Re: [PATCH 1/2] mm, memory_hotplug: cleanup memory offline path

[David Hildenbrand <david@redhat.com>]

On 04.04.19 17:40, Oscar Salvador wrote:
> On Thu, Apr 04, 2019 at 04:47:43PM +0200, David Hildenbrand wrote:
>> On 04.04.19 15:25, Oscar Salvador wrote:
>>> On Thu, Apr 04, 2019 at 03:18:00PM +0200, David Hildenbrand wrote:
>>>>> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
>>>>> index f206b8b66af1..d8a3e9554aec 100644
>>>>> --- a/mm/memory_hotplug.c
>>>>> +++ b/mm/memory_hotplug.c
>>>>> @@ -1451,15 +1451,11 @@ static int
>>>>>  offline_isolated_pages_cb(unsigned long start, unsigned long nr_pages,
>>>>>  			void *data)
>>>>>  {
>>>>> -	__offline_isolated_pages(start, start + nr_pages);
>>>>> -	return 0;
>>>>> -}
>>>>> +	unsigned long offlined_pages;
>>>>>  
>>>>> -static void
>>>>> -offline_isolated_pages(unsigned long start_pfn, unsigned long end_pfn)
>>>>> -{
>>>>> -	walk_system_ram_range(start_pfn, end_pfn - start_pfn, NULL,
>>>>> -				offline_isolated_pages_cb);
>>>>> +	offlined_pages = __offline_isolated_pages(start, start + nr_pages);
>>>>> +	*(unsigned long *)data += offlined_pages;
>>>>
>>>> unsigned long *offlined_pages = data;
>>>>
>>>> *offlined_pages += __offline_isolated_pages(start, start + nr_pages);
>>>
>>> Yeah, more readable.
>>>
>>>> Only nits
>>>
>>> About the identation, I double checked the code and it looks fine to me.
>>> In [1] looks fine too, might be your mail client?
>>>
>>> [1] https://patchwork.kernel.org/patch/10885571/
>>
>> Double checked, alignment on the parameter on the new line is very weird.
> 
> Uhm, are not you confused because we removed the "while (off...)", and
> "ret =" gets idented right below "/*check again*".
> 
> Try to apply the patch and check whether you still see the issue.
> I just checked out the branch and it looks fine to me.


That's what I did and it hurts my eyes (dropping two tabs, converting
tabs to spaces)

your patch:

ret = walk_system_ram_range(start_pfn, end_pfn - start_pfn, NULL,
                                        check_pages_isolated_cb);

vs.

ret = walk_system_ram_range(start_pfn, end_pfn - start_pfn, NULL,
                            check_pages_isolated_cb);


Just so we are on the same page, we usually indent parameters on
additional lines to the start of other parameters. Not to the end
of the previous line.

> 
>> And both lines cross 80 lines per line ... nit :)
> 
> Yeah, 81 characters, but I decided to go with that rather than start doing
> tricky things to accomplish 80 characters.
> Maybe Andrew agrees, or he might slap me.
> 

Why not simply


ret = walk_system_ram_range(start_pfn, end_pfn - start_pfn,
			    NULL, check_pages_isolated_cb);

just as we have in add_memory_resource along with walk_memory_range().


A lot of nit-picking, sorry :)

-- 

Thanks,

David / dhildenb