Re: [Xen-devel] [PATCH RFC 1/7] xen: Relocate mem_access and mem_event into common.

[Tamas K Lengyel <tamas.lengyel@xxxxxxxxxxxx>]

On Tue, Aug 26, 2014 at 2:42 PM, Jan Beulich <JBeulich@xxxxxxxx> wrote:

>>> On 26.08.14 at 12:52, <tamas.lengyel@xxxxxxxxxxxx> wrote:
> On Mon, Aug 25, 2014 at 7:19 PM, Andres Lagar Cavilla <
> andres@xxxxxxxxxxxxxxxx> wrote:
>> On Fri, Aug 22, 2014 at 2:30 AM, Tamas K Lengyel <

>>> -        req->flags |= MEM_EVENT_FLAG_FOREIGN;
>>> -        ASSERT( !(req->flags & MEM_EVENT_FLAG_VCPU_PAUSED) );
>>>
>> Take the opportunity to downgrade the aggressiveness of this at some point
>> in this series. (I'd prefer to keep code motion patches as purely code
>> motion).
>>
>> A faulty tool stack can brick a debug hypervisor. Unpleasant while
>> dev/test.
>>
>
> I'm a little bit hazy in what situations this could arise and what it is
> trying to protect against. I could wrap the condition into an unlikely()
> and have a debug message printed rather than bricking the VMM with ASSERT()
> and/or enable the ASSERT only when we are building with debug=y.

ASSERT()s are enabled only when debug=y (i.e. NDEBUG not defined
at the C level).

Jan

Ah, OK, I'll just switch it to a debug print in a #ifndef NDEBUG block then. Thanks for clarifying that ASSERT behavior!

Tamas

_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
http://lists.xen.org/xen-devel