Re: [PATCH] x86/traps: fix an off-by-one error

[Jan Beulich <jbeulich@xxxxxxxx>]

On 05.05.2020 16:01, Hongyan Xia wrote:
> On Tue, 2020-05-05 at 15:38 +0200, Jan Beulich wrote:
>> On 05.05.2020 13:06, Hongyan Xia wrote:
>>> @@ -358,7 +359,7 @@ static void show_guest_stack(struct vcpu *v,
>>> const struct cpu_user_regs *regs)
>>>      if ( mask == PAGE_SIZE )
>>>      {
>>>          BUILD_BUG_ON(PAGE_SIZE == STACK_SIZE);
>>> -        unmap_domain_page(stack);
>>> +        unmap_domain_page(stack_page);
>>>      }
>>
>> With this I think you want to change the whole construct here to
>>
>>     if ( stack_page )
>>         unmap_domain_page(stack_page);
>>
>> i.e. with the then no longer relevant BUILD_BUG_ON() also dropped.
> 
> I wonder if such a construct is better with UNMAP_DOMAIN_PAGE(), since
> it deals with NULL and will nullify it to prevent misuse.

In the case here I think I agree. For the future may I ask that
you wait with sending a new version until the discussion on the
previous one has settled?

Jan