Re: pre7-6 oops in sys_mount

From: Manfred Spraul (manfreds@colorfullife.com)
Date: Sat May 06 2000 - 15:47:06 EDT

  • Next message: Linus Torvalds: "Re: [PATCH] address_space_operations unification"

    thunder7@xs4all.nl wrote:
    >
    > On Sat, May 06, 2000 at 08:02:54PM +0200, Manfred Spraul wrote:
    > > Alexander Viro wrote:
    > > >
    > > > Total bullshit. Please, use the correct System.map - for one thing, fuser
    > > > doesn't do mounting, for another - ext2_free_inode() _definitely_ doesn't
    > > > call sys_mount().
    > > >
    > > The code sequence and the registers match my crash in __d_path: vfsmnt
    > > is NULL, called by proc_pid_read_maps on a shmfs entry.
    > >
    > proc_pid_read_maps? I'm no kernel expert, but perhaps this is the time
    > to mention that I had top running, from procps-2.0.2, with both crashes.
    >
    You are the third one who encountered that bug: as soon as you read from

            /proc/<pid>/maps

    on a process that uses a SysV shared memory segment, you get an oops.
    There is even a comment in ipc/shm.c, search for MOUNT_REWRITE. I assume
    it's Al comment, and he'll fix it as soon as he has fixed the more
    important bugs. 

    --
	Manfred
P.S.: how did you add the symbols to the oops? ksymoops or klogd? IIRC
there are a few bugs in klogd, it's better to save the oops from dmesg,
and parse it through a recent ksymoops version.

    -
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/

    This archive was generated by hypermail 2b29 : Sat May 06 2000 - 15:49:01 EDT