diff mbox series

[v6,04/14] mm, hugetlbfs: Allow for "high" userspace addresses

Message ID db238c1ca2d46e33c57328f8d450f2563e92f8c2.1639736449.git.christophe.leroy@csgroup.eu (mailing list archive)
State New, archived
Headers show
Series Convert powerpc to default topdown mmap layout | expand

Commit Message

Christophe Leroy Dec. 17, 2021, 10:27 a.m. UTC
This is a complement of 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.

Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Cc: Steve Capper <steve.capper@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
---
 fs/hugetlbfs/inode.c     | 9 +++++----
 include/linux/sched/mm.h | 8 ++++++++
 mm/mmap.c                | 8 --------
 3 files changed, 13 insertions(+), 12 deletions(-)

Comments

Catalin Marinas Jan. 4, 2022, 4:21 p.m. UTC | #1
On Fri, Dec 17, 2021 at 10:27:28AM +0000, Christophe Leroy wrote:
> This is a complement of 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.
> 
> Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
> Cc: Steve Capper <steve.capper@arm.com>
> Cc: Will Deacon <will.deacon@arm.com>
> Cc: Catalin Marinas <catalin.marinas@arm.com>
> ---
>  fs/hugetlbfs/inode.c     | 9 +++++----
>  include/linux/sched/mm.h | 8 ++++++++
>  mm/mmap.c                | 8 --------
>  3 files changed, 13 insertions(+), 12 deletions(-)

This would be an ABI change but I'm fine with it. Basically with this
patch, getting hugetblfs addresses above 48-bit require explicit hint
passed to mmap().

I wonder whether we should add a fixes tag (or at least the cc stable):

Fixes: f6795053dac8 ("mm: mmap: Allow for "high" userspace addresses")
Cc: <stable@vger.kernel.org> # 5.0.x

I think the original commit should have changed
hugetlb_get_unmapped_area() to have the same behaviour as
arch_get_unmapped_area(). Steve, any thoughts?

FWIW,

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
diff mbox series

Patch

diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c
index c7cde4e5924d..a8d3b0899b60 100644
--- a/fs/hugetlbfs/inode.c
+++ b/fs/hugetlbfs/inode.c
@@ -205,7 +205,7 @@  hugetlb_get_unmapped_area_bottomup(struct file *file, unsigned long addr,
 	info.flags = 0;
 	info.length = len;
 	info.low_limit = current->mm->mmap_base;
-	info.high_limit = TASK_SIZE;
+	info.high_limit = arch_get_mmap_end(addr, len, flags);
 	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
 	info.align_offset = 0;
 	return vm_unmapped_area(&info);
@@ -221,7 +221,7 @@  hugetlb_get_unmapped_area_topdown(struct file *file, unsigned long addr,
 	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
 	info.length = len;
 	info.low_limit = max(PAGE_SIZE, mmap_min_addr);
-	info.high_limit = current->mm->mmap_base;
+	info.high_limit = arch_get_mmap_base(addr, current->mm->mmap_base);
 	info.align_mask = PAGE_MASK & ~huge_page_mask(h);
 	info.align_offset = 0;
 	addr = vm_unmapped_area(&info);
@@ -236,7 +236,7 @@  hugetlb_get_unmapped_area_topdown(struct file *file, unsigned long addr,
 		VM_BUG_ON(addr != -ENOMEM);
 		info.flags = 0;
 		info.low_limit = current->mm->mmap_base;
-		info.high_limit = TASK_SIZE;
+		info.high_limit = arch_get_mmap_end(addr, len, flags);
 		addr = vm_unmapped_area(&info);
 	}
 
@@ -251,6 +251,7 @@  generic_hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 	struct mm_struct *mm = current->mm;
 	struct vm_area_struct *vma;
 	struct hstate *h = hstate_file(file);
+	const unsigned long mmap_end = arch_get_mmap_end(addr, len, flags);
 
 	if (len & ~huge_page_mask(h))
 		return -EINVAL;
@@ -266,7 +267,7 @@  generic_hugetlb_get_unmapped_area(struct file *file, unsigned long addr,
 	if (addr) {
 		addr = ALIGN(addr, huge_page_size(h));
 		vma = find_vma(mm, addr);
-		if (TASK_SIZE - len >= addr &&
+		if (mmap_end - len >= addr &&
 		    (!vma || addr + len <= vm_start_gap(vma)))
 			return addr;
 	}
diff --git a/include/linux/sched/mm.h b/include/linux/sched/mm.h
index 2584f7c13f69..cc9d80bd36d5 100644
--- a/include/linux/sched/mm.h
+++ b/include/linux/sched/mm.h
@@ -135,6 +135,14 @@  static inline void mm_update_next_owner(struct mm_struct *mm)
 #endif /* CONFIG_MEMCG */
 
 #ifdef CONFIG_MMU
+#ifndef arch_get_mmap_end
+#define arch_get_mmap_end(addr, len, flags)	(TASK_SIZE)
+#endif
+
+#ifndef arch_get_mmap_base
+#define arch_get_mmap_base(addr, base) (base)
+#endif
+
 extern void arch_pick_mmap_layout(struct mm_struct *mm,
 				  struct rlimit *rlim_stack);
 extern unsigned long
diff --git a/mm/mmap.c b/mm/mmap.c
index ad48f7af7511..c773b5ad9a11 100644
--- a/mm/mmap.c
+++ b/mm/mmap.c
@@ -2112,14 +2112,6 @@  unsigned long vm_unmapped_area(struct vm_unmapped_area_info *info)
 	return addr;
 }
 
-#ifndef arch_get_mmap_end
-#define arch_get_mmap_end(addr, len, flags)	(TASK_SIZE)
-#endif
-
-#ifndef arch_get_mmap_base
-#define arch_get_mmap_base(addr, base) (base)
-#endif
-
 /* Get an address range which is currently unmapped.
  * For shmat() with addr=0.
  *