[Top] [All Lists]

Re: mount() problem triggered by busybox mount

To: Jay Carlson <>
Subject: Re: mount() problem triggered by busybox mount
From: Erik Andersen <>
Date: Mon, 24 Apr 2000 11:53:22 -0600
In-reply-to: <1f9401bfac99$b3928140$0a00000a@decoy>; from on Sat, Apr 22, 2000 at 04:31:11PM -0400
References: <1f9401bfac99$b3928140$0a00000a@decoy>
User-agent: Mutt/1.0.1i
On Sat Apr 22, 2000 at 04:31:11PM -0400, Jay Carlson wrote:
> I don't know whether you think this is a kernel bug or a busybox bug.
> mount() on Linux/MIPS doesn't like stack-allocated arguments, bombing with
> EFAULT.  I suspect this is the fault of fs/super.c:copy_mount_options, but
> I'm not sure.
> This is fixable in busybox by adding strdup()s to specialfile, dir,
> filesystemtype, string_flags in mount.c:do_mount before the status=mount()
> line.

This definately sounds like a kernel bug.  I know that at least for x86,
powerpc, and alpha the current code works fine.  Lemme go ask the SH guy here
to see if he has noticed and mount problems...  Nope, but the SH port passes
everything via registers.  Oh well.  Anyway, I suspect that you are correct in
suspecting copy_mount_options.  Let me know if you get it working,


Erik B. Andersen   email:
--This message was written using 73% post-consumer electrons--

<Prev in Thread] Current Thread [Next in Thread>