
> > > > > Albert refers?
> > > > 
> > > > Yes. Albert, please correct me if I am wrong.
> > > 
> > > It will, but could lead to two different inodes with the same i_ino,
> > > right?
> > 
> > Only if we generate same ID in two different namespaces. Is it currently
> > possible ? 
> 
> Should be nothing stopping it.

(just to be more certain, a quick test showed I can get id 0 for
different keys, and different ids for the same key 0xff, in different
ipc namespaces)

-serge