[RFC] Time for a shared linux-nat.c module?

David S. Miller davem@redhat.com
Tue Apr 23 22:03:00 GMT 2002


   From: Michael Snyder <msnyder@redhat.com>
   Date: Tue, 23 Apr 2002 13:21:22 -0700

   I notice that there are now a bunch of <arch>-linux-tdep.c 
   modules, and that some of them share a bunch of duplicate code.  
   
   I'm wondering if it isn't time to create a linux-nat.c module, 
   shared by all host architectures, into which we can move
   some of this code instead of maintaining separate, duplicate
   versions.
   
The code won't be even Linux specific, I think.

For example the dynamic linker checks are glibc specific.
The commentary refers to it as the "Hurd" resolver, but
in fact it is more of a glibc thing than anything else.

${cpu}_linux_skip_solib_resolver is what I'm specifically talking
about and we can put that as glibc_skip_solib_resolver in glibc-tdep.c
or something like that.

The rest doesn't look all that sharable, as they deal with signal
frame specifics which are quite CPU dependent.



More information about the Gdb mailing list