Re: Problem with VMware 2.0.4 and glibc 2.2.5

[Jeroen Dekkers <jeroen at dekkers dot cx>]

On Wed, Apr 10, 2002 at 06:20:10PM +0200, Thorsten Kukuk wrote:
> On Wed, Apr 10, H . J . Lu wrote:
> 
> > On Wed, Apr 10, 2002 at 02:55:18PM +0200, Jeroen Dekkers wrote:
> > > On Tue, Apr 09, 2002 at 07:32:20PM -0700, H . J . Lu wrote:
> > > > I got a strange problem with glibc 2.2.5-30 from Red Hat and VMware
> > > > 2.0.4. I got
> > > > 
> > > > VMware Workstation PANIC:
> > > > 
> > > > AIO: NOT_IMPLEMENTED F(566): 1081.
> > > > 
> > > > and VMware quitted. glibc 2.2.4-19 works fine. Has anyone seen this?
> > > 
> > > I saw on it on debian-devel, it's a bug in VMWare because they relied
> > > on the bug in nice(). VMWare provided patches, but those are only for
> > > version 3 and because VMWare is non-free software, you can't patch it
> > > yourself. IMHO some people need to get plex86 in an usable state,
> > > relying on crappy non-free software is always bad. And for me a nice
> > > second-hand machine with the same price as a VMWare license works
> > > fine. :)
> > > 
> > 
> > It looks like glibc 2.2.5 changed the ABI for nice. Shouldn't we give
> > it a new version?
> 
> With such a change an the fact, that the old, wrong behaviour is
> descriped in nearly every Linux documentation which contains a
> description for nice, Yes, we should give it a new version.

What do we have to do with Linux documentation? We are glibc and
according to the glibc documentation nice() should return the
priority. Manual pages aren't official glibc documentation. If I write
some documentation saying that nice() should return 72925, do we have
to make another version for that? 
Oh, the manual page on my system says:

NOTES
Note  that  the routine is documented in SUSv2 to return the new nice
value, while the Linux syscall and (g)libc (earlier than glibc 2.2.4)
routines return 0 on success.  The new nice value can be found using
getpriority(2).  Note that an implementation in which  nice  returns
the  new nice value can legitimately return -1.  To reliably detect an
error, set errno to 0 before the call, and check its value when nice
returns -1. 

Other than that, on GNU/Hurd the return value has always been
correct.

Jeroen Dekkers
-- 
Jabber supporter - http://www.jabber.org Jabber ID: jdekkers@jabber.org
Debian GNU supporter - http://www.debian.org http://www.gnu.org
IRC: jeroen@openprojects

Attachment: msg00051/pgp00000.pgp
Description: PGP signature