Input: i8042 - add TUXEDO devices to i8042 quirk tables

[ Upstream commit a6a87c36165e6791eeaed88025cde270536c3198 ]

A lot of modern Clevo barebones have touchpad and/or keyboard issues after
suspend fixable with nomux + reset + noloop + nopnp. Luckily, none of them
have an external PS/2 port so this can safely be set for all of them.

I'm not entirely sure if every device listed really needs all four quirks,
but after testing and production use. No negative effects could be
observed when setting all four.

The list is quite massive as neither the TUXEDO nor the Clevo dmi strings
have been very consistent historically. I tried to keep the list as short
as possible without risking on missing an affected device.

This is revision 3. The Clevo N150CU barebone is still removed as it might
have problems with the fix and needs further investigations. The
SchenkerTechnologiesGmbH System-/Board-Vendor string variations are
added. This is now based in the quirk table refactor. This now also
includes the additional noaux flag for the NS7xMU.

Signed-off-by: Werner Sembach <wse@tuxedocomputers.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20220629112725.12922-5-wse@tuxedocomputers.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Input: i8042 - merge quirk tables

[ Upstream commit ff946268a0813c35b790dfbe07c3bfaa7bfb869c ]

Merge i8042 quirk tables to reduce code duplication for devices that need
more than one quirk. Before every quirk had its own table with devices
needing that quirk. If a new quirk needed to be added a new table had to
be created. When a device needed multiple quirks, it appeared in multiple
tables. Now only one table called i8042_dmi_quirk_table exists. In it every
device has one entry and required quirks are coded in the .driver_data
field of the struct dmi_system_id used by this table. Multiple quirks for
one device can be applied by bitwise-or of the new SERIO_QUIRK_* defines.

Also align quirkable options with command line parameters and make vendor
wide quirks per device overwriteable on a per device basis. The first match
is honored while following matches are ignored. So when a vendor wide quirk
is defined in the table, a device can inserted before and therefore
ignoring the vendor wide define.

Signed-off-by: Werner Sembach <wse@tuxedocomputers.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20220629112725.12922-3-wse@tuxedocomputers.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

Input: i8042 - move __initconst to fix code styling warning

[ Upstream commit 95a9916c909f0b1d95e24b4232b4bc38ff755415 ]

Move __intconst from before i8042_dmi_laptop_table[] to after it for
consistent code styling.

Signed-off-by: Werner Sembach <wse@tuxedocomputers.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/20220629112725.12922-2-wse@tuxedocomputers.com
Signed-off-by: Dmitry Torokhov <dmitry.torokhov@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

btrfs: convert count_max_extents() to use fs_info->max_extent_size

commit 7d7672bc5d1038c745716c397d892d21e29de71c upstream

If count_max_extents() uses BTRFS_MAX_EXTENT_SIZE to calculate the number
of extents needed, btrfs release the metadata reservation too much on its
way to write out the data.

Now that BTRFS_MAX_EXTENT_SIZE is replaced with fs_info->max_extent_size,
convert count_max_extents() to use it instead, and fix the calculation of
the metadata reservation.

CC: stable@vger.kernel.org # 5.12+
Fixes: d8e3fb106f39 ("btrfs: zoned: use ZONE_APPEND write for zoned mode")
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

btrfs: replace BTRFS_MAX_EXTENT_SIZE with fs_info->max_extent_size

commit f7b12a62f008a3041f42f2426983e59a6a0a3c59 upstream

On zoned filesystem, data write out is limited by max_zone_append_size,
and a large ordered extent is split according the size of a bio. OTOH,
the number of extents to be written is calculated using
BTRFS_MAX_EXTENT_SIZE, and that estimated number is used to reserve the
metadata bytes to update and/or create the metadata items.

The metadata reservation is done at e.g, btrfs_buffered_write() and then
released according to the estimation changes. Thus, if the number of extent
increases massively, the reserved metadata can run out.

The increase of the number of extents easily occurs on zoned filesystem
if BTRFS_MAX_EXTENT_SIZE > max_zone_append_size. And, it causes the
following warning on a small RAM environment with disabling metadata
over-commit (in the following patch).

[75721.498492] ------------[ cut here ]------------
[75721.505624] BTRFS: block rsv 1 returned -28
[75721.512230] WARNING: CPU: 24 PID: 2327559 at fs/btrfs/block-rsv.c:537 btrfs_use_block_rsv+0x560/0x760 [btrfs]
[75721.581854] CPU: 24 PID: 2327559 Comm: kworker/u64:10 Kdump: loaded Tainted: G        W         5.18.0-rc2-BTRFS-ZNS+ #109
[75721.597200] Hardware name: Supermicro Super Server/H12SSL-NT, BIOS 2.0 02/22/2021
[75721.607310] Workqueue: btrfs-endio-write btrfs_work_helper [btrfs]
[75721.616209] RIP: 0010:btrfs_use_block_rsv+0x560/0x760 [btrfs]
[75721.646649] RSP: 0018:ffffc9000fbdf3e0 EFLAGS: 00010286
[75721.654126] RAX: 0000000000000000 RBX: 0000000000004000 RCX: 0000000000000000
[75721.663524] RDX: 0000000000000004 RSI: 0000000000000008 RDI: fffff52001f7be6e
[75721.672921] RBP: ffffc9000fbdf420 R08: 0000000000000001 R09: ffff889f8d1fc6c7
[75721.682493] R10: ffffed13f1a3f8d8 R11: 0000000000000001 R12: ffff88980a3c0e28
[75721.692284] R13: ffff889b66590000 R14: ffff88980a3c0e40 R15: ffff88980a3c0e8a
[75721.701878] FS:  0000000000000000(0000) GS:ffff889f8d000000(0000) knlGS:0000000000000000
[75721.712601] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[75721.720726] CR2: 000055d12e05c018 CR3: 0000800193594000 CR4: 0000000000350ee0
[75721.730499] Call Trace:
[75721.735166]  <TASK>
[75721.739886]  btrfs_alloc_tree_block+0x1e1/0x1100 [btrfs]
[75721.747545]  ? btrfs_alloc_logged_file_extent+0x550/0x550 [btrfs]
[75721.756145]  ? btrfs_get_32+0xea/0x2d0 [btrfs]
[75721.762852]  ? btrfs_get_32+0xea/0x2d0 [btrfs]
[75721.769520]  ? push_leaf_left+0x420/0x620 [btrfs]
[75721.776431]  ? memcpy+0x4e/0x60
[75721.781931]  split_leaf+0x433/0x12d0 [btrfs]
[75721.788392]  ? btrfs_get_token_32+0x580/0x580 [btrfs]
[75721.795636]  ? push_for_double_split.isra.0+0x420/0x420 [btrfs]
[75721.803759]  ? leaf_space_used+0x15d/0x1a0 [btrfs]
[75721.811156]  btrfs_search_slot+0x1bc3/0x2790 [btrfs]
[75721.818300]  ? lock_downgrade+0x7c0/0x7c0
[75721.824411]  ? free_extent_buffer.part.0+0x107/0x200 [btrfs]
[75721.832456]  ? split_leaf+0x12d0/0x12d0 [btrfs]
[75721.839149]  ? free_extent_buffer.part.0+0x14f/0x200 [btrfs]
[75721.846945]  ? free_extent_buffer+0x13/0x20 [btrfs]
[75721.853960]  ? btrfs_release_path+0x4b/0x190 [btrfs]
[75721.861429]  btrfs_csum_file_blocks+0x85c/0x1500 [btrfs]
[75721.869313]  ? rcu_read_lock_sched_held+0x16/0x80
[75721.876085]  ? lock_release+0x552/0xf80
[75721.881957]  ? btrfs_del_csums+0x8c0/0x8c0 [btrfs]
[75721.888886]  ? __kasan_check_write+0x14/0x20
[75721.895152]  ? do_raw_read_unlock+0x44/0x80
[75721.901323]  ? _raw_write_lock_irq+0x60/0x80
[75721.907983]  ? btrfs_global_root+0xb9/0xe0 [btrfs]
[75721.915166]  ? btrfs_csum_root+0x12b/0x180 [btrfs]
[75721.921918]  ? btrfs_get_global_root+0x820/0x820 [btrfs]
[75721.929166]  ? _raw_write_unlock+0x23/0x40
[75721.935116]  ? unpin_extent_cache+0x1e3/0x390 [btrfs]
[75721.942041]  btrfs_finish_ordered_io.isra.0+0xa0c/0x1dc0 [btrfs]
[75721.949906]  ? try_to_wake_up+0x30/0x14a0
[75721.955700]  ? btrfs_unlink_subvol+0xda0/0xda0 [btrfs]
[75721.962661]  ? rcu_read_lock_sched_held+0x16/0x80
[75721.969111]  ? lock_acquire+0x41b/0x4c0
[75721.974982]  finish_ordered_fn+0x15/0x20 [btrfs]
[75721.981639]  btrfs_work_helper+0x1af/0xa80 [btrfs]
[75721.988184]  ? _raw_spin_unlock_irq+0x28/0x50
[75721.994643]  process_one_work+0x815/0x1460
[75722.000444]  ? pwq_dec_nr_in_flight+0x250/0x250
[75722.006643]  ? do_raw_spin_trylock+0xbb/0x190
[75722.013086]  worker_thread+0x59a/0xeb0
[75722.018511]  kthread+0x2ac/0x360
[75722.023428]  ? process_one_work+0x1460/0x1460
[75722.029431]  ? kthread_complete_and_exit+0x30/0x30
[75722.036044]  ret_from_fork+0x22/0x30
[75722.041255]  </TASK>
[75722.045047] irq event stamp: 0
[75722.049703] hardirqs last  enabled at (0): [<0000000000000000>] 0x0
[75722.057610] hardirqs last disabled at (0): [<ffffffff8118a94a>] copy_process+0x1c1a/0x66b0
[75722.067533] softirqs last  enabled at (0): [<ffffffff8118a989>] copy_process+0x1c59/0x66b0
[75722.077423] softirqs last disabled at (0): [<0000000000000000>] 0x0
[75722.085335] ---[ end trace 0000000000000000 ]---

To fix the estimation, we need to introduce fs_info->max_extent_size to
replace BTRFS_MAX_EXTENT_SIZE, which allow setting the different size for
regular vs zoned filesystem.

Set fs_info->max_extent_size to BTRFS_MAX_EXTENT_SIZE by default. On zoned
filesystem, it is set to fs_info->max_zone_append_size.

CC: stable@vger.kernel.org # 5.12+
Fixes: d8e3fb106f39 ("btrfs: zoned: use ZONE_APPEND write for zoned mode")
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

btrfs: zoned: revive max_zone_append_bytes

commit c2ae7b772ef4e86c5ddf3fd47bf59045ae96a414 upstream

This patch is basically a revert of commit 5a80d1c6a270 ("btrfs: zoned:
remove max_zone_append_size logic"), but without unnecessary ASSERT and
check. The max_zone_append_size will be used as a hint to estimate the
number of extents to cover delalloc/writeback region in the later commits.

The size of a ZONE APPEND bio is also limited by queue_max_segments(), so
this commit considers it to calculate max_zone_append_size. Technically, a
bio can be larger than queue_max_segments() * PAGE_SIZE if the pages are
contiguous. But, it is safe to consider "queue_max_segments() * PAGE_SIZE"
as an upper limit of an extent size to calculate the number of extents
needed to write data.

Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

block: add bdev_max_segments() helper

commit 65ea1b66482f415d51cd46515b02477257330339 upstream

Add bdev_max_segments() like other queue parameters.

Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Jens Axboe <axboe@kernel.dk>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Naohiro Aota <naohiro.aota@wdc.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

block: add a bdev_max_zone_append_sectors helper

commit 2aba0d19f4d8c8929b4b3b94a9cfde2aa20e6ee2 upstream

Add a helper to check the max supported sectors for zone append based on
the block_device instead of having to poke into the block layer internal
request_queue.

Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Martin K. Petersen <martin.petersen@oracle.com>
Reviewed-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Link: https://lore.kernel.org/r/20220415045258.199825-16-hch@lst.de
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

x86/entry: Move CLD to the start of the idtentry macro

commit c64cc2802a784ecfd25d39945e57e7a147854a5b upstream.

Move it after CLAC.

Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Lai Jiangshan <jiangshan.ljs@antgroup.com>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lore.kernel.org/r/20220503032107.680190-5-jiangshanlai@gmail.com
Signed-off-by: Juergen Gross <jgross@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

kernel/sys_ni: add compat entry for fadvise64_64

commit a8faed3a02eeb75857a3b5d660fa80fe79db77a3 upstream.

When CONFIG_ADVISE_SYSCALLS is not set/enabled and CONFIG_COMPAT is
set/enabled, the riscv compat_syscall_table references
'compat_sys_fadvise64_64', which is not defined:

riscv64-linux-ld: arch/riscv/kernel/compat_syscall_table.o:(.rodata+0x6f8):
undefined reference to `compat_sys_fadvise64_64'

Add 'fadvise64_64' to kernel/sys_ni.c as a conditional COMPAT function so
that when CONFIG_ADVISE_SYSCALLS is not set, there is a fallback function
available.

Link: https://lkml.kernel.org/r/20220807220934.5689-1-rdunlap@infradead.org
Fixes: d3ac21cacc24 ("mm: Support compiling out madvise and fadvise")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Suggested-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Paul Walmsley <paul.walmsley@sifive.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>
Cc: Albert Ou <aou@eecs.berkeley.edu>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

parisc: Fix exception handler for fldw and fstw instructions

commit 7ae1f5508d9a33fd58ed3059bd2d569961e3b8bd upstream.

The exception handler is broken for unaligned memory acceses with fldw
and fstw instructions, because it trashes or uses randomly some other
floating point register than the one specified in the instruction word
on loads and stores.

The instruction "fldw 0(addr),%fr22L" (and the other fldw/fstw
instructions) encode the target register (%fr22) in the rightmost 5 bits
of the instruction word. The 7th rightmost bit of the instruction word
defines if the left or right half of %fr22 should be used.

While processing unaligned address accesses, the FR3() define is used to
extract the offset into the local floating-point register set.  But the
calculation in FR3() was buggy, so that for example instead of %fr22,
register %fr12 [((22 * 2) & 0x1f) = 12] was used.

This bug has been since forever in the parisc kernel and I wonder why it
wasn't detected earlier. Interestingly I noticed this bug just because
the libime debian package failed to build on *native* hardware, while it
successfully built in qemu.

This patch corrects the bitshift and masking calculation in FR3().

Signed-off-by: Helge Deller <deller@gmx.de>
Cc: <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

parisc: Make CONFIG_64BIT available for ARCH=parisc64 only

commit 3dcfb729b5f4a0c9b50742865cd5e6c4dbcc80dc upstream.

With this patch the ARCH= parameter decides if the
CONFIG_64BIT option will be set or not. This means, the
ARCH= parameter will give:

ARCH=parisc -> 32-bit kernel
ARCH=parisc64 -> 64-bit kernel

This simplifies the usage of the other config options like
randconfig, allmodconfig and allyesconfig a lot and produces
the output which is expected for parisc64 (64-bit) vs. parisc (32-bit).

Suggested-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Helge Deller <deller@gmx.de>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Cc: <stable@vger.kernel.org> # 5.15+
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

cgroup: Fix race condition at rebind_subsystems()

commit 763f4fb76e24959c370cdaa889b2492ba6175580 upstream.

Root cause:
The rebind_subsystems() is no lock held when move css object from A
list to B list,then let B's head be treated as css node at
list_for_each_entry_rcu().

Solution:
Add grace period before invalidating the removed rstat_css_node.

Reported-by: Jing-Ting Wu <jing-ting.wu@mediatek.com>
Suggested-by: Michal Koutný <mkoutny@suse.com>
Signed-off-by: Jing-Ting Wu <jing-ting.wu@mediatek.com>
Tested-by: Jing-Ting Wu <jing-ting.wu@mediatek.com>
Link: https://lore.kernel.org/linux-arm-kernel/d8f0bc5e2fb6ed259f9334c83279b4c011283c41.camel@mediatek.com/T/
Acked-by: Mukesh Ojha <quic_mojha@quicinc.com>
Fixes: a7df69b81aac ("cgroup: rstat: support cgroup1")
Cc: stable@vger.kernel.org # v5.13+
Signed-off-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

audit: fix potential double free on error path from fsnotify_add_inode_mark

commit ad982c3be4e60c7d39c03f782733503cbd88fd2a upstream.

Audit_alloc_mark() assign pathname to audit_mark->path, on error path
from fsnotify_add_inode_mark(), fsnotify_put_mark will free memory
of audit_mark->path, but the caller of audit_alloc_mark will free
the pathname again, so there will be double free problem.

Fix this by resetting audit_mark->path to NULL pointer on error path
from fsnotify_add_inode_mark().

Cc: stable@vger.kernel.org
Fixes: 7b1293234084d ("fsnotify: Add group pointer in fsnotify_init_mark()")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

eth: sun: cassini: remove dead code

commit 32329216ca1d6ee29c41215f18b3053bb6158541 upstream.

Fixes the following GCC warning:

drivers/net/ethernet/sun/cassini.c:1316:29: error: comparison between two arrays [-Werror=array-compare]
drivers/net/ethernet/sun/cassini.c:3783:34: error: comparison between two arrays [-Werror=array-compare]

Note that 2 arrays should be compared by comparing of their addresses:
note: use ‘&cas_prog_workaroundtab[0] == &cas_prog_null[0]’ to compare the addresses

Signed-off-by: Martin Liska <mliska@suse.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>
Cc: "Sudip Mukherjee (Codethink)" <sudipm.mukherjee@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

wifi: rtlwifi: remove always-true condition pointed out by GCC 12

commit ee3db469dd317e82f57b13aa3bc61be5cb60c2b4 upstream.

The .value is a two-dim array, not a pointer.

struct iqk_matrix_regs {
bool iqk_done;
        long value[1][IQK_MATRIX_REG_NUM];
};

Acked-by: Kalle Valo <kvalo@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Cc: "Sudip Mukherjee (Codethink)" <sudipm.mukherjee@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

Linux 5.15.63

Link: https://lore.kernel.org/r/20220823080059.091088642@linuxfoundation.org
Tested-by: Shuah Khan <skhan@linuxfoundation.org>
Link: https://lore.kernel.org/r/20220824065913.068916566@linuxfoundation.org
Tested-by: Ron Economos <re@w6rz.net>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Tested-by: Linux Kernel Functional Testing <lkft@linaro.org>
Tested-by: Sudip Mukherjee <sudip.mukherjee@codethink.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: reject crazy array sizes being fed to XFS_IOC_GETBMAP*

[ Upstream commit 29d650f7e3ab55283b89c9f5883d0c256ce478b5 ]

Syzbot tripped over the following complaint from the kernel:

WARNING: CPU: 2 PID: 15402 at mm/util.c:597 kvmalloc_node+0x11e/0x125 mm/util.c:597

While trying to run XFS_IOC_GETBMAP against the following structure:

struct getbmap fubar = {
.bmv_count = 0x22dae649,
};

Obviously, this is a crazy huge value since the next thing that the
ioctl would do is allocate 37GB of memory.  This is enough to make
kvmalloc mad, but isn't large enough to trip the validation functions.
In other words, I'm fussing with checks that were **already sufficient**
because that's easier than dealing with 644 internal bug reports.  Yes,
that's right, six hundred and forty-four.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Allison Henderson <allison.henderson@oracle.com>
Reviewed-by: Catherine Hoang <catherine.hoang@oracle.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: revert "xfs: actually bump warning counts when we send warnings"

[ Upstream commit bc37e4fb5cac2925b2e286b1f1d4fc2b519f7d92 ]

This reverts commit 4b8628d57b725b32616965e66975fcdebe008fe7.

XFS quota has had the concept of a "quota warning limit" since
the earliest Irix implementation, but a mechanism for incrementing
the warning counter was never implemented, as documented in the
xfs_quota(8) man page. We do know from the historical archive that
it was never incremented at runtime during quota reservation
operations.

With this commit, the warning counter quickly increments for every
allocation attempt after the user has crossed a quote soft
limit threshold, and this in turn transitions the user to hard
quota failures, rendering soft quota thresholds and timers useless.
This was reported as a regression by users.

Because the intended behavior of this warning counter has never been
understood or documented, and the result of this change is a regression
in soft quota functionality, revert this commit to make soft quota
limits and timers operable again.

Fixes: 4b8628d57b72 ("xfs: actually bump warning counts when we send warnings)
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: fix soft lockup via spinning in filestream ag selection loop

[ Upstream commit f650df7171b882dca737ddbbeb414100b31f16af ]

The filestream AG selection loop uses pagf data to aid in AG
selection, which depends on pagf initialization. If the in-core
structure is not initialized, the caller invokes the AGF read path
to do so and carries on. If another task enters the loop and finds
a pagf init already in progress, the AGF read returns -EAGAIN and
the task continues the loop. This does not increment the current ag
index, however, which means the task spins on the current AGF buffer
until unlocked.

If the AGF read I/O submitted by the initial task happens to be
delayed for whatever reason, this results in soft lockup warnings
via the spinning task. This is reproduced by xfs/170. To avoid this
problem, fix the AGF trylock failure path to properly iterate to the
next AG. If a task iterates all AGs without making progress, the
trylock behavior is dropped in favor of blocking locks and thus a
soft lockup is no longer possible.

Fixes: f48e2df8a877ca1c ("xfs: make xfs_*read_agf return EAGAIN to ALLOC_FLAG_TRYLOCK callers")
Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: fix overfilling of reserve pool

[ Upstream commit 82be38bcf8a2e056b4c99ce79a3827fa743df6ec ]

Due to cycling of m_sb_lock, it's possible for multiple callers of
xfs_reserve_blocks to race at changing the pool size, subtracting blocks
from fdblocks, and actually putting it in the pool.  The result of all
this is that we can overfill the reserve pool to hilarious levels.

xfs_mod_fdblocks, when called with a positive value, already knows how
to take freed blocks and either fill the reserve until it's full, or put
them in fdblocks.  Use that instead of setting m_resblks_avail directly.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: always succeed at setting the reserve pool size

[ Upstream commit 0baa2657dc4d79202148be79a3dc36c35f425060 ]

Nowadays, xfs_mod_fdblocks will always choose to fill the reserve pool
with freed blocks before adding to fdblocks.  Therefore, we can change
the behavior of xfs_reserve_blocks slightly -- setting the target size
of the pool should always succeed, since a deficiency will eventually
be made up as blocks get freed.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: remove infinite loop when reserving free block pool

[ Upstream commit 15f04fdc75aaaa1cccb0b8b3af1be290e118a7bc ]

Infinite loops in kernel code are scary.  Calls to xfs_reserve_blocks
should be rare (people should just use the defaults!) so we really don't
need to try so hard.  Simplify the logic here by removing the infinite
loop.

Cc: Brian Foster <bfoster@redhat.com>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: reserve quota for target dir expansion when renaming files

[ Upstream commit 41667260bc84db4dfe566e3f6ab6da5293d60d8d ]

XFS does not reserve quota for directory expansion when renaming
children into a directory.  This means that we don't reject the
expansion with EDQUOT when we're at or near a hard limit, which means
that unprivileged userspace can use rename() to exceed quota.

Rename operations don't always expand the target directory, and we allow
a rename to proceed with no space reservation if we don't need to add a
block to the target directory to handle the addition.  Moreover, the
unlink operation on the source directory generally does not expand the
directory (you'd have to free a block and then cause a btree split) and
it's probably of little consequence to leave the corner case that
renaming a file out of a directory can increase its size.

As with link and unlink, there is a further bug in that we do not
trigger the blockgc workers to try to clear space when we're out of
quota.

Because rename is its own special tricky animal, we'll patch xfs_rename
directly to reserve quota to the rename transaction.  We'll leave
cleaning up the rest of xfs_rename for the metadata directory tree
patchset.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: reserve quota for dir expansion when linking/unlinking files

[ Upstream commit 871b9316e7a778ff97bdc34fdb2f2977f616651d ]

XFS does not reserve quota for directory expansion when linking or
unlinking children from a directory.  This means that we don't reject
the expansion with EDQUOT when we're at or near a hard limit, which
means that unprivileged userspace can use link()/unlink() to exceed
quota.

The fix for this is nuanced -- link operations don't always expand the
directory, and we allow a link to proceed with no space reservation if
we don't need to add a block to the directory to handle the addition.
Unlink operations generally do not expand the directory (you'd have to
free a block and then cause a btree split) and we can defer the
directory block freeing if there is no space reservation.

Moreover, there is a further bug in that we do not trigger the blockgc
workers to try to clear space when we're out of quota.

To fix both cases, create a new xfs_trans_alloc_dir function that
allocates the transaction, locks and joins the inodes, and reserves
quota for the directory.  If there isn't sufficient space or quota,
we'll switch the caller to reservationless mode.  This should prevent
quota usage overruns with the least restriction in functionality.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

xfs: flush inodegc workqueue tasks before cancel

[ Upstream commit 6191cf3ad59fda5901160633fef8e41b064a5246 ]

The xfs_inodegc_stop() helper performs a high level flush of pending
work on the percpu queues and then runs a cancel_work_sync() on each
of the percpu work tasks to ensure all work has completed before
returning.  While cancel_work_sync() waits for wq tasks to complete,
it does not guarantee work tasks have started. This means that the
_stop() helper can queue and instantly cancel a wq task without
having completed the associated work. This can be observed by
tracepoint inspection of a simple "rm -f <file>; fsfreeze -f <mnt>"
test:

xfs_destroy_inode: ... ino 0x83 ...
xfs_inode_set_need_inactive: ... ino 0x83 ...
xfs_inodegc_stop: ...
...
xfs_inodegc_start: ...
xfs_inodegc_worker: ...
xfs_inode_inactivating: ... ino 0x83 ...

The first few lines show that the inode is removed and need inactive
state set, but the inactivation work has not completed before the
inodegc mechanism stops. The inactivation doesn't actually occur
until the fs is unfrozen and the gc mechanism starts back up. Note
that this test requires fsfreeze to reproduce because xfs_freeze
indirectly invokes xfs_fs_statfs(), which calls xfs_inodegc_flush().

When this occurs, the workqueue try_to_grab_pending() logic first
tries to steal the pending bit, which does not succeed because the
bit has been set by queue_work_on(). Subsequently, it checks for
association of a pool workqueue from the work item under the pool
lock. This association is set at the point a work item is queued and
cleared when dequeued for processing. If the association exists, the
work item is removed from the queue and cancel_work_sync() returns
true. If the pwq association is cleared, the remove attempt assumes
the task is busy and retries (eventually returning false to the
caller after waiting for the work task to complete).

To avoid this race, we can flush each work item explicitly before
cancel. However, since the _queue_all() already schedules each
underlying work item, the workqueue level helpers are sufficient to
achieve the same ordering effect. E.g., the inodegc enabled flag
prevents scheduling any further work in the _stop() case. Use the
drain_workqueue() helper in this particular case to make the intent
a bit more self explanatory.

Signed-off-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Leah Rumancik <leah.rumancik@gmail.com>
Acked-by: Darrick J. Wong <djwong@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

scsi: ufs: ufs-mediatek: Fix build error and type mismatch

commit f54912b228a8df6c0133e31bc75628677bb8c6e5 upstream.

If CONFIG_PM_SLEEP is not set.

make ARCH=arm64 CROSS_COMPILE=aarch64-linux-gnu-, will fail:

drivers/ufs/host/ufs-mediatek.c: In function ‘ufs_mtk_vreg_fix_vcc’:
drivers/ufs/host/ufs-mediatek.c:688:46: warning: format ‘%u’ expects argument of type ‘unsigned int’, but argument 4 has type ‘long unsigned int’ [-Wformat=]
    snprintf(vcc_name, MAX_VCC_NAME, "vcc-opt%u", res.a1);
                                             ~^   ~~~~~~
                                             %lu
drivers/ufs/host/ufs-mediatek.c: In function ‘ufs_mtk_system_suspend’:
drivers/ufs/host/ufs-mediatek.c:1371:8: error: implicit declaration of function ‘ufshcd_system_suspend’; did you mean ‘ufs_mtk_system_suspend’? [-Werror=implicit-function-declaration]
  ret = ufshcd_system_suspend(dev);
        ^~~~~~~~~~~~~~~~~~~~~
        ufs_mtk_system_suspend
drivers/ufs/host/ufs-mediatek.c: In function ‘ufs_mtk_system_resume’:
drivers/ufs/host/ufs-mediatek.c:1386:9: error: implicit declaration of function ‘ufshcd_system_resume’; did you mean ‘ufs_mtk_system_resume’? [-Werror=implicit-function-declaration]
  return ufshcd_system_resume(dev);
         ^~~~~~~~~~~~~~~~~~~~
         ufs_mtk_system_resume
cc1: some warnings being treated as errors

The declaration of func "ufshcd_system_suspend()" depends on
CONFIG_PM_SLEEP, so the function wrapper ufs_mtk_system_suspend() should
wrapped by CONFIG_PM_SLEEP too.

Link: https://lore.kernel.org/r/20220619115432.205504-1-renzhijie2@huawei.com
Fixes: 3fd23b8dfb54 ("scsi: ufs: ufs-mediatek: Fix the timing of configuring device regulators")
Reported-by: Hulk Robot <hulkci@huawei.com>
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Ren Zhijie <renzhijie2@huawei.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
[only take the suspend/resume portion of the commit - gregkh]
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

can: j1939: j1939_sk_queue_activate_next_locked(): replace WARN_ON_ONCE with netdev_warn_once()

commit 8ef49f7f8244424adcf4a546dba4cbbeb0b09c09 upstream.

We should warn user-space that it is doing something wrong when trying
to activate sessions with identical parameters but WARN_ON_ONCE macro
can not be used here as it serves a different purpose.

So it would be good to replace it with netdev_warn_once() message.

Found by Linux Verification Center (linuxtesting.org) with Syzkaller.

Fixes: 9d71dd0c7009 ("can: add support of SAE J1939 protocol")
Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru>
Signed-off-by: Alexey Khoroshilov <khoroshilov@ispras.ru>
Acked-by: Oleksij Rempel <o.rempel@pengutronix.de>
Link: https://lore.kernel.org/all/20220729143655.1108297-1-pchelkin@ispras.ru
[mkl: fix indention]
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

MIPS: tlbex: Explicitly compare _PAGE_NO_EXEC against 0

[ Upstream commit 74de14fe05dd6b151d73cb0c73c8ec874cbdcde6 ]

When CONFIG_XPA is enabled, Clang warns:

  arch/mips/mm/tlbex.c:629:24: error: converting the result of '<<' to a boolean; did you mean '(1 << _PAGE_NO_EXEC_SHIFT) != 0'? [-Werror,-Wint-in-bool-context]
          if (cpu_has_rixi && !!_PAGE_NO_EXEC) {
                              ^
  arch/mips/include/asm/pgtable-bits.h:174:28: note: expanded from macro '_PAGE_NO_EXEC'
  # define _PAGE_NO_EXEC          (1 << _PAGE_NO_EXEC_SHIFT)
                                     ^
  arch/mips/mm/tlbex.c:2568:24: error: converting the result of '<<' to a boolean; did you mean '(1 << _PAGE_NO_EXEC_SHIFT) != 0'? [-Werror,-Wint-in-bool-context]
          if (!cpu_has_rixi || !_PAGE_NO_EXEC) {
                                ^
  arch/mips/include/asm/pgtable-bits.h:174:28: note: expanded from macro '_PAGE_NO_EXEC'
  # define _PAGE_NO_EXEC          (1 << _PAGE_NO_EXEC_SHIFT)
                                     ^
  2 errors generated.

_PAGE_NO_EXEC can be '0' or '1 << _PAGE_NO_EXEC_SHIFT' depending on the
build and runtime configuration, which is what the negation operators
are trying to convey. To silence the warning, explicitly compare against
0 so the result of the '<<' operator is not implicitly converted to a
boolean.

According to its documentation, GCC enables -Wint-in-bool-context with
-Wall but this warning is not visible when building the same
configuration with GCC. It appears GCC only warns when compiling C++,
not C, although the documentation makes no note of this:
https://godbolt.org/z/x39q3brxf

Reported-by: Sudip Mukherjee (Codethink) <sudipm.mukherjee@gmail.com>
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

video: fbdev: i740fb: Check the argument of i740_calc_vclk()

[ Upstream commit 40bf722f8064f50200b8c4f8946cd625b441dda9 ]

Since the user can control the arguments of the ioctl() from the user
space, under special arguments that may result in a divide-by-zero bug.

If the user provides an improper 'pixclock' value that makes the argumet
of i740_calc_vclk() less than 'I740_RFREQ_FIX', it will cause a
divide-by-zero bug in:
    drivers/video/fbdev/i740fb.c:353 p_best = min(15, ilog2(I740_MAX_VCO_FREQ / (freq / I740_RFREQ_FIX)));

The following log can reveal it:

divide error: 0000 [#1] PREEMPT SMP KASAN PTI
RIP: 0010:i740_calc_vclk drivers/video/fbdev/i740fb.c:353 [inline]
RIP: 0010:i740fb_decode_var drivers/video/fbdev/i740fb.c:646 [inline]
RIP: 0010:i740fb_set_par+0x163f/0x3b70 drivers/video/fbdev/i740fb.c:742
Call Trace:
 fb_set_var+0x604/0xeb0 drivers/video/fbdev/core/fbmem.c:1034
 do_fb_ioctl+0x234/0x670 drivers/video/fbdev/core/fbmem.c:1110
 fb_ioctl+0xdd/0x130 drivers/video/fbdev/core/fbmem.c:1189

Fix this by checking the argument of i740_calc_vclk() first.

Signed-off-by: Zheyu Ma <zheyuma97@gmail.com>
Signed-off-by: Helge Deller <deller@gmx.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

venus: pm_helpers: Fix warning in OPP during probe

[ Upstream commit 1d95af02f23031c2e1cca7607c514b86ce85bc6e ]

Fix the following WARN triggered during Venus driver probe on
5.19.0-rc8-next-20220728:

 WARNING: CPU: 7 PID: 339 at drivers/opp/core.c:2471 dev_pm_opp_set_config+0x49c/0x610
 Modules linked in: qcom_spmi_adc5 rtc_pm8xxx qcom_spmi_adc_tm5 leds_qcom_lpg led_class_multicolor
  qcom_pon qcom_vadc_common venus_core(+) qcom_spmi_temp_alarm v4l2_mem2mem videobuf2_v4l2 msm(+)
  videobuf2_common crct10dif_ce spi_geni_qcom snd_soc_sm8250 i2c_qcom_geni gpu_sched
  snd_soc_qcom_common videodev qcom_q6v5_pas soundwire_qcom drm_dp_aux_bus qcom_stats
  drm_display_helper qcom_pil_info soundwire_bus snd_soc_lpass_va_macro mc qcom_q6v5
  phy_qcom_snps_femto_v2 qcom_rng snd_soc_lpass_macro_common snd_soc_lpass_wsa_macro
  lpass_gfm_sm8250 slimbus qcom_sysmon qcom_common qcom_glink_smem qmi_helpers
  qcom_wdt mdt_loader socinfo icc_osm_l3 display_connector
  drm_kms_helper qnoc_sm8250 drm fuse ip_tables x_tables ipv6
 CPU: 7 PID: 339 Comm: systemd-udevd Not tainted 5.19.0-rc8-next-20220728 #4
 Hardware name: Qualcomm Technologies, Inc. Robotics RB5 (DT)
 pstate: 80400005 (Nzcv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
 pc : dev_pm_opp_set_config+0x49c/0x610
 lr : dev_pm_opp_set_config+0x58/0x610
 sp : ffff8000093c3710
 x29: ffff8000093c3710 x28: ffffbca3959d82b8 x27: ffff8000093c3d00
 x26: ffffbca3959d8e08 x25: ffff4396cac98118 x24: ffff4396c0e24810
 x23: ffff4396c4272c40 x22: ffff4396c0e24810 x21: ffff8000093c3810
 x20: ffff4396cac36800 x19: ffff4396cac96800 x18: 0000000000000000
 x17: 0000000000000003 x16: ffffbca3f4edf198 x15: 0000001cba64a858
 x14: 0000000000000180 x13: 000000000000017e x12: 0000000000000000
 x11: 0000000000000002 x10: 0000000000000a60 x9 : ffff8000093c35c0
 x8 : ffff4396c4273700 x7 : ffff43983efca6c0 x6 : ffff43983efca640
 x5 : 00000000410fd0d0 x4 : ffff4396c4272c40 x3 : ffffbca3f5d1e008
 x2 : 0000000000000000 x1 : ffff4396c2421600 x0 : ffff4396cac96860
 Call trace:
  dev_pm_opp_set_config+0x49c/0x610
  devm_pm_opp_set_config+0x18/0x70
  vcodec_domains_get+0xb8/0x1638 [venus_core]
  core_get_v4+0x1d8/0x218 [venus_core]
  venus_probe+0xf4/0x468 [venus_core]
  platform_probe+0x68/0xd8
  really_probe+0xbc/0x2a8
  __driver_probe_device+0x78/0xe0
  driver_probe_device+0x3c/0xf0
  __driver_attach+0x70/0x120
  bus_for_each_dev+0x70/0xc0
  driver_attach+0x24/0x30
  bus_add_driver+0x150/0x200
  driver_register+0x64/0x120
  __platform_driver_register+0x28/0x38
  qcom_venus_driver_init+0x24/0x1000 [venus_core]
  do_one_initcall+0x54/0x1c8
  do_init_module+0x44/0x1d0
  load_module+0x16c8/0x1aa0
  __do_sys_finit_module+0xbc/0x110
  __arm64_sys_finit_module+0x20/0x30
  invoke_syscall+0x44/0x108
  el0_svc_common.constprop.0+0xcc/0xf0
  do_el0_svc+0x2c/0xb8
  el0_svc+0x2c/0x88
  el0t_64_sync_handler+0xb8/0xc0
  el0t_64_sync+0x18c/0x190
  qcom-venus: probe of aa00000.video-codec failed with error -16

The fix is re-ordering the code related to OPP core. The OPP core
expects all configuration options to be provided before the OPP
table is added.

Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>
Suggested-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Stanimir Varbanov <stanimir.varbanov@linaro.org>
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/64: Init jump labels before parse_early_param()

[ Upstream commit ca829e05d3d4f728810cc5e4b468d9ebc7745eb3 ]

On 64-bit, calling jump_label_init() in setup_feature_keys() is too
late because static keys may be used in subroutines of
parse_early_param() which is again subroutine of early_init_devtree().

For example booting with "threadirqs":

  static_key_enable_cpuslocked(): static key '0xc000000002953260' used before call to jump_label_init()
  WARNING: CPU: 0 PID: 0 at kernel/jump_label.c:166 static_key_enable_cpuslocked+0xfc/0x120
  ...
  NIP static_key_enable_cpuslocked+0xfc/0x120
  LR  static_key_enable_cpuslocked+0xf8/0x120
  Call Trace:
    static_key_enable_cpuslocked+0xf8/0x120 (unreliable)
    static_key_enable+0x30/0x50
    setup_forced_irqthreads+0x28/0x40
    do_early_param+0xa0/0x108
    parse_args+0x290/0x4e0
    parse_early_options+0x48/0x5c
    parse_early_param+0x58/0x84
    early_init_devtree+0xd4/0x518
    early_setup+0xb4/0x214

So call jump_label_init() just before parse_early_param() in
early_init_devtree().

Suggested-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Zhouyi Zhou <zhouzhouyi@gmail.com>
[mpe: Add call trace to change log and minor wording edits.]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220726015747.11754-1-zhouzhouyi@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

smb3: check xattr value length earlier

[ Upstream commit 5fa2cffba0b82336a2244d941322eb1627ff787b ]

Coverity complains about assigning a pointer based on
value length before checking that value length goes
beyond the end of the SMB.  Although this is even more
unlikely as value length is a single byte, and the
pointer is not dereferenced until laterm, it is clearer
to check the lengths first.

Addresses-Coverity: 1467704 ("Speculative execution data leak")
Reviewed-by: Ronnie Sahlberg <lsahlber@redhat.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix to do sanity check on segment type in build_sit_entries()

[ Upstream commit 09beadf289d6e300553e60d6e76f13c0427ecab3 ]

As Wenqing Liu <wenqingliu0120@gmail.com> reported in bugzilla:

https://bugzilla.kernel.org/show_bug.cgi?id=216285

RIP: 0010:memcpy_erms+0x6/0x10
 f2fs_update_meta_page+0x84/0x570 [f2fs]
 change_curseg.constprop.0+0x159/0xbd0 [f2fs]
 f2fs_do_replace_block+0x5c7/0x18a0 [f2fs]
 f2fs_replace_block+0xeb/0x180 [f2fs]
 recover_data+0x1abd/0x6f50 [f2fs]
 f2fs_recover_fsync_data+0x12ce/0x3250 [f2fs]
 f2fs_fill_super+0x4459/0x6190 [f2fs]
 mount_bdev+0x2cf/0x3b0
 legacy_get_tree+0xed/0x1d0
 vfs_get_tree+0x81/0x2b0
 path_mount+0x47e/0x19d0
 do_mount+0xce/0xf0
 __x64_sys_mount+0x12c/0x1a0
 do_syscall_64+0x38/0x90
 entry_SYSCALL_64_after_hwframe+0x63/0xcd

The root cause is segment type is invalid, so in f2fs_do_replace_block(),
f2fs accesses f2fs_sm_info::curseg_array with out-of-range segment type,
result in accessing invalid curseg->sum_blk during memcpy in
f2fs_update_meta_page(). Fix this by adding sanity check on segment type
in build_sit_entries().

Reported-by: Wenqing Liu <wenqingliu0120@gmail.com>
Signed-off-by: Chao Yu <chao.yu@oppo.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

f2fs: fix to avoid use f2fs_bug_on() in f2fs_new_node_page()

[ Upstream commit 141170b759e03958f296033bb7001be62d1d363b ]

As Dipanjan Das <mail.dipanjan.das@gmail.com> reported, syzkaller
found a f2fs bug as below:

RIP: 0010:f2fs_new_node_page+0x19ac/0x1fc0 fs/f2fs/node.c:1295
Call Trace:
 write_all_xattrs fs/f2fs/xattr.c:487 [inline]
 __f2fs_setxattr+0xe76/0x2e10 fs/f2fs/xattr.c:743
 f2fs_setxattr+0x233/0xab0 fs/f2fs/xattr.c:790
 f2fs_xattr_generic_set+0x133/0x170 fs/f2fs/xattr.c:86
 __vfs_setxattr+0x115/0x180 fs/xattr.c:182
 __vfs_setxattr_noperm+0x125/0x5f0 fs/xattr.c:216
 __vfs_setxattr_locked+0x1cf/0x260 fs/xattr.c:277
 vfs_setxattr+0x13f/0x330 fs/xattr.c:303
 setxattr+0x146/0x160 fs/xattr.c:611
 path_setxattr+0x1a7/0x1d0 fs/xattr.c:630
 __do_sys_lsetxattr fs/xattr.c:653 [inline]
 __se_sys_lsetxattr fs/xattr.c:649 [inline]
 __x64_sys_lsetxattr+0xbd/0x150 fs/xattr.c:649
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x46/0xb0

NAT entry and nat bitmap can be inconsistent, e.g. one nid is free
in nat bitmap, and blkaddr in its NAT entry is not NULL_ADDR, it
may trigger BUG_ON() in f2fs_new_node_page(), fix it.

Reported-by: Dipanjan Das <mail.dipanjan.das@gmail.com>
Signed-off-by: Chao Yu <chao.yu@oppo.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ALSA: control: Use deferred fasync helper

[ Upstream commit 4a971e84a7ae10a38d875cd2d4e487c8d1682ca3 ]

For avoiding the potential deadlock via kill_fasync() call, use the
new fasync helpers to defer the invocation from the control API.  Note
that it's merely a workaround.

Another note: although we haven't received reports about the deadlock
with the control API, the deadlock is still potentially possible, and
it's better to align the behavior with other core APIs (PCM and
timer); so let's move altogether.

Link: https://lore.kernel.org/r/20220728125945.29533-5-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ALSA: timer: Use deferred fasync helper

[ Upstream commit 95cc637c1afd83fb7dd3d7c8a53710488f4caf9c ]

For avoiding the potential deadlock via kill_fasync() call, use the
new fasync helpers to defer the invocation from PCI API.  Note that
it's merely a workaround.

Reported-by: syzbot+1ee0910eca9c94f71f25@syzkaller.appspotmail.com
Reported-by: syzbot+49b10793b867871ee26f@syzkaller.appspotmail.com
Reported-by: syzbot+8285e973a41b5aa68902@syzkaller.appspotmail.com
Link: https://lore.kernel.org/r/20220728125945.29533-3-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ALSA: core: Add async signal helpers

[ Upstream commit ef34a0ae7a2654bc9e58675e36898217fb2799d8 ]

Currently the call of kill_fasync() from an interrupt handler might
lead to potential spin deadlocks, as spotted by syzkaller.
Unfortunately, it's not so trivial to fix this lock chain as it's
involved with the tasklist_lock that is touched in allover places.

As a temporary workaround, this patch provides the way to defer the
async signal notification in a work.  The new helper functions,
snd_fasync_helper() and snd_kill_faync() are replacements for
fasync_helper() and kill_fasync(), respectively.  In addition,
snd_fasync_free() needs to be called at the destructor of the relevant
file object.

Link: https://lore.kernel.org/r/20220728125945.29533-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/ioda/iommu/debugfs: Generate unique debugfs entries

[ Upstream commit d73b46c3c1449bf27f793b9d9ee86ed70c7a7163 ]

The iommu_table::it_index is a LIOBN which is not initialized on PowerNV
as it is not used except IOMMU debugfs where it is used for a node name.

This initializes it_index witn a unique number to avoid warnings and
have a node for every iommu_table.

This should not cause any behavioral change without CONFIG_IOMMU_DEBUGFS.

Signed-off-by: Alexey Kardashevskiy <aik@ozlabs.ru>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220714080800.3712998-1-aik@ozlabs.ru
Signed-off-by: Sasha Levin <sashal@kernel.org>

ovl: warn if trusted xattr creation fails

[ Upstream commit b10b85fe5149ee8b39fbbf86095b303632dde2cd ]

When mounting overlayfs in an unprivileged user namespace, trusted xattr
creation will fail.  This will lead to failures in some file operations,
e.g. in the following situation:

  mkdir lower upper work merged
  mkdir lower/directory
  mount -toverlay -olowerdir=lower,upperdir=upper,workdir=work none merged
  rmdir merged/directory
  mkdir merged/directory

The last mkdir will fail:

  mkdir: cannot create directory 'merged/directory': Input/output error

The cause for these failures is currently extremely non-obvious and hard to
debug.  Hence, warn the user and suggest using the userxattr mount option,
if it is not already supplied and xattr creation fails during the
self-check.

Reported-by: Alois Wohlschlager <alois1@gmx-topmail.de>
Signed-off-by: Miklos Szeredi <mszeredi@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/32: Don't always pass -mcpu=powerpc to the compiler

[ Upstream commit 446cda1b21d9a6b3697fe399c6a3a00ff4a285f5 ]

Since commit 4bf4f42a2feb ("powerpc/kbuild: Set default generic
machine type for 32-bit compile"), when building a 32 bits kernel
with a bi-arch version of GCC, or when building a book3s/32 kernel,
the option -mcpu=powerpc is passed to GCC at all time, relying on it
being eventually overriden by a subsequent -mcpu=xxxx.

But when building the same kernel with a 32 bits only version of GCC,
that is not done, relying on gcc being built with the expected default
CPU.

This logic has two problems. First, it is a bit fragile to rely on
whether the GCC version is bi-arch or not, because today we can have
bi-arch versions of GCC configured with a 32 bits default. Second,
there are some versions of GCC which don't support -mcpu=powerpc,
for instance for e500 SPE-only versions.

So, stop relying on this approximative logic and allow the user to
decide whether he/she wants to use the toolchain's default CPU or if
he/she wants to set one, and allow only possible CPUs based on the
selected target.

Reported-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Tested-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Segher Boessenkool <segher@kernel.crashing.org>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/d4df724691351531bf46d685d654689e5dfa0d74.1657549153.git.christophe.leroy@csgroup.eu
Signed-off-by: Sasha Levin <sashal@kernel.org>

powerpc/32: Set an IBAT covering up to _einittext during init

[ Upstream commit 2a0fb3c155c97c75176e557d61f8e66c1bd9b735 ]

Always set an IBAT covering up to _einittext during init because when
CONFIG_MODULES is not selected there is no reason to have an exception
handler for kernel instruction TLB misses.

It implies DBAT and IBAT are now totaly independent, IBATs are set
by setibat() and DBAT by setbat().

This allows to revert commit 9bb162fa26ed ("powerpc/603: Fix
boot failure with DEBUG_PAGEALLOC and KFENCE")

Reported-by: Maxime Bizon <mbizon@freebox.fr>
Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/ce7f04a39593934d9b1ee68c69144ccd3d4da4a1.1655202804.git.christophe.leroy@csgroup.eu
Signed-off-by: Sasha Levin <sashal@kernel.org>

watchdog: export lockup_detector_reconfigure

[ Upstream commit 7c56a8733d0a2a4be2438a7512566e5ce552fccf ]

In some circumstances it may be interesting to reconfigure the watchdog
from inside the kernel.

On PowerPC, this may helpful before and after a LPAR migration (LPM) is
initiated, because it implies some latencies, watchdog, and especially NMI
watchdog is expected to be triggered during this operation. Reconfiguring
the watchdog with a factor, would prevent it to happen too frequently
during LPM.

Rename lockup_detector_reconfigure() as __lockup_detector_reconfigure() and
create a new function lockup_detector_reconfigure() calling
__lockup_detector_reconfigure() under the protection of watchdog_mutex.

Signed-off-by: Laurent Dufour <ldufour@linux.ibm.com>
[mpe: Squash in build fix from Laurent, reported by Sachin]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220713154729.80789-3-ldufour@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

RISC-V: Add fast call path of crash_kexec()

[ Upstream commit 3f1901110a89b0e2e13adb2ac8d1a7102879ea98 ]

Currently, almost all archs (x86, arm64, mips...) support fast call
of crash_kexec() when "regs && kexec_should_crash()" is true. But
RISC-V not, it can only enter crash system via panic(). However panic()
doesn't pass the regs of the real accident scene to crash_kexec(),
it caused we can't get accurate backtrace via gdb,
$ riscv64-linux-gnu-gdb vmlinux vmcore
Reading symbols from vmlinux...
[New LWP 95]
#0  console_unlock () at kernel/printk/printk.c:2557
2557                    if (do_cond_resched)
(gdb) bt
#0  console_unlock () at kernel/printk/printk.c:2557
#1  0x0000000000000000 in ?? ()

With the patch we can get the accurate backtrace,
$ riscv64-linux-gnu-gdb vmlinux vmcore
Reading symbols from vmlinux...
[New LWP 95]
#0  0xffffffe00063a4e0 in test_thread (data=<optimized out>) at drivers/test_crash.c:81
81             *(int *)p = 0xdead;
(gdb)
(gdb) bt
#0  0xffffffe00064d5c0 in test_thread (data=<optimized out>) at drivers/test_crash.c:81
#1  0x0000000000000000 in ?? ()

Test code to produce NULL address dereference in test_crash.c,
void *p = NULL;
*(int *)p = 0xdead;

Reviewed-by: Guo Ren <guoren@kernel.org>
Tested-by: Xianting Tian <xianting.tian@linux.alibaba.com>
Signed-off-by: Xianting Tian <xianting.tian@linux.alibaba.com>
Link: https://lore.kernel.org/r/20220606082308.2883458-1-xianting.tian@linux.alibaba.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

riscv: mmap with PROT_WRITE but no PROT_READ is invalid

[ Upstream commit 2139619bcad7ac44cc8f6f749089120594056613 ]

As mentioned in Table 4.5 in RISC-V spec Volume 2 Section 4.3, write
but not read is "Reserved for future use.". For now, they are not valid.
In the current code, -wx is marked as invalid, but -w- is not marked
as invalid.
This patch refines that judgment.

Reported-by: xctan <xc-tan@outlook.com>
Co-developed-by: dram <dramforever@live.com>
Signed-off-by: dram <dramforever@live.com>
Co-developed-by: Ruizhe Pan <c141028@gmail.com>
Signed-off-by: Ruizhe Pan <c141028@gmail.com>
Signed-off-by: Celeste Liu <coelacanthus@outlook.com>
Link: https://lore.kernel.org/r/PH7PR14MB559464DBDD310E755F5B21E8CEDC9@PH7PR14MB5594.namprd14.prod.outlook.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

riscv: dts: canaan: Add k210 topology information

[ Upstream commit d9d193dea8666bbf69fc21c5bdcdabaa34a466e3 ]

The k210 has no cpu-map node, so tools like hwloc cannot correctly
parse the topology. Add the node using the existing node labels.

Reported-by: Brice Goglin <Brice.Goglin@inria.fr>
Link: https://github.com/open-mpi/hwloc/issues/536
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Reviewed-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Link: https://lore.kernel.org/r/20220705190435.1790466-6-mail@conchuod.ie
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

riscv: dts: sifive: Add fu740 topology information

[ Upstream commit bf6cd1c01c959a31002dfa6784c0d8caffed4cf1 ]

The fu740 has no cpu-map node, so tools like hwloc cannot correctly
parse the topology. Add the node using the existing node labels.

Reported-by: Brice Goglin <Brice.Goglin@inria.fr>
Link: https://github.com/open-mpi/hwloc/issues/536
Signed-off-by: Conor Dooley <conor.dooley@microchip.com>
Link: https://lore.kernel.org/r/20220705190435.1790466-4-mail@conchuod.ie
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ASoC: rsnd: care default case on rsnd_ssiu_busif_err_irq_ctrl()

[ Upstream commit ef30911d3c39fd57884c348c29b9cbff88def155 ]

Before, ssiu.c didn't care SSI5-8, thus,
commit b1384d4c95088d0 ("ASoC: rsnd: care default case on
rsnd_ssiu_busif_err_status_clear()") cares it for status clear.

But we should care it for error irq handling, too.
This patch cares it.

Reported-by: Nguyen Bao Nguyen <nguyen.nguyen.yj@renesas.com>
Reported-by: Nishiyama Kunihiko <kunihiko.nishiyama.dn@renesas.com>
Signed-off-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Link: https://lore.kernel.org/r/871quocio1.wl-kuninori.morimoto.gx@renesas.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

modules: Ensure natural alignment for .altinstructions and __bug_table sections

[ Upstream commit 87c482bdfa79f378297d92af49cdf265be199df5 ]

In the kernel image vmlinux.lds.S linker scripts the .altinstructions
and __bug_table sections are 4- or 8-byte aligned because they hold 32-
and/or 64-bit values.

Most architectures use altinstructions and BUG() or WARN() in modules as
well, but in the module linker script (module.lds.S) those sections are
currently missing. As consequence the linker will store their content
byte-aligned by default, which then can lead to unnecessary unaligned
memory accesses by the CPU when those tables are processed at runtime.

Usually unaligned memory accesses are unnoticed, because either the
hardware (as on x86 CPUs) or in-kernel exception handlers (e.g. on
parisc or sparc) emulate and fix them up at runtime. Nevertheless, such
unaligned accesses introduce a performance penalty and can even crash
the kernel if there is a bug in the unalignment exception handlers
(which happened once to me on the parisc architecture and which is why I
noticed that issue at all).

This patch fixes a non-critical issue and might be backported at any time.
It's trivial and shouldn't introduce any regression because it simply
tells the linker to use a different (8-byte alignment) for those
sections by default.

Signed-off-by: Helge Deller <deller@gmx.de>
Link: https://lore.kernel.org/all/Yr8%2Fgr8e8I7tVX4d@p100/
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

iommu/io-pgtable-arm-v7s: Add a quirk to allow pgtable PA up to 35bit

[ Upstream commit bfdd231374181254742c5e2faef0bef2d30c0ee4 ]

Single memory zone feature will remove ZONE_DMA32 and ZONE_DMA and
cause pgtable PA size larger than 32bit.

Since Mediatek IOMMU hardware support at most 35bit PA in pgtable,
so add a quirk to allow the PA of pgtables support up to bit35.

Signed-off-by: Ning Li <ning.li@mediatek.com>
Signed-off-by: Yunfei Wang <yf.wang@mediatek.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Acked-by: Will Deacon <will@kernel.org>
Link: https://lore.kernel.org/r/20220630092927.24925-2-yf.wang@mediatek.com
Signed-off-by: Joerg Roedel <jroedel@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mips: cavium-octeon: Fix missing of_node_put() in octeon2_usb_clocks_start

[ Upstream commit 7a9f743ceead60ed454c46fbc3085ee9a79cbebb ]

We should call of_node_put() for the reference 'uctl_node' returned by
of_get_parent() which will increase the refcount. Otherwise, there will
be a refcount leak bug.

Signed-off-by: Liang He <windhl@126.com>
Signed-off-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

vfio: Clear the caps->buf to NULL after free

[ Upstream commit 6641085e8d7b3f061911517f79a2a15a0a21b97b ]

On buffer resize failure, vfio_info_cap_add() will free the buffer,
report zero for the size, and return -ENOMEM.  As additional
hardening, also clear the buffer pointer to prevent any chance of a
double free.

Signed-off-by: Schspa Shi <schspa@gmail.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Link: https://lore.kernel.org/r/20220629022948.55608-1-schspa@gmail.com
Signed-off-by: Alex Williamson <alex.williamson@redhat.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

KVM: PPC: Book3S HV: Fix "rm_exit" entry in debugfs timings

[ Upstream commit 9981bace85d816ed8724ac46e49285e8488d29e6 ]

At debugfs/kvm/<pid>/vcpu0/timings we show how long each part of the
code takes to run:

$ cat /sys/kernel/debug/kvm/*-*/vcpu0/timings
rm_entry: 123785 49398892 118 4898
rm_intr: 123780 6075890 22 390
rm_exit: 0 0 0 0                     <-- NOK
guest: 123780 46732919988 402 9997638
cede: 0 0 0 0                        <-- OK, no cede napping in P9

The "rm_exit" is always showing zero because it is the last one and
end_timing does not increment the counter of the previous entry.

We can fix it by calling accumulate_time again instead of
end_timing. That way the counter gets incremented. The rest of the
arithmetic can be ignored because there are no timing points after
this and the accumulators are reset before the next round.

Signed-off-by: Fabiano Rosas <farosas@linux.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20220525130554.2614394-2-farosas@linux.ibm.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

tty: serial: Fix refcount leak bug in ucc_uart.c

[ Upstream commit d24d7bb2cd947676f9b71fb944d045e09b8b282f ]

In soc_info(), of_find_node_by_type() will return a node pointer
with refcount incremented. We should use of_node_put() when it is
not used anymore.

Acked-by: Timur Tabi <timur@kernel.org>
Signed-off-by: Liang He <windhl@126.com>
Link: https://lore.kernel.org/r/20220618060850.4058525-1-windhl@126.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

lib/list_debug.c: Detect uninitialized lists

[ Upstream commit 0cc011c576aaa4de505046f7a6c90933d7c749a9 ]

In some circumstances, attempts are made to add entries to or to remove
entries from an uninitialized list.  A prime example is
amdgpu_bo_vm_destroy(): It is indirectly called from
ttm_bo_init_reserved() if that function fails, and tries to remove an
entry from a list.  However, that list is only initialized in
amdgpu_bo_create_vm() after the call to ttm_bo_init_reserved() returned
success.  This results in crashes such as

 BUG: kernel NULL pointer dereference, address: 0000000000000000
 #PF: supervisor read access in kernel mode
 #PF: error_code(0x0000) - not-present page
 PGD 0 P4D 0
 Oops: 0000 [#1] PREEMPT SMP NOPTI
 CPU: 1 PID: 1479 Comm: chrome Not tainted 5.10.110-15768-g29a72e65dae5
 Hardware name: Google Grunt/Grunt, BIOS Google_Grunt.11031.149.0 07/15/2020
 RIP: 0010:__list_del_entry_valid+0x26/0x7d
 ...
 Call Trace:
  amdgpu_bo_vm_destroy+0x48/0x8b
  ttm_bo_init_reserved+0x1d7/0x1e0
  amdgpu_bo_create+0x212/0x476
  ? amdgpu_bo_user_destroy+0x23/0x23
  ? kmem_cache_alloc+0x60/0x271
  amdgpu_bo_create_vm+0x40/0x7d
  amdgpu_vm_pt_create+0xe8/0x24b
 ...

Check if the list's prev and next pointers are NULL to catch such problems.

Link: https://lkml.kernel.org/r/20220531222951.92073-1-linux@roeck-us.net
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ext4: avoid resizing to a partial cluster size

[ Upstream commit 69cb8e9d8cd97cdf5e293b26d70a9dee3e35e6bd ]

This patch avoids an attempt to resize the filesystem to an
unaligned cluster boundary.  An online resize to a size that is not
integral to cluster size results in the last iteration attempting to
grow the fs by a negative amount, which trips a BUG_ON and leaves the fs
with a corrupted in-memory superblock.

Signed-off-by: Oleg Kiselev <okiselev@amazon.com>
Link: https://lore.kernel.org/r/0E92A0AB-4F16-4F1A-94B7-702CC6504FDE@amazon.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sashal@kernel.org>

ext4: avoid remove directory when directory is corrupted

[ Upstream commit b24e77ef1c6d4dbf42749ad4903c97539cc9755a ]

Now if check directoy entry is corrupted, ext4_empty_dir may return true
then directory will be removed when file system mounted with "errors=continue".
In order not to make things worse just return false when directory is corrupted.

Signed-off-by: Ye Bin <yebin10@huawei.com>
Reviewed-by: Jan Kara <jack@suse.cz>
Link: https://lore.kernel.org/r/20220622090223.682234-1-yebin10@huawei.com
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drivers:md:fix a potential use-after-free bug

[ Upstream commit 104212471b1c1817b311771d817fb692af983173 ]

In line 2884, "raid5_release_stripe(sh);" drops the reference to sh and
may cause sh to be released. However, sh is subsequently used in lines
2886 "if (sh->batch_head && sh != sh->batch_head)". This may result in an
use-after-free bug.

It can be fixed by moving "raid5_release_stripe(sh);" to the bottom of
the function.

Signed-off-by: Wentao_Liang <Wentao_Liang_g@163.com>
Signed-off-by: Song Liu <song@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

nvmet-tcp: fix lockdep complaint on nvmet_tcp_wq flush during queue teardown

[ Upstream commit 533d2e8b4d5e4c89772a0adce913525fb86cbbee ]

We probably need nvmet_tcp_wq to have MEM_RECLAIM as we are
sending/receiving for the socket from works on this workqueue.
Also this eliminates lockdep complaints:
--
[ 6174.010200] workqueue: WQ_MEM_RECLAIM
nvmet-wq:nvmet_tcp_release_queue_work [nvmet_tcp] is flushing
!WQ_MEM_RECLAIM nvmet_tcp_wq:nvmet_tcp_io_work [nvmet_tcp]
[ 6174.010216] WARNING: CPU: 20 PID: 14456 at kernel/workqueue.c:2628
check_flush_dependency+0x110/0x14c

Reported-by: Yi Zhang <yi.zhang@redhat.com>
Signed-off-by: Sagi Grimberg <sagi@grimberg.me>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

md: Notify sysfs sync_completed in md_reap_sync_thread()

[ Upstream commit 9973f0fa7d20269fe6fefe6333997fb5914449c1 ]

The mdadm test 07layouts randomly produces a kernel hung task deadlock.
The deadlock is caused by the suspend_lo/suspend_hi files being set by
the mdadm background process during reshape and not being cleared
because the process hangs. (Leaving aside the issue of the fragility of
freezing kernel tasks by buggy userspace processes...)

When the background mdadm process hangs it, is waiting (without a
timeout) on a change to the sync_completed file signalling that the
reshape has completed. The process is woken up a couple times when
the reshape finishes but it is woken up before MD_RECOVERY_RUNNING
is cleared so sync_completed_show() reports 0 instead of "none".

To fix this, notify the sysfs file in md_reap_sync_thread() after
MD_RECOVERY_RUNNING has been cleared. This wakes up mdadm and causes
it to continue and write to suspend_lo/suspend_hi to allow IO to
continue.

Signed-off-by: Logan Gunthorpe <logang@deltatee.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Song Liu <song@kernel.org>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>

phy: samsung: phy-exynos-pcie: sanitize init/power_on callbacks

[ Upstream commit f2812227bb07e2eaee74253f11cea1576945df31 ]

The exynos-pcie driver called phy_power_on() before phy_init() for some
historical reasons. However the generic PHY framework assumes that the
proper sequence is to call phy_init() first, then phy_power_on(). The
operations done by both functions should be considered as one action and as
such they are called by the exynos-pcie driver (without doing anything
between them). The initialization is just a sequence of register writes,
which cannot be altered without breaking the hardware operation.

To match the generic PHY framework requirement, simply move all register
writes to the phy_init()/phy_exit() and drop power_on()/power_off()
callbacks. This way the driver will also work with the old (incorrect)
PHY initialization call sequence.

Link: https://lore.kernel.org/r/20220628220409.26545-1-m.szyprowski@samsung.com
Reported-by: Bjorn Helgaas <helgaas@kernel.org>
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Reviewed-by: Chanho Park <chanho61.park@samsung.com>
Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Acked-By: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

openrisc: io: Define iounmap argument as volatile

[ Upstream commit 52e0ea900202d23843daee8f7089817e81dd3dd7 ]

When OpenRISC enables PCI it allows for more drivers to be compiled
resulting in exposing the following with -Werror.

    drivers/video/fbdev/riva/fbdev.c: In function 'rivafb_probe':
    drivers/video/fbdev/riva/fbdev.c:2062:42: error:
    passing argument 1 of 'iounmap' discards 'volatile' qualifier from pointer target type

    drivers/video/fbdev/nvidia/nvidia.c: In function 'nvidiafb_probe':
    drivers/video/fbdev/nvidia/nvidia.c:1414:20: error:
    passing argument 1 of 'iounmap' discards 'volatile' qualifier from pointer target type

    drivers/scsi/aic7xxx/aic7xxx_osm.c: In function 'ahc_platform_free':
    drivers/scsi/aic7xxx/aic7xxx_osm.c:1231:41: error:
    passing argument 1 of 'iounmap' discards 'volatile' qualifier from pointer target type

Most architectures define the iounmap argument to be volatile.  To fix this
issue we do the same for OpenRISC.  This patch must go before PCI is enabled on
OpenRISC to avoid any compile failures.

Link: https://lore.kernel.org/lkml/20220729033728.GA2195022@roeck-us.net/
Reported-by: Guenter Roeck <linux@roeck-us.net>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Stafford Horne <shorne@gmail.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: sprd: Cleanup in .remove() after pm_runtime_get_sync() failed

[ Upstream commit 1e42f82cbec7b2cc4873751e7791e6611901c5fc ]

It's not allowed to quit remove early without cleaning up completely.
Otherwise this results in resource leaks that probably yield graver
problems later. Here for example some tasklets might survive the lifetime
of the sprd-dma device and access sdev which is freed after .remove()
returns.

As none of the device freeing requires an active device, just ignore the
return value of pm_runtime_get_sync().

Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
Reviewed-by: Baolin Wang <baolin.wang7@gmail.com>
Link: https://lore.kernel.org/r/20220721204054.323602-1-u.kleine-koenig@pengutronix.de
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

selftests/kprobe: Do not test for GRP/ without event failures

[ Upstream commit f5eab65ff2b76449286d18efc7fee3e0b72f7d9b ]

A new feature is added where kprobes (and other probes) do not need to
explicitly state the event name when creating a probe. The event name will
come from what is being attached.

That is:

  # echo 'p:foo/ vfs_read' > kprobe_events

Will no longer error, but instead create an event:

  # cat kprobe_events
 p:foo/p_vfs_read_0 vfs_read

This should not be tested as an error case anymore. Remove it from the
selftest as now this feature "breaks" the selftest as it no longer fails
as expected.

Link: https://lore.kernel.org/all/1656296348-16111-1-git-send-email-quic_linyyuan@quicinc.com/
Link: https://lkml.kernel.org/r/20220712161707.6dc08a14@gandalf.local.home
Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

csky/kprobe: reclaim insn_slot on kprobe unregistration

[ Upstream commit a2310c74d418deca0f1d749c45f1f43162510f51 ]

On kprobe registration kernel allocate one insn_slot for new kprobe,
but it forget to reclaim the insn_slot on unregistration, leading to a
potential leakage.

Reported-by: Chen Guokai <chenguokai17@mails.ucas.ac.cn>
Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Signed-off-by: Liao Chang <liaochang1@huawei.com>
Signed-off-by: Guo Ren <guoren@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

RDMA/rxe: Limit the number of calls to each tasklet

[ Upstream commit eff6d998ca297cb0b2e53b032a56cf8e04dd8b17 ]

Limit the maximum number of calls to each tasklet from rxe_do_task()
before yielding the cpu. When the limit is reached reschedule the tasklet
and exit the calling loop. This patch prevents one tasklet from consuming
100% of a cpu core and causing a deadlock or soft lockup.

Link: https://lore.kernel.org/r/20220630190425.2251-9-rpearsonhpe@gmail.com
Signed-off-by: Bob Pearson <rpearsonhpe@gmail.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: dw-axi-dmac: ignore interrupt if no descriptor

[ Upstream commit 820f5ce999d2f99961e88c16d65cd26764df0590 ]

If the channel has no descriptor and the interrupt is raised then the
kernel will OOPS. Check the result of vchan_next_desc() in the handler
axi_chan_block_xfer_complete() to avoid the error happening.

Signed-off-by: Ben Dooks <ben.dooks@sifive.com>
Link: https://lore.kernel.org/r/20220708170153.269991-4-ben.dooks@sifive.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

dmaengine: dw-axi-dmac: do not print NULL LLI during error

[ Upstream commit 86cb0defe0e275453bc39e856bb523eb425a6537 ]

During debugging we have seen an issue where axi_chan_dump_lli()
is passed a NULL LLI pointer which ends up causing an OOPS due
to trying to get fields from it. Simply print NULL LLI and exit
to avoid this.

Signed-off-by: Ben Dooks <ben.dooks@sifive.com>
Link: https://lore.kernel.org/r/20220708170153.269991-3-ben.dooks@sifive.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

um: add "noreboot" command line option for PANIC_TIMEOUT=-1 setups

[ Upstream commit dda520d07b95072a0b63f6c52a8eb566d08ea897 ]

QEMU has a -no-reboot option, which halts instead of reboots when the
guest asks to reboot. This is invaluable when used with
CONFIG_PANIC_TIMEOUT=-1 (and panic_on_warn), because it allows panics
and warnings to be caught immediately in CI. Implement this in UML too,
by way of a basic setup param.

Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Richard Weinberger <richard@nod.at>
Signed-off-by: Sasha Levin <sashal@kernel.org>

PCI/ACPI: Guard ARM64-specific mcfg_quirks

[ Upstream commit 40a6cc141b4b9580de140bcb3e893445708acc5d ]

Guard ARM64-specific quirks with CONFIG_ARM64 to avoid build errors,
since mcfg_quirks will be shared by more than one architectures.

Link: https://lore.kernel.org/r/20220714124216.1489304-2-chenhuacai@loongson.cn
Signed-off-by: Huacai Chen <chenhuacai@loongson.cn>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

cxl: Fix a memory leak in an error handling path

[ Upstream commit 3a15b45b5454da862376b5d69a4967f5c6fa1368 ]

A bitmap_zalloc() must be balanced by a corresponding bitmap_free() in the
error handling path of afu_allocate_irqs().

Acked-by: Andrew Donnellan <ajd@linux.ibm.com>
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/ce5869418f5838187946eb6b11a52715a93ece3d.1657566849.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

pinctrl: intel: Check against matching data instead of ACPI companion

[ Upstream commit c551bd81d198bf1dcd4398d5454acdc0309dbe77 ]

In some cases we may get a platform device that has ACPI companion
which is different to the pin control described in the ACPI tables.
This is primarily happens when device is instantiated by board file.

In order to allow this device being enumerated, refactor
intel_pinctrl_get_soc_data() to check the matching data instead of
ACPI companion.

Reported-by: Henning Schild <henning.schild@siemens.com>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Tested-by: Henning Schild <henning.schild@siemens.com>
Acked-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Mika Westerberg <mika.westerberg@linux.intel.com>
Acked-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Lee Jones <lee@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

mmc: tmio: avoid glitches when resetting

[ Upstream commit 2e586f8a5b0ed4a525014a692923ac96f6647816 ]

If we reset because of an error, we need to preserve values for the
clock frequency. Otherwise, glitches may be seen on the bus.

To achieve that, we introduce a 'preserve' parameter to the reset
function and the IP core specific reset callbacks to handle everything
accordingly.

Reported-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Tested-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20220625131722.1397-1-wsa@kernel.org
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

habanalabs/gaudi: mask constant value before cast

[ Upstream commit e3f49437a2e0221a387ecd192d742ae1434e1e3a ]

This fixes a sparse warning of
"cast truncates bits from constant value"

Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

habanalabs/gaudi: fix shift out of bounds

[ Upstream commit 01622098aeb05a5efbb727199bbc2a4653393255 ]

When validating NIC queues, queue offset calculation must be
performed only for NIC queues.

Signed-off-by: Ofir Bitton <obitton@habana.ai>
Reviewed-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Oded Gabbay <ogabbay@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

coresight: etm4x: avoid build failure with unrolled loops

[ Upstream commit 4d45bc82df667ad9e9cb8361830e54fc1264e993 ]

When the following configs are enabled:
* CORESIGHT
* CORESIGHT_SOURCE_ETM4X
* UBSAN
* UBSAN_TRAP

Clang fails assemble the kernel with the error:
<instantiation>:1:7: error: expected constant expression in '.inst' directive
.inst (0xd5200000|((((2) << 19) | ((1) << 16) | (((((((((((0x160 + (i * 4))))) >> 2))) >> 7) & 0x7)) << 12) | ((((((((((0x160 + (i * 4))))) >> 2))) & 0xf)) << 8) | (((((((((((0x160 + (i * 4))))) >> 2))) >> 4) & 0x7)) << 5)))|(.L__reg_num_x8))
      ^
drivers/hwtracing/coresight/coresight-etm4x-core.c:702:4: note: while in
macro instantiation
etm4x_relaxed_read32(csa, TRCCNTVRn(i));
^
drivers/hwtracing/coresight/coresight-etm4x.h:403:4: note: expanded from
macro 'etm4x_relaxed_read32'
read_etm4x_sysreg_offset((offset), false)))
^
drivers/hwtracing/coresight/coresight-etm4x.h:383:12: note: expanded
from macro 'read_etm4x_sysreg_offset'
__val = read_etm4x_sysreg_const_offset((offset));       \
        ^
drivers/hwtracing/coresight/coresight-etm4x.h:149:2: note: expanded from
macro 'read_etm4x_sysreg_const_offset'
READ_ETM4x_REG(ETM4x_OFFSET_TO_REG(offset))
^
drivers/hwtracing/coresight/coresight-etm4x.h:144:2: note: expanded from
macro 'READ_ETM4x_REG'
read_sysreg_s(ETM4x_REG_NUM_TO_SYSREG((reg)))
^
arch/arm64/include/asm/sysreg.h:1108:15: note: expanded from macro
'read_sysreg_s'
asm volatile(__mrs_s("%0", r) : "=r" (__val));                  \
             ^
arch/arm64/include/asm/sysreg.h:1074:2: note: expanded from macro '__mrs_s'
"       mrs_s " v ", " __stringify(r) "\n"                      \
 ^

Consider the definitions of TRCSSCSRn and TRCCNTVRn:
drivers/hwtracing/coresight/coresight-etm4x.h:56
 #define TRCCNTVRn(n)      (0x160 + (n * 4))
drivers/hwtracing/coresight/coresight-etm4x.h:81
 #define TRCSSCSRn(n)      (0x2A0 + (n * 4))

Where the macro parameter is expanded to i; a loop induction variable
from etm4_disable_hw.

When any compiler can determine that loops may be unrolled, then the
__builtin_constant_p check in read_etm4x_sysreg_offset() defined in
drivers/hwtracing/coresight/coresight-etm4x.h may evaluate to true. This
can lead to the expression `(0x160 + (i * 4))` being passed to
read_etm4x_sysreg_const_offset. Via the trace above, this is passed
through READ_ETM4x_REG, read_sysreg_s, and finally to __mrs_s where it
is string-ified and used directly in inline asm.

Regardless of which compiler or compiler options determine whether a
loop can or can't be unrolled, which determines whether
__builtin_constant_p evaluates to true when passed an expression using a
loop induction variable, it is NEVER safe to allow the preprocessor to
construct inline asm like:
  asm volatile (".inst (0x160 + (i * 4))" : "=r"(__val));
                                 ^ expected constant expression

Instead of read_etm4x_sysreg_offset() using __builtin_constant_p(), use
__is_constexpr from include/linux/const.h instead to ensure only
expressions that are valid integer constant expressions get passed
through to read_sysreg_s().

This is not a bug in clang; it's a potentially unsafe use of the macro
arguments in read_etm4x_sysreg_offset dependent on __builtin_constant_p.

Link: https://github.com/ClangBuiltLinux/linux/issues/1310
Reported-by: Arnd Bergmann <arnd@kernel.org>
Reported-by: Tao Zhang <quic_taozha@quicinc.com>
Signed-off-by: Nick Desaulniers <ndesaulniers@google.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Suzuki K Poulose <suzuki.poulose@arm.com>
Link: https://lore.kernel.org/r/20220708231520.3958391-1-ndesaulniers@google.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

gadgetfs: ep_io - wait until IRQ finishes

[ Upstream commit 04cb742d4d8f30dc2e83b46ac317eec09191c68e ]

after usb_ep_queue() if wait_for_completion_interruptible() is
interrupted we need to wait until IRQ gets finished.

Otherwise complete() from epio_complete() can corrupt stack.

Signed-off-by: Jozef Martiniak <jomajm@gmail.com>
Link: https://lore.kernel.org/r/20220708070645.6130-1-jomajm@gmail.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: lpfc: Fix possible memory leak when failing to issue CMF WQE

[ Upstream commit 2f67dc7970bce3529edce93a0a14234d88b3fcd5 ]

There is no corresponding free routine if lpfc_sli4_issue_wqe fails to
issue the CMF WQE in lpfc_issue_cmf_sync_wqe.

If ret_val is non-zero, then free the iocbq request structure.

Link: https://lore.kernel.org/r/20220701211425.2708-6-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: lpfc: Prevent buffer overflow crashes in debugfs with malformed user input

[ Upstream commit f8191d40aa612981ce897e66cda6a88db8df17bb ]

Malformed user input to debugfs results in buffer overflow crashes.  Adapt
input string lengths to fit within internal buffers, leaving space for NULL
terminators.

Link: https://lore.kernel.org/r/20220701211425.2708-3-jsmart2021@gmail.com
Co-developed-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: Justin Tee <justin.tee@broadcom.com>
Signed-off-by: James Smart <jsmart2021@gmail.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

clk: qcom: clk-alpha-pll: fix clk_trion_pll_configure description

[ Upstream commit 94bed9bb05c7850ff5d80b87cc29004901f37956 ]

After merging lucid and trion pll functions in commit 0b01489475c6
("clk: qcom: clk-alpha-pll: same regs and ops for trion and lucid")
the function clk_trion_pll_configure() is left with an old description
header, which results in a W=2 compile time warning, fix it.

Acked-by: Stephen Boyd <sboyd@kernel.org>
Reviewed-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Vladimir Zapolskiy <vladimir.zapolskiy@linaro.org>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220701062711.2757855-1-vladimir.zapolskiy@linaro.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

zram: do not lookup algorithm in backends table

[ Upstream commit dc89997264de565999a1cb55db3f295d3a8e457b ]

Always use crypto_has_comp() so that crypto can lookup module, call
usermodhelper to load the modules, wait for usermodhelper to finish and so
on.  Otherwise crypto will do all of these steps under CPU hot-plug lock
and this looks like too much stuff to handle under the CPU hot-plug lock.
Besides this can end up in a deadlock when usermodhelper triggers a code
path that attempts to lock the CPU hot-plug lock, that zram already holds.

An example of such deadlock:

- path A. zram grabs CPU hot-plug lock, execs /sbin/modprobe from crypto
  and waits for modprobe to finish

disksize_store
 zcomp_create
  __cpuhp_state_add_instance
   __cpuhp_state_add_instance_cpuslocked
    zcomp_cpu_up_prepare
     crypto_alloc_base
      crypto_alg_mod_lookup
       call_usermodehelper_exec
        wait_for_completion_killable
         do_wait_for_common
          schedule

- path B. async work kthread that brings in scsi device. It wants to
  register CPUHP states at some point, and it needs the CPU hot-plug
  lock for that, which is owned by zram.

async_run_entry_fn
 scsi_probe_and_add_lun
  scsi_mq_alloc_queue
   blk_mq_init_queue
    blk_mq_init_allocated_queue
     blk_mq_realloc_hw_ctxs
      __cpuhp_state_add_instance
       __cpuhp_state_add_instance_cpuslocked
        mutex_lock
         schedule

- path C. modprobe sleeps, waiting for all aync works to finish.

load_module
 do_init_module
  async_synchronize_full
   async_synchronize_cookie_domain
    schedule

[senozhatsky@chromium.org: add comment]
Link: https://lkml.kernel.org/r/20220624060606.1014474-1-senozhatsky@chromium.org
Link: https://lkml.kernel.org/r/20220622023501.517125-1-senozhatsky@chromium.org
Signed-off-by: Sergey Senozhatsky <senozhatsky@chromium.org>
Cc: Minchan Kim <minchan@kernel.org>
Cc: Nitin Gupta <ngupta@vflare.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

uacce: Handle parent device removal or parent driver module rmmod

[ Upstream commit 80fc671bcc0173836e9032b0c698ea74c13b9d7c ]

The uacce driver must deal with a possible removal of the parent device
or parent driver module rmmod at any time.

Although uacce_remove(), called on device removal and on driver unbind,
prevents future use of the uacce fops by removing the cdev, fops that
were called before that point may still be running.

Serialize uacce_fops_open() and uacce_remove() with uacce->mutex.
Serialize other fops against uacce_remove() with q->mutex.
Since we need to protect uacce_fops_poll() which gets called on the fast
path, replace uacce->queues_lock with q->mutex to improve scalability.
The other fops are only used during setup.

uacce_queue_is_valid(), checked under q->mutex or uacce->mutex, denotes
whether uacce_remove() has disabled all queues. If that is the case,
don't go any further since the parent device is being removed and
uacce->ops should not be called anymore.

Reported-by: Yang Shen <shenyang39@huawei.com>
Signed-off-by: Zhangfei Gao <zhangfei.gao@linaro.org>
Signed-off-by: Jean-Philippe Brucker <jean-philippe@linaro.org>
Link: https://lore.kernel.org/r/20220701034843.7502-1-zhangfei.gao@linaro.org
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

clk: qcom: ipq8074: dont disable gcc_sleep_clk_src

[ Upstream commit 1bf7305e79aab095196131bdc87a97796e0e3fac ]

Once the usb sleep clocks are disabled, clock framework is trying to
disable the sleep clock source also.

However, it seems that it cannot be disabled and trying to do so produces:
[  245.436390] ------------[ cut here ]------------
[  245.441233] gcc_sleep_clk_src status stuck at 'on'
[  245.441254] WARNING: CPU: 2 PID: 223 at clk_branch_wait+0x130/0x140
[  245.450435] Modules linked in: xhci_plat_hcd xhci_hcd dwc3 dwc3_qcom leds_gpio
[  245.456601] CPU: 2 PID: 223 Comm: sh Not tainted 5.18.0-rc4 #215
[  245.463889] Hardware name: Xiaomi AX9000 (DT)
[  245.470050] pstate: 204000c5 (nzCv daIF +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  245.474307] pc : clk_branch_wait+0x130/0x140
[  245.481073] lr : clk_branch_wait+0x130/0x140
[  245.485588] sp : ffffffc009f2bad0
[  245.489838] x29: ffffffc009f2bad0 x28: ffffff8003e6c800 x27: 0000000000000000
[  245.493057] x26: 0000000000000000 x25: 0000000000000000 x24: ffffff800226ef20
[  245.500175] x23: ffffffc0089ff550 x22: 0000000000000000 x21: ffffffc008476ad0
[  245.507294] x20: 0000000000000000 x19: ffffffc00965ac70 x18: fffffffffffc51a7
[  245.514413] x17: 68702e3030303837 x16: 3a6d726f6674616c x15: ffffffc089f2b777
[  245.521531] x14: ffffffc0095c9d18 x13: 0000000000000129 x12: 0000000000000129
[  245.528649] x11: 00000000ffffffea x10: ffffffc009621d18 x9 : 0000000000000001
[  245.535767] x8 : 0000000000000001 x7 : 0000000000017fe8 x6 : 0000000000000001
[  245.542885] x5 : ffffff803fdca6d8 x4 : 0000000000000000 x3 : 0000000000000027
[  245.550002] x2 : 0000000000000027 x1 : 0000000000000023 x0 : 0000000000000026
[  245.557122] Call trace:
[  245.564229]  clk_branch_wait+0x130/0x140
[  245.566490]  clk_branch2_disable+0x2c/0x40
[  245.570656]  clk_core_disable+0x60/0xb0
[  245.574561]  clk_core_disable+0x68/0xb0
[  245.578293]  clk_disable+0x30/0x50
[  245.582113]  dwc3_qcom_remove+0x60/0xc0 [dwc3_qcom]
[  245.585588]  platform_remove+0x28/0x60
[  245.590361]  device_remove+0x4c/0x80
[  245.594179]  device_release_driver_internal+0x1dc/0x230
[  245.597914]  device_driver_detach+0x18/0x30
[  245.602861]  unbind_store+0xec/0x110
[  245.607027]  drv_attr_store+0x24/0x40
[  245.610847]  sysfs_kf_write+0x44/0x60
[  245.614405]  kernfs_fop_write_iter+0x128/0x1c0
[  245.618052]  new_sync_write+0xc0/0x130
[  245.622391]  vfs_write+0x1d4/0x2a0
[  245.626123]  ksys_write+0x58/0xe0
[  245.629508]  __arm64_sys_write+0x1c/0x30
[  245.632895]  invoke_syscall.constprop.0+0x5c/0x110
[  245.636890]  do_el0_svc+0xa0/0x150
[  245.641488]  el0_svc+0x18/0x60
[  245.644872]  el0t_64_sync_handler+0xa4/0x130
[  245.647914]  el0t_64_sync+0x174/0x178
[  245.652340] ---[ end trace 0000000000000000 ]---

So, add CLK_IS_CRITICAL flag to the clock so that the kernel won't try
to disable the sleep clock.

Signed-off-by: Robert Marko <robimarko@gmail.com>
Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20220515210048.483898-10-robimarko@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>

vboxguest: Do not use devm for irq

[ Upstream commit 6169525b76764acb81918aa387ac168fb9a55575 ]

When relying on devm it doesn't get freed early enough which causes the
following warning when unloading the module:

[249348.837181] remove_proc_entry: removing non-empty directory 'irq/20', leaking at least 'vboxguest'
[249348.837219] WARNING: CPU: 0 PID: 6708 at fs/proc/generic.c:715 remove_proc_entry+0x119/0x140

[249348.837379] Call Trace:
[249348.837385]  unregister_irq_proc+0xbd/0xe0
[249348.837392]  free_desc+0x23/0x60
[249348.837396]  irq_free_descs+0x4a/0x70
[249348.837401]  irq_domain_free_irqs+0x160/0x1a0
[249348.837452]  mp_unmap_irq+0x5c/0x60
[249348.837458]  acpi_unregister_gsi_ioapic+0x29/0x40
[249348.837463]  acpi_unregister_gsi+0x17/0x30
[249348.837467]  acpi_pci_irq_disable+0xbf/0xe0
[249348.837473]  pcibios_disable_device+0x20/0x30
[249348.837478]  pci_disable_device+0xef/0x120
[249348.837482]  vbg_pci_remove+0x6c/0x70 [vboxguest]

Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Pascal Terjan <pterjan@google.com>
Link: https://lore.kernel.org/r/20220612133744.4030602-1-pterjan@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: dwc2: gadget: remove D+ pull-up while no vbus with usb-role-switch

[ Upstream commit db638c6500abaffb8f7770b2a69c40d003d54ae1 ]

When using usb-role-switch, D+ pull-up is set as soon as DTCL_SFTDISCON is
cleared, whatever the vbus valid signal state is. The pull-up should not
be set when vbus isn't present (this is determined by the drd controller).

This patch ensures that B-Session (so Peripheral role + vbus valid signal)
is valid before clearing the DCTL_SFTDISCON bit when role switch is used.
Keep original behavior when usb-role-switch isn't used.

Acked-by: Minas Harutyunyan <hminas@synopsys.com>
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Signed-off-by: Fabrice Gasnier <fabrice.gasnier@foss.st.com>
Link: https://lore.kernel.org/r/20220622160717.314580-1-fabrice.gasnier@foss.st.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: renesas: Fix refcount leak bug

[ Upstream commit 9d6d5303c39b8bc182475b22f45504106a07f086 ]

In usbhs_rza1_hardware_init(), of_find_node_by_name() will return
a node pointer with refcount incremented. We should use of_node_put()
when it is not used anymore.

Signed-off-by: Liang He <windhl@126.com>
Link: https://lore.kernel.org/r/20220618023205.4056548-1-windhl@126.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: host: ohci-ppc-of: Fix refcount leak bug

[ Upstream commit 40a959d7042bb7711e404ad2318b30e9f92c6b9b ]

In ohci_hcd_ppc_of_probe(), of_find_compatible_node() will return
a node pointer with refcount incremented. We should use of_node_put()
when it is not used anymore.

Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Liang He <windhl@126.com>
Link: https://lore.kernel.org/r/20220617034637.4003115-1-windhl@126.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

scsi: ufs: ufs-mediatek: Fix the timing of configuring device regulators

[ Upstream commit 3fd23b8dfb54d9b74eba6dfdd3225db3ac116785 ]

Currently the LPM configurations of device regulators may not work since
VCC is not disabled yet while ufs_mtk_vreg_set_lpm() is executed.

Fix this by changing the timing of invoking ufs_mtk_vreg_set_lpm().

Link: https://lore.kernel.org/r/20220616053725.5681-5-stanley.chu@mediatek.com
Reviewed-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Po-Wen Kao <powen.kao@mediatek.com>
Signed-off-by: Stanley Chu <stanley.chu@mediatek.com>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

clk: ti: Stop using legacy clkctrl names for omap4 and 5

[ Upstream commit 255584b138343d4a28c6d25bd82d04b09460d672 ]

With the addition of clock-output-names, we can now unify the internal
clock naming for omap4 and 5 to follow the other TI SoCs.

We are still using legacy clkctrl names for omap4 and 5 based on the clock
manager name which is wrong. Instead, we want to use the clkctrl clock
based naming.

We must now also drop the legacy TI_CLK_CLKCTRL_COMPAT quirk for the
clkctrl clock.

This change will allow further devicetree warning cleanup as already
done for am3/4 and dra7.

Cc: linux-clk@vger.kernel.org
Cc: Stephen Boyd <sboyd@kernel.org>
Cc: Tero Kristo <kristo@kernel.org>
Signed-off-by: Tony Lindgren <tony@atomide.com>
Link: https://lore.kernel.org/r/20220615064306.22254-1-tony@atomide.com
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

drm/meson: Fix overflow implicit truncation warnings

[ Upstream commit 98692f52c588225034cbff458622c2c06dfcb544 ]

Fix -Woverflow warnings for drm/meson driver which is a result
of moving arm64 custom MMIO accessor macros to asm-generic function
implementations giving a bonus type-checking now and uncovering these
overflow warnings.

drivers/gpu/drm/meson/meson_viu.c: In function ‘meson_viu_init’:
drivers/gpu/drm/meson/meson_registers.h:1826:48: error: large integer implicitly truncated to unsigned type [-Werror=overflow]
 #define  VIU_OSD_BLEND_REORDER(dest, src)      ((src) << (dest * 4))
                                                ^
drivers/gpu/drm/meson/meson_viu.c:472:18: note: in expansion of macro ‘VIU_OSD_BLEND_REORDER’
   writel_relaxed(VIU_OSD_BLEND_REORDER(0, 1) |
                  ^~~~~~~~~~~~~~~~~~~~~

Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Sai Prakash Ranjan <quic_saipraka@quicinc.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Neil Armstrong <narmstrong@baylibre.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

irqchip/tegra: Fix overflow implicit truncation warnings

[ Upstream commit 443685992bda9bb4f8b17fc02c9f6c60e62b1461 ]

Fix -Woverflow warnings for tegra irqchip driver which is a result
of moving arm64 custom MMIO accessor macros to asm-generic function
implementations giving a bonus type-checking now and uncovering these
overflow warnings.

drivers/irqchip/irq-tegra.c: In function ‘tegra_ictlr_suspend’:
drivers/irqchip/irq-tegra.c:151:18: warning: large integer implicitly truncated to unsigned type [-Woverflow]
   writel_relaxed(~0ul, ictlr + ICTLR_COP_IER_CLR);
                  ^

Suggested-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Sai Prakash Ranjan <quic_saipraka@quicinc.com>
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Cc: Marc Zyngier <maz@kernel.org>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>

PCI: aardvark: Fix reporting Slot capabilities on emulated bridge

[ Upstream commit bcdb6fd4f3e9ac1097698c8d8f56b70853b49873 ]

Slot capabilities are currently not reported because emulated bridge does
not report the PCI_EXP_FLAGS_SLOT flag.

Set PCI_EXP_FLAGS_SLOT to let the kernel know that PCI_EXP_SLT* registers
are supported.

Move setting of PCI_EXP_SLTCTL register from "dynamic" pcie_conf_read
function to static buffer as it is only statically filled the
PCI_EXP_SLTSTA_PDS flag and dynamic read callback is not needed for this
register.

Set Presence State Bit to 1 since there is no support for unplugging the
card and there is currently no platform able to detect presence of a card -
in such a case the bit needs to be set to 1.

Finally correctly set Physical Slot Number to 1 since there is only one
port and zero value is reserved for ports within the same silicon as Root
Port which is not our case for Aardvark HW.

Link: https://lore.kernel.org/r/20220524132827.8837-3-kabel@kernel.org
Signed-off-by: Pali Rohár <pali@kernel.org>
Signed-off-by: Marek Behún <kabel@kernel.org>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: gadget: uvc: call uvc uvcg_warn on completed status instead of uvcg_info

[ Upstream commit a725d0f6dfc5d3739d6499f30ec865305ba3544d ]

Likewise to the uvcvideo hostside driver, this patch is changing the
usb_request message of an non zero completion handler call from dev_info
to dev_warn.

Reviewed-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
Link: https://lore.kernel.org/r/20220529223848.105914-4-m.grzeschik@pengutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: gadget: uvc: calculate the number of request depending on framesize

[ Upstream commit 87d76b5f1d8eeb49efa16e2018e188864cbb9401 ]

The current limitation of possible number of requests being handled is
dependent on the gadget speed. It makes more sense to depend on the
typical frame size when calculating the number of requests. This patch
is changing this and is using the previous limits as boundaries for
reasonable minimum and maximum number of requests.

For a 1080p jpeg encoded video stream with a maximum imagesize of
e.g. 800kB with a maxburst of 8 and an multiplier of 1 the resulting
number of requests is calculated to 49.

        800768         1
nreqs = ------ * -------------- ~= 49
          2      (1024 * 8 * 1)

Tested-by: Dan Vacura <w36195@motorola.com>
Signed-off-by: Michael Grzeschik <m.grzeschik@pengutronix.de>
Link: https://lore.kernel.org/r/20220529223848.105914-2-m.grzeschik@pengutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: cdns3: fix random warning message when driver load

[ Upstream commit 8659ab3d936fcf0084676f98b75b317017aa8f82 ]

Warning log:
[    4.141392] Unexpected gfp: 0x4 (GFP_DMA32). Fixing up to gfp: 0xa20 (GFP_ATOMIC). Fix your code!
[    4.150340] CPU: 1 PID: 175 Comm: 1-0050 Not tainted 5.15.5-00039-g2fd9ae1b568c #20
[    4.158010] Hardware name: Freescale i.MX8QXP MEK (DT)
[    4.163155] Call trace:
[    4.165600]  dump_backtrace+0x0/0x1b0
[    4.169286]  show_stack+0x18/0x68
[    4.172611]  dump_stack_lvl+0x68/0x84
[    4.176286]  dump_stack+0x18/0x34
[    4.179613]  kmalloc_fix_flags+0x60/0x88
[    4.183550]  new_slab+0x334/0x370
[    4.186878]  ___slab_alloc.part.108+0x4d4/0x748
[    4.191419]  __slab_alloc.isra.109+0x30/0x78
[    4.195702]  kmem_cache_alloc+0x40c/0x420
[    4.199725]  dma_pool_alloc+0xac/0x1f8
[    4.203486]  cdns3_allocate_trb_pool+0xb4/0xd0

pool_alloc_page(struct dma_pool *pool, gfp_t mem_flags)
{
...
page = kmalloc(sizeof(*page), mem_flags);
page->vaddr = dma_alloc_coherent(pool->dev, pool->allocation,
 &page->dma, mem_flags);
...
}

kmalloc was called with mem_flags, which is passed down in
cdns3_allocate_trb_pool() and have GFP_DMA32 flags.
kmall_fix_flags() report warning.

GFP_DMA32 is not useful at all. dma_alloc_coherent() will handle
DMA memory region correctly by pool->dev. GFP_DMA32 can be removed
safely.

Signed-off-by: Frank Li <Frank.Li@nxp.com>
Link: https://lore.kernel.org/r/20220609154456.2871672-1-Frank.Li@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

usb: cdns3 fix use-after-free at workaround 2

[ Upstream commit 7d602f30149a117eea260208b1661bc404c21dfd ]

BUG: KFENCE: use-after-free read in __list_del_entry_valid+0x10/0xac

cdns3_wa2_remove_old_request()
{
...
kfree(priv_req->request.buf);
cdns3_gadget_ep_free_request(&priv_ep->endpoint, &priv_req->request);
list_del_init(&priv_req->list);
^^^ use after free
...
}

cdns3_gadget_ep_free_request() free the space pointed by priv_req,
but priv_req is used in the following list_del_init().

This patch move list_del_init() before cdns3_gadget_ep_free_request().

Signed-off-by: Frank Li <Frank.Li@nxp.com>
Signed-off-by: Faqiang Zhu <faqiang.zhu@nxp.com>
Link: https://lore.kernel.org/r/20220608190430.2814358-1-Frank.Li@nxp.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>

platform/chrome: cros_ec_proto: don't show MKBP version if unsupported

[ Upstream commit b36f0643ff14a2fb281b105418e4e73c9d7c11d0 ]

It wrongly showed the following message when it doesn't support MKBP:
"MKBP support version 4294967295".

Fix it.

Reviewed-by: Guenter Roeck <groeck@chromium.org>
Signed-off-by: Tzung-Bi Shih <tzungbi@kernel.org>
Link: https://lore.kernel.org/r/20220609084957.3684698-14-tzungbi@kernel.org
Signed-off-by: Sasha Levin <sashal@kernel.org>

PCI: Add ACS quirk for Broadcom BCM5750x NICs

[ Upstream commit afd306a65cedb9589564bdb23a0c368abc4215fd ]

The Broadcom BCM5750x NICs may be multi-function devices.  They do not
advertise ACS capability. Peer-to-peer transactions are not possible
between the individual functions, so it is safe to treat them as fully
isolated.

Add an ACS quirk for these devices so the functions can be in independent
IOMMU groups and attached individually to userspace applications using
VFIO.

Link: https://lore.kernel.org/r/1654796507-28610-1-git-send-email-michael.chan@broadcom.com
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>

HID: multitouch: new device class fix Lenovo X12 trackpad sticky

[ Upstream commit 54eed5c7b938dc4ef6b14d4ee048bbdafdbce352 ]

The trackpad of the given device sends continuous report of pointers
status as per wxn8 spec. However, the spec did not clarify when the
fingers are lifted so fast that between the interval of two report
frames fingers on pad reduced from >=2 to 0. The second last report
contains >=2 fingers with tip state 1 and the last report contains only
1 finger with tip state 0. Although this can happen unfrequently, a
  quick fix will be improve the consistency to 100%. A quick fix is to
disable MT_QUIRK_ALWAYS_VALID and enable MT_QUIRK_NOT_SEEN_MEANS_UP.

Test for hid-tools is added in [1]

In addition to this, I2C device 04CA:00B1 may also need similar class
but with MT_QUIRK_FORCE_MULTI_INPUT disabled (but it does not harm to
 enable it on non-multi-input device either). The respective owner has
been notified and a patch may coming soon after test.

[1]: https://gitlab.freedesktop.org/libevdev/hid-tools/-/merge_requests/130

Signed-off-by: Tao Jin <tao-j@outlook.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Sasha Levin <sashal@kernel.org>

KVM: arm64: Reject 32bit user PSTATE on asymmetric systems

[ Upstream commit b10d86fb8e46cc812171728bcd326df2f34e9ed5 ]

KVM does not support AArch32 EL0 on asymmetric systems. To that end,
prevent userspace from configuring a vCPU in such a state through
setting PSTATE.

It is already ABI that KVM rejects such a write on a system where
AArch32 EL0 is unsupported. Though the kernel's definition of a 32bit
system changed in commit 2122a833316f ("arm64: Allow mismatched
32-bit EL0 support"), KVM's did not.

Fixes: 2122a833316f ("arm64: Allow mismatched 32-bit EL0 support")
Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
Signed-off-by: Marc Zyngier <maz@kernel.org>
Link: https://lore.kernel.org/r/20220816192554.1455559-3-oliver.upton@linux.dev
Signed-off-by: Sasha Levin <sashal@kernel.org>