mm: always expand the stack with the mmap write lock held
authorLinus Torvalds <torvalds@linux-foundation.org>
Sat, 24 Jun 2023 20:45:51 +0000 (13:45 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 27 Jun 2023 16:41:30 +0000 (09:41 -0700)
commit8d7071af890768438c14db6172cc8f9f4d04e184
tree75db95708563f58bf8d1790ca77bde5b306f5dc0
parentf313c51d26aa87e69633c9b46efb37a930faca71
mm: always expand the stack with the mmap write lock held

This finishes the job of always holding the mmap write lock when
extending the user stack vma, and removes the 'write_locked' argument
from the vm helper functions again.

For some cases, we just avoid expanding the stack at all: drivers and
page pinning really shouldn't be extending any stacks.  Let's see if any
strange users really wanted that.

It's worth noting that architectures that weren't converted to the new
lock_mm_and_find_vma() helper function are left using the legacy
"expand_stack()" function, but it has been changed to drop the mmap_lock
and take it for writing while expanding the vma.  This makes it fairly
straightforward to convert the remaining architectures.

As a result of dropping and re-taking the lock, the calling conventions
for this function have also changed, since the old vma may no longer be
valid.  So it will now return the new vma if successful, and NULL - and
the lock dropped - if the area could not be extended.

Tested-by: Vegard Nossum <vegard.nossum@oracle.com>
Tested-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> # ia64
Tested-by: Frank Scheiner <frank.scheiner@web.de> # ia64
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
17 files changed:
arch/ia64/mm/fault.c
arch/m68k/mm/fault.c
arch/microblaze/mm/fault.c
arch/openrisc/mm/fault.c
arch/parisc/mm/fault.c
arch/s390/mm/fault.c
arch/sparc/mm/fault_64.c
arch/um/kernel/trap.c
drivers/iommu/amd/iommu_v2.c
drivers/iommu/iommu-sva.c
fs/binfmt_elf.c
fs/exec.c
include/linux/mm.h
mm/gup.c
mm/memory.c
mm/mmap.c
mm/nommu.c