Re: [PATCH v3 12/12] mm/debug_vm_pgtable: Fix corrupted page flag

[Gavin Shan]

Hi Anshuman,

On 7/21/21 8:18 PM, Anshuman Khandual wrote:
> On 7/19/21 6:36 PM, Gavin Shan wrote:
> > In page table entry modifying tests, set_xxx_at() are used to populate
> > the page table entries. On ARM64, PG_arch_1 is set to the target page
> > flag if execution permission is given. The page flag is kept when the
> > page is free'd to buddy's free area list. However, it will trigger page
> > checking failure when it's pulled from the buddy's free area list, as
> > the following warning messages indicate.
> >
> >     BUG: Bad page state in process memhog  pfn:08000
> >     page:0000000015c0a628 refcount:0 mapcount:0 \
> >          mapping:0000000000000000 index:0x1 pfn:0x8000
> >     flags: 0x7ffff8000000800(arch_1|node=0|zone=0|lastcpupid=0xfffff)
> >     raw: 07ffff8000000800 dead000000000100 dead000000000122 0000000000000000
> >     raw: 0000000000000001 0000000000000000 00000000ffffffff 0000000000000000
> >     page dumped because: PAGE_FLAGS_CHECK_AT_PREP flag(s) set
> >
> > This fixes the issue by clearing PG_arch_1 through flush_dcache_page()
> > after set_xxx_at() is called.
>
> Could you please add comments before each flush_dcache_page() instance
> explaining why this is needed for arm64 platforms with relevant PG_arch_1
> context and how this does not have any adverse effect on other platforms ?
> It should be easy for some one looking at this code after a while to figure
> out from where flush_dcache_page() came from.

Good point. I will improve chage log to include the commit ID in v4 where the
page flag (PG_arch_1) is used and explain how. In that case, it's much clearer
to understand the reason why we need flush_dcache_page() after set_xxx_at() on
ARM64.

Thanks,
Gavin