On 2/10/2016 7:11 PM, Jan Beulich wrote:
On 10.02.16 at 18:04, <cz...@bitdefender.com> wrote:
On 2/10/2016 6:18 PM, Jan Beulich wrote:
On 10.02.16 at 16:50, <cz...@bitdefender.com> wrote:
--- a/xen/include/asm-x86/hvm/event.h
+++ b/xen/include/asm-x86/hvm/event.h
@@ -17,6 +17,12 @@
#ifndef __ASM_X86_HVM_EVENT_H__
#define __ASM_X86_HVM_EVENT_H__
+enum hvm_event_breakpoint_type
+{
+ HVM_EVENT_SOFTWARE_BREAKPOINT,
+ HVM_EVENT_SINGLESTEP_BREAKPOINT,
+};
I don't see what good it does to put existing constants into an
enum.
As Andrew pointed out, an enum was requested in v1 instead of the
single_step param.
One could use the already existing VM_EVENT_REASON_* constants, but
conceptually this
function only involves a subset of those (i.e. *breakpoint vm-events*).
Re-using existing constants would seem fine to me.
I only now realize that I've made a mistake while looking at the
above - the capitals made it implicitly "obvious" to me that they're
on the right side of an assignment. Please use capitals only for
#define-d constants, not enumerated ones.
Jan
Also, why this bias towards #define vs enums? I usually always prefer
the latter,
they make the code more readable, e.g. it's clearer what can or cannot
be passed
to a corresponding function parameter. And from an IDE user's
point-of-view, I don't
understand why I have to read a comment and search for a list of identifiers
rather than just clicking on a type name and getting there in a flash.
They also make debugging easier, etc.. :).
Corneliu.
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
