On 10/24/2011 09:51 AM, Jeremy Huddleston wrote:
While possibly safe in practice, this doesn't look like the correct fix. It is
possible that this will result in a close(0) if HAVE_MTRR is defined and
mtrr_fd was just never set.
HAVE_MTRR is defined ifasm/mtrr.h exists.
mtrr_fd is set if
in linux_sysfs.c itself. That will allow for better abstraction
down the road as other implementations might want to do something similar.
Sounds good. I'll send another patch shortly.
Thanks,
Nithin.
On Oct 24, 2011, at 10:18, Nithin Sujir wrote:
On 10/24/2011 09:51 AM, Jeremy Huddleston