Re: [PATCH] mm: page_ext: allocate page extension though first PFN is invalid
From: Jaewon Kim
Date: Mon Nov 06 2017 - 17:30:12 EST
2017-11-02 17:02 GMT+09:00 Michal Hocko <mhocko@xxxxxxxxxx>:
> On Thu 02-11-17 15:35:07, Jaewon Kim wrote:
>> online_page_ext and page_ext_init allocate page_ext for each section, but
>> they do not allocate if the first PFN is !pfn_present(pfn) or
>> !pfn_valid(pfn). Then section->page_ext remains as NULL. lookup_page_ext
>> checks NULL only if CONFIG_DEBUG_VM is enabled. For a valid PFN,
>> __set_page_owner will try to get page_ext through lookup_page_ext.
>> Without CONFIG_DEBUG_VM lookup_page_ext will misuse NULL pointer as value
>> 0. This incurrs invalid address access.
>>
>> This is the panic example when PFN 0x100000 is not valid but PFN 0x13FC00
>> is being used for page_ext. section->page_ext is NULL, get_entry returned
>> invalid page_ext address as 0x1DFA000 for a PFN 0x13FC00.
>>
>> <1>[ 11.618085] Unable to handle kernel paging request at virtual address 01dfa014
>> <1>[ 11.618140] pgd = ffffffc0c6dc9000
>> <1>[ 11.618174] [01dfa014] *pgd=0000000000000000, *pud=0000000000000000
>> <4>[ 11.618240] ------------[ cut here ]------------
>> <2>[ 11.618278] Kernel BUG at ffffff80082371e0 [verbose debug info unavailable]
>> <0>[ 11.618338] Internal error: Oops: 96000045 [#1] PREEMPT SMP
>> <4>[ 11.618381] Modules linked in:
>> <4>[ 11.618524] task: ffffffc0c6ec9180 task.stack: ffffffc0c6f40000
>> <4>[ 11.618569] PC is at __set_page_owner+0x48/0x78
>> <4>[ 11.618607] LR is at __set_page_owner+0x44/0x78
>> <4>[ 11.626025] [<ffffff80082371e0>] __set_page_owner+0x48/0x78
>> <4>[ 11.626071] [<ffffff80081df9f0>] get_page_from_freelist+0x880/0x8e8
>> <4>[ 11.626118] [<ffffff80081e00a4>] __alloc_pages_nodemask+0x14c/0xc48
>> <4>[ 11.626165] [<ffffff80081e610c>] __do_page_cache_readahead+0xdc/0x264
>> <4>[ 11.626214] [<ffffff80081d8824>] filemap_fault+0x2ac/0x550
>> <4>[ 11.626259] [<ffffff80082e5cf8>] ext4_filemap_fault+0x3c/0x58
>> <4>[ 11.626305] [<ffffff800820a2f8>] __do_fault+0x80/0x120
>> <4>[ 11.626347] [<ffffff800820eb4c>] handle_mm_fault+0x704/0xbb0
>> <4>[ 11.626393] [<ffffff800809ba70>] do_page_fault+0x2e8/0x394
>> <4>[ 11.626437] [<ffffff8008080be4>] do_mem_abort+0x88/0x124
>>
>> Though the first page is not valid, page_ext could be useful for other
>> pages in the section. But checking all PFNs in a section may be time
>> consuming job. Let's check each (section count / 16) PFN, then prepare
>> page_ext if any PFN is present or valid. And remove the CONFIG_DEBUG_VM in
>> lookup_page_ext to avoid panic.
>
> So I would split this patch into two. First one to address the panic
> which sounds like a stable material and then the enhancement which will
> most likely need a further discussion.
Hello Michal Hocko
Thank you for your comment.
I think checking Null by erasing #if defined(CONFIG_DEBUG_VM) is the
stable material.
I wonder if you want me to split and resend the 2 patches, or if you
will use this mail thread for the further discussion.
Thank you
Jaewon Kim
>
>> Signed-off-by: Jaewon Kim <jaewon31.kim@xxxxxxxxxxx>
>> ---
>> mm/page_ext.c | 29 ++++++++++++++++++++++-------
>> 1 file changed, 22 insertions(+), 7 deletions(-)
>>
>> diff --git a/mm/page_ext.c b/mm/page_ext.c
>> index 32f18911deda..bf9c99beb312 100644
>> --- a/mm/page_ext.c
>> +++ b/mm/page_ext.c
>> @@ -124,7 +124,6 @@ struct page_ext *lookup_page_ext(struct page *page)
>> struct page_ext *base;
>>
>> base = NODE_DATA(page_to_nid(page))->node_page_ext;
>> -#if defined(CONFIG_DEBUG_VM)
>> /*
>> * The sanity checks the page allocator does upon freeing a
>> * page can reach here before the page_ext arrays are
>> @@ -133,7 +132,6 @@ struct page_ext *lookup_page_ext(struct page *page)
>> */
>> if (unlikely(!base))
>> return NULL;
>> -#endif
>> index = pfn - round_down(node_start_pfn(page_to_nid(page)),
>> MAX_ORDER_NR_PAGES);
>> return get_entry(base, index);
>> @@ -198,7 +196,6 @@ struct page_ext *lookup_page_ext(struct page *page)
>> {
>> unsigned long pfn = page_to_pfn(page);
>> struct mem_section *section = __pfn_to_section(pfn);
>> -#if defined(CONFIG_DEBUG_VM)
>> /*
>> * The sanity checks the page allocator does upon freeing a
>> * page can reach here before the page_ext arrays are
>> @@ -207,7 +204,6 @@ struct page_ext *lookup_page_ext(struct page *page)
>> */
>> if (!section->page_ext)
>> return NULL;
>> -#endif
>> return get_entry(section->page_ext, pfn);
>> }
>>
>> @@ -312,7 +308,17 @@ static int __meminit online_page_ext(unsigned long start_pfn,
>> }
>>
>> for (pfn = start; !fail && pfn < end; pfn += PAGES_PER_SECTION) {
>> - if (!pfn_present(pfn))
>> + unsigned long t_pfn = pfn;
>> + bool present = false;
>> +
>> + while (t_pfn < ALIGN(pfn + 1, PAGES_PER_SECTION)) {
>> + if (pfn_present(t_pfn)) {
>> + present = true;
>> + break;
>> + }
>> + t_pfn = ALIGN(pfn + 1, PAGES_PER_SECTION >> 4);
>> + }
>> + if (!present)
>> continue;
>> fail = init_section_page_ext(pfn, nid);
>> }
>> @@ -391,8 +397,17 @@ void __init page_ext_init(void)
>> */
>> for (pfn = start_pfn; pfn < end_pfn;
>> pfn = ALIGN(pfn + 1, PAGES_PER_SECTION)) {
>> -
>> - if (!pfn_valid(pfn))
>> + unsigned long t_pfn = pfn;
>> + bool valid = false;
>> +
>> + while (t_pfn < ALIGN(pfn + 1, PAGES_PER_SECTION)) {
>> + if (pfn_valid(t_pfn)) {
>> + valid = true;
>> + break;
>> + }
>> + t_pfn = ALIGN(pfn + 1, PAGES_PER_SECTION >> 4);
>> + }
>> + if (!valid)
>> continue;
>> /*
>> * Nodes's pfns can be overlapping.
>> --
>> 2.13.0
>
> --
> Michal Hocko
> SUSE Labs