2019-06-20 05:59:20

by Alexandre Ghiti

[permalink] [raw]
Subject: [PATCH RESEND 0/8] Fix mmap base in bottom-up mmap

This series fixes the fallback of the top-down mmap: in case of
failure, a bottom-up scheme can be tried as a last resort between
the top-down mmap base and the stack, hoping for a large unused stack
limit.

Lots of architectures and even mm code start this fallback
at TASK_UNMAPPED_BASE, which is useless since the top-down scheme
already failed on the whole address space: instead, simply use
mmap_base.

Along the way, it allows to get rid of of mmap_legacy_base and
mmap_compat_legacy_base from mm_struct.

Note that arm and mips already implement this behaviour.

Alexandre Ghiti (8):
s390: Start fallback of top-down mmap at mm->mmap_base
sh: Start fallback of top-down mmap at mm->mmap_base
sparc: Start fallback of top-down mmap at mm->mmap_base
x86, hugetlbpage: Start fallback of top-down mmap at mm->mmap_base
mm: Start fallback top-down mmap at mm->mmap_base
parisc: Use mmap_base, not mmap_legacy_base, as low_limit for
bottom-up mmap
x86: Use mmap_*base, not mmap_*legacy_base, as low_limit for bottom-up
mmap
mm: Remove mmap_legacy_base and mmap_compat_legacy_code fields from
mm_struct

arch/parisc/kernel/sys_parisc.c | 8 +++-----
arch/s390/mm/mmap.c | 2 +-
arch/sh/mm/mmap.c | 2 +-
arch/sparc/kernel/sys_sparc_64.c | 2 +-
arch/sparc/mm/hugetlbpage.c | 2 +-
arch/x86/include/asm/elf.h | 2 +-
arch/x86/kernel/sys_x86_64.c | 4 ++--
arch/x86/mm/hugetlbpage.c | 7 ++++---
arch/x86/mm/mmap.c | 20 +++++++++-----------
include/linux/mm_types.h | 2 --
mm/debug.c | 4 ++--
mm/mmap.c | 2 +-
12 files changed, 26 insertions(+), 31 deletions(-)

--
2.20.1


2019-06-20 06:03:42

by Alexandre Ghiti

[permalink] [raw]
Subject: [PATCH RESEND 4/8] x86, hugetlbpage: Start fallback of top-down mmap at mm->mmap_base

In case of mmap failure in top-down mode, there is no need to go through
the whole address space again for the bottom-up fallback: the goal of this
fallback is to find, as a last resort, space between the top-down mmap base
and the stack, which is the only place not covered by the top-down mmap.

Signed-off-by: Alexandre Ghiti <[email protected]>
---
arch/x86/mm/hugetlbpage.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/arch/x86/mm/hugetlbpage.c b/arch/x86/mm/hugetlbpage.c
index fab095362c50..4b90339aef50 100644
--- a/arch/x86/mm/hugetlbpage.c
+++ b/arch/x86/mm/hugetlbpage.c
@@ -106,11 +106,12 @@ static unsigned long hugetlb_get_unmapped_area_topdown(struct file *file,
{
struct hstate *h = hstate_file(file);
struct vm_unmapped_area_info info;
+ unsigned long mmap_base = get_mmap_base(0);

info.flags = VM_UNMAPPED_AREA_TOPDOWN;
info.length = len;
info.low_limit = PAGE_SIZE;
- info.high_limit = get_mmap_base(0);
+ info.high_limit = mmap_base;

/*
* If hint address is above DEFAULT_MAP_WINDOW, look for unmapped area
@@ -132,7 +133,7 @@ static unsigned long hugetlb_get_unmapped_area_topdown(struct file *file,
if (addr & ~PAGE_MASK) {
VM_BUG_ON(addr != -ENOMEM);
info.flags = 0;
- info.low_limit = TASK_UNMAPPED_BASE;
+ info.low_limit = mmap_base;
info.high_limit = TASK_SIZE_LOW;
addr = vm_unmapped_area(&info);
}
--
2.20.1

2019-07-30 13:48:59

by Alexandre Ghiti

[permalink] [raw]
Subject: Re: [PATCH RESEND 0/8] Fix mmap base in bottom-up mmap

On 6/20/19 7:03 AM, Alexandre Ghiti wrote:
> This series fixes the fallback of the top-down mmap: in case of
> failure, a bottom-up scheme can be tried as a last resort between
> the top-down mmap base and the stack, hoping for a large unused stack
> limit.
>
> Lots of architectures and even mm code start this fallback
> at TASK_UNMAPPED_BASE, which is useless since the top-down scheme
> already failed on the whole address space: instead, simply use
> mmap_base.
>
> Along the way, it allows to get rid of of mmap_legacy_base and
> mmap_compat_legacy_base from mm_struct.
>
> Note that arm and mips already implement this behaviour.
>
> Alexandre Ghiti (8):
> s390: Start fallback of top-down mmap at mm->mmap_base
> sh: Start fallback of top-down mmap at mm->mmap_base
> sparc: Start fallback of top-down mmap at mm->mmap_base
> x86, hugetlbpage: Start fallback of top-down mmap at mm->mmap_base
> mm: Start fallback top-down mmap at mm->mmap_base
> parisc: Use mmap_base, not mmap_legacy_base, as low_limit for
> bottom-up mmap
> x86: Use mmap_*base, not mmap_*legacy_base, as low_limit for bottom-up
> mmap
> mm: Remove mmap_legacy_base and mmap_compat_legacy_code fields from
> mm_struct
>
> arch/parisc/kernel/sys_parisc.c | 8 +++-----
> arch/s390/mm/mmap.c | 2 +-
> arch/sh/mm/mmap.c | 2 +-
> arch/sparc/kernel/sys_sparc_64.c | 2 +-
> arch/sparc/mm/hugetlbpage.c | 2 +-
> arch/x86/include/asm/elf.h | 2 +-
> arch/x86/kernel/sys_x86_64.c | 4 ++--
> arch/x86/mm/hugetlbpage.c | 7 ++++---
> arch/x86/mm/mmap.c | 20 +++++++++-----------
> include/linux/mm_types.h | 2 --
> mm/debug.c | 4 ++--
> mm/mmap.c | 2 +-
> 12 files changed, 26 insertions(+), 31 deletions(-)
>

Hi everyone,

This is just a preparatory series for the merging of x86 mmap top-down
functions with
the generic ones (those should get into v5.3), if you could take some
time to take a look,
that would be great :)

Thanks,

Alex

2019-08-26 07:36:53

by Alexandre Ghiti

[permalink] [raw]
Subject: Re: [PATCH RESEND 0/8] Fix mmap base in bottom-up mmap

On 6/20/19 7:03 AM, Alexandre Ghiti wrote:
> This series fixes the fallback of the top-down mmap: in case of
> failure, a bottom-up scheme can be tried as a last resort between
> the top-down mmap base and the stack, hoping for a large unused stack
> limit.
>
> Lots of architectures and even mm code start this fallback
> at TASK_UNMAPPED_BASE, which is useless since the top-down scheme
> already failed on the whole address space: instead, simply use
> mmap_base.
>
> Along the way, it allows to get rid of of mmap_legacy_base and
> mmap_compat_legacy_base from mm_struct.
>
> Note that arm and mips already implement this behaviour.
>
> Alexandre Ghiti (8):
> s390: Start fallback of top-down mmap at mm->mmap_base
> sh: Start fallback of top-down mmap at mm->mmap_base
> sparc: Start fallback of top-down mmap at mm->mmap_base
> x86, hugetlbpage: Start fallback of top-down mmap at mm->mmap_base
> mm: Start fallback top-down mmap at mm->mmap_base
> parisc: Use mmap_base, not mmap_legacy_base, as low_limit for
> bottom-up mmap
> x86: Use mmap_*base, not mmap_*legacy_base, as low_limit for bottom-up
> mmap
> mm: Remove mmap_legacy_base and mmap_compat_legacy_code fields from
> mm_struct
>
> arch/parisc/kernel/sys_parisc.c | 8 +++-----
> arch/s390/mm/mmap.c | 2 +-
> arch/sh/mm/mmap.c | 2 +-
> arch/sparc/kernel/sys_sparc_64.c | 2 +-
> arch/sparc/mm/hugetlbpage.c | 2 +-
> arch/x86/include/asm/elf.h | 2 +-
> arch/x86/kernel/sys_x86_64.c | 4 ++--
> arch/x86/mm/hugetlbpage.c | 7 ++++---
> arch/x86/mm/mmap.c | 20 +++++++++-----------
> include/linux/mm_types.h | 2 --
> mm/debug.c | 4 ++--
> mm/mmap.c | 2 +-
> 12 files changed, 26 insertions(+), 31 deletions(-)
>

Hi everyone,

Any thoughts about that series ? As said before, this is just a
preparatory patchset in order to
merge x86 mmap top down code with the generic version.

Thanks for taking a look,

Alex

2019-08-26 22:41:39

by Helge Deller

[permalink] [raw]
Subject: Re: [PATCH RESEND 0/8] Fix mmap base in bottom-up mmap

On 26.08.19 09:34, Alexandre Ghiti wrote:
> On 6/20/19 7:03 AM, Alexandre Ghiti wrote:
>> This series fixes the fallback of the top-down mmap: in case of
>> failure, a bottom-up scheme can be tried as a last resort between
>> the top-down mmap base and the stack, hoping for a large unused stack
>> limit.
>>
>> Lots of architectures and even mm code start this fallback
>> at TASK_UNMAPPED_BASE, which is useless since the top-down scheme
>> already failed on the whole address space: instead, simply use
>> mmap_base.
>>
>> Along the way, it allows to get rid of of mmap_legacy_base and
>> mmap_compat_legacy_base from mm_struct.
>>
>> Note that arm and mips already implement this behaviour.
>>
>> Alexandre Ghiti (8):
>>    s390: Start fallback of top-down mmap at mm->mmap_base
>>    sh: Start fallback of top-down mmap at mm->mmap_base
>>    sparc: Start fallback of top-down mmap at mm->mmap_base
>>    x86, hugetlbpage: Start fallback of top-down mmap at mm->mmap_base
>>    mm: Start fallback top-down mmap at mm->mmap_base
>>    parisc: Use mmap_base, not mmap_legacy_base, as low_limit for
>>      bottom-up mmap
>>    x86: Use mmap_*base, not mmap_*legacy_base, as low_limit for bottom-up
>>      mmap
>>    mm: Remove mmap_legacy_base and mmap_compat_legacy_code fields from
>>      mm_struct
>>
>>   arch/parisc/kernel/sys_parisc.c  |  8 +++-----
>>   arch/s390/mm/mmap.c              |  2 +-
>>   arch/sh/mm/mmap.c                |  2 +-
>>   arch/sparc/kernel/sys_sparc_64.c |  2 +-
>>   arch/sparc/mm/hugetlbpage.c      |  2 +-
>>   arch/x86/include/asm/elf.h       |  2 +-
>>   arch/x86/kernel/sys_x86_64.c     |  4 ++--
>>   arch/x86/mm/hugetlbpage.c        |  7 ++++---
>>   arch/x86/mm/mmap.c               | 20 +++++++++-----------
>>   include/linux/mm_types.h         |  2 --
>>   mm/debug.c                       |  4 ++--
>>   mm/mmap.c                        |  2 +-
>>   12 files changed, 26 insertions(+), 31 deletions(-)
>>
>
> Any thoughts about that series ? As said before, this is just a preparatory patchset in order to
> merge x86 mmap top down code with the generic version.

I just tested your patch series successfully on the parisc
architeture. You may add:

Tested-by: Helge Deller <[email protected]> # parisc

Thanks!
Helge

2019-08-28 04:54:32

by Alexandre Ghiti

[permalink] [raw]
Subject: Re: [PATCH RESEND 0/8] Fix mmap base in bottom-up mmap

On 8/26/19 6:37 PM, Helge Deller wrote:
> On 26.08.19 09:34, Alexandre Ghiti wrote:
>> On 6/20/19 7:03 AM, Alexandre Ghiti wrote:
>>> This series fixes the fallback of the top-down mmap: in case of
>>> failure, a bottom-up scheme can be tried as a last resort between
>>> the top-down mmap base and the stack, hoping for a large unused stack
>>> limit.
>>>
>>> Lots of architectures and even mm code start this fallback
>>> at TASK_UNMAPPED_BASE, which is useless since the top-down scheme
>>> already failed on the whole address space: instead, simply use
>>> mmap_base.
>>>
>>> Along the way, it allows to get rid of of mmap_legacy_base and
>>> mmap_compat_legacy_base from mm_struct.
>>>
>>> Note that arm and mips already implement this behaviour.
>>>
>>> Alexandre Ghiti (8):
>>>    s390: Start fallback of top-down mmap at mm->mmap_base
>>>    sh: Start fallback of top-down mmap at mm->mmap_base
>>>    sparc: Start fallback of top-down mmap at mm->mmap_base
>>>    x86, hugetlbpage: Start fallback of top-down mmap at mm->mmap_base
>>>    mm: Start fallback top-down mmap at mm->mmap_base
>>>    parisc: Use mmap_base, not mmap_legacy_base, as low_limit for
>>>      bottom-up mmap
>>>    x86: Use mmap_*base, not mmap_*legacy_base, as low_limit for
>>> bottom-up
>>>      mmap
>>>    mm: Remove mmap_legacy_base and mmap_compat_legacy_code fields from
>>>      mm_struct
>>>
>>>   arch/parisc/kernel/sys_parisc.c  |  8 +++-----
>>>   arch/s390/mm/mmap.c              |  2 +-
>>>   arch/sh/mm/mmap.c                |  2 +-
>>>   arch/sparc/kernel/sys_sparc_64.c |  2 +-
>>>   arch/sparc/mm/hugetlbpage.c      |  2 +-
>>>   arch/x86/include/asm/elf.h       |  2 +-
>>>   arch/x86/kernel/sys_x86_64.c     |  4 ++--
>>>   arch/x86/mm/hugetlbpage.c        |  7 ++++---
>>>   arch/x86/mm/mmap.c               | 20 +++++++++-----------
>>>   include/linux/mm_types.h         |  2 --
>>>   mm/debug.c                       |  4 ++--
>>>   mm/mmap.c                        |  2 +-
>>>   12 files changed, 26 insertions(+), 31 deletions(-)
>>>
>>
>> Any thoughts about that series ? As said before, this is just a
>> preparatory patchset in order to
>> merge x86 mmap top down code with the generic version.
>
> I just tested your patch series successfully on the parisc
> architeture. You may add:
>
> Tested-by: Helge Deller <[email protected]> # parisc

Thanks again Helge !

Alex


>
> Thanks!
> Helge