This is the mail archive of the
gdb@sourceware.org
mailing list for the GDB project.
Re: regression test coverage for sim/common ?
- From: Laurent GUERBY <laurent at guerby dot net>
- To: Joern Rennecke <amylaar at spamcop dot net>, Joel Sherrill <joel dot sherrill at oarcorp dot com>
- Cc: gdb at sourceware dot org
- Date: Mon, 09 Mar 2009 23:41:10 +0100
- Subject: Re: regression test coverage for sim/common ?
- References: <20090309183049.k1gtv5m4744k04wo-nzlynne@webmail.spamcop.net>
On Mon, 2009-03-09 at 18:30 -0400, Joern Rennecke wrote:
> I have a patch that speeds up simulator startup for simulators that
> use code in sim/common to create a new simulator instance.
>
> Instead of xmalloc / memset I use xcalloc to obtain zeroed memory;
> on operating systems like a recent glibc/Linux system, this allows
> the pages to me faultyed in as if they had been mmapped.
>
> What would be a suitable simulator target for a regression test?
I'm cc-ing Joel Sherrill, he has setup GCC tests for lots *-rtems
targets for GCC and those use heavily GDB simulators. His testing
process builds binutils, GDB and other stuff completely from source (I
believe his test framework is even commited and documented somewhere, in
addition to running on the compile farm :).
Sincerely,
Laurent