diff mbox

[v2] arm64: mm: Fix memmap to be initialized for the entire section

Message ID ffaa77c7-8982-524f-f624-e4463a463daf@huawei.com (mailing list archive)
State New, archived
Headers show

Commit Message

Xie Yisheng Dec. 9, 2016, 1:15 p.m. UTC 
On 2016/12/9 20:23, Hanjun Guo wrote:
> On 12/09/2016 08:19 PM, Ard Biesheuvel wrote:
>> On 9 December 2016 at 12:14, Yisheng Xie <xieyisheng1@huawei.com> wrote:
>>> Hi Robert,
>>> We have merged your patch to 4.9.0-rc8, however we still meet the similar problem
>>> on our D05 board:
>>>
>>
>> To be clear: does this issue also occur on D05 *without* the patch?
> 
> It boots ok on D05 without this patch.
> 
> But I think the problem Robert described in the commit message is
> still there, just not triggered in the boot. we met this problem
> when having LTP stress memory test and hit the same BUG_ON.
> 
That's right. for D05's case, when trigger the BUG_ON on:
1863         VM_BUG_ON(page_zone(start_page) != page_zone(end_page));

the end_page is not nomap, but BIOS reserved. here is the log I got:

[    0.000000] efi:   0x00003fc00000-0x00003fffffff [Reserved           |   |  |  |  |  |  |  |   |  |  |  |  ]
[...]
[    5.081443] move_freepages: phys(start_page: 0x20000000,end_page:0x3fff0000)

For invalid pages, their zone and node information is not initialized, and it
do have risk to trigger the BUG_ON, so I have a silly question,
why not just change the BUG_ON:
-----------

Thanks,
Yisheng Xie

> Thanks
> Hanjun
> 
> .
>

Comments

Richter, Robert Dec. 9, 2016, 2:52 p.m. UTC | #1 
On 09.12.16 21:15:12, Yisheng Xie wrote:
> For invalid pages, their zone and node information is not initialized, and it
> do have risk to trigger the BUG_ON, so I have a silly question,
> why not just change the BUG_ON:

We need to get the page handling correct. Modifying the BUG_ON() just
hides that something is wrong.

-Robert

> -----------
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 6de9440..af199b8 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -1860,12 +1860,13 @@ int move_freepages(struct zone *zone,
>          * Remove at a later date when no bug reports exist related to
>          * grouping pages by mobility
>          */
> -       VM_BUG_ON(page_zone(start_page) != page_zone(end_page));
> +       VM_BUG_ON(early_pfn_valid(start_page) && early_pfn_valid(end_page) &&
> +                       page_zone(start_page) != page_zone(end_page));
>  #endif
> 
>         for (page = start_page; page <= end_page;) {
>                 /* Make sure we are not inadvertently changing nodes */
> -               VM_BUG_ON_PAGE(page_to_nid(page) != zone_to_nid(zone), page);
> +               VM_BUG_ON_PAGE(early_pfn_valid(page) && (page_to_nid(page) != zone_to_nid(zone)), page);
> 
>                 if (!pfn_valid_within(page_to_pfn(page))) {
>                         page++;
Ard Biesheuvel Dec. 12, 2016, 12:02 p.m. UTC | #2 
On 9 December 2016 at 14:52, Robert Richter <robert.richter@cavium.com> wrote:
> On 09.12.16 21:15:12, Yisheng Xie wrote:
>> For invalid pages, their zone and node information is not initialized, and it
>> do have risk to trigger the BUG_ON, so I have a silly question,
>> why not just change the BUG_ON:
>
> We need to get the page handling correct. Modifying the BUG_ON() just
> hides that something is wrong.
>

Actually, I think this is a reasonable question. We are trying very
hard to work around the BUG_ON(), which arguably does something wrong
by calling page_to_nid() on a struct page without checking if it is a
valid page.

Looking at commit 344c790e3821 ("mm: make setup_zone_migrate_reserve()
aware of overlapping nodes"), the BUG_ON() has a specific purpose
related to adjacent zones.

What will go wrong if we ignore this check?


>
>> -----------
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index 6de9440..af199b8 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -1860,12 +1860,13 @@ int move_freepages(struct zone *zone,
>>          * Remove at a later date when no bug reports exist related to
>>          * grouping pages by mobility
>>          */
>> -       VM_BUG_ON(page_zone(start_page) != page_zone(end_page));
>> +       VM_BUG_ON(early_pfn_valid(start_page) && early_pfn_valid(end_page) &&
>> +                       page_zone(start_page) != page_zone(end_page));
>>  #endif
>>
>>         for (page = start_page; page <= end_page;) {
>>                 /* Make sure we are not inadvertently changing nodes */
>> -               VM_BUG_ON_PAGE(page_to_nid(page) != zone_to_nid(zone), page);
>> +               VM_BUG_ON_PAGE(early_pfn_valid(page) && (page_to_nid(page) != zone_to_nid(zone)), page);
>>
>>                 if (!pfn_valid_within(page_to_pfn(page))) {
>>                         page++;
>
diff mbox

Patch

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 6de9440..af199b8 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -1860,12 +1860,13 @@  int move_freepages(struct zone *zone,
         * Remove at a later date when no bug reports exist related to
         * grouping pages by mobility
         */
-       VM_BUG_ON(page_zone(start_page) != page_zone(end_page));
+       VM_BUG_ON(early_pfn_valid(start_page) && early_pfn_valid(end_page) &&
+                       page_zone(start_page) != page_zone(end_page));
 #endif

        for (page = start_page; page <= end_page;) {
                /* Make sure we are not inadvertently changing nodes */
-               VM_BUG_ON_PAGE(page_to_nid(page) != zone_to_nid(zone), page);
+               VM_BUG_ON_PAGE(early_pfn_valid(page) && (page_to_nid(page) != zone_to_nid(zone)), page);

                if (!pfn_valid_within(page_to_pfn(page))) {
                        page++;