muen/linux.git
10 months agomm: workingset: fix vmstat counters for shadow nodes
Roman Gushchin [Tue, 13 Aug 2019 22:37:41 +0000 (15:37 -0700)]
mm: workingset: fix vmstat counters for shadow nodes

Memcg counters for shadow nodes are broken because the memcg pointer is
obtained in a wrong way. The following approach is used:
        virt_to_page(xa_node)->mem_cgroup

Since commit 4d96ba353075 ("mm: memcg/slab: stop setting
page->mem_cgroup pointer for slab pages") page->mem_cgroup pointer isn't
set for slab pages, so memcg_from_slab_page() should be used instead.

Also I doubt that it ever worked correctly: virt_to_head_page() should
be used instead of virt_to_page().  Otherwise objects residing on tail
pages are not accounted, because only the head page contains a valid
mem_cgroup pointer.  That was a case since the introduction of these
counters by the commit 68d48e6a2df5 ("mm: workingset: add vmstat counter
for shadow nodes").

Link: http://lkml.kernel.org/r/20190801233532.138743-1-guro@fb.com
Fixes: 4d96ba353075 ("mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages")
Signed-off-by: Roman Gushchin <guro@fb.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm/usercopy: use memory range to be accessed for wraparound check
Isaac J. Manjarres [Tue, 13 Aug 2019 22:37:37 +0000 (15:37 -0700)]
mm/usercopy: use memory range to be accessed for wraparound check

Currently, when checking to see if accessing n bytes starting at address
"ptr" will cause a wraparound in the memory addresses, the check in
check_bogus_address() adds an extra byte, which is incorrect, as the
range of addresses that will be accessed is [ptr, ptr + (n - 1)].

This can lead to incorrectly detecting a wraparound in the memory
address, when trying to read 4 KB from memory that is mapped to the the
last possible page in the virtual address space, when in fact, accessing
that range of memory would not cause a wraparound to occur.

Use the memory range that will actually be accessed when considering if
accessing a certain amount of bytes will cause the memory address to
wrap around.

Link: http://lkml.kernel.org/r/1564509253-23287-1-git-send-email-isaacm@codeaurora.org
Fixes: f5509cc18daa ("mm: Hardened usercopy")
Signed-off-by: Prasad Sodagudi <psodagud@codeaurora.org>
Signed-off-by: Isaac J. Manjarres <isaacm@codeaurora.org>
Co-developed-by: Prasad Sodagudi <psodagud@codeaurora.org>
Reviewed-by: William Kucharski <william.kucharski@oracle.com>
Acked-by: Kees Cook <keescook@chromium.org>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Trilok Soni <tsoni@codeaurora.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm: kmemleak: disable early logging in case of error
Catalin Marinas [Tue, 13 Aug 2019 22:37:34 +0000 (15:37 -0700)]
mm: kmemleak: disable early logging in case of error

If an error occurs during kmemleak_init() (e.g.  kmem cache cannot be
created), kmemleak is disabled but kmemleak_early_log remains enabled.
Subsequently, when the .init.text section is freed, the log_early()
function no longer exists.  To avoid a page fault in such scenario,
ensure that kmemleak_disable() also disables early logging.

Link: http://lkml.kernel.org/r/20190731152302.42073-1-catalin.marinas@arm.com
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Reported-by: Qian Cai <cai@lca.pw>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm/vmalloc.c: fix percpu free VM area search criteria
Kuppuswamy Sathyanarayanan [Tue, 13 Aug 2019 22:37:31 +0000 (15:37 -0700)]
mm/vmalloc.c: fix percpu free VM area search criteria

Recent changes to the vmalloc code by commit 68ad4a330433
("mm/vmalloc.c: keep track of free blocks for vmap allocation") can
cause spurious percpu allocation failures.  These, in turn, can result
in panic()s in the slub code.  One such possible panic was reported by
Dave Hansen in following link https://lkml.org/lkml/2019/6/19/939.
Another related panic observed is,

 RIP: 0033:0x7f46f7441b9b
 Call Trace:
  dump_stack+0x61/0x80
  pcpu_alloc.cold.30+0x22/0x4f
  mem_cgroup_css_alloc+0x110/0x650
  cgroup_apply_control_enable+0x133/0x330
  cgroup_mkdir+0x41b/0x500
  kernfs_iop_mkdir+0x5a/0x90
  vfs_mkdir+0x102/0x1b0
  do_mkdirat+0x7d/0xf0
  do_syscall_64+0x5b/0x180
  entry_SYSCALL_64_after_hwframe+0x44/0xa9

VMALLOC memory manager divides the entire VMALLOC space (VMALLOC_START
to VMALLOC_END) into multiple VM areas (struct vm_areas), and it mainly
uses two lists (vmap_area_list & free_vmap_area_list) to track the used
and free VM areas in VMALLOC space.  And pcpu_get_vm_areas(offsets[],
sizes[], nr_vms, align) function is used for allocating congruent VM
areas for percpu memory allocator.  In order to not conflict with
VMALLOC users, pcpu_get_vm_areas allocates VM areas near the end of the
VMALLOC space.  So the search for free vm_area for the given requirement
starts near VMALLOC_END and moves upwards towards VMALLOC_START.

Prior to commit 68ad4a330433, the search for free vm_area in
pcpu_get_vm_areas() involves following two main steps.

Step 1:
    Find a aligned "base" adress near VMALLOC_END.
    va = free vm area near VMALLOC_END
Step 2:
    Loop through number of requested vm_areas and check,
        Step 2.1:
           if (base < VMALLOC_START)
              1. fail with error
        Step 2.2:
           // end is offsets[area] + sizes[area]
           if (base + end > va->vm_end)
               1. Move the base downwards and repeat Step 2
        Step 2.3:
           if (base + start < va->vm_start)
              1. Move to previous free vm_area node, find aligned
                 base address and repeat Step 2

But Commit 68ad4a330433 removed Step 2.2 and modified Step 2.3 as below:

        Step 2.3:
           if (base + start < va->vm_start || base + end > va->vm_end)
              1. Move to previous free vm_area node, find aligned
                 base address and repeat Step 2

Above change is the root cause of spurious percpu memory allocation
failures.  For example, consider a case where a relatively large vm_area
(~ 30 TB) was ignored in free vm_area search because it did not pass the
base + end < vm->vm_end boundary check.  Ignoring such large free
vm_area's would lead to not finding free vm_area within boundary of
VMALLOC_start to VMALLOC_END which in turn leads to allocation failures.

So modify the search algorithm to include Step 2.2.

Link: http://lkml.kernel.org/r/20190729232139.91131-1-sathyanarayanan.kuppuswamy@linux.intel.com
Fixes: 68ad4a330433 ("mm/vmalloc.c: keep track of free blocks for vmap allocation")
Signed-off-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Reported-by: Dave Hansen <dave.hansen@intel.com>
Acked-by: Dennis Zhou <dennis@kernel.org>
Reviewed-by: Uladzislau Rezki (Sony) <urezki@gmail.com>
Cc: Roman Gushchin <guro@fb.com>
Cc: sathyanarayanan kuppuswamy <sathyanarayanan.kuppuswamy@linux.intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm/memcontrol.c: fix use after free in mem_cgroup_iter()
Miles Chen [Tue, 13 Aug 2019 22:37:28 +0000 (15:37 -0700)]
mm/memcontrol.c: fix use after free in mem_cgroup_iter()

This patch is sent to report an use after free in mem_cgroup_iter()
after merging commit be2657752e9e ("mm: memcg: fix use after free in
mem_cgroup_iter()").

I work with android kernel tree (4.9 & 4.14), and commit be2657752e9e
("mm: memcg: fix use after free in mem_cgroup_iter()") has been merged
to the trees.  However, I can still observe use after free issues
addressed in the commit be2657752e9e.  (on low-end devices, a few times
this month)

backtrace:
        css_tryget <- crash here
        mem_cgroup_iter
        shrink_node
        shrink_zones
        do_try_to_free_pages
        try_to_free_pages
        __perform_reclaim
        __alloc_pages_direct_reclaim
        __alloc_pages_slowpath
        __alloc_pages_nodemask

To debug, I poisoned mem_cgroup before freeing it:

  static void __mem_cgroup_free(struct mem_cgroup *memcg)
        for_each_node(node)
        free_mem_cgroup_per_node_info(memcg, node);
        free_percpu(memcg->stat);
  +     /* poison memcg before freeing it */
  +     memset(memcg, 0x78, sizeof(struct mem_cgroup));
        kfree(memcg);
  }

The coredump shows the position=0xdbbc2a00 is freed.

  (gdb) p/x ((struct mem_cgroup_per_node *)0xe5009e00)->iter[8]
  $13 = {position = 0xdbbc2a00, generation = 0x2efd}

  0xdbbc2a00:     0xdbbc2e00      0x00000000      0xdbbc2800      0x00000100
  0xdbbc2a10:     0x00000200      0x78787878      0x00026218      0x00000000
  0xdbbc2a20:     0xdcad6000      0x00000001      0x78787800      0x00000000
  0xdbbc2a30:     0x78780000      0x00000000      0x0068fb84      0x78787878
  0xdbbc2a40:     0x78787878      0x78787878      0x78787878      0xe3fa5cc0
  0xdbbc2a50:     0x78787878      0x78787878      0x00000000      0x00000000
  0xdbbc2a60:     0x00000000      0x00000000      0x00000000      0x00000000
  0xdbbc2a70:     0x00000000      0x00000000      0x00000000      0x00000000
  0xdbbc2a80:     0x00000000      0x00000000      0x00000000      0x00000000
  0xdbbc2a90:     0x00000001      0x00000000      0x00000000      0x00100000
  0xdbbc2aa0:     0x00000001      0xdbbc2ac8      0x00000000      0x00000000
  0xdbbc2ab0:     0x00000000      0x00000000      0x00000000      0x00000000
  0xdbbc2ac0:     0x00000000      0x00000000      0xe5b02618      0x00001000
  0xdbbc2ad0:     0x00000000      0x78787878      0x78787878      0x78787878
  0xdbbc2ae0:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2af0:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b00:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b10:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b20:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b30:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b40:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b50:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b60:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b70:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2b80:     0x78787878      0x78787878      0x00000000      0x78787878
  0xdbbc2b90:     0x78787878      0x78787878      0x78787878      0x78787878
  0xdbbc2ba0:     0x78787878      0x78787878      0x78787878      0x78787878

In the reclaim path, try_to_free_pages() does not setup
sc.target_mem_cgroup and sc is passed to do_try_to_free_pages(), ...,
shrink_node().

In mem_cgroup_iter(), root is set to root_mem_cgroup because
sc->target_mem_cgroup is NULL.  It is possible to assign a memcg to
root_mem_cgroup.nodeinfo.iter in mem_cgroup_iter().

        try_to_free_pages
         struct scan_control sc = {...}, target_mem_cgroup is 0x0;
        do_try_to_free_pages
        shrink_zones
        shrink_node
          mem_cgroup *root = sc->target_mem_cgroup;
          memcg = mem_cgroup_iter(root, NULL, &reclaim);
        mem_cgroup_iter()
         if (!root)
         root = root_mem_cgroup;
         ...

         css = css_next_descendant_pre(css, &root->css);
         memcg = mem_cgroup_from_css(css);
         cmpxchg(&iter->position, pos, memcg);

My device uses memcg non-hierarchical mode.  When we release a memcg:
invalidate_reclaim_iterators() reaches only dead_memcg and its parents.
If non-hierarchical mode is used, invalidate_reclaim_iterators() never
reaches root_mem_cgroup.

  static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg)
  {
        struct mem_cgroup *memcg = dead_memcg;

        for (; memcg; memcg = parent_mem_cgroup(memcg)
        ...
  }

So the use after free scenario looks like:

  CPU1 CPU2

  try_to_free_pages
  do_try_to_free_pages
  shrink_zones
  shrink_node
  mem_cgroup_iter()
      if (!root)
       root = root_mem_cgroup;
      ...
      css = css_next_descendant_pre(css, &root->css);
      memcg = mem_cgroup_from_css(css);
      cmpxchg(&iter->position, pos, memcg);

         invalidate_reclaim_iterators(memcg);
         ...
         __mem_cgroup_free()
         kfree(memcg);

  try_to_free_pages
  do_try_to_free_pages
  shrink_zones
  shrink_node
  mem_cgroup_iter()
      if (!root)
       root = root_mem_cgroup;
      ...
      mz = mem_cgroup_nodeinfo(root, reclaim->pgdat->node_id);
      iter = &mz->iter[reclaim->priority];
      pos = READ_ONCE(iter->position);
      css_tryget(&pos->css) <- use after free

To avoid this, we should also invalidate root_mem_cgroup.nodeinfo.iter
in invalidate_reclaim_iterators().

[cai@lca.pw: fix -Wparentheses compilation warning]
Link: http://lkml.kernel.org/r/1564580753-17531-1-git-send-email-cai@lca.pw
Link: http://lkml.kernel.org/r/20190730015729.4406-1-miles.chen@mediatek.com
Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting")
Signed-off-by: Miles Chen <miles.chen@mediatek.com>
Signed-off-by: Qian Cai <cai@lca.pw>
Acked-by: Michal Hocko <mhocko@suse.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm/z3fold.c: fix z3fold_destroy_pool() race condition
Henry Burns [Tue, 13 Aug 2019 22:37:25 +0000 (15:37 -0700)]
mm/z3fold.c: fix z3fold_destroy_pool() race condition

The constraint from the zpool use of z3fold_destroy_pool() is there are
no outstanding handles to memory (so no active allocations), but it is
possible for there to be outstanding work on either of the two wqs in
the pool.

Calling z3fold_deregister_migration() before the workqueues are drained
means that there can be allocated pages referencing a freed inode,
causing any thread in compaction to be able to trip over the bad pointer
in PageMovable().

Link: http://lkml.kernel.org/r/20190726224810.79660-2-henryburns@google.com
Fixes: 1f862989b04a ("mm/z3fold.c: support page migration")
Signed-off-by: Henry Burns <henryburns@google.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Reviewed-by: Jonathan Adams <jwadams@google.com>
Cc: Vitaly Vul <vitaly.vul@sony.com>
Cc: Vitaly Wool <vitalywool@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Henry Burns <henrywolfeburns@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm/z3fold.c: fix z3fold_destroy_pool() ordering
Henry Burns [Tue, 13 Aug 2019 22:37:21 +0000 (15:37 -0700)]
mm/z3fold.c: fix z3fold_destroy_pool() ordering

The constraint from the zpool use of z3fold_destroy_pool() is there are
no outstanding handles to memory (so no active allocations), but it is
possible for there to be outstanding work on either of the two wqs in
the pool.

If there is work queued on pool->compact_workqueue when it is called,
z3fold_destroy_pool() will do:

   z3fold_destroy_pool()
     destroy_workqueue(pool->release_wq)
     destroy_workqueue(pool->compact_wq)
       drain_workqueue(pool->compact_wq)
         do_compact_page(zhdr)
           kref_put(&zhdr->refcount)
             __release_z3fold_page(zhdr, ...)
               queue_work_on(pool->release_wq, &pool->work) *BOOM*

So compact_wq needs to be destroyed before release_wq.

Link: http://lkml.kernel.org/r/20190726224810.79660-1-henryburns@google.com
Fixes: 5d03a6613957 ("mm/z3fold.c: use kref to prevent page free/compact race")
Signed-off-by: Henry Burns <henryburns@google.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Reviewed-by: Jonathan Adams <jwadams@google.com>
Cc: Vitaly Vul <vitaly.vul@sony.com>
Cc: Vitaly Wool <vitalywool@gmail.com>
Cc: David Howells <dhowells@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Al Viro <viro@zeniv.linux.org.uk
Cc: Henry Burns <henrywolfeburns@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm: mempolicy: handle vma with unmovable pages mapped correctly in mbind
Yang Shi [Tue, 13 Aug 2019 22:37:18 +0000 (15:37 -0700)]
mm: mempolicy: handle vma with unmovable pages mapped correctly in mbind

When running syzkaller internally, we ran into the below bug on 4.9.x
kernel:

  kernel BUG at mm/huge_memory.c:2124!
  invalid opcode: 0000 [#1] SMP KASAN
  CPU: 0 PID: 1518 Comm: syz-executor107 Not tainted 4.9.168+ #2
  Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 0.5.1 01/01/2011
  task: ffff880067b34900 task.stack: ffff880068998000
  RIP: split_huge_page_to_list+0x8fb/0x1030 mm/huge_memory.c:2124
  Call Trace:
    split_huge_page include/linux/huge_mm.h:100 [inline]
    queue_pages_pte_range+0x7e1/0x1480 mm/mempolicy.c:538
    walk_pmd_range mm/pagewalk.c:50 [inline]
    walk_pud_range mm/pagewalk.c:90 [inline]
    walk_pgd_range mm/pagewalk.c:116 [inline]
    __walk_page_range+0x44a/0xdb0 mm/pagewalk.c:208
    walk_page_range+0x154/0x370 mm/pagewalk.c:285
    queue_pages_range+0x115/0x150 mm/mempolicy.c:694
    do_mbind mm/mempolicy.c:1241 [inline]
    SYSC_mbind+0x3c3/0x1030 mm/mempolicy.c:1370
    SyS_mbind+0x46/0x60 mm/mempolicy.c:1352
    do_syscall_64+0x1d2/0x600 arch/x86/entry/common.c:282
    entry_SYSCALL_64_after_swapgs+0x5d/0xdb
  Code: c7 80 1c 02 00 e8 26 0a 76 01 <0f> 0b 48 c7 c7 40 46 45 84 e8 4c
  RIP  [<ffffffff81895d6b>] split_huge_page_to_list+0x8fb/0x1030 mm/huge_memory.c:2124
   RSP <ffff88006899f980>

with the below test:

  uint64_t r[1] = {0xffffffffffffffff};

  int main(void)
  {
        syscall(__NR_mmap, 0x20000000, 0x1000000, 3, 0x32, -1, 0);
                                intptr_t res = 0;
        res = syscall(__NR_socket, 0x11, 3, 0x300);
        if (res != -1)
                r[0] = res;
        *(uint32_t*)0x20000040 = 0x10000;
        *(uint32_t*)0x20000044 = 1;
        *(uint32_t*)0x20000048 = 0xc520;
        *(uint32_t*)0x2000004c = 1;
        syscall(__NR_setsockopt, r[0], 0x107, 0xd, 0x20000040, 0x10);
        syscall(__NR_mmap, 0x20fed000, 0x10000, 0, 0x8811, r[0], 0);
        *(uint64_t*)0x20000340 = 2;
        syscall(__NR_mbind, 0x20ff9000, 0x4000, 0x4002, 0x20000340, 0x45d4, 3);
        return 0;
  }

Actually the test does:

  mmap(0x20000000, 16777216, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x20000000
  socket(AF_PACKET, SOCK_RAW, 768)        = 3
  setsockopt(3, SOL_PACKET, PACKET_TX_RING, {block_size=65536, block_nr=1, frame_size=50464, frame_nr=1}, 16) = 0
  mmap(0x20fed000, 65536, PROT_NONE, MAP_SHARED|MAP_FIXED|MAP_POPULATE|MAP_DENYWRITE, 3, 0) = 0x20fed000
  mbind(..., MPOL_MF_STRICT|MPOL_MF_MOVE) = 0

The setsockopt() would allocate compound pages (16 pages in this test)
for packet tx ring, then the mmap() would call packet_mmap() to map the
pages into the user address space specified by the mmap() call.

When calling mbind(), it would scan the vma to queue the pages for
migration to the new node.  It would split any huge page since 4.9
doesn't support THP migration, however, the packet tx ring compound
pages are not THP and even not movable.  So, the above bug is triggered.

However, the later kernel is not hit by this issue due to commit
d44d363f6578 ("mm: don't assume anonymous pages have SwapBacked flag"),
which just removes the PageSwapBacked check for a different reason.

But, there is a deeper issue.  According to the semantic of mbind(), it
should return -EIO if MPOL_MF_MOVE or MPOL_MF_MOVE_ALL was specified and
MPOL_MF_STRICT was also specified, but the kernel was unable to move all
existing pages in the range.  The tx ring of the packet socket is
definitely not movable, however, mbind() returns success for this case.

Although the most socket file associates with non-movable pages, but XDP
may have movable pages from gup.  So, it sounds not fine to just check
the underlying file type of vma in vma_migratable().

Change migrate_page_add() to check if the page is movable or not, if it
is unmovable, just return -EIO.  But do not abort pte walk immediately,
since there may be pages off LRU temporarily.  We should migrate other
pages if MPOL_MF_MOVE* is specified.  Set has_unmovable flag if some
paged could not be not moved, then return -EIO for mbind() eventually.

With this change the above test would return -EIO as expected.

[yang.shi@linux.alibaba.com: fix review comments from Vlastimil]
Link: http://lkml.kernel.org/r/1563556862-54056-3-git-send-email-yang.shi@linux.alibaba.com
Link: http://lkml.kernel.org/r/1561162809-59140-3-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm: mempolicy: make the behavior consistent when MPOL_MF_MOVE* and MPOL_MF_STRICT...
Yang Shi [Tue, 13 Aug 2019 22:37:15 +0000 (15:37 -0700)]
mm: mempolicy: make the behavior consistent when MPOL_MF_MOVE* and MPOL_MF_STRICT were specified

When both MPOL_MF_MOVE* and MPOL_MF_STRICT was specified, mbind() should
try best to migrate misplaced pages, if some of the pages could not be
migrated, then return -EIO.

There are three different sub-cases:
 1. vma is not migratable
 2. vma is migratable, but there are unmovable pages
 3. vma is migratable, pages are movable, but migrate_pages() fails

If #1 happens, kernel would just abort immediately, then return -EIO,
after a7f40cfe3b7a ("mm: mempolicy: make mbind() return -EIO when
MPOL_MF_STRICT is specified").

If #3 happens, kernel would set policy and migrate pages with
best-effort, but won't rollback the migrated pages and reset the policy
back.

Before that commit, they behaves in the same way.  It'd better to keep
their behavior consistent.  But, rolling back the migrated pages and
resetting the policy back sounds not feasible, so just make #1 behave as
same as #3.

Userspace will know that not everything was successfully migrated (via
-EIO), and can take whatever steps it deems necessary - attempt
rollback, determine which exact page(s) are violating the policy, etc.

Make queue_pages_range() return 1 to indicate there are unmovable pages
or vma is not migratable.

The #2 is not handled correctly in the current kernel, the following
patch will fix it.

[yang.shi@linux.alibaba.com: fix review comments from Vlastimil]
Link: http://lkml.kernel.org/r/1563556862-54056-2-git-send-email-yang.shi@linux.alibaba.com
Link: http://lkml.kernel.org/r/1561162809-59140-2-git-send-email-yang.shi@linux.alibaba.com
Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com>
Reviewed-by: Vlastimil Babka <vbabka@suse.cz>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm/hmm: fix bad subpage pointer in try_to_unmap_one
Ralph Campbell [Tue, 13 Aug 2019 22:37:11 +0000 (15:37 -0700)]
mm/hmm: fix bad subpage pointer in try_to_unmap_one

When migrating an anonymous private page to a ZONE_DEVICE private page,
the source page->mapping and page->index fields are copied to the
destination ZONE_DEVICE struct page and the page_mapcount() is
increased.  This is so rmap_walk() can be used to unmap and migrate the
page back to system memory.

However, try_to_unmap_one() computes the subpage pointer from a swap pte
which computes an invalid page pointer and a kernel panic results such
as:

  BUG: unable to handle page fault for address: ffffea1fffffffc8

Currently, only single pages can be migrated to device private memory so
no subpage computation is needed and it can be set to "page".

[rcampbell@nvidia.com: add comment]
Link: http://lkml.kernel.org/r/20190724232700.23327-4-rcampbell@nvidia.com
Link: http://lkml.kernel.org/r/20190719192955.30462-4-rcampbell@nvidia.com
Fixes: a5430dda8a3a1c ("mm/migrate: support un-addressable ZONE_DEVICE page in migration")
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: John Hubbard <jhubbard@nvidia.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm/hmm: fix ZONE_DEVICE anon page mapping reuse
Ralph Campbell [Tue, 13 Aug 2019 22:37:07 +0000 (15:37 -0700)]
mm/hmm: fix ZONE_DEVICE anon page mapping reuse

When a ZONE_DEVICE private page is freed, the page->mapping field can be
set.  If this page is reused as an anonymous page, the previous value
can prevent the page from being inserted into the CPU's anon rmap table.
For example, when migrating a pte_none() page to device memory:

  migrate_vma(ops, vma, start, end, src, dst, private)
    migrate_vma_collect()
      src[] = MIGRATE_PFN_MIGRATE
    migrate_vma_prepare()
      /* no page to lock or isolate so OK */
    migrate_vma_unmap()
      /* no page to unmap so OK */
    ops->alloc_and_copy()
      /* driver allocates ZONE_DEVICE page for dst[] */
    migrate_vma_pages()
      migrate_vma_insert_page()
        page_add_new_anon_rmap()
          __page_set_anon_rmap()
            /* This check sees the page's stale mapping field */
            if (PageAnon(page))
              return
            /* page->mapping is not updated */

The result is that the migration appears to succeed but a subsequent CPU
fault will be unable to migrate the page back to system memory or worse.

Clear the page->mapping field when freeing the ZONE_DEVICE page so stale
pointer data doesn't affect future page use.

Link: http://lkml.kernel.org/r/20190719192955.30462-3-rcampbell@nvidia.com
Fixes: b7a523109fb5c9d2d6dd ("mm: don't clear ->mapping in hmm_devmem_free")
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Cc: Dan Williams <dan.j.williams@intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Logan Gunthorpe <logang@deltatee.com>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Jan Kara <jack@suse.cz>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Cc: "Jérôme Glisse" <jglisse@redhat.com>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm: document zone device struct page field usage
Ralph Campbell [Tue, 13 Aug 2019 22:37:04 +0000 (15:37 -0700)]
mm: document zone device struct page field usage

Patch series "mm/hmm: fixes for device private page migration", v3.

Testing the latest linux git tree turned up a few bugs with page
migration to and from ZONE_DEVICE private and anonymous pages.
Hopefully it clarifies how ZONE_DEVICE private struct page uses the same
mapping and index fields from the source anonymous page mapping.

This patch (of 3):

Struct page for ZONE_DEVICE private pages uses the page->mapping and and
page->index fields while the source anonymous pages are migrated to
device private memory.  This is so rmap_walk() can find the page when
migrating the ZONE_DEVICE private page back to system memory.
ZONE_DEVICE pmem backed fsdax pages also use the page->mapping and
page->index fields when files are mapped into a process address space.

Add comments to struct page and remove the unused "_zd_pad_1" field to
make this more clear.

Link: http://lkml.kernel.org/r/20190724232700.23327-2-rcampbell@nvidia.com
Signed-off-by: Ralph Campbell <rcampbell@nvidia.com>
Reviewed-by: John Hubbard <jhubbard@nvidia.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Christoph Lameter <cl@linux.com>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Jérôme Glisse <jglisse@redhat.com>
Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Andrey Ryabinin <aryabinin@virtuozzo.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agoLinux 5.3-rc4 v5.3-rc4
Linus Torvalds [Sun, 11 Aug 2019 20:26:41 +0000 (13:26 -0700)]
Linux 5.3-rc4

10 months agoMerge tag 'dax-fixes-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm...
Linus Torvalds [Sun, 11 Aug 2019 20:15:10 +0000 (13:15 -0700)]
Merge tag 'dax-fixes-5.3-rc4' of git://git./linux/kernel/git/nvdimm/nvdimm

Pull dax fixes from Dan Williams:
 "A filesystem-dax and device-dax fix for v5.3.

  The filesystem-dax fix is tagged for stable as the implementation has
  been mistakenly throwing away all cow pages on any truncate or hole
  punch operation as part of the solution to coordinate device-dma vs
  truncate to dax pages.

  The device-dax change fixes up a regression this cycle from the
  introduction of a common 'internal per-cpu-ref' implementation.

  Summary:

   - Fix dax_layout_busy_page() to not discard private cow pages of
     fs/dax private mappings.

   - Update the memremap_pages core to properly cleanup on behalf of
     internal reference-count users like device-dax"

* tag 'dax-fixes-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  mm/memremap: Fix reuse of pgmap instances with internal references
  dax: dax_layout_busy_page() should not unmap cow pages

10 months agoMerge tag 'ntb-5.3-bugfixes' of git://github.com/jonmason/ntb
Linus Torvalds [Sun, 11 Aug 2019 17:13:53 +0000 (10:13 -0700)]
Merge tag 'ntb-5.3-bugfixes' of git://github.com/jonmason/ntb

Pull NTB fix from Jon Mason:
 "Bug fix for NTB MSI kernel compile warning"

* tag 'ntb-5.3-bugfixes' of git://github.com/jonmason/ntb:
  NTB/msi: remove incorrect MODULE defines

10 months agoMerge tag 'riscv/for-v5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv...
Linus Torvalds [Sat, 10 Aug 2019 23:31:47 +0000 (16:31 -0700)]
Merge tag 'riscv/for-v5.3-rc4' of git://git./linux/kernel/git/riscv/linux

Pull RISC-V updates from Paul Walmsley:
 "A few minor RISC-V updates for v5.3-rc4:

   - Remove __udivdi3() from the 32-bit Linux port, converting the only
     upstream user to use do_div(), per Linux policy

   - Convert the RISC-V standard clocksource away from per-cpu data
     structures, since only one is used by Linux, even on a multi-CPU
     system

   - A set of DT binding updates that remove an obsolete text binding in
     favor of a YAML binding, fix a bogus compatible string in the
     schema (thus fixing a "make dtbs_check" warning), and clarifies the
     future values expected in one of the RISC-V CPU properties"

* tag 'riscv/for-v5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/riscv/linux:
  dt-bindings: riscv: fix the schema compatible string for the HiFive Unleashed board
  dt-bindings: riscv: remove obsolete cpus.txt
  RISC-V: Remove udivdi3
  riscv: delay: use do_div() instead of __udivdi3()
  dt-bindings: Update the riscv,isa string description
  RISC-V: Remove per cpu clocksource

10 months agoMerge branch 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 10 Aug 2019 23:24:03 +0000 (16:24 -0700)]
Merge branch 'x86-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull x86 fixes from Thomas Gleixner:
 "A few fixes for x86:

   - Don't reset the carefully adjusted build flags for the purgatory
     and remove the unwanted flags instead. The 'reset all' approach led
     to build fails under certain circumstances.

   - Unbreak CLANG build of the purgatory by avoiding the builtin
     memcpy/memset implementations.

   - Address missing prototype warnings by including the proper header

   - Fix yet more fall-through issues"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/lib/cpu: Address missing prototypes warning
  x86/purgatory: Use CFLAGS_REMOVE rather than reset KBUILD_CFLAGS
  x86/purgatory: Do not use __builtin_memcpy and __builtin_memset
  x86: mtrr: cyrix: Mark expected switch fall-through
  x86/ptrace: Mark expected switch fall-through

10 months agoMerge branch 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 10 Aug 2019 23:19:02 +0000 (16:19 -0700)]
Merge branch 'perf-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull perf tooling fixes from Thomas Gleixner:
 "Perf tooling fixes all over the place:

   - Fix the selection of the main thread COMM in db-export

   - Fix the disassemmbly display for BPF in annotate

   - Fix cpumap mask setup in perf ftrace when only one CPU is present

   - Add the missing 'cpu_clk_unhalted.core' event

   - Fix CPU 0 bindings in NUMA benchmarks

   - Fix the module size calculations for s390

   - Handle the gap between kernel end and module start on s390
     correctly

   - Build and typo fixes"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf pmu-events: Fix missing "cpu_clk_unhalted.core" event
  perf annotate: Fix s390 gap between kernel end and module start
  perf record: Fix module size on s390
  perf tools: Fix include paths in ui directory
  perf tools: Fix a typo in a variable name in the Documentation Makefile
  perf cpumap: Fix writing to illegal memory in handling cpumap mask
  perf ftrace: Fix failure to set cpumask when only one cpu is present
  perf db-export: Fix thread__exec_comm()
  perf annotate: Fix printing of unaugmented disassembled instructions from BPF
  perf bench numa: Fix cpu0 binding

10 months agoMerge branch 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 10 Aug 2019 22:48:02 +0000 (15:48 -0700)]
Merge branch 'sched-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull scheduler fixes from Thomas Gleixner:
 "Three fixlets for the scheduler:

   - Avoid double bandwidth accounting in the push & pull code

   - Use a sane FIFO priority for the Pressure Stall Information (PSI)
     thread.

   - Avoid permission checks when setting the scheduler params for the
     PSI thread"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/psi: Do not require setsched permission from the trigger creator
  sched/psi: Reduce psimon FIFO priority
  sched/deadline: Fix double accounting of rq/running bw in push & pull

10 months agoMerge branch 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 10 Aug 2019 22:46:25 +0000 (15:46 -0700)]
Merge branch 'irq-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull irq fix from Thomas Gleixner:
 "A small fix for the affinity spreading code.

  It failed to handle situations where a single vector was requested
  either due to only one CPU being available or vector exhaustion
  causing only a single interrupt to be granted.

  The fix is to simply remove the requirement in the affinity spreading
  code for more than one interrupt being available"

* 'irq-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  genirq/affinity: Create affinity mask for single vector

10 months agoMerge branch 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 10 Aug 2019 22:44:09 +0000 (15:44 -0700)]
Merge branch 'core-urgent-for-linus' of git://git./linux/kernel/git/tip/tip

Pull objtool warning fix from Thomas Gleixner:
 "The recent objtool fixes/enhancements unearthed a unbalanced CLAC in
  the i915 driver.

  Chris asked me to pick the fix up and route it through"

* 'core-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  drm/i915: Remove redundant user_access_end() from __copy_from_user() error path

10 months agoMerge tag 'gfs2-v5.3-rc3.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 10 Aug 2019 22:41:15 +0000 (15:41 -0700)]
Merge tag 'gfs2-v5.3-rc3.fixes' of git://git./linux/kernel/git/gfs2/linux-gfs2

Pull gfs2 fix from Andreas Gruenbacher:
 "Fix incorrect lseek / fiemap results"

* tag 'gfs2-v5.3-rc3.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
  gfs2: gfs2_walk_metadata fix

10 months agoMakefile: Convert -Wimplicit-fallthrough=3 to just -Wimplicit-fallthrough for clang
Joe Perches [Mon, 5 Aug 2019 22:11:15 +0000 (15:11 -0700)]
Makefile: Convert -Wimplicit-fallthrough=3 to just -Wimplicit-fallthrough for clang

A compilation -Wimplicit-fallthrough warning was enabled by commit
a035d552a93b ("Makefile: Globally enable fall-through warning")

Even though clang 10.0.0 does not currently support this warning without
a patch, clang currently does not support a value for this option.

Link: https://bugs.llvm.org/show_bug.cgi?id=39382
The gcc default for this warning is 3 so removing the =3 has no effect
for gcc and enables the warning for patched versions of clang.

Also remove the =3 from an existing use in a parisc Makefile:
arch/parisc/math-emu/Makefile

Signed-off-by: Joe Perches <joe@perches.com>
Reviewed-and-tested-by: Nathan Chancellor <natechancellor@gmail.com>
Cc: Gustavo A. R. Silva <gustavo@embeddedor.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agoMerge tag 'char-misc-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 10 Aug 2019 19:24:20 +0000 (12:24 -0700)]
Merge tag 'char-misc-5.3-rc4' of git://git./linux/kernel/git/gregkh/char-misc

Pull char/misc driver fixes Greg KH:
 "Here are some small char/misc driver fixes for 5.3-rc4.

  Two of these are for the habanalabs driver for issues found when
  running on a big-endian system (are they still alive?) The others are
  tiny fixes reported by people, and a MAINTAINERS update about the
  location of the fpga development tree.

  All of these have been in linux-next for a while with no reported
  issues"

* tag 'char-misc-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc:
  coresight: Fix DEBUG_LOCKS_WARN_ON for uninitialized attribute
  MAINTAINERS: Move linux-fpga tree to new location
  nvmem: Use the same permissions for eeprom as for nvmem
  habanalabs: fix host memory polling in BE architecture
  habanalabs: fix F/W download in BE architecture

10 months agoMerge tag 'driver-core-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 10 Aug 2019 19:20:02 +0000 (12:20 -0700)]
Merge tag 'driver-core-5.3-rc4' of git://git./linux/kernel/git/gregkh/driver-core

Pull driver core fixes from Greg KH:
 "Here are two small fixes for some driver core issues that have been
  reported. There is also a kernfs "fix" here, which was then reverted
  because it was found to cause problems in linux-next.

  The driver core fixes both resolve reported issues, one with gpioint
  stuff that showed up in 5.3-rc1, and the other finally (and hopefully)
  resolves a very long standing race when removing glue directories.
  It's nice to get that issue finally resolved and the developers
  involved should be applauded for the persistence it took to get this
  patch finally accepted.

  All of these have been in linux-next for a while with no reported
  issues. Well, the one reported issue, hence the revert :)"

* tag 'driver-core-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  Revert "kernfs: fix memleak in kernel_ops_readdir()"
  kernfs: fix memleak in kernel_ops_readdir()
  driver core: Fix use-after-free and double free on glue directory
  driver core: platform: return -ENXIO for missing GpioInt

10 months agoMerge tag 'tty-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sat, 10 Aug 2019 19:17:12 +0000 (12:17 -0700)]
Merge tag 'tty-5.3-rc4' of git://git./linux/kernel/git/gregkh/tty

Pull tty fix from Greg KH:
 "Here is a single tty kgdb fix for 5.3-rc4.

  It fixes an annoying log message that has caused kdb to become
  useless. It's another fallout from commit ddde3c18b700 ("vt: More
  locking checks") which tries to enforce locking checks more strictly
  in the tty layer, unfortunatly when kdb is stopped, there's no need
  for locks :)

  This patch has been linux-next for a while with no reported issues"

* tag 'tty-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  kgdboc: disable the console lock when in kgdb

10 months agoMerge tag 'staging-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sat, 10 Aug 2019 19:13:39 +0000 (12:13 -0700)]
Merge tag 'staging-5.3-rc4' of git://git./linux/kernel/git/gregkh/staging

Pull staging / IIO driver fixes from Greg KH:
 "Here are some small staging and IIO driver fixes for 5.3-rc4.

  Nothing major, just resolutions for a number of small reported issues,
  full details in the shortlog.

  All have been in linux-next for a while with no reported issues"

* tag 'staging-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  iio: adc: gyroadc: fix uninitialized return code
  docs: generic-counter.rst: fix broken references for ABI file
  staging: android: ion: Bail out upon SIGKILL when allocating memory.
  Staging: fbtft: Fix GPIO handling
  staging: unisys: visornic: Update the description of 'poll_for_irq()'
  staging: wilc1000: flush the workqueue before deinit the host
  staging: gasket: apex: fix copy-paste typo
  Staging: fbtft: Fix reset assertion when using gpio descriptor
  Staging: fbtft: Fix probing of gpio descriptor
  iio: imu: mpu6050: add missing available scan masks
  iio: cros_ec_accel_legacy: Fix incorrect channel setting
  IIO: Ingenic JZ47xx: Set clock divider on probe
  iio: adc: max9611: Fix misuse of GENMASK macro

10 months agoMerge tag 'usb-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sat, 10 Aug 2019 18:59:57 +0000 (11:59 -0700)]
Merge tag 'usb-5.3-rc4' of git://git./linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are some small USB fixes for 5.3-rc4.

  The "biggest" one here is moving code from one file to another in
  order to fix a long-standing race condition with the creation of sysfs
  files for USB devices. Turns out that there are now userspace tools
  out there that are hitting this long-known bug, so it's time to fix
  them. Thankfully the tool-maker in this case fixed the issue :)

  The other patches in here are all fixes for reported issues. Now that
  syzbot knows how to fuzz USB drivers better, and is starting to now
  fuzz the userspace facing side of them at the same time, there will be
  more and more small fixes like these coming, which is a good thing.

  All of these have been in linux-next with no reported issues"

* tag 'usb-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  usb: setup authorized_default attributes using usb_bus_notify
  usb: iowarrior: fix deadlock on disconnect
  Revert "USB: rio500: simplify locking"
  usb: usbfs: fix double-free of usb memory upon submiturb error
  usb: yurex: Fix use-after-free in yurex_delete
  usb: typec: tcpm: Ignore unsupported/unknown alternate mode requests
  xhci: Fix NULL pointer dereference at endpoint zero reset.
  usb: host: xhci-rcar: Fix timeout in xhci_suspend()
  usb: typec: ucsi: ccg: Fix uninitilized symbol error
  usb: typec: tcpm: remove tcpm dir if no children
  usb: typec: tcpm: free log buf memory when remove debug file
  usb: typec: tcpm: Add NULL check before dereferencing config

10 months agoMerge tag 'pinctrl-v5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw...
Linus Torvalds [Sat, 10 Aug 2019 17:21:25 +0000 (10:21 -0700)]
Merge tag 'pinctrl-v5.3-2' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull pin control fixes from Linus Walleij:

 - Delay acquisition of regmaps in the Aspeed G5 driver.

 - Make a symbol static to reduce compiler noise.

* tag 'pinctrl-v5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
  pinctrl: aspeed: Make aspeed_pinmux_ips static
  pinctrl: aspeed-g5: Delay acquisition of regmaps

10 months agoMerge tag 'powerpc-5.3-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc...
Linus Torvalds [Sat, 10 Aug 2019 17:17:19 +0000 (10:17 -0700)]
Merge tag 'powerpc-5.3-4' of git://git./linux/kernel/git/powerpc/linux

Pull powerpc fix from Michael Ellerman:
 "Just one fix, a revert of a commit that was meant to be a minor
  improvement to some inline asm, but ended up having no real benefit
  with GCC and broke booting 32-bit machines when using Clang.

  Thanks to: Arnd Bergmann, Christophe Leroy, Nathan Chancellor, Nick
  Desaulniers, Segher Boessenkool"

* tag 'powerpc-5.3-4' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  Revert "powerpc: slightly improve cache helpers"

10 months agoMerge tag 'Wimplicit-fallthrough-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kerne...
Linus Torvalds [Sat, 10 Aug 2019 17:10:33 +0000 (10:10 -0700)]
Merge tag 'Wimplicit-fallthrough-5.3-rc4' of git://git./linux/kernel/git/gustavoars/linux

Pull fall-through fixes from Gustavo A. R. Silva:
 "Mark more switch cases where we are expecting to fall through, fixing
  fall-through warnings in arm, sparc64, mips, i386 and s390"

* tag 'Wimplicit-fallthrough-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/gustavoars/linux:
  ARM: ep93xx: Mark expected switch fall-through
  scsi: fas216: Mark expected switch fall-throughs
  pcmcia: db1xxx_ss: Mark expected switch fall-throughs
  video: fbdev: omapfb_main: Mark expected switch fall-throughs
  watchdog: riowd: Mark expected switch fall-through
  s390/net: Mark expected switch fall-throughs
  crypto: ux500/crypt: Mark expected switch fall-throughs
  watchdog: wdt977: Mark expected switch fall-through
  watchdog: scx200_wdt: Mark expected switch fall-through
  watchdog: Mark expected switch fall-throughs
  ARM: signal: Mark expected switch fall-through
  mfd: omap-usb-host: Mark expected switch fall-throughs
  mfd: db8500-prcmu: Mark expected switch fall-throughs
  ARM: OMAP: dma: Mark expected switch fall-throughs
  ARM: alignment: Mark expected switch fall-throughs
  ARM: tegra: Mark expected switch fall-through
  ARM/hw_breakpoint: Mark expected switch fall-throughs

10 months agoMerge tag 'kbuild-fixes-v5.3-3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 10 Aug 2019 03:31:04 +0000 (20:31 -0700)]
Merge tag 'kbuild-fixes-v5.3-3' of git://git./linux/kernel/git/masahiroy/linux-kbuild

Pull Kbuild fixes from Masahiro Yamada:

 - revive single target %.ko

 - do not create built-in.a where it is unneeded

 - do not create modules.order where it is unneeded

 - show a warning if subdir-y/m is used to visit a module Makefile

* tag 'kbuild-fixes-v5.3-3' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild:
  kbuild: show hint if subdir-y/m is used to visit module Makefile
  kbuild: generate modules.order only in directories visited by obj-y/m
  kbuild: fix false-positive need-builtin calculation
  kbuild: revive single target %.ko

10 months agoARM: ep93xx: Mark expected switch fall-through
Gustavo A. R. Silva [Thu, 8 Aug 2019 02:34:48 +0000 (21:34 -0500)]
ARM: ep93xx: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

Fix the following warnings (Building: arm-ep93xx_defconfig arm):

arch/arm/mach-ep93xx/crunch.c: In function 'crunch_do':
arch/arm/mach-ep93xx/crunch.c:46:3: warning: this statement may
fall through [-Wimplicit-fallthrough=]
      memset(crunch_state, 0, sizeof(*crunch_state));
      ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
   arch/arm/mach-ep93xx/crunch.c:53:2: note: here
     case THREAD_NOTIFY_EXIT:
     ^~~~

Notice that, in this particular case, the code comment is
modified in accordance with what GCC is expecting to find.

Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agoscsi: fas216: Mark expected switch fall-throughs
Gustavo A. R. Silva [Tue, 6 Aug 2019 08:26:15 +0000 (03:26 -0500)]
scsi: fas216: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

Fix the following warnings (Building: rpc_defconfig arm):

drivers/scsi/arm/fas216.c: In function ‘fas216_disconnect_intr’:
drivers/scsi/arm/fas216.c:913:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (fas216_get_last_msg(info, info->scsi.msgin_fifo) == ABORT) {
      ^
drivers/scsi/arm/fas216.c:919:2: note: here
  default:    /* huh?     */
  ^~~~~~~
drivers/scsi/arm/fas216.c: In function ‘fas216_kick’:
drivers/scsi/arm/fas216.c:1959:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
   fas216_allocate_tag(info, SCpnt);
   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/scsi/arm/fas216.c:1960:2: note: here
  case TYPE_OTHER:
  ^~~~
drivers/scsi/arm/fas216.c: In function ‘fas216_busservice_intr’:
drivers/scsi/arm/fas216.c:1413:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
   fas216_stoptransfer(info);
   ^~~~~~~~~~~~~~~~~~~~~~~~~
drivers/scsi/arm/fas216.c:1414:2: note: here
  case STATE(STAT_STATUS, PHASE_SELSTEPS):/* Sel w/ steps -> Status       */
  ^~~~
drivers/scsi/arm/fas216.c:1424:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
   fas216_stoptransfer(info);
   ^~~~~~~~~~~~~~~~~~~~~~~~~
drivers/scsi/arm/fas216.c:1425:2: note: here
  case STATE(STAT_MESGIN, PHASE_COMMAND): /* Command -> Message In */
  ^~~~
drivers/scsi/arm/fas216.c: In function ‘fas216_funcdone_intr’:
drivers/scsi/arm/fas216.c:1573:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if ((stat & STAT_BUSMASK) == STAT_MESGIN) {
      ^
drivers/scsi/arm/fas216.c:1579:2: note: here
  default:
  ^~~~~~~
drivers/scsi/arm/fas216.c: In function ‘fas216_handlesync’:
drivers/scsi/arm/fas216.c:605:20: warning: this statement may fall through [-Wimplicit-fallthrough=]
   info->scsi.phase = PHASE_MSGOUT_EXPECT;
   ~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~
drivers/scsi/arm/fas216.c:607:2: note: here
  case async:
  ^~~~

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agopcmcia: db1xxx_ss: Mark expected switch fall-throughs
Gustavo A. R. Silva [Mon, 5 Aug 2019 19:47:35 +0000 (14:47 -0500)]
pcmcia: db1xxx_ss: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings (Building: db1xxx_defconfig mips):

drivers/pcmcia/db1xxx_ss.c:257:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/pcmcia/db1xxx_ss.c:269:3: warning: this statement may fall through [-Wimplicit-fallthrough=]

Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agovideo: fbdev: omapfb_main: Mark expected switch fall-throughs
Gustavo A. R. Silva [Fri, 2 Aug 2019 19:10:29 +0000 (14:10 -0500)]
video: fbdev: omapfb_main: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warning (Building: omap1_defconfig arm):

drivers/watchdog/wdt285.c:170:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/watchdog/ar7_wdt.c:237:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:449:23: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:1549:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:1547:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:1545:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:1543:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:1540:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:1538:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
drivers/video/fbdev/omap/omapfb_main.c:1535:3: warning: this statement may fall through [-Wimplicit-fallthrough=]

Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agowatchdog: riowd: Mark expected switch fall-through
Gustavo A. R. Silva [Tue, 30 Jul 2019 01:45:29 +0000 (20:45 -0500)]
watchdog: riowd: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings (Building: sparc64):

drivers/watchdog/riowd.c: In function ‘riowd_ioctl’:
drivers/watchdog/riowd.c:136:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
   riowd_writereg(p, riowd_timeout, WDTO_INDEX);
   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/watchdog/riowd.c:139:2: note: here
  case WDIOC_GETTIMEOUT:
  ^~~~

Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agos390/net: Mark expected switch fall-throughs
Gustavo A. R. Silva [Tue, 30 Jul 2019 00:13:54 +0000 (19:13 -0500)]
s390/net: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings (Building: s390):

drivers/s390/net/ctcm_fsms.c: In function ‘ctcmpc_chx_attnbusy’:
drivers/s390/net/ctcm_fsms.c:1703:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (grp->changed_side == 1) {
      ^
drivers/s390/net/ctcm_fsms.c:1707:2: note: here
  case MPCG_STATE_XID0IOWAIX:
  ^~~~

drivers/s390/net/ctcm_mpc.c: In function ‘ctc_mpc_alloc_channel’:
drivers/s390/net/ctcm_mpc.c:358:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (callback)
      ^
drivers/s390/net/ctcm_mpc.c:360:2: note: here
  case MPCG_STATE_XID0IOWAIT:
  ^~~~

drivers/s390/net/ctcm_mpc.c: In function ‘mpc_action_timeout’:
drivers/s390/net/ctcm_mpc.c:1469:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if ((fsm_getstate(rch->fsm) == CH_XID0_PENDING) &&
      ^
drivers/s390/net/ctcm_mpc.c:1472:2: note: here
  default:
  ^~~~~~~
drivers/s390/net/ctcm_mpc.c: In function ‘mpc_send_qllc_discontact’:
drivers/s390/net/ctcm_mpc.c:2087:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (grp->estconnfunc) {
      ^
drivers/s390/net/ctcm_mpc.c:2092:2: note: here
  case MPCG_STATE_FLOWC:
  ^~~~

drivers/s390/net/qeth_l2_main.c: In function ‘qeth_l2_process_inbound_buffer’:
drivers/s390/net/qeth_l2_main.c:328:7: warning: this statement may fall through [-Wimplicit-fallthrough=]
    if (IS_OSN(card)) {
       ^
drivers/s390/net/qeth_l2_main.c:337:3: note: here
   default:
   ^~~~~~~

Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agocrypto: ux500/crypt: Mark expected switch fall-throughs
Gustavo A. R. Silva [Mon, 29 Jul 2019 22:35:45 +0000 (17:35 -0500)]
crypto: ux500/crypt: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warning (Building: arm):

drivers/crypto/ux500/cryp/cryp.c: In function ‘cryp_save_device_context’:
drivers/crypto/ux500/cryp/cryp.c:316:16: warning: this statement may fall through [-Wimplicit-fallthrough=]
   ctx->key_4_r = readl_relaxed(&src_reg->key_4_r);
drivers/crypto/ux500/cryp/cryp.c:318:2: note: here
  case CRYP_KEY_SIZE_192:
  ^~~~
drivers/crypto/ux500/cryp/cryp.c:320:16: warning: this statement may fall through [-Wimplicit-fallthrough=]
   ctx->key_3_r = readl_relaxed(&src_reg->key_3_r);
drivers/crypto/ux500/cryp/cryp.c:322:2: note: here
  case CRYP_KEY_SIZE_128:
  ^~~~
drivers/crypto/ux500/cryp/cryp.c:324:16: warning: this statement may fall through [-Wimplicit-fallthrough=]
   ctx->key_2_r = readl_relaxed(&src_reg->key_2_r);
drivers/crypto/ux500/cryp/cryp.c:326:2: note: here
  default:
  ^~~~~~~
In file included from ./include/linux/io.h:13:0,
                 from drivers/crypto/ux500/cryp/cryp_p.h:14,
                 from drivers/crypto/ux500/cryp/cryp.c:15:
drivers/crypto/ux500/cryp/cryp.c: In function ‘cryp_restore_device_context’:
./arch/arm/include/asm/io.h:92:22: warning: this statement may fall through [-Wimplicit-fallthrough=]
 #define __raw_writel __raw_writel
                      ^
./arch/arm/include/asm/io.h:299:29: note: in expansion of macro ‘__raw_writel’
 #define writel_relaxed(v,c) __raw_writel((__force u32) cpu_to_le32(v),c)
                             ^~~~~~~~~~~~
drivers/crypto/ux500/cryp/cryp.c:363:3: note: in expansion of macro ‘writel_relaxed’
   writel_relaxed(ctx->key_4_r, &reg->key_4_r);
   ^~~~~~~~~~~~~~
drivers/crypto/ux500/cryp/cryp.c:365:2: note: here
  case CRYP_KEY_SIZE_192:
  ^~~~
In file included from ./include/linux/io.h:13:0,
                 from drivers/crypto/ux500/cryp/cryp_p.h:14,
                 from drivers/crypto/ux500/cryp/cryp.c:15:
./arch/arm/include/asm/io.h:92:22: warning: this statement may fall through [-Wimplicit-fallthrough=]
 #define __raw_writel __raw_writel
                      ^
./arch/arm/include/asm/io.h:299:29: note: in expansion of macro ‘__raw_writel’
 #define writel_relaxed(v,c) __raw_writel((__force u32) cpu_to_le32(v),c)
                             ^~~~~~~~~~~~
drivers/crypto/ux500/cryp/cryp.c:367:3: note: in expansion of macro ‘writel_relaxed’
   writel_relaxed(ctx->key_3_r, &reg->key_3_r);
   ^~~~~~~~~~~~~~
drivers/crypto/ux500/cryp/cryp.c:369:2: note: here
  case CRYP_KEY_SIZE_128:
  ^~~~
In file included from ./include/linux/io.h:13:0,
                 from drivers/crypto/ux500/cryp/cryp_p.h:14,
                 from drivers/crypto/ux500/cryp/cryp.c:15:
./arch/arm/include/asm/io.h:92:22: warning: this statement may fall through [-Wimplicit-fallthrough=]
 #define __raw_writel __raw_writel
                      ^
./arch/arm/include/asm/io.h:299:29: note: in expansion of macro ‘__raw_writel’
 #define writel_relaxed(v,c) __raw_writel((__force u32) cpu_to_le32(v),c)
                             ^~~~~~~~~~~~
drivers/crypto/ux500/cryp/cryp.c:371:3: note: in expansion of macro ‘writel_relaxed’
   writel_relaxed(ctx->key_2_r, &reg->key_2_r);
   ^~~~~~~~~~~~~~
drivers/crypto/ux500/cryp/cryp.c:373:2: note: here
  default:
  ^~~~~~~

Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agowatchdog: wdt977: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 29 Jul 2019 22:30:57 +0000 (17:30 -0500)]
watchdog: wdt977: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

This patch fixes the following warning (Building: arm):

drivers/watchdog/wdt977.c: In function ‘wdt977_ioctl’:
  LD [M]  drivers/media/platform/vicodec/vicodec.o
drivers/watchdog/wdt977.c:400:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
   wdt977_keepalive();
   ^~~~~~~~~~~~~~~~~~
drivers/watchdog/wdt977.c:403:2: note: here
  case WDIOC_GETTIMEOUT:
  ^~~~

Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agowatchdog: scx200_wdt: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 29 Jul 2019 20:03:49 +0000 (15:03 -0500)]
watchdog: scx200_wdt: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

This patch fixes the following warning (Building: i386):

drivers/watchdog/scx200_wdt.c: In function ‘scx200_wdt_ioctl’:
drivers/watchdog/scx200_wdt.c:188:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
   scx200_wdt_ping();
   ^~~~~~~~~~~~~~~~~
drivers/watchdog/scx200_wdt.c:189:2: note: here
  case WDIOC_GETTIMEOUT:
  ^~~~

Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agowatchdog: Mark expected switch fall-throughs
Gustavo A. R. Silva [Mon, 29 Jul 2019 15:08:05 +0000 (10:08 -0500)]
watchdog: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings:

drivers/watchdog/ar7_wdt.c: warning: this statement may fall
through [-Wimplicit-fallthrough=]:  => 237:3
drivers/watchdog/pcwd.c: warning: this statement may fall
through [-Wimplicit-fallthrough=]:  => 653:3
drivers/watchdog/sb_wdog.c: warning: this statement may fall
through [-Wimplicit-fallthrough=]:  => 204:3
drivers/watchdog/wdt.c: warning: this statement may fall
through [-Wimplicit-fallthrough=]:  => 391:3

Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agoARM: signal: Mark expected switch fall-through
Gustavo A. R. Silva [Mon, 29 Jul 2019 00:11:11 +0000 (19:11 -0500)]
ARM: signal: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

This patch fixes the following warning:

arch/arm/kernel/signal.c: In function 'do_signal':
arch/arm/kernel/signal.c:598:12: warning: this statement may fall through [-Wimplicit-fallthrough=]
    restart -= 2;
    ~~~~~~~~^~~~
arch/arm/kernel/signal.c:599:3: note: here
   case -ERESTARTNOHAND:
   ^~~~

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agomfd: omap-usb-host: Mark expected switch fall-throughs
Gustavo A. R. Silva [Sun, 28 Jul 2019 23:57:05 +0000 (18:57 -0500)]
mfd: omap-usb-host: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings:

drivers/mfd/omap-usb-host.c: In function 'usbhs_runtime_resume':
drivers/mfd/omap-usb-host.c:303:7: warning: this statement may fall through [-Wimplicit-fallthrough=]
    if (!IS_ERR(omap->hsic480m_clk[i])) {
       ^
drivers/mfd/omap-usb-host.c:313:3: note: here
   case OMAP_EHCI_PORT_MODE_TLL:
   ^~~~
drivers/mfd/omap-usb-host.c: In function 'usbhs_runtime_suspend':
drivers/mfd/omap-usb-host.c:345:7: warning: this statement may fall through [-Wimplicit-fallthrough=]
    if (!IS_ERR(omap->hsic480m_clk[i]))
       ^
drivers/mfd/omap-usb-host.c:349:3: note: here
   case OMAP_EHCI_PORT_MODE_TLL:
   ^~~~

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agomfd: db8500-prcmu: Mark expected switch fall-throughs
Gustavo A. R. Silva [Sun, 28 Jul 2019 23:53:15 +0000 (18:53 -0500)]
mfd: db8500-prcmu: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings:

drivers/mfd/db8500-prcmu.c: In function 'dsiclk_rate':
drivers/mfd/db8500-prcmu.c:1592:7: warning: this statement may fall through [-Wimplicit-fallthrough=]
   div *= 2;
   ~~~~^~~~
drivers/mfd/db8500-prcmu.c:1593:2: note: here
  case PRCM_DSI_PLLOUT_SEL_PHI_2:
  ^~~~
drivers/mfd/db8500-prcmu.c:1594:7: warning: this statement may fall through [-Wimplicit-fallthrough=]
   div *= 2;
   ~~~~^~~~
drivers/mfd/db8500-prcmu.c:1595:2: note: here
  case PRCM_DSI_PLLOUT_SEL_PHI:
  ^~~~

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agoARM: OMAP: dma: Mark expected switch fall-throughs
Gustavo A. R. Silva [Sun, 28 Jul 2019 23:19:41 +0000 (18:19 -0500)]
ARM: OMAP: dma: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings:

arch/arm/plat-omap/dma.c: In function 'omap_set_dma_src_burst_mode':
arch/arm/plat-omap/dma.c:384:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (dma_omap2plus()) {
      ^
arch/arm/plat-omap/dma.c:393:2: note: here
  case OMAP_DMA_DATA_BURST_16:
  ^~~~
arch/arm/plat-omap/dma.c:394:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (dma_omap2plus()) {
      ^
arch/arm/plat-omap/dma.c:402:2: note: here
  default:
  ^~~~~~~
arch/arm/plat-omap/dma.c: In function 'omap_set_dma_dest_burst_mode':
arch/arm/plat-omap/dma.c:473:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (dma_omap2plus()) {
      ^
arch/arm/plat-omap/dma.c:481:2: note: here
  default:
  ^~~~~~~

Notice that, in this particular case, the code comment is
modified in accordance with what GCC is expecting to find.

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agoARM: alignment: Mark expected switch fall-throughs
Gustavo A. R. Silva [Sun, 28 Jul 2019 23:17:18 +0000 (18:17 -0500)]
ARM: alignment: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings:

arch/arm/mm/alignment.c: In function 'thumb2arm':
arch/arm/mm/alignment.c:688:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if ((tinstr & (3 << 9)) == 0x0400) {
      ^
arch/arm/mm/alignment.c:700:2: note: here
  default:
  ^~~~~~~
arch/arm/mm/alignment.c: In function 'do_alignment_t32_to_handler':
arch/arm/mm/alignment.c:753:15: warning: this statement may fall through [-Wimplicit-fallthrough=]
   poffset->un = (tinst2 & 0xff) << 2;
   ~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~
arch/arm/mm/alignment.c:754:2: note: here
  case 0xe940:
  ^~~~

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agoARM: tegra: Mark expected switch fall-through
Gustavo A. R. Silva [Sun, 28 Jul 2019 23:10:21 +0000 (18:10 -0500)]
ARM: tegra: Mark expected switch fall-through

Mark switch cases where we are expecting to fall through.

This patch fixes the following warning:

arch/arm/mach-tegra/reset.c: In function 'tegra_cpu_reset_handler_enable':
arch/arm/mach-tegra/reset.c:72:3: warning: this statement may fall through [-Wimplicit-fallthrough=]
   tegra_cpu_reset_handler_set(reset_address);
   ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
arch/arm/mach-tegra/reset.c:74:2: note: here
  case 0:
  ^~~~

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agoARM/hw_breakpoint: Mark expected switch fall-throughs
Gustavo A. R. Silva [Sun, 28 Jul 2019 23:02:41 +0000 (18:02 -0500)]
ARM/hw_breakpoint: Mark expected switch fall-throughs

Mark switch cases where we are expecting to fall through.

This patch fixes the following warnings:

arch/arm/kernel/hw_breakpoint.c: In function 'hw_breakpoint_arch_parse':
arch/arm/kernel/hw_breakpoint.c:609:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (hw->ctrl.len == ARM_BREAKPOINT_LEN_2)
      ^
arch/arm/kernel/hw_breakpoint.c:611:2: note: here
  case 3:
  ^~~~
arch/arm/kernel/hw_breakpoint.c:613:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (hw->ctrl.len == ARM_BREAKPOINT_LEN_1)
      ^
arch/arm/kernel/hw_breakpoint.c:615:2: note: here
  default:
  ^~~~~~~
arch/arm/kernel/hw_breakpoint.c: In function 'arch_build_bp_info':
arch/arm/kernel/hw_breakpoint.c:544:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if ((hw->ctrl.type != ARM_BREAKPOINT_EXECUTE)
      ^
arch/arm/kernel/hw_breakpoint.c:547:2: note: here
  default:
  ^~~~~~~
In file included from include/linux/kernel.h:11,
                 from include/linux/list.h:9,
                 from include/linux/preempt.h:11,
                 from include/linux/hardirq.h:5,
                 from arch/arm/kernel/hw_breakpoint.c:16:
arch/arm/kernel/hw_breakpoint.c: In function 'hw_breakpoint_pending':
include/linux/compiler.h:78:22: warning: this statement may fall through [-Wimplicit-fallthrough=]
 # define unlikely(x) __builtin_expect(!!(x), 0)
                      ^~~~~~~~~~~~~~~~~~~~~~~~~~
include/asm-generic/bug.h:136:2: note: in expansion of macro 'unlikely'
  unlikely(__ret_warn_on);     \
  ^~~~~~~~
arch/arm/kernel/hw_breakpoint.c:863:3: note: in expansion of macro 'WARN'
   WARN(1, "Asynchronous watchpoint exception taken. Debugging results may be unreliable\n");
   ^~~~
arch/arm/kernel/hw_breakpoint.c:864:2: note: here
  case ARM_ENTRY_SYNC_WATCHPOINT:
  ^~~~
arch/arm/kernel/hw_breakpoint.c: In function 'core_has_os_save_restore':
arch/arm/kernel/hw_breakpoint.c:910:6: warning: this statement may fall through [-Wimplicit-fallthrough=]
   if (oslsr & ARM_OSLSR_OSLM0)
      ^
arch/arm/kernel/hw_breakpoint.c:912:2: note: here
  default:
  ^~~~~~~

Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Gustavo A. R. Silva <gustavo@embeddedor.com>
10 months agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Fri, 9 Aug 2019 22:46:29 +0000 (15:46 -0700)]
Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:
 "Bugfixes (arm and x86) and cleanups"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  selftests: kvm: Adding config fragments
  KVM: selftests: Update gitignore file for latest changes
  kvm: remove unnecessary PageReserved check
  KVM: arm/arm64: vgic: Reevaluate level sensitive interrupts on enable
  KVM: arm: Don't write junk to CP15 registers on reset
  KVM: arm64: Don't write junk to sysregs on reset
  KVM: arm/arm64: Sync ICH_VMCR_EL2 back when about to block
  x86: kvm: remove useless calls to kvm_para_available
  KVM: no need to check return value of debugfs_create functions
  KVM: remove kvm_arch_has_vcpu_debugfs()
  KVM: Fix leak vCPU's VMCS value into other pCPU
  KVM: Check preempted_in_kernel for involuntary preemption
  KVM: LAPIC: Don't need to wakeup vCPU twice afer timer fire
  arm64: KVM: hyp: debug-sr: Mark expected switch fall-through
  KVM: arm64: Update kvm_arm_exception_class and esr_class_str for new EC
  KVM: arm: vgic-v3: Mark expected switch fall-through
  arm64: KVM: regmap: Fix unexpected switch fall-through
  KVM: arm/arm64: Introduce kvm_pmu_vcpu_init() to setup PMU counter index

10 months agoMerge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Linus Torvalds [Fri, 9 Aug 2019 22:31:19 +0000 (15:31 -0700)]
Merge branch 'for-linus' of git://git./linux/kernel/git/dtor/input

Pull input updates from Dmitry Torokhov:

 - newer systems with Elan touchpads will be switched over to SMBus

 - HP Spectre X360 will be using SMbus/RMI4

 - checks for invalid USB descriptors in kbtab and iforce

 - build fixes for applespi driver (misconfigs)

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: iforce - add sanity checks
  Input: applespi - use struct_size() helper
  Input: kbtab - sanity check for endpoint type
  Input: usbtouchscreen - initialize PM mutex before using it
  Input: applespi - add dependency on LEDS_CLASS
  Input: synaptics - enable RMI mode for HP Spectre X360
  Input: elantech - annotate fall-through case in elantech_use_host_notify()
  Input: elantech - enable SMBus on new (2018+) systems
  Input: applespi - fix trivial typo in struct description
  Input: applespi - select CRC16 module
  Input: applespi - fix warnings detected by sparse

10 months agomm/memremap: Fix reuse of pgmap instances with internal references
Dan Williams [Thu, 8 Aug 2019 21:43:49 +0000 (14:43 -0700)]
mm/memremap: Fix reuse of pgmap instances with internal references

Currently, attempts to shutdown and re-enable a device-dax instance
trigger:

    Missing reference count teardown definition
    WARNING: CPU: 37 PID: 1608 at mm/memremap.c:211 devm_memremap_pages+0x234/0x850
    [..]
    RIP: 0010:devm_memremap_pages+0x234/0x850
    [..]
    Call Trace:
     dev_dax_probe+0x66/0x190 [device_dax]
     really_probe+0xef/0x390
     driver_probe_device+0xb4/0x100
     device_driver_attach+0x4f/0x60

Given that the setup path initializes pgmap->ref, arrange for it to be
also torn down so devm_memremap_pages() is ready to be called again and
not be mistaken for the 3rd-party per-cpu-ref case.

Fixes: 24917f6b1041 ("memremap: provide an optional internal refcount in struct dev_pagemap")
Reported-by: Fan Du <fan.du@intel.com>
Tested-by: Vishal Verma <vishal.l.verma@intel.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Ira Weiny <ira.weiny@intel.com>
Cc: Jason Gunthorpe <jgg@mellanox.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Link: https://lore.kernel.org/r/156530042781.2068700.8733813683117819799.stgit@dwillia2-desk3.amr.corp.intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
10 months agodrm/i915: Remove redundant user_access_end() from __copy_from_user() error path
Josh Poimboeuf [Thu, 25 Jul 2019 20:29:57 +0000 (15:29 -0500)]
drm/i915: Remove redundant user_access_end() from __copy_from_user() error path

Objtool reports:

  drivers/gpu/drm/i915/gem/i915_gem_execbuffer.o: warning: objtool: .altinstr_replacement+0x36: redundant UACCESS disable

__copy_from_user() already does both STAC and CLAC, so the
user_access_end() in its error path adds an extra unnecessary CLAC.

Fixes: 0b2c8f8b6b0c ("i915: fix missing user_access_end() in page fault exception case")
Reported-by: Thomas Gleixner <tglx@linutronix.de>
Reported-by: Sedat Dilek <sedat.dilek@gmail.com>
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://github.com/ClangBuiltLinux/linux/issues/617
Link: https://lkml.kernel.org/r/51a4155c5bc2ca847a9cbe85c1c11918bb193141.1564086017.git.jpoimboe@redhat.com
10 months agokbuild: show hint if subdir-y/m is used to visit module Makefile
Masahiro Yamada [Thu, 8 Aug 2019 11:21:11 +0000 (20:21 +0900)]
kbuild: show hint if subdir-y/m is used to visit module Makefile

Since commit ff9b45c55b26 ("kbuild: modpost: read modules.order instead
of $(MODVERDIR)/*.mod"), a module is no longer built in the following
pattern:

  [Makefile]
  subdir-y := some-module

  [some-module/Makefile]
  obj-m := some-module.o

You cannot write Makefile this way in upstream because modules.order is
not correctly generated. subdir-y is used to descend to a sub-directory
that builds tools, device trees, etc.

For external modules, the modules order does not matter. So, the
Makefile above was known to work.

I believe the Makefile should be re-written as follows:

  [Makefile]
  obj-m := some-module/

  [some-module/Makefile]
  obj-m := some-module.o

However, people will have no idea if their Makefile suddenly stops
working. In fact, I received questions from multiple people.

Show a warning for a while if obj-m is specified in a Makefile visited
by subdir-y or subdir-m.

I touched the %/ rule to avoid false-positive warnings for the single
target.

Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Tom Stonecypher <thomas.edwardx.stonecypher@intel.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Tested-by: Jan Kiszka <jan.kiszka@siemens.com>
10 months agokbuild: generate modules.order only in directories visited by obj-y/m
Masahiro Yamada [Tue, 6 Aug 2019 10:03:22 +0000 (19:03 +0900)]
kbuild: generate modules.order only in directories visited by obj-y/m

The modules.order files in directories visited by the chain of obj-y
or obj-m are merged to the upper-level ones, and become parts of the
top-level modules.order. On the other hand, there is no need to
generate modules.order in directories visited by subdir-y or subdir-m
since they would become orphan anyway.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
10 months agokbuild: fix false-positive need-builtin calculation
Masahiro Yamada [Tue, 6 Aug 2019 10:03:21 +0000 (19:03 +0900)]
kbuild: fix false-positive need-builtin calculation

The current implementation of need-builtin is false-positive,
for example, in the following Makefile:

  obj-m := foo/
  obj-y := foo/bar/

..., where foo/built-in.a is not required.

Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
10 months agokbuild: revive single target %.ko
Masahiro Yamada [Fri, 2 Aug 2019 10:23:58 +0000 (19:23 +0900)]
kbuild: revive single target %.ko

I removed the single target %.ko in commit ff9b45c55b26 ("kbuild:
modpost: read modules.order instead of $(MODVERDIR)/*.mod") because
the modpost stage does not work reliably. For instance, the module
dependency, modversion, etc. do not work if we lack symbol information
from the other modules.

Yet, some people still want to build only one module in their interest,
and it may be still useful if it is used within those limitations.

Fixes: ff9b45c55b26 ("kbuild: modpost: read modules.order instead of $(MODVERDIR)/*.mod")
Reported-by: Don Brace <don.brace@microsemi.com>
Reported-by: Arend Van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
10 months agoMerge tag 'drm-fixes-2019-08-09' of git://anongit.freedesktop.org/drm/drm
Linus Torvalds [Fri, 9 Aug 2019 16:35:23 +0000 (09:35 -0700)]
Merge tag 'drm-fixes-2019-08-09' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Usual fixes roundup. Nothing too crazy or serious, one non-released
  ioctl is removed in the amdkfd driver.

  core:
   - mode parser strncpy fix

  i915:
   - GLK DSI escape clock setting
   - HDCP memleak fix

  tegra:
   - one gpiod/of regression fix

  amdgpu:
   - fix VCN to handle the latest navi10 firmware
   - fix for fan control on navi10
   - properly handle SMU metrics table on navi10
   - fix a resume regression on Stoney
   - kfd revert a GWS ioctl

  vmwgfx:
   - memory leak fix

  rockchip:
   - suspend fix"

* tag 'drm-fixes-2019-08-09' of git://anongit.freedesktop.org/drm/drm:
  drm/vmwgfx: fix memory leak when too many retries have occurred
  Revert "drm/amdkfd: New IOCTL to allocate queue GWS"
  Revert "drm/amdgpu: fix transform feedback GDS hang on gfx10 (v2)"
  drm/amdgpu: pin the csb buffer on hw init for gfx v8
  drm/rockchip: Suspend DP late
  drm/i915: Fix wrong escape clock divisor init for GLK
  drm/i915: fix possible memory leak in intel_hdcp_auth_downstream()
  drm/modes: Fix unterminated strncpy
  drm/amd/powerplay: correct navi10 vcn powergate
  drm/amd/powerplay: honor hw limit on fetching metrics data for navi10
  drm/amd/powerplay: Allow changing of fan_control in smu_v11_0
  drm/amd/amdgpu/vcn_v2_0: Move VCN 2.0 specific dec ring test to vcn_v2_0
  drm/amd/amdgpu/vcn_v2_0: Mark RB commands as KMD commands
  drm/tegra: Fix gpiod_get_from_of_node() regression

10 months agoMerge tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux
Linus Torvalds [Fri, 9 Aug 2019 16:31:40 +0000 (09:31 -0700)]
Merge tag 'arm64-fixes' of git://git./linux/kernel/git/arm64/linux

Pull arm64 fix from Catalin Marinas:
 "Fix bad_pte warning caused by pte_mkdevmap() not setting PTE_SPECIAL"

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64: mm: add missing PTE_SPECIAL in pte_mkdevmap on arm64

10 months agoMerge tag 's390-5.3-5' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Fri, 9 Aug 2019 16:30:00 +0000 (09:30 -0700)]
Merge tag 's390-5.3-5' of git://git./linux/kernel/git/s390/linux

Pull s390 fixes from Vasily Gorbik:

 - Map vdso also for statically linked binaries like all other
   architectures.

 - Fix no .bss usage compile-time check to account common objects with
   the help of binutils size tool. Top level Makefile change acked-by
   Masahiro.

 - A fix to make perf happy with _etext symbol type.

 - Fix dump_pagetables which is broken since p*d_offset implementation
   change to comply with mm/gup.c expectations.

 - Revert memory sharing for diag calls in protected virtualization,
   since this is not required after all.

 - Couple of other minor code cleanups.

* tag 's390-5.3-5' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/vdso: map vdso also for statically linked binaries
  s390/build: use size command to perform empty .bss check
  kbuild: add OBJSIZE variable for the size tool
  s390: put _stext and _etext into .text section
  s390/head64: cleanup unused labels
  s390/unwind: remove stack recursion warning
  s390/setup: adjust start_code of init_mm to _text
  s390/mm: fix dump_pagetables top level page table walking
  s390/protvirt: avoid memory sharing for diag 308 set/store

10 months agoMerge tag 'for-linus-20190809' of git://git.kernel.dk/linux-block
Linus Torvalds [Fri, 9 Aug 2019 16:28:18 +0000 (09:28 -0700)]
Merge tag 'for-linus-20190809' of git://git.kernel.dk/linux-block

Pull block fixes from Jens Axboe:

 - Revert of a bcache patch that caused an oops for some (Coly)

 - ata rb532 unused warning fix (Gustavo)

 - AoE kernel crash fix (He)

 - Error handling fixup for blkdev_get() (Jan)

 - libata read/write translation and SFF PIO fix (me)

 - Use after free and error handling fix for O_DIRECT fragments. There's
   still a nowait + sync oddity in there, we'll nail that start next
   week. If all else fails, I'll queue a revert of the NOWAIT change.
   (me)

 - Loop GFP_KERNEL -> GFP_NOIO deadlock fix (Mikulas)

 - Two BFQ regression fixes that caused crashes (Paolo)

* tag 'for-linus-20190809' of git://git.kernel.dk/linux-block:
  bcache: Revert "bcache: use sysfs_match_string() instead of __sysfs_match_string()"
  loop: set PF_MEMALLOC_NOIO for the worker thread
  bdev: Fixup error handling in blkdev_get()
  block, bfq: handle NULL return value by bfq_init_rq()
  block, bfq: move update of waker and woken list to queue freeing
  block, bfq: reset last_completed_rq_bfqq if the pointed queue is freed
  block: aoe: Fix kernel crash due to atomic sleep when exiting
  libata: add SG safety checks in SFF pio transfers
  libata: have ata_scsi_rw_xlat() fail invalid passthrough requests
  block: fix O_DIRECT error handling for bio fragments
  ata: rb532_cf: Fix unused variable warning in rb532_pata_driver_probe

10 months agoMerge tag 'mmc-v5.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc
Linus Torvalds [Fri, 9 Aug 2019 16:26:47 +0000 (09:26 -0700)]
Merge tag 'mmc-v5.3-rc3' of git://git./linux/kernel/git/ulfh/mmc

Pull MMC fixes from Ulf Hansson:

 - cavium: Fix DMA support

 - sdhci-sprd: Fix soft reset when runtime resuming"

* tag 'mmc-v5.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/ulfh/mmc:
  mmc: cavium: Add the missing dma unmap when the dma has finished.
  mmc: cavium: Set the correct dma max segment size for mmc_host
  mmc: sdhci-sprd: Fix the incorrect soft reset operation when runtime resuming

10 months agoMerge tag 'fbdev-v5.3-rc4' of git://github.com/bzolnier/linux
Linus Torvalds [Fri, 9 Aug 2019 16:24:49 +0000 (09:24 -0700)]
Merge tag 'fbdev-v5.3-rc4' of git://github.com/bzolnier/linux

Pull fbdev fix from Bartlomiej Zolnierkiewicz:
 "fbdev patches will now go to upstream through drm-misc tree for
  improved maintainership and better integration testing so update
  MAINTAINERS file accordingly"

* tag 'fbdev-v5.3-rc4' of git://github.com/bzolnier/linux:
  MAINTAINERS: handle fbdev changes through drm-misc tree

10 months agoMerge tag 'pwm/for-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry...
Linus Torvalds [Fri, 9 Aug 2019 16:23:23 +0000 (09:23 -0700)]
Merge tag 'pwm/for-5.3-rc4' of git://git./linux/kernel/git/thierry.reding/linux-pwm

Pull pwm fix from Thierry Reding:
 "A single fix for a backlight brightness regression introduced in
  this merge window"

* tag 'pwm/for-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/thierry.reding/linux-pwm:
  pwm: Fallback to the static lookup-list when acpi_pwm_get fails

10 months agoMerge tag 'sound-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Fri, 9 Aug 2019 16:21:27 +0000 (09:21 -0700)]
Merge tag 'sound-5.3-rc4' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "Lots of small fixes at this time since we've received the ASoC fix
  batch now.

   - Some coverage in ASoC core mostly for minor issues like NULL checks
     for DPCM and proper error handling in DAI instantiation

   - A collection of small device-specific changes in various ASoC codec
     and platform drivers

   - OF-tree refcount fixes in a few ASoC drivers

   - Fixes of memory leaks in the error paths of various ASoC / ALSA
     drivers

   - A workaround for a long-standing issue on AMD HD-audio device

   - Updates of MAINTAINERS, mail addresses, file permission fixups"

* tag 'sound-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound: (38 commits)
  ALSA: firewire: fix a memory leak bug
  sound: fix a memory leak bug
  ALSA: hda - Workaround for crackled sound on AMD controller (1022:1457)
  ALSA: hiface: fix multiple memory leak bugs
  ALSA: hda - Don't override global PCM hw info flag
  ALSA: usb-audio: fix a memory leak bug
  ASoC: max98373: Remove executable bits
  ASoC: amd: acp3x: use dma address for acp3x dma driver
  ASoC: amd: acp3x: use dma_ops of parent device for acp3x dma driver
  ASoC: max98373: add 88200 and 96000 sampling rate support
  ASoC: sun4i-i2s: Incorrect SR and WSS computation
  MAINTAINERS: Update Intel ASoC drivers maintainers
  ASoC: ti: davinci-mcasp: Correct slot_width posed constraint
  ASoC: rockchip: Fix mono capture
  ASoC: Intel: Fix some acpi vs apci typo in somme comments
  ASoC: ti: davinci-mcasp: Fix clk PDIR handling for i2s master mode
  ASoC: Fail card instantiation if DAI format setup fails
  ASoC: SOF: Intel: hda: remove misleading error trace from IRQ thread
  ASoC: qcom: apq8016_sbc: Fix oops with multiple DAI links
  ASoC: dapm: fix a memory leak bug
  ...

10 months agoMerge tag 'media/v5.3-3' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab...
Linus Torvalds [Fri, 9 Aug 2019 16:19:53 +0000 (09:19 -0700)]
Merge tag 'media/v5.3-3' of git://git./linux/kernel/git/mchehab/linux-media

Pull media fix from Mauro Carvalho Chehab:
 "A fix at the vivid CEC support"

* tag 'media/v5.3-3' of git://git.kernel.org/pub/scm/linux/kernel/git/mchehab/linux-media:
  media: vivid: fix missing cec adapter name

10 months agoMerge tag 'pm-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Linus Torvalds [Fri, 9 Aug 2019 16:18:36 +0000 (09:18 -0700)]
Merge tag 'pm-5.3-rc4' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management fix from Rafael Wysocki:
 "Revert a recent PCI power management change that caused problems to
  occur on multiple systems (Mika Westerberg)"

* tag 'pm-5.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  Revert "PCI: Add missing link delays required by the PCIe spec"

10 months agoMerge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Fri, 9 Aug 2019 16:17:05 +0000 (09:17 -0700)]
Merge branch 'linus' of git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fixes from Herbert Xu:
 "Fix a number of bugs in the ccp driver"

* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: ccp - Ignore tag length when decrypting GCM ciphertext
  crypto: ccp - Add support for valid authsize values less than 16
  crypto: ccp - Fix oops by properly managing allocated structures

10 months agogfs2: gfs2_walk_metadata fix
Andreas Gruenbacher [Mon, 5 Aug 2019 11:22:03 +0000 (12:22 +0100)]
gfs2: gfs2_walk_metadata fix

It turns out that the current version of gfs2_metadata_walker suffers
from multiple problems that can cause gfs2_hole_size to report an
incorrect size.  This will confuse fiemap as well as lseek with the
SEEK_DATA flag.

Fix that by changing gfs2_hole_walker to compute the metapath to the
first data block after the hole (if any), and compute the hole size
based on that.

Fixes xfstest generic/490.

Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
Reviewed-by: Bob Peterson <rpeterso@redhat.com>
Cc: stable@vger.kernel.org # v4.18+
10 months agoMerge tag 'kvmarm-fixes-for-5.3-2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Paolo Bonzini [Fri, 9 Aug 2019 14:53:50 +0000 (16:53 +0200)]
Merge tag 'kvmarm-fixes-for-5.3-2' of git://git./linux/kernel/git/kvmarm/kvmarm into HEAD

KVM/arm fixes for 5.3, take #2

- Fix our system register reset so that we stop writing
  non-sensical values to them, and track which registers
  get reset instead.
- Sync VMCR back from the GIC on WFI so that KVM has an
  exact vue of PMR.
- Reevaluate state of HW-mapped, level triggered interrupts
  on enable.

10 months agoMerge tag 'kvmarm-fixes-for-5.3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Paolo Bonzini [Fri, 9 Aug 2019 14:53:39 +0000 (16:53 +0200)]
Merge tag 'kvmarm-fixes-for-5.3' of git://git./linux/kernel/git/kvmarm/kvmarm into HEAD

KVM/arm fixes for 5.3

- A bunch of switch/case fall-through annotation, fixing one actual bug
- Fix PMU reset bug
- Add missing exception class debug strings

10 months agoselftests: kvm: Adding config fragments
Naresh Kamboju [Wed, 7 Aug 2019 13:58:14 +0000 (13:58 +0000)]
selftests: kvm: Adding config fragments

selftests kvm test cases need pre-required kernel configs for the test
to get pass.

Signed-off-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 months agoKVM: selftests: Update gitignore file for latest changes
Thomas Huth [Wed, 31 Jul 2019 14:28:51 +0000 (16:28 +0200)]
KVM: selftests: Update gitignore file for latest changes

The kvm_create_max_vcpus test has been moved to the main directory,
and sync_regs_test is now available on s390x, too.

Signed-off-by: Thomas Huth <thuth@redhat.com>
Acked-by: Shuah Khan <skhan@linuxfoundation.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 months agokvm: remove unnecessary PageReserved check
Paolo Bonzini [Mon, 5 Aug 2019 09:11:08 +0000 (11:11 +0200)]
kvm: remove unnecessary PageReserved check

The same check is already done in kvm_is_reserved_pfn.

Reviewed-by: Sean Christopherson <sean.j.christopherson@intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
10 months agoMAINTAINERS: handle fbdev changes through drm-misc tree
Bartlomiej Zolnierkiewicz [Fri, 9 Aug 2019 13:46:40 +0000 (15:46 +0200)]
MAINTAINERS: handle fbdev changes through drm-misc tree

fbdev patches will now go to upstream through drm-misc tree (IOW
starting with v5.4 merge window fbdev changes will be included in
DRM pull request) for improved maintainership and better integration
testing. Update MAINTAINERS file accordingly.

Acked-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>
10 months agobcache: Revert "bcache: use sysfs_match_string() instead of __sysfs_match_string()"
Coly Li [Fri, 9 Aug 2019 06:14:05 +0000 (14:14 +0800)]
bcache: Revert "bcache: use sysfs_match_string() instead of __sysfs_match_string()"

This reverts commit 89e0341af082dbc170019f908846f4a424efc86b.

In drivers/md/bcache/sysfs.c:bch_snprint_string_list(), NULL pointer at
the end of list is necessary. Remove the NULL from last element of each
lists will cause the following panic,

[ 4340.455652] bcache: register_cache() registered cache device nvme0n1
[ 4340.464603] bcache: register_bdev() registered backing device sdk
[ 4421.587335] bcache: bch_cached_dev_run() cached dev sdk is running already
[ 4421.587348] bcache: bch_cached_dev_attach() Caching sdk as bcache0 on set 354e1d46-d99f-4d8b-870b-078b80dc88a6
[ 5139.247950] general protection fault: 0000 [#1] SMP NOPTI
[ 5139.247970] CPU: 9 PID: 5896 Comm: cat Not tainted 4.12.14-95.29-default #1 SLE12-SP4
[ 5139.247988] Hardware name: HPE ProLiant DL380 Gen10/ProLiant DL380 Gen10, BIOS U30 04/18/2019
[ 5139.248006] task: ffff888fb25c0b00 task.stack: ffff9bbacc704000
[ 5139.248021] RIP: 0010:string+0x21/0x70
[ 5139.248030] RSP: 0018:ffff9bbacc707bf0 EFLAGS: 00010286
[ 5139.248043] RAX: ffffffffa7e432e3 RBX: ffff8881c20da02a RCX: ffff0a00ffffff04
[ 5139.248058] RDX: 3f00656863616362 RSI: ffff8881c20db000 RDI: ffffffffffffffff
[ 5139.248075] RBP: ffff8881c20db000 R08: 0000000000000000 R09: ffff8881c20da02a
[ 5139.248090] R10: 0000000000000004 R11: 0000000000000000 R12: ffff9bbacc707c48
[ 5139.248104] R13: 0000000000000fd6 R14: ffffffffc0665855 R15: ffffffffc0665855
[ 5139.248119] FS:  00007faf253b8700(0000) GS:ffff88903f840000(0000) knlGS:0000000000000000
[ 5139.248137] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 5139.248149] CR2: 00007faf25395008 CR3: 0000000f72150006 CR4: 00000000007606e0
[ 5139.248164] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 5139.248179] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 5139.248193] PKRU: 55555554
[ 5139.248200] Call Trace:
[ 5139.248210]  vsnprintf+0x1fb/0x510
[ 5139.248221]  snprintf+0x39/0x40
[ 5139.248238]  bch_snprint_string_list.constprop.15+0x5b/0x90 [bcache]
[ 5139.248256]  __bch_cached_dev_show+0x44d/0x5f0 [bcache]
[ 5139.248270]  ? __alloc_pages_nodemask+0xb2/0x210
[ 5139.248284]  bch_cached_dev_show+0x2c/0x50 [bcache]
[ 5139.248297]  sysfs_kf_seq_show+0xbb/0x190
[ 5139.248308]  seq_read+0xfc/0x3c0
[ 5139.248317]  __vfs_read+0x26/0x140
[ 5139.248327]  vfs_read+0x87/0x130
[ 5139.248336]  SyS_read+0x42/0x90
[ 5139.248346]  do_syscall_64+0x74/0x160
[ 5139.248358]  entry_SYSCALL_64_after_hwframe+0x3d/0xa2
[ 5139.248370] RIP: 0033:0x7faf24eea370
[ 5139.248379] RSP: 002b:00007fff82d03f38 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
[ 5139.248395] RAX: ffffffffffffffda RBX: 0000000000020000 RCX: 00007faf24eea370
[ 5139.248411] RDX: 0000000000020000 RSI: 00007faf25396000 RDI: 0000000000000003
[ 5139.248426] RBP: 00007faf25396000 R08: 00000000ffffffff R09: 0000000000000000
[ 5139.248441] R10: 000000007c9d4d41 R11: 0000000000000246 R12: 00007faf25396000
[ 5139.248456] R13: 0000000000000003 R14: 0000000000000000 R15: 0000000000000fff
[ 5139.248892] Code: ff ff ff 0f 1f 80 00 00 00 00 49 89 f9 48 89 cf 48 c7 c0 e3 32 e4 a7 48 c1 ff 30 48 81 fa ff 0f 00 00 48 0f 46 d0 48 85 ff 74 45 <44> 0f b6 02 48 8d 42 01 45 84 c0 74 38 48 01 fa 4c 89 cf eb 0e

The simplest way to fix is to revert commit 89e0341af082 ("bcache: use
sysfs_match_string() instead of __sysfs_match_string()").

This bug was introduced in Linux v5.2, so this fix only applies to
Linux v5.2 is enough for stable tree maintainer.

Fixes: 89e0341af082 ("bcache: use sysfs_match_string() instead of __sysfs_match_string()")
Cc: stable@vger.kernel.org
Cc: Alexandru Ardelean <alexandru.ardelean@analog.com>
Reported-by: Peifeng Lin <pflin@suse.com>
Acked-by: Alexandru Ardelean <alexandru.ardelean@analog.com>
Signed-off-by: Coly Li <colyli@suse.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
10 months agos390/vdso: map vdso also for statically linked binaries
Heiko Carstens [Wed, 7 Aug 2019 12:31:59 +0000 (14:31 +0200)]
s390/vdso: map vdso also for statically linked binaries

s390 does not map the vdso for statically linked binaries, assuming
that this doesn't make sense. See commit fc5243d98ac2 ("[S390]
arch_setup_additional_pages arguments").

However with glibc commit d665367f596d ("linux: Enable vDSO for static
linking as default (BZ#19767)") and commit 5e855c895401 ("s390: Enable
VDSO for static linking") the vdso is also used for statically linked
binaries - if the kernel would make it available.

Therefore map the vdso always, just like all other architectures.

Reported-by: Stefan Liebler <stli@linux.ibm.com>
Signed-off-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Vasily Gorbik <gor@linux.ibm.com>
10 months agoKVM: arm/arm64: vgic: Reevaluate level sensitive interrupts on enable
Alexandru Elisei [Wed, 7 Aug 2019 09:53:20 +0000 (10:53 +0100)]
KVM: arm/arm64: vgic: Reevaluate level sensitive interrupts on enable

A HW mapped level sensitive interrupt asserted by a device will not be put
into the ap_list if it is disabled at the VGIC level. When it is enabled
again, it will be inserted into the ap_list and written to a list register
on guest entry regardless of the state of the device.

We could argue that this can also happen on real hardware, when the command
to enable the interrupt reached the GIC before the device had the chance to
de-assert the interrupt signal; however, we emulate the distributor and
redistributors in software and we can do better than that.

Signed-off-by: Alexandru Elisei <alexandru.elisei@arm.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
10 months agoKVM: arm: Don't write junk to CP15 registers on reset
Marc Zyngier [Mon, 5 Aug 2019 09:34:51 +0000 (10:34 +0100)]
KVM: arm: Don't write junk to CP15 registers on reset

At the moment, the way we reset CP15 registers is mildly insane:
We write junk to them, call the reset functions, and then check that
we have something else in them.

The "fun" thing is that this can happen while the guest is running
(PSCI, for example). If anything in KVM has to evaluate the state
of a CP15 register while junk is in there, bad thing may happen.

Let's stop doing that. Instead, we track that we have called a
reset function for that register, and assume that the reset
function has done something.

In the end, the very need of this reset check is pretty dubious,
as it doesn't check everything (a lot of the CP15 reg leave outside
of the cp15_regs[] array). It may well be axed in the near future.

Signed-off-by: Marc Zyngier <maz@kernel.org>
10 months agoKVM: arm64: Don't write junk to sysregs on reset
Marc Zyngier [Mon, 5 Aug 2019 09:34:51 +0000 (10:34 +0100)]
KVM: arm64: Don't write junk to sysregs on reset

At the moment, the way we reset system registers is mildly insane:
We write junk to them, call the reset functions, and then check that
we have something else in them.

The "fun" thing is that this can happen while the guest is running
(PSCI, for example). If anything in KVM has to evaluate the state
of a system register while junk is in there, bad thing may happen.

Let's stop doing that. Instead, we track that we have called a
reset function for that register, and assume that the reset
function has done something. This requires fixing a couple of
sysreg refinition in the trap table.

In the end, the very need of this reset check is pretty dubious,
as it doesn't check everything (a lot of the sysregs leave outside of
the sys_regs[] array). It may well be axed in the near future.

Tested-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
10 months agoMerge tag 'drm-intel-fixes-2019-08-08' of git://anongit.freedesktop.org/drm/drm-intel...
Dave Airlie [Fri, 9 Aug 2019 05:46:09 +0000 (15:46 +1000)]
Merge tag 'drm-intel-fixes-2019-08-08' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes

drm/i915 fixes for v5.3-rc4:
- Fix GLK DSI escape clock setting
- Fix a memleak on HDCP revoked Ksv error path

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/87pnlghz79.fsf@intel.com
10 months agoMerge tag 'drm-misc-fixes-2019-08-08' of git://anongit.freedesktop.org/drm/drm-misc...
Dave Airlie [Fri, 9 Aug 2019 05:45:02 +0000 (15:45 +1000)]
Merge tag 'drm-misc-fixes-2019-08-08' of git://anongit.freedesktop.org/drm/drm-misc into drm-fixes

drm-misc-fixes for v5.3-rc4:
- Suspend fix for rockchip
- Fix unterminated strncpy cmdline mode parser

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/ace294a6-6bb2-d9b1-695d-3260e1d60831@linux.intel.com
10 months agoMerge branch 'vmwgfx-fixes-5.3' of git://people.freedesktop.org/~thomash/linux into...
Dave Airlie [Fri, 9 Aug 2019 05:40:06 +0000 (15:40 +1000)]
Merge branch 'vmwgfx-fixes-5.3' of git://people.freedesktop.org/~thomash/linux into drm-fixes

One single memory leak fix.

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Thomas Hellstrom "VMware" <thomas@shipmail.org>
Link: https://patchwork.freedesktop.org/patch/msgid/20190808094615.31040-1-thomas@shipmail.org
10 months agodt-bindings: riscv: fix the schema compatible string for the HiFive Unleashed board
Paul Walmsley [Thu, 8 Aug 2019 22:36:44 +0000 (15:36 -0700)]
dt-bindings: riscv: fix the schema compatible string for the HiFive Unleashed board

The YAML binding document for SiFive boards has an incorrect
compatible string for the HiFive Unleashed board.  Change it to match
the name of the board on the SiFive web site:

   https://www.sifive.com/boards/hifive-unleashed

which also matches the contents of the board DT data file:

   https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/arch/riscv/boot/dts/sifive/hifive-unleashed-a00.dts#n13

Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Acked-by: Rob Herring <robh@kernel.org>
10 months agodt-bindings: riscv: remove obsolete cpus.txt
Paul Walmsley [Thu, 8 Aug 2019 21:46:58 +0000 (14:46 -0700)]
dt-bindings: riscv: remove obsolete cpus.txt

Remove the now-obsolete riscv/cpus.txt DT binding document, since we
are using YAML binding documentation instead.

While doing so, transfer the explanatory text about 'harts' (with some
edits) into the YAML file, at Rob's request.

Link: https://lore.kernel.org/linux-riscv/CAL_JsqJs6MtvmuyAknsUxQymbmoV=G+=JfS1PQj9kNHV7fjC9g@mail.gmail.com/
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Rob Herring <robh@kernel.org>
Reviewed-by: Rob Herring <robh@kernel.org>
10 months agoRISC-V: Remove udivdi3
Palmer Dabbelt [Wed, 7 Aug 2019 00:05:08 +0000 (17:05 -0700)]
RISC-V: Remove udivdi3

This should never have landed in the first place: it was added as part
of 64-bit divide support for 32-bit systems, but the kernel doesn't
allow this sort of division.  I must have forgotten to remove it.

This patch removes the support.  Since this routine only worked on
64-bit platforms but was only built on 32-bit platforms, it's
essentially just nonsense anyway.

Signed-off-by: Palmer Dabbelt <palmer@sifive.com>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Link: https://lore.kernel.org/linux-riscv/nycvar.YSQ.7.76.1908061413360.19480@knanqh.ubzr/T/#t
Reported-by: Eric Lin <tesheng@andestech.com>
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
10 months agoriscv: delay: use do_div() instead of __udivdi3()
Paul Walmsley [Wed, 7 Aug 2019 01:28:33 +0000 (18:28 -0700)]
riscv: delay: use do_div() instead of __udivdi3()

In preparation for removing __udivdi3() from the RISC-V
architecture-specific files, convert its one user to use do_div().
This avoids breaking the RV32 build after __udivdi3() is removed.

This second version removes the assignment of the remainder to an
unused temporary variable.  Thanks to Nicolas Pitre <nico@fluxnic.net>
for the suggestion.

Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Nicolas Pitre <nico@fluxnic.net>
10 months agodt-bindings: Update the riscv,isa string description
Atish Patra [Sat, 3 Aug 2019 04:27:23 +0000 (21:27 -0700)]
dt-bindings: Update the riscv,isa string description

Since the RISC-V specification states that ISA description strings are
case-insensitive, there's no functional difference between mixed-case,
upper-case, and lower-case ISA strings. Thus, to simplify parsing,
specify that the letters present in "riscv,isa" must be all lowercase.

Suggested-by: Paul Walmsley <paul.walmsley@sifive.com>
Signed-off-by: Atish Patra <atish.patra@wdc.com>
Signed-off-by: Paul Walmsley <paul.walmsley@sifive.com>
10 months agoMerge tag 'nfs-for-5.3-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs
Linus Torvalds [Thu, 8 Aug 2019 21:47:19 +0000 (14:47 -0700)]
Merge tag 'nfs-for-5.3-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client fixes from Trond Myklebust:
 "Highlights include:

  Stable fixes:

   - NFSv4: Ensure we check the return value of update_open_stateid() so
     we correctly track active open state.

   - NFSv4: Fix for delegation state recovery to ensure we recover all
     open modes that are active.

   - NFSv4: Fix an Oops in nfs4_do_setattr

  Fixes:

   - NFS: Fix regression whereby fscache errors are appearing on 'nofsc'
     mounts

   - NFSv4: Fix a potential sleep while atomic in nfs4_do_reclaim()

   - NFSv4: Fix a credential refcount leak in nfs41_check_delegation_stateid

   - pNFS: Report errors from the call to nfs4_select_rw_stateid()

   - NFSv4: Various other delegation and open stateid recovery fixes

   - NFSv4: Fix state recovery behaviour when server connection times
     out"

* tag 'nfs-for-5.3-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  NFSv4: Ensure state recovery handles ETIMEDOUT correctly
  NFS: Fix regression whereby fscache errors are appearing on 'nofsc' mounts
  NFSv4: Fix an Oops in nfs4_do_setattr
  NFSv4: Fix a potential sleep while atomic in nfs4_do_reclaim()
  NFSv4: Check the return value of update_open_stateid()
  NFSv4.1: Only reap expired delegations
  NFSv4.1: Fix open stateid recovery
  NFSv4: Report the error from nfs4_select_rw_stateid()
  NFSv4: When recovering state fails with EAGAIN, retry the same recovery
  NFSv4: Print an error in the syslog when state is marked as irrecoverable
  NFSv4: Fix delegation state recovery
  NFSv4: Fix a credential refcount leak in nfs41_check_delegation_stateid

10 months agoMerge tag 'perf-urgent-for-mingo-5.3-20190808' of git://git.kernel.org/pub/scm/linux...
Thomas Gleixner [Thu, 8 Aug 2019 20:12:01 +0000 (22:12 +0200)]
Merge tag 'perf-urgent-for-mingo-5.3-20190808' of git://git./linux/kernel/git/acme/linux into perf/urgent

perf/urgent fixes:

db-export:

  Adrian Hunter:

  - Fix thread__exec_comm() picking of main thread COMM for pre-existing,
    synthesized from /proc data records.

annotate:

  Arnaldo Carvalho de Melo:

  - Fix printing of unaugmented disassembled instructions from BPF, some
    lines were leaving leftovers from the previous screen, due to use of
    newlines by binutils's libopcode disassembler.

perf ftrace:

  He Zhe:

  - Fix cpumask problems when only one CPU is present.

PMU events:

  Jin Yao:

  - Add missing "cpu_clk_unhalted.core" event.

perf bench:

  Jiri Olsa:

  - Fix cpu0 binding in the NUMA benchmarks.

s390:

  Thomas Richter:

  - Fix module size calculations.

build system:

  Masanari Iida:

  - Fix a typo in a variable name in the Documentation Makefile

misc:

  Ian Rogers:

  - Fix include paths in ui directory.

Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf pmu-events: Fix missing "cpu_clk_unhalted.core" event
Jin Yao [Mon, 29 Jul 2019 07:27:55 +0000 (15:27 +0800)]
perf pmu-events: Fix missing "cpu_clk_unhalted.core" event

The events defined in pmu-events JSON are parsed and added into perf
tool. For fixed counters, we handle the encodings between JSON and perf
by using a static array fixed[].

But the fixed[] has missed an important event "cpu_clk_unhalted.core".

For example, on the Tremont platform,

  [root@localhost ~]# perf stat -e cpu_clk_unhalted.core -a
  event syntax error: 'cpu_clk_unhalted.core'
                       \___ parser error

With this patch, the event cpu_clk_unhalted.core can be parsed.

  [root@localhost perf]# ./perf stat -e cpu_clk_unhalted.core -a -vvv
  ------------------------------------------------------------
  perf_event_attr:
    type                             4
    size                             112
    config                           0x3c
    sample_type                      IDENTIFIER
    read_format                      TOTAL_TIME_ENABLED|TOTAL_TIME_RUNNING
    disabled                         1
    inherit                          1
    exclude_guest                    1
  ------------------------------------------------------------
...

Signed-off-by: Jin Yao <yao.jin@linux.intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jin Yao <yao.jin@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20190729072755.2166-1-yao.jin@linux.intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf annotate: Fix s390 gap between kernel end and module start
Thomas Richter [Wed, 24 Jul 2019 12:27:03 +0000 (14:27 +0200)]
perf annotate: Fix s390 gap between kernel end and module start

During execution of command 'perf top' the error message:

   Not enough memory for annotating '__irf_end' symbol!)

is emitted from this call sequence:
  __cmd_top
    perf_top__mmap_read
      perf_top__mmap_read_idx
        perf_event__process_sample
          hist_entry_iter__add
            hist_iter__top_callback
              perf_top__record_precise_ip
                hist_entry__inc_addr_samples
                  symbol__inc_addr_samples
                    symbol__get_annotation
                      symbol__alloc_hist

In this function the size of symbol __irf_end is calculated. The size of
a symbol is the difference between its start and end address.

When the symbol was read the first time, its start and end was set to:

   symbol__new: __irf_end 0xe954d0-0xe954d0

which is correct and maps with /proc/kallsyms:

   root@s8360046:~/linux-4.15.0/tools/perf# fgrep _irf_end /proc/kallsyms
   0000000000e954d0 t __irf_end
   root@s8360046:~/linux-4.15.0/tools/perf#

In function symbol__alloc_hist() the end of symbol __irf_end is

  symbol__alloc_hist sym:__irf_end start:0xe954d0 end:0x3ff80045a8

which is identical with the first module entry in /proc/kallsyms

This results in a symbol size of __irf_req for histogram analyses of
70334140059072 bytes and a malloc() for this requested size fails.

The root cause of this is function
  __dso__load_kallsyms()
  +-> symbols__fixup_end()

Function symbols__fixup_end() enlarges the last symbol in the kallsyms
map:

   # fgrep __irf_end /proc/kallsyms
   0000000000e954d0 t __irf_end
   #

to the start address of the first module:
   # cat /proc/kallsyms | sort  | egrep ' [tT] '
   ....
   0000000000e952d0 T __security_initcall_end
   0000000000e954d0 T __initramfs_size
   0000000000e954d0 t __irf_end
   000003ff800045a8 T fc_get_event_number       [scsi_transport_fc]
   000003ff800045d0 t store_fc_vport_disable    [scsi_transport_fc]
   000003ff800046a8 T scsi_is_fc_rport  [scsi_transport_fc]
   000003ff800046d0 t fc_target_setup   [scsi_transport_fc]

On s390 the kernel is located around memory address 0x200, 0x10000 or
0x100000, depending on linux version. Modules however start some- where
around 0x3ff xxxx xxxx.

This is different than x86 and produces a large gap for which histogram
allocation fails.

Fix this by detecting the kernel's last symbol and do no adjustment for
it. Introduce a weak function and handle s390 specifics.

Reported-by: Klaus Theurich <klaus.theurich@de.ibm.com>
Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Hendrik Brueckner <brueckner@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: stable@vger.kernel.org
Link: http://lkml.kernel.org/r/20190724122703.3996-2-tmricht@linux.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf record: Fix module size on s390
Thomas Richter [Wed, 24 Jul 2019 12:27:02 +0000 (14:27 +0200)]
perf record: Fix module size on s390

On s390 the modules loaded in memory have the text segment located after
the GOT and Relocation table. This can be seen with this output:

  [root@m35lp76 perf]# fgrep qeth /proc/modules
  qeth 151552 1 qeth_l2, Live 0x000003ff800b2000
  ...
  [root@m35lp76 perf]# cat /sys/module/qeth/sections/.text
  0x000003ff800b3990
  [root@m35lp76 perf]#

There is an offset of 0x1990 bytes. The size of the qeth module is
151552 bytes (0x25000 in hex).

The location of the GOT/relocation table at the beginning of a module is
unique to s390.

commit 203d8a4aa6ed ("perf s390: Fix 'start' address of module's map")
adjusts the start address of a module in the map structures, but does
not adjust the size of the modules. This leads to overlapping of module
maps as this example shows:

[root@m35lp76 perf] # ./perf report -D
     0 0 0xfb0 [0xa0]: PERF_RECORD_MMAP -1/0: [0x3ff800b3990(0x25000)
          @ 0]:  x /lib/modules/.../qeth.ko.xz
     0 0 0x1050 [0xb0]: PERF_RECORD_MMAP -1/0: [0x3ff800d85a0(0x8000)
          @ 0]:  x /lib/modules/.../ip6_tables.ko.xz

The module qeth.ko has an adjusted start address modified to b3990, but
its size is unchanged and the module ends at 0x3ff800d8990.  This end
address overlaps with the next modules start address of 0x3ff800d85a0.

When the size of the leading GOT/Relocation table stored in the
beginning of the text segment (0x1990 bytes) is subtracted from module
qeth end address, there are no overlaps anymore:

   0x3ff800d8990 - 0x1990 = 0x0x3ff800d7000

which is the same as

   0x3ff800b2000 + 0x25000 = 0x0x3ff800d7000.

To fix this issue, also adjust the modules size in function
arch__fix_module_text_start(). Add another function parameter named size
and reduce the size of the module when the text segment start address is
changed.

Output after:
     0 0 0xfb0 [0xa0]: PERF_RECORD_MMAP -1/0: [0x3ff800b3990(0x23670)
          @ 0]:  x /lib/modules/.../qeth.ko.xz
     0 0 0x1050 [0xb0]: PERF_RECORD_MMAP -1/0: [0x3ff800d85a0(0x7a60)
          @ 0]:  x /lib/modules/.../ip6_tables.ko.xz

Reported-by: Stefan Liebler <stli@linux.ibm.com>
Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Hendrik Brueckner <brueckner@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Cc: stable@vger.kernel.org
Fixes: 203d8a4aa6ed ("perf s390: Fix 'start' address of module's map")
Link: http://lkml.kernel.org/r/20190724122703.3996-1-tmricht@linux.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf tools: Fix include paths in ui directory
Ian Rogers [Wed, 31 Jul 2019 22:54:41 +0000 (15:54 -0700)]
perf tools: Fix include paths in ui directory

These paths point to the wrong location but still work because they get
picked up by a -I flag that happens to direct to the correct file. Fix
paths to point to the correct location without -I flags.

Signed-off-by: Ian Rogers <irogers@google.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Link: http://lkml.kernel.org/r/20190731225441.233800-1-irogers@google.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf tools: Fix a typo in a variable name in the Documentation Makefile
Masanari Iida [Thu, 1 Aug 2019 03:28:12 +0000 (12:28 +0900)]
perf tools: Fix a typo in a variable name in the Documentation Makefile

This patch fix a spelling typo in a variable name in the Documentation Makefile.

Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Reviewed-by: Mukesh Ojha <mojha@codeaurora.org>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Link: http://lkml.kernel.org/r/20190801032812.25018-1-standby24x7@gmail.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf cpumap: Fix writing to illegal memory in handling cpumap mask
He Zhe [Fri, 2 Aug 2019 08:29:52 +0000 (16:29 +0800)]
perf cpumap: Fix writing to illegal memory in handling cpumap mask

cpu_map__snprint_mask() would write to illegal memory pointed by
zalloc(0) when there is only one cpu.

This patch fixes the calculation and adds sanity check against the input
parameters.

Signed-off-by: He Zhe <zhe.he@windriver.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Budankov <alexey.budankov@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Fixes: 4400ac8a9a90 ("perf cpumap: Introduce cpu_map__snprint_mask()")
Link: http://lkml.kernel.org/r/1564734592-15624-2-git-send-email-zhe.he@windriver.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf ftrace: Fix failure to set cpumask when only one cpu is present
He Zhe [Fri, 2 Aug 2019 08:29:51 +0000 (16:29 +0800)]
perf ftrace: Fix failure to set cpumask when only one cpu is present

The buffer containing the string used to set cpumask is overwritten at
the end of the string later in cpu_map__snprint_mask due to not enough
memory space, when there is only one cpu.

And thus causes the following failure:

  $ perf ftrace ls
  failed to reset ftrace
  $

This patch fixes the calculation of the cpumask string size.

Signed-off-by: He Zhe <zhe.he@windriver.com>
Tested-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Alexey Budankov <alexey.budankov@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Stephane Eranian <eranian@google.com>
Fixes: dc23103278c5 ("perf ftrace: Add support for -a and -C option")
Link: http://lkml.kernel.org/r/1564734592-15624-1-git-send-email-zhe.he@windriver.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf db-export: Fix thread__exec_comm()
Adrian Hunter [Thu, 8 Aug 2019 06:48:23 +0000 (09:48 +0300)]
perf db-export: Fix thread__exec_comm()

Threads synthesized from /proc have comms with a start time of zero, and
not marked as "exec". Currently, there can be 2 such comms. The first is
created by processing a synthesized fork event and is set to the
parent's comm string, and the second by processing a synthesized comm
event set to the thread's current comm string.

In the absence of an "exec" comm, thread__exec_comm() picks the last
(oldest) comm, which, in the case above, is the parent's comm string.
For a main thread, that is very probably wrong. Use the second-to-last
in that case.

This affects only db-export because it is the only user of
thread__exec_comm().

Example:

  $ sudo perf record -a -o pt-a-sleep-1 -e intel_pt//u -- sleep 1
  $ sudo chown ahunter pt-a-sleep-1

Before:

  $ perf script -i pt-a-sleep-1 --itrace=bep -s tools/perf/scripts/python/export-to-sqlite.py pt-a-sleep-1.db branches calls
  $ sqlite3 -header -column pt-a-sleep-1.db 'select * from comm_threads_view'
  comm_id     command     thread_id   pid         tid
  ----------  ----------  ----------  ----------  ----------
  1           swapper     1           0           0
  2           rcu_sched   2           10          10
  3           kthreadd    3           78          78
  5           sudo        4           15180       15180
  5           sudo        5           15180       15182
  7           kworker/4:  6           10335       10335
  8           kthreadd    7           55          55
  10          systemd     8           865         865
  10          systemd     9           865         875
  13          perf        10          15181       15181
  15          sleep       10          15181       15181
  16          kworker/3:  11          14179       14179
  17          kthreadd    12          29376       29376
  19          systemd     13          746         746
  21          systemd     14          401         401
  23          systemd     15          879         879
  23          systemd     16          879         945
  25          kthreadd    17          556         556
  27          kworker/u1  18          14136       14136
  28          kworker/u1  19          15021       15021
  29          kthreadd    20          509         509
  31          systemd     21          836         836
  31          systemd     22          836         967
  33          systemd     23          1148        1148
  33          systemd     24          1148        1163
  35          kworker/2:  25          17988       17988
  36          kworker/0:  26          13478       13478

After:

  $ perf script -i pt-a-sleep-1 --itrace=bep -s tools/perf/scripts/python/export-to-sqlite.py pt-a-sleep-1b.db branches calls
  $ sqlite3 -header -column pt-a-sleep-1b.db 'select * from comm_threads_view'
  comm_id     command     thread_id   pid         tid
  ----------  ----------  ----------  ----------  ----------
  1           swapper     1           0           0
  2           rcu_sched   2           10          10
  3           kswapd0     3           78          78
  4           perf        4           15180       15180
  4           perf        5           15180       15182
  6           kworker/4:  6           10335       10335
  7           kcompactd0  7           55          55
  8           accounts-d  8           865         865
  8           accounts-d  9           865         875
  10          perf        10          15181       15181
  12          sleep       10          15181       15181
  13          kworker/3:  11          14179       14179
  14          kworker/1:  12          29376       29376
  15          haveged     13          746         746
  16          systemd-jo  14          401         401
  17          NetworkMan  15          879         879
  17          NetworkMan  16          879         945
  19          irq/131-iw  17          556         556
  20          kworker/u1  18          14136       14136
  21          kworker/u1  19          15021       15021
  22          kworker/u1  20          509         509
  23          thermald    21          836         836
  23          thermald    22          836         967
  25          unity-sett  23          1148        1148
  25          unity-sett  24          1148        1163
  27          kworker/2:  25          17988       17988
  28          kworker/0:  26          13478       13478

Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: stable@vger.kernel.org
Fixes: 65de51f93ebf ("perf tools: Identify which comms are from exec")
Link: http://lkml.kernel.org/r/20190808064823.14846-1-adrian.hunter@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoperf annotate: Fix printing of unaugmented disassembled instructions from BPF
Arnaldo Carvalho de Melo [Tue, 6 Aug 2019 14:24:09 +0000 (11:24 -0300)]
perf annotate: Fix printing of unaugmented disassembled instructions from BPF

The code to disassemble BPF programs uses binutil's disassembling
routines, and those use in turn fprintf to print to a memstream FILE,
adding a newline at the end of each line, which ends up confusing the
TUI routines called from:

  annotate_browser__write()
    annotate_line__write()
      annotate_browser__printf()
        ui_browser__vprintf()
          SLsmg_vprintf()

The SLsmg_vprintf() function in the slang library gets confused with the
terminating newline, so make the disasm_line__parse() function that
parses the lines produced by the BPF specific disassembler (that uses
binutil's libopcodes) and the lines produced by the objdump based
disassembler used for everything else (and that doesn't adds this
terminating newline) trim the end of the line in addition of the
beginning.

This way when disasm_line->ops.raw, i.e. for instructions without a
special scnprintf() method, we'll not have that \n getting in the way of
filling the screen right after the instruction with spaces to avoid
leaving what was on the screen before and thus garbling the annotation
screen, breaking scrolling, etc.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Song Liu <songliubraving@fb.com>
Fixes: 6987561c9e86 ("perf annotate: Enable annotation of BPF programs")
Link: https://lkml.kernel.org/n/tip-unbr5a5efakobfr6rhxq99ta@git.kernel.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
10 months agoarm64: mm: add missing PTE_SPECIAL in pte_mkdevmap on arm64
Jia He [Wed, 7 Aug 2019 04:58:51 +0000 (12:58 +0800)]
arm64: mm: add missing PTE_SPECIAL in pte_mkdevmap on arm64

Without this patch, the MAP_SYNC test case will cause a print_bad_pte
warning on arm64 as follows:

[   25.542693] BUG: Bad page map in process mapdax333 pte:2e8000448800f53 pmd:41ff5f003
[   25.546360] page:ffff7e0010220000 refcount:1 mapcount:-1 mapping:ffff8003e29c7440 index:0x0
[   25.550281] ext4_dax_aops
[   25.550282] name:"__aaabbbcccddd__"
[   25.551553] flags: 0x3ffff0000001002(referenced|reserved)
[   25.555802] raw: 03ffff0000001002 ffff8003dfffa908 0000000000000000 ffff8003e29c7440
[   25.559446] raw: 0000000000000000 0000000000000000 00000001fffffffe 0000000000000000
[   25.563075] page dumped because: bad pte
[   25.564938] addr:0000ffffbe05b000 vm_flags:208000fb anon_vma:0000000000000000 mapping:ffff8003e29c7440 index:0
[   25.574272] file:__aaabbbcccddd__ fault:ext4_dax_fault mmmmap:ext4_file_mmap readpage:0x0
[   25.578799] CPU: 1 PID: 1180 Comm: mapdax333 Not tainted 5.2.0+ #21
[   25.581702] Hardware name: QEMU KVM Virtual Machine, BIOS 0.0.0 02/06/2015
[   25.585624] Call trace:
[   25.587008]  dump_backtrace+0x0/0x178
[   25.588799]  show_stack+0x24/0x30
[   25.590328]  dump_stack+0xa8/0xcc
[   25.591901]  print_bad_pte+0x18c/0x218
[   25.593628]  unmap_page_range+0x778/0xc00
[   25.595506]  unmap_single_vma+0x94/0xe8
[   25.597304]  unmap_vmas+0x90/0x108
[   25.598901]  unmap_region+0xc0/0x128
[   25.600566]  __do_munmap+0x284/0x3f0
[   25.602245]  __vm_munmap+0x78/0xe0
[   25.603820]  __arm64_sys_munmap+0x34/0x48
[   25.605709]  el0_svc_common.constprop.0+0x78/0x168
[   25.607956]  el0_svc_handler+0x34/0x90
[   25.609698]  el0_svc+0x8/0xc
[...]

The root cause is in _vm_normal_page, without the PTE_SPECIAL bit,
the return value will be incorrectly set to pfn_to_page(pfn) instead
of NULL. Besides, this patch also rewrite the pmd_mkdevmap to avoid
setting PTE_SPECIAL for pmd

The MAP_SYNC test case is as follows(Provided by Yibo Cai)
$#include <stdio.h>
$#include <string.h>
$#include <unistd.h>
$#include <sys/file.h>
$#include <sys/mman.h>

$#ifndef MAP_SYNC
$#define MAP_SYNC 0x80000
$#endif

/* mount -o dax /dev/pmem0 /mnt */
$#define F "/mnt/__aaabbbcccddd__"

int main(void)
{
    int fd;
    char buf[4096];
    void *addr;

    if ((fd = open(F, O_CREAT|O_TRUNC|O_RDWR, 0644)) < 0) {
        perror("open1");
        return 1;
    }

    if (write(fd, buf, 4096) != 4096) {
        perror("lseek");
        return 1;
    }

    addr = mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_SHARED|MAP_SYNC, fd, 0);
    if (addr == MAP_FAILED) {
        perror("mmap");
        printf("did you mount with '-o dax'?\n");
        return 1;
    }

    memset(addr, 0x55, 4096);

    if (munmap(addr, 4096) == -1) {
        perror("munmap");
        return 1;
    }

    close(fd);

    return 0;
}

Fixes: 73b20c84d42d ("arm64: mm: implement pte_devmap support")
Reported-by: Yibo Cai <Yibo.Cai@arm.com>
Acked-by: Will Deacon <will@kernel.org>
Acked-by: Robin Murphy <Robin.Murphy@arm.com>
Signed-off-by: Jia He <justin.he@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>