Because it's a mess.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Now that we've internalised the two-phase inode allocation, we can
now easily make the AG selection and allocation atomic from the
perspective of a single perag context. This will ensure AGs going
offline/away cannot occur between the selection and allocation
steps.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

This is just a simple wrapper around the per-ag inode allocation
that doesn't need to exist. The internal mechanism to select and
allocate within an AG does not need to be exposed outside
xfs_ialloc.c, and it being exposed simply makes it harder to follow
the code and simplify it.

This is simplified by internalising xf_dialloc_select_ag() and
xfs_dialloc_ag() into a single xfs_dialloc() function and then
xfs_dir_ialloc() can go away.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

xfs_dialloc_select_ag() does a lot of repetitive work. It first
calls xfs_ialloc_ag_select() to select the AG to start allocation
attempts in, which can do up to two entire loops across the perags
that inodes can be allocated in. This is simply checking if there is
spce available to allocate inodes in an AG, and it returns when it
finds the first candidate AG.

xfs_dialloc_select_ag() then does it's own iterative walk across
all the perags locking the AGIs and trying to allocate inodes from
the locked AG. It also doesn't limit the search to mp->m_maxagi,
so it will walk all AGs whether they can allocate inodes or not.

Hence if we are really low on inodes, we could do almost 3 entire
walks across the whole perag range before we find an allocation
group we can allocate inodes in or report ENOSPC.

Because xfs_ialloc_ag_select() returns on the first candidate AG it
finds, we can simply do these checks directly in
xfs_dialloc_select_ag() before we lock and try to allocate inodes.
This reduces the inode allocation pass down to 2 perag sweeps at
most - one for aligned inode cluster allocation and if we can't
allocate full, aligned inode clusters anywhere we'll do another pass
trying to do sparse inode cluster allocation.

This also removes a big chunk of duplicate code.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

The only caller of xfs_dialloc_select_ag() will always return
-ENOSPC to it's caller if the agbp returned from
xfs_dialloc_select_ag() is NULL. IOWs, failure to find a candidate
AGI we can allocate inodes from is always an ENOSPC condition, so
move this logic up into xfs_dialloc_select_ag() so we can simplify
the return logic in this function.

xfs_dialloc_select_ag() now only ever returns 0 with a locked
agbp, or an error with no agbp.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Now that everything passes a perag, the agno is not needed anymore.
Convert all the users to use pag->pag_agno instead and remove the
agno from the cursor. This was largely done as an automated search
and replace.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Which will eventually completely replace the agno in it.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>

Needs a [from, to] ranged AG walk, and the perag to be stuffed into
the info structure for callouts to use.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

We currently pass an agno from the AG reservation functions to the
individual feature accounting functions, which in future may have to
do perag lookups to access per-AG state. Instead, pre-emptively
plumb the perag through from the highest AG reservation layer to the
feature callouts so they won't have to look it up again.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Brian Foster <bfoster@redhat.com>

All of the callers of the busy extent API either have perag
references available to use so we can pass a perag to the busy
extent functions rather than having them have to do unnecessary
lookups.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Clean up the last external manual AG walk.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Rather than manually walking the ags and passing agnunbers around,
pass the perag for the AG we are currently working on around in the
iwalk structure.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Convert the raw walks to an iterator, pulling the current AG out of
pag->pag_agno instead of the loop iterator variable.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

for_each_perag_tag() is defined in xfs_icache.c for local use.
Promote this to xfs_ag.h and define equivalent iteration functions
so that we can use them to iterate AGs instead to replace open coded
perag walks and perag lookups.

We also convert as many of the straight forward open coded AG walks
to use these iterators as possible. Anything that is not a direct
conversion to an iterator is ignored and will be updated in future
commits.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Move the xfs_perag infrastructure to the libxfs files that contain
all the per AG infrastructure. This helps set up for passing perags
around all the code instead of bare agnos with minimal extra
includes for existing files.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

The perag structures really need to be defined with the rest of the
AG support infrastructure. The struct xfs_perag and init/teardown
has been placed in xfs_mount.[ch] because there are differences in
the structure between kernel and userspace. Mainly that userspace
doesn't have a lot of the internal stuff that the kernel has for
caches and discard and other such structures.

However, it makes more sense to move this to libxfs than to keep
this separation because we are now moving to use struct perags
everywhere in the code instead of passing raw agnumber_t values
about. Hence we shoudl really move the support infrastructure to
libxfs/xfs_ag.[ch].

To do this without breaking userspace, first we need to rearrange
the structures and code so that all the kernel specific code is
located together. This makes it simple for userspace to ifdef out
the all the parts it does not need, minimising the code differences
between kernel and userspace. The next commit will do the move...
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

They are AG functions, not superblock functions, so move them to the
appropriate location.
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>

Merge tag 'driver-core-5.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core

Pull driver core fixes from Greg KH:
 "Here are two driver fixes for driver core changes that happened in
  5.13-rc1.

  The clk driver fix resolves a many-reported issue with booting some
  devices, and the USB typec fix resolves the reported problem of USB
  systems on some embedded boards.

  Both of these have been in linux-next this week with no reported
  issues"

* tag 'driver-core-5.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  clk: Skip clk provider registration when np is NULL
  usb: typec: tcpm: Don't block probing of consumers of "connector" nodes

Pull staging and IIO driver fixes from Greg KH:
 "Here are some small IIO driver fixes and one Staging driver fix for
  5.13-rc2.

  Nothing major, just some resolutions for reported problems:

   - gcc-11 bogus warning fix for rtl8723bs

   - iio driver tiny fixes

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

* tag 'staging-5.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  iio: tsl2583: Fix division by a zero lux_val
  iio: core: return ENODEV if ioctl is unknown
  iio: core: fix ioctl handlers removal
  iio: gyro: mpu3050: Fix reported temperature value
  iio: hid-sensors: select IIO_TRIGGERED_BUFFER under HID_SENSOR_IIO_TRIGGER
  iio: proximity: pulsedlight: Fix rumtime PM imbalance on error
  iio: light: gp2ap002: Fix rumtime PM imbalance on error
  staging: rtl8723bs: avoid bogus gcc warning

Pull USB fixes from Greg KH:
 "Here are some small USB fixes for 5.13-rc2. They consist of a number
  of resolutions for reported issues:

   - typec fixes for found problems

   - xhci fixes and quirk additions

   - dwc3 driver fixes

   - minor fixes found by Coverity

   - cdc-wdm fixes for reported problems

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

* tag 'usb-5.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb: (28 commits)
  usb: core: hub: fix race condition about TRSMRCY of resume
  usb: typec: tcpm: Fix SINK_DISCOVERY current limit for Rp-default
  xhci: Add reset resume quirk for AMD xhci controller.
  usb: xhci: Increase timeout for HC halt
  xhci: Do not use GFP_KERNEL in (potentially) atomic context
  xhci: Fix giving back cancelled URBs even if halted endpoint can't reset
  xhci-pci: Allow host runtime PM as default for Intel Alder Lake xHCI
  usb: musb: Fix an error message
  usb: typec: tcpm: Fix wrong handling for Not_Supported in VDM AMS
  usb: typec: tcpm: Send DISCOVER_IDENTITY from dedicated work
  usb: typec: ucsi: Retrieve all the PDOs instead of just the first 4
  usb: fotg210-hcd: Fix an error message
  docs: usb: function: Modify path name
  usb: dwc3: omap: improve extcon initialization
  usb: typec: ucsi: Put fwnode in any case during ->probe()
  usb: typec: tcpm: Fix wrong handling in GET_SINK_CAP
  usb: dwc2: Remove obsolete MODULE_ constants from platform.c
  usb: dwc3: imx8mp: fix error return code in dwc3_imx8mp_probe()
  usb: dwc3: imx8mp: detect dwc3 core node via compatible string
  usb: dwc3: gadget: Return success always for kick transfer in ep queue
  ...

Pull timer fixes from Thomas Gleixner:
 "Two fixes for timers:

   - Use the ALARM feature check in the alarmtimer core code insted of
     the old method of checking for the set_alarm() callback.

     Drivers can have that callback set but the feature bit cleared. If
     such a RTC device is selected then alarms wont work.

   - Use a proper define to let the preprocessor check whether Hyper-V
     VDSO clocksource should be active.

     The code used a constant in an enum with #ifdef, which evaluates to
     always false and disabled the clocksource for VDSO"

* tag 'timers-urgent-2021-05-16' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  clocksource/drivers/hyper-v: Re-enable VDSO_CLOCKMODE_HVCLOCK on X86
  alarmtimer: Check RTC features instead of ops

Pull xen fixes from Juergen Gross:

 - two patches for error path fixes

 - a small series for fixing a regression with swiotlb with Xen on Arm

* tag 'for-linus-5.13b-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
  xen/swiotlb: check if the swiotlb has already been initialized
  arm64: do not set SWIOTLB_NO_FORCE when swiotlb is required
  xen/arm: move xen_swiotlb_detect to arm/swiotlb-xen.h
  xen/unpopulated-alloc: fix error return code in fill_list()
  xen/gntdev: fix gntdev_mmap() error exit path

Pull x86 fixes from Borislav Petkov:
 "The three SEV commits are not really urgent material. But we figured
  since getting them in now will avoid a huge amount of conflicts
  between future SEV changes touching tip, the kvm and probably other
  trees, sending them to you now would be best.

  The idea is that the tip, kvm etc branches for 5.14 will all base
  ontop of -rc2 and thus everything will be peachy. What is more, those
  changes are purely mechanical and defines movement so they should be
  fine to go now (famous last words).

  Summary:

   - Enable -Wundef for the compressed kernel build stage

   - Reorganize SEV code to streamline and simplify future development"

* tag 'x86_urgent_for_v5.13_rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/boot/compressed: Enable -Wundef
  x86/msr: Rename MSR_K8_SYSCFG to MSR_AMD64_SYSCFG
  x86/sev: Move GHCB MSR protocol and NAE definitions in a common header
  x86/sev-es: Rename sev-es.{ch} to sev.{ch}

Pull powerpc fixes from Michael Ellerman:

 - Fix a regression in the conversion of the 64-bit BookE interrupt
   entry to C.

 - Fix KVM hosts running with the hash MMU since the recent KVM gfn
   changes.

 - Fix a deadlock in our paravirt spinlocks when hcall tracing is
   enabled.

 - Several fixes for oopses in our runtime code patching for security
   mitigations.

 - A couple of minor fixes for the recent conversion of 32-bit interrupt
   entry/exit to C.

 - Fix __get_user() causing spurious crashes in sigreturn due to a bad
   inline asm constraint, spotted with GCC 11.

 - A fix for the way we track IRQ masking state vs NMI interrupts when
   using the new scv system call entry path.

 - A couple more minor fixes.

Thanks to Cédric Le Goater, Christian Zigotzky, Christophe Leroy,
Naveen N. Rao, Nicholas Piggin Paul Menzel, and Sean Christopherson.

* tag 'powerpc-5.13-3' of git://git.kernel.org/pub/scm/linux/kernel/git/powerpc/linux:
  powerpc/64e/interrupt: Fix nvgprs being clobbered
  powerpc/64s: Make NMI record implicitly soft-masked code as irqs disabled
  powerpc/64s: Fix stf mitigation patching w/strict RWX & hash
  powerpc/64s: Fix entry flush patching w/strict RWX & hash
  powerpc/64s: Fix crashes when toggling entry flush barrier
  powerpc/64s: Fix crashes when toggling stf barrier
  KVM: PPC: Book3S HV: Fix kvm_unmap_gfn_range_hv() for Hash MMU
  powerpc/legacy_serial: Fix UBSAN: array-index-out-of-bounds
  powerpc/signal: Fix possible build failure with unsafe_copy_fpr_{to/from}_user
  powerpc/uaccess: Fix __get_user() with CONFIG_CC_HAS_ASM_GOTO_OUTPUT
  powerpc/pseries: warn if recursing into the hcall tracing code
  powerpc/pseries: use notrace hcall variant for H_CEDE idle
  powerpc/pseries: Don't trace hcall tracing wrapper
  powerpc/pseries: Fix hcall tracing recursion in pv queued spinlocks
  powerpc/syscall: Calling kuap_save_and_lock() is wrong
  powerpc/interrupts: Fix kuep_unlock() call

Pull scheduler fixes from Ingo Molnar:
 "Fix an idle CPU selection bug, and an AMD Ryzen maximum frequency
  enumeration bug"

* tag 'sched-urgent-2021-05-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86, sched: Fix the AMD CPPC maximum performance value on certain AMD Ryzen generations
  sched/fair: Fix clearing of has_idle_cores flag in select_idle_cpu()

Pull objtool fixes from Ingo Molnar:
 "Fix a couple of endianness bugs that crept in"

* tag 'objtool-urgent-2021-05-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  objtool/x86: Fix elf_add_alternative() endianness
  objtool: Fix elf_create_undef_symbol() endianness

Pull irq fix from Ingo Molnar:
 "Fix build warning on SH"

* tag 'irq-urgent-2021-05-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  sh: Remove unused variable

Pull x86 stack randomization fix from Ingo Molnar:
 "Fix an assembly constraint that affected LLVM up to version 12"

* tag 'core-urgent-2021-05-15' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  stack: Replace "o" output with "r" input constraint

Merge misc fixes from Andrew Morton:
 "13 patches.

  Subsystems affected by this patch series: resource, squashfs, hfsplus,
  modprobe, and mm (hugetlb, slub, userfaultfd, ksm, pagealloc, kasan,
  pagemap, and ioremap)"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  mm/ioremap: fix iomap_max_page_shift
  docs: admin-guide: update description for kernel.modprobe sysctl
  hfsplus: prevent corruption in shrinking truncate
  mm/filemap: fix readahead return types
  kasan: fix unit tests with CONFIG_UBSAN_LOCAL_BOUNDS enabled
  mm: fix struct page layout on 32-bit systems
  ksm: revert "use GET_KSM_PAGE_NOLOCK to get ksm page in remove_rmap_item_from_tree()"
  userfaultfd: release page in error path to avoid BUG_ON
  squashfs: fix divide error in calculate_skip()
  kernel/resource: fix return code check in __request_free_mem_region
  mm, slub: move slub_debug static key enabling outside slab_mutex
  mm/hugetlb: fix cow where page writtable in child
  mm/hugetlb: fix F_SEAL_FUTURE_WRITE

Pull ARC fixes from Vineet Gupta:

 - PAE fixes

 - syscall num check off-by-one bug

 - misc fixes

* tag 'arc-5.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/vgupta/arc:
  ARC: mm: Use max_high_pfn as a HIGHMEM zone border
  ARC: mm: PAE: use 40-bit physical page mask
  ARC: entry: fix off-by-one error in syscall number validation
  ARC: kgdb: add 'fallthrough' to prevent a warning
  arc: Fix typos/spellos

Pull block fixes from Jens Axboe:

 - Fix for shared tag set exit (Bart)

 - Correct ioctl range for zoned ioctls (Damien)

 - Removed dead/unused function (Lin)

 - Fix perf regression for shared tags (Ming)

 - Fix out-of-bounds issue with kyber and preemption (Omar)

 - BFQ merge fix (Paolo)

 - Two error handling fixes for nbd (Sun)

 - Fix weight update in blk-iocost (Tejun)

 - NVMe pull request (Christoph):
      - correct the check for using the inline bio in nvmet (Chaitanya
        Kulkarni)
      - demote unsupported command warnings (Chaitanya Kulkarni)
      - fix corruption due to double initializing ANA state (me, Hou Pu)
      - reset ns->file when open fails (Daniel Wagner)
      - fix a NULL deref when SEND is completed with error in nvmet-rdma
        (Michal Kalderon)

 - Fix kernel-doc warning (Bart)

* tag 'block-5.13-2021-05-14' of git://git.kernel.dk/linux-block:
  block/partitions/efi.c: Fix the efi_partition() kernel-doc header
  blk-mq: Swap two calls in blk_mq_exit_queue()
  blk-mq: plug request for shared sbitmap
  nvmet: use new ana_log_size instead the old one
  nvmet: seset ns->file when open fails
  nbd: share nbd_put and return by goto put_nbd
  nbd: Fix NULL pointer in flush_workqueue
  blkdev.h: remove unused codes blk_account_rq
  block, bfq: avoid circular stable merges
  blk-iocost: fix weight updates of inner active iocgs
  nvmet: demote fabrics cmd parse err msg to debug
  nvmet: use helper to remove the duplicate code
  nvmet: demote discovery cmd parse err msg to debug
  nvmet-rdma: Fix NULL deref when SEND is completed with error
  nvmet: fix inline bio check for passthru
  nvmet: fix inline bio check for bdev-ns
  nvme-multipath: fix double initialization of ANA state
  kyber: fix out of bounds access when preempted
  block: uapi: fix comment about block device ioctl

Pull io_uring fixes from Jens Axboe:
 "Just a few minor fixes/changes:

   - Fix issue with double free race for linked timeout completions

   - Fix reference issue with timeouts

   - Remove last few places that make SQPOLL special, since it's just an
     io thread now.

   - Bump maximum allowed registered buffers, as we don't allocate as
     much anymore"

* tag 'io_uring-5.13-2021-05-14' of git://git.kernel.dk/linux-block:
  io_uring: increase max number of reg buffers
  io_uring: further remove sqpoll limits on opcodes
  io_uring: fix ltout double free on completion race
  io_uring: fix link timeout refs

Pull erofs fixes from Gao Xiang:
 "This mainly fixes 1 lcluster-sized pclusters for the big pcluster
  feature, which can be forcely generated by mkfs as a specific on-disk
  case for per-(sub)file compression strategies but missed to handle in
  runtime properly.

  Also, documentation updates are included to fix the broken
  illustration due to the ReST conversion by accident and complete the
  big pcluster introduction.

  Summary:

   - update documentation to fix the broken illustration due to ReST
     conversion by accident at that time and complete the big pcluster
     introduction

   - fix 1 lcluster-sized pclusters for the big pcluster feature"

* tag 'erofs-for-5.13-rc2-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs:
  erofs: fix 1 lcluster-sized pcluster for big pcluster
  erofs: update documentation about data compression
  erofs: fix broken illustration in documentation

Pull libnvdimm fixes from Dan Williams:
 "A regression fix for a bootup crash condition introduced in this merge
  window and some other minor fixups:

   - Fix regression in ACPI NFIT table handling leading to crashes and
     driver load failures.

   - Move the nvdimm mailing list

   - Miscellaneous minor fixups"

* tag 'libnvdimm-fixes-5.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  ACPI: NFIT: Fix support for variable 'SPA' structure size
  MAINTAINERS: Move nvdimm mailing list
  tools/testing/nvdimm: Make symbol '__nfit_test_ioremap' static
  libnvdimm: Remove duplicate struct declaration

Pull dax fixes from Dan Williams:
 "A fix for a hang condition due to missed wakeups in the filesystem-dax
  core when exercised by virtiofs.

  This bug has been there from the beginning, but the condition has
  not triggered on other filesystems since they hold a lock over
  invalidation events"

* tag 'dax-fixes-5.13-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/nvdimm/nvdimm:
  dax: Wake up all waiters after invalidating dax entry
  dax: Add a wakeup mode parameter to put_unlocked_entry()
  dax: Add an enum for specifying dax wakup mode