Re: glibc 2.1.92 tst-getdate failure on linuxppc

[Geoff Keating <geoffk at cygnus dot com>]

> From: "Hendricks, Kevin" <khendricks@ivey.uwo.ca>
> Cc: geoffk@cygnus.com, Franz.Sirl-kernel@lauterbach.com, howarth@fuse.net
> Date: Mon, 28 Aug 2000 11:09:38 -0400

> But do we ever really want to be fiddling with bits inside signal handler
> stack frames?  Given the race problems, problems with overwriting a user
> signal handler even temporarily, and the dangerous precedent it sets in
> general, I really don't think this is a good solution.

Nor do I, otherwise I'd have done this long ago.

> Why don't we simply explain the issues to Paul and Cort and get them to add
> the proper syscall to do this from inside the kernel where the msr can be
> played with in supervisor mode?
> 
> Are these check failures really that important to ppc that we can't let it
> be broken for awhile longer until the kernel guys do something about it.

I don't believe it will ever be fixed if it is just left alone.  This
has been outstanding for more than three years.

-- 
- Geoffrey Keating <geoffk@cygnus.com>