RE: [Valgrind-users] BerkeleyDB gets bogus errno in Valgrind

[Troels W. H. <tr...@th...>]

I just dicovered that RH has updated the glibc errata for RH8 to version
2.3.2-4.80.6.
https://rhn.redhat.com/errata/RHSA-2003-089.html

There is also a 2.3.2-27.9 version for RH9:
https://rhn.redhat.com/errata/RHBA-2003-136.html

The changelog for the RH8 package contains the entry below, and Valgrind
1.9.6 now works unpatched. The bugzilla entry for that change is at=20
http://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=3D86437

"* Sun Apr 06 2003 Jakub Jelinek <ja...@re...> 2.3.2-4.80.5
- make sure all calls to __errno_location() and __h_errno_location()
  jump through .plt to make older wine happy"

This is probably no help to you Julian (since I guess this is a RH
specific compability patch that won't be included in mainline glibc),
but it certainly makes me happy.

Troels

> -----Original Message-----
> From: val...@li...=20
> [mailto:val...@li...] On Behalf=20
> Of Julian Seward
> Sent: 15. mai 2003 01:13
> To: Tom Hughes; val...@li...
> Subject: Re: [Valgrind-users] BerkeleyDB gets bogus errno in Valgrind
>=20
>=20
>=20
> > This is why wine now deliberately writes a jmp instruction over the
> > first instruction of glibc's internal errno address getting routine
> > so that it guarantee to get control at that point...
>=20
> Well, I've peered at glibc-2.3.2 internals regarding errno=20
> and concluded that it's a *$%*=A3@ nightmare.