Re: [PATCH v3 16/19] xen/arm: Introduce a macro to synchronize SError

[Wei Chen <Wei.Chen@arm.com>]

Hi Julien,

On 2017/4/5 15:29, Julien Grall wrote:
>
>
> On 05/04/2017 08:14, Wei Chen wrote:
>>> Because the ASSERT I suggested was wrong, sorry for that. It should have
>>> been:
>>>
>>> ASSERT(!cpus_have_cap(feat) && local_abort_is_enabled());
>>>
>>> This is because we want abort enabled when the "feature" is not present.
>>>

Think more about this assert, I feel confused.

Currently, enable abort or not has not combined with the "feature".
In my testing, the abort is always enabled in the places where we want
to use this macro.

Use "SKIP_SYNCHRONIZE_SERROR_ENTRY_EXIT" for example,
	ASSERT(cpus_have_cap(feat) && local_abort_is_enabled());
will panic by when option == DIVERSE, and 	
	ASSERT(!cpus_have_cap(feat) && local_abort_is_enabled());
will panic by when option != DIVERSE

I can't see any significance about this change.

>>> This series looks good so far, so I would be happy if you send a
>>> follow-up patch to add the ASSERT rather than modifying this patch.
>>>
>>
>> I will send a follow-up patch after this series has been done.
>
> Well, you have to resend this series. So why don't you add the ASSERT in
> the new version?
>
> Cheers,
>


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