->readpage looks in file->private_data to try and find out how the
userspace program set "count" in read(2) or with "dd bs=" or whatever.

->readpage uses "count" and inode->i_size to calculate how much
data Orangefs should deposit in the Orangefs shared buffer, and
remembers which slot the data is in.

After copying data from the Orangefs shared buffer slot into
"the page", readpage tries to increment through the pagecache index
and fill as many pages as it can from the extra data in the shared
buffer. Hopefully these extra pages will soon be needed by the vfs,
and they'll be in the pagecache already.
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
Signed-off-by: Martin Brandenburg <martin@omnibond.com>

When userspace deposits more than a page of data into the shared buffer,
we'll need to know which slot it is in when we get back to readpage
so that we can try to use the extra data to fill some extra pages.
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
Signed-off-by: Martin Brandenburg <martin@omnibond.com>

Orangefs wins when it can do IO on large (up to four meg) blocks at a time,
and looses when it has to do tiny "small io" reads and writes. Accessing
Orangefs through the pagecache with the kernel module helps with small io,
both reading and writing, a great deal. Readpage generally tries to fetch a
page (four k) at a time. We'll let users use "count" (as in read(2) or
pread(2) for example) as a knob to control how much data they get from
Orangefs at a time and we'll try to use the data to fill extra
pagecache pages when we get to ->readpage, hopefully resulting in
fewer calls to readpage and Orangefs userspace.

We need a way to remember how they set count so that we can still have
it available when we get to ->readpage.

 - We'll use file->private_data to keep track of "count".
   We'll wrap generic_file_open with orangefs_file_open and
   initialize private_data to NULL there.

 - In ->read_iter we have access to both "count" and file, so
   we'll kmalloc some space onto file->private_data and store
   "count" there.

 - We'll kfree file->private_data each time we visit ->flush and
   reinitialize it to NULL.
Signed-off-by: Mike Marshall <hubcap@omnibond.com>
Signed-off-by: Martin Brandenburg <martin@omnibond.com>

This is modeled after NFS, except our method is different.  We use a
simple timer to determine whether to invalidate the page cache.  This
is bound to perform.

This addes a sysfs parameter cache_timeout_msecs which controls the time
between page cache invalidations.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Go through pages and look for a consecutive writable region.  After
finding a number of consecutive writable pages or when finding that
the next page's dirty range is not contiguous and cannot be written
as one request, send the write to the server.

The number of pages is determined by the client-core's buffer size.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Attach the actual range of bytes written to plus the responsible uid/gid
to each dirty page.  This information must be sent to the server when
the page is written out.

Now write_begin, page_mkwrite, and invalidatepage keep up with this
information.  There are several conditions where they must write out the
page immediately to store the new range.  Two non-contiguous ranges
cannot be stored on a single page.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Without this, an fsync call is sent to the server even if no data
changed.  This resulted in a rather severe (50%) performance regression
under certain metadata-heavy workloads.

In the past, everything was direct IO.  Nothing happend on a close call.
An explicit fsync call would send an fsync request to the server which
in turn fsynced the underlying file.

Now there are cached writes.  Then fsync began writing out dirty pages
in addition to making an fsync request to the server, and close began
calling fsync.

With this commit, close only writes out dirty pages, and does not make
the fsync request.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Would happen if an inode is dirty but whatever happened is not something
that can be written out to OrangeFS.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

direct_IO was the only caller and all direct_IO did was call it,
so there's no use in having the code spread out into so many functions.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Now orangefs_inode_getattr fills from cache if an inode has dirty pages.

also if attr_valid and dirty pages and !flags, we spin on inode writeback
before returning if pages still dirty after: should it be other way
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Remove orangefs_inode_read.  It was used by readpage.  Calling
wait_for_direct_io directly serves the purpose just as well.  There is
now no check of the bufmap size in the readpage path.  There are already
other places the bufmap size is assumed to be greater than PAGE_SIZE.

Important to call truncate_inode_pages now in the write path so a
subsequent read sees the new data.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

It's a copy of the loop which would run in read_pages from
mm/readahead.c.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

OrangeFS accepts a mask indicating which attributes were changed.  The
kernel must not set any bits except those that were actually changed.
The kernel must set the uid/gid of the request to the actual uid/gid
responsible for the change.

Code path for notify_change initiated setattrs is

orangefs_setattr(dentry, iattr)
-> __orangefs_setattr(inode, iattr)

In kernel changes are initiated by calling __orangefs_setattr.

Code path for writeback is

orangefs_write_inode
-> orangefs_inode_setattr

attr_valid and attr_uid and attr_gid change together under i_lock.
I_DIRTY changes separately.

__orangefs_setattr
	lock
	if needs to be cleaned first, unlock and retry
	set attr_valid
	copy data in
	unlock
	mark_inode_dirty

orangefs_inode_setattr
	lock
	copy attributes out
	unlock
	clear getattr_time
	# __writeback_single_inode clears dirty

orangefs_inode_getattr
	# possible to get here with attr_valid set and not dirty
	lock
	if getattr_time ok or attr_valid set, unlock and return
	unlock
	do server operation
	# another thread may getattr or setattr, so check for that
	lock
	if getattr_time ok or attr_valid, unlock and return
	else, copy in
	update getattr_time
	unlock
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

This is a fairly big change, but ultimately it's not a lot of code.

Implement write_inode and then avoid the call to orangefs_inode_setattr
within orangefs_setattr.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

This should be a no-op now.  When inode writeback works, this will
prevent a getattr from overwriting inode data while an inode is
transitioning to dirty.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

This should be a no-op now, but once inode writeback works, it'll be
necessary to have the correct attribute in the dirty inode.

Previously the attribute fetch timeout was marked invalid and the server
provided the updated attribute.  When the inode is dirty, the server
cannot be consulted since it does not yet know the pending setattr.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

No need to store the received mask.  It is either STATX_BASIC_STATS or
STATX_BASIC_STATS & ~STATX_SIZE.  If STATX_SIZE is requested, the cache
is bypassed anyway, so the cached mask is unnecessary to decide whether
to do a real getattr.

This is a change.  Previously a getattr would want size and use the
cached size.  All of the in-kernel callers that wanted size did not want
a cached size.  Now a getattr cannot use the cached size if it wants
size at all.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

When an inode is created, we fetch attributes from the server.  There is
no need to turn around and invalidate them.

No need to initialize attributes after the getattr either.  Either it'll
be exactly the same, or it'll be something else and wrong.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

This uses the same timeout as the getattr cache.  This substantially
increases performance when writing files with smaller buffer sizes.

When writing, the size is (often) changed, which causes a call to
notify_change which calls security_inode_need_killpriv which needs a
getxattr.  Caching it reduces traffic to the server.
Signed-off-by: Martin Brandenburg <martin@omnibond.com>
Signed-off-by: Mike Marshall <hubcap@omnibond.com>

Pull ARM fixes from Russell King:
 "A small number of ARM fixes

   - Fix function tracer and unwinder dependencies so that we don't end
     up building kernels that will crash

   - Fix ARMv7M nommu initialisation (missing register initialisation)

   - Fix EFI decompressor entry (ensuring barrier instructions are
     enabled prior to use)"

* tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm:
  ARM: 8857/1: efi: enable CP15 DMB instructions before cleaning the cache
  ARM: 8856/1: NOMMU: Fix CCR register faulty initialization when MPU is disabled
  ARM: fix function graph tracer and unwinder dependencies

Pull powerpc fixes from Michael Ellerman:
 "A one-liner to make our Radix MMU support depend on HUGETLB_PAGE. We
  use some of the hugetlb inlines (eg. pud_huge()) when operating on the
  linear mapping and if they're compiled into empty wrappers we can
  corrupt memory.

  Then two fixes to our VFIO IOMMU code. The first is not a regression
  but fixes the locking to avoid a user-triggerable deadlock.

  The second does fix a regression since rc1, and depends on the first
  fix. It makes it possible to run guests with large amounts of memory
  again (~256GB).

  Thanks to Alexey Kardashevskiy"

* tag 'powerpc-5.1-6' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/mm_iommu: Allow pinning large regions
  powerpc/mm_iommu: Fix potential deadlock
  powerpc/mm/radix: Make Radix require HUGETLB_PAGE

Pull block fixes from Jens Axboe:
 "A set of io_uring fixes that should go into this release. In
  particular, this contains:

   - The mutex lock vs ctx ref count fix (me)

   - Removal of a dead variable (me)

   - Two race fixes (Stefan)

   - Ring head/tail condition fix for poll full SQ detection (Stefan)"

* tag 'for-linus-20190428' of git://git.kernel.dk/linux-block:
  io_uring: remove 'state' argument from io_{read,write} path
  io_uring: fix poll full SQ detection
  io_uring: fix race condition when sq threads goes sleeping
  io_uring: fix race condition reading SQ entries
  io_uring: fail io_uring_register(2) on a dying io_uring instance

Pull rdma fixes from Jason Gunthorpe:
 "One core bug fix and a few driver ones

   - FRWR memory registration for hfi1/qib didn't work with with some
     iovas causing a NFSoRDMA failure regression due to a fix in the NFS
     side

   - A command flow error in mlx5 allowed user space to send a corrupt
     command (and also smash the kernel stack we've since learned)

   - Fix a regression and some bugs with device hot unplug that was
     discovered while reviewing Andrea's patches

   - hns has a failure if the user asks for certain QP configurations"

* tag 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma:
  RDMA/hns: Bugfix for mapping user db
  RDMA/ucontext: Fix regression with disassociate
  RDMA/mlx5: Use rdma_user_map_io for mapping BAR pages
  RDMA/mlx5: Do not allow the user to write to the clock page
  IB/mlx5: Fix scatter to CQE in DCT QP creation
  IB/rdmavt: Fix frwr memory registration

Pull dmaengine fixes from Vinod Koul:

 - fix for wrong register use in mediatek driver

 - fix in sh driver for glitch is tx_status and treating 0 a valid
   residue for cyclic

 - fix in bcm driver for using right memory allocation flag

* tag 'dmaengine-fix-5.1-rc7' of git://git.infradead.org/users/vkoul/slave-dma:
  dmaengine: mediatek-cqdma: fix wrong register usage in mtk_cqdma_start
  dmaengine: sh: rcar-dmac: Fix glitch in dmaengine_tx_status
  dmaengine: sh: rcar-dmac: With cyclic DMA residue 0 is valid
  dmaengine: bcm2835: Avoid GFP_KERNEL in device_prep_slave_sg

Pull input fixes from Dmitry Torokhov:
 "Just a couple of fixups for Synaptics RMI4 driver and allowing
  snvs_pwrkey to be selected on more boards"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
  Input: synaptics-rmi4 - write config register values to the right offset
  Input: synaptics-rmi4 - fix possible double free
  Input: snvs_pwrkey - make it depend on ARCH_MXC

Pull x86 fixes from Ingo Molnar:

 - Fix an early boot crash in the RSDP parsing code by effectively
   turning off the parsing call - we ran out of time but want to fix the
   regression. The more involved fix is being worked on.

 - Fix a crash that can trigger in the kmemlek code.

* 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/mm: Fix a crash with kmemleak_scan()
  x86/boot: Disable RSDP parsing temporarily

Pull scheduler fix from Ingo Molnar:
 "Fix a division by zero bug that can trigger in the NUMA placement
  code"

* 'sched-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sched/numa: Fix a possible divide-by-zero

Pull perf fix from Ingo Molnar:
 "A cstate event enumeration fix for Kaby/Coffee Lake CPUs"

* 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/intel: Update KBL Package C-state events to also include PC8/PC9/PC10 counters

This way, slhc_free() accepts what slhc_init() returns, whether that is
an error or not.

In particular, the pattern in sl_alloc_bufs() is

        slcomp = slhc_init(16, 16);
        ...
        slhc_free(slcomp);

for the error handling path, and rather than complicate that code, just
make it ok to always free what was returned by the init function.

That's what the code used to do before commit 4ab42d78 ("ppp, slip:
Validate VJ compression slot parameters completely") when slhc_init()
just returned NULL for the error case, with no actual indication of the
details of the error.

Reported-by: syzbot+45474c076a4927533d2e@syzkaller.appspotmail.com
Fixes: 4ab42d78 ("ppp, slip: Validate VJ compression slot parameters completely")
Acked-by: Ben Hutchings <ben@decadent.org.uk>
Cc: David Miller <davem@davemloft.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

Merge misc fixes from Andrew Morton:
 "9 fixes"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  fs/proc/proc_sysctl.c: Fix a NULL pointer dereference
  mm/page_alloc.c: fix never set ALLOC_NOFRAGMENT flag
  mm/page_alloc.c: avoid potential NULL pointer dereference
  mm, page_alloc: always use a captured page regardless of compaction result
  mm: do not boost watermarks to avoid fragmentation for the DISCONTIG memory model
  lib/test_vmalloc.c: do not create cpumask_t variable on stack
  lib/Kconfig.debug: fix build error without CONFIG_BLOCK
  zram: pass down the bvec we need to read into in the work struct
  mm/memory_hotplug.c: drop memory device reference after find_memory_block()

Currently any changed config register values don't take effect, as the
function to write them back is called with the wrong register offset.

Fixes: ff8f8370 (Input: synaptics-rmi4 - add support for 2D
                     sensors and F11)
Signed-off-by: Lucas Stach <l.stach@pengutronix.de>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Cc: stable@vger.kernel.org
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>

Pull arm64 fixes from Catalin Marinas:

 - keep the tail of an unaligned initrd reserved

 - adjust ftrace_make_call() to deal with the relative nature of PLTs

* tag 'arm64-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux:
  arm64/module: ftrace: deal with place relative nature of PLTs
  arm64: mm: Ensure tail of unaligned initrd is reserved

Pull tracing fixes from Steven Rostedt:
 "Three tracing fixes:

   - Use "nosteal" for ring buffer splice pages

   - Memory leak fix in error path of trace_pid_write()

   - Fix preempt_enable_no_resched() (use preempt_enable()) in ring
     buffer code"

* tag 'trace-v5.1-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
  trace: Fix preempt_enable_no_resched() abuse
  tracing: Fix a memory leak by early error exit in trace_pid_write()
  tracing: Fix buffer_ref pipe ops

Pull GPIO fixes from Linus Walleij:
 "Not much to say about them, regular fixes:

   - Fix a bug on the errorpath of gpiochip_add_data_with_key()

   - IRQ type setting on the spreadtrum GPIO driver"

* tag 'gpio-v5.1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-gpio:
  gpio: Fix gpiochip_add_data_with_key() error path
  gpio: eic: sprd: Fix incorrect irq type setting for the sync EIC

Pull drm fixes from Dave Airlie:
 "Regular drm fixes, nothing too outstanding, I'm guessing Easter was
  slowing people down.

  i915:
   - FEC enable fix
   - BXT display lanes fix

  ttm:
   - fix reinit for reloading drivers regression

  imx:
   - DP CSC fix

  sun4i:
   - module unload/load fix

  vc4:
   - memory leak fix
   - compile fix

  dw-hdmi:
   - rockchip scdc overflow fix

  sched:
   - docs fix

  vmwgfx:
   - dma api layering fix"

* tag 'drm-fixes-2019-04-26' of git://anongit.freedesktop.org/drm/drm:
  drm/bridge: dw-hdmi: fix SCDC configuration for ddc-i2c-bus
  drm/vmwgfx: Fix dma API layer violation
  drm/vc4: Fix compilation error reported by kbuild test bot
  drm/sun4i: Unbind components before releasing DRM and memory
  drm/vc4: Fix memory leak during gpu reset.
  drm/sched: Fix description of drm_sched_stop
  drm/imx: don't skip DP channel disable for background plane
  gpu: ipu-v3: dp: fix CSC handling
  drm/ttm: fix re-init of global structures
  drm/sun4i: Fix component unbinding and component master deletion
  drm/sun4i: Set device driver data at bind time for use in unbind
  drm/sun4i: Add missing drm_atomic_helper_shutdown at driver unbind
  drm/i915: Restore correct bxt_ddi_phy_calc_lane_lat_optim_mask() calculation
  drm/i915: Do not enable FEC without DSC
  drm: bridge: dw-hdmi: Fix overflow workaround for Rockchip SoCs

Pull btrfs fix from David Sterba:
 "One patch to fix a crash in io submission path, due to memory
  allocation errors.

  In short, the multipage bio work that landed in 5.1 caused larger bios
  that in turn require larger temporary memory for checksums. The patch
  is a workaround, we're going to rework the allocation so it does not
  require the vmalloc fallback.

  It took a while to identify that it's caused by patches in 5.1 and not
  a patchset that did some changes in error handling in the code. I've
  tested it on various memory/cpu combinations, it could hit OOM but
  does not crash.

  The timestamp of the patch is less than a day due to updates in the
  changelog, tests were running meanwhile"

* tag 'for-5.1-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: Switch memory allocations in async csum calculation path to kvmalloc