Do this by grouping fields used for cached writes and putting them into a
union with fileds used for cached readdir (with obviously no overlap, since
we don't have hybrid objects).
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Use the internal iversion counter to make sure modifications of the
directory through this filesystem are not missed by the mtime check (due to
mtime granularity).
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Store the modification time of the directory in the cache, obtained before
starting to fill the cache.

When reading the cache, verify that the directory hasn't changed, by
checking if current modification time is the same as the one stored in the
cache.

This only needs to be done when the current file position is at the
beginning of the directory, as mandated by POSIX.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Allow the cache to be invalidated when page(s) have gone missing.  In this
case increment the version of the cache and reset to an empty state.

Add a version number to the directory stream in struct fuse_file as well,
indicating the version of the cache it's supposed to be reading.  If the
cache version doesn't match the stream's version, then reset the stream to
the beginning of the cache.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

The cache is only used if it's completed, not while it's still being
filled; this constraint could be lifted later, if it turns out to be
useful.

Introduce state in struct fuse_file that indicates the position within the
cache.  After a seek, reset the position to the beginning of the cache and
search the cache for the current position.  If the current position is not
found in the cache, then fall back to uncached readdir.

It can also happen that page(s) disappear from the cache, in which case we
must also fall back to uncached readdir.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

This patch just adds the cache filling functions, which are invoked if
FOPEN_CACHE_DIR flag is set in the OPENDIR reply.

Cache reading and cache invalidation are added by subsequent patches.

The directory cache uses the page cache.  Directory entries are packed into
a page in the same format as in the READDIR reply.  A page only contains
whole entries, the space at the end of the page is cleared.  The page is
locked while being modified.

Multiple parallel readdirs on the same directory can fill the cache; the
only constraint is that continuity must be maintained (d_off of last entry
points to position of current entry).
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Prepare for cache filling by introducing a helper for emitting a single
directory entry.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Add flag returned by OPENDIR request to allow kernel to cache directory
contents in page cache.  The effect of FOPEN_CACHE_DIR is twofold:

 a) if not already cached, it writes entries into the cache

 b) if already cached, it allows reading entries from the cache

The FOPEN_KEEP_CACHE has the same effect as on regular files: unless this
flag is given the cache is cleared upon completion of open.

So FOPEN_KEEP_CACHE and FOPEN_KEEP_CACHE flags should be used together to
make use of the directory caching facility introduced in the following
patches.

The FUSE_AUTO_INVAL_DATA flag returned in INIT reply also has the same
affect on the directory cache as it has on data cache for regular files.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Directory reading code is about to grow larger, so split it out from dir.c
into a new source file.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

We noticed the performance bottleneck in FUSE running our Virtuozzo storage
over rdma. On some types of workload we observe 20% of times spent in
request_find() in profiler.  This function is iterating over long requests
list, and it scales bad.

The patch introduces hash table to reduce the number of iterations, we do
in this function. Hash generating algorithm is taken from hash_add()
function, while 256 lines table is used to store pending requests.  This
fixes problem and improves the performance.
Reported-by: NAlexey Kuznetsov <kuznet@virtuozzo.com>
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

This field is not needed after the previous patch, since we can easily
convert request ID to interrupt request ID and vice versa.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Using of two unconnected IDs req->in.h.unique and req->intr_unique does not
allow to link requests to a hash table. We need can't use none of them as a
key to calculate hash.

This patch changes the algorithm of allocation of IDs for a request. Plain
requests obtain even ID, while interrupt requests are encoded in the low
bit. So, in next patches we will be able to use the rest of ID bits to
calculate hash, and the hash will be the same for plain and interrupt
requests.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Currently, we take fc->lock there only to check for fc->connected.
But this flag is changed only on connection abort, which is very
rare operation.

So allow checking fc->connected under just fc->bg_lock and use this lock
(as well as fc->lock) when resetting fc->connected.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

To reduce contention of fc->lock, this patch introduces bg_lock for
protection of fields related to background queue. These are:
max_background, congestion_threshold, num_background, active_background,
bg_queue and blocked.

This allows next patch to make async reads not requiring fc->lock, so async
reads and writes will have better performance executed in parallel.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Functions sequences like request_end()->flush_bg_queue() require that
max_background and congestion_threshold are constant during their
execution. Otherwise, checks like

	if (fc->num_background == fc->max_background)

made in different time may behave not like expected.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Since they are of unsigned int type, it's allowed to read them
unlocked during reporting to userspace. Let's underline this fact
with READ_ONCE() macroses.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

This cleanup patch makes the function to use the primitive
instead of direct dereferencing.

Also, move fiq dereferencing out of cycle, since it's
always constant.
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

There are several FUSE filesystems that can implement server-side copy
or other efficient copy/duplication/clone methods. The copy_file_range()
syscall is the standard interface that users have access to while not
depending on external libraries that bypass FUSE.
Signed-off-by: NNiels de Vos <ndevos@redhat.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>

Using waitqueue_active() is racy.  Make sure we issue a wake_up()
unconditionally after storing into fc->blocked.  After that it's okay to
optimize with waitqueue_active() since the first wake up provides the
necessary barrier for all waiters, not the just the woken one.
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Fixes: 3c18ef81 ("fuse: optimize wake_up")
Cc: <stable@vger.kernel.org> # v3.10

Otherwise fuse_dev_do_write() could come in and finish off the request, and
the set_bit(FR_SENT, ...) could trigger the WARN_ON(test_bit(FR_SENT, ...))
in request_end().
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Reported-by: syzbot+ef054c4d3f64cd7f7cec@syzkaller.appspotmai
Fixes: 46c34a34 ("fuse: no fc->lock for pqueue parts")
Cc: <stable@vger.kernel.org> # v4.2

After we found req in request_find() and released the lock,
everything may happen with the req in parallel:

cpu0                              cpu1
fuse_dev_do_write()               fuse_dev_do_write()
  req = request_find(fpq, ...)    ...
  spin_unlock(&fpq->lock)         ...
  ...                             req = request_find(fpq, oh.unique)
  ...                             spin_unlock(&fpq->lock)
  queue_interrupt(&fc->iq, req);   ...
  ...                              ...
  ...                              ...
  request_end(fc, req);
    fuse_put_request(fc, req);
  ...                              queue_interrupt(&fc->iq, req);
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Fixes: 46c34a34 ("fuse: no fc->lock for pqueue parts")
Cc: <stable@vger.kernel.org> # v4.2

We may pick freed req in this way:

[cpu0]                                  [cpu1]
fuse_dev_do_read()                      fuse_dev_do_write()
   list_move_tail(&req->list, ...);     ...
   spin_unlock(&fpq->lock);             ...
   ...                                  request_end(fc, req);
   ...                                    fuse_put_request(fc, req);
   if (test_bit(FR_INTERRUPTED, ...))
         queue_interrupt(fiq, req);

Fix that by keeping req alive until we finish all manipulations.

Reported-by: syzbot+4e975615ca01f2277bdd@syzkaller.appspotmail.com
Signed-off-by: NKirill Tkhai <ktkhai@virtuozzo.com>
Signed-off-by: NMiklos Szeredi <mszeredi@redhat.com>
Fixes: 46c34a34 ("fuse: no fc->lock for pqueue parts")
Cc: <stable@vger.kernel.org> # v4.2

Lee writes:
  "MFD fixes for v4.19
   - Fix Dialog DA9063 regulator constraints issue causing failure in
     probe
   - Fix OMAP Device Tree compatible strings to match DT"

* tag 'mfd-fixes-4.19' of git://git.kernel.org/pub/scm/linux/kernel/git/lee/mfd:
  mfd: omap-usb-host: Fix dts probe of children
  mfd: da9063: Fix DT probing with constraints

Juergen writes:
  "xen:
   Two small fixes for xen drivers."

* tag 'for-linus-4.19d-rc5-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen: issue warning message when out of grant maptrack entries
  xen/x86/vpmu: Zero struct pt_regs before calling into sample handling code

Jens writes:
  "Just a single fix in this pull request, fixing a regression in
  /proc/diskstats caused by the unification of timestamps."

* tag 'for-linus-20180922' of git://git.kernel.dk/linux-block:
  block: use nanosecond resolution for iostat

Thomas writes:
  "A set of fixes for x86:

   - Resolve the kvmclock regression on AMD systems with memory
     encryption enabled. The rework of the kvmclock memory allocation
     during early boot results in encrypted storage, which is not
     shareable with the hypervisor. Create a new section for this data
     which is mapped unencrypted and take care that the later
     allocations for shared kvmclock memory is unencrypted as well.

   - Fix the build regression in the paravirt code introduced by the
     recent spectre v2 updates.

   - Ensure that the initial static page tables cover the fixmap space
     correctly so early console always works. This worked so far by
     chance, but recent modifications to the fixmap layout can -
     depending on kernel configuration - move the relevant entries to a
     different place which is not covered by the initial static page
     tables.

   - Address the regressions and issues which got introduced with the
     recent extensions to the Intel Recource Director Technology code.

   - Update maintainer entries to document reality"

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/mm: Expand static page table for fixmap space
  MAINTAINERS: Add X86 MM entry
  x86/intel_rdt: Add Reinette as co-maintainer for RDT
  MAINTAINERS: Add Borislav to the x86 maintainers
  x86/paravirt: Fix some warning messages
  x86/intel_rdt: Fix incorrect loop end condition
  x86/intel_rdt: Fix exclusive mode handling of MBA resource
  x86/intel_rdt: Fix incorrect loop end condition
  x86/intel_rdt: Do not allow pseudo-locking of MBA resource
  x86/intel_rdt: Fix unchecked MSR access
  x86/intel_rdt: Fix invalid mode warning when multiple resources are managed
  x86/intel_rdt: Global closid helper to support future fixes
  x86/intel_rdt: Fix size reporting of MBA resource
  x86/intel_rdt: Fix data type in parsing callbacks
  x86/kvm: Use __bss_decrypted attribute in shared variables
  x86/mm: Add .bss..decrypted section to hold shared variables

Thomas writes:
  "- Provide a strerror_r wrapper so lib/bpf can be built on systems
     without _GNU_SOURCE
   - Unbreak the man page generator when building out of tree"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf Documentation: Fix out-of-tree asciidoctor man page generation
  tools lib bpf: Provide wrapper for strerror_r to build in !_GNU_SOURCE systems

Thomas writes:
  "Make the EFI arm stub device tree loader default on to unbreak
  existing EFI boot loaders which do not have DTB support."

* 'efi-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  efi/libstub/arm: default EFI_ARMSTUB_DTB_LOADER to y

Klaus Kusche reported that the I/O busy time in /proc/diskstats was not
updating properly on 4.18. This is because we started using ktime to
track elapsed time, and we convert nanoseconds to jiffies when we update
the partition counter. However, this gets rounded down, so any I/Os that
take less than a jiffy are not accounted for. Previously in this case,
the value of jiffies would sometimes increment while we were doing I/O,
so at least some I/Os were accounted for.

Let's convert the stats to use nanoseconds internally. We still report
milliseconds as before, now more accurately than ever. The value is
still truncated to 32 bits for backwards compatibility.

Fixes: 522a7775 ("block: consolidate struct request timestamp fields")
Cc: stable@vger.kernel.org
Reported-by: NKlaus Kusche <klaus.kusche@computerix.info>
Signed-off-by: NOmar Sandoval <osandov@fb.com>
Signed-off-by: NJens Axboe <axboe@kernel.dk>

Linus writes:
  "Pin control fixes for v4.19:
   - Two fixes for the Intel pin controllers than cause
     problems on laptops."

* tag 'pinctrl-v4.19-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
  pinctrl: intel: Do pin translation in other GPIO operations as well
  pinctrl: cannonlake: Fix gpio base for GPP-E

Paolo writes:
  "It's mostly small bugfixes and cleanups, mostly around x86 nested
   virtualization.  One important change, not related to nested
   virtualization, is that the ability for the guest kernel to trap
   CPUID instructions (in Linux that's the ARCH_SET_CPUID arch_prctl) is
   now masked by default.  This is because the feature is detected
   through an MSR; a very bad idea that Intel seems to like more and
   more.  Some applications choke if the other fields of that MSR are
   not initialized as on real hardware, hence we have to disable the
   whole MSR by default, as was the case before Linux 4.12."

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm: (23 commits)
  KVM: nVMX: Fix bad cleanup on error of get/set nested state IOCTLs
  kvm: selftests: Add platform_info_test
  KVM: x86: Control guest reads of MSR_PLATFORM_INFO
  KVM: x86: Turbo bits in MSR_PLATFORM_INFO
  nVMX x86: Check VPID value on vmentry of L2 guests
  nVMX x86: check posted-interrupt descriptor addresss on vmentry of L2
  KVM: nVMX: Wake blocked vCPU in guest-mode if pending interrupt in virtual APICv
  KVM: VMX: check nested state and CR4.VMXE against SMM
  kvm: x86: make kvm_{load|put}_guest_fpu() static
  x86/hyper-v: rename ipi_arg_{ex,non_ex} structures
  KVM: VMX: use preemption timer to force immediate VMExit
  KVM: VMX: modify preemption timer bit only when arming timer
  KVM: VMX: immediately mark preemption timer expired only for zero value
  KVM: SVM: Switch to bitmap_zalloc()
  KVM/MMU: Fix comment in walk_shadow_page_lockless_end()
  kvm: selftests: use -pthread instead of -lpthread
  KVM: x86: don't reset root in kvm_mmu_setup()
  kvm: mmu: Don't read PDPTEs when paging is not enabled
  x86/kvm/lapic: always disable MMIO interface in x2APIC mode
  KVM: s390: Make huge pages unavailable in ucontrol VMs
  ...

Richard writes:
  "This pull request contains fixes for UBIFS:
   - A wrong UBIFS assertion in mount code
   - Fix for a NULL pointer deref in mount code
   - Revert of a bad fix for xattrs"

* tag 'upstream-4.19-rc4' of git://git.infradead.org/linux-ubifs:
  Revert "ubifs: xattr: Don't operate on deleted inodes"
  ubifs: drop false positive assertion
  ubifs: Check for name being NULL while mounting

Jens writes:
  "Storage fixes for 4.19-rc5

  - Fix for leaking kernel pointer in floppy ioctl (Andy Whitcroft)

  - NVMe pull request from Christoph, and a single ANA log page fix
    (Hannes)

  - Regression fix for libata qd32 support, where we trigger an illegal
    active command transition. This fixes a CD-ROM detection issue that
    was reported, but could also trigger premature completion of the
    internal tag (me)"

* tag 'for-linus-20180920' of git://git.kernel.dk/linux-block:
  floppy: Do not copy a kernel pointer to user memory in FDGETPRM ioctl
  libata: mask swap internal and hardware tag
  nvme: count all ANA groups for ANA Log page

David writes:
  "drm fixes for 4.19-rc5:

   - core: fix debugfs for atomic, fix the check for atomic for
     non-modesetting drivers
   - amdgpu: adds a new PCI id, some kfd fixes and a sdma fix
   - i915: a bunch of GVT fixes.
   - vc4: scaling fix
   - vmwgfx: modesetting fixes and a old buffer eviction fix
   - udl: framebuffer destruction fix
   - sun4i: disable on R40 fix until next kernel
   - pl111: NULL termination on table fix"

* tag 'drm-fixes-2018-09-21' of git://anongit.freedesktop.org/drm/drm: (21 commits)
  drm/amdkfd: Fix ATS capablity was not reported correctly on some APUs
  drm/amdkfd: Change the control stack MTYPE from UC to NC on GFX9
  drm/amdgpu: Fix SDMA HQD destroy error on gfx_v7
  drm/vmwgfx: Fix buffer object eviction
  drm/vmwgfx: Don't impose STDU limits on framebuffer size
  drm/vmwgfx: limit mode size for all display unit to texture_max
  drm/vmwgfx: limit screen size to stdu_max during check_modeset
  drm/vmwgfx: don't check for old_crtc_state enable status
  drm/amdgpu: add new polaris pci id
  drm: sun4i: drop second PLL from A64 HDMI PHY
  drm: fix drm_drv_uses_atomic_modeset on non modesetting drivers.
  drm/i915/gvt: clear ggtt entries when destroy vgpu
  drm/i915/gvt: request srcu_read_lock before checking if one gfn is valid
  drm/i915/gvt: Add GEN9_CLKGATE_DIS_4 to default BXT mmio handler
  drm/i915/gvt: Init PHY related registers for BXT
  drm/atomic: Use drm_drv_uses_atomic_modeset() for debugfs creation
  drm/fb-helper: Remove set but not used variable 'connector_funcs'
  drm: udl: Destroy framebuffer only if it was initialized
  drm/sun4i: Remove R40 display pipeline compatibles
  drm/pl111: Make sure of_device_id tables are NULL terminated
  ...

A few fixes for 4.19:
- Add a new polaris pci id
- KFD fixes for raven and gfx7
Signed-off-by: NDave Airlie <airlied@redhat.com>
From: Alex Deucher <alexdeucher@gmail.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180920155850.5455-1-alexander.deucher@amd.com

A couple of modesetting fixes and a fix for a long-standing buffer-eviction
problem cc'd stable.
Signed-off-by: NDave Airlie <airlied@redhat.com>
From: Thomas Hellstrom <thellstrom@vmware.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180920063935.35492-1-thellstrom@vmware.com

We met a kernel panic when enabling earlycon, which is due to the fixmap
address of earlycon is not statically setup.

Currently the static fixmap setup in head_64.S only covers 2M virtual
address space, while it actually could be in 4M space with different
kernel configurations, e.g. when VSYSCALL emulation is disabled.

So increase the static space to 4M for now by defining FIXMAP_PMD_NUM to 2,
and add a build time check to ensure that the fixmap is covered by the
initial static page tables.

Fixes: 1ad83c85 ("x86_64,vsyscall: Make vsyscall emulation configurable")
Suggested-by: NThomas Gleixner <tglx@linutronix.de>
Signed-off-by: NFeng Tang <feng.tang@intel.com>
Signed-off-by: NThomas Gleixner <tglx@linutronix.de>
Tested-by: Nkernel test robot <rong.a.chen@intel.com>
Reviewed-by: Juergen Gross <jgross@suse.com> (Xen parts)
Cc: H Peter Anvin <hpa@linux.intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Yinghai Lu <yinghai@kernel.org>
Cc: Dave Hansen <dave.hansen@intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Andy Lutomirsky <luto@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20180920025828.23699-1-feng.tang@intel.com

While reading block, it is possible that io error return due to underlying
storage issue, in this case, BH_NeedsValidate was left in the buffer head.
Then when reading the very block next time, if it was already linked into
journal, that will trigger the following panic.

[203748.702517] kernel BUG at fs/ocfs2/buffer_head_io.c:342!
[203748.702533] invalid opcode: 0000 [#1] SMP
[203748.702561] Modules linked in: ocfs2 ocfs2_dlmfs ocfs2_stack_o2cb ocfs2_dlm ocfs2_nodemanager ocfs2_stackglue configfs sunrpc dm_switch dm_queue_length dm_multipath bonding be2iscsi iscsi_boot_sysfs bnx2i cnic uio cxgb4i iw_cxgb4 cxgb4 cxgb3i libcxgbi iw_cxgb3 cxgb3 mdio ib_iser rdma_cm ib_cm iw_cm ib_sa ib_mad ib_core ib_addr ipv6 iscsi_tcp libiscsi_tcp libiscsi scsi_transport_iscsi ipmi_devintf iTCO_wdt iTCO_vendor_support dcdbas ipmi_ssif i2c_core ipmi_si ipmi_msghandler acpi_pad pcspkr sb_edac edac_core lpc_ich mfd_core shpchp sg tg3 ptp pps_core ext4 jbd2 mbcache2 sr_mod cdrom sd_mod ahci libahci megaraid_sas wmi dm_mirror dm_region_hash dm_log dm_mod
[203748.703024] CPU: 7 PID: 38369 Comm: touch Not tainted 4.1.12-124.18.6.el6uek.x86_64 #2
[203748.703045] Hardware name: Dell Inc. PowerEdge R620/0PXXHP, BIOS 2.5.2 01/28/2015
[203748.703067] task: ffff880768139c00 ti: ffff88006ff48000 task.ti: ffff88006ff48000
[203748.703088] RIP: 0010:[<ffffffffa05e9f09>]  [<ffffffffa05e9f09>] ocfs2_read_blocks+0x669/0x7f0 [ocfs2]
[203748.703130] RSP: 0018:ffff88006ff4b818  EFLAGS: 00010206
[203748.703389] RAX: 0000000008620029 RBX: ffff88006ff4b910 RCX: 0000000000000000
[203748.703885] RDX: 0000000000000001 RSI: 0000000000000000 RDI: 00000000023079fe
[203748.704382] RBP: ffff88006ff4b8d8 R08: 0000000000000000 R09: ffff8807578c25b0
[203748.704877] R10: 000000000f637376 R11: 000000003030322e R12: 0000000000000000
[203748.705373] R13: ffff88006ff4b910 R14: ffff880732fe38f0 R15: 0000000000000000
[203748.705871] FS:  00007f401992c700(0000) GS:ffff880bfebc0000(0000) knlGS:0000000000000000
[203748.706370] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[203748.706627] CR2: 00007f4019252440 CR3: 00000000a621e000 CR4: 0000000000060670
[203748.707124] Stack:
[203748.707371]  ffff88006ff4b828 ffffffffa0609f52 ffff88006ff4b838 0000000000000001
[203748.707885]  0000000000000000 0000000000000000 ffff880bf67c3800 ffffffffa05eca00
[203748.708399]  00000000023079ff ffffffff81c58b80 0000000000000000 0000000000000000
[203748.708915] Call Trace:
[203748.709175]  [<ffffffffa0609f52>] ? ocfs2_inode_cache_io_unlock+0x12/0x20 [ocfs2]
[203748.709680]  [<ffffffffa05eca00>] ? ocfs2_empty_dir_filldir+0x80/0x80 [ocfs2]
[203748.710185]  [<ffffffffa05ec0cb>] ocfs2_read_dir_block_direct+0x3b/0x200 [ocfs2]
[203748.710691]  [<ffffffffa05f0fbf>] ocfs2_prepare_dx_dir_for_insert.isra.57+0x19f/0xf60 [ocfs2]
[203748.711204]  [<ffffffffa065660f>] ? ocfs2_metadata_cache_io_unlock+0x1f/0x30 [ocfs2]
[203748.711716]  [<ffffffffa05f4f3a>] ocfs2_prepare_dir_for_insert+0x13a/0x890 [ocfs2]
[203748.712227]  [<ffffffffa05f442e>] ? ocfs2_check_dir_for_entry+0x8e/0x140 [ocfs2]
[203748.712737]  [<ffffffffa061b2f2>] ocfs2_mknod+0x4b2/0x1370 [ocfs2]
[203748.713003]  [<ffffffffa061c385>] ocfs2_create+0x65/0x170 [ocfs2]
[203748.713263]  [<ffffffff8121714b>] vfs_create+0xdb/0x150
[203748.713518]  [<ffffffff8121b225>] do_last+0x815/0x1210
[203748.713772]  [<ffffffff812192e9>] ? path_init+0xb9/0x450
[203748.714123]  [<ffffffff8121bca0>] path_openat+0x80/0x600
[203748.714378]  [<ffffffff811bcd45>] ? handle_pte_fault+0xd15/0x1620
[203748.714634]  [<ffffffff8121d7ba>] do_filp_open+0x3a/0xb0
[203748.714888]  [<ffffffff8122a767>] ? __alloc_fd+0xa7/0x130
[203748.715143]  [<ffffffff81209ffc>] do_sys_open+0x12c/0x220
[203748.715403]  [<ffffffff81026ddb>] ? syscall_trace_enter_phase1+0x11b/0x180
[203748.715668]  [<ffffffff816f0c9f>] ? system_call_after_swapgs+0xe9/0x190
[203748.715928]  [<ffffffff8120a10e>] SyS_open+0x1e/0x20
[203748.716184]  [<ffffffff816f0d5e>] system_call_fastpath+0x18/0xd7
[203748.716440] Code: 00 00 48 8b 7b 08 48 83 c3 10 45 89 f8 44 89 e1 44 89 f2 4c 89 ee e8 07 06 11 e1 48 8b 03 48 85 c0 75 df 8b 5d c8 e9 4d fa ff ff <0f> 0b 48 8b 7d a0 e8 dc c6 06 00 48 b8 00 00 00 00 00 00 00 10
[203748.717505] RIP  [<ffffffffa05e9f09>] ocfs2_read_blocks+0x669/0x7f0 [ocfs2]
[203748.717775]  RSP <ffff88006ff4b818>

Joesph ever reported a similar panic.
Link: https://oss.oracle.com/pipermail/ocfs2-devel/2013-May/008931.html

Link: http://lkml.kernel.org/r/20180912063207.29484-1-junxiao.bi@oracle.comSigned-off-by: NJunxiao Bi <junxiao.bi@oracle.com>
Cc: Joseph Qi <jiangqi903@gmail.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Changwei Ge <ge.changwei@h3c.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>

9092c71b ("mm: use sc->priority for slab shrink targets") changed the
way that the target slab pressure is calculated and made it
priority-based:

    delta = freeable >> priority;
    delta *= 4;
    do_div(delta, shrinker->seeks);

The problem is that on a default priority (which is 12) no pressure is
applied at all, if the number of potentially reclaimable objects is less
than 4096 (1<<12).

This causes the last objects on slab caches of no longer used cgroups to
(almost) never get reclaimed.  It's obviously a waste of memory.

It can be especially painful, if these stale objects are holding a
reference to a dying cgroup.  Slab LRU lists are reparented on memcg
offlining, but corresponding objects are still holding a reference to the
dying cgroup.  If we don't scan these objects, the dying cgroup can't go
away.  Most likely, the parent cgroup hasn't any directly charged objects,
only remaining objects from dying children cgroups.  So it can easily hold
a reference to hundreds of dying cgroups.

If there are no big spikes in memory pressure, and new memory cgroups are
created and destroyed periodically, this causes the number of dying
cgroups grow steadily, causing a slow-ish and hard-to-detect memory
"leak".  It's not a real leak, as the memory can be eventually reclaimed,
but it could not happen in a real life at all.  I've seen hosts with a
steadily climbing number of dying cgroups, which doesn't show any signs of
a decline in months, despite the host is loaded with a production
workload.

It is an obvious waste of memory, and to prevent it, let's apply a minimal
pressure even on small shrinker lists.  E.g.  if there are freeable
objects, let's scan at least min(freeable, scan_batch) objects.

This fix significantly improves a chance of a dying cgroup to be
reclaimed, and together with some previous patches stops the steady growth
of the dying cgroups number on some of our hosts.

Link: http://lkml.kernel.org/r/20180905230759.12236-1-guro@fb.com
Fixes: 9092c71b ("mm: use sc->priority for slab shrink targets")
Signed-off-by: NRoman Gushchin <guro@fb.com>
Acked-by: NRik van Riel <riel@surriel.com>
Cc: Josef Bacik <jbacik@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Shakeel Butt <shakeelb@google.com>
Cc: Michal Hocko <mhocko@kernel.org>
Cc: Vladimir Davydov <vdavydov.dev@gmail.com>
Signed-off-by: NAndrew Morton <akpm@linux-foundation.org>
Signed-off-by: NGreg Kroah-Hartman <gregkh@linuxfoundation.org>