cifs: allocate buffer in the caller of build_path_from_dentry()

build_path_from_dentry() open-codes dentry_path_raw().  The reason
we can't use dentry_path_raw() in there (and postprocess the
result as needed) is that the callers of build_path_from_dentry()
expect that the object to be freed on cleanup and the string to
be used are at the same address.  That's painful, since the path
is naturally built end-to-beginning - we start at the leaf and
go through the ancestors, accumulating the pathname.

Life would be easier if we left the buffer allocation to callers.
It wouldn't be exact-sized buffer, but none of the callers keep
the result for long - it's always freed before the caller returns.
So there's no need to do exact-sized allocation; better use
__getname()/__putname(), same as we do for pathname arguments
of syscalls.  What's more, there's no need to do allocation under
spinlocks, so GFP_ATOMIC is not needed.

Next patch will replace the open-coded dentry_path_raw() (in
build_path_from_dentry_optional_prefix()) with calling the real
thing.  This patch only introduces wrappers for allocating/freeing
the buffers and switches to new calling conventions:
build_path_from_dentry(dentry, buf)
expects buf to be address of a page-sized object or NULL,
return value is a pathname built inside that buffer on success,
ERR_PTR(-ENOMEM) if buf is NULL and ERR_PTR(-ENAMETOOLONG) if
the pathname won't fit into page.  Note that we don't need to
check for failure when allocating the buffer in the caller -
build_path_from_dentry() will do the right thing.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: make build_path_from_dentry() return const char *

... and adjust the callers.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: constify pathname arguments in a bunch of helpers

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: constify path argument of ->make_node()

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: constify get_normalized_path() properly

As it is, it takes const char * and, in some cases, stores it in
caller's variable that is plain char *.  Fortunately, none of the
callers actually proceeded to modify the string via now-non-const
alias, but that's trouble waiting to happen.

It's easy to do properly, anyway...

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: don't cargo-cult strndup()

strndup(s, strlen(s)) is a highly unidiomatic way to spell strdup(s);
it's *NOT* safer in any way, since strlen() is just as sensitive to
NUL-termination as strdup() is.

strndup() is for situations when you need a copy of a known-sized
substring, not a magic security juju to drive the bad spirits away.

Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Steve French <stfrench@microsoft.com>

SMB3: update structures for new compression protocol definitions

Protocol has been extended for additional compression headers.
See MS-SMB2 section 2.2.42

Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: remove old dead code

While reviewing a patch clarifying locks and locking hierarchy I
realized some locks were unused.

This commit removes old data and code that isn't actually used
anywhere, or hidden in ifdefs which cannot be enabled from the kernel
config.

* The uid/gid trees and associated locks are left-overs from when
  uid/sid mapping had an extra caching layer on top of the keyring and
  are now unused.
  See commit faa65f07d21e ("cifs: simplify id_to_sid and sid_to_id mapping code")
  from 2012.

* cifs_oplock_break_ops is a left-over from when slow_work was remplaced
  by regular workqueue and is now unused.
  See commit 9b646972467f ("cifs: use workqueue instead of slow-work")
  from 2010.

* CIFSSMBSetAttrLegacy is SMB1 cruft dealing with some legacy
  NT4/Win9x behaviour.

* Remove CONFIG_CIFS_DNOTIFY_EXPERIMENTAL left-overs. This was already
  partially removed in 392e1c5dc9cc ("cifs: rename and clarify CIFS_ASYNC_OP and CIFS_NO_RESP")
  from 2019. Kill it completely.

* Another candidate that was considered but spared is
  CONFIG_CIFS_NFSD_EXPORT which has an empty implementation and cannot
  be enabled by a config option (although it is listed but disabled with
  "BROKEN" as a dep). It's unclear whether this could even function
  today in its current form but it has it's own .c file and Kconfig
  entry which is a bit more involved to remove and might make a come
  back?

Signed-off-by: Aurelien Aptel <aaptel@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: cifspdu.h: Replace one-element array with flexible-array member

There is a regular need in the kernel to provide a way to declare having
a dynamically sized set of trailing elements in a structure. Kernel code
should always use “flexible array members”[1] for these cases. The older
style of one-element or zero-length arrays should no longer be used[2].

Also, this helps with the ongoing efforts to enable -Warray-bounds by
fixing the following warning:

  CC [M]  fs/cifs/cifssmb.o
fs/cifs/cifssmb.c: In function ‘CIFSFindNext’:
fs/cifs/cifssmb.c:4636:23: warning: array subscript 1 is above array bounds of ‘char[1]’ [-Warray-bounds]
 4636 |   pSMB->ResumeFileName[name_len+1] = 0;
      |   ~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~

[1] https://en.wikipedia.org/wiki/Flexible_array_member
[2] https://www.kernel.org/doc/html/v5.10/process/deprecated.html#zero-length-and-one-element-arrays

Link: https://github.com/KSPP/linux/issues/79
Link: https://github.com/KSPP/linux/issues/109
Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Reviewed-by: Aurelien Aptel <aaptel@suse.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

fs: cifs: Remove repeated struct declaration

struct cifs_writedata is declared twice.
One is declared at 209th line.
And struct cifs_writedata is defined blew.
The declaration hear is not needed. Remove the duplicate.

Signed-off-by: Wan Jiabing <wanjiabing@vivo.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

Documentation/admin-guide/cifs: document open_files and dfscache

Add missing documentation for open_files and dfscache /proc files.

Signed-off-by: Aurelien Aptel <aaptel@suse.com>
Reviewed-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: simplify SWN code with dummy funcs instead of ifdefs

This commit doesn't change the logic of SWN.

Add dummy implementation of SWN functions when SWN is disabled instead
of using ifdef sections.

The dummy functions get optimized out, this leads to clearer code and
compile time type-checking regardless of config options with no
runtime penalty.

Leave the simple ifdefs section as-is.

A single bitfield (bool foo:1) on its own will use up one int. Move
tcon->use_witness out of ifdefs with the other tcon bitfields.

Signed-off-by: Aurelien Aptel <aaptel@suse.com>
Reviewed-by: Samuel Cabrero <scabrero@suse.de>
Signed-off-by: Steve French <stfrench@microsoft.com>

smb3: update protocol header definitions based to include new flags

[MS-SMB2] protocol specification was recently updated to include
new flags, new negotiate context and some minor changes to fields.
Update smb2pdu.h structure definitions to match the newest version
of the protocol specification.  Updates to the compression context
values will be in a followon patch.

Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: correct comments explaining internal semaphore usage in the module

A few of the semaphores had been removed, and one additional one
needed to be noted in the comments.

Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: Remove useless variable

Fix the following gcc warning:

fs/cifs/cifsacl.c:1097:8: warning: variable ‘nmode’ set but not used
[-Wunused-but-set-variable].

Reported-by: Abaci Robot <abaci@linux.alibaba.com>
Signed-off-by: Jiapeng Chong <jiapeng.chong@linux.alibaba.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

cifs: Fix spelling of 'security'

secuirty -> security

Signed-off-by: jack1.li_cp <liliu1@yulong.com>
Signed-off-by: Steve French <stfrench@microsoft.com>

Linux 5.12

Merge tag 'perf-tools-fixes-for-v5.12-2021-04-25' of git://git./linux/kernel/git/acme/linux

Pull perf tools fixes from Arnaldo Carvalho de Melo:

 - Fix potential NULL pointer dereference in the auxtrace option parser

 - Fix access to PID in an array when setting a PID filter in 'perf ftrace'

 - Fix error return code in the 'perf data' tool and in maps__clone(),
   found using a static analysis tool from Huawei

* tag 'perf-tools-fixes-for-v5.12-2021-04-25' of git://git.kernel.org/pub/scm/linux/kernel/git/acme/linux:
  perf map: Fix error return code in maps__clone()
  perf ftrace: Fix access to pid in array when setting a pid filter
  perf auxtrace: Fix potential NULL pointer dereference
  perf data: Fix error return code in perf_data__create_dir()

Merge tag 'perf_urgent_for_v5.12' of git://git./linux/kernel/git/tip/tip

Pull x86 perf fixes from Borislav Petkov:

 - Fix Broadwell Xeon's stepping in the PEBS isolation table of CPUs

 - Fix a panic when initializing perf uncore machinery on Haswell and
   Broadwell servers

* tag 'perf_urgent_for_v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  perf/x86/kvm: Fix Broadwell Xeon stepping in isolation_ucodes[]
  perf/x86/intel/uncore: Remove uncore extra PCI dev HSWEP_PCI_PCU_3

Merge tag 'locking_urgent_for_v5.12' of git://git./linux/kernel/git/tip/tip

Pull locking fix from Borislav Petkov:
 "Fix ordering in the queued writer lock's slowpath"

* tag 'locking_urgent_for_v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  locking/qrwlock: Fix ordering in queued_write_lock_slowpath()

Merge tag 'sched_urgent_for_v5.12' of git://git./linux/kernel/git/tip/tip

Pull scheduler fix from Borislav Petkov:
 "Fix a typo in a macro ifdeffery"

* tag 'sched_urgent_for_v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  preempt/dynamic: Fix typo in macro conditional statement

Merge tag 'x86_urgent_for_v5.12' of git://git./linux/kernel/git/tip/tip

Pull x86 fix from Borislav Petkov:
 "Fix an out-of-bounds memory access when setting up a crash kernel with
  kexec"

* tag 'x86_urgent_for_v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/crash: Fix crash_setup_memmap_entries() out-of-bounds access

Merge tag 'irqchip-5.13' of git://git./linux/kernel/git/maz/arm-platforms into irq/core

Pull irqchip and irqdomain updates from Marc Zyngier:

 New HW support:

  - New driver for the Nuvoton WPCM450 interrupt controller
  - New driver for the IDT 79rc3243x interrupt controller
  - Add support for interrupt trigger configuration to the MStar irqchip
  - Add more external interrupt support to the STM32 irqchip
  - Add new compatible strings for QCOM SC7280 to the qcom-pdc binding

 Fixes and cleanups:

  - Drop irq_create_strict_mappings() and irq_create_identity_mapping()
    from the irqdomain API, with cleanups in a couple of drivers
  - Fix nested NMI issue with spurious interrupts on GICv3
  - Don't allow GICv4.1 vSGIs when the CPU doesn't support them
  - Various cleanups and minor fixes

Link: https://lore.kernel.org/r/20210424094640.1731920-1-maz@kernel.org

Merge tag 'for-linus' of git://git./virt/kvm/kvm

Pull kvm fix from Paolo Bonzini:
 "Fix SRCU bug introduced in the merge window"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: x86/xen: Take srcu lock when accessing kvm_memslots()

Revert "net/rds: Avoid potential use after free in rds_send_remove_from_sock"

This reverts commit 0c85a7e87465f2d4cbc768e245f4f45b2f299b05.

The games with 'rm' are on (two separate instances) of a local variable,
and make no difference.

Quoting Aditya Pakki:
 "I was the author of the patch and it was the cause of the giant UMN
  revert.

  The patch is garbage and I was unaware of the steps involved in
  retracting it. I *believed* the maintainers would pull it, given it
  was already under Greg's list. The patch does not introduce any bugs
  but is pointless and is stupid. I accept my incompetence and for not
  requesting a revert earlier."

Link: https://lwn.net/Articles/854319/
Requested-by: Aditya Pakki <pakki001@umn.edu>
Cc: Santosh Shilimkar <santosh.shilimkar@oracle.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

usb: Fix up movement of USB core kerneldoc location

Commit 855b35ea96c4 ("usb: common: move function's kerneldoc next to its
definition") moved the USB common function documentation out of the
linux/usb/ch9.h header file into drivers/usb/common/common.c and
drivers/usb/common/debug.c, which causes the following 'make htmldocs'
build warning:

include/linux/usb/ch9.h:1: warning: no structured comments found

Fix that up by pointing the documentation at the correct location.

Fixes: 855b35ea96c4 ("usb: common: move function's kerneldoc next to its definition")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Signed-off-by: Fabio Estevam <festevam@gmail.com>
Link: https://lore.kernel.org/r/20210424135103.2476670-1-festevam@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

irqchip/xilinx: Expose Kconfig option for Zynq/ZynqMP

Previously the XILINX_INTC config option was hidden and only
auto-selected on the MicroBlaze platform. However, this IP can also be
used on the Zynq and ZynqMP platforms as a secondary cascaded
controller. Allow this option to be user-enabled on those platforms.

Signed-off-by: Robert Hancock <robert.hancock@calian.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210423185853.2556087-1-robert.hancock@calian.com

Merge tag 'pinctrl-v5.12-3' of git://git./linux/kernel/git/linusw/linux-pinctrl

Pull pin control fixes from Linus Walleij:
 "Late pin control fixes, would have been in the main pull request
  normally but hey I got lucky and we got another week to polish up
  v5.12 so here we go.

  One driver fix and one making the core debugfs work:

   - Fix the number of pins in the community of the Intel Lewisburg SoC

   - Show pin numbers for controllers with base = 0 in the new debugfs
     feature"

* tag 'pinctrl-v5.12-3' of git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl:
  pinctrl: core: Show pin numbers for the controllers with base = 0
  pinctrl: lewisburg: Update number of pins in community

Merge branch 'akpm' (patches from Andrew)

Merge misc fixes from Andrew Morton:
 "5 patches.

  Subsystems affected by this patch series: coda, overlayfs, and
  mm (pagecache and memcg)"

* emailed patches from Andrew Morton <akpm@linux-foundation.org>:
  tools/cgroup/slabinfo.py: updated to work on current kernel
  mm/filemap: fix mapping_seek_hole_data on THP & 32-bit
  mm/filemap: fix find_lock_entries hang on 32-bit THP
  ovl: fix reference counting in ovl_mmap error path
  coda: fix reference counting in coda_file_mmap error path

Merge tag 'block-5.12-2021-04-23' of git://git.kernel.dk/linux-block

Pull block fix from Jens Axboe:
 "A single fix for a behavioral regression in this series, when
  re-reading the partition table with partitions open"

* tag 'block-5.12-2021-04-23' of git://git.kernel.dk/linux-block:
  block: return -EBUSY when there are open partitions in blkdev_reread_part

tools/cgroup/slabinfo.py: updated to work on current kernel

slabinfo.py script does not work with actual kernel version.

First, it was unable to recognise SLUB susbsytem, and when I specified
it manually it failed again with

  AttributeError: 'struct page' has no member 'obj_cgroups'

.. and then again with

  File "tools/cgroup/memcg_slabinfo.py", line 221, in main
    memcg.kmem_caches.address_of_(),
  AttributeError: 'struct mem_cgroup' has no member 'kmem_caches'

Link: https://lkml.kernel.org/r/cec1a75e-43b4-3d64-2084-d9f98fda037f@virtuozzo.com
Signed-off-by: Vasily Averin <vvs@virtuozzo.com>
Tested-by: Roman Gushchin <guro@fb.com>
Acked-by: Roman Gushchin <guro@fb.com>
Cc: Michal Hocko <mhocko@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/filemap: fix mapping_seek_hole_data on THP & 32-bit

No problem on 64-bit, or without huge pages, but xfstests generic/285
and other SEEK_HOLE/SEEK_DATA tests have regressed on huge tmpfs, and on
32-bit architectures, with the new mapping_seek_hole_data().  Several
different bugs turned out to need fixing.

u64 cast to stop losing bits when converting unsigned long to loff_t
(and let's use shifts throughout, rather than mixed with * and /).

Use round_up() when advancing pos, to stop assuming that pos was already
THP-aligned when advancing it by THP-size.  (This use of round_up()
assumes that any THP has THP-aligned index: true at present and true
going forward, but could be recoded to avoid the assumption.)

Use xas_set() when iterating away from a THP, so that xa_index stays in
synch with start, instead of drifting away to return bogus offset.

Check start against end to avoid wrapping 32-bit xa_index to 0 (and to
handle these additional cases, seek_data or not, it's easier to break
the loop than goto: so rearrange exit from the function).

[hughd@google.com: remove unneeded u64 casts, per Matthew]
Link: https://lkml.kernel.org/r/alpine.LSU.2.11.2104221347240.1170@eggly.anvils
Link: https://lkml.kernel.org/r/alpine.LSU.2.11.2104211737410.3299@eggly.anvils
Fixes: 41139aa4c3a3 ("mm/filemap: add mapping_seek_hole_data")
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Dave Chinner <dchinner@redhat.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Yang Shi <yang.shi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

mm/filemap: fix find_lock_entries hang on 32-bit THP

No problem on 64-bit, or without huge pages, but xfstests generic/308
hung uninterruptibly on 32-bit huge tmpfs.

Since commit 0cc3b0ec23ce ("Clarify (and fix) in 4.13 MAX_LFS_FILESIZE
macros"), MAX_LFS_FILESIZE is only a PAGE_SIZE away from wrapping 32-bit
xa_index to 0, so the new find_lock_entries() has to be extra careful
when handling a THP.

Link: https://lkml.kernel.org/r/alpine.LSU.2.11.2104211735430.3299@eggly.anvils
Fixes: 5c211ba29deb ("mm: add and use find_lock_entries")
Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: Matthew Wilcox <willy@infradead.org>
Cc: William Kucharski <william.kucharski@oracle.com>
Cc: Christoph Hellwig <hch@lst.de>
Cc: Jan Kara <jack@suse.cz>
Cc: Dave Chinner <dchinner@redhat.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com>
Cc: Yang Shi <yang.shi@linux.alibaba.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

ovl: fix reference counting in ovl_mmap error path

mmap_region() now calls fput() on the vma->vm_file.

Fix this by using vma_set_file() so it doesn't need to be handled
manually here any more.

Link: https://lkml.kernel.org/r/20210421132012.82354-2-christian.koenig@amd.com
Fixes: 1527f926fd04 ("mm: mmap: fix fput in error path v2")
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Jan Harkes <jaharkes@cs.cmu.edu>
Cc: Miklos Szeredi <miklos@szeredi.hu>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: <stable@vger.kernel.org> [5.11+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

coda: fix reference counting in coda_file_mmap error path

mmap_region() now calls fput() on the vma->vm_file.

So we need to drop the extra reference on the coda file instead of the
host file.

Link: https://lkml.kernel.org/r/20210421132012.82354-1-christian.koenig@amd.com
Fixes: 1527f926fd04 ("mm: mmap: fix fput in error path v2")
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Acked-by: Jan Harkes <jaharkes@cs.cmu.edu>
Cc: Miklos Szeredi <miklos@szeredi.hu>
Cc: Jason Gunthorpe <jgg@ziepe.ca>
Cc: <stable@vger.kernel.org> [5.11+]
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

KVM: x86/xen: Take srcu lock when accessing kvm_memslots()

kvm_memslots() will be called by kvm_write_guest_offset_cached() so we should
take the srcu lock. Let's pull the srcu lock operation from kvm_steal_time_set_preempted()
again to fix xen part.

Fixes: 30b5c851af7 ("KVM: x86/xen: Add support for vCPU runstate information")
Signed-off-by: Wanpeng Li <wanpengli@tencent.com>
Message-Id: <1619166200-9215-1-git-send-email-wanpengli@tencent.com>
Reviewed-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

Merge tag 'arm-fixes-5.12-4' of git://git./linux/kernel/git/soc/soc

Pull ARM SoC fixes from Arnd Bergmann:
 "These should be the final fixes for v5.12.

  There is one fix for SD card detection on one Allwinner board, and a
  few fixes for the Tegra platform that I had already queued up for
  v5.13 due to a communication problem. This addresses MMC device
  ordering on multiple machines, audio support on Jetson AGX Xavier and
  suspend/resume on Jetson TX2"

* tag 'arm-fixes-5.12-4' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:
  arm64: dts: allwinner: Revert SD card CD GPIO for Pine64-LTS
  arm64: tegra: Move clocks from RT5658 endpoint to device node
  arm64: tegra: Fix mmc0 alias for Jetson Xavier NX
  arm64: tegra: Set fw_devlink=on for Jetson TX2
  arm64: tegra: Add unit-address for ACONNECT on Tegra186

perf map: Fix error return code in maps__clone()

Although 'err' has been initialized to -ENOMEM, but it will be reassigned
by the "err = unwind__prepare_access(...)" statement in the for loop. So
that, the value of 'err' is unknown when map__clone() failed.

Fixes: 6c502584438bda63 ("perf unwind: Call unwind__prepare_access for forked thread")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: zhen lei <thunder.leizhen@huawei.com>
Link: http://lore.kernel.org/lkml/20210415092744.3793-1-thunder.leizhen@huawei.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf ftrace: Fix access to pid in array when setting a pid filter

Command 'perf ftrace -v -- ls' fails in s390 (at least 5.12.0rc6).

The root cause is a missing pointer dereference which causes an
array element address to be used as PID.

Fix this by extracting the PID.

Output before:
  # ./perf ftrace -v -- ls
  function_graph tracer is used
  write '-263732416' to tracing/set_ftrace_pid failed: Invalid argument
  failed to set ftrace pid
  #

Output after:
   ./perf ftrace -v -- ls
   function_graph tracer is used
   # tracer: function_graph
   #
   # CPU  DURATION                  FUNCTION CALLS
   # |     |   |                     |   |   |   |
   4)               |  rcu_read_lock_sched_held() {
   4)   0.552 us    |    rcu_lockdep_current_cpu_online();
   4)   6.124 us    |  }

Reported-by: Alexander Schmidt <alexschm@de.ibm.com>
Signed-off-by: Thomas Richter <tmricht@linux.ibm.com>
Acked-by: Namhyung Kim <namhyung@kernel.org>
Cc: Heiko Carstens <hca@linux.ibm.com>
Cc: Sumanth Korikkar <sumanthk@linux.ibm.com>
Cc: Sven Schnelle <svens@linux.ibm.com>
Cc: Vasily Gorbik <gor@linux.ibm.com>
Link: http://lore.kernel.org/lkml/20210421120400.2126433-1-tmricht@linux.ibm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

perf auxtrace: Fix potential NULL pointer dereference

In the function auxtrace_parse_snapshot_options(), the callback pointer
"itr->parse_snapshot_options" can be NULL if it has not been set during
the AUX record initialization.  This can cause tool crashing if the
callback pointer "itr->parse_snapshot_options" is dereferenced without
performing NULL check.

Add a NULL check for the pointer "itr->parse_snapshot_options" before
invoke the callback.

Fixes: d20031bb63dd6dde ("perf tools: Add AUX area tracing Snapshot Mode")
Signed-off-by: Leo Yan <leo.yan@linaro.org>
Acked-by: Adrian Hunter <adrian.hunter@intel.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Andi Kleen <ak@linux.intel.com>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Tiezhu Yang <yangtiezhu@loongson.cn>
Link: http://lore.kernel.org/lkml/20210420151554.2031768-1-leo.yan@linaro.org
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>

Merge remote-tracking branch 'regulator/for-5.13' into regulator-next

Merge remote-tracking branch 'regulator/for-5.12' into regulator-linus

Merge tag 'drm-fixes-2021-04-23' of git://anongit.freedesktop.org/drm/drm

Pull drm fixes from Dave Airlie:
 "Just some small i915 and amdgpu fixes this week, should be all until
  you open the merge window.

  amdgpu:
   - Fix gpuvm page table update issue
   - Modifier fixes
   - Register fix for dimgrey cavefish

  i915:
   - GVT's BDW regression fix for cmd parser
   - Fix modesetting in case of unexpected AUX timeouts"

* tag 'drm-fixes-2021-04-23' of git://anongit.freedesktop.org/drm/drm:
  drm/amdgpu: fix GCR_GENERAL_CNTL offset for dimgrey_cavefish
  amd/display: allow non-linear multi-planar formats
  drm/amd/display: Update modifier list for gfx10_3
  drm/amdgpu: reserve fence slot to update page table
  drm/i915: Fix modesetting in case of unexpected AUX timeouts
  drm/i915/gvt: Fix BDW command parser regression

Merge tag 'gpio-fixes-for-v5.12' of git://git./linux/kernel/git/brgl/linux

Pull gpio fix from Bartosz Golaszewski:
 "Save and restore the sysconfig register in gpio-omap to fix a
  power-management issue"

* tag 'gpio-fixes-for-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
  gpio: omap: Save and restore sysconfig

platform/x86: gigabyte-wmi: add support for B550M AORUS PRO-P

From: Alexey Klimov <klimov.linux@gmail.com>

Add the B550M AORUS PRO-P motherboard description to
gigabyte_wmi_known_working_platforms.

Signed-off-by: Alexey Klimov <klimov.linux@gmail.com>
Acked-by: Thomas Weißschuh <thomas@weissschuh.net>
Link: https://lore.kernel.org/r/20210421234156.3942343-1-aklimov@redhat.com
Signed-off-by: Hans de Goede <hdegoede@redhat.com>

spi: Rename enable1 to activate in spi_set_cs()

The enable1 is confusing name. Change it to clearly show what is
the intention behind it. No functional changes.

Fixes: 25093bdeb6bc ("spi: implement SW control for CS times")
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Link: https://lore.kernel.org/r/20210420131846.75983-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>

regulator: Add binding for TCS4525

Add a compatible string to support TCS4525/TCS4526 devices,
which are compatible with Fairchild FAN53555 regulators.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Link: https://lore.kernel.org/r/20210421210338.43819-2-ezequiel@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

regulator: fan53555: Add TCS4525 DCDC support

TCS4525 main features:

- 2.7V to 5.5V Input Voltage Range;
- 3MHz Constant Switching Frequency;
- 5A Available Load Current;
- Programmable Output Voltage: 0.6V to 1.4V in 6.25mV Steps;
- PFM/PWM Operation for Optimum Increased Efficiency;

Signed-off-by: Joseph Chen <chenjh@rock-chips.com>
[Ezequiel: Forward port]
Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Link: https://lore.kernel.org/r/20210421210338.43819-3-ezequiel@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

dt-bindings: vendor-prefixes: Add Tang Cheng (TCS)

Shenzhen City Tang Cheng Technology (http://www.tctek.cn/)
is a power management IC manufacturer.

Signed-off-by: Ezequiel Garcia <ezequiel@collabora.com>
Link: https://lore.kernel.org/r/20210421210338.43819-1-ezequiel@collabora.com
Signed-off-by: Mark Brown <broonie@kernel.org>

Merge branch 'tegra/dt64' into arm/fixes

arm64: tegra: Device tree fixes for v5.12-rc6

This contains a couple of device tree fixes for the v5.12 release cycle.
These are needed for proper audio support on Jetson AGX Xavier, to boot
the Jetson Xavier NX from an SD card and to be able to suspend/resume
the Jetson TX2.

* tegra/dt64:
  arm64: tegra: Move clocks from RT5658 endpoint to device node
  arm64: tegra: Fix mmc0 alias for Jetson Xavier NX
  arm64: tegra: Set fw_devlink=on for Jetson TX2
  arm64: tegra: Add unit-address for ACONNECT on Tegra186

Link: https://lore.kernel.org/linux-arm-kernel/YILD4yyPXuiYbHW1@orome.fritz.box/
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

irqchip/gic-v3: Do not enable irqs when handling spurious interrups

We triggered the following error while running our 4.19 kernel
with the pseudo-NMI patches backported to it:

[   14.816231] ------------[ cut here ]------------
[   14.816231] kernel BUG at irq.c:99!
[   14.816232] Internal error: Oops - BUG: 0 [#1] SMP
[   14.816232] Process swapper/0 (pid: 0, stack limit = 0x(____ptrval____))
[   14.816233] CPU: 0 PID: 0 Comm: swapper/0 Tainted: G           O      4.19.95.aarch64 #14
[   14.816233] Hardware name: evb (DT)
[   14.816234] pstate: 80400085 (Nzcv daIf +PAN -UAO)
[   14.816234] pc : asm_nmi_enter+0x94/0x98
[   14.816235] lr : asm_nmi_enter+0x18/0x98
[   14.816235] sp : ffff000008003c50
[   14.816235] pmr_save: 00000070
[   14.816237] x29: ffff000008003c50 x28: ffff0000095f56c0
[   14.816238] x27: 0000000000000000 x26: ffff000008004000
[   14.816239] x25: 00000000015e0000 x24: ffff8008fb916000
[   14.816240] x23: 0000000020400005 x22: ffff0000080817cc
[   14.816241] x21: ffff000008003da0 x20: 0000000000000060
[   14.816242] x19: 00000000000003ff x18: ffffffffffffffff
[   14.816243] x17: 0000000000000008 x16: 003d090000000000
[   14.816244] x15: ffff0000095ea6c8 x14: ffff8008fff5ab40
[   14.816244] x13: ffff8008fff58b9d x12: 0000000000000000
[   14.816245] x11: ffff000008c8a200 x10: 000000008e31fca5
[   14.816246] x9 : ffff000008c8a208 x8 : 000000000000000f
[   14.816247] x7 : 0000000000000004 x6 : ffff8008fff58b9e
[   14.816248] x5 : 0000000000000000 x4 : 0000000080000000
[   14.816249] x3 : 0000000000000000 x2 : 0000000080000000
[   14.816250] x1 : 0000000000120000 x0 : ffff0000095f56c0
[   14.816251] Call trace:
[   14.816251]  asm_nmi_enter+0x94/0x98
[   14.816251]  el1_irq+0x8c/0x180                    (IRQ C)
[   14.816252]  gic_handle_irq+0xbc/0x2e4
[   14.816252]  el1_irq+0xcc/0x180                    (IRQ B)
[   14.816253]  arch_timer_handler_virt+0x38/0x58
[   14.816253]  handle_percpu_devid_irq+0x90/0x240
[   14.816253]  generic_handle_irq+0x34/0x50
[   14.816254]  __handle_domain_irq+0x68/0xc0
[   14.816254]  gic_handle_irq+0xf8/0x2e4
[   14.816255]  el1_irq+0xcc/0x180                    (IRQ A)
[   14.816255]  arch_cpu_idle+0x34/0x1c8
[   14.816255]  default_idle_call+0x24/0x44
[   14.816256]  do_idle+0x1d0/0x2c8
[   14.816256]  cpu_startup_entry+0x28/0x30
[   14.816256]  rest_init+0xb8/0xc8
[   14.816257]  start_kernel+0x4c8/0x4f4
[   14.816257] Code: 940587f1 d5384100 b9401001 36a7fd01 (d4210000)
[   14.816258] Modules linked in: start_dp(O) smeth(O)
[   15.103092] ---[ end trace 701753956cb14aa8 ]---
[   15.103093] Kernel panic - not syncing: Fatal exception in interrupt
[   15.103099] SMP: stopping secondary CPUs
[   15.103100] Kernel Offset: disabled
[   15.103100] CPU features: 0x36,a2400218
[   15.103100] Memory Limit: none

which is cause by a 'BUG_ON(in_nmi())' in nmi_enter().

From the call trace, we can find three interrupts (noted A, B, C above):
interrupt (A) is preempted by (B), which is further interrupted by (C).

Subsequent investigations show that (B) results in nmi_enter() being
called, but that it actually is a spurious interrupt. Furthermore,
interrupts are reenabled in the context of (B), and (C) fires with
NMI priority. We end-up with a nested NMI situation, something
we definitely do not want to (and cannot) handle.

The bug here is that spurious interrupts should never result in any
state change, and we should just return to the interrupted context.
Moving the handling of spurious interrupts as early as possible in
the GICv3 handler fixes this issue.

Fixes: 3f1f3234bc2d ("irqchip/gic-v3: Switch to PMR masking before calling IRQ handler")
Acked-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: He Ying <heying24@huawei.com>
[maz: rewrote commit message, corrected Fixes: tag]
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210423083516.170111-1-heying24@huawei.com
Cc: stable@vger.kernel.org

spi: Convert Freescale QSPI binding to json schema

Convert the Freescale QSPI binding to DT schema format using json-schema.

Signed-off-by: Kuldeep Singh <kuldeep.singh@nxp.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210312054038.3586706-1-kuldeep.singh@nxp.com
Signed-off-by: Mark Brown <broonie@kernel.org>

regulator: core: Fix off_on_delay handling

The jiffies-based off_on_delay implementation has a couple of problems
that cause it to sometimes not actually delay for the required time:

 (1) If, for example, the off_on_delay time is equivalent to one jiffy,
     and the ->last_off_jiffy is set just before a new jiffy starts,
     then _regulator_do_enable() does not wait at all since it checks
     using time_before().

 (2) When jiffies overflows, the value of "remaining" becomes higher
     than "max_delay" and the code simply proceeds without waiting.

Fix these problems by changing it to use ktime_t instead.

[Note that since jiffies doesn't start at zero but at INITIAL_JIFFIES
 ("-5 minutes"), (2) above also led to the code not delaying if
 the first regulator_enable() is called when the ->last_off_jiffy is not
 initialised, such as for regulators with ->constraints->boot_on set.
 It's not clear to me if this was intended or not, but I've preserved
 this behaviour explicitly with the check for a non-zero ->last_off.]

Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Link: https://lore.kernel.org/r/20210423114524.26414-1-vincent.whitchurch@axis.com
Signed-off-by: Mark Brown <broonie@kernel.org>

Merge branch 'for-5.12' of https://git./linux/kernel/git/broonie/regulator into regulator-5.13

bluetooth: eliminate the potential race condition when removing the HCI controller

There is a possible race condition vulnerability between issuing a HCI
command and removing the cont.  Specifically, functions hci_req_sync()
and hci_dev_do_close() can race each other like below:

thread-A in hci_req_sync()      |   thread-B in hci_dev_do_close()
                                |   hci_req_sync_lock(hdev);
test_bit(HCI_UP, &hdev->flags); |
...                             |   test_and_clear_bit(HCI_UP, &hdev->flags)
hci_req_sync_lock(hdev);        |
                                |
In this commit we alter the sequence in function hci_req_sync(). Hence,
the thread-A cannot issue th.

Signed-off-by: Lin Ma <linma@zju.edu.cn>
Cc: Marcel Holtmann <marcel@holtmann.org>
Fixes: 7c6a329e4447 ("[Bluetooth] Fix regression from using default link policy")
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: dwc3: gadget: Handle DEV_TXF_FLUSH_BYPASS capability

DWC_usb32 IP introduces a new behavior when handling NoStream event for
IN endpoints. If the controller is capable of DEV_TXF_FLUSH_BYPASS, then
the driver does not need to force to restart stream for IN endpoints.
The controller will generate ERDY and restart the stream periodically.

Acked-by: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
Link: https://lore.kernel.org/r/65c3070c666cd6b8beeee62d7f8e3e704ebf2d32.1619134559.git.Thinh.Nguyen@synopsys.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

usb: dwc3: Capture new capability register GHWPARAMS9

DWC_usb32 introduces a new HW capability register GHWPARAMS9. Capture
this in the dwc->hwparams.hwparams9 field.

Acked-by: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Thinh Nguyen <Thinh.Nguyen@synopsys.com>
Link: https://lore.kernel.org/r/f76cc4a9c8c4ab325f5babe03c57b039166360b0.1619134559.git.Thinh.Nguyen@synopsys.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xen/arm: introduce XENFEAT_direct_mapped and XENFEAT_not_direct_mapped

Newer Xen versions expose two Xen feature flags to tell us if the domain
is directly mapped or not. Only when a domain is directly mapped it
makes sense to enable swiotlb-xen on ARM.

Introduce a function on ARM to check the new Xen feature flags and also
to deal with the legacy case. Call the function xen_swiotlb_detect.

Signed-off-by: Stefano Stabellini <stefano.stabellini@xilinx.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Link: https://lore.kernel.org/r/20210319200140.12512-1-sstabellini@kernel.org
Signed-off-by: Juergen Gross <jgross@suse.com>

soc: aspeed: fix a ternary sign expansion bug

The intent here was to return negative error codes but it actually
returns positive values.  The problem is that type promotion with
ternary operations is quite complicated.

"ret" is an int.  "copied" is a u32.  And the snoop_file_read() function
returns long.  What happens is that "ret" is cast to u32 and becomes
positive then it's cast to long and it's still positive.

Fix this by removing the ternary so that "ret" is type promoted directly
to long.

Fixes: 3772e5da4454 ("drivers/misc: Aspeed LPC snoop output using misc chardev")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Patrick Venture <venture@google.com>
Link: https://lore.kernel.org/r/YIE90PSXsMTa2Y8n@mwanda
Link: https://lore.kernel.org/r/20210423000919.1249474-1-joel@jms.id.au'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

xen-pciback: simplify vpci's find hook

There's no point in comparing SBDF - we can simply compare the struct
pci_dev pointers. If they weren't the same for a given device, we'd have
bigger problems from having stored a stale pointer.

Signed-off-by: Jan Beulich <jbeulich@suse.com>
Reviewed-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Link: https://lore.kernel.org/r/158273a2-d1b9-3545-b25d-affca867376c@suse.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen-blkfront: Fix 'physical' typos

Fix misspelling of "physical".

Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Roger Pau Monné <roger.pau@citrix.com>
Link: https://lore.kernel.org/r/20210126205509.2917606-1-helgaas@kernel.org
Signed-off-by: Juergen Gross <jgross@suse.com>

xen-blkback: fix compatibility bug with single page rings

Prior to commit 4a8c31a1c6f5 ("xen/blkback: rework connect_ring() to avoid
inconsistent xenstore 'ring-page-order' set by malicious blkfront"), the
behaviour of xen-blkback when connecting to a frontend was:

- read 'ring-page-order'
- if not present then expect a single page ring specified by 'ring-ref'
- else expect a ring specified by 'ring-refX' where X is between 0 and
  1 << ring-page-order

This was correct behaviour, but was broken by the afforementioned commit to
become:

- read 'ring-page-order'
- if not present then expect a single page ring (i.e. ring-page-order = 0)
- expect a ring specified by 'ring-refX' where X is between 0 and
  1 << ring-page-order
- if that didn't work then see if there's a single page ring specified by
  'ring-ref'

This incorrect behaviour works most of the time but fails when a frontend
that sets 'ring-page-order' is unloaded and replaced by one that does not
because, instead of reading 'ring-ref', xen-blkback will read the stale
'ring-ref0' left around by the previous frontend will try to map the wrong
grant reference.

This patch restores the original behaviour.

Fixes: 4a8c31a1c6f5 ("xen/blkback: rework connect_ring() to avoid inconsistent xenstore 'ring-page-order' set by malicious blkfront")
Signed-off-by: Paul Durrant <pdurrant@amazon.com>
Reviewed-by: Dongli Zhang <dongli.zhang@oracle.com>
Reviewed-by: "Roger Pau Monné" <roger.pau@citrix.com>
Link: https://lore.kernel.org/r/20210202175659.18452-1-paul@xen.org
Signed-off-by: Juergen Gross <jgross@suse.com>

xen: Remove support for PV ACPI cpu/memory hotplug

Commit 76fc253723ad ("xen/acpi-stub: Disable it b/c the acpi_processor_add
is no longer called.") declared as BROKEN support for Xen ACPI stub (which
is required for xen-acpi-{cpu|memory}-hotplug) and suggested that this
is temporary and will be soon fixed. This was in March 2013.

Further, commit cfafae940381 ("xen: rename dom0_op to platform_op")
renamed an interface used by memory hotplug code without updating that
code (as it was BROKEN and therefore not compiled). This was
in November 2015 and has gone unnoticed for over 5 year.

It is now clear that this code is of no interest to anyone and therefore
should be removed.

Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Link: https://lore.kernel.org/r/1618336344-3162-1-git-send-email-boris.ostrovsky@oracle.com
Signed-off-by: Juergen Gross <jgross@suse.com>

xen/pciback: Fix incorrect type warnings

Correct enum pci_channel_io_normal should be used instead of putting
integer value 1.

Fix following smatch warnings:
drivers/xen/xen-pciback/pci_stub.c:805:40: warning: incorrect type in argument 2 (different base types)
drivers/xen/xen-pciback/pci_stub.c:805:40:    expected restricted pci_channel_state_t [usertype] state
drivers/xen/xen-pciback/pci_stub.c:805:40:    got int
drivers/xen/xen-pciback/pci_stub.c:862:40: warning: incorrect type in argument 2 (different base types)
drivers/xen/xen-pciback/pci_stub.c:862:40:    expected restricted pci_channel_state_t [usertype] state
drivers/xen/xen-pciback/pci_stub.c:862:40:    got int
drivers/xen/xen-pciback/pci_stub.c:973:31: warning: incorrect type in argument 2 (different base types)
drivers/xen/xen-pciback/pci_stub.c:973:31:    expected restricted pci_channel_state_t [usertype] state
drivers/xen/xen-pciback/pci_stub.c:973:31:    got int

Signed-off-by: Muhammad Usama Anjum <musamaanjum@gmail.com>
Reviewed-by: Juergen Gross <jgross@suse.com>
Link: https://lore.kernel.org/r/20210326181442.GA1735905@LEGION
Signed-off-by: Juergen Gross <jgross@suse.com>

Merge tag 'drm-intel-fixes-2021-04-22' of git://anongit.freedesktop.org/drm/drm-intel into drm-fixes

- GVT's BDW regression fix for cmd parser (Zhenyu)
- Fix modesetting in case of unexpected AUX timeouts (Imre)

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/YIGZ3pQPgPQtZtyI@intel.com

Merge tag 'amd-drm-fixes-5.12-2021-04-21' of https://gitlab.freedesktop.org/agd5f/linux into drm-fixes

amd-drm-fixes-5.12-2021-04-21:

amdgpu:
- Fix gpuvm page table update issue
- Modifier fixes
- Register fix for dimgrey cavefish

Signed-off-by: Dave Airlie <airlied@redhat.com>
From: Alex Deucher <alexander.deucher@amd.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20210421220456.3839-1-alexander.deucher@amd.com

Merge tag 'for_linus' of git://git./linux/kernel/git/mst/vhost

Pull virtio fixes from Michael Tsirkin:
 "Very late in the cycle but both risky if left unfixed and more or less
  obvious.."

* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
  vdpa/mlx5: Set err = -ENOMEM in case dma_map_sg_attrs fails
  vhost-vdpa: protect concurrent access to vhost device iotlb

vdpa/mlx5: Set err = -ENOMEM in case dma_map_sg_attrs fails

Set err = -ENOMEM if dma_map_sg_attrs() fails so the function reutrns
error.

Fixes: 94abbccdf291 ("vdpa/mlx5: Add shared memory registration code")
Signed-off-by: Eli Cohen <elic@nvidia.com>
Reported-by: kernel test robot <lkp@intel.com>
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20210411083646.910546-1-elic@nvidia.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>

vhost-vdpa: protect concurrent access to vhost device iotlb

Protect vhost device iotlb by vhost_dev->mutex. Otherwise,
it might cause corruption of the list and interval tree in
struct vhost_iotlb if userspace sends the VHOST_IOTLB_MSG_V2
message concurrently.

Fixes: 4c8cf318("vhost: introduce vDPA-based backend")
Cc: stable@vger.kernel.org
Signed-off-by: Xie Yongji <xieyongji@bytedance.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20210412095512.178-1-xieyongji@bytedance.com
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

dt-bindings: mali-bifrost: add dma-coherent

Add optional dma-coherent property to binding doc.

Found by 'make dtbs_check' on arm64/amlogic DT files.

Signed-off-by: Kevin Hilman <khilman@baylibre.com>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210421204833.18523-2-khilman@baylibre.com'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

arm64: dts: amlogic: misc DT schema fixups

Take a pass at cleaning up a bunch of warnings
from 'make dtbs_check' that have crept in.

Signed-off-by: Kevin Hilman <khilman@baylibre.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210421204833.18523-1-khilman@baylibre.com'
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'sunxi-fixes-for-5.12-2' of git://git./linux/kernel/git/sunxi/linux into arm/fixes

One fix for the MMC card detect on the Pine H64 board

* tag 'sunxi-fixes-for-5.12-2' of git://git.kernel.org/pub/scm/linux/kernel/git/sunxi/linux:
  arm64: dts: allwinner: Revert SD card CD GPIO for Pine64-LTS

Link: https://lore.kernel.org/r/45fc5e4d-ef48-4729-a869-79a8f288bb83.lettre@localhost
Signed-off-by: Arnd Bergmann <arnd@arndb.de>

Merge tag 'fixes' of git://git./linux/kernel/git/jejb/tpmdd

Pull tpm fix from James Bottomley:
 "This is an urgent regression fix for a tpm patch set that went in this
  merge window. It looks like a rebase before the original pull request
  lost a tpm_try_get_ops() so we have a lock imbalance in our code which
  is causing oopses. The original patch was correct on the mailing list.

  I'm sending this in agreement with Mimi (as joint maintainers of
  trusted keys) because Jarkko is off communing with the Reindeer or
  whatever it is Finns do when on holiday"

* tag 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/tpmdd:
  KEYS: trusted: Fix TPM reservation for seal/unseal

spi: stm32-qspi: fix debug format string

Printing size_t needs a special %zx format modifier to avoid a
warning like:

drivers/spi/spi-stm32-qspi.c:481:41: note: format string is defined here
  481 |         dev_dbg(qspi->dev, "%s len = 0x%x offs = 0x%llx buf = 0x%p\n", __func__, len, offs, buf);

Patrice already tried to fix this, but picked %lx instead of %zx,
which fixed some architectures but broke others in the same way.
Using %zx works everywhere.

Fixes: 18674dee3cd6 ("spi: stm32-qspi: Add dirmap support")
Fixes: 1b8a7d4282c0 ("spi: stm32-qspi: Fix compilation warning in ARM64")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/r/20210422134955.1988316-1-arnd@kernel.org
Signed-off-by: Mark Brown <broonie@kernel.org>

spi: tools: make a symbolic link to the header file spi.h

The header file spi.h in include/uapi/linux/spi is needed for spidev.h,
so we also need make a symbolic link to it to eliminate the error message
as below:

In file included from spidev_test.c:24:
include/linux/spi/spidev.h:28:10: fatal error: linux/spi/spi.h: No such file or directory
   28 | #include <linux/spi/spi.h>
      |          ^~~~~~~~~~~~~~~~~
compilation terminated.

Fixes: f7005142dace ("spi: uapi: unify SPI modes into a single spi.h")
Signed-off-by: Quanyang Wang <quanyang.wang@windriver.com>
Link: https://lore.kernel.org/r/20210422102604.3034217-1-quanyang.wang@windriver.com
Signed-off-by: Mark Brown <broonie@kernel.org>

regulator: core: Respect off_on_delay at startup

We currently do not respect off_on_delay the first time we turn on a
regulator.  This is problematic since the regulator could have been
turned off by the bootloader, or it could it have been turned off during
the probe of the regulator driver (such as when regulator-fixed requests
the enable GPIO), either of which could potentially have happened less
than off_on_delay microseconds ago before the first time a client
requests for the regulator to be turned on.

We can't know exactly when the regulator was turned off, but initialise
off_on_delay to the current time when registering the regulator, so that
we guarantee that we respect the off_on_delay in all cases.

Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
Link: https://lore.kernel.org/r/20210422083044.11479-1-vincent.whitchurch@axis.com
Signed-off-by: Mark Brown <broonie@kernel.org>

dt-bindings: interrupt-controller: Add IDT 79RC3243x Interrupt Controller

Document DT bindings for IDT 79RC3243x Interrupt Controller.

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210422145330.73452-2-tsbogend@alpha.franken.de

irqchip: Add support for IDT 79rc3243x interrupt controller

IDT 79rc3243x SoCs have rather simple interrupt controllers connected
to the MIPS CPU interrupt lines. Each of them has room for up to
32 interrupts.

Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210422145330.73452-1-tsbogend@alpha.franken.de

irqdomain: Drop references to recusive irqdomain setup

It was never completely implemented, and was removed a long time
ago. Adjust the documentation to reflect this.

Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210406093557.1073423-8-maz@kernel.org

irqdomain: Get rid of irq_create_strict_mappings()

No user of this helper is left, remove it.

Signed-off-by: Marc Zyngier <maz@kernel.org>

irqchip/jcore-aic: Kill use of irq_create_strict_mappings()

irq_create_strict_mappings() is a poor way to allow the use of
a linear IRQ domain as a legacy one. Let's be upfront about it.

Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210406093557.1073423-4-maz@kernel.org

ARM: PXA: Kill use of irq_create_strict_mappings()

irq_create_strict_mappings() is a poor way to allow the use of
a linear IRQ domain as a legacy one. Let's be upfront about
it and use a legacy domain when appropriate.

Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210406093557.1073423-3-maz@kernel.org

irqchip/gic-v4.1: Disable vSGI upon (GIC CPUIF < v4.1) detection

GIC CPU interfaces versions predating GIC v4.1 were not built to
accommodate vINTID within the vSGI range; as reported in the GIC
specifications (8.2 "Changes to the CPU interface"), it is
CONSTRAINED UNPREDICTABLE to deliver a vSGI to a PE with
ID_AA64PFR0_EL1.GIC < b0011.

Check the GIC CPUIF version by reading the SYS_ID_AA64_PFR0_EL1.

Disable vSGIs if a CPUIF version < 4.1 is detected to prevent using
vSGIs on systems where they may misbehave.

Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Marc Zyngier <maz@kernel.org>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210317100719.3331-2-lorenzo.pieralisi@arm.com

PM: wakeup: remove redundant assignment to variable retval

The variable retval is being initialized with a value that is
never read and it is being updated later with a new value. The
initialization is redundant and can be removed.

Addresses-Coverity: ("Unused value")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

irqchip/tb10x: Use 'fallthrough' to eliminate a warning

Use the 'fallthrough' macro to document that this switch case
does indeed fall through to the next case.

../drivers/irqchip/irq-tb10x.c: In function 'tb10x_irq_set_type':
../drivers/irqchip/irq-tb10x.c:62:13: warning: this statement may fall through [-Wimplicit-fallthrough=]
   62 |   flow_type = IRQ_TYPE_LEVEL_LOW;
../drivers/irqchip/irq-tb10x.c:63:2: note: here
   63 |  case IRQ_TYPE_LEVEL_LOW:
      |  ^~~~

Fixes: b06eb0173ef1 ("irqchip: Add TB10x interrupt controller driver")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Marc Zyngier <maz@kernel.org>
Cc: Christian Ruppert <christian.ruppert@abilis.com>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20210422051620.23021-1-rdunlap@infradead.org

perf/x86/kvm: Fix Broadwell Xeon stepping in isolation_ucodes[]

The only stepping of Broadwell Xeon parts is stepping 1. Fix the
relevant isolation_ucodes[] entry, which previously enumerated
stepping 2.

Although the original commit was characterized as an optimization, it
is also a workaround for a correctness issue.

If a PMI arrives between kvm's call to perf_guest_get_msrs() and the
subsequent VM-entry, a stale value for the IA32_PEBS_ENABLE MSR may be
restored at the next VM-exit. This is because, unbeknownst to kvm, PMI
throttling may clear bits in the IA32_PEBS_ENABLE MSR. CPUs with "PEBS
isolation" don't suffer from this issue, because perf_guest_get_msrs()
doesn't report the IA32_PEBS_ENABLE value.

Fixes: 9b545c04abd4f ("perf/x86/kvm: Avoid unnecessary work in guest filtering")
Signed-off-by: Jim Mattson <jmattson@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Peter Shier <pshier@google.com>
Acked-by: Andi Kleen <ak@linux.intel.com>
Link: https://lkml.kernel.org/r/20210422001834.1748319-1-jmattson@google.com

usb: gadget: prevent a ternary sign expansion bug

The problem is that "req->actual" is a u32, "req->status" is an int, and
iocb->ki_complete() takes a long.  We would expect that a negative error
code in "req->status" would translate to a negative long value.

But what actually happens is that because "req->actual" is a u32, the
error codes is type promoted to a high positive value and then remains
a positive value when it is cast to long.  (No sign expansion).

We can fix this by casting "req->status" to long.

Acked-by: Felipe Balbi <balbi@kernel.org>
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/YIE7RrBPLWc3XtMg@mwanda
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: extend compile-test coverage

Allow more drivers to be compile tested more easily, for example, when
doing subsystem-wide changes.

Verified on X86_64 as well as arm, powerpc and m68k with minimal configs
in order to catch missing implicit build dependencies (e.g. MAILBOX for
SERIAL_TEGRA_TCU).

Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210422080211.29326-1-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: stm32: add FIFO threshold configuration

Add the support for two optional DT properties, to configure RX and TX
FIFO thresholds:
- rx-threshold
- tx-threshold
This replaces hard-coded 8 bytes threshold. Keep 8 as the default value if
not specified, for backward compatibility.

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Changes in v2:
Change added properties naming as proposed by Rob Herring.
Link: https://lore.kernel.org/r/20210413174015.23011-5-erwan.leray@foss.st.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dt-bindings: serial: 8250: update TX FIFO trigger level

Remove data type from tx-threshold trigger level as defined now as a
serial generic property.

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Link: https://lore.kernel.org/r/20210413174015.23011-4-erwan.leray@foss.st.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dt-bindings: serial: stm32: override FIFO threshold properties

Override rx-threshold and tx-threshold properties:
- extend description
- provide default and expected values

Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Changes in v2:
Change added properties naming and factorize it in serial.yaml as proposed
by Rob Herring.

Reviewed-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20210413174015.23011-3-erwan.leray@foss.st.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

dt-bindings: serial: add RX and TX FIFO properties

Add two optional DT properties to configure RX and TX FIFO thresholds:
- rx-threshold
- tx-threshold

Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Erwan Le Ray <erwan.leray@foss.st.com>
Link: https://lore.kernel.org/r/20210413174015.23011-2-erwan.leray@foss.st.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: xilinx_uartps: drop low-latency workaround

Commit c8dbdc842d30 ("serial: xuartps: Rewrite the interrupt handling
logic") reworked the driver interrupt processing but also, without
comment, added an unnecessary workaround for the infamous low_latency
behaviour of tty_flip_buffer_push() which had been removed years
before.

Specifically, since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks around.

Cc: Michal Simek <michal.simek@xilinx.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-27-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: vt8500: drop low-latency workaround

Commit de49df58366f ("tty: serial: vt8500: drop uart_port->lock before
calling tty_flip_buffer_push()") claimed to address a locking
issue but only provided a dubious lockdep splat from an unrelated
driver, which in the end turned out to be due a broken local change
carried by the author.

Unfortunately these patches were merged before the issue had been
analysed properly so the commit messages makes no sense whatsoever.

The real issue was first seen on RT which at the time effectively always
set the low_latency flag for all serial drivers by patching
tty_flip_buffer_push(). This in turn revealed that many drivers did not
handle the infamous low_latency behaviour which meant that data was
pushed immediately to the line discipline instead of being deferred to a
work queue.

Since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks
around.

Link: https://lore.kernel.org/linux-serial/cover.1376923198.git.viresh.kumar@linaro.org/
Cc: Tony Prisk <linux@prisktech.co.nz>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-26-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: timbuart: drop low-latency workaround

The timbuart driver has always carried a workaround for the infamous
low_latency behaviour of tty_flip_buffer_push() which required not
holding the port lock when the low_latency flag was set.

Since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks
around.

Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-25-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: sunsu: drop low-latency workaround

The sunsu driver has been carrying a workaround for the infamous
low_latency behaviour of tty_flip_buffer_push() by dropping and
reacquiring the port lock in the interrupt handler since 2004.

Since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks around.

Cc: "David S. Miller" <davem@davemloft.net>
Acked-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-24-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: sifive: drop low-latency workaround

The sifive driver has always carried an unnecessary workaround for the
infamous low_latency behaviour of tty_flip_buffer_push() which had been
removed years before the driver was added by commit 45c054d0815b ("tty:
serial: add driver for the SiFive UART").

Specifically, since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks around.

Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-23-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: txx9: drop low-latency workaround

Commit f5ee56cc184e ("[PATCH] txx9 serial update") worked around the
infamous low_latency behaviour of tty_flip_buffer_push() by simply
dropping and reacquiring the port lock in the interrupt handler.

Since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks around.

Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-22-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: sa1100: drop low-latency workaround

Commit 53e0e6706c76 ("tty: serial: sa1100: drop uart_port->lock before
calling tty_flip_buffer_push()") claimed to address a locking
issue but only provided a dubious lockdep splat from an unrelated
driver, which in the end turned out to be due a broken local change
carried by the author.

Unfortunately these patches were merged before the issue had been
analysed properly so the commit messages makes no sense whatsoever.

The real issue was first seen on RT which at the time effectively always
set the low_latency flag for all serial drivers by patching
tty_flip_buffer_push(). This in turn revealed that many drivers did not
handle the infamous low_latency behaviour which meant that data was
pushed immediately to the line discipline instead of being deferred to a
work queue.

Since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks
around.

Link: https://lore.kernel.org/linux-serial/cover.1376923198.git.viresh.kumar@linaro.org/
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-21-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

serial: rp2: drop low-latency workaround

Commit de7053c77123 ("tty: serial: rp2: drop uart_port->lock before
calling tty_flip_buffer_push()") claimed to address a locking
issue but only provided a dubious lockdep splat from an unrelated
driver, which in the end turned out to be due a broken local change
carried by the author.

Unfortunately these patches were merged before the issue had been
analysed properly so the commit messages makes no sense whatsoever.

The real issue was first seen on RT which at the time effectively always
set the low_latency flag for all serial drivers by patching
tty_flip_buffer_push(). This in turn revealed that many drivers did not
handle the infamous low_latency behaviour which meant that data was
pushed immediately to the line discipline instead of being deferred to a
work queue.

Since commit a9c3f68f3cd8 ("tty: Fix low_latency BUG"),
tty_flip_buffer_push() always schedules a work item to push data to the
line discipline and there's no need to keep any low_latency hacks
around.

Link: https://lore.kernel.org/linux-serial/cover.1376923198.git.viresh.kumar@linaro.org/
Cc: Kevin Cernekee <cernekee@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Link: https://lore.kernel.org/r/20210421095509.3024-20-johan@kernel.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>