Re: CVS commit: src/lib/libc/arch/alpha/gen

2013-03-13 Thread Martin Husemann
On Wed, Mar 13, 2013 at 08:44:03AM -0400, Christos Zoulas wrote: > But most of the structure is explicitly initialized. It is probably better > to explicitly set the missing members that calling memset(). I have verified that only argument and scratch registers remain uninitialized, and so removed

Re: CVS commit: src/lib/libc/arch/alpha/gen

2013-03-13 Thread Christos Zoulas
On Mar 13, 8:33am, mar...@duskware.de (Martin Husemann) wrote: -- Subject: Re: CVS commit: src/lib/libc/arch/alpha/gen | I have a bad feeling setting random values to registers, and we will | do a system call right afterwards, so I think it is no big deal performance | wise. We can carfully null

Re: CVS commit: src/lib/libc/arch/alpha/gen

2013-03-13 Thread Martin Husemann
On Tue, Mar 12, 2013 at 09:20:44PM +, Christos Zoulas wrote: > In article <20130312193820.a36b017...@cvs.netbsd.org>, > Martin Husemann wrote: > > is the memset() in longjmp desired? It will just slow down things. I have a bad feeling setting random values to registers, and we will do a syst