Re: [PATCH v5 10/14] x86/mm: Simplify clear_page_*

[Ankur Arora <ankur.a.arora@oracle.com>]

David Hildenbrand <david@redhat.com> writes:

> On 11.07.25 19:26, Ankur Arora wrote:
>> David Hildenbrand <david@redhat.com> writes:
>>
>>> On 10.07.25 02:59, Ankur Arora wrote:
>>>> clear_page_rep() and clear_page_erms() are wrappers around "REP; STOS"
>>>> variations. Inlining gets rid of an unnecessary CALL/RET (which isn't
>>>> free when using RETHUNK speculative execution mitigations.)
>>>> Fixup and rename clear_page_orig() to adapt to the changed calling
>>>> convention.
>>>> And, add a comment from Dave Hansen detailing various clearing mechanisms
>>>> used in clear_page().
>>>> Signed-off-by: Ankur Arora <ankur.a.arora@oracle.com>
>>>> ---
>>>>    arch/x86/include/asm/page_32.h |  6 +++++
>>>>    arch/x86/include/asm/page_64.h | 42 ++++++++++++++++++++++++++--------
>>>>    arch/x86/lib/clear_page_64.S   | 39 +++++++------------------------
>>>>    3 files changed, 46 insertions(+), 41 deletions(-)
>>>> diff --git a/arch/x86/include/asm/page_32.h b/arch/x86/include/asm/page_32.h
>>>> index 0c623706cb7e..a8ff43bb9652 100644
>>>> --- a/arch/x86/include/asm/page_32.h
>>>> +++ b/arch/x86/include/asm/page_32.h
>>>> @@ -17,6 +17,12 @@ extern unsigned long __phys_addr(unsigned long);
>>>>      #include <linux/string.h>
>>>>    +/*
>>>
>>> /** if this was supposed to be kernel doc (which it looks like it is)
>>>
>>>> + * clear_page() - clear kernel page.
>>>
>>> "clear a kernel page"
>>>
>>> Although I am not sure what a "kernel page" is.
>>>
>>> Did you mean "clear a page using a kernel virtual address" ?
>> Thanks. Yes, this makes way more sense.
>
> FWIW, most clear_user_page() will just call clear_page(). The ones that don't,
> have to deal with the dcache flushing, as the page might have cache alias from
> another (mm) address space.
>
> So clear_page() is just about clearing page content using a kernel virtual
> address, and clear_user_page() is additionally taking care of any dcache
> aliases.

Yeah that's a good point.

Right now I'm assuming that clear_user_page() and clear_page() are
effectively identical. Which as you say, isn't true for all the
architectures that can't disambiguate dcache aliases.

So maybe I should be adjusting the naming of the helpers a bit?

Specifically, by calling clear_user_pages() from clear_pages_resched().
Which on x86 etc can just call clear_pages().

--
ankur