Re: [PATCH v10] mm, hugetlbfs: Allow for "high" userspace addresses

[Christophe Leroy <christophe.leroy@csgroup.eu>]

Hi Catalin,

Le 16/04/2022 à 00:09, Andrew Morton a écrit :
> On Fri, 15 Apr 2022 16:45:13 +0200 Christophe Leroy <christophe.leroy@csgroup.eu> wrote:
> 
>> This is a fix for commit f6795053dac8 ("mm: mmap: Allow for "high"
>> userspace addresses") for hugetlb.
>>
>> This patch adds support for "high" userspace addresses that are
>> optionally supported on the system and have to be requested via a hint
>> mechanism ("high" addr parameter to mmap).
>>
>> Architectures such as powerpc and x86 achieve this by making changes to
>> their architectural versions of hugetlb_get_unmapped_area() function.
>> However, arm64 uses the generic version of that function.
>>
>> So take into account arch_get_mmap_base() and arch_get_mmap_end() in
>> hugetlb_get_unmapped_area(). To allow that, move those two macros
>> out of mm/mmap.c into include/linux/sched/mm.h
>>
>> If these macros are not defined in architectural code then they default
>> to (TASK_SIZE) and (base) so should not introduce any behavioural
>> changes to architectures that do not define them.
>>
>> For the time being, only ARM64 is affected by this change.
>>
>> >From Catalin (ARM64):
>>    We should have fixed hugetlb_get_unmapped_area() as well when we
>> added support for 52-bit VA. The reason for commit f6795053dac8 was to
>> prevent normal mmap() from returning addresses above 48-bit by default
>> as some user-space had hard assumptions about this.
>>
>> It's a slight ABI change if you do this for hugetlb_get_unmapped_area()
>> but I doubt anyone would notice. It's more likely that the current
>> behaviour would cause issues, so I'd rather have them consistent.
> 
> I'm struggling to understand the need for a -stable backport from the
> above text.
> 
> Could we please get a simple statement of the end-user visible effects
> of the shortcoming?  Target audience is -stable tree maintainers, and
> people who we've never heard of who will be wondering whether they should
> add this to their organization's older kernel.

Catalin, can you help answering this question ? It was your 
recommendation to tag this patch for stable in 
https://patchwork.ozlabs.org/project/linuxppc-dev/patch/db238c1ca2d46e33c57328f8d450f2563e92f8c2.1639736449.git.christophe.leroy@csgroup.eu/

> 
>>   fs/hugetlbfs/inode.c     | 9 +++++----
>>   include/linux/sched/mm.h | 8 ++++++++
>>   mm/mmap.c                | 8 --------
>>   3 files changed, 13 insertions(+), 12 deletions(-)
> 
> I'm a bit surprised that this has reached version 10!  Was it really
> that tricky?
> 

Well, that's the series it was part of that has reached v10. This patch 
was introduced in the series in v6

v6: 
https://patchwork.ozlabs.org/project/linuxppc-dev/patch/db238c1ca2d46e33c57328f8d450f2563e92f8c2.1639736449.git.christophe.leroy@csgroup.eu/
v7: 
https://patchwork.ozlabs.org/project/linuxppc-dev/patch/6c95091eab9f58cee58da3762a4dc4c56ab700e7.1642752946.git.christophe.leroy@csgroup.eu/
v8: 
https://patchwork.ozlabs.org/project/linuxppc-dev/patch/c234ceaf81ff37447fec5c9813d4ba5fc472a355.1646847562.git.christophe.leroy@csgroup.eu/
v9: 
https://patchwork.ozlabs.org/project/linuxppc-dev/patch/3bb944642140841c065f1cd6eae73f084fc026d2.1649401201.git.christophe.leroy@csgroup.eu/

Thanks
Christophe