[mirror_ubuntu-hirsute-kernel.git] / Documentation / core-api / gfp_mask-from-fs-io.rst

=================================
GFP masks used from FS/IO context
=================================

:Date: May, 2018
:Author: Michal Hocko <mhocko@kernel.org>

Introduction
============

Code paths in the filesystem and IO stacks must be careful when
allocating memory to prevent recursion deadlocks caused by direct
memory reclaim calling back into the FS or IO paths and blocking on
already held resources (e.g. locks - most commonly those used for the
transaction context).

The traditional way to avoid this deadlock problem is to clear __GFP_FS
respectively __GFP_IO (note the latter implies clearing the first as well) in
the gfp mask when calling an allocator. GFP_NOFS respectively GFP_NOIO can be
used as shortcut. It turned out though that above approach has led to
abuses when the restricted gfp mask is used "just in case" without a
deeper consideration which leads to problems because an excessive use
of GFP_NOFS/GFP_NOIO can lead to memory over-reclaim or other memory
reclaim issues.

New API
========

Since 4.12 we do have a generic scope API for both NOFS and NOIO context
``memalloc_nofs_save``, ``memalloc_nofs_restore`` respectively ``memalloc_noio_save``,
``memalloc_noio_restore`` which allow to mark a scope to be a critical
section from a filesystem or I/O point of view. Any allocation from that
scope will inherently drop __GFP_FS respectively __GFP_IO from the given
mask so no memory allocation can recurse back in the FS/IO.

.. kernel-doc:: include/linux/sched/mm.h
   :functions: memalloc_nofs_save memalloc_nofs_restore
.. kernel-doc:: include/linux/sched/mm.h
   :functions: memalloc_noio_save memalloc_noio_restore

FS/IO code then simply calls the appropriate save function before
any critical section with respect to the reclaim is started - e.g.
lock shared with the reclaim context or when a transaction context
nesting would be possible via reclaim. The restore function should be
called when the critical section ends. All that ideally along with an
explanation what is the reclaim context for easier maintenance.

Please note that the proper pairing of save/restore functions
allows nesting so it is safe to call ``memalloc_noio_save`` or
``memalloc_noio_restore`` respectively from an existing NOIO or NOFS
scope.

What about __vmalloc(GFP_NOFS)
==============================

vmalloc doesn't support GFP_NOFS semantic because there are hardcoded
GFP_KERNEL allocations deep inside the allocator which are quite non-trivial
to fix up. That means that calling ``vmalloc`` with GFP_NOFS/GFP_NOIO is
almost always a bug. The good news is that the NOFS/NOIO semantic can be
achieved by the scope API.

In the ideal world, upper layers should already mark dangerous contexts
and so no special care is required and vmalloc should be called without
any problems. Sometimes if the context is not really clear or there are
layering violations then the recommended way around that is to wrap ``vmalloc``
by the scope API with a comment explaining the problem.
Commit	Line	Data
46ca3599 MH	1	=================================
	2	GFP masks used from FS/IO context
	3	=================================
	4
	5	:Date: May, 2018
	6	:Author: Michal Hocko <mhocko@kernel.org>
	7
	8	Introduction
	9	============
	10
	11	Code paths in the filesystem and IO stacks must be careful when
	12	allocating memory to prevent recursion deadlocks caused by direct
	13	memory reclaim calling back into the FS or IO paths and blocking on
	14	already held resources (e.g. locks - most commonly those used for the
	15	transaction context).
	16
	17	The traditional way to avoid this deadlock problem is to clear __GFP_FS
	18	respectively __GFP_IO (note the latter implies clearing the first as well) in
	19	the gfp mask when calling an allocator. GFP_NOFS respectively GFP_NOIO can be
	20	used as shortcut. It turned out though that above approach has led to
	21	abuses when the restricted gfp mask is used "just in case" without a
	22	deeper consideration which leads to problems because an excessive use
	23	of GFP_NOFS/GFP_NOIO can lead to memory over-reclaim or other memory
	24	reclaim issues.
	25
	26	New API
	27	========
	28
	29	Since 4.12 we do have a generic scope API for both NOFS and NOIO context
	30	``memalloc_nofs_save``, ``memalloc_nofs_restore`` respectively ``memalloc_noio_save``,
	31	``memalloc_noio_restore`` which allow to mark a scope to be a critical
	32	section from a filesystem or I/O point of view. Any allocation from that
	33	scope will inherently drop __GFP_FS respectively __GFP_IO from the given
	34	mask so no memory allocation can recurse back in the FS/IO.
	35
d43f2c98 JC	36	.. kernel-doc:: include/linux/sched/mm.h
	37	:functions: memalloc_nofs_save memalloc_nofs_restore
	38	.. kernel-doc:: include/linux/sched/mm.h
	39	:functions: memalloc_noio_save memalloc_noio_restore
	40
46ca3599 MH	41	FS/IO code then simply calls the appropriate save function before
	42	any critical section with respect to the reclaim is started - e.g.
	43	lock shared with the reclaim context or when a transaction context
	44	nesting would be possible via reclaim. The restore function should be
	45	called when the critical section ends. All that ideally along with an
	46	explanation what is the reclaim context for easier maintenance.
	47
	48	Please note that the proper pairing of save/restore functions
	49	allows nesting so it is safe to call ``memalloc_noio_save`` or
	50	``memalloc_noio_restore`` respectively from an existing NOIO or NOFS
	51	scope.
	52
	53	What about __vmalloc(GFP_NOFS)
	54	==============================
	55
	56	vmalloc doesn't support GFP_NOFS semantic because there are hardcoded
	57	GFP_KERNEL allocations deep inside the allocator which are quite non-trivial
	58	to fix up. That means that calling ``vmalloc`` with GFP_NOFS/GFP_NOIO is
	59	almost always a bug. The good news is that the NOFS/NOIO semantic can be
	60	achieved by the scope API.
	61
	62	In the ideal world, upper layers should already mark dangerous contexts
	63	and so no special care is required and vmalloc should be called without
	64	any problems. Sometimes if the context is not really clear or there are
	65	layering violations then the recommended way around that is to wrap ``vmalloc``
	66	by the scope API with a comment explaining the problem.