Re: [PATCH] A little wad of patches for 2.3.10

david parsons (o.r.c@p.e.l.l.p.o.r.t.l.a.n.d.o.r.u.s)
29 Jul 1999 01:02:59 -0700

In article <linux.kernel.199907290625.CAA05743@jupiter.cs.uml.edu>,
Albert D. Cahalan <acahalan@cs.uml.edu> wrote:
>
>david parsons (o.r.c@p.e.l.l.p.o.r.t.l.a.n.d.o.r.u.s)
>> Matthew Wilcox <Matthew.Wilcox@genedata.com> wrote:
>>> On Wed, Jul 21, 1999 at 11:53:46PM -0700, david parsons wrote:
>
>>>> * Puts back old-style fcntl locking, so that old a.out gdbm libraries
>>>> won't explode and die.
>>>
>>> I thought this was taken out because it was a broken design and could
>>> lead to data corruption?
>
>Yes.
>
>> Well, I'm not so sure about that, unless doing new-style flocking
>> will also cause data corruption. Basically, all the patch with
>> the locks does is to map SHLCK -> RDLCK and EXLCK -> WRLCK.
>
>The software you have was written to use one API.
>Old kernels did not provide that API.
>Old libc used a similar API as a substitute.
>Lock inheritance did not work as apps expected.
>Stuff broke. People lost email and other data.

Can you point me at some articles that provide test cases for this? I
may be missing something, but RDLCK is RDLCK, and it shouldn't matter if
that userland passes it in via a compatability interface.

Anecdotally, I've never encountered any problems like you describe with
the old libc4 locking scheme; it doesn't mean that they don't exist,
of course, but it does mean that I'm less likely take IT'S BROKEN at
face value.

____
david parsons \bi/ call me skeptical.
\/

>
>Now we have the desired API implemented correctly.
>Anything using the old API is suspect.
>Some old Linux-specific apps would work fine.
>Apps that really needed what was asked for...
>
>-
>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/

-
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/