Re: [PATCH] fix floppy.c to store correct ro/rw status in underlying gendisk

[Jon Masters <jonmasters@gmail.com>]

On 10/31/05, Rob Landley <rob@landley.net> wrote:
> On Monday 31 October 2005 05:57, Evgeny Stambulchik wrote:
> > Jon Masters wrote:
> > > Let me know if this fixes it for you - should bomb out now if you try.
> > > The error isn't the cleanest (blame mount), but it does fail.
> >
> > This works fine, thanks! For what it worth, though, mount -o remount,rw
> > says remounting read-only yet still returns success. (Opposite to
> > busybox, which now says "Permission denied" - rather misleading, but at
> > least it fails).
>
> That sounds like the string translation of EPERM returned by libc's
> strerror().  (At busybox we're frugal bastards; we don't include text
> messages when we can get the C library to provide them for us. :)

Indeed. Reminds me that I should clean up and send a form parser I
wrote for busybox to handle multi-part mime form posts in its
webserver while I'm at it. That's something it could do with even if
it's being frugal.

> But yeah, we're sticklers for correct behavior, and only attempt to remount
> readonly if we get EACCES or EROFS, not _just_ because we attempted a
> read/write mount and it failed.  (And yes, I personally tested this corner
> case.  We haven't started on an automated regression test script for mount
> yet because running it would require root access, but it's on the todo list
> as we upgrade the test suite in our Copious Free Time...)

You looked at running this inside a qemu environment (scratchbox), huh?

Jon.