epoll: convert internal api to timespec64

Patch series "add epoll_pwait2 syscall", v4.

Enable nanosecond timeouts for epoll.

Analogous to pselect and ppoll, introduce an epoll_wait syscall
variant that takes a struct timespec instead of int timeout.

This patch (of 4):

Make epoll more consistent with select/poll: pass along the timeout as
timespec64 pointer.

In anticipation of additional changes affecting all three polling
mechanisms:

- add epoll_pwait2 syscall with timespec semantics,
  and share poll_select_set_timeout implementation.
- compute slack before conversion to absolute time,
  to save one ktime_get_ts64 call.

Link: https://lkml.kernel.org/r/20201121144401.3727659-1-willemdebruijn.kernel@gmail.com
Link: https://lkml.kernel.org/r/20201121144401.3727659-2-willemdebruijn.kernel@gmail.com
Signed-off-by: Willem de Bruijn <willemb@google.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: eliminate unnecessary lock for zero timeout

We call ep_events_available() under lock when timeout is 0, and then call
it without locks in the loop for the other cases.

Instead, call ep_events_available() without lock for all cases.  For
non-zero timeouts, we will recheck after adding the thread to the wait
queue.  For zero timeout cases, by definition, user is opportunistically
polling and will have to call epoll_wait again in the future.

Note that this lock was kept in c5a282e9635e9 because the whole loop was
historically under lock.

This patch results in a 1% CPU/RPC reduction in RPC benchmarks.

Link: https://lkml.kernel.org/r/20201106231635.3528496-9-soheil.kdev@gmail.com
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Suggested-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Cc: Guantao Liu <guantaol@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: replace gotos with a proper loop

The existing loop is pointless, and the labels make it really hard to
follow the structure.

Replace that control structure with a simple loop that returns when there
are new events, there is a signal, or the thread has timed out.

Link: https://lkml.kernel.org/r/20201106231635.3528496-8-soheil.kdev@gmail.com
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Cc: Guantao Liu <guantaol@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: pull all code between fetch_events and send_event into the loop

This is a no-op change which simplifies the follow up patches.

Link: https://lkml.kernel.org/r/20201106231635.3528496-7-soheil.kdev@gmail.com
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Cc: Guantao Liu <guantaol@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: simplify and optimize busy loop logic

ep_events_available() is called multiple times around the busy loop logic,
even though the logic is generally not used.  ep_reset_busy_poll_napi_id()
is similarly always called, even when busy loop is not used.

Eliminate ep_reset_busy_poll_napi_id() and inline it inside
ep_busy_loop().  Make ep_busy_loop() return whether there are any events
available after the busy loop.  This will eliminate unnecessary loads and
branches, and simplifies the loop.

Link: https://lkml.kernel.org/r/20201106231635.3528496-6-soheil.kdev@gmail.com
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Cc: Guantao Liu <guantaol@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: move eavail next to the list_empty_careful check

This is a no-op change and simply to make the code more coherent.

Link: https://lkml.kernel.org/r/20201106231635.3528496-5-soheil.kdev@gmail.com
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Cc: Guantao Liu <guantaol@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: pull fatal signal checks into ep_send_events()

To simplify the code, pull in checking the fatal signals into
ep_send_events().  ep_send_events() is called only from ep_poll().

Note that, previously, we were always checking fatal events, but it is
checked only if eavail is true.  This should be fine because the goal of
that check is to quickly return from epoll_wait() when there is a pending
fatal signal.

Link: https://lkml.kernel.org/r/20201106231635.3528496-4-soheil.kdev@gmail.com
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Suggested-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Cc: Guantao Liu <guantaol@google.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: simplify signal handling

Check signals before locking ep->lock, and immediately return -EINTR if
there is any signal pending.

This saves a few loads, stores, and branches from the hot path and
simplifies the loop structure for follow up patches.

Link: https://lkml.kernel.org/r/20201106231635.3528496-3-soheil.kdev@gmail.com
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Cc: Guantao Liu <guantaol@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

epoll: check for events when removing a timed out thread from the wait queue

Patch series "simplify ep_poll".

This patch series is a followup based on the suggestions and feedback by
Linus:
https://lkml.kernel.org/r/CAHk-=wizk=OxUyQPbO8MS41w2Pag1kniUV5WdD5qWL-gq1kjDA@mail.gmail.com

The first patch in the series is a fix for the epoll race in presence of
timeouts, so that it can be cleanly backported to all affected stable
kernels.

The rest of the patch series simplify the ep_poll() implementation.  Some
of these simplifications result in minor performance enhancements as well.
We have kept these changes under self tests and internal benchmarks for a
few days, and there are minor (1-2%) performance enhancements as a result.

This patch (of 8):

After abc610e01c66 ("fs/epoll: avoid barrier after an epoll_wait(2)
timeout"), we break out of the ep_poll loop upon timeout, without checking
whether there is any new events available.  Prior to that patch-series we
always called ep_events_available() after exiting the loop.

This can cause races and missed wakeups.  For example, consider the
following scenario reported by Guantao Liu:

Suppose we have an eventfd added using EPOLLET to an epollfd.

Thread 1: Sleeps for just below 5ms and then writes to an eventfd.
Thread 2: Calls epoll_wait with a timeout of 5 ms. If it sees an
          event of the eventfd, it will write back on that fd.
Thread 3: Calls epoll_wait with a negative timeout.

Prior to abc610e01c66, it is guaranteed that Thread 3 will wake up either
by Thread 1 or Thread 2.  After abc610e01c66, Thread 3 can be blocked
indefinitely if Thread 2 sees a timeout right before the write to the
eventfd by Thread 1.  Thread 2 will be woken up from
schedule_hrtimeout_range and, with evail 0, it will not call
ep_send_events().

To fix this issue:
1) Simplify the timed_out case as suggested by Linus.
2) while holding the lock, recheck whether the thread was woken up
   after its time out has reached.

Note that (2) is different from Linus' original suggestion: It do not set
"eavail = ep_events_available(ep)" to avoid unnecessary contention (when
there are too many timed-out threads and a small number of events), as
well as races mentioned in the discussion thread.

This is the first patch in the series so that the backport to stable
releases is straightforward.

Link: https://lkml.kernel.org/r/20201106231635.3528496-1-soheil.kdev@gmail.com
Link: https://lkml.kernel.org/r/CAHk-=wizk=OxUyQPbO8MS41w2Pag1kniUV5WdD5qWL-gq1kjDA@mail.gmail.com
Link: https://lkml.kernel.org/r/20201106231635.3528496-2-soheil.kdev@gmail.com
Fixes: abc610e01c66 ("fs/epoll: avoid barrier after an epoll_wait(2) timeout")
Signed-off-by: Soheil Hassas Yeganeh <soheil@google.com>
Tested-by: Guantao Liu <guantaol@google.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Reported-by: Guantao Liu <guantaol@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Reviewed-by: Khazhismel Kumykov <khazhy@google.com>
Reviewed-by: Davidlohr Bueso <dbueso@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/memcontrol:rewrite mem_cgroup_page_lruvec()

mem_cgroup_page_lruvec() in memcontrol.c and mem_cgroup_lruvec() in
memcontrol.h is very similar except for the param(page and memcg) which
also can be convert to each other.

So rewrite mem_cgroup_page_lruvec() with mem_cgroup_lruvec().

[alex.shi@linux.alibaba.com: add missed warning in mem_cgroup_lruvec]
Link: https://lkml.kernel.org/r/94f17bb7-ec61-5b72-3555-fabeb5a4d73b@linux.alibaba.com
[lstoakes@gmail.com: warn on missing memcg on mem_cgroup_page_lruvec()]
Link: https://lkml.kernel.org/r/20201125112202.387009-1-lstoakes@gmail.com
Link: https://lkml.kernel.org/r/20201108143731.GA74138@rlk
Signed-off-by: Hui Su <sh_def@163.com>
Signed-off-by: Alex Shi <alex.shi@linux.alibaba.com>
Signed-off-by: Lorenzo Stoakes <lstoakes@gmail.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Roman Gushchin <guro@fb.com>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Cc: Yafang Shao <laoar.shao@gmail.com>
Cc: Chris Down <chris@chrisdown.name>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm, kvm: account kvm_vcpu_mmap to kmemcg

A VCPU of a VM can allocate couple of pages which can be mmap'ed by the
user space application. At the moment this memory is not charged to the
memcg of the VMM. On a large machine running large number of VMs or
small number of VMs having large number of VCPUs, this unaccounted
memory can be very significant. So, charge this memory to the memcg of
the VMM. Please note that lifetime of these allocations corresponds to
the lifetime of the VMM.

Link: https://lkml.kernel.org/r/20201106202923.2087414-1-shakeelb@google.com
Signed-off-by: Shakeel Butt <shakeelb@google.com>
Acked-by: Roman Gushchin <guro@fb.com>
Acked-by: Paolo Bonzini <pbonzini@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/memcg: remove unused definitions

Some definitions are left unused, just clean them.

Link: https://lkml.kernel.org/r/20201108003834.12669-1-richard.weiyang@gmail.com
Signed-off-by: Wei Yang <richard.weiyang@gmail.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Reviewed-by: Roman Gushchin <guro@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/memcg: warning on !memcg after readahead page charged

Add VM_WARN_ON_ONCE_PAGE() macro.

Since readahead page is charged on memcg too, in theory we don't have to
check this exception now.  Before safely remove them all, add a warning
for the unexpected !memcg.

Link: https://lkml.kernel.org/r/1604283436-18880-3-git-send-email-alex.shi@linux.alibaba.com
Signed-off-by: Alex Shi <alex.shi@linux.alibaba.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Hugh Dickins <hughd@google.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/memcg: bail early from swap accounting if memcg disabled

Patch series "bail out early for memcg disable".

These 2 patches are indepenedent from per memcg lru lock, and may
encounter unexpected warning, so let's move out them from per memcg
lru locking patchset.

This patch (of 2):

We could bail out early when memcg wasn't enabled.

Link: https://lkml.kernel.org/r/1604283436-18880-1-git-send-email-alex.shi@linux.alibaba.com
Link: https://lkml.kernel.org/r/1604283436-18880-2-git-send-email-alex.shi@linux.alibaba.com
Signed-off-by: Alex Shi <alex.shi@linux.alibaba.com>
Reviewed-by: Roman Gushchin <guro@fb.com>
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Hugh Dickins <hughd@google.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

selftests/core: add regression test for CLOSE_RANGE_UNSHARE | CLOSE_RANGE_CLOEXEC

This test is a minimalized version of the reproducer given by syzbot
(cf. [1]).

After introducing CLOSE_RANGE_CLOEXEC syzbot reported a crash when
CLOSE_RANGE_CLOEXEC is specified in conjunction with
CLOSE_RANGE_UNSHARE. When CLOSE_RANGE_UNSHARE is specified the caller
will receive a private file descriptor table in case their file
descriptor table is currently shared.
For the case where the caller has requested all file descriptors to be
actually closed via e.g. close_range(3, ~0U, 0) the kernel knows that
the caller does not need any of the file descriptors anymore and will
optimize the close operation by only copying all files in the range from
0 to 3 and no others.

However, if the caller requested CLOSE_RANGE_CLOEXEC together with
CLOSE_RANGE_UNSHARE the caller wants to still make use of the file
descriptors so the kernel needs to copy all of them and can't optimize.

The original patch didn't account for this and thus could cause oopses
as evidenced by the syzbot report. Add tests for this regression.

We first create a huge gap in the fd table. When we now call
CLOSE_RANGE_UNSHARE with a shared fd table and and with ~0U as upper
bound the kernel will only copy up to fd1 file descriptors into the new
fd table. If the kernel is buggy and doesn't handle CLOSE_RANGE_CLOEXEC
correctly it will not have copied all file descriptors and we will oops!

This test passes on a fixed kernel and will trigger an oops on a buggy
kernel.

[1]: https://syzkaller.appspot.com/text?tag=KernelConfig&x=db720fe37a6a41d8

Cc: Giuseppe Scrivano <gscrivan@redhat.com>
Cc: linux-fsdevel@vger.kernel.org
Link: syzbot+96cfd2b22b3213646a93@syzkaller.appspotmail.com
Link: https://lore.kernel.org/r/20201218145415.801063-4-christian.brauner@ubuntu.com
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

selftests/core: add test for CLOSE_RANGE_UNSHARE | CLOSE_RANGE_CLOEXEC

Add a test to verify that CLOSE_RANGE_UNSHARE works correctly when combined
with CLOSE_RANGE_CLOEXEC for the single-threaded case.

Cc: Giuseppe Scrivano <gscrivan@redhat.com>
Cc: linux-fsdevel@vger.kernel.org
Link: https://lore.kernel.org/r/20201218145415.801063-3-christian.brauner@ubuntu.com
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

selftests/core: handle missing syscall number for close_range

This improves the syscall number handling in the close_range()
selftests. This should handle any architecture.

Cc: Giuseppe Scrivano <gscrivan@redhat.com>
Cc: linux-fsdevel@vger.kernel.org
Link: https://lore.kernel.org/r/20201218145415.801063-2-christian.brauner@ubuntu.com
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

selftests/core: fix close_range_test build after XFAIL removal

XFAIL was removed in commit 9847d24af95c ("selftests/harness: Refactor
XFAIL into SKIP") and its use in close_range_test was already replaced
by commit 1d44d0dd61b6 ("selftests: core: use SKIP instead of XFAIL in
close_range_test.c"). However, commit 23afeaeff3d9 ("selftests: core:
add tests for CLOSE_RANGE_CLOEXEC") introduced usage of XFAIL in
TEST(close_range_cloexec). Use SKIP there as well.

Fixes: 23afeaeff3d9 ("selftests: core: add tests for CLOSE_RANGE_CLOEXEC")
Cc: Giuseppe Scrivano <gscrivan@redhat.com>
Cc: linux-fsdevel@vger.kernel.org
Signed-off-by: Tobias Klauser <tklauser@distanz.ch>
Acked-by: Christian Brauner <christian.brauner@ubuntu.com>
Link: https://lore.kernel.org/r/20201218112428.13662-1-tklauser@distanz.ch
Link: https://lore.kernel.org/r/20201218145415.801063-1-christian.brauner@ubuntu.com
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

close_range: unshare all fds for CLOSE_RANGE_UNSHARE | CLOSE_RANGE_CLOEXEC

After introducing CLOSE_RANGE_CLOEXEC syzbot reported a crash when
CLOSE_RANGE_CLOEXEC is specified in conjunction with CLOSE_RANGE_UNSHARE.
When CLOSE_RANGE_UNSHARE is specified the caller will receive a private
file descriptor table in case their file descriptor table is currently
shared.

For the case where the caller has requested all file descriptors to be
actually closed via e.g. close_range(3, ~0U, 0) the kernel knows that
the caller does not need any of the file descriptors anymore and will
optimize the close operation by only copying all files in the range from
0 to 3 and no others.

However, if the caller requested CLOSE_RANGE_CLOEXEC together with
CLOSE_RANGE_UNSHARE the caller wants to still make use of the file
descriptors so the kernel needs to copy all of them and can't optimize.

The original patch didn't account for this and thus could cause oopses
as evidenced by the syzbot report because it assumed that all fds had
been copied. Fix this by handling the CLOSE_RANGE_CLOEXEC case.

syzbot reported
==================================================================
BUG: KASAN: null-ptr-deref in instrument_atomic_read include/linux/instrumented.h:71 [inline]
BUG: KASAN: null-ptr-deref in atomic64_read include/asm-generic/atomic-instrumented.h:837 [inline]
BUG: KASAN: null-ptr-deref in atomic_long_read include/asm-generic/atomic-long.h:29 [inline]
BUG: KASAN: null-ptr-deref in filp_close+0x22/0x170 fs/open.c:1274
Read of size 8 at addr 0000000000000077 by task syz-executor511/8522

CPU: 1 PID: 8522 Comm: syz-executor511 Not tainted 5.10.0-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Call Trace:
 __dump_stack lib/dump_stack.c:79 [inline]
 dump_stack+0x107/0x163 lib/dump_stack.c:120
 __kasan_report mm/kasan/report.c:549 [inline]
 kasan_report.cold+0x5/0x37 mm/kasan/report.c:562
 check_memory_region_inline mm/kasan/generic.c:186 [inline]
 check_memory_region+0x13d/0x180 mm/kasan/generic.c:192
 instrument_atomic_read include/linux/instrumented.h:71 [inline]
 atomic64_read include/asm-generic/atomic-instrumented.h:837 [inline]
 atomic_long_read include/asm-generic/atomic-long.h:29 [inline]
 filp_close+0x22/0x170 fs/open.c:1274
 close_files fs/file.c:402 [inline]
 put_files_struct fs/file.c:417 [inline]
 put_files_struct+0x1cc/0x350 fs/file.c:414
 exit_files+0x12a/0x170 fs/file.c:435
 do_exit+0xb4f/0x2a00 kernel/exit.c:818
 do_group_exit+0x125/0x310 kernel/exit.c:920
 get_signal+0x428/0x2100 kernel/signal.c:2792
 arch_do_signal_or_restart+0x2a8/0x1eb0 arch/x86/kernel/signal.c:811
 handle_signal_work kernel/entry/common.c:147 [inline]
 exit_to_user_mode_loop kernel/entry/common.c:171 [inline]
 exit_to_user_mode_prepare+0x124/0x200 kernel/entry/common.c:201
 __syscall_exit_to_user_mode_work kernel/entry/common.c:291 [inline]
 syscall_exit_to_user_mode+0x19/0x50 kernel/entry/common.c:302
 entry_SYSCALL_64_after_hwframe+0x44/0xa9
RIP: 0033:0x447039
Code: Unable to access opcode bytes at RIP 0x44700f.
RSP: 002b:00007f1b1225cdb8 EFLAGS: 00000246 ORIG_RAX: 00000000000000ca
RAX: 0000000000000001 RBX: 00000000006dbc28 RCX: 0000000000447039
RDX: 00000000000f4240 RSI: 0000000000000081 RDI: 00000000006dbc2c
RBP: 00000000006dbc20 R08: 0000000000000000 R09: 0000000000000000
R10: 0000000000000000 R11: 0000000000000246 R12: 00000000006dbc2c
R13: 00007fff223b6bef R14: 00007f1b1225d9c0 R15: 00000000006dbc2c
==================================================================

syzbot has tested the proposed patch and the reproducer did not trigger any issue:

Reported-and-tested-by: syzbot+96cfd2b22b3213646a93@syzkaller.appspotmail.com
Tested on:

commit:         10f7cddd selftests/core: add regression test for CLOSE_RAN..
git tree:       git://git.kernel.org/pub/scm/linux/kernel/git/brauner/linux.git vfs
kernel config:  https://syzkaller.appspot.com/x/.config?x=5d42216b510180e3
link: https://syzkaller.appspot.com/bug?extid=96cfd2b22b3213646a93
compiler:       gcc (GCC) 10.1.0-syz 20200507

Reported-by: syzbot+96cfd2b22b3213646a93@syzkaller.appspotmail.com
Fixes: 582f1fb6b721 ("fs, close_range: add flag CLOSE_RANGE_CLOEXEC")
Cc: Giuseppe Scrivano <gscrivan@redhat.com>
Cc: linux-fsdevel@vger.kernel.org
Link: https://lore.kernel.org/r/20201217213303.722643-1-christian.brauner@ubuntu.com
Signed-off-by: Christian Brauner <christian.brauner@ubuntu.com>

xen: Kconfig: remove X86_64 depends from XEN_512GB

commit bfda93aee0ec ("xen: Kconfig: nest Xen guest options")
accidentally re-added X86_64 as a dependency to XEN_512GB.  It was
originally removed in commit a13f2ef168cb ("x86/xen: remove 32-bit Xen
PV guest support").  Remove it again.

Fixes: bfda93aee0ec ("xen: Kconfig: nest Xen guest options")
Reported-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Signed-off-by: Jason Andryuk <jandryuk@gmail.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/20201216140838.16085-1-jandryuk@gmail.com
Signed-off-by: Juergen Gross <jgross@suse.com>

Add SMB 2 support for getting and setting SACLs

Fix passing of the additional security info via version
operations. Force new open when getting SACL and avoid
reuse of files that were previously open without
sufficient privileges to access SACLs.

Signed-off-by: Boris Protopopov <pboris@amazon.com>
Reviewed-by: Shyam Prasad N <sprasad@microsoft.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

rtc: pcf2127: only use watchdog when explicitly available

Most boards using the pcf2127 chip (in my bubble) don't make use of the
watchdog functionality and the respective output is not connected. The
effect on such a board is that there is a watchdog device provided that
doesn't work.

So only register the watchdog if the device tree has a "reset-source"
property.

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
[RV: s/has-watchdog/reset-source/]
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/20201218101054.25416-3-rasmus.villemoes@prevas.dk

dt-bindings: rtc: add reset-source property

Some RTCs, e.g. the pcf2127, can be used as a hardware watchdog. But
if the reset pin is not actually wired up, the driver exposes a
watchdog device that doesn't actually work.

Provide a standard binding that can be used to indicate that a given
RTC can perform a reset of the machine, similar to wakeup-source.

Suggested-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Link: https://lore.kernel.org/r/20201218101054.25416-2-rasmus.villemoes@prevas.dk

mm/filemap: fix infinite loop in generic_file_buffered_read()

If iter->count is 0 and iocb->ki_pos is page aligned, this causes
nr_pages to be 0.

Then in generic_file_buffered_read_get_pages() find_get_pages_contig()
returns 0 - because we asked for 0 pages, so we call
generic_file_buffered_read_no_cached_page() which attempts to add a page
to the page cache, which fails with -EEXIST, and then we loop. Oops...

Signed-off-by: Kent Overstreet <kent.overstreet@gmail.com>
Reported-by: Jens Axboe <axboe@kernel.dk>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

dt-bindings/display: abt,y030xx067a: Fix binding

The binding should use "unevaluatedProperties" instead of
"additionalProperties", since it is a SPI device and may have
SPI-related Device Tree properties, for instance the "spi-max-frequency"
property that is present in the example.

Fixes: e366a644c69d ("dt-bindings: display: Add ABT Y030XX067A panel bindings")
Signed-off-by: Paul Cercueil <paul@crapouillou.net>
Reviewed-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://lore.kernel.org/r/20201217005945.335111-1-paul%40crapouillou.net
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: clock: imx8qxp-lpcg: eliminate yamllint warnings

Eliminate the following yamllint warnings:
./Documentation/devicetree/bindings/clock/imx8qxp-lpcg.yaml
:32:13:[warning] wrong indentation: expected 14 but found 12 (indentation)
:35:9: [warning] wrong indentation: expected 10 but found 8 (indentation)

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Reviewed-by: Stephen Boyd <sboyd@kernel.org>
Link: https://lore.kernel.org/r/20201207045527.1607-2-thunder.leizhen@huawei.com
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: display: eliminate yamllint warnings

Eliminate the following yamllint warnings:
./Documentation/devicetree/bindings/display/bridge/analogix,anx7625.yaml
:52:9: [warning] wrong indentation: expected 6 but found 8 (indentation)

./Documentation/devicetree/bindings/display/bridge/intel,keembay-dsi.yaml
:42:8: [warning] wrong indentation: expected 8 but found 7 (indentation)
:45:8: [warning] wrong indentation: expected 8 but found 7 (indentation)

./Documentation/devicetree/bindings/display/intel,keembay-msscam.yaml
:21:6: [warning] wrong indentation: expected 6 but found 5 (indentation)

./Documentation/devicetree/bindings/display/panel/novatek,nt36672a.yaml
:25:10: [warning] wrong indentation: expected 10 but found 9 (indentation)

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20201207044830.1551-2-thunder.leizhen@huawei.com
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: media: nokia,smia: eliminate yamllint warnings

Eliminate the following yamllint warnings:
./Documentation/devicetree/bindings/media/i2c/mipi-ccs.yaml
:4:1: [error] missing document start "---" (document-start)
:29:9: [warning] wrong indentation: expected 10 but found 8 (indentation)
:32:9: [warning] wrong indentation: expected 10 but found 8 (indentation)

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20201207042400.1498-3-thunder.leizhen@huawei.com
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: devapc: add the required property 'additionalProperties'

When I do dt_binding_check for any YAML file, below wanring is always
reported:

xxx/soc/mediatek/devapc.yaml: 'additionalProperties' is a required property
xxx/soc/mediatek/devapc.yaml: ignoring, error in schema:
warning: no schema found in file: xxx/soc/mediatek/devapc.yaml

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20201204093813.1275-5-thunder.leizhen@huawei.com
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: soc: add the required property 'additionalProperties'

When I do dt_binding_check for any YAML file, below wanring is always
reported:

xxx/soc/litex/litex,soc-controller.yaml: 'additionalProperties' is a required property
xxx/soc/litex/litex,soc-controller.yaml: ignoring, error in schema:
warning: no schema found in file: xxx/soc/litex/litex,soc-controller.yaml

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20201204093813.1275-4-thunder.leizhen@huawei.com
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: serial: add the required property 'additionalProperties'

When I do dt_binding_check for any YAML file, below wanring is always
reported:

xxx/serial/litex,liteuart.yaml: 'additionalProperties' is a required property
xxx/serial/litex,liteuart.yaml: ignoring, error in schema:
warning: no schema found in file: xxx/serial/litex,liteuart.yaml

Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20201204093813.1275-3-thunder.leizhen@huawei.com
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: xlnx,vcu-settings: fix dt_binding_check warnings

When running make dt_binding_check, the xlnx,vcu-settings binding
triggers the following two warnings:

'additionalProperties' is a required property

example-0: vcu@a0041000:reg:0: [0, 2684620800, 0, 4096] is too long

Fix the binding and make the checker happy.

Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20201202090522.251607-1-m.tretter@pengutronix.de
Signed-off-by: Rob Herring <robh@kernel.org>

media: dt-bindings: coda: Add missing 'additionalProperties'

'additionalProperties' is now required by the meta-schema. Add it for
coda. As a result, 'interrupts', 'interrupt-names' and 'power-domains'
need to be reworked to be defined at the top level.

Cc: Philipp Zabel <p.zabel@pengutronix.de>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: linux-media@vger.kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Link: https://lore.kernel.org/r/20201117200752.4004368-1-robh@kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>

dt-bindings: Fix JSON pointers

The correct syntax for JSON pointers begins with a '/' after the '#'.
Without a '/', the string should be interpreted as a subschema
identifier. The jsonschema module currently doesn't handle subschema
identifiers and incorrectly allows JSON pointers to begin without a '/'.
Let's fix this before it becomes a problem when jsonschema module is
fixed.

Converted with:
perl -p -i -e 's/yaml#definitions/yaml#\/definitions/g' `find Documentation/devicetree/bindings/ -name "*.yaml"`

Cc: Maxime Ripard <mripard@kernel.org>
Cc: Jonathan Cameron <jic23@kernel.org>
Cc: Lars-Peter Clausen <lars@metafoo.de>
Cc: Daniel Thompson <daniel.thompson@linaro.org>
Cc: Jingoo Han <jingoohan1@gmail.com>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Mark Brown <broonie@kernel.org>
Cc: netdev@vger.kernel.org
Acked-By: Vinod Koul <vkoul@kernel.org>
Acked-by: Lee Jones <lee.jones@linaro.org>
Acked-by: Guenter Roeck <linux@roeck-us.net>
Acked-by: Sebastian Reichel <sre@kernel.org>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Acked-by: Mark Brown <broonie@kernel.org>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20201217223429.354283-1-robh@kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>

vdpa: Use simpler version of ida allocation

vdpa doesn't have any specific need to define start and end range of the
device index.
Hence use the simper version of the ida allocator.

Signed-off-by: Parav Pandit <parav@nvidia.com>
Reviewed-by: Eli Cohen <elic@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/20201112064005.349268-3-parav@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: Add missing comment for virtqueue count

Add missing comment for number of virtqueue.

Signed-off-by: Parav Pandit <parav@nvidia.com>
Reviewed-by: Eli Cohen <elic@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/20201112064005.349268-2-parav@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

uapi: virtio_ids: add missing device type IDs from OASIS spec

The OASIS virtio spec (1.1) defines several IDs that aren't reflected
in the header yet. Fixing this by adding the missing IDs, even though
they're not yet used by the kernel yet.

Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Link: https://lore.kernel.org/r/20201202111931.31953-2-info@metux.net
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

uapi: virtio_ids.h: consistent indentions

Fixing the differing indentions to be consistent and properly aligned.

Signed-off-by: Enrico Weigelt, metux IT consult <info@metux.net>
Link: https://lore.kernel.org/r/20201202111931.31953-1-info@metux.net
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vhost scsi: fix error return code in vhost_scsi_set_endpoint()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 25b98b64e284 ("vhost scsi: alloc cmds per vq instead of session")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1607071411-33484-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio_ring: Fix two use after free bugs

The "vq" struct is added to the "vdev->vqs" list prematurely.  If we
encounter an error later in the function then the "vq" is freed, but
since it is still on the list that could lead to a use after free bug.

Fixes: cbeedb72b97a ("virtio_ring: allocate desc state for split ring separately")
Reported-by: Robert Buhren <robert.buhren@sect.tu-berlin.de>
Reported-by: Felicitas Hetzelt <file@sect.tu-berlin.de>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/X8pGaG/zkI3jk8mk@mwanda
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>

virtio_net: Fix error code in probe()

Set a negative error code intead of returning success if the MTU has
been changed to something invalid.

Fixes: fe36cbe0671e ("virtio_net: clear MTU when out of range")
Reported-by: Robert Buhren <robert.buhren@sect.tu-berlin.de>
Reported-by: Felicitas Hetzelt <file@sect.tu-berlin.de>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/X8pGVJSeeCdII1Ys@mwanda
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>

virtio_ring: Cut and paste bugs in vring_create_virtqueue_packed()

There is a copy and paste bug in the error handling of this code and
it uses "ring_dma_addr" three times instead of "device_event_dma_addr"
and "driver_event_dma_addr".

Fixes: 1ce9e6055fa0 (" virtio_ring: introduce packed ring support")
Reported-by: Robert Buhren <robert.buhren@sect.tu-berlin.de>
Reported-by: Felicitas Hetzelt <file@sect.tu-berlin.de>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/X8pGRJlEzyn+04u2@mwanda
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>

tools/virtio: add barrier for aarch64

Add barrier for aarch64 for cross compiling, and  most are from Linux Kernel.

Signed-off-by: Peng Fan <peng.fan@nxp.com>
Link: https://lore.kernel.org/r/20201209084205.24062-4-peng.fan@oss.nxp.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tools/virtio: add krealloc_array

krealloc_array is used in drivers/vhost/vringh.c, add it to avoid build
failure.

Drop WARN_ON_ONCE, because duplicated with the one in bug.h

Signed-off-by: Peng Fan <peng.fan@nxp.com>
Link: https://lore.kernel.org/r/20201209084205.24062-3-peng.fan@oss.nxp.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

tools/virtio: include asm/bug.h

WARN_ON is used in drivers/vhost/vringh.c, to avoid build failure,
need include asm/bug.h

Signed-off-by: Peng Fan <peng.fan@nxp.com>
Link: https://lore.kernel.org/r/20201209084205.24062-2-peng.fan@oss.nxp.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa/mlx5: Use write memory barrier after updating CQ index

Make sure to put dma write memory barrier after updating CQ consumer
index so the hardware knows that there are available CQE slots in the
queue.

Failure to do this can cause the update of the RX doorbell record to get
updated before the CQ consumer index resulting in CQ overrun.

Fixes: 1a86b377aa21 ("vdpa/mlx5: Add VDPA driver for supported mlx5 devices")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Link: https://lore.kernel.org/r/20201209140004.15892-1-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: split vdpasim to core and net modules

Introduce new vdpa_sim_net and vdpa_sim (core) drivers. This is a
preparation for adding a vdpa simulator module for block devices.

Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
[sgarzare: various cleanups/fixes]
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-19-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: split vdpasim_virtqueue's iov field in out_iov and in_iov

vringh_getdesc_iotlb() manages 2 iovs for writable and readable
descriptors. This is very useful for the block device, where for
each request we have both types of descriptor.

Let's split the vdpasim_virtqueue's iov field in out_iov and
in_iov to use them with vringh_getdesc_iotlb().

We are using VIRTIO terminology for "out" (readable by the device)
and "in" (writable by the device) descriptors.

Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-18-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: make vdpasim->buffer size configurable

Allow each device to specify the size of the buffer allocated
in vdpa_sim.

Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-17-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: use kvmalloc to allocate vdpasim->buffer

The next patch will make the buffer size configurable from each
device.
Since the buffer could be larger than a page, we use kvmalloc()
instead of kmalloc().

Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-16-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: set vringh notify callback

Instead of calling the vq callback directly, we can leverage the
vringh_notify() function, adding vdpasim_vq_notify() and setting it
in the vringh notify callback.

Suggested-by: Jason Wang <jasowang@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-15-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: add set_config callback in vdpasim_dev_attr

The set_config callback can be used by the device to parse the
config structure modified by the driver.

The callback will be invoked, if set, in vdpasim_set_config() after
copying bytes from caller buffer into vdpasim->config buffer.

Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-14-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: add get_config callback in vdpasim_dev_attr

The get_config callback can be used by the device to fill the
config structure.
The callback will be invoked in vdpasim_get_config() before copying
bytes into caller buffer.

Move vDPA-net config updates from vdpasim_set_features() in the
new vdpasim_net_get_config() callback.
This is safe since in vdpa_get_config() we already check that
.set_features() callback is called before .get_config().

Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-13-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: make 'config' generic and usable for any device type

Add new 'config_size' attribute in 'vdpasim_dev_attr' and allocates
'config' dynamically to support any device types.

Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-12-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: store parsed MAC address in a buffer

As preparation for the next patches, we store the MAC address,
parsed during the vdpasim_create(), in a buffer that will be used
to fill 'config' together with other configurations.

Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-11-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: add work_fn in vdpasim_dev_attr

Rename vdpasim_work() in vdpasim_net_work() and add it to
the vdpasim_dev_attr structure.

Co-developed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-10-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: add supported_features field in vdpasim_dev_attr

Introduce a new VDPASIM_FEATURES macro with the generic features
supported by the vDPA simulator, and VDPASIM_NET_FEATURES macro with
vDPA-net features.

Add 'supported_features' field in vdpasim_dev_attr, to allow devices
to specify their features.

Co-developed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-9-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: add device id field in vdpasim_dev_attr

Remove VDPASIM_DEVICE_ID macro and add 'id' field in vdpasim_dev_attr,
that will be returned by vdpasim_get_device_id().

Use VIRTIO_ID_NET for vDPA-net simulator device id.

Co-developed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-8-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: add struct vdpasim_dev_attr for device attributes

vdpasim_dev_attr will contain device specific attributes. We starting
moving the number of virtqueues (i.e. nvqs) to vdpasim_dev_attr.

vdpasim_create() creates a new vDPA simulator following the device
attributes defined in the vdpasim_dev_attr parameter.

Co-developed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-7-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: rename vdpasim_config_ops variables

These variables store generic callbacks used by the vDPA simulator
core, so we can remove the 'net' word in their names.

Co-developed-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-6-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: make IOTLB entries limit configurable

Some devices may require a higher limit for the number of IOTLB
entries, so let's make it configurable through a module parameter.

By default, it's initialized with the current limit (2048).

Suggested-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-5-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: remove hard-coded virtq count

Add a new attribute that will define the number of virt queues to be
created for the vdpasim device.

Signed-off-by: Max Gurtovoy <mgurtovoy@nvidia.com>
[sgarzare: replace kmalloc_array() with kcalloc()]
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-4-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa_sim: remove unnecessary headers inclusion

Some headers are not necessary, so let's remove them to do
some cleaning.

Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-3-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: remove unnecessary 'default n' in Kconfig entries

'default n' is not necessary since it is already the default when
nothing is specified.

Suggested-by: Jason Wang <jasowang@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201215144256.155342-2-sgarzare@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

vdpa: ifcvf: Use dma_set_mask_and_coherent to simplify code

'pci_set_dma_mask()' + 'pci_set_consistent_dma_mask()' can be replaced by
an equivalent 'dma_set_mask_and_coherent()' which is much less verbose.

While at it, fix a typo (s/confiugration/configuration)

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/20201129125434.1462638-1-christophe.jaillet@wanadoo.fr
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>

vhost_vdpa: switch to vmemdup_user()

Replace opencoded alloc and copy with vmemdup_user()

Signed-off-by: Tian Tao <tiantao6@hisilicon.com>
Link: https://lore.kernel.org/r/1605057288-60400-1-git-send-email-tiantao6@hisilicon.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>

virtio-mem: Big Block Mode (BBM) - safe memory hotunplug

Let's add a safe mechanism to unplug memory, avoiding long/endless loops
when trying to offline memory - similar to in SBM.

Fake-offline all memory (via alloc_contig_range()) before trying to
offline+remove it. Use this mode as default, but allow to enable the other
mode explicitly (which could give better memory hotunplug guarantees in
some environments).

The "unsafe" mode can be enabled e.g., via virtio_mem.bbm_safe_unplug=0
on the cmdline.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-30-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: Big Block Mode (BBM) - basic memory hotunplug

Let's try to unplug completely offline big blocks first. Then, (if
enabled via unplug_offline) try to offline and remove whole big blocks.

No locking necessary - we can deal with concurrent onlining/offlining
just fine.

Note1: This is sub-optimal and might be dangerous in some environments: we
could end up in an infinite loop when offlining (e.g., long-term pinnings),
similar as with DIMMs. We'll introduce safe memory hotunplug via
fake-offlining next, and use this basic mode only when explicitly enabled.

Note2: Without ZONE_MOVABLE, memory unplug will be extremely unreliable
with bigger block sizes.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-29-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

mm/memory_hotplug: extend offline_and_remove_memory() to handle more than one memory block

virtio-mem soon wants to use offline_and_remove_memory() memory that
exceeds a single Linux memory block (memory_block_size_bytes()). Let's
remove that restriction.

Let's remember the old state and try to restore that if anything goes
wrong. While re-onlining can, in general, fail, it's highly unlikely to
happen (usually only when a notifier fails to allocate memory, and these
are rather rare).

This will be used by virtio-mem to offline+remove memory ranges that are
bigger than a single memory block - for example, with a device block
size of 1 GiB (e.g., gigantic pages in the hypervisor) and a Linux memory
block size of 128MB.

While we could compress the state into 2 bit, using 8 bit is much
easier.

This handling is similar, but different to acpi_scan_try_to_offline():

a) We don't try to offline twice. I am not sure if this CONFIG_MEMCG
optimization is still relevant - it should only apply to ZONE_NORMAL
(where we have no guarantees). If relevant, we can always add it.

b) acpi_scan_try_to_offline() simply onlines all memory in case
something goes wrong. It doesn't restore previous online type. Let's do
that, so we won't overwrite what e.g., user space configured.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-28-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Andrew Morton <akpm@linux-foundation.org>

virtio-mem: allow to force Big Block Mode (BBM) and set the big block size

Let's allow to force BBM, even if subblocks would be possible. Take care
of properly calculating the first big block id, because the start
address might no longer be aligned to the big block size.

Also, allow to manually configure the size of Big Blocks.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-27-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: Big Block Mode (BBM) memory hotplug

Currently, we do not support device block sizes that exceed the Linux
memory block size. For example, having a device block size of 1 GiB (e.g.,
gigantic pages in the hypervisor) won't work with 128 MiB Linux memory
blocks.

Let's implement Big Block Mode (BBM), whereby we add/remove at least
one Linux memory block at a time. With a 1 GiB device block size, a Big
Block (BB) will cover 8 Linux memory blocks.

We'll keep registering the online_page_callback machinery, it will be used
for safe memory hotunplug in BBM next.

Note: BBM is properly prepared for variable-sized Linux memory
blocks that we might see in the future. So we won't care how many Linux
memory blocks a big block actually spans, and how the memory notifier is
called.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Oscar Salvador <osalvador@suse.de>
Cc: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-26-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: factor out adding/removing memory from Linux

Let's use wrappers for the low-level functions that dev_dbg/dev_warn
and work on addr + size, such that we can reuse them for adding/removing
in other granularity.

We only warn when adding memory failed, because that's something to pay
attention to. We won't warn when removing failed, we'll reuse that in
racy context soon (and we do have proper BUG_ON() statements in the
current cases where it must never happen).

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-25-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: memory notifier callbacks are specific to Sub Block Mode (SBM)

Let's rename accordingly.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-24-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virito-mem: existing (un)plug functions are specific to Sub Block Mode (SBM)

Let's rename them accordingly. virtio_mem_plug_request() and
virtio_mem_unplug_request() will be handled separately.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-23-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: memory block ids are specific to Sub Block Mode (SBM)

Let's move first_mb_id/next_mb_id/last_usable_mb_id accordingly.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-22-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: nb_sb_per_mb and subblock_size are specific to Sub Block Mode (SBM)

Let's rename to "sbs_per_mb" and "sb_size" and move accordingly.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-21-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virito-mem: subblock states are specific to Sub Block Mode (SBM)

Let's rename and move accordingly. While at it, rename sb_bitmap to
"sb_states".

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-20-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: memory block states are specific to Sub Block Mode (SBM)

let's use a new "sbm" sub-struct to hold SBM-specific state and rename +
move applicable definitions, functions, and variables (related to
memory block states).

While at it:
- Drop the "_STATE" part from memory block states
- Rename "nb_mb_state" to "mb_count"
- "set_mb_state" / "get_mb_state" vs. "mb_set_state" / "mb_get_state"
- Don't use lengthy "enum virtio_mem_smb_mb_state", simply use "uint8_t"

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-19-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virito-mem: document Sub Block Mode (SBM)

Let's add some documentation for the current mode - Sub Block Mode (SBM) -
to prepare for a new mode - Big Block Mode (BBM).

Follow-up patches will properly factor out the existing Sub Block Mode
(SBM) and implement Big Block Mode (BBM).

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-18-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: generalize handling when memory is getting onlined deferred

We don't want to add too much memory when it's not getting onlined
immediately, to avoid running OOM. Generalize the handling, to avoid
making use of memory block states. Use a threshold of 1 GiB for now.

Properly adjust the offline size when adding/removing memory. As we are
not always protected by a lock when touching the offline size, use an
atomic64_t. We don't care about races (e.g., someone offlining memory
while we are adding more), only about consistent values.

(1 GiB needs a memmap of ~16MiB - which sounds reasonable even for
 setups with little boot memory and (possibly) one virtio-mem device per
 node)

We don't want to retrigger when onlining is caused immediately by our
action (e.g., adding memory which immediately gets onlined), so use a
flag to indicate if the workqueue is active and use that as an
indicator whether to trigger a retry. This will also be especially relevant
for Big Block Mode (BBM), whereby we might re-online memory in case
offlining of another memory block failed.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-17-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: don't always trigger the workqueue when offlining memory

Let's trigger from offlining code only when we're not allowed to unplug
online memory. Handle the other case (memmap possibly freeing up another
memory block) when actually removing memory. We now also properly handle
the case when removing already offline memory blocks via
virtio_mem_mb_remove(). When removing via virtio_mem_remove(), when
unloading the driver, virtio_mem_retry() is a NOP and safe to use.

While at it, move retry handling when offlining out of
virtio_mem_notify_offline(), to share it with Big Block Mode (BBM)
soon.

This is a preparation for Big Block Mode (BBM), whereby we can see some
temporary offlining of memory blocks without actually making progress.
Imagine you have a Big Block that spans to Linux memory blocks. Assume
the first Linux memory blocks has no unmovable data on it. When we would
call offline_and_remove_memory() on the big block, we would
1. Try to offline the first block. Works, notifiers triggered.
   virtio_mem_retry() called.
2. Try to offline the second block. Does not work.
3. Re-online first block.
4. Exit to main loop, exit workqueue.
5. Retry immediately (due to virtio_mem_retry()), go to 1.
The result are endless retries.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-16-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: drop last_mb_id

No longer used, let's drop it.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-15-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: generalize virtio_mem_overlaps_range()

Avoid using memory block ids. While at it, use uint64_t for
address/size.

This is a preparation for Big Block Mode (BBM).

Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-14-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: generalize virtio_mem_owned_mb()

Avoid using memory block ids. Rename it to virtio_mem_contains_range().

This is a preparation for Big Block Mode (BBM).

Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-13-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: generalize check for added memory

Let's check by traversing busy system RAM resources instead, to avoid
relying on memory block states.

Don't use walk_system_ram_range(), as that works on pages and we want to
use the bare addresses we have easily at hand.

This is a preparation for Big Block Mode (BBM), which won't have memory
block states.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-12-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: retry fake-offlining via alloc_contig_range() on ZONE_MOVABLE

ZONE_MOVABLE is supposed to give some guarantees, yet,
alloc_contig_range() isn't prepared to properly deal with some racy
cases properly (e.g., temporary page pinning when exiting processed, PCP).

Retry 5 times for now. There is certainly room for improvement in the
future.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-11-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: factor out handling of fake-offline pages in memory notifier

Let's factor out the core pieces and place the implementation next to
virtio_mem_fake_offline(). We'll reuse this functionality soon.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-10-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: factor out fake-offlining into virtio_mem_fake_offline()

... which now matches virtio_mem_fake_online(). We'll reuse this
functionality soon.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-9-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: print debug messages from virtio_mem_send_*_request()

Let's move the existing dev_dbg() into the functions, print if something
went wrong, and also print for virtio_mem_send_unplug_all_request().

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-8-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: factor out calculation of the bit number within the subblock bitmap

The calculation is already complicated enough, let's limit it to one
location.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-7-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: use "unsigned long" for nr_pages when fake onlining/offlining

No harm done, but let's be consistent.

Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-6-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: drop rc2 in virtio_mem_mb_plug_and_add()

We can drop rc2, we don't actually need the value.

Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-5-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: simplify MAX_ORDER - 1 / pageblock_order handling

Let's use pageblock_nr_pages and MAX_ORDER_NR_PAGES instead where
possible to simplify.

Add a comment why we have that restriction for now.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-4-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: more precise calculation in virtio_mem_mb_state_prepare_next_mb()

We actually need one byte less (next_mb_id is exclusive, first_mb_id is
inclusive). While at it, compact the code.

Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-3-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

virtio-mem: determine nid only once using memory_add_physaddr_to_nid()

Let's determine the target nid only once in case we have none specified -
usually, we'll end up with node 0 either way.

Reviewed-by: Wei Yang <richard.weiyang@linux.alibaba.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Jason Wang <jasowang@redhat.com>
Cc: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Signed-off-by: David Hildenbrand <david@redhat.com>
Link: https://lore.kernel.org/r/20201112133815.13332-2-david@redhat.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

Merge tag 'xfs-5.11-merge-4' of git://git./fs/xfs/xfs-linux

Pull xfs updates from Darrick Wong:
 "In this release we add the ability to set a 'needsrepair' flag
  indicating that we /know/ the filesystem requires xfs_repair, but
  other than that, it's the usual strengthening of metadata validation
  and miscellaneous cleanups.

  Summary:

   - Introduce a "needsrepair" "feature" to flag a filesystem as needing
     a pass through xfs_repair. This is key to enabling filesystem
     upgrades (in xfs_db) that require xfs_repair to make minor
     adjustments to metadata.

   - Refactor parameter checking of recovered log intent items so that
     we actually use the same validation code as them that generate the
     intent items.

   - Various fixes to online scrub not reacting correctly to directory
     entries pointing to inodes that cannot be igetted.

   - Refactor validation helpers for data and rt volume extents.

   - Refactor XFS_TRANS_DQ_DIRTY out of existence.

   - Fix a longstanding bug where mounting with "uqnoenforce" would
     start user quotas in non-enforcing mode but /proc/mounts would
     display "usrquota", implying that they are being enforced.

   - Don't flag dax+reflink inodes as corruption since that is a valid
     (but not fully functional) combination right now.

   - Clean up raid stripe validation functions.

   - Refactor the inode allocation code to be more straightforward.

   - Small prep cleanup for idmapping support.

   - Get rid of the xfs_buf_t typedef"

* tag 'xfs-5.11-merge-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (40 commits)
  xfs: remove xfs_buf_t typedef
  fs/xfs: convert comma to semicolon
  xfs: open code updating i_mode in xfs_set_acl
  xfs: remove xfs_vn_setattr_nonsize
  xfs: kill ialloced in xfs_dialloc()
  xfs: spilt xfs_dialloc() into 2 functions
  xfs: move xfs_dialloc_roll() into xfs_dialloc()
  xfs: move on-disk inode allocation out of xfs_ialloc()
  xfs: introduce xfs_dialloc_roll()
  xfs: convert noroom, okalloc in xfs_dialloc() to bool
  xfs: don't catch dax+reflink inodes as corruption in verifier
  xfs: fix the forward progress assertion in xfs_iwalk_run_callbacks
  xfs: remove unneeded return value check for *init_cursor()
  xfs: introduce xfs_validate_stripe_geometry()
  xfs: show the proper user quota options
  xfs: remove the unused XFS_B_FSB_OFFSET macro
  xfs: remove unnecessary null check in xfs_generic_create
  xfs: directly return if the delta equal to zero
  xfs: check tp->t_dqinfo value instead of the XFS_TRANS_DQ_DIRTY flag
  xfs: delete duplicated tp->t_dqinfo null check and allocation
  ...

Merge tag 'ktest-v5.11' of git://git./linux/kernel/git/rostedt/linux-ktest

Pull ktest updates from Steven Rostedt:
 "No new features. Just a couple of fixes that I had in my local
  repository that fixed issues with sending the result emails"

* tag 'ktest-v5.11' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-ktest:
  ktest.pl: Fix the logic for truncating the size of the log file for email
  ktest.pl: If size of log is too big to email, email error message

Merge tag 'drm-next-2020-12-18' of git://anongit.freedesktop.org/drm/drm

Pull more drm updates from Daniel Vetter:
 "UAPI Changes:

   - Only enable char/agp uapi when CONFIG_DRM_LEGACY is set

  Cross-subsystem Changes:

   - vma_set_file helper to make vma->vm_file changing less brittle,
     acked by Andrew

  Core Changes:

   - dma-buf heaps improvements

   - pass full atomic modeset state to driver callbacks

   - shmem helpers: cached bo by default

   - cleanups for fbdev, fb-helpers

   - better docs for drm modes and SCALING_FITLER uapi

   - ttm: fix dma32 page pool regression

  Driver Changes:

   - multi-hop regression fixes for amdgpu, radeon, nouveau

   - lots of small amdgpu hw enabling fixes (display, pm, ...)

   - fixes for imx, mcde, meson, some panels, virtio, qxl, i915, all
     fairly minor

   - some cleanups for legacy drm/fbdev drivers"

* tag 'drm-next-2020-12-18' of git://anongit.freedesktop.org/drm/drm: (117 commits)
  drm/qxl: don't allocate a dma_address array
  drm/nouveau: fix multihop when move doesn't work.
  drm/i915/tgl: Fix REVID macros for TGL to fetch correct stepping
  drm/i915: Fix mismatch between misplaced vma check and vma insert
  drm/i915/perf: also include Gen11 in OATAILPTR workaround
  Revert "drm/i915: re-order if/else ladder for hpd_irq_setup"
  drm/amdgpu/disply: fix documentation warnings in display manager
  drm/amdgpu: print mmhub client name for dimgrey_cavefish
  drm/amdgpu: set mode1 reset as default for dimgrey_cavefish
  drm/amd/display: Add get_dig_frontend implementation for DCEx
  drm/radeon: remove h from printk format specifier
  drm/amdgpu: remove h from printk format specifier
  drm/amdgpu: Fix spelling mistake "Heterogenous" -> "Heterogeneous"
  drm/amdgpu: fix regression in vbios reservation handling on headless
  drm/amdgpu/SRIOV: Extend VF reset request wait period
  drm/amdkfd: correct amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu log.
  drm/amd/display: Adding prototype for dccg21_update_dpp_dto()
  drm/amdgpu: print what method we are using for runtime pm
  drm/amdgpu: simplify logic in atpx resume handling
  drm/amdgpu: no need to call pci_ignore_hotplug for _PR3
  ...

tools headers UAPI: Update asm-generic/unistd.h

Just a comment change, trivial.

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

tools headers cpufeatures: Sync with the kernel sources

To pick the changes in:

  e7b6385b01d8e9fb ("x86/cpufeatures: Add Intel SGX hardware bits")

That causes only these 'perf bench' objects to rebuild:

  CC       /tmp/build/perf/bench/mem-memcpy-x86-64-asm.o
  CC       /tmp/build/perf/bench/mem-memset-x86-64-asm.o

And addresses these perf build warnings:

  Warning: Kernel ABI header at 'tools/arch/x86/include/asm/disabled-features.h' differs from latest version at 'arch/x86/include/asm/disabled-features.h'
  diff -u tools/arch/x86/include/asm/disabled-features.h arch/x86/include/asm/disabled-features.h
  Warning: Kernel ABI header at 'tools/arch/x86/include/asm/cpufeatures.h' differs from latest version at 'arch/x86/include/asm/cpufeatures.h'
  diff -u tools/arch/x86/include/asm/cpufeatures.h arch/x86/include/asm/cpufeatures.h

Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@suse.de>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Sean Christopherson <seanjc@google.com>
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>