Re: [PATCH v2 2/2] kexec: Provide arch_kexec_protect(unprotect)_crashkres()

[Xunlei Pang <xlpang@redhat.com>]

On 01/07/2016 at 05:20 PM, Petr Tesarik wrote:
> On Thu, 7 Jan 2016 13:08:21 +0800
> Xunlei Pang <xlpang@redhat.com> wrote:
>
>> On 01/07/2016 at 10:36 AM, Minfei Huang wrote:
>>> On 01/07/16 at 10:14am, Xunlei Pang wrote:
>>>>>> +static int
>>>>>> +kexec_mark_range(unsigned long start, unsigned long end, bool protect)
>>>>>> +{
>>>>>> +	struct page *page;
>>>>>> +	unsigned int nr_pages;
>>>>>> +
>>>>>> +	/* For physical range: [start, end] */
>>>>>> +	if (!start || !end || start > end)
>>>>>> +		return 0;
>>>>> Hi, Xunlei.
>>>>>
>>>>>         if (start > end)
>>>>>                 return 0;
>>>> If both start and end are zero, we want to return directly, so the two
>>>> more check doesn't hurt.
>>> How about if the start is equal to 0, and end is larger than 0? It is
>>> better to make code more robust, although it never happen in currect
>>> kexec code.
>> Hmm, this will be better:
>>
>> 	if (!end || start > end)
>> 		return 0;
>>
>> it handles the common case not using crash_low_res(start and end are 0).
> Hm, if both start and end are 0, then what about using this condition:
>
>  	if (start >= end)
>  		return 0;
>
> I think it's good enough, because if start is equal to end, then
> there's nothing to protect anyway.

In theory, start==end(not 0) still means we have 1B to protect :-)
But in practice there are no such cases, so I think this is ok.

Regards,
Xunlei

>
> Regards,
> Petr Tesarik
>
> _______________________________________________
> kexec mailing list
> kexec@lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/kexec