Re: [PATCH v2 1/6] mm: Handle page->mapping better in dump_page

[John Hubbard <jhubbard@nvidia.com>]

On 2020-07-10 18:30, John Hubbard wrote:
> On 2020-07-09 13:21, Matthew Wilcox (Oracle) wrote:
>> If we can't call page_mapping() to get the page mapping, handle the
>> anon/ksm/movable bits correctly.
> 
> Hi Matthew,
> 
> I see why you did this, as otherwise the new printing won't work.

Oh, and I forget to add: my only comment was just a documentation
suggestion, not critical at all, and so:

     Reviewed-by: John Hubbard <jhubbard@nvidia.com>

thanks,
-- 
John Hubbard
NVIDIA
> 
>>
>> Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
>> ---
>>   mm/debug.c | 7 ++++++-
>>   1 file changed, 6 insertions(+), 1 deletion(-)
>>
>> diff --git a/mm/debug.c b/mm/debug.c
>> index 4f376514744d..e5de63406b59 100644
>> --- a/mm/debug.c
>> +++ b/mm/debug.c
>> @@ -70,7 +70,12 @@ void __dump_page(struct page *page, const char *reason)
>>       if (page < head || (page >= head + MAX_ORDER_NR_PAGES)) {
>>           /* Corrupt page, cannot call page_mapping */
> 
> It's tricky to step lightly through the debris of corrupted page
> pointer here, but how about this updated comment, to replace the one-liner
> above:
> 
>          /*
>           * Corrupt page, so we cannot call page_mapping. Instead, do a
>           * safe subset of the steps that page_mapping() does. Caution:
>           * this will be misleading for tail pages, PageSwapCache pages,
>           * and potentially other situations. (See the page_mapping()
>           * implementation for what's missing here.)
>           */
> 
> 
>> -        mapping = page->mapping;
>> +        unsigned long tmp = (unsigned long)page->mapping;
>> +
>> +        if (tmp & PAGE_MAPPING_ANON)
>> +            mapping = NULL;
>> +        else
>> +            mapping = (void *)(tmp & ~PAGE_MAPPING_FLAGS);
>>           head = page;
>>           compound = false;
>>       } else {
>>
>