This patch introduces 'write_backward' bit to perf_event_attr, which
controls the direction of a ring buffer. After set, the corresponding
ring buffer is written from end to beginning. This feature is design to
support reading from overwritable ring buffer.
Ring buffer can be created by mapping a perf event fd. Kernel puts event
records into ring buffer, user tooling like perf fetch them from
address returned by mmap(). To prevent racing between kernel and tooling,
they communicate to each other through 'head' and 'tail' pointers.
Kernel maintains 'head' pointer, points it to the next free area (tail
of the last record). Tooling maintains 'tail' pointer, points it to the
tail of last consumed record (record has already been fetched). Kernel
determines the available space in a ring buffer using these two
pointers to avoid overwrite unfetched records.
By mapping without 'PROT_WRITE', an overwritable ring buffer is created.
Different from normal ring buffer, tooling is unable to maintain 'tail'
pointer because writing is forbidden. Therefore, for this type of ring
buffers, kernel overwrite old records unconditionally, works like flight
recorder. This feature would be useful if reading from overwritable ring
buffer were as easy as reading from normal ring buffer. However,
there's an obscure problem.
The following figure demonstrates a full overwritable ring buffer. In
this figure, the 'head' pointer points to the end of last record, and a
long record 'E' is pending. For a normal ring buffer, a 'tail' pointer
would have pointed to position (X), so kernel knows there's no more
space in the ring buffer. However, for an overwritable ring buffer,
kernel ignore the 'tail' pointer.
(X) head
. |
. V
+------+-------+----------+------+---+
|A....A|B.....B|C........C|D....D| |
+------+-------+----------+------+---+
Record 'A' is overwritten by event 'E':
head
|
V
+--+---+-------+----------+------+---+
|.E|..A|B.....B|C........C|D....D|E..|
+--+---+-------+----------+------+---+
Now tooling decides to read from this ring buffer. However, none of these
two natural positions, 'head' and the start of this ring buffer, are
pointing to the head of a record. Even the full ring buffer can be
accessed by tooling, it is unable to find a position to start decoding.
The first attempt tries to solve this problem AFAIK can be found from
[1]. It makes kernel to maintain 'tail' pointer: updates it when ring
buffer is half full. However, this approach introduces overhead to
fast path. Test result shows a 1% overhead [2]. In addition, this method
utilizes no more tham 50% records.
Another attempt can be found from [3], which allows putting the size of
an event at the end of each record. This approach allows tooling to find
records in a backward manner from 'head' pointer by reading size of a
record from its tail. However, because of alignment requirement, it
needs 8 bytes to record the size of a record, which is a huge waste. Its
performance is also not good, because more data need to be written.
This approach also introduces some extra branch instructions to fast
path.
'write_backward' is a better solution to this problem.
Following figure demonstrates the state of the overwritable ring buffer
when 'write_backward' is set before overwriting:
head
|
V
+---+------+----------+-------+------+
| |D....D|C........C|B.....B|A....A|
+---+------+----------+-------+------+
and after overwriting:
head
|
V
+---+------+----------+-------+---+--+
|..E|D....D|C........C|B.....B|A..|E.|
+---+------+----------+-------+---+--+
In each situation, 'head' points to the beginning of the newest record.
From this record, tooling can iterate over the full ring buffer and fetch
records one by one.
The only limitation that needs to be considered is back-to-back reading.
Due to the non-deterministic of user programs, it is impossible to ensure
the ring buffer keeps stable during reading. Consider an extreme situation:
tooling is scheduled out after reading record 'D', then a burst of events
come, eat up the whole ring buffer (one or multiple rounds). When the
tooling process comes back, reading after 'D' is incorrect now.
To prevent this problem, we need to find a way to ensure the ring buffer
is stable during reading. ioctl(PERF_EVENT_IOC_PAUSE_OUTPUT) is
suggested because its overhead is lower than
ioctl(PERF_EVENT_IOC_ENABLE).
By carefully verifying 'header' pointer, reader can avoid pausing the
ring-buffer. For example:
/* A union of all possible events */
union perf_event event;
p = head = perf_mmap__read_head();
while (true) {
/* copy header of next event */
fetch(&event.header, p, sizeof(event.header));
/* read 'head' pointer */
head = perf_mmap__read_head();
/* check overwritten: is the header good? */
if (!verify(sizeof(event.header), p, head))
break;
/* copy the whole event */
fetch(&event, p, event.header.size);
/* read 'head' pointer again */
head = perf_mmap__read_head();
/* is the whole event good? */
if (!verify(event.header.size, p, head))
break;
p += event.header.size;
}
However, the overhead is high because:
a) In-place decoding is not safe.
Copying-verifying-decoding is required.
b) Fetching 'head' pointer requires additional synchronization.
(From Alexei Starovoitov:
Even when this trick works, pause is needed for more than stability of
reading. When we collect the events into overwrite buffer we're waiting
for some other trigger (like all cpu utilization spike or just one cpu
running and all others are idle) and when it happens the buffer has
valuable info from the past. At this point new events are no longer
interesting and buffer should be paused, events read and unpaused until
next trigger comes.)
This patch utilizes event's default overflow_handler introduced
previously. perf_event_output_backward() is created as the default
overflow handler for backward ring buffers. To avoid extra overhead to
fast path, original perf_event_output() becomes __perf_event_output()
and marked '__always_inline'. In theory, there's no extra overhead
introduced to fast path.
Performance testing:
Calling 3000000 times of 'close(-1)', use gettimeofday() to check
duration. Use 'perf record -o /dev/null -e raw_syscalls:*' to capture
system calls. In ns.
Testing environment:
CPU : Intel(R) Core(TM) i7-4790 CPU @ 3.60GHz
Kernel : v4.5.0
MEAN STDVAR
BASE 800214.950 2853.083
PRE1 2253846.700 9997.014
PRE2 2257495.540 8516.293
POST 2250896.100 8933.921
Where 'BASE' is pure performance without capturing. 'PRE1' is test
result of pure 'v4.5.0' kernel. 'PRE2' is test result before this
patch. 'POST' is test result after this patch. See [4] for the detailed
experimental setup.
Considering the stdvar, this patch doesn't introduce performance
overhead to the fast path.
[1] http://lkml.iu.edu/hypermail/linux/kernel/1304.1/04584.html
[2] http://lkml.iu.edu/hypermail/linux/kernel/1307.1/00535.html
[3] http://lkml.iu.edu/hypermail/linux/kernel/1512.0/01265.html
[4] http://lkml.kernel.org/g/56F89DCD.1040202@huawei.com
Signed-off-by: Wang Nan <wangnan0@huawei.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Alexei Starovoitov <ast@kernel.org>
Cc: <acme@kernel.org>
Cc: <pi3orama@163.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Brendan Gregg <brendan.d.gregg@gmail.com>
Cc: He Kuang <hekuang@huawei.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Vince Weaver <vincent.weaver@maine.edu>
Cc: Zefan Li <lizefan@huawei.com>
Link: http://lkml.kernel.org/r/1459865478-53413-1-git-send-email-wangnan0@huawei.com
[ Fixed the changelog some more. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull USB driver fixes from Greg KH:
"Here are some small USB fixes for 4.6-rc4.
Mostly xhci fixes for reported issues, a UAS bug that has hit a number
of people, including stable tree users, and a few other minor things.
All have been in linux-next for a while with no reported issues"
* tag 'usb-4.6-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
usb: hcd: out of bounds access in for_each_companion
USB: uas: Add a new NO_REPORT_LUNS quirk
USB: uas: Limit qdepth at the scsi-host level
doc: usb: Fix typo in gadget_multi documentation
usb: host: xhci-plat: Make enum xhci_plat_type start at a non zero value
xhci: fix 10 second timeout on removal of PCI hotpluggable xhci controllers
usb: xhci: fix wild pointers in xhci_mem_cleanup
usb: host: xhci-plat: fix cannot work if R-Car Gen2/3 run on above 4GB phys
usb: host: xhci: add a new quirk XHCI_NO_64BIT_SUPPORT
xhci: resume USB 3 roothub first
usb: xhci: applying XHCI_PME_STUCK_QUIRK to Intel BXT B0 host
cdc-acm: fix crash if flushed with nothing buffered
Pull libnvdimm fixes from Ross Zwisler:
"Two fixes:
- Fix memcpy_from_pmem() to fallback to memcpy() for architectures
where CONFIG_ARCH_HAS_PMEM_API=n.
- Add a comment explaining why we write data twice when clearing
poison in pmem_do_bvec().
This has passed a boot test on an X86_32 config, which was the
architecture where issue #1 above was first noticed"
Dan Williams adds:
"We're giving this multi-maintainer setup a shot, so expect libnvdimm
pull requests from either Ross or I going forward"
* 'libnvdimm-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
libnvdimm, pmem: clarify the write+clear_poison+write flow
pmem: fix BUG() error in pmem.h:48 on X86_32
Pull mm gup cleanup from Ingo Molnar:
"This removes the ugly get-user-pages API hack, now that all upstream
code has been migrated to it"
("ugly" is putting it mildly. But it worked.. - Linus)
* 'mm-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
mm/gup: Remove the macro overload API migration helpers from the get_user*() APIs
Pull f2fs/fscrypto fixes from Jaegeuk Kim:
"In addition to f2fs/fscrypto fixes, I've added one patch which
prevents RCU mode lookup in d_revalidate, as Al mentioned.
These patches fix f2fs and fscrypto based on -rc3 bug fixes in ext4
crypto, which have not yet been fully propagated as follows.
- use of dget_parent and file_dentry to avoid crashes
- disallow RCU-mode lookup in d_invalidate
- disallow -ENOMEM in the core data encryption path"
* tag 'for-linus-4.6-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/jaegeuk/f2fs:
ext4/fscrypto: avoid RCU lookup in d_revalidate
fscrypto: don't let data integrity writebacks fail with ENOMEM
f2fs: use dget_parent and file_dentry in f2fs_file_open
fscrypto: use dget_parent() in fscrypt_d_revalidate()
A lot of seqfile users seem to be using things like %pK that uses the
credentials of the current process, but that is actually completely
wrong for filesystem interfaces.
The unix semantics for permission checking files is to check permissions
at _open_ time, not at read or write time, and that is not just a small
detail: passing off stdin/stdout/stderr to a suid application and making
the actual IO happen in privileged context is a classic exploit
technique.
So if we want to be able to look at permissions at read time, we need to
use the file open credentials, not the current ones. Normal file
accesses can just use "f_cred" (or any of the helper functions that do
that, like file_ns_capable()), but the seqfile interfaces do not have
any such options.
It turns out that seq_file _does_ save away the user_ns information of
the file, though. Since user_ns is just part of the full credential
information, replace that special case with saving off the cred pointer
instead, and suddenly seq_file has all the permission information it
needs.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
After 'commit fc0c202813 ("x86, pmem: use memcpy_mcsafe()
for memcpy_from_pmem()")', probing a PMEM device hits the BUG()
error below on X86_32 kernel.
kernel BUG at include/linux/pmem.h:48!
memcpy_from_pmem() calls arch_memcpy_from_pmem(), which is
unimplemented since CONFIG_ARCH_HAS_PMEM_API is undefined on
X86_32.
Fix the BUG() error by adding default_memcpy_from_pmem().
Acked-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Toshi Kani <toshi.kani@hpe.com>
Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Ross Zwisler <ross.zwisler@linux.intel.com>
This patch fixes the issue introduced by the ext4 crypto fix in a same manner.
For F2FS, however, we flush the pending IOs and wait for a while to acquire free
memory.
Fixes: c9af28fdd4 ("ext4 crypto: don't let data integrity writebacks fail with ENOMEM")
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Pull networking fixes from David Miller:
1) Stale SKB data pointer access across pskb_may_pull() calls in L2TP,
from Haishuang Yan.
2) Fix multicast frame handling in mac80211 AP code, from Felix
Fietkau.
3) mac80211 station hashtable insert errors not handled properly, fix
from Johannes Berg.
4) Fix TX descriptor count limit handling in e1000, from Alexander
Duyck.
5) Revert a buggy netdev refcount fix in netpoll, from Bjorn Helgaas.
6) Must assign rtnl_link_ops of the device before registering it, fix
in ip6_tunnel from Thadeu Lima de Souza Cascardo.
7) Memory leak fix in tc action net exit, from WANG Cong.
8) Add missing AF_KCM entries to name tables, from Dexuan Cui.
9) Fix regression in GRE handling of csums wrt. FOU, from Alexander
Duyck.
10) Fix memory allocation alignment and congestion map corruption in
RDS, from Shamir Rabinovitch.
11) Fix default qdisc regression in tuntap driver, from Jason Wang.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (44 commits)
bridge, netem: mark mailing lists as moderated
tuntap: restore default qdisc
mpls: find_outdev: check for err ptr in addition to NULL check
ipv6: Count in extension headers in skb->network_header
RDS: fix congestion map corruption for PAGE_SIZE > 4k
RDS: memory allocated must be align to 8
GRE: Disable segmentation offloads w/ CSUM and we are encapsulated via FOU
net: add the AF_KCM entries to family name tables
MAINTAINERS: intel-wired-lan list is moderated
lib/test_bpf: Add additional BPF_ADD tests
lib/test_bpf: Add test to check for result of 32-bit add that overflows
lib/test_bpf: Add tests for unsigned BPF_JGT
lib/test_bpf: Fix JMP_JSET tests
VSOCK: Detach QP check should filter out non matching QPs.
stmmac: fix adjust link call in case of a switch is attached
af_packet: tone down the Tx-ring unsupported spew.
net_sched: fix a memory leak in tc action
samples/bpf: Enable powerpc support
samples/bpf: Use llc in PATH, rather than a hardcoded value
samples/bpf: Fix build breakage with map_perf_test_user.c
...
Pull IOMMU fixes from Joerg Roedel:
- compile-time fixes (warnings and failures)
- a bug in iommu core code which could cause the group->domain pointer
to be falsly cleared
- fix in scatterlist handling of the ARM common DMA-API code
- stall detection fix for the Rockchip IOMMU driver
* tag 'iommu-fixes-v4.6-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/joro/iommu:
iommu/vt-d: Silence an uninitialized variable warning
iommu/rockchip: Fix "is stall active" check
iommu: Don't overwrite domain pointer when there is no default_domain
iommu/dma: Restore scatterlist offsets correctly
iommu: provide of_xlate pointer unconditionally
Pull ext4 bugfixes from Ted Ts'o:
"These changes contains a fix for overlayfs interacting with some
(badly behaved) dentry code in various file systems. These have been
reviewed by Al and the respective file system mtinainers and are going
through the ext4 tree for convenience.
This also has a few ext4 encryption bug fixes that were discovered in
Android testing (yes, we will need to get these sync'ed up with the
fs/crypto code; I'll take care of that). It also has some bug fixes
and a change to ignore the legacy quota options to allow for xfstests
regression testing of ext4's internal quota feature and to be more
consistent with how xfs handles this case"
* tag 'ext4_for_linus_stable' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/ext4:
ext4: ignore quota mount options if the quota feature is enabled
ext4 crypto: fix some error handling
ext4: avoid calling dquot_get_next_id() if quota is not enabled
ext4: retry block allocation for failed DIO and DAX writes
ext4: add lockdep annotations for i_data_sem
ext4: allow readdir()'s of large empty directories to be interrupted
btrfs: fix crash/invalid memory access on fsync when using overlayfs
ext4 crypto: use dget_parent() in ext4_d_revalidate()
ext4: use file_dentry()
ext4: use dget_parent() in ext4_file_open()
nfs: use file_dentry()
fs: add file_dentry()
ext4 crypto: don't let data integrity writebacks fail with ENOMEM
ext4: check if in-inode xattr is corrupted in ext4_expand_extra_isize_ea()
This patch fixes an issue I found in which we were dropping frames if we
had enabled checksums on GRE headers that were encapsulated by either FOU
or GUE. Without this patch I was barely able to get 1 Gb/s of throughput.
With this patch applied I am now at least getting around 6 Gb/s.
The issue is due to the fact that with FOU or GUE applied we do not provide
a transport offset pointing to the GRE header, nor do we offload it in
software as the GRE header is completely skipped by GSO and treated like a
VXLAN or GENEVE type header. As such we need to prevent the stack from
generating it and also prevent GRE from generating it via any interface we
create.
Fixes: c3483384ee ("gro: Allow tunnel stacking in the case of FOU/GUE")
Signed-off-by: Alexander Duyck <aduyck@mirantis.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The pkeys changes brought about a truly hideous set of macros in:
cde70140fe ("mm/gup: Overload get_user_pages() functions")
... which macros are (ab-)using the fact that __VA_ARGS__ can be used
to shift parameter positions in macro arguments without breaking the
build and so can be used to call separate C functions depending on
the number of arguments of the macro.
This allowed easy migration of these 3 GUP APIs, as both these variants
worked at the C level:
old:
ret = get_user_pages(current, current->mm, address, 1, 1, 0, &page, NULL);
new:
ret = get_user_pages(address, 1, 1, 0, &page, NULL);
... while we also generated a (functionally harmless but noticeable) build
time warning if the old API was used. As there are over 300 uses of these
APIs, this trick eased the migration of the API and avoided excessive
migration pain in linux-next.
Now, with its work done, get rid of all of that complication and ugliness:
3 files changed, 16 insertions(+), 140 deletions(-)
... where the linecount of the migration hack was further inflated by the
fact that there are NOMMU variants of these GUP APIs as well.
Much of the conversion was done in linux-next over the past couple of months,
and Linus recently removed all remaining old API uses from the upstream tree
in the following upstrea commit:
cb107161df ("Convert straggling drivers to new six-argument get_user_pages()")
There was one more old-API usage in mm/gup.c, in the CONFIG_HAVE_GENERIC_RCU_GUP
code path that ARM, ARM64 and PowerPC uses.
After this commit any old API usage will break the build.
[ Also fixed a PowerPC/HAVE_GENERIC_RCU_GUP warning reported by Stephen Rothwell. ]
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Dave Hansen <dave@sr71.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Cc: linux-mm@kvack.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull KVM fixes from Paolo Bonzini:
"Miscellaneous bugfixes.
The ARM and s390 fixes are for new regressions from the merge window,
others are usual stable material"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
compiler-gcc: disable -ftracer for __noclone functions
kvm: x86: make lapic hrtimer pinned
s390/mm/kvm: fix mis-merge in gmap handling
kvm: set page dirty only if page has been writable
KVM: x86: reduce default value of halt_poll_ns parameter
KVM: Hyper-V: do not do hypercall userspace exits if SynIC is disabled
KVM: x86: Inject pending interrupt even if pending nmi exist
arm64: KVM: Register CPU notifiers when the kernel runs at HYP
arm64: kvm: 4.6-rc1: Fix VTCR_EL2 VS setting
-ftracer can duplicate asm blocks causing compilation to fail in
noclone functions. For example, KVM declares a global variable
in an asm like
asm("2: ... \n
.pushsection data \n
.global vmx_return \n
vmx_return: .long 2b");
and -ftracer causes a double declaration.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Michal Marek <mmarek@suse.cz>
Cc: stable@vger.kernel.org
Cc: kvm@vger.kernel.org
Reported-by: Linda Walsh <lkml@tlinx.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
iommu drivers that support the standard DT bindings use a of_xlate
callback pointer, but that is only part of struct iommu_ops when
CONFIG_OF_IOMMU is enabled, leading to build errors in randconfig
builds when that is not provided:
drivers/iommu/mtk_iommu.c:497:2: error: unknown field 'of_xlate' specified in initializer
.of_xlate = mtk_iommu_of_xlate,
^
drivers/iommu/mtk_iommu.c:497:14: error: initialization from incompatible pointer type [-Werror=incompatible-pointer-types]
.of_xlate = mtk_iommu_of_xlate,
^~~~~~~~~~~~~~~~~~
drivers/iommu/mtk_iommu.c:497:14: note: (near initialization for 'mtk_iommu_ops.domain_get_attr')
We can work around it by adding more #ifdefs in each driver, but
it seems nicer to just allow setting the pointer even if it is
unused. This makes the driver code look nicer, and it gives better
compile-time coverage when test building on other architectures.
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 0df4fabe20 ("iommu/mediatek: Add mt8173 IOMMU driver")
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Merge PAGE_CACHE_SIZE removal patches from Kirill Shutemov:
"PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time
ago with promise that one day it will be possible to implement page
cache with bigger chunks than PAGE_SIZE.
This promise never materialized. And unlikely will.
Let's stop pretending that pages in page cache are special. They are
not.
The first patch with most changes has been done with coccinelle. The
second is manual fixups on top.
The third patch removes macros definition"
[ I was planning to apply this just before rc2, but then I spaced out,
so here it is right _after_ rc2 instead.
As Kirill suggested as a possibility, I could have decided to only
merge the first two patches, and leave the old interfaces for
compatibility, but I'd rather get it all done and any out-of-tree
modules and patches can trivially do the converstion while still also
working with older kernels, so there is little reason to try to
maintain the redundant legacy model. - Linus ]
* PAGE_CACHE_SIZE-removal:
mm: drop PAGE_CACHE_* and page_cache_{get,release} definition
mm, fs: remove remaining PAGE_CACHE_* and page_cache_{get,release} usage
mm, fs: get rid of PAGE_CACHE_* and page_cache_{get,release} macros
PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} macros were introduced *long* time
ago with promise that one day it will be possible to implement page
cache with bigger chunks than PAGE_SIZE.
This promise never materialized. And unlikely will.
We have many places where PAGE_CACHE_SIZE assumed to be equal to
PAGE_SIZE. And it's constant source of confusion on whether
PAGE_CACHE_* or PAGE_* constant should be used in a particular case,
especially on the border between fs and mm.
Global switching to PAGE_CACHE_SIZE != PAGE_SIZE would cause to much
breakage to be doable.
Let's stop pretending that pages in page cache are special. They are
not.
The changes are pretty straight-forward:
- <foo> << (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>;
- <foo> >> (PAGE_CACHE_SHIFT - PAGE_SHIFT) -> <foo>;
- PAGE_CACHE_{SIZE,SHIFT,MASK,ALIGN} -> PAGE_{SIZE,SHIFT,MASK,ALIGN};
- page_cache_get() -> get_page();
- page_cache_release() -> put_page();
This patch contains automated changes generated with coccinelle using
script below. For some reason, coccinelle doesn't patch header files.
I've called spatch for them manually.
The only adjustment after coccinelle is revert of changes to
PAGE_CAHCE_ALIGN definition: we are going to drop it later.
There are few places in the code where coccinelle didn't reach. I'll
fix them manually in a separate patch. Comments and documentation also
will be addressed with the separate patch.
virtual patch
@@
expression E;
@@
- E << (PAGE_CACHE_SHIFT - PAGE_SHIFT)
+ E
@@
expression E;
@@
- E >> (PAGE_CACHE_SHIFT - PAGE_SHIFT)
+ E
@@
@@
- PAGE_CACHE_SHIFT
+ PAGE_SHIFT
@@
@@
- PAGE_CACHE_SIZE
+ PAGE_SIZE
@@
@@
- PAGE_CACHE_MASK
+ PAGE_MASK
@@
expression E;
@@
- PAGE_CACHE_ALIGN(E)
+ PAGE_ALIGN(E)
@@
expression E;
@@
- page_cache_get(E)
+ get_page(E)
@@
expression E;
@@
- page_cache_release(E)
+ put_page(E)
Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull core kernel fixes from Ingo Molnar:
"This contains the nohz/atomic cleanup/fix for the fetch_or() ugliness
you noted during the original nohz pull request, plus there's also
misc fixes:
- fix liblockdep build bug
- fix uapi header build bug
- print more lockdep hash collision info to help debug recent reports
of hash collisions
- update MAINTAINERS email address"
* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
MAINTAINERS: Update my email address
locking/lockdep: Print chain_key collision information
uapi/linux/stddef.h: Provide __always_inline to userspace headers
tools/lib/lockdep: Fix unsupported 'basename -s' in run_tests.sh
locking/atomic, sched: Unexport fetch_or()
timers/nohz: Convert tick dependency mask to atomic_t
locking/atomic: Introduce atomic_fetch_or()
Pull configfs fix from Christoph Hellwig:
"A trivial fix to the recently introduced binary attribute helper
macros"
* tag 'configfs-for-linus-2' of git://git.infradead.org/users/hch/configfs:
configfs: fix CONFIGFS_BIN_ATTR_[RW]O definitions
Pull networking fixes from David Miller:
1) Missing device reference in IPSEC input path results in crashes
during device unregistration. From Subash Abhinov Kasiviswanathan.
2) Per-queue ISR register writes not being done properly in macb
driver, from Cyrille Pitchen.
3) Stats accounting bugs in bcmgenet, from Patri Gynther.
4) Lightweight tunnel's TTL and TOS were swapped in netlink dumps, from
Quentin Armitage.
5) SXGBE driver has off-by-one in probe error paths, from Rasmus
Villemoes.
6) Fix race in save/swap/delete options in netfilter ipset, from
Vishwanath Pai.
7) Ageing time of bridge not set properly when not operating over a
switchdev device. Fix from Haishuang Yan.
8) Fix GRO regression wrt nested FOU/GUE based tunnels, from Alexander
Duyck.
9) IPV6 UDP code bumps wrong stats, from Eric Dumazet.
10) FEC driver should only access registers that actually exist on the
given chipset, fix from Fabio Estevam.
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (73 commits)
net: mvneta: fix changing MTU when using per-cpu processing
stmmac: fix MDIO settings
Revert "stmmac: Fix 'eth0: No PHY found' regression"
stmmac: fix TX normal DESC
net: mvneta: use cache_line_size() to get cacheline size
net: mvpp2: use cache_line_size() to get cacheline size
net: mvpp2: fix maybe-uninitialized warning
tun, bpf: fix suspicious RCU usage in tun_{attach, detach}_filter
net: usb: cdc_ncm: adding Telit LE910 V2 mobile broadband card
rtnl: fix msg size calculation in if_nlmsg_size()
fec: Do not access unexisting register in Coldfire
net: mvneta: replace MVNETA_CPU_D_CACHE_LINE_SIZE with L1_CACHE_BYTES
net: mvpp2: replace MVPP2_CPU_D_CACHE_LINE_SIZE with L1_CACHE_BYTES
net: dsa: mv88e6xxx: Clear the PDOWN bit on setup
net: dsa: mv88e6xxx: Introduce _mv88e6xxx_phy_page_{read, write}
bpf: make padding in bpf_tunnel_key explicit
ipv6: udp: fix UDP_MIB_IGNOREDMULTI updates
bnxt_en: Fix ethtool -a reporting.
bnxt_en: Fix typo in bnxt_hwrm_set_pause_common().
bnxt_en: Implement proper firmware message padding.
...
Initially the phy_bus_name was added to manipulate the
driver name but it was recently just used to manage the
fixed-link and then to take some decision at run-time.
So the patch uses the is_pseudo_fixed_link and removes
the phy_bus_name variable not necessary anymore.
The driver can manage the mdio registration by using phy-handle,
dwmac-mdio and own parameter e.g. snps,phy-addr.
This patch takes care about all these possible configurations
and fixes the mdio registration in case of there is a real
transceiver or a switch (that needs to be managed by using
fixed-link).
Signed-off-by: Giuseppe Cavallaro <peppe.cavallaro@st.com>
Reviewed-by: Andreas Färber <afaerber@suse.de>
Tested-by: Frank Schäfer <fschaefer.oss@googlemail.com>
Cc: Gabriel Fernandez <gabriel.fernandez@linaro.org>
Cc: Dinh Nguyen <dinh.linux@gmail.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Phil Reid <preid@electromag.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Sasha Levin reported a suspicious rcu_dereference_protected() warning
found while fuzzing with trinity that is similar to this one:
[ 52.765684] net/core/filter.c:2262 suspicious rcu_dereference_protected() usage!
[ 52.765688] other info that might help us debug this:
[ 52.765695] rcu_scheduler_active = 1, debug_locks = 1
[ 52.765701] 1 lock held by a.out/1525:
[ 52.765704] #0: (rtnl_mutex){+.+.+.}, at: [<ffffffff816a64b7>] rtnl_lock+0x17/0x20
[ 52.765721] stack backtrace:
[ 52.765728] CPU: 1 PID: 1525 Comm: a.out Not tainted 4.5.0+ #264
[...]
[ 52.765768] Call Trace:
[ 52.765775] [<ffffffff813e488d>] dump_stack+0x85/0xc8
[ 52.765784] [<ffffffff810f2fa5>] lockdep_rcu_suspicious+0xd5/0x110
[ 52.765792] [<ffffffff816afdc2>] sk_detach_filter+0x82/0x90
[ 52.765801] [<ffffffffa0883425>] tun_detach_filter+0x35/0x90 [tun]
[ 52.765810] [<ffffffffa0884ed4>] __tun_chr_ioctl+0x354/0x1130 [tun]
[ 52.765818] [<ffffffff8136fed0>] ? selinux_file_ioctl+0x130/0x210
[ 52.765827] [<ffffffffa0885ce3>] tun_chr_ioctl+0x13/0x20 [tun]
[ 52.765834] [<ffffffff81260ea6>] do_vfs_ioctl+0x96/0x690
[ 52.765843] [<ffffffff81364af3>] ? security_file_ioctl+0x43/0x60
[ 52.765850] [<ffffffff81261519>] SyS_ioctl+0x79/0x90
[ 52.765858] [<ffffffff81003ba2>] do_syscall_64+0x62/0x140
[ 52.765866] [<ffffffff817d563f>] entry_SYSCALL64_slow_path+0x25/0x25
Same can be triggered with PROVE_RCU (+ PROVE_RCU_REPEATEDLY) enabled
from tun_attach_filter() when user space calls ioctl(tun_fd, TUN{ATTACH,
DETACH}FILTER, ...) for adding/removing a BPF filter on tap devices.
Since the fix in f91ff5b9ff ("net: sk_{detach|attach}_filter() rcu
fixes") sk_attach_filter()/sk_detach_filter() now dereferences the
filter with rcu_dereference_protected(), checking whether socket lock
is held in control path.
Since its introduction in 9940516259 ("tun: socket filter support"),
tap filters are managed under RTNL lock from __tun_chr_ioctl(). Thus the
sock_owned_by_user(sk) doesn't apply in this specific case and therefore
triggers the false positive.
Extend the BPF API with __sk_attach_filter()/__sk_detach_filter() pair
that is used by tap filters and pass in lockdep_rtnl_is_held() for the
rcu_dereference_protected() checks instead.
Reported-by: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch functionally reverts:
5fd7a09cfb ("atomic: Export fetch_or()")
During the merge Linus observed that the generic version of fetch_or()
was messy:
" This makes the ugly "fetch_or()" macro that the scheduler used
internally a new generic helper, and does a bad job at it. "
e23604edac Merge branch 'timers-nohz-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Now that we have introduced atomic_fetch_or(), fetch_or() is only used
by the scheduler in order to deal with thread_info flags which type
can vary across architectures.
Lets confine fetch_or() back to the scheduler so that we encourage
future users to use the more robust and well typed atomic_t version
instead.
While at it, fetch_or() gets robustified, pasting improvements from a
previous patch by Ingo Molnar that avoids needless expression
re-evaluations in the loop.
Reported-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/1458830281-4255-4-git-send-email-fweisbec@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
This fix adds a new reference counter (ref_netlink) for the struct ip_set.
The other reference counter (ref) can be swapped out by ip_set_swap and we
need a separate counter to keep track of references for netlink events
like dump. Using the same ref counter for dump causes a race condition
which can be demonstrated by the following script:
ipset create hash_ip1 hash:ip family inet hashsize 1024 maxelem 500000 \
counters
ipset create hash_ip2 hash:ip family inet hashsize 300000 maxelem 500000 \
counters
ipset create hash_ip3 hash:ip family inet hashsize 1024 maxelem 500000 \
counters
ipset save &
ipset swap hash_ip3 hash_ip2
ipset destroy hash_ip3 /* will crash the machine */
Swap will exchange the values of ref so destroy will see ref = 0 instead of
ref = 1. With this fix in place swap will not succeed because ipset save
still has ref_netlink on the set (ip_set_swap doesn't swap ref_netlink).
Both delete and swap will error out if ref_netlink != 0 on the set.
Note: The changes to *_head functions is because previously we would
increment ref whenever we called these functions, we don't do that
anymore.
Reviewed-by: Joshua Hunt <johunt@akamai.com>
Signed-off-by: Vishwanath Pai <vpai@akamai.com>
Signed-off-by: Jozsef Kadlecsik <kadlec@blackhole.kfki.hu>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Pull Ceph updates from Sage Weil:
"There is quite a bit here, including some overdue refactoring and
cleanup on the mon_client and osd_client code from Ilya, scattered
writeback support for CephFS and a pile of bug fixes from Zheng, and a
few random cleanups and fixes from others"
[ I already decided not to pull this because of it having been rebased
recently, but ended up changing my mind after all. Next time I'll
really hold people to it. Oh well. - Linus ]
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client: (34 commits)
libceph: use KMEM_CACHE macro
ceph: use kmem_cache_zalloc
rbd: use KMEM_CACHE macro
ceph: use lookup request to revalidate dentry
ceph: kill ceph_get_dentry_parent_inode()
ceph: fix security xattr deadlock
ceph: don't request vxattrs from MDS
ceph: fix mounting same fs multiple times
ceph: remove unnecessary NULL check
ceph: avoid updating directory inode's i_size accidentally
ceph: fix race during filling readdir cache
libceph: use sizeof_footer() more
ceph: kill ceph_empty_snapc
ceph: fix a wrong comparison
ceph: replace CURRENT_TIME by current_fs_time()
ceph: scattered page writeback
libceph: add helper that duplicates last extent operation
libceph: enable large, variable-sized OSD requests
libceph: osdc->req_mempool should be backed by a slab pool
libceph: make r_request msg_size calculation clearer
...
This series fixes bugs in nfs and ext4 due to 4bacc9c923 ("overlayfs:
Make f_path always point to the overlay and f_inode to the underlay").
Regular files opened on overlayfs will result in the file being opened on
the underlying filesystem, while f_path points to the overlayfs
mount/dentry.
This confuses filesystems which get the dentry from struct file and assume
it's theirs.
Add a new helper, file_dentry() [*], to get the filesystem's own dentry
from the file. This checks file->f_path.dentry->d_flags against
DCACHE_OP_REAL, and returns file->f_path.dentry if DCACHE_OP_REAL is not
set (this is the common, non-overlayfs case).
In the uncommon case it will call into overlayfs's ->d_real() to get the
underlying dentry, matching file_inode(file).
The reason we need to check against the inode is that if the file is copied
up while being open, d_real() would return the upper dentry, while the open
file comes from the lower dentry.
[*] If possible, it's better simply to use file_inode() instead.
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Tested-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: Trond Myklebust <trond.myklebust@primarydata.com>
Cc: <stable@vger.kernel.org> # v4.2
Cc: David Howells <dhowells@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Daniel Axtens <dja@axtens.net>
Pull NTB bug fixes from Jon Mason:
"NTB bug fixes for tasklet from spinning forever, link errors,
translation window setup, NULL ptr dereference, and ntb-perf errors.
Also, a modification to the driver API that makes _addr functions
optional"
* tag 'ntb-4.6' of git://github.com/jonmason/ntb:
NTB: Remove _addr functions from ntb_hw_amd
NTB: Make _addr functions optional in the API
NTB: Fix incorrect clean up routine in ntb_perf
NTB: Fix incorrect return check in ntb_perf
ntb: fix possible NULL dereference
ntb: add missing setup of translation window
ntb: stop link work when we do not have memory
ntb: stop tasklet from spinning forever during shutdown.
ntb: perf test: fix address space confusion
Implement the stack depot and provide CONFIG_STACKDEPOT. Stack depot
will allow KASAN store allocation/deallocation stack traces for memory
chunks. The stack traces are stored in a hash table and referenced by
handles which reside in the kasan_alloc_meta and kasan_free_meta
structures in the allocated memory chunks.
IRQ stack traces are cut below the IRQ entry point to avoid unnecessary
duplication.
Right now stackdepot support is only enabled in SLAB allocator. Once
KASAN features in SLAB are on par with those in SLUB we can switch SLUB
to stackdepot as well, thus removing the dependency on SLUB stack
bookkeeping, which wastes a lot of memory.
This patch is based on the "mm: kasan: stack depots" patch originally
prepared by Dmitry Chernenkov.
Joonsoo has said that he plans to reuse the stackdepot code for the
mm/page_owner.c debugging facility.
[akpm@linux-foundation.org: s/depot_stack_handle/depot_stack_handle_t]
[aryabinin@virtuozzo.com: comment style fixes]
Signed-off-by: Alexander Potapenko <glider@google.com>
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Andrey Konovalov <adech.fo@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Konstantin Serebryany <kcc@google.com>
Cc: Dmitry Chernenkov <dmitryc@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
KASAN needs to know whether the allocation happens in an IRQ handler.
This lets us strip everything below the IRQ entry point to reduce the
number of unique stack traces needed to be stored.
Move the definition of __irq_entry to <linux/interrupt.h> so that the
users don't need to pull in <linux/ftrace.h>. Also introduce the
__softirq_entry macro which is similar to __irq_entry, but puts the
corresponding functions to the .softirqentry.text section.
Signed-off-by: Alexander Potapenko <glider@google.com>
Acked-by: Steven Rostedt <rostedt@goodmis.org>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Andrey Konovalov <adech.fo@gmail.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Konstantin Serebryany <kcc@google.com>
Cc: Dmitry Chernenkov <dmitryc@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
"oom, oom_reaper: disable oom_reaper for oom_kill_allocating_task" tried
to protect oom_reaper_list using MMF_OOM_KILLED flag. But we can do it
by simply checking tsk->oom_reaper_list != NULL.
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Tetsuo has reported that oom_kill_allocating_task=1 will cause
oom_reaper_list corruption because oom_kill_process doesn't follow
standard OOM exclusion (aka ignores TIF_MEMDIE) and allows to enqueue
the same task multiple times - e.g. by sacrificing the same child
multiple times.
This patch fixes the issue by introducing a new MMF_OOM_KILLED mm flag
which is set in oom_kill_process atomically and oom reaper is disabled
if the flag was already set.
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reported-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: David Rientjes <rientjes@google.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
wake_oom_reaper has allowed only 1 oom victim to be queued. The main
reason for that was the simplicity as other solutions would require some
way of queuing. The current approach is racy and that was deemed
sufficient as the oom_reaper is considered a best effort approach to
help with oom handling when the OOM victim cannot terminate in a
reasonable time. The race could lead to missing an oom victim which can
get stuck
out_of_memory
wake_oom_reaper
cmpxchg // OK
oom_reaper
oom_reap_task
__oom_reap_task
oom_victim terminates
atomic_inc_not_zero // fail
out_of_memory
wake_oom_reaper
cmpxchg // fails
task_to_reap = NULL
This race requires 2 OOM invocations in a short time period which is not
very likely but certainly not impossible. E.g. the original victim
might have not released a lot of memory for some reason.
The situation would improve considerably if wake_oom_reaper used a more
robust queuing. This is what this patch implements. This means adding
oom_reaper_list list_head into task_struct (eat a hole before embeded
thread_struct for that purpose) and a oom_reaper_lock spinlock for
queuing synchronization. wake_oom_reaper will then add the task on the
queue and oom_reaper will dequeue it.
Signed-off-by: Michal Hocko <mhocko@suse.com>
Cc: Vladimir Davydov <vdavydov@virtuozzo.com>
Cc: Andrea Argangeli <andrea@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
When oom_reaper manages to unmap all the eligible vmas there shouldn't
be much of the freable memory held by the oom victim left anymore so it
makes sense to clear the TIF_MEMDIE flag for the victim and allow the
OOM killer to select another task.
The lack of TIF_MEMDIE also means that the victim cannot access memory
reserves anymore but that shouldn't be a problem because it would get
the access again if it needs to allocate and hits the OOM killer again
due to the fatal_signal_pending resp. PF_EXITING check. We can safely
hide the task from the OOM killer because it is clearly not a good
candidate anymore as everyhing reclaimable has been torn down already.
This patch will allow to cap the time an OOM victim can keep TIF_MEMDIE
and thus hold off further global OOM killer actions granted the oom
reaper is able to take mmap_sem for the associated mm struct. This is
not guaranteed now but further steps should make sure that mmap_sem for
write should be blocked killable which will help to reduce such a lock
contention. This is not done by this patch.
Note that exit_oom_victim might be called on a remote task from
__oom_reap_task now so we have to check and clear the flag atomically
otherwise we might race and underflow oom_victims or wake up waiters too
early.
Signed-off-by: Michal Hocko <mhocko@suse.com>
Suggested-by: Johannes Weiner <hannes@cmpxchg.org>
Suggested-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Andrea Argangeli <andrea@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This patch (of 5):
This is based on the idea from Mel Gorman discussed during LSFMM 2015
and independently brought up by Oleg Nesterov.
The OOM killer currently allows to kill only a single task in a good
hope that the task will terminate in a reasonable time and frees up its
memory. Such a task (oom victim) will get an access to memory reserves
via mark_oom_victim to allow a forward progress should there be a need
for additional memory during exit path.
It has been shown (e.g. by Tetsuo Handa) that it is not that hard to
construct workloads which break the core assumption mentioned above and
the OOM victim might take unbounded amount of time to exit because it
might be blocked in the uninterruptible state waiting for an event (e.g.
lock) which is blocked by another task looping in the page allocator.
This patch reduces the probability of such a lockup by introducing a
specialized kernel thread (oom_reaper) which tries to reclaim additional
memory by preemptively reaping the anonymous or swapped out memory owned
by the oom victim under an assumption that such a memory won't be needed
when its owner is killed and kicked from the userspace anyway. There is
one notable exception to this, though, if the OOM victim was in the
process of coredumping the result would be incomplete. This is
considered a reasonable constrain because the overall system health is
more important than debugability of a particular application.
A kernel thread has been chosen because we need a reliable way of
invocation so workqueue context is not appropriate because all the
workers might be busy (e.g. allocating memory). Kswapd which sounds
like another good fit is not appropriate as well because it might get
blocked on locks during reclaim as well.
oom_reaper has to take mmap_sem on the target task for reading so the
solution is not 100% because the semaphore might be held or blocked for
write but the probability is reduced considerably wrt. basically any
lock blocking forward progress as described above. In order to prevent
from blocking on the lock without any forward progress we are using only
a trylock and retry 10 times with a short sleep in between. Users of
mmap_sem which need it for write should be carefully reviewed to use
_killable waiting as much as possible and reduce allocations requests
done with the lock held to absolute minimum to reduce the risk even
further.
The API between oom killer and oom reaper is quite trivial.
wake_oom_reaper updates mm_to_reap with cmpxchg to guarantee only
NULL->mm transition and oom_reaper clear this atomically once it is done
with the work. This means that only a single mm_struct can be reaped at
the time. As the operation is potentially disruptive we are trying to
limit it to the ncessary minimum and the reaper blocks any updates while
it operates on an mm. mm_struct is pinned by mm_count to allow parallel
exit_mmap and a race is detected by atomic_inc_not_zero(mm_users).
Signed-off-by: Michal Hocko <mhocko@suse.com>
Suggested-by: Oleg Nesterov <oleg@redhat.com>
Suggested-by: Mel Gorman <mgorman@suse.de>
Acked-by: Mel Gorman <mgorman@suse.de>
Acked-by: David Rientjes <rientjes@google.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Andrea Argangeli <andrea@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
