openwrt/staging/blogic.git
6 years agoproc: use set_puts() at /proc/*/wchan
Alexey Dobriyan [Tue, 10 Apr 2018 23:31:34 +0000 (16:31 -0700)]
proc: use set_puts() at /proc/*/wchan

Link: http://lkml.kernel.org/r/20180217072011.GB16074@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: check permissions earlier for /proc/*/wchan
Alexey Dobriyan [Tue, 10 Apr 2018 23:31:30 +0000 (16:31 -0700)]
proc: check permissions earlier for /proc/*/wchan

get_wchan() accesses stack page before permissions are checked, let's
not play this game.

Link: http://lkml.kernel.org/r/20180217071923.GA16074@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: replace seq_printf by seq_put_smth to speed up /proc/pid/status
Andrei Vagin [Tue, 10 Apr 2018 23:31:26 +0000 (16:31 -0700)]
proc: replace seq_printf by seq_put_smth to speed up /proc/pid/status

seq_printf() works slower than seq_puts, seq_puts, etc.

== test_proc.c
int main(int argc, char **argv)
{
int n, i, fd;
char buf[16384];

n = atoi(argv[1]);
for (i = 0; i < n; i++) {
fd = open(argv[2], O_RDONLY);
if (fd < 0)
return 1;
if (read(fd, buf, sizeof(buf)) <= 0)
return 1;
close(fd);
}

return 0;
}
==

$ time ./test_proc  1000000 /proc/1/status

== Before path ==
real 0m5.171s
user 0m0.328s
sys 0m4.783s

== After patch ==
real 0m4.761s
user 0m0.334s
sys 0m4.366s

Link: http://lkml.kernel.org/r/20180212074931.7227-4-avagin@openvz.org
Signed-off-by: Andrei Vagin <avagin@openvz.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: optimize single-symbol delimiters to spead up seq_put_decimal_ull
Andrei Vagin [Tue, 10 Apr 2018 23:31:23 +0000 (16:31 -0700)]
proc: optimize single-symbol delimiters to spead up seq_put_decimal_ull

A delimiter is a string which is printed before a number.  A
syngle-symbol delimiters can be printed by set_putc() and this works
faster than printing by set_puts().

== test_proc.c

int main(int argc, char **argv)
{
int n, i, fd;
char buf[16384];

n = atoi(argv[1]);
for (i = 0; i < n; i++) {
fd = open(argv[2], O_RDONLY);
if (fd < 0)
return 1;
if (read(fd, buf, sizeof(buf)) <= 0)
return 1;
close(fd);
}

return 0;
}
==

$ time ./test_proc  1000000 /proc/1/stat

== Before patch ==
  real 0m3.820s
  user 0m0.337s
  sys 0m3.394s

== After patch ==
  real 0m3.110s
  user 0m0.324s
  sys 0m2.700s

Link: http://lkml.kernel.org/r/20180212074931.7227-3-avagin@openvz.org
Signed-off-by: Andrei Vagin <avagin@openvz.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: replace seq_printf on seq_putc to speed up /proc/pid/smaps
Andrei Vagin [Tue, 10 Apr 2018 23:31:19 +0000 (16:31 -0700)]
proc: replace seq_printf on seq_putc to speed up /proc/pid/smaps

seq_putc() works much faster than seq_printf()

== Before patch ==
  $ time python test_smaps.py
  real    0m3.828s
  user    0m0.413s
  sys     0m3.408s

== After patch ==
  $ time python test_smaps.py
  real 0m3.405s
  user 0m0.401s
  sys 0m3.003s

== Before patch ==
-   75.51%     4.62%  python   [kernel.kallsyms]    [k] show_smap.isra.33
   - 70.88% show_smap.isra.33
      + 24.82% seq_put_decimal_ull_aligned
      + 19.78% __walk_page_range
      + 12.74% seq_printf
      + 11.08% show_map_vma.isra.23
      + 1.68% seq_puts

== After patch ==
-   69.16%     5.70%  python   [kernel.kallsyms]    [k] show_smap.isra.33
   - 63.46% show_smap.isra.33
      + 25.98% seq_put_decimal_ull_aligned
      + 20.90% __walk_page_range
      + 12.60% show_map_vma.isra.23
        1.56% seq_putc
      + 1.55% seq_puts

Link: http://lkml.kernel.org/r/20180212074931.7227-2-avagin@openvz.org
Signed-off-by: Andrei Vagin <avagin@openvz.org>
Reviewed-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: add seq_put_decimal_ull_width to speed up /proc/pid/smaps
Andrei Vagin [Tue, 10 Apr 2018 23:31:16 +0000 (16:31 -0700)]
proc: add seq_put_decimal_ull_width to speed up /proc/pid/smaps

seq_put_decimal_ull_w(m, str, val, width) prints a decimal number with a
specified minimal field width.

It is equivalent of seq_printf(m, "%s%*d", str, width, val), but it
works much faster.

== test_smaps.py
  num = 0
  with open("/proc/1/smaps") as f:
          for x in xrange(10000):
                  data = f.read()
                  f.seek(0, 0)
==

== Before patch ==
  $ time python test_smaps.py
  real    0m4.593s
  user    0m0.398s
  sys     0m4.158s

== After patch ==
  $ time python test_smaps.py
  real    0m3.828s
  user    0m0.413s
  sys     0m3.408s

$ perf -g record python test_smaps.py
== Before patch ==
-   79.01%     3.36%  python   [kernel.kallsyms]    [k] show_smap.isra.33
   - 75.65% show_smap.isra.33
      + 48.85% seq_printf
      + 15.75% __walk_page_range
      + 9.70% show_map_vma.isra.23
        0.61% seq_puts

== After patch ==
-   75.51%     4.62%  python   [kernel.kallsyms]    [k] show_smap.isra.33
   - 70.88% show_smap.isra.33
      + 24.82% seq_put_decimal_ull_w
      + 19.78% __walk_page_range
      + 12.74% seq_printf
      + 11.08% show_map_vma.isra.23
      + 1.68% seq_puts

[akpm@linux-foundation.org: fix drivers/of/unittest.c build]
Link: http://lkml.kernel.org/r/20180212074931.7227-1-avagin@openvz.org
Signed-off-by: Andrei Vagin <avagin@openvz.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: account "struct pde_opener"
Alexey Dobriyan [Tue, 10 Apr 2018 23:31:12 +0000 (16:31 -0700)]
proc: account "struct pde_opener"

The allocation is persistent in fact as any fool can open a file in
/proc and sit on it.

Link: http://lkml.kernel.org/r/20180214082409.GC17157@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: move "struct pde_opener" to kmem cache
Alexey Dobriyan [Tue, 10 Apr 2018 23:31:09 +0000 (16:31 -0700)]
proc: move "struct pde_opener" to kmem cache

"struct pde_opener" is fixed size and we can have more granular approach
to debugging.

For those who don't know, per cache SLUB poisoning and red zoning don't
work if there is at least one object allocated which is hopeless in case
of kmalloc-64 but not in case of standalone cache.  Although systemd
opens 2 files from the get go, so it is hopeless after all.

Link: http://lkml.kernel.org/r/20180214082306.GB17157@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: randomize "struct pde_opener"
Alexey Dobriyan [Tue, 10 Apr 2018 23:31:05 +0000 (16:31 -0700)]
proc: randomize "struct pde_opener"

The more the merrier.

Link: http://lkml.kernel.org/r/20180214081935.GA17157@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: faster open/close of files without ->release hook
Alexey Dobriyan [Tue, 10 Apr 2018 23:31:01 +0000 (16:31 -0700)]
proc: faster open/close of files without ->release hook

The whole point of code in fs/proc/inode.c is to make sure ->release
hook is called either at close() or at rmmod time.

All if it is unnecessary if there is no ->release hook.

Save allocation+list manipulations under spinlock in that case.

Link: http://lkml.kernel.org/r/20180214063033.GA15579@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: move /proc/sysvipc creation to where it belongs
Alexey Dobriyan [Tue, 10 Apr 2018 23:30:58 +0000 (16:30 -0700)]
proc: move /proc/sysvipc creation to where it belongs

Move the proc_mkdir() call within the sysvipc subsystem such that we
avoid polluting proc_root_init() with petty cpp.

[dave@stgolabs.net: contributed changelog]
Link: http://lkml.kernel.org/r/20180216161732.GA10297@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: Davidlohr Bueso <dave@stgolabs.net>
Cc: Manfred Spraul <manfred@colorfullife.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: do less stuff under ->pde_unload_lock
Alexey Dobriyan [Tue, 10 Apr 2018 23:30:54 +0000 (16:30 -0700)]
proc: do less stuff under ->pde_unload_lock

Commit ca469f35a8e9ef ("deal with races between remove_proc_entry() and
proc_reg_release()") moved too much stuff under ->pde_unload_lock making
a problem described at series "[PATCH v5] procfs: Improve Scaling in
proc" worse.

While RCU is being figured out, move kfree() out of ->pde_unload_lock.

On my potato, difference is only 0.5% speedup with concurrent
open+read+close of /proc/cmdline, but the effect should be more
noticeable on more capable machines.

$ perf stat -r 16 -- ./proc-j 16

 Performance counter stats for './proc-j 16' (16 runs):

     130569.502377      task-clock (msec)         #   15.872 CPUs utilized            ( +-  0.05% )
            19,169      context-switches          #    0.147 K/sec                    ( +-  0.18% )
                15      cpu-migrations            #    0.000 K/sec                    ( +-  3.27% )
               437      page-faults               #    0.003 K/sec                    ( +-  1.25% )
   300,172,097,675      cycles                    #    2.299 GHz                      ( +-  0.05% )
    96,793,267,308      instructions              #    0.32  insn per cycle           ( +-  0.04% )
    22,798,342,298      branches                  #  174.607 M/sec                    ( +-  0.04% )
       111,764,687      branch-misses             #    0.49% of all branches          ( +-  0.47% )

       8.226574400 seconds time elapsed                                          ( +-  0.05% )
       ^^^^^^^^^^^

$ perf stat -r 16 -- ./proc-j 16

 Performance counter stats for './proc-j 16' (16 runs):

     129866.777392      task-clock (msec)         #   15.869 CPUs utilized            ( +-  0.04% )
            19,154      context-switches          #    0.147 K/sec                    ( +-  0.66% )
                14      cpu-migrations            #    0.000 K/sec                    ( +-  1.73% )
               431      page-faults               #    0.003 K/sec                    ( +-  1.09% )
   298,556,520,546      cycles                    #    2.299 GHz                      ( +-  0.04% )
    96,525,366,833      instructions              #    0.32  insn per cycle           ( +-  0.04% )
    22,730,194,043      branches                  #  175.027 M/sec                    ( +-  0.04% )
       111,506,074      branch-misses             #    0.49% of all branches          ( +-  0.18% )

       8.183629778 seconds time elapsed                                          ( +-  0.04% )
       ^^^^^^^^^^^

Link: http://lkml.kernel.org/r/20180213132911.GA24298@avx2
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoproc: get rid of task lock/unlock pair to read umask for the "status" file
Mateusz Guzik [Tue, 10 Apr 2018 23:30:51 +0000 (16:30 -0700)]
proc: get rid of task lock/unlock pair to read umask for the "status" file

get_task_umask locks/unlocks the task on its own.  The only caller does
the same thing immediately after.

Utilize the fact the task has to be locked anyway and just do it once.
Since there are no other users and the code is short, fold it in.

Link: http://lkml.kernel.org/r/1517995608-23683-1-git-send-email-mguzik@redhat.com
Signed-off-by: Mateusz Guzik <mguzik@redhat.com>
Reviewed-by: Alexey Dobriyan <adobriyan@gmail.com>
Cc: Konstantin Khlebnikov <koct9i@gmail.com>
Cc: Jerome Marchand <jmarchan@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoprocfs: optimize seq_pad() to speed up /proc/pid/maps
Andrei Vagin [Tue, 10 Apr 2018 23:30:47 +0000 (16:30 -0700)]
procfs: optimize seq_pad() to speed up /proc/pid/maps

seq_printf() is slow and it can be replaced by memset() in this case.

== test.py
  num = 0
  with open("/proc/1/maps") as f:
          while num < 10000 :
                  data = f.read()
                  f.seek(0, 0)
                  num = num + 1
==

== Before patch ==
  $  time python test.py
  real 0m0.986s
  user 0m0.279s
  sys 0m0.707s

== After patch ==
  $ time python test.py
  real 0m0.932s
  user 0m0.261s
  sys 0m0.669s

$ perf record -g python test.py
== Before patch ==
-   47.35%     3.38%  python   [kernel.kallsyms] [k] show_map_vma.isra.23
   - 43.97% show_map_vma.isra.23
      + 20.84% seq_path
      - 15.73% show_vma_header_prefix
      + 6.96% seq_pad
   + 2.94% __GI___libc_read

== After patch ==
-   44.01%     0.34%  python   [kernel.kallsyms] [k] show_pid_map
   - 43.67% show_pid_map
      - 42.91% show_map_vma.isra.23
         + 21.55% seq_path
         - 15.68% show_vma_header_prefix
         + 2.08% seq_pad
        0.55% seq_putc

Link: http://lkml.kernel.org/r/20180112185812.7710-2-avagin@openvz.org
Signed-off-by: Andrei Vagin <avagin@openvz.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoprocfs: add seq_put_hex_ll to speed up /proc/pid/maps
Andrei Vagin [Tue, 10 Apr 2018 23:30:44 +0000 (16:30 -0700)]
procfs: add seq_put_hex_ll to speed up /proc/pid/maps

seq_put_hex_ll() prints a number in hexadecimal notation and works
faster than seq_printf().

== test.py
  num = 0
  with open("/proc/1/maps") as f:
          while num < 10000 :
                  data = f.read()
                  f.seek(0, 0)
                 num = num + 1
==

== Before patch ==
  $  time python test.py

  real 0m1.561s
  user 0m0.257s
  sys 0m1.302s

== After patch ==
  $ time python test.py

  real 0m0.986s
  user 0m0.279s
  sys 0m0.707s

$ perf -g record python test.py:

== Before patch ==
-   67.42%     2.82%  python   [kernel.kallsyms] [k] show_map_vma.isra.22
   - 64.60% show_map_vma.isra.22
      - 44.98% seq_printf
         - seq_vprintf
            - vsnprintf
               + 14.85% number
               + 12.22% format_decode
                 5.56% memcpy_erms
      + 15.06% seq_path
      + 4.42% seq_pad
   + 2.45% __GI___libc_read

== After patch ==
-   47.35%     3.38%  python   [kernel.kallsyms] [k] show_map_vma.isra.23
   - 43.97% show_map_vma.isra.23
      + 20.84% seq_path
      - 15.73% show_vma_header_prefix
           10.55% seq_put_hex_ll
         + 2.65% seq_put_decimal_ull
           0.95% seq_putc
      + 6.96% seq_pad
   + 2.94% __GI___libc_read

[avagin@openvz.org: use unsigned int instead of int where it is suitable]
Link: http://lkml.kernel.org/r/20180214025619.4005-1-avagin@openvz.org
[avagin@openvz.org: v2]
Link: http://lkml.kernel.org/r/20180117082050.25406-1-avagin@openvz.org
Link: http://lkml.kernel.org/r/20180112185812.7710-1-avagin@openvz.org
Signed-off-by: Andrei Vagin <avagin@openvz.org>
Cc: Alexey Dobriyan <adobriyan@gmail.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokasan: prevent compiler from optimizing away memset in tests
Andrey Konovalov [Tue, 10 Apr 2018 23:30:39 +0000 (16:30 -0700)]
kasan: prevent compiler from optimizing away memset in tests

A compiler can optimize away memset calls by replacing them with mov
instructions.  There are KASAN tests that specifically test that KASAN
correctly handles memset calls so we don't want this optimization to
happen.

The solution is to add -fno-builtin flag to test_kasan.ko

Link: http://lkml.kernel.org/r/105ec9a308b2abedb1a0d1fdced0c22d765e4732.1519924383.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Nick Terrell <terrelln@fb.com>
Cc: Chris Mason <clm@fb.com>
Cc: Yury Norov <ynorov@caviumnetworks.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Jeff Layton <jlayton@redhat.com>
Cc: "Jason A . Donenfeld" <Jason@zx2c4.com>
Cc: Kostya Serebryany <kcc@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokasan: fix invalid-free test crashing the kernel
Andrey Konovalov [Tue, 10 Apr 2018 23:30:35 +0000 (16:30 -0700)]
kasan: fix invalid-free test crashing the kernel

When an invalid-free is triggered by one of the KASAN tests, the object
doesn't actually get freed.  This later leads to a BUG failure in
kmem_cache_destroy that checks that there are no allocated objects in
the cache that is being destroyed.

Fix this by calling kmem_cache_free with the proper object address after
the call that triggers invalid-free.

Link: http://lkml.kernel.org/r/286eaefc0a6c3fa9b83b87e7d6dc0fbb5b5c9926.1519924383.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Nick Terrell <terrelln@fb.com>
Cc: Chris Mason <clm@fb.com>
Cc: Yury Norov <ynorov@caviumnetworks.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: "Luis R . Rodriguez" <mcgrof@kernel.org>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Jeff Layton <jlayton@redhat.com>
Cc: "Jason A . Donenfeld" <Jason@zx2c4.com>
Cc: Kostya Serebryany <kcc@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agokasan, slub: fix handling of kasan_slab_free hook
Andrey Konovalov [Tue, 10 Apr 2018 23:30:31 +0000 (16:30 -0700)]
kasan, slub: fix handling of kasan_slab_free hook

The kasan_slab_free hook's return value denotes whether the reuse of a
slab object must be delayed (e.g.  when the object is put into memory
qurantine).

The current way SLUB handles this hook is by ignoring its return value
and hardcoding checks similar (but not exactly the same) to the ones
performed in kasan_slab_free, which is prone to making mistakes.

The main difference between the hardcoded checks and the ones in
kasan_slab_free is whether we want to perform a free in case when an
invalid-free or a double-free was detected (we don't).

This patch changes the way SLUB handles this by:
1. taking into account the return value of kasan_slab_free for each of
   the objects, that are being freed;
2. reconstructing the freelist of objects to exclude the ones, whose
   reuse must be delayed.

[andreyknvl@google.com: eliminate unnecessary branch in slab_free]
Link: http://lkml.kernel.org/r/a62759a2545fddf69b0c034547212ca1eb1b3ce2.1520359686.git.andreyknvl@google.com
Link: http://lkml.kernel.org/r/083f58501e54731203801d899632d76175868e97.1519400992.git.andreyknvl@google.com
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Acked-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: Alexander Potapenko <glider@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Kostya Serebryany <kcc@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/thp: don't count ZONE_MOVABLE as the target for freepage reserving
Joonsoo Kim [Tue, 10 Apr 2018 23:30:27 +0000 (16:30 -0700)]
mm/thp: don't count ZONE_MOVABLE as the target for freepage reserving

There was a regression report for "mm/cma: manage the memory of the CMA
area by using the ZONE_MOVABLE" [1] and I think that it is related to
this problem.  CMA patchset makes the system use one more zone
(ZONE_MOVABLE) and then increases min_free_kbytes.  It reduces usable
memory and it could cause regression.

ZONE_MOVABLE only has movable pages so we don't need to keep enough
freepages to avoid or deal with fragmentation.  So, don't count it.

This changes min_free_kbytes and thus min_watermark greatly if
ZONE_MOVABLE is used.  It will make the user uses more memory.

System:
  22GB ram, fakenuma, 2 nodes. 5 zones are used.

Before:
  min_free_kbytes: 112640

  zone_info (min_watermark):
  Node 0, zone      DMA
          min      19
  Node 0, zone    DMA32
          min      3778
  Node 0, zone   Normal
          min      10191
  Node 0, zone  Movable
          min      0
  Node 0, zone   Device
          min      0
  Node 1, zone      DMA
          min      0
  Node 1, zone    DMA32
          min      0
  Node 1, zone   Normal
          min      14043
  Node 1, zone  Movable
          min      127
  Node 1, zone   Device
          min      0

After:
  min_free_kbytes: 90112

  zone_info (min_watermark):
  Node 0, zone      DMA
          min      15
  Node 0, zone    DMA32
          min      3022
  Node 0, zone   Normal
          min      8152
  Node 0, zone  Movable
          min      0
  Node 0, zone   Device
          min      0
  Node 1, zone      DMA
          min      0
  Node 1, zone    DMA32
          min      0
  Node 1, zone   Normal
          min      11234
  Node 1, zone  Movable
          min      102
  Node 1, zone   Device
          min      0

[1] (lkml.kernel.org/r/20180102063528.GG30397%20()%20yexl-desktop)

Link: http://lkml.kernel.org/r/1522913236-15776-1-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoARM: CMA: avoid double mapping to the CMA area if CONFIG_HIGHMEM=y
Joonsoo Kim [Tue, 10 Apr 2018 23:30:23 +0000 (16:30 -0700)]
ARM: CMA: avoid double mapping to the CMA area if CONFIG_HIGHMEM=y

CMA area is now managed by the separate zone, ZONE_MOVABLE, to fix many
MM related problems.  In this implementation, if CONFIG_HIGHMEM = y,
then ZONE_MOVABLE is considered as HIGHMEM and the memory of the CMA
area is also considered as HIGHMEM.  That means that they are considered
as the page without direct mapping.  However, CMA area could be in a
lowmem and the memory could have direct mapping.

In ARM, when establishing a new mapping for DMA, direct mapping should
be cleared since two mapping with different cache policy could cause
unknown problem.  With this patch, PageHighmem() for the CMA memory
located in lowmem returns true so that the function for DMA mapping
cannot notice whether it needs to clear direct mapping or not,
correctly.  To handle this situation, this patch always clears direct
mapping for such CMA memory.

Link: http://lkml.kernel.org/r/1512114786-5085-4-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Cc: "Aneesh Kumar K . V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Laura Abbott <lauraa@codeaurora.org>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/cma: remove ALLOC_CMA
Joonsoo Kim [Tue, 10 Apr 2018 23:30:19 +0000 (16:30 -0700)]
mm/cma: remove ALLOC_CMA

Now, all reserved pages for CMA region are belong to the ZONE_MOVABLE
and it only serves for a request with GFP_HIGHMEM && GFP_MOVABLE.

Therefore, we don't need to maintain ALLOC_CMA at all.

Link: http://lkml.kernel.org/r/1512114786-5085-3-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Laura Abbott <lauraa@codeaurora.org>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/cma: manage the memory of the CMA area by using the ZONE_MOVABLE
Joonsoo Kim [Tue, 10 Apr 2018 23:30:15 +0000 (16:30 -0700)]
mm/cma: manage the memory of the CMA area by using the ZONE_MOVABLE

Patch series "mm/cma: manage the memory of the CMA area by using the
ZONE_MOVABLE", v2.

0. History

This patchset is the follow-up of the discussion about the "Introduce
ZONE_CMA (v7)" [1].  Please reference it if more information is needed.

1. What does this patch do?

This patch changes the management way for the memory of the CMA area in
the MM subsystem.  Currently the memory of the CMA area is managed by
the zone where their pfn is belong to.  However, this approach has some
problems since MM subsystem doesn't have enough logic to handle the
situation that different characteristic memories are in a single zone.
To solve this issue, this patch try to manage all the memory of the CMA
area by using the MOVABLE zone.  In MM subsystem's point of view,
characteristic of the memory on the MOVABLE zone and the memory of the
CMA area are the same.  So, managing the memory of the CMA area by using
the MOVABLE zone will not have any problem.

2. Motivation

There are some problems with current approach.  See following.  Although
these problem would not be inherent and it could be fixed without this
conception change, it requires many hooks addition in various code path
and it would be intrusive to core MM and would be really error-prone.
Therefore, I try to solve them with this new approach.  Anyway,
following is the problems of the current implementation.

o CMA memory utilization

First, following is the freepage calculation logic in MM.

 - For movable allocation: freepage = total freepage
 - For unmovable allocation: freepage = total freepage - CMA freepage

Freepages on the CMA area is used after the normal freepages in the zone
where the memory of the CMA area is belong to are exhausted.  At that
moment that the number of the normal freepages is zero, so

 - For movable allocation: freepage = total freepage = CMA freepage
 - For unmovable allocation: freepage = 0

If unmovable allocation comes at this moment, allocation request would
fail to pass the watermark check and reclaim is started.  After reclaim,
there would exist the normal freepages so freepages on the CMA areas
would not be used.

FYI, there is another attempt [2] trying to solve this problem in lkml.
And, as far as I know, Qualcomm also has out-of-tree solution for this
problem.

Useless reclaim:

There is no logic to distinguish CMA pages in the reclaim path.  Hence,
CMA page is reclaimed even if the system just needs the page that can be
usable for the kernel allocation.

Atomic allocation failure:

This is also related to the fallback allocation policy for the memory of
the CMA area.  Consider the situation that the number of the normal
freepages is *zero* since the bunch of the movable allocation requests
come.  Kswapd would not be woken up due to following freepage
calculation logic.

- For movable allocation: freepage = total freepage = CMA freepage

If atomic unmovable allocation request comes at this moment, it would
fails due to following logic.

- For unmovable allocation: freepage = total freepage - CMA freepage = 0

It was reported by Aneesh [3].

Useless compaction:

Usual high-order allocation request is unmovable allocation request and
it cannot be served from the memory of the CMA area.  In compaction,
migration scanner try to migrate the page in the CMA area and make
high-order page there.  As mentioned above, it cannot be usable for the
unmovable allocation request so it's just waste.

3. Current approach and new approach

Current approach is that the memory of the CMA area is managed by the
zone where their pfn is belong to.  However, these memory should be
distinguishable since they have a strong limitation.  So, they are
marked as MIGRATE_CMA in pageblock flag and handled specially.  However,
as mentioned in section 2, the MM subsystem doesn't have enough logic to
deal with this special pageblock so many problems raised.

New approach is that the memory of the CMA area is managed by the
MOVABLE zone.  MM already have enough logic to deal with special zone
like as HIGHMEM and MOVABLE zone.  So, managing the memory of the CMA
area by the MOVABLE zone just naturally work well because constraints
for the memory of the CMA area that the memory should always be
migratable is the same with the constraint for the MOVABLE zone.

There is one side-effect for the usability of the memory of the CMA
area.  The use of MOVABLE zone is only allowed for a request with
GFP_HIGHMEM && GFP_MOVABLE so now the memory of the CMA area is also
only allowed for this gfp flag.  Before this patchset, a request with
GFP_MOVABLE can use them.  IMO, It would not be a big issue since most
of GFP_MOVABLE request also has GFP_HIGHMEM flag.  For example, file
cache page and anonymous page.  However, file cache page for blockdev
file is an exception.  Request for it has no GFP_HIGHMEM flag.  There is
pros and cons on this exception.  In my experience, blockdev file cache
pages are one of the top reason that causes cma_alloc() to fail
temporarily.  So, we can get more guarantee of cma_alloc() success by
discarding this case.

Note that there is no change in admin POV since this patchset is just
for internal implementation change in MM subsystem.  Just one minor
difference for admin is that the memory stat for CMA area will be
printed in the MOVABLE zone.  That's all.

4. Result

Following is the experimental result related to utilization problem.

8 CPUs, 1024 MB, VIRTUAL MACHINE
make -j16

<Before>
  CMA area:               0 MB            512 MB
  Elapsed-time:           92.4 186.5
  pswpin:                 82 18647
  pswpout:                160 69839

<After>
  CMA        :            0 MB            512 MB
  Elapsed-time:           93.1 93.4
  pswpin:                 84 46
  pswpout:                183 92

akpm: "kernel test robot" reported a 26% improvement in
vm-scalability.throughput:
http://lkml.kernel.org/r/20180330012721.GA3845@yexl-desktop

[1]: lkml.kernel.org/r/1491880640-9944-1-git-send-email-iamjoonsoo.kim@lge.com
[2]: https://lkml.org/lkml/2014/10/15/623
[3]: http://www.spinics.net/lists/linux-mm/msg100562.html

Link: http://lkml.kernel.org/r/1512114786-5085-2-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Laura Abbott <lauraa@codeaurora.org>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/page_alloc: don't reserve ZONE_HIGHMEM for ZONE_MOVABLE request
Joonsoo Kim [Tue, 10 Apr 2018 23:30:11 +0000 (16:30 -0700)]
mm/page_alloc: don't reserve ZONE_HIGHMEM for ZONE_MOVABLE request

Freepage on ZONE_HIGHMEM doesn't work for kernel memory so it's not that
important to reserve.  When ZONE_MOVABLE is used, this problem would
theorectically cause to decrease usable memory for GFP_HIGHUSER_MOVABLE
allocation request which is mainly used for page cache and anon page
allocation.  So, fix it by setting 0 to
sysctl_lowmem_reserve_ratio[ZONE_HIGHMEM].

And, defining sysctl_lowmem_reserve_ratio array by MAX_NR_ZONES - 1 size
makes code complex.  For example, if there is highmem system, following
reserve ratio is activated for *NORMAL ZONE* which would be easyily
misleading people.

 #ifdef CONFIG_HIGHMEM
 32
 #endif

This patch also fixes this situation by defining
sysctl_lowmem_reserve_ratio array by MAX_NR_ZONES and place "#ifdef" to
right place.

Link: http://lkml.kernel.org/r/1504672525-17915-1-git-send-email-iamjoonsoo.kim@lge.com
Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Tested-by: Tony Lindgren <tony@atomide.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: "Aneesh Kumar K . V" <aneesh.kumar@linux.vnet.ibm.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Laura Abbott <lauraa@codeaurora.org>
Cc: Marek Szyprowski <m.szyprowski@samsung.com>
Cc: Michal Nazarewicz <mina86@mina86.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: Will Deacon <will.deacon@arm.com>
Cc: <linux-api@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm: unclutter THP migration
Michal Hocko [Tue, 10 Apr 2018 23:30:07 +0000 (16:30 -0700)]
mm: unclutter THP migration

THP migration is hacked into the generic migration with rather
surprising semantic.  The migration allocation callback is supposed to
check whether the THP can be migrated at once and if that is not the
case then it allocates a simple page to migrate.  unmap_and_move then
fixes that up by spliting the THP into small pages while moving the head
page to the newly allocated order-0 page.  Remaning pages are moved to
the LRU list by split_huge_page.  The same happens if the THP allocation
fails.  This is really ugly and error prone [1].

I also believe that split_huge_page to the LRU lists is inherently wrong
because all tail pages are not migrated.  Some callers will just work
around that by retrying (e.g.  memory hotplug).  There are other pfn
walkers which are simply broken though.  e.g. madvise_inject_error will
migrate head and then advances next pfn by the huge page size.
do_move_page_to_node_array, queue_pages_range (migrate_pages, mbind),
will simply split the THP before migration if the THP migration is not
supported then falls back to single page migration but it doesn't handle
tail pages if the THP migration path is not able to allocate a fresh THP
so we end up with ENOMEM and fail the whole migration which is a
questionable behavior.  Page compaction doesn't try to migrate large
pages so it should be immune.

This patch tries to unclutter the situation by moving the special THP
handling up to the migrate_pages layer where it actually belongs.  We
simply split the THP page into the existing list if unmap_and_move fails
with ENOMEM and retry.  So we will _always_ migrate all THP subpages and
specific migrate_pages users do not have to deal with this case in a
special way.

[1] http://lkml.kernel.org/r/20171121021855.50525-1-zi.yan@sent.com

Link: http://lkml.kernel.org/r/20180103082555.14592-4-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Reviewed-by: Zi Yan <zi.yan@cs.rutgers.edu>
Cc: Andrea Reale <ar@linux.vnet.ibm.com>
Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm, migrate: remove reason argument from new_page_t
Michal Hocko [Tue, 10 Apr 2018 23:30:03 +0000 (16:30 -0700)]
mm, migrate: remove reason argument from new_page_t

No allocation callback is using this argument anymore.  new_page_node
used to use this parameter to convey node_id resp.  migration error up
to move_pages code (do_move_page_to_node_array).  The error status never
made it into the final status field and we have a better way to
communicate node id to the status field now.  All other allocation
callbacks simply ignored the argument so we can drop it finally.

[mhocko@suse.com: fix migration callback]
Link: http://lkml.kernel.org/r/20180105085259.GH2801@dhcp22.suse.cz
[akpm@linux-foundation.org: fix alloc_misplaced_dst_page()]
[mhocko@kernel.org: fix build]
Link: http://lkml.kernel.org/r/20180103091134.GB11319@dhcp22.suse.cz
Link: http://lkml.kernel.org/r/20180103082555.14592-3-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Zi Yan <zi.yan@cs.rutgers.edu>
Cc: Andrea Reale <ar@linux.vnet.ibm.com>
Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm, numa: rework do_pages_move
Michal Hocko [Tue, 10 Apr 2018 23:29:59 +0000 (16:29 -0700)]
mm, numa: rework do_pages_move

Patch series "unclutter thp migration"

Motivation:

THP migration is hacked into the generic migration with rather
surprising semantic.  The migration allocation callback is supposed to
check whether the THP can be migrated at once and if that is not the
case then it allocates a simple page to migrate.  unmap_and_move then
fixes that up by splitting the THP into small pages while moving the
head page to the newly allocated order-0 page.  Remaining pages are
moved to the LRU list by split_huge_page.  The same happens if the THP
allocation fails.  This is really ugly and error prone [2].

I also believe that split_huge_page to the LRU lists is inherently wrong
because all tail pages are not migrated.  Some callers will just work
around that by retrying (e.g.  memory hotplug).  There are other pfn
walkers which are simply broken though.  e.g. madvise_inject_error will
migrate head and then advances next pfn by the huge page size.
do_move_page_to_node_array, queue_pages_range (migrate_pages, mbind),
will simply split the THP before migration if the THP migration is not
supported then falls back to single page migration but it doesn't handle
tail pages if the THP migration path is not able to allocate a fresh THP
so we end up with ENOMEM and fail the whole migration which is a
questionable behavior.  Page compaction doesn't try to migrate large
pages so it should be immune.

The first patch reworks do_pages_move which relies on a very ugly
calling semantic when the return status is pushed to the migration path
via private pointer.  It uses pre allocated fixed size batching to
achieve that.  We simply cannot do the same if a THP is to be split
during the migration path which is done in the patch 3.  Patch 2 is
follow up cleanup which removes the mentioned return status calling
convention ugliness.

On a side note:

There are some semantic issues I have encountered on the way when
working on patch 1 but I am not addressing them here.  E.g. trying to
move THP tail pages will result in either success or EBUSY (the later
one more likely once we isolate head from the LRU list).  Hugetlb
reports EACCESS on tail pages.  Some errors are reported via status
parameter but migration failures are not even though the original
`reason' argument suggests there was an intention to do so.  From a
quick look into git history this never worked.  I have tried to keep the
semantic unchanged.

Then there is a relatively minor thing that the page isolation might
fail because of pages not being on the LRU - e.g. because they are
sitting on the per-cpu LRU caches.  Easily fixable.

This patch (of 3):

do_pages_move is supposed to move user defined memory (an array of
addresses) to the user defined numa nodes (an array of nodes one for
each address).  The user provided status array then contains resulting
numa node for each address or an error.  The semantic of this function
is little bit confusing because only some errors are reported back.
Notably migrate_pages error is only reported via the return value.  This
patch doesn't try to address these semantic nuances but rather change
the underlying implementation.

Currently we are processing user input (which can be really large) in
batches which are stored to a temporarily allocated page.  Each address
is resolved to its struct page and stored to page_to_node structure
along with the requested target numa node.  The array of these
structures is then conveyed down the page migration path via private
argument.  new_page_node then finds the corresponding structure and
allocates the proper target page.

What is the problem with the current implementation and why to change
it? Apart from being quite ugly it also doesn't cope with unexpected
pages showing up on the migration list inside migrate_pages path.  That
doesn't happen currently but the follow up patch would like to make the
thp migration code more clear and that would need to split a THP into
the list for some cases.

How does the new implementation work? Well, instead of batching into a
fixed size array we simply batch all pages that should be migrated to
the same node and isolate all of them into a linked list which doesn't
require any additional storage.  This should work reasonably well
because page migration usually migrates larger ranges of memory to a
specific node.  So the common case should work equally well as the
current implementation.  Even if somebody constructs an input where the
target numa nodes would be interleaved we shouldn't see a large
performance impact because page migration alone doesn't really benefit
from batching.  mmap_sem batching for the lookup is quite questionable
and isolate_lru_page which would benefit from batching is not using it
even in the current implementation.

Link: http://lkml.kernel.org/r/20180103082555.14592-2-mhocko@kernel.org
Signed-off-by: Michal Hocko <mhocko@suse.com>
Acked-by: Kirill A. Shutemov <kirill@shutemov.name>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com>
Cc: Zi Yan <zi.yan@cs.rutgers.edu>
Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrea Reale <ar@linux.vnet.ibm.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/swapfile.c: make pointer swap_avail_heads static
Colin Ian King [Tue, 10 Apr 2018 23:29:55 +0000 (16:29 -0700)]
mm/swapfile.c: make pointer swap_avail_heads static

The pointer swap_avail_heads is local to the source and does not need to
be in global scope, so make it static.

Cleans up sparse warning:

  mm/swapfile.c:88:19: warning: symbol 'swap_avail_heads' was not declared. Should it be static?

Link: http://lkml.kernel.org/r/20180206215836.12366-1-colin.king@canonical.com
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Acked-by: "Huang, Ying" <ying.huang@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomemcg: fix per_node_info cleanup
Michal Hocko [Tue, 10 Apr 2018 23:29:52 +0000 (16:29 -0700)]
memcg: fix per_node_info cleanup

syzbot has triggered a NULL ptr dereference when allocation fault
injection enforces a failure and alloc_mem_cgroup_per_node_info
initializes memcg->nodeinfo only half way through.

But __mem_cgroup_free still tries to free all per-node data and
dereferences pn->lruvec_stat_cpu unconditioanlly even if the specific
per-node data hasn't been initialized.

The bug is quite unlikely to hit because small allocations do not fail
and we would need quite some numa nodes to make struct
mem_cgroup_per_node large enough to cross the costly order.

Link: http://lkml.kernel.org/r/20180406100906.17790-1-mhocko@kernel.org
Reported-by: syzbot+8a5de3cce7cdc70e9ebe@syzkaller.appspotmail.com
Fixes: 00f3ca2c2d66 ("mm: memcontrol: per-lruvec stats infrastructure")
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoswap: divide-by-zero when zero length swap file on ssd
Tom Abraham [Tue, 10 Apr 2018 23:29:48 +0000 (16:29 -0700)]
swap: divide-by-zero when zero length swap file on ssd

Calling swapon() on a zero length swap file on SSD can lead to a
divide-by-zero.

Although creating such files isn't possible with mkswap and they woud be
considered invalid, it would be better for the swapon code to be more
robust and handle this condition gracefully (return -EINVAL).
Especially since the fix is small and straightforward.

To help with wear leveling on SSD, the swapon syscall calculates a
random position in the swap file using modulo p->highest_bit, which is
set to maxpages - 1 in read_swap_header.

If the swap file is zero length, read_swap_header sets maxpages=1 and
last_page=0, resulting in p->highest_bit=0 and we divide-by-zero when we
modulo p->highest_bit in swapon syscall.

This can be prevented by having read_swap_header return zero if
last_page is zero.

Link: http://lkml.kernel.org/r/5AC747C1020000A7001FA82C@prv-mh.provo.novell.com
Signed-off-by: Thomas Abraham <tabraham@suse.com>
Reported-by: <Mark.Landis@Teradata.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm: memcg: make sure memory.events is uptodate when waking pollers
Johannes Weiner [Tue, 10 Apr 2018 23:29:45 +0000 (16:29 -0700)]
mm: memcg: make sure memory.events is uptodate when waking pollers

Commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in
memory.stat reporting") added per-cpu drift to all memory cgroup stats
and events shown in memory.stat and memory.events.

For memory.stat this is acceptable.  But memory.events issues file
notifications, and somebody polling the file for changes will be
confused when the counters in it are unchanged after a wakeup.

Luckily, the events in memory.events - MEMCG_LOW, MEMCG_HIGH, MEMCG_MAX,
MEMCG_OOM - are sufficiently rare and high-level that we don't need
per-cpu buffering for them: MEMCG_HIGH and MEMCG_MAX would be the most
frequent, but they're counting invocations of reclaim, which is a
complex operation that touches many shared cachelines.

This splits memory.events from the generic VM events and tracks them in
their own, unbuffered atomic counters.  That's also cleaner, as it
eliminates the ugly enum nesting of VM and cgroup events.

[hannes@cmpxchg.org: "array subscript is above array bounds"]
Link: http://lkml.kernel.org/r/20180406155441.GA20806@cmpxchg.org
Link: http://lkml.kernel.org/r/20180405175507.GA24817@cmpxchg.org
Fixes: a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting")
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Reported-by: Tejun Heo <tj@kernel.org>
Acked-by: Tejun Heo <tj@kernel.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: Rik van Riel <riel@surriel.com>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/ksm.c: fix inconsistent accounting of zero pages
Claudio Imbrenda [Tue, 10 Apr 2018 23:29:41 +0000 (16:29 -0700)]
mm/ksm.c: fix inconsistent accounting of zero pages

When using KSM with use_zero_pages, we replace anonymous pages
containing only zeroes with actual zero pages, which are not anonymous.
We need to do proper accounting of the mm counters, otherwise we will
get wrong values in /proc and a BUG message in dmesg when tearing down
the mm.

Link: http://lkml.kernel.org/r/1522931274-15552-1-git-send-email-imbrenda@linux.vnet.ibm.com
Fixes: e86c59b1b1 ("mm/ksm: improve deduplication of zero pages with colouring")
Signed-off-by: Claudio Imbrenda <imbrenda@linux.vnet.ibm.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: Hugh Dickins <hughd@google.com>
Cc: Christian Borntraeger <borntraeger@de.ibm.com>
Cc: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/z3fold.c: use gfpflags_allow_blocking
Matthew Wilcox [Tue, 10 Apr 2018 23:29:37 +0000 (16:29 -0700)]
mm/z3fold.c: use gfpflags_allow_blocking

We have a perfectly good macro to determine whether the gfp flags allow
you to sleep or not; use it instead of trying to infer it.

Link: http://lkml.kernel.org/r/20180408062206.GC16007@bombadil.infradead.org
Signed-off-by: Matthew Wilcox <mawilcox@microsoft.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Vitaly Wool <vitalywool@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoz3fold: fix memory leak
Xidong Wang [Tue, 10 Apr 2018 23:29:34 +0000 (16:29 -0700)]
z3fold: fix memory leak

In z3fold_create_pool(), the memory allocated by __alloc_percpu() is not
released on the error path that pool->compact_wq , which holds the
return value of create_singlethread_workqueue(), is NULL.  This will
result in a memory leak bug.

[akpm@linux-foundation.org: fix oops on kzalloc() failure, check __alloc_percpu() retval]
Link: http://lkml.kernel.org/r/1522803111-29209-1-git-send-email-wangxidong_97@163.com
Signed-off-by: Xidong Wang <wangxidong_97@163.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Vitaly Wool <vitalywool@gmail.com>
Cc: Mike Rapoport <rppt@linux.vnet.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomemcg, thp: do not invoke oom killer on thp charges
Michal Hocko [Tue, 10 Apr 2018 23:29:30 +0000 (16:29 -0700)]
memcg, thp: do not invoke oom killer on thp charges

A THP memcg charge can trigger the oom killer since 2516035499b9 ("mm,
thp: remove __GFP_NORETRY from khugepaged and madvised allocations").
We have used an explicit __GFP_NORETRY previously which ruled the OOM
killer automagically.

Memcg charge path should be semantically compliant with the allocation
path and that means that if we do not trigger the OOM killer for costly
orders which should do the same in the memcg charge path as well.
Otherwise we are forcing callers to distinguish the two and use
different gfp masks which is both non-intuitive and bug prone.  As soon
as we get a costly high order kmalloc user we even do not have any means
to tell the memcg specific gfp mask to prevent from OOM because the
charging is deep within guts of the slab allocator.

The unexpected memcg OOM on THP has already been fixed upstream by
9d3c3354bb85 ("mm, thp: do not cause memcg oom for thp") but this is a
one-off fix rather than a generic solution.  Teach mem_cgroup_oom to
bail out on costly order requests to fix the THP issue as well as any
other costly OOM eligible allocations to be added in future.

Also revert 9d3c3354bb85 because special gfp for THP is no longer
needed.

Link: http://lkml.kernel.org/r/20180403193129.22146-1-mhocko@kernel.org
Fixes: 2516035499b9 ("mm, thp: remove __GFP_NORETRY from khugepaged and madvised allocations")
Signed-off-by: Michal Hocko <mhocko@suse.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/migrate: properly preserve write attribute in special migrate entry
Ralph Campbell [Tue, 10 Apr 2018 23:29:27 +0000 (16:29 -0700)]
mm/migrate: properly preserve write attribute in special migrate entry

Use of pte_write(pte) is only valid for present pte, the common code
which set the migration entry can be reach for both valid present pte
and special swap entry (for device memory).  Fix the code to use the
mpfn value which properly handle both cases.

On x86 this did not have any bad side effect because pte write bit is
below PAGE_BIT_GLOBAL and thus special swap entry have it set to 0 which
in turn means we were always creating read only special migration entry.

So once migration did finish we always write protected the CPU page
table entry (moreover this is only an issue when migrating from device
memory to system memory).  End effect is that CPU write access would
fault again and restore write permission.

This behaviour isn't too bad; it just burns CPU cycles by forcing CPU to
take a second fault on write access. ie, double faulting the same
address.  There is no corruption or incorrect states (it behaves as a
COWed page from a fork with a mapcount of 1).

Link: http://lkml.kernel.org/r/20180402023506.12180-1-jglisse@redhat.com
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm: check __highest_present_section_nr directly in memory_dev_init()
Wei Yang [Tue, 10 Apr 2018 23:29:23 +0000 (16:29 -0700)]
mm: check __highest_present_section_nr directly in memory_dev_init()

__highest_present_section_nr is a more strict boundary than
NR_MEM_SECTIONS.  So checking __highest_present_section_nr directly is
enough.

Link: http://lkml.kernel.org/r/20180330032044.21647-1-richard.weiyang@gmail.com
Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agosched/numa: avoid trapping faults and attempting migration of file-backed dirty pages
Mel Gorman [Tue, 10 Apr 2018 23:29:20 +0000 (16:29 -0700)]
sched/numa: avoid trapping faults and attempting migration of file-backed dirty pages

change_pte_range is called from task work context to mark PTEs for
receiving NUMA faulting hints.  If the marked pages are dirty then
migration may fail.  Some filesystems cannot migrate dirty pages without
blocking so are skipped in MIGRATE_ASYNC mode which just wastes CPU.
Even when they can, it can be a waste of cycles when the pages are
shared forcing higher scan rates.  This patch avoids marking shared
dirty pages for hinting faults but also will skip a migration if the
page was dirtied after the scanner updated a clean page.

This is most noticeable running the NASA Parallel Benchmark when backed
by btrfs, the default root filesystem for some distributions, but also
noticeable when using XFS.

The following are results from a 4-socket machine running a 4.16-rc4
kernel with some scheduler patches that are pending for the next merge
window.

                        4.16.0-rc4             4.16.0-rc4
                 schedtip-20180309          nodirty-v1
  Time cg.D      459.07 (   0.00%)      444.21 (   3.24%)
  Time ep.D       76.96 (   0.00%)       77.69 (  -0.95%)
  Time is.D       25.55 (   0.00%)       27.85 (  -9.00%)
  Time lu.D      601.58 (   0.00%)      596.87 (   0.78%)
  Time mg.D      107.73 (   0.00%)      108.22 (  -0.45%)

is.D regresses slightly in terms of absolute time but note that that
particular load varies quite a bit from run to run.  The more relevant
observation is the total system CPU usage.

            4.16.0-rc4  4.16.0-rc4
          schedtip-20180309 nodirty-v1
  User        71471.91    70627.04
  System      11078.96     8256.13
  Elapsed       661.66      632.74

That is a substantial drop in system CPU usage and overall the workload
completes faster.  The NUMA balancing statistics are also interesting

  NUMA base PTE updates        111407972   139848884
  NUMA huge PMD updates           206506      264869
  NUMA page range updates      217139044   275461812
  NUMA hint faults               4300924     3719784
  NUMA hint local faults         3012539     3416618
  NUMA hint local percent             70          91
  NUMA pages migrated            1517487     1358420

While more PTEs are scanned due to changes in what faults are gathered,
it's clear that a far higher percentage of faults are local as the bulk
of the remote hits were dirty pages that, in this case with btrfs, had
no chance of migrating.

The following is a comparison when using XFS as that is a more realistic
filesystem choice for a data partition

                        4.16.0-rc4             4.16.0-rc4
                 schedtip-20180309          nodirty-v1r47
  Time cg.D      485.28 (   0.00%)      442.62 (   8.79%)
  Time ep.D       77.68 (   0.00%)       77.54 (   0.18%)
  Time is.D       26.44 (   0.00%)       24.79 (   6.24%)
  Time lu.D      597.46 (   0.00%)      597.11 (   0.06%)
  Time mg.D      142.65 (   0.00%)      105.83 (  25.81%)

That is a reasonable gain on two relatively long-lived workloads.  While
not presented, there is also a substantial drop in system CPu usage and
the NUMA balancing stats show similar improvements in locality as btrfs
did.

Link: http://lkml.kernel.org/r/20180326094334.zserdec62gwmmfqf@techsingularity.net
Signed-off-by: Mel Gorman <mgorman@techsingularity.net>
Reviewed-by: Rik van Riel <riel@surriel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoDocumentation/vm/hmm.txt: typos and syntaxes fixes
Jérôme Glisse [Tue, 10 Apr 2018 23:29:16 +0000 (16:29 -0700)]
Documentation/vm/hmm.txt: typos and syntaxes fixes

This fix typos and syntaxes, thanks to Randy Dunlap for pointing them out
(they were all my faults).

Link: http://lkml.kernel.org/r/20180409151859.4713-1-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: Ralph Campbell <rcampbell@nvidia.com>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: fix header file if/else/endif maze, again
Arnd Bergmann [Tue, 10 Apr 2018 23:29:13 +0000 (16:29 -0700)]
mm/hmm: fix header file if/else/endif maze, again

The last fix was still wrong, as we need the inline dummy functions also
for the case that CONFIG_HMM is enabled but CONFIG_HMM_MIRROR is not:

  kernel/fork.o: In function `__mmdrop':
  fork.c:(.text+0x14f6): undefined reference to `hmm_mm_destroy'

This adds back the second copy of the dummy functions, hopefully
this time in the right place.

Link: http://lkml.kernel.org/r/20180404110236.804484-1-arnd@arndb.de
Fixes: 8900d06a277a ("mm/hmm: fix header file if/else/endif maze")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Jérôme Glisse <jglisse@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm.c: remove superfluous RCU protection around radix tree lookup
Tejun Heo [Tue, 10 Apr 2018 23:29:09 +0000 (16:29 -0700)]
mm/hmm.c: remove superfluous RCU protection around radix tree lookup

hmm_devmem_find() requires rcu_read_lock_held() but there's nothing which
actually uses the RCU protection.  The only caller is
hmm_devmem_pages_create() which already grabs the mutex and does
superfluous rcu_read_lock/unlock() around the function.

This doesn't add anything and just adds to confusion.  Remove the RCU
protection and open-code the radix tree lookup.  If this needs to become
more sophisticated in the future, let's add them back when necessary.

Link: http://lkml.kernel.org/r/20180314194515.1661824-4-tj@kernel.org
Signed-off-by: Tejun Heo <tj@kernel.org>
Reviewed-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Benjamin LaHaise <bcrl@kvack.org>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Kent Overstreet <kent.overstreet@gmail.com>
Cc: Matthew Wilcox <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: use device driver encoding for HMM pfn
Jérôme Glisse [Tue, 10 Apr 2018 23:29:06 +0000 (16:29 -0700)]
mm/hmm: use device driver encoding for HMM pfn

Users of hmm_vma_fault() and hmm_vma_get_pfns() provide a flags array and
pfn shift value allowing them to define their own encoding for HMM pfn
that are fill inside the pfns array of the hmm_range struct.  With this
device driver can get pfn that match their own private encoding out of HMM
without having to do any conversion.

[rcampbell@nvidia.com: don't ignore specific pte fault flag in hmm_vma_fault()]
Link: http://lkml.kernel.org/r/20180326213009.2460-2-jglisse@redhat.com
[rcampbell@nvidia.com: clarify fault logic for device private memory]
Link: http://lkml.kernel.org/r/20180326213009.2460-3-jglisse@redhat.com
Link: http://lkml.kernel.org/r/20180323005527.758-16-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: change hmm_vma_fault() to allow write fault on page basis
Jérôme Glisse [Tue, 10 Apr 2018 23:29:02 +0000 (16:29 -0700)]
mm/hmm: change hmm_vma_fault() to allow write fault on page basis

This changes hmm_vma_fault() to not take a global write fault flag for a
range but instead rely on caller to populate HMM pfns array with proper
fault flag ie HMM_PFN_VALID if driver want read fault for that address or
HMM_PFN_VALID and HMM_PFN_WRITE for write.

Moreover by setting HMM_PFN_DEVICE_PRIVATE the device driver can ask for
device private memory to be migrated back to system memory through page
fault.

This is more flexible API and it better reflects how device handles and
reports fault.

Link: http://lkml.kernel.org/r/20180323005527.758-15-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: factor out pte and pmd handling to simplify hmm_vma_walk_pmd()
Jérôme Glisse [Tue, 10 Apr 2018 23:28:59 +0000 (16:28 -0700)]
mm/hmm: factor out pte and pmd handling to simplify hmm_vma_walk_pmd()

No functional change, just create one function to handle pmd and one to
handle pte (hmm_vma_handle_pmd() and hmm_vma_handle_pte()).

Link: http://lkml.kernel.org/r/20180323005527.758-14-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: move hmm_pfns_clear() closer to where it is used
Jérôme Glisse [Tue, 10 Apr 2018 23:28:54 +0000 (16:28 -0700)]
mm/hmm: move hmm_pfns_clear() closer to where it is used

Move hmm_pfns_clear() closer to where it is used to make it clear it is
not use by page table walkers.

Link: http://lkml.kernel.org/r/20180323005527.758-13-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: rename HMM_PFN_DEVICE_UNADDRESSABLE to HMM_PFN_DEVICE_PRIVATE
Jérôme Glisse [Tue, 10 Apr 2018 23:28:49 +0000 (16:28 -0700)]
mm/hmm: rename HMM_PFN_DEVICE_UNADDRESSABLE to HMM_PFN_DEVICE_PRIVATE

Make naming consistent across code, DEVICE_PRIVATE is the name use outside
HMM code so use that one.

Link: http://lkml.kernel.org/r/20180323005527.758-12-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: do not differentiate between empty entry or missing directory
Jérôme Glisse [Tue, 10 Apr 2018 23:28:46 +0000 (16:28 -0700)]
mm/hmm: do not differentiate between empty entry or missing directory

There is no point in differentiating between a range for which there is
not even a directory (and thus entries) and empty entry (pte_none() or
pmd_none() returns true).

Simply drop the distinction ie remove HMM_PFN_EMPTY flag and merge now
duplicate hmm_vma_walk_hole() and hmm_vma_walk_clear() functions.

Link: http://lkml.kernel.org/r/20180323005527.758-11-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: cleanup special vma handling (VM_SPECIAL)
Jérôme Glisse [Tue, 10 Apr 2018 23:28:42 +0000 (16:28 -0700)]
mm/hmm: cleanup special vma handling (VM_SPECIAL)

Special vma (one with any of the VM_SPECIAL flags) can not be access by
device because there is no consistent model across device drivers on those
vma and their backing memory.

This patch directly use hmm_range struct for hmm_pfns_special() argument
as it is always affecting the whole vma and thus the whole range.

It also make behavior consistent after this patch both hmm_vma_fault() and
hmm_vma_get_pfns() returns -EINVAL when facing such vma.  Previously
hmm_vma_fault() returned 0 and hmm_vma_get_pfns() return -EINVAL but both
were filling the HMM pfn array with special entry.

Link: http://lkml.kernel.org/r/20180323005527.758-10-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: use uint64_t for HMM pfn instead of defining hmm_pfn_t to ulong
Jérôme Glisse [Tue, 10 Apr 2018 23:28:38 +0000 (16:28 -0700)]
mm/hmm: use uint64_t for HMM pfn instead of defining hmm_pfn_t to ulong

All device driver we care about are using 64bits page table entry.  In
order to match this and to avoid useless define convert all HMM pfn to
directly use uint64_t.  It is a first step on the road to allow driver to
directly use pfn value return by HMM (saving memory and CPU cycles use for
conversion between the two).

Link: http://lkml.kernel.org/r/20180323005527.758-9-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: remove HMM_PFN_READ flag and ignore peculiar architecture
Jérôme Glisse [Tue, 10 Apr 2018 23:28:34 +0000 (16:28 -0700)]
mm/hmm: remove HMM_PFN_READ flag and ignore peculiar architecture

Only peculiar architecture allow write without read thus assume that any
valid pfn do allow for read.  Note we do not care for write only because
it does make sense with thing like atomic compare and exchange or any
other operations that allow you to get the memory value through them.

Link: http://lkml.kernel.org/r/20180323005527.758-8-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: use struct for hmm_vma_fault(), hmm_vma_get_pfns() parameters
Jérôme Glisse [Tue, 10 Apr 2018 23:28:30 +0000 (16:28 -0700)]
mm/hmm: use struct for hmm_vma_fault(), hmm_vma_get_pfns() parameters

Both hmm_vma_fault() and hmm_vma_get_pfns() were taking a hmm_range struct
as parameter and were initializing that struct with others of their
parameters.  Have caller of those function do this as they are likely to
already do and only pass this struct to both function this shorten
function signature and make it easier in the future to add new parameters
by simply adding them to the structure.

Link: http://lkml.kernel.org/r/20180323005527.758-7-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: hmm_pfns_bad() was accessing wrong struct
Jérôme Glisse [Tue, 10 Apr 2018 23:28:27 +0000 (16:28 -0700)]
mm/hmm: hmm_pfns_bad() was accessing wrong struct

The private field of mm_walk struct point to an hmm_vma_walk struct and
not to the hmm_range struct desired.  Fix to get proper struct pointer.

Link: http://lkml.kernel.org/r/20180323005527.758-6-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: unregister mmu_notifier when last HMM client quit
Jérôme Glisse [Tue, 10 Apr 2018 23:28:23 +0000 (16:28 -0700)]
mm/hmm: unregister mmu_notifier when last HMM client quit

This code was lost in translation at one point.  This properly call
mmu_notifier_unregister_no_release() once last user is gone.  This fix the
zombie mm_struct as without this patch we do not drop the refcount we have
on it.

Link: http://lkml.kernel.org/r/20180323005527.758-5-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: HMM should have a callback before MM is destroyed
Ralph Campbell [Tue, 10 Apr 2018 23:28:19 +0000 (16:28 -0700)]
mm/hmm: HMM should have a callback before MM is destroyed

hmm_mirror_register() registers a callback for when the CPU pagetable is
modified.  Normally, the device driver will call hmm_mirror_unregister()
when the process using the device is finished.  However, if the process
exits uncleanly, the struct_mm can be destroyed with no warning to the
device driver.

Link: http://lkml.kernel.org/r/20180323005527.758-4-jglisse@redhat.com
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: fix header file if/else/endif maze
Jérôme Glisse [Tue, 10 Apr 2018 23:28:15 +0000 (16:28 -0700)]
mm/hmm: fix header file if/else/endif maze

The #if/#else/#endif for IS_ENABLED(CONFIG_HMM) were wrong.  Because of
this after multiple include there was multiple definition of both
hmm_mm_init() and hmm_mm_destroy() leading to build failure if HMM was
enabled (CONFIG_HMM set).

Link: http://lkml.kernel.org/r/20180323005527.758-3-jglisse@redhat.com
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Acked-by: Balbir Singh <bsingharora@gmail.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Ralph Campbell <rcampbell@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/hmm: documentation editorial update to HMM documentation
Ralph Campbell [Tue, 10 Apr 2018 23:28:11 +0000 (16:28 -0700)]
mm/hmm: documentation editorial update to HMM documentation

Update the documentation for HMM to fix minor typos and phrasing to be a
bit more readable.

Link: http://lkml.kernel.org/r/20180323005527.758-2-jglisse@redhat.com
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Signed-off-by: Jérôme Glisse <jglisse@redhat.com>
Cc: Stephen Bates <sbates@raithlin.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Evgeny Baskakov <ebaskakov@nvidia.com>
Cc: Mark Hairgrove <mhairgrove@nvidia.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm, vmscan, tracing: use pointer to reclaim_stat struct in trace event
Steven Rostedt [Tue, 10 Apr 2018 23:28:07 +0000 (16:28 -0700)]
mm, vmscan, tracing: use pointer to reclaim_stat struct in trace event

The trace event trace_mm_vmscan_lru_shrink_inactive() currently has 12
parameters! Seven of them are from the reclaim_stat structure.  This
structure is currently local to mm/vmscan.c.  By moving it to the global
vmstat.h header, we can also reference it from the vmscan tracepoints.
In moving it, it brings down the overhead of passing so many arguments
to the trace event.  In the future, we may limit the number of arguments
that a trace event may pass (ideally just 6, but more realistically it
may be 8).

Before this patch, the code to call the trace event is this:

 0f 83 aa fe ff ff       jae    ffffffff811e6261 <shrink_inactive_list+0x1e1>
 48 8b 45 a0             mov    -0x60(%rbp),%rax
 45 8b 64 24 20          mov    0x20(%r12),%r12d
 44 8b 6d d4             mov    -0x2c(%rbp),%r13d
 8b 4d d0                mov    -0x30(%rbp),%ecx
 44 8b 75 cc             mov    -0x34(%rbp),%r14d
 44 8b 7d c8             mov    -0x38(%rbp),%r15d
 48 89 45 90             mov    %rax,-0x70(%rbp)
 8b 83 b8 fe ff ff       mov    -0x148(%rbx),%eax
 8b 55 c0                mov    -0x40(%rbp),%edx
 8b 7d c4                mov    -0x3c(%rbp),%edi
 8b 75 b8                mov    -0x48(%rbp),%esi
 89 45 80                mov    %eax,-0x80(%rbp)
 65 ff 05 e4 f7 e2 7e    incl   %gs:0x7ee2f7e4(%rip)        # 15bd0 <__preempt_count>
 48 8b 05 75 5b 13 01    mov    0x1135b75(%rip),%rax        # ffffffff8231bf68 <__tracepoint_mm_vmscan_lru_shrink_inactive+0x28>
 48 85 c0                test   %rax,%rax
 74 72                   je     ffffffff811e646a <shrink_inactive_list+0x3ea>
 48 89 c3                mov    %rax,%rbx
 4c 8b 10                mov    (%rax),%r10
 89 f8                   mov    %edi,%eax
 48 89 85 68 ff ff ff    mov    %rax,-0x98(%rbp)
 89 f0                   mov    %esi,%eax
 48 89 85 60 ff ff ff    mov    %rax,-0xa0(%rbp)
 89 c8                   mov    %ecx,%eax
 48 89 85 78 ff ff ff    mov    %rax,-0x88(%rbp)
 89 d0                   mov    %edx,%eax
 48 89 85 70 ff ff ff    mov    %rax,-0x90(%rbp)
 8b 45 8c                mov    -0x74(%rbp),%eax
 48 8b 7b 08             mov    0x8(%rbx),%rdi
 48 83 c3 18             add    $0x18,%rbx
 50                      push   %rax
 41 54                   push   %r12
 41 55                   push   %r13
 ff b5 78 ff ff ff       pushq  -0x88(%rbp)
 41 56                   push   %r14
 41 57                   push   %r15
 ff b5 70 ff ff ff       pushq  -0x90(%rbp)
 4c 8b 8d 68 ff ff ff    mov    -0x98(%rbp),%r9
 4c 8b 85 60 ff ff ff    mov    -0xa0(%rbp),%r8
 48 8b 4d 98             mov    -0x68(%rbp),%rcx
 48 8b 55 90             mov    -0x70(%rbp),%rdx
 8b 75 80                mov    -0x80(%rbp),%esi
 41 ff d2                callq  *%r10

After the patch:

 0f 83 a8 fe ff ff       jae    ffffffff811e626d <shrink_inactive_list+0x1cd>
 8b 9b b8 fe ff ff       mov    -0x148(%rbx),%ebx
 45 8b 64 24 20          mov    0x20(%r12),%r12d
 4c 8b 6d a0             mov    -0x60(%rbp),%r13
 65 ff 05 f5 f7 e2 7e    incl   %gs:0x7ee2f7f5(%rip)        # 15bd0 <__preempt_count>
 4c 8b 35 86 5b 13 01    mov    0x1135b86(%rip),%r14        # ffffffff8231bf68 <__tracepoint_mm_vmscan_lru_shrink_inactive+0x28>
 4d 85 f6                test   %r14,%r14
 74 2a                   je     ffffffff811e6411 <shrink_inactive_list+0x371>
 49 8b 06                mov    (%r14),%rax
 8b 4d 8c                mov    -0x74(%rbp),%ecx
 49 8b 7e 08             mov    0x8(%r14),%rdi
 49 83 c6 18             add    $0x18,%r14
 4c 89 ea                mov    %r13,%rdx
 45 89 e1                mov    %r12d,%r9d
 4c 8d 45 b8             lea    -0x48(%rbp),%r8
 89 de                   mov    %ebx,%esi
 51                      push   %rcx
 48 8b 4d 98             mov    -0x68(%rbp),%rcx
 ff d0                   callq  *%rax

Link: http://lkml.kernel.org/r/2559d7cb-ec60-1200-2362-04fa34fd02bb@fb.com
Link: http://lkml.kernel.org/r/20180322121003.4177af15@gandalf.local.home
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Reported-by: Alexei Starovoitov <ast@fb.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Alexei Starovoitov <ast@fb.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/vmscan: don't mess with pgdat->flags in memcg reclaim
Andrey Ryabinin [Tue, 10 Apr 2018 23:28:03 +0000 (16:28 -0700)]
mm/vmscan: don't mess with pgdat->flags in memcg reclaim

memcg reclaim may alter pgdat->flags based on the state of LRU lists in
cgroup and its children.  PGDAT_WRITEBACK may force kswapd to sleep
congested_wait(), PGDAT_DIRTY may force kswapd to writeback filesystem
pages.  But the worst here is PGDAT_CONGESTED, since it may force all
direct reclaims to stall in wait_iff_congested().  Note that only kswapd
have powers to clear any of these bits.  This might just never happen if
cgroup limits configured that way.  So all direct reclaims will stall as
long as we have some congested bdi in the system.

Leave all pgdat->flags manipulations to kswapd.  kswapd scans the whole
pgdat, only kswapd can clear pgdat->flags once node is balanced, thus
it's reasonable to leave all decisions about node state to kswapd.

Why only kswapd? Why not allow to global direct reclaim change these
flags? It is because currently only kswapd can clear these flags.  I'm
less worried about the case when PGDAT_CONGESTED falsely not set, and
more worried about the case when it falsely set.  If direct reclaimer
sets PGDAT_CONGESTED, do we have guarantee that after the congestion
problem is sorted out, kswapd will be woken up and clear the flag? It
seems like there is no such guarantee.  E.g.  direct reclaimers may
eventually balance pgdat and kswapd simply won't wake up (see
wakeup_kswapd()).

Moving pgdat->flags manipulation to kswapd, means that cgroup2 recalim
now loses its congestion throttling mechanism.  Add per-cgroup
congestion state and throttle cgroup2 reclaimers if memcg is in
congestion state.

Currently there is no need in per-cgroup PGDAT_WRITEBACK and PGDAT_DIRTY
bits since they alter only kswapd behavior.

The problem could be easily demonstrated by creating heavy congestion in
one cgroup:

    echo "+memory" > /sys/fs/cgroup/cgroup.subtree_control
    mkdir -p /sys/fs/cgroup/congester
    echo 512M > /sys/fs/cgroup/congester/memory.max
    echo $$ > /sys/fs/cgroup/congester/cgroup.procs
    /* generate a lot of diry data on slow HDD */
    while true; do dd if=/dev/zero of=/mnt/sdb/zeroes bs=1M count=1024; done &
    ....
    while true; do dd if=/dev/zero of=/mnt/sdb/zeroes bs=1M count=1024; done &

and some job in another cgroup:

    mkdir /sys/fs/cgroup/victim
    echo 128M > /sys/fs/cgroup/victim/memory.max

    # time cat /dev/sda > /dev/null
    real    10m15.054s
    user    0m0.487s
    sys     1m8.505s

According to the tracepoint in wait_iff_congested(), the 'cat' spent 50%
of the time sleeping there.

With the patch, cat don't waste time anymore:

    # time cat /dev/sda > /dev/null
    real    5m32.911s
    user    0m0.411s
    sys     0m56.664s

[aryabinin@virtuozzo.com: congestion state should be per-node]
Link: http://lkml.kernel.org/r/20180406135215.10057-1-aryabinin@virtuozzo.com
[ayabinin@virtuozzo.com: make congestion state per-cgroup-per-node instead of just per-cgroup[
Link: http://lkml.kernel.org/r/20180406180254.8970-2-aryabinin@virtuozzo.com
Link: http://lkml.kernel.org/r/20180323152029.11084-5-aryabinin@virtuozzo.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Tejun Heo <tj@kernel.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/vmscan: don't change pgdat state on base of a single LRU list state
Andrey Ryabinin [Tue, 10 Apr 2018 23:27:59 +0000 (16:27 -0700)]
mm/vmscan: don't change pgdat state on base of a single LRU list state

We have separate LRU list for each memory cgroup.  Memory reclaim
iterates over cgroups and calls shrink_inactive_list() every inactive
LRU list.  Based on the state of a single LRU shrink_inactive_list() may
flag the whole node as dirty,congested or under writeback.  This is
obviously wrong and hurtful.  It's especially hurtful when we have
possibly small congested cgroup in system.  Than *all* direct reclaims
waste time by sleeping in wait_iff_congested().  And the more memcgs in
the system we have the longer memory allocation stall is, because
wait_iff_congested() called on each lru-list scan.

Sum reclaim stats across all visited LRUs on node and flag node as
dirty, congested or under writeback based on that sum.  Also call
congestion_wait(), wait_iff_congested() once per pgdat scan, instead of
once per lru-list scan.

This only fixes the problem for global reclaim case.  Per-cgroup reclaim
may alter global pgdat flags too, which is wrong.  But that is separate
issue and will be addressed in the next patch.

This change will not have any effect on a systems with all workload
concentrated in a single cgroup.

[aryabinin@virtuozzo.com: check nr_writeback against all nr_taken, not just file]
Link: http://lkml.kernel.org/r/20180406180254.8970-1-aryabinin@virtuozzo.com
Link: http://lkml.kernel.org/r/20180323152029.11084-4-aryabinin@virtuozzo.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Tejun Heo <tj@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/vmscan: remove redundant current_may_throttle() check
Andrey Ryabinin [Tue, 10 Apr 2018 23:27:55 +0000 (16:27 -0700)]
mm/vmscan: remove redundant current_may_throttle() check

Only kswapd can have non-zero nr_immediate, and current_may_throttle()
is always true for kswapd (PF_LESS_THROTTLE bit is never set) thus it's
enough to check stat.nr_immediate only.

Link: http://lkml.kernel.org/r/20180315164553.17856-4-aryabinin@virtuozzo.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Tejun Heo <tj@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm/vmscan: update stale comments
Andrey Ryabinin [Tue, 10 Apr 2018 23:27:51 +0000 (16:27 -0700)]
mm/vmscan: update stale comments

Update some comments that became stale since transiton from per-zone to
per-node reclaim.

Link: http://lkml.kernel.org/r/20180315164553.17856-2-aryabinin@virtuozzo.com
Signed-off-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Tejun Heo <tj@kernel.org>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm: treat indirectly reclaimable memory as free in overcommit logic
Roman Gushchin [Tue, 10 Apr 2018 23:27:47 +0000 (16:27 -0700)]
mm: treat indirectly reclaimable memory as free in overcommit logic

Indirectly reclaimable memory can consume a significant part of total
memory and it's actually reclaimable (it will be released under actual
memory pressure).

So, the overcommit logic should treat it as free.

Otherwise, it's possible to cause random system-wide memory allocation
failures by consuming a significant amount of memory by indirectly
reclaimable memory, e.g.  dentry external names.

If overcommit policy GUESS is used, it might be used for denial of
service attack under some conditions.

The following program illustrates the approach.  It causes the kernel to
allocate an unreclaimable kmalloc-256 chunk for each stat() call, so
that at some point the overcommit logic may start blocking large
allocation system-wide.

  int main()
  {
   char buf[256];
   unsigned long i;
   struct stat statbuf;

   buf[0] = '/';
   for (i = 1; i < sizeof(buf); i++)
   buf[i] = '_';

   for (i = 0; 1; i++) {
   sprintf(&buf[248], "%8lu", i);
   stat(buf, &statbuf);
   }

   return 0;
  }

This patch in combination with related indirectly reclaimable memory
patches closes this issue.

Link: http://lkml.kernel.org/r/20180313130041.8078-1-guro@fb.com
Signed-off-by: Roman Gushchin <guro@fb.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agodcache: account external names as indirectly reclaimable memory
Roman Gushchin [Tue, 10 Apr 2018 23:27:44 +0000 (16:27 -0700)]
dcache: account external names as indirectly reclaimable memory

I received a report about suspicious growth of unreclaimable slabs on
some machines.  I've found that it happens on machines with low memory
pressure, and these unreclaimable slabs are external names attached to
dentries.

External names are allocated using generic kmalloc() function, so they
are accounted as unreclaimable.  But they are held by dentries, which
are reclaimable, and they will be reclaimed under the memory pressure.

In particular, this breaks MemAvailable calculation, as it doesn't take
unreclaimable slabs into account.  This leads to a silly situation, when
a machine is almost idle, has no memory pressure and therefore has a big
dentry cache.  And the resulting MemAvailable is too low to start a new
workload.

To address the issue, the NR_INDIRECTLY_RECLAIMABLE_BYTES counter is
used to track the amount of memory, consumed by external names.  The
counter is increased in the dentry allocation path, if an external name
structure is allocated; and it's decreased in the dentry freeing path.

To reproduce the problem I've used the following Python script:

  import os

  for iter in range (0, 10000000):
      try:
          name = ("/some_long_name_%d" % iter) + "_" * 220
          os.stat(name)
      except Exception:
          pass

Without this patch:
  $ cat /proc/meminfo | grep MemAvailable
  MemAvailable:    7811688 kB
  $ python indirect.py
  $ cat /proc/meminfo | grep MemAvailable
  MemAvailable:    2753052 kB

With the patch:
  $ cat /proc/meminfo | grep MemAvailable
  MemAvailable:    7809516 kB
  $ python indirect.py
  $ cat /proc/meminfo | grep MemAvailable
  MemAvailable:    7749144 kB

[guro@fb.com: fix indirectly reclaimable memory accounting for CONFIG_SLOB]
Link: http://lkml.kernel.org/r/20180312194140.19517-1-guro@fb.com
[guro@fb.com: fix indirectly reclaimable memory accounting]
Link: http://lkml.kernel.org/r/20180313125701.7955-1-guro@fb.com
Link: http://lkml.kernel.org/r/20180305133743.12746-5-guro@fb.com
Signed-off-by: Roman Gushchin <guro@fb.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm: treat indirectly reclaimable memory as available in MemAvailable
Roman Gushchin [Tue, 10 Apr 2018 23:27:40 +0000 (16:27 -0700)]
mm: treat indirectly reclaimable memory as available in MemAvailable

Adjust /proc/meminfo MemAvailable calculation by adding the amount of
indirectly reclaimable memory (rounded to the PAGE_SIZE).

Link: http://lkml.kernel.org/r/20180305133743.12746-4-guro@fb.com
Signed-off-by: Roman Gushchin <guro@fb.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agomm: introduce NR_INDIRECTLY_RECLAIMABLE_BYTES
Roman Gushchin [Tue, 10 Apr 2018 23:27:36 +0000 (16:27 -0700)]
mm: introduce NR_INDIRECTLY_RECLAIMABLE_BYTES

Patch series "indirectly reclaimable memory", v2.

This patchset introduces the concept of indirectly reclaimable memory
and applies it to fix the issue of when a big number of dentries with
external names can significantly affect the MemAvailable value.

This patch (of 3):

Introduce a concept of indirectly reclaimable memory and adds the
corresponding memory counter and /proc/vmstat item.

Indirectly reclaimable memory is any sort of memory, used by the kernel
(except of reclaimable slabs), which is actually reclaimable, i.e.  will
be released under memory pressure.

The counter is in bytes, as it's not always possible to count such
objects in pages.  The name contains BYTES by analogy to
NR_KERNEL_STACK_KB.

Link: http://lkml.kernel.org/r/20180305133743.12746-2-guro@fb.com
Signed-off-by: Roman Gushchin <guro@fb.com>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Alexander Viro <viro@zeniv.linux.org.uk>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoMerge tag 'for-linus' of git://linux-c6x.org/git/projects/linux-c6x-upstreaming
Linus Torvalds [Tue, 10 Apr 2018 18:50:14 +0000 (11:50 -0700)]
Merge tag 'for-linus' of git://linux-c6x.org/git/projects/linux-c6x-upstreaming

Pull c6x updates from Mark Salter.

* tag 'for-linus' of git://linux-c6x.org/git/projects/linux-c6x-upstreaming:
  c6x: pass endianness info to sparse
  c6x: fix platforms/plldata.c get_coreid build error
  c6x: remove unused KTHREAD_SIZE definition

6 years agoMerge tag 'mips_4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/mips
Linus Torvalds [Tue, 10 Apr 2018 18:39:22 +0000 (11:39 -0700)]
Merge tag 'mips_4.17' of git://git./linux/kernel/git/jhogan/mips

Pull MIPS updates from James Hogan:
 "These are the main MIPS changes for 4.17. Rough overview:

   (1) generic platform: Add support for Microsemi Ocelot SoCs

   (2) crypto: Add CRC32 and CRC32C HW acceleration module

   (3) Various cleanups and misc improvements

  More detailed summary:

  Miscellaneous:
   - hang more efficiently on halt/powerdown/restart
   - pm-cps: Block system suspend when a JTAG probe is present
   - expand make help text for generic defconfigs
   - refactor handling of legacy defconfigs
   - determine the entry point from the ELF file header to fix microMIPS
     for certain toolchains
   - introduce isa-rev.h for MIPS_ISA_REV and use to simplify other code

  Minor cleanups:
   - DTS: boston/ci20: Unit name cleanups and correction
   - kdump: Make the default for PHYSICAL_START always 64-bit
   - constify gpio_led in Alchemy, AR7, and TXX9
   - silence a couple of W=1 warnings
   - remove duplicate includes

  Platform support:
  Generic platform:
   - add support for Microsemi Ocelot
   - dt-bindings: Add vendor prefix for Microsemi Corporation
   - dt-bindings: Add bindings for Microsemi SoCs
   - add ocelot SoC & PCB123 board DTS files
   - MAINTAINERS: Add entry for Microsemi MIPS SoCs
   - enable crc32-mips on r6 configs

  ath79:
   - fix AR724X_PLL_REG_PCIE_CONFIG offset

  BCM47xx:
   - firmware: Use mac_pton() for MAC address parsing
   - add Luxul XAP1500/XWR1750 WiFi LEDs
   - use standard reset button for Luxul XWR-1750

  BMIPS:
   - enable CONFIG_BRCMSTB_PM in bmips_stb_defconfig for build coverage
   - add STB PM, wake-up timer, watchdog DT nodes

  Octeon:
   - drop '.' after newlines in printk calls

  ralink:
   - pci-mt7621: Enable PCIe on MT7688"

* tag 'mips_4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/jhogan/mips: (37 commits)
  MIPS: BCM47XX: Use standard reset button for Luxul XWR-1750
  MIPS: BCM47XX: Add Luxul XAP1500/XWR1750 WiFi LEDs
  MIPS: Make the default for PHYSICAL_START always 64-bit
  MIPS: Use the entry point from the ELF file header
  MAINTAINERS: Add entry for Microsemi MIPS SoCs
  MIPS: generic: Add support for Microsemi Ocelot
  MIPS: mscc: Add ocelot PCB123 device tree
  MIPS: mscc: Add ocelot dtsi
  dt-bindings: mips: Add bindings for Microsemi SoCs
  dt-bindings: Add vendor prefix for Microsemi Corporation
  MIPS: ath79: Fix AR724X_PLL_REG_PCIE_CONFIG offset
  MIPS: pci-mt7620: Enable PCIe on MT7688
  MIPS: pm-cps: Block system suspend when a JTAG probe is present
  MIPS: VDSO: Replace __mips_isa_rev with MIPS_ISA_REV
  MIPS: BPF: Replace __mips_isa_rev with MIPS_ISA_REV
  MIPS: cpu-features.h: Replace __mips_isa_rev with MIPS_ISA_REV
  MIPS: Introduce isa-rev.h to define MIPS_ISA_REV
  MIPS: Hang more efficiently on halt/powerdown/restart
  FIRMWARE: bcm47xx_nvram: Replace mac address parsing
  MIPS: BMIPS: Add Broadcom STB watchdog nodes
  ...

6 years agoMerge tag 'trace-v4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt...
Linus Torvalds [Tue, 10 Apr 2018 18:27:30 +0000 (11:27 -0700)]
Merge tag 'trace-v4.17' of git://git./linux/kernel/git/rostedt/linux-trace

Pull tracing updates from Steven Rostedt:
 "New features:

   - Tom Zanussi's extended histogram work.

     This adds the synthetic events to have histograms from multiple
     event data Adds triggers "onmatch" and "onmax" to call the
     synthetic events Several updates to the histogram code from this

   - Allow way to nest ring buffer calls in the same context

   - Allow absolute time stamps in ring buffer

   - Rewrite of filter code parsing based on Al Viro's suggestions

   - Setting of trace_clock to global if TSC is unstable (on boot)

   - Better OOM handling when allocating large ring buffers

   - Added initcall tracepoints (consolidated initcall_debug code with
     them)

  And other various fixes and clean ups"

* tag 'trace-v4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace: (68 commits)
  init: Have initcall_debug still work without CONFIG_TRACEPOINTS
  init, tracing: Have printk come through the trace events for initcall_debug
  init, tracing: instrument security and console initcall trace events
  init, tracing: Add initcall trace events
  tracing: Add rcu dereference annotation for test func that touches filter->prog
  tracing: Add rcu dereference annotation for filter->prog
  tracing: Fixup logic inversion on setting trace_global_clock defaults
  tracing: Hide global trace clock from lockdep
  ring-buffer: Add set/clear_current_oom_origin() during allocations
  ring-buffer: Check if memory is available before allocation
  lockdep: Add print_irqtrace_events() to __warn
  vsprintf: Do not preprocess non-dereferenced pointers for bprintf (%px and %pK)
  tracing: Uninitialized variable in create_tracing_map_fields()
  tracing: Make sure variable string fields are NULL-terminated
  tracing: Add action comparisons when testing matching hist triggers
  tracing: Don't add flag strings when displaying variable references
  tracing: Fix display of hist trigger expressions containing timestamps
  ftrace: Drop a VLA in module_exists()
  tracing: Mention trace_clock=global when warning about unstable clocks
  tracing: Default to using trace_global_clock if sched_clock is unstable
  ...

6 years agoMerge tag 'libnvdimm-for-4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdim...
Linus Torvalds [Tue, 10 Apr 2018 17:25:57 +0000 (10:25 -0700)]
Merge tag 'libnvdimm-for-4.17' of git://git./linux/kernel/git/nvdimm/nvdimm

Pull libnvdimm updates from Dan Williams:
 "This cycle was was not something I ever want to repeat as there were
  several late changes that have only now just settled.

  Half of the branch up to commit d2c997c0f145 ("fs, dax: use
  page->mapping to warn...") have been in -next for several releases.
  The of_pmem driver and the address range scrub rework were late
  arrivals, and the dax work was scaled back at the last moment.

  The of_pmem driver missed a previous merge window due to an oversight.
  A sense of obligation to rectify that miss is why it is included for
  4.17. It has acks from PowerPC folks. Stephen reported a build failure
  that only occurs when merging it with your latest tree, for now I have
  fixed that up by disabling modular builds of of_pmem. A test merge
  with your tree has received a build success report from the 0day robot
  over 156 configs.

  An initial version of the ARS rework was submitted before the merge
  window. It is self contained to libnvdimm, a net code reduction, and
  passing all unit tests.

  The filesystem-dax changes are based on the wait_var_event()
  functionality from tip/sched/core. However, late review feedback
  showed that those changes regressed truncate performance to a large
  degree. The branch was rewound to drop the truncate behavior change
  and now only includes preparation patches and cleanups (with full acks
  and reviews). The finalization of this dax-dma-vs-trnucate work will
  need to wait for 4.18.

  Summary:

   - A rework of the filesytem-dax implementation provides for detection
     of unmap operations (truncate / hole punch) colliding with
     in-progress device-DMA. A fix for these collisions remains a
     work-in-progress pending resolution of truncate latency and
     starvation regressions.

   - The of_pmem driver expands the users of libnvdimm outside of x86
     and ACPI to describe an implementation of persistent memory on
     PowerPC with Open Firmware / Device tree.

   - Address Range Scrub (ARS) handling is completely rewritten to
     account for the fact that ARS may run for 100s of seconds and there
     is no platform defined way to cancel it. ARS will now no longer
     block namespace initialization.

   - The NVDIMM Namespace Label implementation is updated to handle
     label areas as small as 1K, down from 128K.

   - Miscellaneous cleanups and updates to unit test infrastructure"

* tag 'libnvdimm-for-4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm: (39 commits)
  libnvdimm, of_pmem: workaround OF_NUMA=n build error
  nfit, address-range-scrub: add module option to skip initial ars
  nfit, address-range-scrub: rework and simplify ARS state machine
  nfit, address-range-scrub: determine one platform max_ars value
  powerpc/powernv: Create platform devs for nvdimm buses
  doc/devicetree: Persistent memory region bindings
  libnvdimm: Add device-tree based driver
  libnvdimm: Add of_node to region and bus descriptors
  libnvdimm, region: quiet region probe
  libnvdimm, namespace: use a safe lookup for dimm device name
  libnvdimm, dimm: fix dpa reservation vs uninitialized label area
  libnvdimm, testing: update the default smart ctrl_temperature
  libnvdimm, testing: Add emulation for smart injection commands
  nfit, address-range-scrub: introduce nfit_spa->ars_state
  libnvdimm: add an api to cast a 'struct nd_region' to its 'struct device'
  nfit, address-range-scrub: fix scrub in-progress reporting
  dax, dm: allow device-mapper to operate without dax support
  dax: introduce CONFIG_DAX_DRIVER
  fs, dax: use page->mapping to warn if truncate collides with a busy page
  ext2, dax: introduce ext2_dax_aops
  ...

6 years agoMerge tag 'rtc-4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux
Linus Torvalds [Tue, 10 Apr 2018 17:22:27 +0000 (10:22 -0700)]
Merge tag 'rtc-4.17' of git://git./linux/kernel/git/abelloni/linux

Pull RTC updates from Alexandre Belloni:
 "This contains a few series that have been in preparation for a while
  and that will help systems with RTCs that will fail in 2038, 2069 or
  2100.

  Subsystem:
   - Add tracepoints
   - Rework of the RTC/nvmem API to allow drivers to discard struct
     nvmem_config after registration
   - New range API, drivers can now expose the useful range of the RTC
   - New offset API the core is now able to add an offset to the RTC
     time, modifying the supported range.
   - Multiple rtc_time64_to_tm fixes
   - Handle time_t overflow on 32 bit platforms in the core instead of
     letting drivers do crazy things.
   - remove rtc_control API

  New driver:
   - Intersil ISL12026

  Drivers:
   - Drivers exposing the RTC non volatile memory have been converted to
     use nvmem
   - Removed useless time and date validation
   - Removed an indirection pattern that was a cargo cult from ancient
     drivers
   - Removed VLA usage
   - Fixed a possible race condition in probe functions
   - AB8540 support is dropped from ab8500
   - pcf85363 now has alarm support"

* tag 'rtc-4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/abelloni/linux: (128 commits)
  rtc: snvs: Fix usage of snvs_rtc_enable
  rtc: mt7622: fix module autoloading for OF platform drivers
  rtc: isl12022: use true and false for boolean values
  rtc: ab8500: Drop AB8540 support
  rtc: remove a warning during scripts/kernel-doc step
  rtc: 88pm860x: remove artificial limitation
  rtc: 88pm80x: remove artificial limitation
  rtc: st-lpc: remove artificial limitation
  rtc: mrst: remove artificial limitation
  rtc: mv: remove artificial limitation
  rtc: hctosys: Ensure system time doesn't overflow time_t
  parisc: time: stop validating rtc_time in .read_time
  rtc: pcf85063: fix clearing bits in pcf85063_start_clock
  rtc: at91sam9: Set name of regmap_config
  rtc: s5m: Remove VLA usage
  rtc: s5m: Move enum from rtc.h to rtc-s5m.c
  rtc: remove VLA usage
  rtc: Add useful timestamp definitions
  rtc: Add one offset seconds to expand RTC range
  rtc: Factor out the RTC range validation into rtc_valid_range()
  ...

6 years agoMerge tag 'fbdev-v4.17' of git://github.com/bzolnier/linux
Linus Torvalds [Tue, 10 Apr 2018 17:20:00 +0000 (10:20 -0700)]
Merge tag 'fbdev-v4.17' of git://github.com/bzolnier/linux

Pull fbdev updates from Bartlomiej Zolnierkiewicz:
 "There is nothing really major here, just a couple of small bugfixes,
  improvements and cleanups:

   - make it possible to load radeonfb driver when offb driver is loaded
     first (Mathieu Malaterre)

   - fix memory leak in offb driver (Mathieu Malaterre)

   - fix unaligned access in udlfb driver (Ladislav Michl)

   - convert atmel_lcdfb driver to use GPIO descriptors (Ludovic
     Desroches)

   - avoid mismatched prototypes in sisfb driver (Arnd Bergmann)

   - remove VLA usage from viafb driver (Gustavo A. R. Silva)

   - add missing help text to FB_I810_I2 config option (Ulf Magnusson)

   - misc fixes (Gustavo A. R. Silva, Colin Ian King, Markus Elfring)

   - remove dead code from s3c-fb driver for Exynos and S5PV210
     platforms

   - misc cleanups (Corentin Labbe, Ladislav Michl, Ulf Magnusson,
     Vladimir Zapolskiy, Markus Elfring)"

* tag 'fbdev-v4.17' of git://github.com/bzolnier/linux: (32 commits)
  video: fbdev: s3c-fb: remove dead platform code for Exynos and S5PV210 platforms
  video: au1100fb: Delete an unnecessary variable initialisation in au1100fb_drv_probe()
  video: au1100fb: Improve a size determination in au1100fb_drv_probe()
  video: au1100fb: Delete an error message for a failed memory allocation in au1100fb_drv_probe()
  video/console/sticore: Delete an error message for a failed memory allocation in sti_try_rom_generic()
  video: ARM CLCD: Improve a size determination in clcdfb_probe()
  video: ARM CLCD: Delete an error message for a failed memory allocation in clcdfb_probe()
  video: matroxfb: Delete an error message for a failed memory allocation in matroxfb_crtc2_probe()
  video: s3c-fb: Improve a size determination in s3c_fb_probe()
  video: s3c-fb: Delete an error message for a failed memory allocation in s3c_fb_probe()
  video: fsl-diu-fb: Delete an error message for a failed memory allocation in fsl_diu_init()
  video: ssd1307fb: Improve a size determination in ssd1307fb_probe()
  video: smscufx: Delete an error message for a failed memory allocation in ufx_realloc_framebuffer()
  video: smscufx: Return an error code only as a constant in ufx_realloc_framebuffer()
  video: smscufx: Less checks in ufx_usb_probe() after error detection
  video: udlfb: Return an error code only as a constant in dlfb_realloc_framebuffer()
  video/fbdev/stifb: Delete an error message for a failed memory allocation in stifb_init_fb()
  video/fbdev/stifb: Return -ENOMEM after a failed kzalloc() in stifb_init_fb()
  video: fbdev: aty128fb: use true and false for boolean values
  fbdev: aty: fix missing indentation in if statement
  ...

6 years agoMerge tag 'sound-fix-4.17-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Tue, 10 Apr 2018 17:16:04 +0000 (10:16 -0700)]
Merge tag 'sound-fix-4.17-rc1' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "The main purpose of this pull request is a fix for a regression in the
  recent PCM OSS emulation code that may lead to RCU stall. Since
  syzkaller hits this too often, I send the pull request now with a
  minimal collection. Possibly another pull request may follow before
  RC1.

  The other fixes here are for USB-audio class 2 and 3 to improve the
  parser for the clock descriptors. These are rather cleanups but good
  for security, too.

  Last but not least, another included fix is the trivial one to remove
  superfluous WARN_ON() that annoyed syzbot"

* tag 'sound-fix-4.17-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: pcm: Remove WARN_ON() at snd_pcm_hw_params() error
  ALSA: pcm: Fix endless loop for XRUN recovery in OSS emulation
  ALSA: usb-audio: Add sanity checks in UAC3 clock parsers
  ALSA: usb-audio: More strict sanity checks for clock parsers
  ALSA: usb-audio: Refactor clock finder helpers

6 years agoMerge tag 'media/v4.17-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Tue, 10 Apr 2018 17:10:30 +0000 (10:10 -0700)]
Merge tag 'media/v4.17-2' of git://git./linux/kernel/git/mchehab/linux-media

Pull media fixes from Mauro Carvalho Chehab:
 "A series of media updates/fixes for 4.17.

  There are two important core fix patches in this series:

   - A regression fix on Kernel 4.16 with causes it to not work with
     some input devices that depend on media core

   - A fix at compat32 bits with causes it to OOPS on overlay, and
     affects the Kernels where the CVE-2017-13166 was backported

  The remaining ones are other random fixes at the documentation and on
  drivers.

  The biggest part of this series is a set of 18 patches for the Intel
  atomisp driver. Currently, it produces hundreds of warnings/errors on
  sparse/smatch, causing me to sometimes ignore new warnings on other
  drivers that are not so broken. This driver is on really poor state,
  even for staging standards: it has several layers of abstraction on
  it, and it supports two different hardware. Selecting between them
  require to add a define (there isn't even a Kconfig option for such
  purpose). Just on this smatch cleanup, I could easily get rid of 8
  "do-nothing" files. So, I'm seriously considering its removal from
  upstream, if I don't see any real work on addressing the problems
  there along this year"

* tag 'media/v4.17-2' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media: (48 commits)
  media: v4l2-core: fix size of devnode_nums[] bitarray
  media: v4l2-compat-ioctl32: don't oops on overlay
  media: i2c: adv748x: afe: fix sparse warning
  media: extended-controls.rst: transmitter -> receiver
  media: staging: atomisp: stop duplicating input format types
  media: staging: atomisp: get rid of an unused var
  media: staging: atomisp: stop mixing enum types
  media: staging: atomisp: get rid of some static warnings
  media: staging: atomisp: use %p to print pointers
  media: staging: atomisp: remove an useless check
  media: staging: atomisp: avoid a warning if 32 bits build
  media: staging: atomisp: don't access a NULL var
  media: staging: atomisp: Get rid of *default.host.[ch]
  media: staging: atomisp: get rid of an unused function
  media: staging: atomisp: remove unused set_pd_base()
  media: staging: atomisp: fix endianess issues
  media: staging: atomisp: add a missing include
  media: staging: atomisp: get rid of stupid statements
  media: staging: atomisp: declare static vars as such
  media: staging: atomisp: ia_css_output.host: don't use var before check
  ...

6 years agoc6x: pass endianness info to sparse
Luc Van Oostenryck [Tue, 31 Oct 2017 17:22:02 +0000 (18:22 +0100)]
c6x: pass endianness info to sparse

c6x depends on the macro '_BIG_ENDIAN' being defined or not
to correctly select or define endian-specific macros, structures
or pieces of code.

This macro is predefined by the compiler but sparse knows nothing
about it and thus may pre-process files differently from what
gcc would.

Fix this by passing '-D_BIG_ENDIAN' when compiling a big-endian
kernel, like GCC would have done.

To: Mark Salter <msalter@redhat.com>
To: Aurelien Jacquiot <a-jacquiot@ti.com>
CC: linux-c6x-dev@linux-c6x.org
Signed-off-by: Luc Van Oostenryck <luc.vanoostenryck@gmail.com>
Signed-off-by: Mark Salter <msalter@redhat.com>
6 years agoc6x: fix platforms/plldata.c get_coreid build error
Randy Dunlap [Fri, 19 Jan 2018 19:19:33 +0000 (11:19 -0800)]
c6x: fix platforms/plldata.c get_coreid build error

Fix build error reported by the 0day bot by including the header
file for that macro.

Fixes this build error: (should fix; not tested)
arch/c6x/platforms/plldata.c: In function 'c6472_setup_clocks':
arch/c6x/platforms/plldata.c:279:33: error: implicit declaration of function 'get_coreid'; did you mean 'get_order'? [-Werror=implicit-function-declaration]
      c6x_core_clk.parent = &sysclks[get_coreid() + 1];

Reported-by: Fengguang Wu <fengguang.wu@intel.com>
Cc: Mark Salter <msalter@redhat.com>
Cc: Aurelien Jacquiot <jacquiot.aurelien@gmail.com>
Cc: linux-c6x-dev@linux-c6x.org
Cc: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Mark Salter <msalter@redhat.com>
6 years agoc6x: remove unused KTHREAD_SIZE definition
Jérémy Lefaure [Mon, 4 Sep 2017 19:14:12 +0000 (15:14 -0400)]
c6x: remove unused KTHREAD_SIZE definition

KTHREAD_SIZE has never been used since it has been defined for c6x arch.
Let's remove this useless definition.

Signed-off-by: Jérémy Lefaure <jeremy.lefaure@lse.epita.fr>
Signed-off-by: Mark Salter <msalter@redhat.com>
6 years agoMerge tag 'hwmon-for-linus-v4.17' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Tue, 10 Apr 2018 02:59:54 +0000 (19:59 -0700)]
Merge tag 'hwmon-for-linus-v4.17' of git://git./linux/kernel/git/groeck/linux-staging

Pull hwmon updates from Guenter Roeck:

 - added chip support: new Centaur CPUs, ADM1272, NCT6796D

 - ucd9000: added debugfs attributes, gpio support

 - cleanup and minor bug fixes

* tag 'hwmon-for-linus-v4.17' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (via-cputemp) support new centaur CPUs
  hwmon: (nct6775) Fix writing pwmX_mode
  hwmon: (lm92) Add max6635 to lm92_id[]
  hwmon: (pmbus/adm1275) Add support for ADM1272
  hwmon: (lm92) Do not try to detect MAX6635
  hwmon: (ucd9000) Add debugfs attributes to provide mfr_status
  hwmon: (ucd9000) Add gpio chip interface
  hwmon: (nct6775) Add support for NCT6796D
  hwmon: (nct6775) Initialize boolean variables with declaration
  hwmon: (nct6775) Improve fan6/pwm6 support
  hwmon: (nct6775) Use NUM_FAN consistently
  hwmon: (g762) handle cleanup with devm_add_action
  hwmon: (sht3x) Update data sheet URL
  hwmon: (sht21) Update data sheet URLs
  hwmon: (pmbus/adm1275) Accept negative page register values
  hwmon: (pmbus/max8688) Accept negative page register values

6 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net
Linus Torvalds [Tue, 10 Apr 2018 00:04:10 +0000 (17:04 -0700)]
Merge git://git./linux/kernel/git/davem/net

Pull networking fixes from David Miller:

 1) The sockmap code has to free socket memory on close if there is
    corked data, from John Fastabend.

 2) Tunnel names coming from userspace need to be length validated. From
    Eric Dumazet.

 3) arp_filter() has to take VRFs properly into account, from Miguel
    Fadon Perlines.

 4) Fix oops in error path of tcf_bpf_init(), from Davide Caratti.

 5) Missing idr_remove() in u32_delete_key(), from Cong Wang.

 6) More syzbot stuff. Several use of uninitialized value fixes all
    over, from Eric Dumazet.

 7) Do not leak kernel memory to userspace in sctp, also from Eric
    Dumazet.

 8) Discard frames from unused ports in DSA, from Andrew Lunn.

 9) Fix DMA mapping and reset/failover problems in ibmvnic, from Thomas
    Falcon.

10) Do not access dp83640 PHY registers prematurely after reset, from
    Esben Haabendal.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net: (46 commits)
  vhost-net: set packet weight of tx polling to 2 * vq size
  net: thunderx: rework mac addresses list to u64 array
  inetpeer: fix uninit-value in inet_getpeer
  dp83640: Ensure against premature access to PHY registers after reset
  devlink: convert occ_get op to separate registration
  ARM: dts: ls1021a: Specify TBIPA register address
  net/fsl_pq_mdio: Allow explicit speficition of TBIPA address
  ibmvnic: Do not reset CRQ for Mobility driver resets
  ibmvnic: Fix failover case for non-redundant configuration
  ibmvnic: Fix reset scheduler error handling
  ibmvnic: Zero used TX descriptor counter on reset
  ibmvnic: Fix DMA mapping mistakes
  tipc: use the right skb in tipc_sk_fill_sock_diag()
  sctp: sctp_sockaddr_af must check minimal addr length for AF_INET6
  net: dsa: Discard frames from unused ports
  sctp: do not leak kernel memory to user space
  soreuseport: initialise timewait reuseport field
  ipv4: fix uninit-value in ip_route_output_key_hash_rcu()
  dccp: initialize ireq->ir_mark
  net: fix uninit-value in __hw_addr_add_ex()
  ...

6 years agoMerge branch 'work.namei' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs
Linus Torvalds [Mon, 9 Apr 2018 19:48:05 +0000 (12:48 -0700)]
Merge branch 'work.namei' of git://git./linux/kernel/git/viro/vfs

Pull vfs namei updates from Al Viro:

 - make lookup_one_len() safe with parent locked only shared(incoming
   afs series wants that)

 - fix of getname_kernel() regression from 2015 (-stable fodder, that
   one).

* 'work.namei' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  getname_kernel() needs to make sure that ->name != ->iname in long case
  make lookup_one_len() safe to use with directory locked shared
  new helper: __lookup_slow()
  merge common parts of lookup_one_len{,_unlocked} into common helper

6 years agoMerge tag 'for-linus-4.17-ofs' of git://git.kernel.org/pub/scm/linux/kernel/git/hubca...
Linus Torvalds [Mon, 9 Apr 2018 19:45:47 +0000 (12:45 -0700)]
Merge tag 'for-linus-4.17-ofs' of git://git./linux/kernel/git/hubcap/linux

Pull orangefs updates from Mike Marshall:
 "Fixes and cleanups:

   - Documentation cleanups

   - removal of unused code

   - make some structs static

   - implement Orangefs vm_operations fault callout

   - eliminate two single-use functions and put their cleaned up code in
     line.

   - replace a vmalloc/memset instance with vzalloc

   - fix a race condition bug in wait code"

* tag 'for-linus-4.17-ofs' of git://git.kernel.org/pub/scm/linux/kernel/git/hubcap/linux:
  Orangefs: documentation updates
  orangefs: document package install and xfstests procedure
  orangefs: remove unused code
  orangefs: make several *_operations structs static
  orangefs: implement vm_ops->fault
  orangefs: open code short single-use functions
  orangefs: replace vmalloc and memset with vzalloc
  orangefs: bug fix for a race condition when getting a slot

6 years agoMerge tag 'pstore-v4.17-rc1-fix' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Mon, 9 Apr 2018 19:43:18 +0000 (12:43 -0700)]
Merge tag 'pstore-v4.17-rc1-fix' of git://git./linux/kernel/git/kees/linux

Pull pstore fix from Kees Cook:
 "Fix another compression Kconfig combination missed in testing (Tobias
  Regnery)"

* tag 'pstore-v4.17-rc1-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  pstore: fix crypto dependencies without compression

6 years agoselinux: fix missing dput() before selinuxfs unmount
Stephen Smalley [Mon, 9 Apr 2018 18:36:05 +0000 (14:36 -0400)]
selinux: fix missing dput() before selinuxfs unmount

Commit 0619f0f5e36f ("selinux: wrap selinuxfs state") triggers a BUG
when SELinux is runtime-disabled (i.e. systemd or equivalent disables
SELinux before initial policy load via /sys/fs/selinux/disable based on
/etc/selinux/config SELINUX=disabled).

This does not manifest if SELinux is disabled via kernel command line
argument or if SELinux is enabled (permissive or enforcing).

Before:
  SELinux:  Disabled at runtime.
  BUG: Dentry 000000006d77e5c7{i=17,n=null}  still in use (1) [unmount of selinuxfs selinuxfs]

After:
  SELinux:  Disabled at runtime.

Fixes: 0619f0f5e36f ("selinux: wrap selinuxfs state")
Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Stephen Smalley <sds@tycho.nsa.gov>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Mon, 9 Apr 2018 18:42:31 +0000 (11:42 -0700)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm updates from Paolo Bonzini:
 "ARM:
   - VHE optimizations

   - EL2 address space randomization

   - speculative execution mitigations ("variant 3a", aka execution past
     invalid privilege register access)

   - bugfixes and cleanups

  PPC:
   - improvements for the radix page fault handler for HV KVM on POWER9

  s390:
   - more kvm stat counters

   - virtio gpu plumbing

   - documentation

   - facilities improvements

  x86:
   - support for VMware magic I/O port and pseudo-PMCs

   - AMD pause loop exiting

   - support for AMD core performance extensions

   - support for synchronous register access

   - expose nVMX capabilities to userspace

   - support for Hyper-V signaling via eventfd

   - use Enlightened VMCS when running on Hyper-V

   - allow userspace to disable MWAIT/HLT/PAUSE vmexits

   - usual roundup of optimizations and nested virtualization bugfixes

  Generic:
   - API selftest infrastructure (though the only tests are for x86 as
     of now)"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (174 commits)
  kvm: x86: fix a prototype warning
  kvm: selftests: add sync_regs_test
  kvm: selftests: add API testing infrastructure
  kvm: x86: fix a compile warning
  KVM: X86: Add Force Emulation Prefix for "emulate the next instruction"
  KVM: X86: Introduce handle_ud()
  KVM: vmx: unify adjacent #ifdefs
  x86: kvm: hide the unused 'cpu' variable
  KVM: VMX: remove bogus WARN_ON in handle_ept_misconfig
  Revert "KVM: X86: Fix SMRAM accessing even if VM is shutdown"
  kvm: Add emulation for movups/movupd
  KVM: VMX: raise internal error for exception during invalid protected mode state
  KVM: nVMX: Optimization: Dont set KVM_REQ_EVENT when VMExit with nested_run_pending
  KVM: nVMX: Require immediate-exit when event reinjected to L2 and L1 event pending
  KVM: x86: Fix misleading comments on handling pending exceptions
  KVM: x86: Rename interrupt.pending to interrupt.injected
  KVM: VMX: No need to clear pending NMI/interrupt on inject realmode interrupt
  x86/kvm: use Enlightened VMCS when running on Hyper-V
  x86/hyper-v: detect nested features
  x86/hyper-v: define struct hv_enlightened_vmcs and clean field bits
  ...

6 years agoMerge branch 'for-4.17/dax' into libnvdimm-for-next
Dan Williams [Mon, 9 Apr 2018 17:50:17 +0000 (10:50 -0700)]
Merge branch 'for-4.17/dax' into libnvdimm-for-next

6 years agoMerge branch 'for-4.17/libnvdimm' into libnvdimm-for-next
Dan Williams [Mon, 9 Apr 2018 17:50:08 +0000 (10:50 -0700)]
Merge branch 'for-4.17/libnvdimm' into libnvdimm-for-next

6 years agoFix subtle macro variable shadowing in min_not_zero()
Linus Torvalds [Mon, 9 Apr 2018 17:34:07 +0000 (10:34 -0700)]
Fix subtle macro variable shadowing in min_not_zero()

Commit 3c8ba0d61d04 ("kernel.h: Retain constant expression output for
max()/min()") rewrote our min/max macros to be very clever, but in the
meantime resurrected a variable name shadow issue that we had had
previously fixed in commit 589a9785ee3a ("min/max: remove sparse
warnings when they're nested").

That commit talks about the sparse warnings that this shadowing causes,
which we ignored as just a minor annoyance.  But it turns out that the
sparse warning is the least of our problems.  We actually have a real
bug due to the shadowing through the interaction with "min_not_zero()",
which ends up doing

   min(__x, __y)

internally, and then the new declaration of "__x" and "__y" as new
variables in __cmp_once() results in a complete mess of an expression,
and "min_not_zero()" doesn't work at all.

For some odd reason, this only ever caused (reported) problems on s390,
even though it is a generic issue and most of the (obviously successful)
testing of the problematic commit had happened on other architectures.

Quoting Sebastian Ott:
 "What happened is that the bio build by the partition detection code
  was attempted to be split by the block layer because the block queue
  had a max_sector setting of 0. blk_queue_max_hw_sectors uses
  min_not_zero."

So re-introduce the use of __UNIQUE_ID() to make sure that the min/max
macros do not have these kinds of clashes.

[ That said, __UNIQUE_ID() itself has several issues that make it less
  than wonderful.

  In particular, the "uniqueness" has a fallback on the line number,
  which means that it's not actually unique in more complex cases if you
  don't build with gcc or clang (which have working unique counters that
  aren't tied to line numbers).

  That historical broken fallback also means that we have that pointless
  "prefix" argument that doesn't actually make much sense _except_ for
  the known-broken case. Oh well. ]

Fixes: 3c8ba0d61d04 ("kernel.h: Retain constant expression output for max()/min()")
Reported-and-tested-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Cc: Kees Cook <keescook@chromium.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
6 years agoMerge branch 'for-linus-sa1100' of git://git.armlinux.org.uk/~rmk/linux-arm
Linus Torvalds [Mon, 9 Apr 2018 16:26:36 +0000 (09:26 -0700)]
Merge branch 'for-linus-sa1100' of git://git.armlinux.org.uk/~rmk/linux-arm

Pull ARM SA1100 updates from Russell King:
 "We have support for arbitary MMIO registers providing platform GPIOs,
  which allows us to abstract some of the SA11x0 CF support.

  This set of updates makes that change"

* 'for-linus-sa1100' of git://git.armlinux.org.uk/~rmk/linux-arm:
  ARM: sa1100/simpad: switch simpad CF to use gpiod APIs
  ARM: sa1100/shannon: convert to generic CF sockets
  ARM: sa1100/nanoengine: convert to generic CF sockets
  ARM: sa1100/h3xxx: switch h3xxx PCMCIA to use gpiod APIs
  ARM: sa1100/cerf: convert to generic CF sockets
  ARM: sa1100/assabet: convert to generic CF sockets
  ARM: sa1100: provide infrastructure to support generic CF sockets
  pcmcia: sa1100: provide generic CF support

6 years agoMerge branch 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm
Linus Torvalds [Mon, 9 Apr 2018 16:19:30 +0000 (09:19 -0700)]
Merge branch 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm

Pull ARM updates from Russell King:
 "A number of core ARM changes:

   - Refactoring linker script by Nicolas Pitre

   - Enable source fortification

   - Add support for Cortex R8"

* 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm:
  ARM: decompressor: fix warning introduced in fortify patch
  ARM: 8751/1: Add support for Cortex-R8 processor
  ARM: 8749/1: Kconfig: Add ARCH_HAS_FORTIFY_SOURCE
  ARM: simplify and fix linker script for TCM
  ARM: linker script: factor out TCM bits
  ARM: linker script: factor out vectors and stubs
  ARM: linker script: factor out unwinding table sections
  ARM: linker script: factor out stuff for the .text section
  ARM: linker script: factor out stuff for the DISCARD section
  ARM: linker script: factor out some common definitions between XIP and non-XIP

6 years agoMerge branch 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu
Linus Torvalds [Mon, 9 Apr 2018 16:15:46 +0000 (09:15 -0700)]
Merge branch 'for-next' of git://git./linux/kernel/git/gerg/m68knommu

Pull m68knommu update from Greg Ungerer:
 "Only a single fix to set the DMA masks in the ColdFire FEC platform
  data structure.

  This stops the warning from dma-mapping.h at boot time"

* 'for-next' of git://git.kernel.org/pub/scm/linux/kernel/git/gerg/m68knommu:
  m68k: set dma and coherent masks for platform FEC ethernets

6 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mattst88...
Linus Torvalds [Mon, 9 Apr 2018 16:11:32 +0000 (09:11 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/mattst88/alpha

Pull alpha updates from Matt Turner:
 "A few small changes for alpha"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mattst88/alpha:
  alpha: io: reorder barriers to guarantee writeX() and iowriteX() ordering
  alpha: Implement CPU vulnerabilities sysfs functions.
  alpha: rtc: stop validating rtc_time in .read_time
  alpha: rtc: remove unused set_mmss ops

6 years agolibnvdimm, of_pmem: workaround OF_NUMA=n build error
Dan Williams [Mon, 9 Apr 2018 16:10:22 +0000 (09:10 -0700)]
libnvdimm, of_pmem: workaround OF_NUMA=n build error

Stephen reports that an x86 allmodconfig build fails to build the
of_pmem driver due to a missing definition of of_node_to_nid(). That
helper is currently only exported in the OF_NUMA=y case. In other cases,
ppc and sparc, it is a weak symbol, and outside of those platforms it is
a static inline.

Until an OF_NUMA=n configuration can reliably support usage of
of_node_to_nid() in modules across architectures, mark this driver as
'bool' instead of 'tristate'.

Cc: Rob Herring <robh@kernel.org>
Cc: Oliver O'Halloran <oohall@gmail.com>
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
6 years agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Mon, 9 Apr 2018 16:04:10 +0000 (09:04 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/s390/linux

Pull s390 updates from Martin Schwidefsky:

 - Improvements for the spectre defense:
    * The spectre related code is consolidated to a single file
      nospec-branch.c
    * Automatic enable/disable for the spectre v2 defenses (expoline vs.
      nobp)
    * Syslog messages for specve v2 are added
    * Enable CONFIG_GENERIC_CPU_VULNERABILITIES and define the attribute
      functions for spectre v1 and v2

 - Add helper macros for assembler alternatives and use them to shorten
   the code in entry.S.

 - Add support for persistent configuration data via the SCLP Store Data
   interface. The H/W interface requires a page table that uses 4K pages
   only, the code to setup such an address space is added as well.

 - Enable virtio GPU emulation in QEMU. To do this the depends
   statements for a few common Kconfig options are modified.

 - Add support for format-3 channel path descriptors and add a binary
   sysfs interface to export the associated utility strings.

 - Add a sysfs attribute to control the IFCC handling in case of
   constant channel errors.

 - The vfio-ccw changes from Cornelia.

 - Bug fixes and cleanups.

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux: (40 commits)
  s390/kvm: improve stack frame constants in entry.S
  s390/lpp: use assembler alternatives for the LPP instruction
  s390/entry.S: use assembler alternatives
  s390: add assembler macros for CPU alternatives
  s390: add sysfs attributes for spectre
  s390: report spectre mitigation via syslog
  s390: add automatic detection of the spectre defense
  s390: move nobp parameter functions to nospec-branch.c
  s390/cio: add util_string sysfs attribute
  s390/chsc: query utility strings via fmt3 channel path descriptor
  s390/cio: rename struct channel_path_desc
  s390/cio: fix unbind of io_subchannel_driver
  s390/qdio: split up CCQ handling for EQBS / SQBS
  s390/qdio: don't retry EQBS after CCQ 96
  s390/qdio: restrict buffer merging to eligible devices
  s390/qdio: don't merge ERROR output buffers
  s390/qdio: simplify math in get_*_buffer_frontier()
  s390/decompressor: trim uncompressed image head during the build
  s390/crypto: Fix kernel crash on aes_s390 module remove.
  s390/defkeymap: fix global init to zero
  ...

6 years agoALSA: pcm: Remove WARN_ON() at snd_pcm_hw_params() error
Takashi Iwai [Mon, 9 Apr 2018 15:12:16 +0000 (17:12 +0200)]
ALSA: pcm: Remove WARN_ON() at snd_pcm_hw_params() error

snd_pcm_hw_params() (more exactly snd_pcm_hw_params_choose()) contains
a check of the return error from snd_pcm_hw_param_first() and _last()
with snd_BUG_ON() -- i.e. it may trigger WARN_ON() depending on the
kconfig.

This was a valid check in the past, as these functions shouldn't
return any error if the parameters have been already refined via
snd_pcm_hw_refine() beforehand.  However, the recent rewrite
introduced a kmalloc() in snd_pcm_hw_refine() for removing VLA, and
this brought a possibility to trigger an error.  As a result, syzbot
caught lots of superfluous kernel WARN_ON() and paniced via fault
injection.

As the WARN_ON() is no longer valid with the introduction of
kmalloc(), let's drop snd_BUG_ON() check, in order to make the world
peaceful place again.

Reported-by: syzbot+803e0047ac3a3096bb4f@syzkaller.appspotmail.com
Fixes: 5730f9f744cf ("ALSA: pcm: Remove VLA usage")
Signed-off-by: Takashi Iwai <tiwai@suse.de>
6 years agovhost-net: set packet weight of tx polling to 2 * vq size
haibinzhang(张海斌) [Mon, 9 Apr 2018 07:22:17 +0000 (07:22 +0000)]
vhost-net: set packet weight of tx polling to 2 * vq size

handle_tx will delay rx for tens or even hundreds of milliseconds when tx busy
polling udp packets with small length(e.g. 1byte udp payload), because setting
VHOST_NET_WEIGHT takes into account only sent-bytes but no single packet length.

Ping-Latencies shown below were tested between two Virtual Machines using
netperf (UDP_STREAM, len=1), and then another machine pinged the client:

vq size=256
Packet-Weight   Ping-Latencies(millisecond)
                   min      avg       max
Origin           3.319   18.489    57.303
64               1.643    2.021     2.552
128              1.825    2.600     3.224
256              1.997    2.710     4.295
512              1.860    3.171     4.631
1024             2.002    4.173     9.056
2048             2.257    5.650     9.688
4096             2.093    8.508    15.943

vq size=512
Packet-Weight   Ping-Latencies(millisecond)
                   min      avg       max
Origin           6.537   29.177    66.245
64               2.798    3.614     4.403
128              2.861    3.820     4.775
256              3.008    4.018     4.807
512              3.254    4.523     5.824
1024             3.079    5.335     7.747
2048             3.944    8.201    12.762
4096             4.158   11.057    19.985

Seems pretty consistent, a small dip at 2 VQ sizes.
Ring size is a hint from device about a burst size it can tolerate. Based on
benchmarks, set the weight to 2 * vq size.

To evaluate this change, another tests were done using netperf(RR, TX) between
two machines with Intel(R) Xeon(R) Gold 6133 CPU @ 2.50GHz, and vq size was
tweaked through qemu. Results shown below does not show obvious changes.

vq size=256 TCP_RR                vq size=512 TCP_RR
size/sessions/+thu%/+normalize%   size/sessions/+thu%/+normalize%
   1/       1/  -7%/        -2%      1/       1/   0%/        -2%
   1/       4/  +1%/         0%      1/       4/  +1%/         0%
   1/       8/  +1%/        -2%      1/       8/   0%/        +1%
  64/       1/  -6%/         0%     64/       1/  +7%/        +3%
  64/       4/   0%/        +2%     64/       4/  -1%/        +1%
  64/       8/   0%/         0%     64/       8/  -1%/        -2%
 256/       1/  -3%/        -4%    256/       1/  -4%/        -2%
 256/       4/  +3%/        +4%    256/       4/  +1%/        +2%
 256/       8/  +2%/         0%    256/       8/  +1%/        -1%

vq size=256 UDP_RR                vq size=512 UDP_RR
size/sessions/+thu%/+normalize%   size/sessions/+thu%/+normalize%
   1/       1/  -5%/        +1%      1/       1/  -3%/        -2%
   1/       4/  +4%/        +1%      1/       4/  -2%/        +2%
   1/       8/  -1%/        -1%      1/       8/  -1%/         0%
  64/       1/  -2%/        -3%     64/       1/  +1%/        +1%
  64/       4/  -5%/        -1%     64/       4/  +2%/         0%
  64/       8/   0%/        -1%     64/       8/  -2%/        +1%
 256/       1/  +7%/        +1%    256/       1/  -7%/         0%
 256/       4/  +1%/        +1%    256/       4/  -3%/        -4%
 256/       8/  +2%/        +2%    256/       8/  +1%/        +1%

vq size=256 TCP_STREAM            vq size=512 TCP_STREAM
size/sessions/+thu%/+normalize%   size/sessions/+thu%/+normalize%
  64/       1/   0%/        -3%     64/       1/   0%/         0%
  64/       4/  +3%/        -1%     64/       4/  -2%/        +4%
  64/       8/  +9%/        -4%     64/       8/  -1%/        +2%
 256/       1/  +1%/        -4%    256/       1/  +1%/        +1%
 256/       4/  -1%/        -1%    256/       4/  -3%/         0%
 256/       8/  +7%/        +5%    256/       8/  -3%/         0%
 512/       1/  +1%/         0%    512/       1/  -1%/        -1%
 512/       4/  +1%/        -1%    512/       4/   0%/         0%
 512/       8/  +7%/        -5%    512/       8/  +6%/        -1%
1024/       1/   0%/        -1%   1024/       1/   0%/        +1%
1024/       4/  +3%/         0%   1024/       4/  +1%/         0%
1024/       8/  +8%/        +5%   1024/       8/  -1%/         0%
2048/       1/  +2%/        +2%   2048/       1/  -1%/         0%
2048/       4/  +1%/         0%   2048/       4/   0%/        -1%
2048/       8/  -2%/         0%   2048/       8/   5%/        -1%
4096/       1/  -2%/         0%   4096/       1/  -2%/         0%
4096/       4/  +2%/         0%   4096/       4/   0%/         0%
4096/       8/  +9%/        -2%   4096/       8/  -5%/        -1%

Acked-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Haibin Zhang <haibinzhang@tencent.com>
Signed-off-by: Yunfang Tai <yunfangtai@tencent.com>
Signed-off-by: Lidong Chen <lidongchen@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
6 years agonet: thunderx: rework mac addresses list to u64 array
Vadim Lomovtsev [Mon, 9 Apr 2018 13:24:48 +0000 (06:24 -0700)]
net: thunderx: rework mac addresses list to u64 array

It is too expensive to pass u64 values via linked list, instead
allocate array for them by overall number of mac addresses from netdev.

This eventually removes multiple kmalloc() calls, aviod memory
fragmentation and allow to put single null check on kmalloc
return value in order to prevent a potential null pointer dereference.

Addresses-Coverity-ID: 1467429 ("Dereference null return value")
Fixes: 37c3347eb247 ("net: thunderx: add ndo_set_rx_mode callback implementation for VF")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Vadim Lomovtsev <Vadim.Lomovtsev@cavium.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
6 years agoinetpeer: fix uninit-value in inet_getpeer
Eric Dumazet [Mon, 9 Apr 2018 13:43:27 +0000 (06:43 -0700)]
inetpeer: fix uninit-value in inet_getpeer

syzbot/KMSAN reported that p->dtime was read while it was
not yet initialized in :

delta = (__u32)jiffies - p->dtime;
if (delta < ttl || !refcount_dec_if_one(&p->refcnt))
gc_stack[i] = NULL;

This is a false positive, because the inetpeer wont be erased
from rb-tree if the refcount_dec_if_one(&p->refcnt) does not
succeed. And this wont happen before first inet_putpeer() call
for this inetpeer has been done, and ->dtime field is written
exactly before the refcount_dec_and_test(&p->refcnt).

The KMSAN report was :

BUG: KMSAN: uninit-value in inet_peer_gc net/ipv4/inetpeer.c:163 [inline]
BUG: KMSAN: uninit-value in inet_getpeer+0x1567/0x1e70 net/ipv4/inetpeer.c:228
CPU: 0 PID: 9494 Comm: syz-executor5 Not tainted 4.16.0+ #82
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:17 [inline]
 dump_stack+0x185/0x1d0 lib/dump_stack.c:53
 kmsan_report+0x142/0x240 mm/kmsan/kmsan.c:1067
 __msan_warning_32+0x6c/0xb0 mm/kmsan/kmsan_instr.c:676
 inet_peer_gc net/ipv4/inetpeer.c:163 [inline]
 inet_getpeer+0x1567/0x1e70 net/ipv4/inetpeer.c:228
 inet_getpeer_v4 include/net/inetpeer.h:110 [inline]
 icmpv4_xrlim_allow net/ipv4/icmp.c:330 [inline]
 icmp_send+0x2b44/0x3050 net/ipv4/icmp.c:725
 ip_options_compile+0x237c/0x29f0 net/ipv4/ip_options.c:472
 ip_rcv_options net/ipv4/ip_input.c:284 [inline]
 ip_rcv_finish+0xda8/0x16d0 net/ipv4/ip_input.c:365
 NF_HOOK include/linux/netfilter.h:288 [inline]
 ip_rcv+0x119d/0x16f0 net/ipv4/ip_input.c:493
 __netif_receive_skb_core+0x47cf/0x4a80 net/core/dev.c:4562
 __netif_receive_skb net/core/dev.c:4627 [inline]
 netif_receive_skb_internal+0x49d/0x630 net/core/dev.c:4701
 netif_receive_skb+0x230/0x240 net/core/dev.c:4725
 tun_rx_batched drivers/net/tun.c:1555 [inline]
 tun_get_user+0x6d88/0x7580 drivers/net/tun.c:1962
 tun_chr_write_iter+0x1d4/0x330 drivers/net/tun.c:1990
 do_iter_readv_writev+0x7bb/0x970 include/linux/fs.h:1776
 do_iter_write+0x30d/0xd40 fs/read_write.c:932
 vfs_writev fs/read_write.c:977 [inline]
 do_writev+0x3c9/0x830 fs/read_write.c:1012
 SYSC_writev+0x9b/0xb0 fs/read_write.c:1085
 SyS_writev+0x56/0x80 fs/read_write.c:1082
 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x3d/0xa2
RIP: 0033:0x455111
RSP: 002b:00007fae0365cba0 EFLAGS: 00000293 ORIG_RAX: 0000000000000014
RAX: ffffffffffffffda RBX: 000000000000002e RCX: 0000000000455111
RDX: 0000000000000001 RSI: 00007fae0365cbf0 RDI: 00000000000000fc
RBP: 0000000020000040 R08: 00000000000000fc R09: 0000000000000000
R10: 000000000000002e R11: 0000000000000293 R12: 00000000ffffffff
R13: 0000000000000658 R14: 00000000006fc8e0 R15: 0000000000000000

Uninit was created at:
 kmsan_save_stack_with_flags mm/kmsan/kmsan.c:278 [inline]
 kmsan_internal_poison_shadow+0xb8/0x1b0 mm/kmsan/kmsan.c:188
 kmsan_kmalloc+0x94/0x100 mm/kmsan/kmsan.c:314
 kmem_cache_alloc+0xaab/0xb90 mm/slub.c:2756
 inet_getpeer+0xed8/0x1e70 net/ipv4/inetpeer.c:210
 inet_getpeer_v4 include/net/inetpeer.h:110 [inline]
 ip4_frag_init+0x4d1/0x740 net/ipv4/ip_fragment.c:153
 inet_frag_alloc net/ipv4/inet_fragment.c:369 [inline]
 inet_frag_create net/ipv4/inet_fragment.c:385 [inline]
 inet_frag_find+0x7da/0x1610 net/ipv4/inet_fragment.c:418
 ip_find net/ipv4/ip_fragment.c:275 [inline]
 ip_defrag+0x448/0x67a0 net/ipv4/ip_fragment.c:676
 ip_check_defrag+0x775/0xda0 net/ipv4/ip_fragment.c:724
 packet_rcv_fanout+0x2a8/0x8d0 net/packet/af_packet.c:1447
 deliver_skb net/core/dev.c:1897 [inline]
 deliver_ptype_list_skb net/core/dev.c:1912 [inline]
 __netif_receive_skb_core+0x314a/0x4a80 net/core/dev.c:4545
 __netif_receive_skb net/core/dev.c:4627 [inline]
 netif_receive_skb_internal+0x49d/0x630 net/core/dev.c:4701
 netif_receive_skb+0x230/0x240 net/core/dev.c:4725
 tun_rx_batched drivers/net/tun.c:1555 [inline]
 tun_get_user+0x6d88/0x7580 drivers/net/tun.c:1962
 tun_chr_write_iter+0x1d4/0x330 drivers/net/tun.c:1990
 do_iter_readv_writev+0x7bb/0x970 include/linux/fs.h:1776
 do_iter_write+0x30d/0xd40 fs/read_write.c:932
 vfs_writev fs/read_write.c:977 [inline]
 do_writev+0x3c9/0x830 fs/read_write.c:1012
 SYSC_writev+0x9b/0xb0 fs/read_write.c:1085
 SyS_writev+0x56/0x80 fs/read_write.c:1082
 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287
 entry_SYSCALL_64_after_hwframe+0x3d/0xa2

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
6 years agoMerge branches 'devel-stable' and 'misc' into for-linus
Russell King [Mon, 9 Apr 2018 09:08:51 +0000 (10:08 +0100)]
Merge branches 'devel-stable' and 'misc' into for-linus

6 years agodp83640: Ensure against premature access to PHY registers after reset
Esben Haabendal [Sun, 8 Apr 2018 20:17:01 +0000 (22:17 +0200)]
dp83640: Ensure against premature access to PHY registers after reset

The datasheet specifies a 3uS pause after performing a software
reset. The default implementation of genphy_soft_reset() does not
provide this, so implement soft_reset with the needed pause.

Signed-off-by: Esben Haabendal <eha@deif.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
6 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
David S. Miller [Sun, 8 Apr 2018 23:51:41 +0000 (19:51 -0400)]
Merge git://git./pub/scm/linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2018-04-09

The following pull-request contains BPF updates for your *net* tree.

The main changes are:

1) Two sockmap fixes: i) fix a potential warning when a socket with
   pending cork data is closed by freeing the memory right when the
   socket is closed instead of seeing still outstanding memory at
   garbage collector time, ii) fix a NULL pointer deref in case of
   duplicates release calls, so make sure to only reset the sk_prot
   pointer when it's in a valid state to do so, both from John.

2) Fix a compilation warning in bpf_prog_attach_check_attach_type()
   by moving the function under CONFIG_CGROUP_BPF ifdef since only
   used there, from Anders.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
6 years agoMerge branch 'for-upstream' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetoot...
David S. Miller [Sun, 8 Apr 2018 21:19:15 +0000 (17:19 -0400)]
Merge branch 'for-upstream' of git://git./linux/kernel/git/bluetooth/bluetooth

Johan Hedberg says:

====================
pull request: bluetooth 2018-04-08

Here's one important Bluetooth fix for the 4.17-rc series that's needed
to pass several Bluetooth qualification test cases.

Let me know if there are any issues pulling. Thanks.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
6 years agodevlink: convert occ_get op to separate registration
Jiri Pirko [Thu, 5 Apr 2018 20:13:21 +0000 (22:13 +0200)]
devlink: convert occ_get op to separate registration

This resolves race during initialization where the resources with
ops are registered before driver and the structures used by occ_get
op is initialized. So keep occ_get callbacks registered only when
all structs are initialized.

The example flows, as it is in mlxsw:
1) driver load/asic probe:
   mlxsw_core
      -> mlxsw_sp_resources_register
        -> mlxsw_sp_kvdl_resources_register
          -> devlink_resource_register IDX
   mlxsw_spectrum
      -> mlxsw_sp_kvdl_init
        -> mlxsw_sp_kvdl_parts_init
          -> mlxsw_sp_kvdl_part_init
            -> devlink_resource_size_get IDX (to get the current setup
                                              size from devlink)
        -> devlink_resource_occ_get_register IDX (register current
                                                  occupancy getter)
2) reload triggered by devlink command:
  -> mlxsw_devlink_core_bus_device_reload
    -> mlxsw_sp_fini
      -> mlxsw_sp_kvdl_fini
-> devlink_resource_occ_get_unregister IDX
    (struct mlxsw_sp *mlxsw_sp is freed at this point, call to occ get
     which is using mlxsw_sp would cause use-after free)
    -> mlxsw_sp_init
      -> mlxsw_sp_kvdl_init
        -> mlxsw_sp_kvdl_parts_init
          -> mlxsw_sp_kvdl_part_init
            -> devlink_resource_size_get IDX (to get the current setup
                                              size from devlink)
        -> devlink_resource_occ_get_register IDX (register current
                                                  occupancy getter)

Fixes: d9f9b9a4d05f ("devlink: Add support for resource abstraction")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>