Add a final dma_wmb() barrier before triggering the transmit request
(TCCR_TSRQ) to ensure all descriptor and buffer writes are visible to
the DMA engine.
According to the hardware manual, a read-back operation is required
before writing to the doorbell register to guarantee completion of
previous writes. Instead of performing a dummy read, a dma_wmb() is
used to both enforce the same ordering semantics on the CPU side and
also to ensure completion of writes.
Ensure the TX descriptor type fields are published in a safe order so the
DMA engine never begins processing a descriptor chain before all descriptor
fields are fully initialised.
For multi-descriptor transmits the driver writes DT_FEND into the last
descriptor and DT_FSTART into the first. The DMA engine begins processing
when it observes DT_FSTART. Move the dma_wmb() barrier so it executes
immediately after DT_FEND and immediately before writing DT_FSTART
(and before DT_FSINGLE in the single-descriptor case). This guarantees
that all prior CPU writes to the descriptor memory are visible to the
device before DT_FSTART is seen.
This avoids a situation where compiler/CPU reordering could publish
DT_FSTART ahead of DT_FEND or other descriptor fields, allowing the DMA to
start on a partially initialised chain and causing corrupted transmissions
or TX timeouts. Such a failure was observed on RZ/G2L with an RT kernel as
transmit queue timeouts and device resets.
TX frames aren't padded and unknown memory is sent into the ether.
Theoretically, it isn't even guaranteed that the extra memory exists
and can be sent out, which could cause further problems. In practice,
I found that plenty of tailroom exists in the skb itself (in my test
with ping at least) and skb_padto() easily succeeds, so use it here.
In the event of -ENOMEM drop the frame like other drivers do.
The use of one more padding byte instead of a USB zero-length packet
is retained to avoid regression. I have a dodgy Etron xHCI controller
which doesn't seem to support sending ZLPs at all.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Cc: stable@vger.kernel.org Signed-off-by: Michal Pecio <michal.pecio@gmail.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20251014203528.3f9783c4.michal.pecio@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On all platforms set_clock_selection() writes to a GRF register. This
requires certain clocks running and thus should happen before the
clocks are disabled.
This has been noticed on RK3576 Sige5, which hangs during system suspend
when trying to suspend the second network interface. Note, that
suspending the first interface works, because the second device ensures
that the necessary clocks for the GRF are enabled.
Cc: stable@vger.kernel.org Fixes: 2f2b60a0ec28 ("net: ethernet: stmmac: dwmac-rk: Add gmac support for rk3588") Signed-off-by: Sebastian Reichel <sebastian.reichel@collabora.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20251014-rockchip-network-clock-fix-v1-1-c257b4afdf75@collabora.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Syzbot reported a potential lock inversion deadlock between
vsock_register_mutex and sk_lock-AF_VSOCK when vsock_linger() is called.
The issue was introduced by commit 687aa0c5581b ("vsock: Fix
transport_* TOCTOU") which added vsock_register_mutex locking in
vsock_assign_transport() around the transport->release() call, that can
call vsock_linger(). vsock_assign_transport() can be called with sk_lock
held. vsock_linger() calls sk_wait_event() that temporarily releases and
re-acquires sk_lock. During this window, if another thread hold
vsock_register_mutex while trying to acquire sk_lock, a circular
dependency is created.
Fix this by releasing vsock_register_mutex before calling
transport->release() and vsock_deassign_transport(). This is safe
because we don't need to hold vsock_register_mutex while releasing the
old transport, and we ensure the new transport won't disappear by
obtaining a module reference first via try_module_get().
The extent map cache can become stale when extents are moved or
defragmented, causing subsequent operations to see outdated extent flags.
This triggers a BUG_ON in ocfs2_refcount_cal_cow_clusters().
The problem occurs when:
1. copy_file_range() creates a reflinked extent with OCFS2_EXT_REFCOUNTED
2. ioctl(FITRIM) triggers ocfs2_move_extents()
3. __ocfs2_move_extents_range() reads and caches the extent (flags=0x2)
4. ocfs2_move_extent()/ocfs2_defrag_extent() calls __ocfs2_move_extent()
which clears OCFS2_EXT_REFCOUNTED flag on disk (flags=0x0)
5. The extent map cache is not invalidated after the move
6. Later write() operations read stale cached flags (0x2) but disk has
updated flags (0x0), causing a mismatch
7. BUG_ON(!(rec->e_flags & OCFS2_EXT_REFCOUNTED)) triggers
Fix by clearing the extent map cache after each extent move/defrag
operation in __ocfs2_move_extents_range(). This ensures subsequent
operations read fresh extent data from disk.
Link: https://lore.kernel.org/all/20251009142917.517229-1-kartikey406@gmail.com/T/ Link: https://lkml.kernel.org/r/20251009154903.522339-1-kartikey406@gmail.com Fixes: 53069d4e7695 ("Ocfs2/move_extents: move/defrag extents within a certain range.") Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com> Reported-by: syzbot+6fdd8fa3380730a4b22c@syzkaller.appspotmail.com Tested-by: syzbot+6fdd8fa3380730a4b22c@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?id=2959889e1f6e216585ce522f7e8bc002b46ad9e7 Reviewed-by: Mark Fasheh <mark@fasheh.com> Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com> Cc: Joel Becker <jlbec@evilplan.org> Cc: Junxiao Bi <junxiao.bi@oracle.com> Cc: Changwei Ge <gechangwei@live.cn> Cc: Jun Piao <piaojun@huawei.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
MIPS Malta platform code registers the PCI southbridge legacy port I/O
PS/2 keyboard range as a standard resource marked as busy. It prevents
the i8042 driver from registering as it fails to claim the resource in
a call to i8042_platform_init(). Consequently PS/2 keyboard and mouse
devices cannot be used with this platform.
Fix the issue by removing the busy marker from the standard reservation,
making the driver register successfully:
serio: i8042 KBD port at 0x60,0x64 irq 1
serio: i8042 AUX port at 0x60,0x64 irq 12
and the resource show up as expected among the legacy devices:
If the i8042 driver has not been configured, then the standard resource
will remain there preventing any conflicting dynamic assignment of this
PCI port I/O address range.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Maciej W. Rozycki <macro@orcam.me.uk> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Reviewed-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Acked-by: Thomas Bogendoerfer <tsbogend@alpha.franken.de> Cc: stable@vger.kernel.org Link: https://patch.msgid.link/alpine.DEB.2.21.2510211919240.8377@angie.orcam.me.uk Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix TCP_Server_Info::credits to be signed, just as echo_credits and
oplock_credits are. This also fixes what ought to get at least a
compilation warning if not an outright error in *get_credits_field() as a
pointer to the unsigned server->credits field is passed back as a pointer
to a signed int.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-cifs@vger.kernel.org Cc: stable@vger.kernel.org Acked-by: Paulo Alcantara (Red Hat) <pc@manguebit.org> Acked-by: Pavel Shilovskiy <pshilovskiy@microsoft.com> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Since the commit c1f3f9797c1f ("can: netlink: can_changelink(): fix NULL
pointer deref of struct can_priv::do_set_mode"), the automatic restart
delay can only be set for devices that implement the restart handler struct
can_priv::do_set_mode. As it makes no sense to configure a automatic
restart for devices that doesn't support it.
However, since systemd commit 13ce5d4632e3 ("network/can: properly handle
CAN.RestartSec=0") [1], systemd-networkd correctly handles a restart delay
of "0" (i.e. the restart is disabled). Which means that a disabled restart
is always configured in the kernel.
On systems with both changes active this causes that CAN interfaces that
don't implement a restart handler cannot be brought up by systemd-networkd.
Solve this problem by allowing a delay of "0" to be configured, even if the
device does not implement a restart handler.
If two competing threads enter alloc_slab_obj_exts(), and the one that
allocates the vector wins the cmpxchg(), the other thread that failed
allocation mistakenly assumes that slab->obj_exts is still empty due to
its own allocation failure. This will then trigger warnings with
CONFIG_MEM_ALLOC_PROFILING_DEBUG checks in the subsequent free path.
Therefore, let's check the result of cmpxchg() to see if marking the
allocation as failed was successful. If it wasn't, check whether the
winning side has succeeded its allocation (it might have been also
marking it as failed) and if yes, return success.
Suggested-by: Harry Yoo <harry.yoo@oracle.com> Fixes: f7381b911640 ("slab: mark slab->obj_exts allocation failures unconditionally") Cc: <stable@vger.kernel.org> Signed-off-by: Hao Ge <gehao@kylinos.cn> Link: https://patch.msgid.link/20251023143313.1327968-1-hao.ge@linux.dev Reviewed-by: Suren Baghdasaryan <surenb@google.com> Reviewed-by: Harry Yoo <harry.yoo@oracle.com> Signed-off-by: Vlastimil Babka <vbabka@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If two competing threads enter alloc_slab_obj_exts() and one of them
fails to allocate the object extension vector, it might override the
valid slab->obj_exts allocated by the other thread with
OBJEXTS_ALLOC_FAIL. This will cause the thread that lost this race and
expects a valid pointer to dereference a NULL pointer later on.
Update slab->obj_exts atomically using cmpxchg() to avoid
slab->obj_exts overrides by racing threads.
Thanks for Vlastimil and Suren's help with debugging.
It is reported that commit 85975daeaa4d ("cpuidle: menu: Avoid discarding
useful information") led to a performance regression on Intel Jasper Lake
systems because it reduced the time spent by CPUs in idle state C7 which
is correlated to the maximum frequency the CPUs can get to because of an
average running power limit [1].
Before that commit, get_typical_interval() would have returned UINT_MAX
whenever it had been unable to make a high-confidence prediction which
had led to selecting the deepest available idle state too often and
both power and performance had been inadequate as a result of that on
some systems. However, this had not been a problem on systems with
relatively aggressive average running power limits, like the Jasper Lake
systems in question, because on those systems it was compensated by the
ability to run CPUs faster.
It was addressed by causing get_typical_interval() to return a number
based on the recent idle duration information available to it even if it
could not make a high-confidence prediction, but that clearly did not
take the possible correlation between idle power and available CPU
capacity into account.
For this reason, revert most of the changes made by commit 85975daeaa4d,
except for one cosmetic cleanup, and add a comment explaining the
rationale for returning UINT_MAX from get_typical_interval() when it
is unable to make a high-confidence prediction.
Fixes: 85975daeaa4d ("cpuidle: menu: Avoid discarding useful information") Closes: https://lore.kernel.org/linux-pm/36iykr223vmcfsoysexug6s274nq2oimcu55ybn6ww4il3g3cv@cohflgdbpnq7/ [1] Reported-by: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: All applicable <stable@vger.kernel.org> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Link: https://patch.msgid.link/3663603.iIbC2pHGDl@rafael.j.wysocki Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On a filesystem with parent pointers, xchk_nlinks_collect_dir walks both
the directory entries (data fork) and the parent pointers (attr fork) to
determine the correct link count. Unfortunately I forgot to update the
lock mode logic to handle the case of a directory whose attr fork is in
btree format and has not yet been loaded *and* whose data fork doesn't
need loading.
This leads to a bunch of assertions from xfs/286 in xfs_iread_extents
because we only took ILOCK_SHARED, not ILOCK_EXCL. You'd need the rare
happenstance of a directory with a large number of non-pptr extended
attributes set and enough memory pressure to cause the directory to be
evicted and partially reloaded from disk.
I /think/ this only started in 6.18-rc1 because I've started seeing OOM
errors with the maple tree slab using 70% of memory, and this didn't
happen in 6.17. Yay dynamic systems!
Cc: stable@vger.kernel.org # v6.10 Fixes: 77ede5f44b0d86 ("xfs: walk directory parent pointers to determine backref count") Signed-off-by: Darrick J. Wong <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Carlos Maiolino <cem@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Attempting to load the 104-idio-16 module fails during regmap
initialization with a return error -EINVAL. This is a result of the
regmap cache failing initialization. Set the idio_16_regmap_config
max_register member to fix this failure.
Fixes: 2c210c9a34a3 ("gpio: 104-idio-16: Migrate to the regmap API") Reported-by: Mark Cave-Ayland <mark.caveayland@nutanix.com> Closes: https://lore.kernel.org/r/9b0375fd-235f-4ee1-a7fa-daca296ef6bf@nutanix.com Suggested-by: Mark Cave-Ayland <mark.caveayland@nutanix.com> Cc: stable@vger.kernel.org Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: William Breathitt Gray <wbg@kernel.org> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Link: https://lore.kernel.org/r/20251020-fix-gpio-idio-16-regmap-v2-1-ebeb50e93c33@kernel.org Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Attempting to load the pci-idio-16 module fails during regmap
initialization with a return error -EINVAL. This is a result of the
regmap cache failing initialization. Set the idio_16_regmap_config
max_register member to fix this failure.
Fixes: 73d8f3efc5c2 ("gpio: pci-idio-16: Migrate to the regmap API") Reported-by: Mark Cave-Ayland <mark.caveayland@nutanix.com> Closes: https://lore.kernel.org/r/9b0375fd-235f-4ee1-a7fa-daca296ef6bf@nutanix.com Suggested-by: Mark Cave-Ayland <mark.caveayland@nutanix.com> Cc: stable@vger.kernel.org Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com> Signed-off-by: William Breathitt Gray <wbg@kernel.org> Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Link: https://lore.kernel.org/r/20251020-fix-gpio-idio-16-regmap-v2-2-ebeb50e93c33@kernel.org Signed-off-by: Bartosz Golaszewski <bartosz.golaszewski@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If fs_info->super_copy or fs_info->super_for_commit allocated failed in
btrfs_get_tree_subvol(), then no need to call btrfs_free_fs_info().
Otherwise btrfs_check_leaked_roots() would access NULL pointer because
fs_info->allocated_roots had not been initialised.
Fix incorrect use of PTR_ERR_OR_ZERO() in topology_parse_cpu_capacity()
which causes the code to proceed with NULL clock pointers. The current
logic uses !PTR_ERR_OR_ZERO(cpu_clk) which evaluates to true for both
valid pointers and NULL, leading to potential NULL pointer dereference
in clk_get_rate().
Per include/linux/err.h documentation, PTR_ERR_OR_ZERO(ptr) returns:
"The error code within @ptr if it is an error pointer; 0 otherwise."
This means PTR_ERR_OR_ZERO() returns 0 for both valid pointers AND NULL
pointers. Therefore !PTR_ERR_OR_ZERO(cpu_clk) evaluates to true (proceed)
when cpu_clk is either valid or NULL, causing clk_get_rate(NULL) to be
called when of_clk_get() returns NULL.
Replace with !IS_ERR_OR_NULL(cpu_clk) which only proceeds for valid
pointers, preventing potential NULL pointer dereference in clk_get_rate().
Commit 370645f41e6e ("dma-mapping: force bouncing if the kmalloc() size is
not cache-line-aligned") introduced DMA_BOUNCE_UNALIGNED_KMALLOC feature
and permitted architecture specific code configure kmalloc slabs with
sizes smaller than the value of dma_get_cache_alignment().
When that feature is enabled, the physical address of some small
kmalloc()-ed buffers might be not aligned to the CPU cachelines, thus not
really suitable for typical DMA. To properly handle that case a SWIOTLB
buffer bouncing is used, so no CPU cache corruption occurs. When that
happens, there is no point reporting a false-positive DMA-API warning that
the buffer is not properly aligned, as this is not a client driver fault.
[m.szyprowski@samsung.com: replace is_swiotlb_allocated() with is_swiotlb_active(), per Catalin] Link: https://lkml.kernel.org/r/20251010173009.3916215-1-m.szyprowski@samsung.com Link: https://lkml.kernel.org/r/20251009141508.2342138-1-m.szyprowski@samsung.com Fixes: 370645f41e6e ("dma-mapping: force bouncing if the kmalloc() size is not cache-line-aligned") Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com> Reviewed-by: Catalin Marinas <catalin.marinas@arm.com> Cc: Christoph Hellwig <hch@lst.de> Cc: Inki Dae <m.szyprowski@samsung.com> Cc: Robin Murohy <robin.murphy@arm.com> Cc: "Isaac J. Manjarres" <isaacmanjarres@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
If the send_peer_notif counter and the peer event notify are not synchronized.
It may cause problems such as the loss or dup of peer notify event.
Before this patch:
- If should_notify_peers is true and the lock for send_peer_notif-- fails, peer
event may be sent again in next mii_monitor loop, because should_notify_peers
is still true.
- If should_notify_peers is true and the lock for send_peer_notif-- succeeded,
but the lock for peer event fails, the peer event will be lost.
This patch locks the RTNL for send_peer_notif, events, and commit simultaneously.
Fixes: 07a4ddec3ce9 ("bonding: add an option to specify a delay between peer notifications") Cc: Jay Vosburgh <jv@jvosburgh.net> Cc: Andrew Lunn <andrew+netdev@lunn.ch> Cc: Eric Dumazet <edumazet@google.com> Cc: Jakub Kicinski <kuba@kernel.org> Cc: Paolo Abeni <pabeni@redhat.com> Cc: Hangbin Liu <liuhangbin@gmail.com> Cc: Nikolay Aleksandrov <razor@blackwall.org> Cc: Vincent Bernat <vincent@bernat.ch> Cc: <stable@vger.kernel.org> Signed-off-by: Tonghao Zhang <tonghao@bamaicloud.com> Acked-by: Jay Vosburgh <jv@jvosburgh.net> Link: https://patch.msgid.link/20251021050933.46412-1-tonghao@bamaicloud.com Signed-off-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixes: c45beebfde34 ("ovl: support encoding fid from inode with no alias") Signed-off-by: Jakub Acs <acsjakub@amazon.de> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Christian Brauner <brauner@kernel.org> Cc: linux-unionfs@vger.kernel.org Cc: linux-fsdevel@vger.kernel.org Cc: linux-kernel@vger.kernel.org Cc: stable@vger.kernel.org Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When we do mlx5e_detach_netdev() we eventually disable blocking events
notifier, among those events are IPsec MPV events from IB to core.
So before disabling those blocking events, make sure to also unregister
the devcom device and mark all this device operations as complete,
in order to prevent the other device from using invalid netdev
during future devcom events which could cause the trace below.
Currently, during the LAN8814 PTP probe shared->phydev is only set if PTP
clock gets actually set, otherwise the function will return before setting
it.
This is an issue as shared->phydev is unconditionally being used when IRQ
is being handled, especially in lan8814_gpio_process_cap and since it was
not set it will cause a NULL pointer exception and crash the kernel.
So, simply always set shared->phydev to avoid the NULL pointer exception.
Fixes: b3f1a08fcf0d ("net: phy: micrel: Add support for PTP_PF_EXTTS for lan8814") Signed-off-by: Robert Marko <robert.marko@sartura.hr> Tested-by: Horatiu Vultur <horatiu.vultur@microchip.com> Link: https://patch.msgid.link/20251021132034.983936-1-robert.marko@sartura.hr Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
chunk->skb pointer is dereferenced in the if-block where it's supposed
to be NULL only.
chunk->skb can only be NULL if chunk->head_skb is not. Check for frag_list
instead and do it just before replacing chunk->skb. We're sure that
otherwise chunk->skb is non-NULL because of outer if() condition.
In ptp_ocp_sma_fb_init(), the code mistakenly used bp->sma[1]
instead of bp->sma[i] inside a for-loop, which caused only SMA[1]
to have its DIRECTION_CAN_CHANGE capability cleared. This led to
inconsistent capability flags across SMA pins.
Current pte_mkwrite_novma() makes PTE dirty unconditionally. This may
mark some pages that are never written dirty wrongly. For example,
do_swap_page() may map the exclusive pages with writable and clean PTEs
if the VMA is writable and the page fault is for read access.
However, current pte_mkwrite_novma() implementation always dirties the
PTE. This may cause unnecessary disk writing if the pages are
never written before being reclaimed.
So, change pte_mkwrite_novma() to clear the PTE_RDONLY bit only if the
PTE_DIRTY bit is set to make it possible to make the PTE writable and
clean.
The current behavior was introduced in commit 73e86cb03cf2 ("arm64:
Move PTE_RDONLY bit handling out of set_pte_at()"). Before that,
pte_mkwrite() only sets the PTE_WRITE bit, while set_pte_at() only
clears the PTE_RDONLY bit if both the PTE_WRITE and the PTE_DIRTY bits
are set.
To test the performance impact of the patch, on an arm64 server
machine, run 16 redis-server processes on socket 1 and 16
memtier_benchmark processes on socket 0 with mostly get
transactions (that is, redis-server will mostly read memory only).
The memory footprint of redis-server is larger than the available
memory, so swap out/in will be triggered. Test results show that the
patch can avoid most swapping out because the pages are mostly clean.
And the benchmark throughput improves ~23.9% in the test.
Fixes: 73e86cb03cf2 ("arm64: Move PTE_RDONLY bit handling out of set_pte_at()") Signed-off-by: Huang Ying <ying.huang@linux.alibaba.com> Cc: Will Deacon <will@kernel.org> Cc: Anshuman Khandual <anshuman.khandual@arm.com> Cc: Ryan Roberts <ryan.roberts@arm.com> Cc: Gavin Shan <gshan@redhat.com> Cc: Ard Biesheuvel <ardb@kernel.org> Cc: Matthew Wilcox (Oracle) <willy@infradead.org> Cc: Yicong Yang <yangyicong@hisilicon.com> Cc: linux-arm-kernel@lists.infradead.org Cc: linux-kernel@vger.kernel.org Reviewed-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Resolve race conditions in timestamp events list handling between TX
and RX paths causing missed timestamps.
The current implementation uses a single events list for both TX and RX
timestamps. The am65_cpts_find_ts() function acquires the lock,
splices all events (TX as well as RX events) to a temporary list,
and releases the lock. This function performs matching of timestamps
for TX packets only. Before it acquires the lock again to put the
non-TX events back to the main events list, a concurrent RX
processing thread could acquire the lock (as observed in practice),
find an empty events list, and fail to attach timestamp to it,
even though a relevant event exists in the spliced list which is yet to
be restored to the main list.
Fix this by creating separate events lists to handle TX and RX
timestamps independently.
With CONFIG_DEBUG_LOCK_ALLOC=y, the smc->clcsock is unexpectedly changed
in inet_init_csk_locks(). The INET_PROTOSW_ICSK flag is no need by smc,
just remove it.
After removing the INET_PROTOSW_ICSK flag, this patch alse revert
commit 6fd27ea183c2 ("net/smc: fix lacks of icsk_syn_mss with IPPROTO_SMC")
to avoid casting smc_sock to inet_connection_sock.
Reported-by: syzbot+f775be4458668f7d220e@syzkaller.appspotmail.com Closes: https://syzkaller.appspot.com/bug?extid=f775be4458668f7d220e Tested-by: syzbot+f775be4458668f7d220e@syzkaller.appspotmail.com Fixes: d25a92ccae6b ("net/smc: Introduce IPPROTO_SMC") Signed-off-by: Wang Liang <wangliang74@huawei.com> Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Reviewed-by: D. Wythe <alibuda@linux.alibaba.com> Link: https://patch.msgid.link/20251017024827.3137512-1-wangliang74@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
XDP programs can change the layout of an xdp_buff through
bpf_xdp_adjust_tail() and bpf_xdp_adjust_head(). Therefore, the driver
cannot assume the size of the linear data area nor fragments. Fix the
bug in mlx5 by generating skb according to xdp_buff after XDP programs
run.
Currently, when handling multi-buf XDP, the mlx5 driver assumes the
layout of an xdp_buff to be unchanged. That is, the linear data area
continues to be empty and fragments remain the same. This may cause
the driver to generate erroneous skb or triggering a kernel
warning. When an XDP program added linear data through
bpf_xdp_adjust_head(), the linear data will be ignored as
mlx5e_build_linear_skb() builds an skb without linear data and then
pull data from fragments to fill the linear data area. When an XDP
program has shrunk the non-linear data through bpf_xdp_adjust_tail(),
the delta passed to __pskb_pull_tail() may exceed the actual nonlinear
data size and trigger the BUG_ON in it.
To fix the issue, first record the original number of fragments. If the
number of fragments changes after the XDP program runs, rewind the end
fragment pointer by the difference and recalculate the truesize. Then,
build the skb with the linear data area matching the xdp_buff. Finally,
only pull data in if there is non-linear data and fill the linear part
up to 256 bytes.
Fixes: f52ac7028bec ("net/mlx5e: RX, Add XDP multi-buffer support in Striding RQ") Signed-off-by: Amery Hung <ameryhung@gmail.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Link: https://patch.msgid.link/1760644540-899148-3-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
XDP programs can release xdp_buff fragments when calling
bpf_xdp_adjust_tail(). The driver currently assumes the number of
fragments to be unchanged and may generate skb with wrong truesize or
containing invalid frags. Fix the bug by generating skb according to
xdp_buff after the XDP program runs.
Fixes: ea5d49bdae8b ("net/mlx5e: Add XDP multi buffer support to the non-linear legacy RQ") Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com> Signed-off-by: Amery Hung <ameryhung@gmail.com> Signed-off-by: Tariq Toukan <tariqt@nvidia.com> Link: https://patch.msgid.link/1760644540-899148-2-git-send-email-tariqt@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
CONFIG_INIT_STACK_ALL_ZERO introduces a performance cost by
zero-initializing all stack variables on function entry. The mlx5 XDP
RX path previously allocated a struct mlx5e_xdp_buff on the stack per
received CQE, resulting in measurable performance degradation under
this config.
This patch reuses a mlx5e_xdp_buff stored in the mlx5e_rq struct,
avoiding per-CQE stack allocations and repeated zeroing.
With this change, XDP_DROP and XDP_TX performance matches that of
kernels built without CONFIG_INIT_STACK_ALL_ZERO.
Performance was measured on a ConnectX-6Dx using a single RX channel
(1 CPU at 100% usage) at ~50 Mpps. The baseline results were taken from
net-next-6.15.
TEST 12: bind vrf-2 & 1 in server, connect from client 1 & 2, N [FAIL]
not ok 1 selftests: net: sctp_vrf.sh # exit=3
The failure happens when the server bind in a new run conflicts with an
existing association from the previous run:
[1] ip netns exec $SERVER_NS ./sctp_hello server ...
[2] ip netns exec $CLIENT_NS ./sctp_hello client ...
[3] ip netns exec $SERVER_NS pkill sctp_hello ...
[4] ip netns exec $SERVER_NS ./sctp_hello server ...
It occurs if the client in [2] sends a message and closes immediately.
With the message unacked, no SHUTDOWN is sent. Killing the server in [3]
triggers a SHUTDOWN the client also ignores due to the unacked message,
leaving the old association alive. This causes the bind at [4] to fail
until the message is acked and the client responds to a second SHUTDOWN
after the server’s T2 timer expires (3s).
This patch fixes the issue by preventing the client from sending data.
Instead, the client blocks on recv() and waits for the server to close.
It also waits until both the server and the client sockets are fully
released in stop_server and wait_client before restarting.
Additionally, replace 2>&1 >/dev/null with -q in sysctl and grep, and
drop other redundant 2>&1 >/dev/null redirections, and fix a typo from
N to Y (connect successfully) in the description of the last test.
Fixes: a61bd7b9fef3 ("selftests: add a selftest for sctp vrf") Reported-by: Hangbin Liu <liuhangbin@gmail.com> Tested-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Xin Long <lucien.xin@gmail.com> Link: https://patch.msgid.link/be2dacf52d0917c4ba5e2e8c5a9cb640740ad2b6.1760731574.git.lucien.xin@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In addition to can_dropped_invalid_skb(), the helper function
can_dev_dropped_skb() checks whether the device is in listen-only mode and
discards the skb accordingly.
Replace can_dropped_invalid_skb() by can_dev_dropped_skb() to also drop
skbs in for listen-only mode.
Reported-by: Marc Kleine-Budde <mkl@pengutronix.de> Closes: https://lore.kernel.org/all/20251017-bizarre-enchanted-quokka-f3c704-mkl@pengutronix.de/ Fixes: ff60bfbaf67f ("can: rockchip_canfd: add driver for Rockchip CAN-FD controller") Link: https://patch.msgid.link/20251017-fix-skb-drop-check-v1-3-556665793fa4@pengutronix.de Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
In addition to can_dropped_invalid_skb(), the helper function
can_dev_dropped_skb() checks whether the device is in listen-only mode and
discards the skb accordingly.
Replace can_dropped_invalid_skb() by can_dev_dropped_skb() to also drop
skbs in for listen-only mode.
Reported-by: Marc Kleine-Budde <mkl@pengutronix.de> Closes: https://lore.kernel.org/all/20251017-bizarre-enchanted-quokka-f3c704-mkl@pengutronix.de/ Fixes: 9721866f07e1 ("can: esd: add support for esd GmbH PCIe/402 CAN interface family") Link: https://patch.msgid.link/20251017-fix-skb-drop-check-v1-2-556665793fa4@pengutronix.de Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
In addition to can_dropped_invalid_skb(), the helper function
can_dev_dropped_skb() checks whether the device is in listen-only mode and
discards the skb accordingly.
Replace can_dropped_invalid_skb() by can_dev_dropped_skb() to also drop
skbs in for listen-only mode.
Reported-by: Marc Kleine-Budde <mkl@pengutronix.de> Closes: https://lore.kernel.org/all/20251017-bizarre-enchanted-quokka-f3c704-mkl@pengutronix.de/ Fixes: f00647d8127b ("can: bxcan: add support for ST bxCAN controller") Link: https://patch.msgid.link/20251017-fix-skb-drop-check-v1-1-556665793fa4@pengutronix.de Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
The blamed commit increased the needed headroom to account for
alignment. This means that the size required to always align a Tx buffer
was added inside the dpaa2_eth_needed_headroom() function. By doing
that, a manual adjustment of the pointer passed to PTR_ALIGN() was no
longer correct since the 'buffer_start' variable was already pointing
to the start of the skb's memory.
The behavior of the dpaa2-eth driver without this patch was to drop
frames on Tx even when the headroom was matching the 128 bytes
necessary. Fix this by removing the manual adjust of 'buffer_start' from
the PTR_MODE call.
Closes: https://lore.kernel.org/netdev/70f0dcd9-1906-4d13-82df-7bbbbe7194c6@app.fastmail.com/T/#u Fixes: f422abe3f23d ("dpaa2-eth: increase the needed headroom to account for alignment") Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com> Tested-by: Mathew McBride <matt@traverse.com.au> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20251016135807.360978-1-ioana.ciornei@nxp.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The ENETC RX ring uses the page halves flipping mechanism, each page is
split into two halves for the RX ring to use. And ENETC_RXB_TRUESIZE is
defined to 2048 to indicate the size of half a page. However, the page
size is configurable, for ARM64 platform, PAGE_SIZE is default to 4K,
but it could be configured to 16K or 64K.
When PAGE_SIZE is set to 16K or 64K, ENETC_RXB_TRUESIZE is not correct,
and the RX ring will always use the first half of the page. This is not
consistent with the description in the relevant kernel doc and commit
messages.
This issue is invisible in most cases, but if users want to increase
PAGE_SIZE to receive a Jumbo frame with a single buffer for some use
cases, it will not work as expected, because the buffer size of each
RX BD is fixed to 2048 bytes.
Based on the above two points, we expect to correct ENETC_RXB_TRUESIZE
to (PAGE_SIZE >> 1), as described in the comment.
After applying the workaround for err050089, the LS1028A platform
experiences RCU stalls on RT kernel. This issue is caused by the
recursive acquisition of the read lock enetc_mdio_lock. Here list some
of the call stacks identified under the enetc_poll path that may lead to
a deadlock:
After enetc_poll acquires the read lock, a higher-priority writer attempts
to acquire the lock, causing preemption. The writer detects that a
read lock is already held and is scheduled out. However, readers under
enetc_poll cannot acquire the read lock again because a writer is already
waiting, leading to a thread hang.
Currently, the deadlock is avoided by adjusting enetc_lock_mdio to prevent
recursive lock acquisition.
Fixes: 6d36ecdbc441 ("net: enetc: take the MDIO lock only once per NAPI poll cycle") Signed-off-by: Jianpeng Chang <jianpeng.chang.cn@windriver.com> Acked-by: Wei Fang <wei.fang@nxp.com> Link: https://patch.msgid.link/20251015021427.180757-1-jianpeng.chang.cn@windriver.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Creating FDB entries is possible from a non-initial user namespace when
having CAP_NET_ADMIN, yet, when deleting FDB entries, processes receive
an EPERM because the capability is always checked against the initial
user namespace. This restricts the FDB management from unprivileged
containers.
Drop the netlink_capable check in rtnl_fdb_del as it was originally
dropped in c5c351088ae7 and reintroduced in 1690be63a27b without
intention.
This patch was tested using a container on GyroidOS, where it was
possible to delete FDB entries from an unprivileged user namespace and
private network namespace.
Fixes: 1690be63a27b ("bridge: Add vlan support to static neighbors") Reviewed-by: Michael Weiß <michael.weiss@aisec.fraunhofer.de> Tested-by: Harshal Gohel <hg@simonwunderlich.de> Signed-off-by: Johannes Wiesböck <johannes.wiesboeck@aisec.fraunhofer.de> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Nikolay Aleksandrov <razor@blackwall.org> Link: https://patch.msgid.link/20251015201548.319871-1-johannes.wiesboeck@aisec.fraunhofer.de Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When building with Clang 20 or newer, there are some objtool warnings
from unexpected fallthroughs to other functions:
vmlinux.o: warning: objtool: mlx5e_mpwrq_mtts_per_wqe() falls through to next function mlx5e_mpwrq_max_num_entries()
vmlinux.o: warning: objtool: mlx5e_mpwrq_max_log_rq_size() falls through to next function mlx5e_get_linear_rq_headroom()
LLVM 20 contains an (admittedly problematic [1]) optimization [2] to
convert divide by zero into the equivalent of __builtin_unreachable(),
which invokes undefined behavior and destroys code generation when it is
encountered in a control flow graph.
mlx5e_mpwrq_umr_entry_size() returns 0 in the default case of an
unrecognized mlx5e_mpwrq_umr_mode value. mlx5e_mpwrq_mtts_per_wqe(),
which is inlined into mlx5e_mpwrq_max_log_rq_size(), uses the result of
mlx5e_mpwrq_umr_entry_size() in a divide operation without checking for
zero, so LLVM is able to infer there will be a divide by zero in this
case and invokes undefined behavior. While there is some proposed work
to isolate this undefined behavior and avoid the destructive code
generation that results in these objtool warnings, code should still be
defensive against divide by zero.
As the WARN_ONCE() implies that an invalid value should be handled
gracefully, return 1 instead of 0 in the default case so that the
results of this division operation is always valid.
Commit e3f1164fc9ee ("PM: EM: Support late CPUs booting and capacity
adjustment") added a mechanism to handle CPUs that come up late by
retrying when any of the `cpufreq_cpu_get()` call fails.
However, if there are holes in the CPU topology (offline CPUs, e.g.
nosmt), the first missing CPU causes the loop to break, preventing
subsequent online CPUs from being updated.
Instead of aborting on the first missing CPU policy, loop through all
and retry if any were missing.
Fixes: e3f1164fc9ee ("PM: EM: Support late CPUs booting and capacity adjustment") Suggested-by: Kenneth Crudup <kenneth.crudup@gmail.com> Reported-by: Kenneth Crudup <kenneth.crudup@gmail.com> Link: https://lore.kernel.org/linux-pm/40212796-734c-4140-8a85-854f72b8144d@panix.com/ Cc: 6.9+ <stable@vger.kernel.org> # 6.9+ Signed-off-by: Christian Loehle <christian.loehle@arm.com> Link: https://patch.msgid.link/20250831214357.2020076-1-christian.loehle@arm.com
[ rjw: Drop the new pr_debug() message which is not very useful ] Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move the check of the CPU capacity currently stored in the energy model
against the arch_scale_cpu_capacity() value to em_adjust_new_capacity()
so it will be done regardless of where the latter is called from.
This will be useful when a new em_adjust_new_capacity() caller is added
subsequently.
While at it, move the pd local variable declaration in
em_check_capacity_update() into the loop in which it is used.
No intentional functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Reviewed-by: Lukasz Luba <lukasz.luba@arm.com> Tested-by: Christian Loehle <christian.loehle@arm.com> Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com> Link: https://patch.msgid.link/7810787.EvYhyI6sBW@rjwysocki.net
Stable-dep-of: 1ebe8f7e7825 ("PM: EM: Fix late boot with holes in CPU topology") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Every iteration of the loop over all possible CPUs in
em_check_capacity_update() causes get_cpu_device() to be called twice
for the same CPU, once indirectly via em_cpu_get() and once directly.
Get rid of the indirect get_cpu_device() call by moving the direct
invocation of it earlier and using em_pd_get() instead of em_cpu_get()
to get a pd pointer for the dev one returned by it.
This also exposes the fact that dev is needed to get a pd, so the code
becomes somewhat easier to follow after it.
No functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Reviewed-by: Lukasz Luba <lukasz.luba@arm.com> Link: https://patch.msgid.link/1925950.tdWV9SEqCh@rjwysocki.net
Stable-dep-of: 1ebe8f7e7825 ("PM: EM: Fix late boot with holes in CPU topology") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The max_cap parameter is never used in em_adjust_new_capacity(), so
drop it.
No functional impact.
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Reviewed-by: Lukasz Luba <lukasz.luba@arm.com> Link: https://patch.msgid.link/2369979.ElGaqSPkdT@rjwysocki.net
Stable-dep-of: 1ebe8f7e7825 ("PM: EM: Fix late boot with holes in CPU topology") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This pattern isn't very documented, and apparently not used much outside
of 'make tools/help', but it has existed for over a decade (since commit ea01fa9f63ae: "tools: Connect to the kernel build system").
However, it doesn't work very well for most cases, particularly the
useful "tools/all" target, because it overrides the LDFLAGS value with
an empty one.
And once overridden, 'make' will then not honor the tooling makefiles
trying to change it - which then makes any LDFLAGS use in the tooling
directory break, typically causing odd link errors.
Remove that LDFLAGS override, since it seems to be entirely historical.
The core kernel makefiles no longer modify LDFLAGS as part of the build,
and use kernel-specific link flags instead (eg 'KBUILD_LDFLAGS' and
friends).
This allows more of the 'make tools/*' cases to work. I say 'more',
because some of the tooling build rules make various other assumptions
or have other issues, so it's still a bit hit-or-miss. But those issues
tend to show up with the 'make -C tools xyz' pattern too, so now it's no
longer an issue of this particular 'tools/*' build rule being special.
If we want to invalidate a remote key we should do that as soon as
possible, so do it in the first send work request.
Acked-by: Namjae Jeon <linkinjeon@kernel.org> Cc: Steve French <smfrench@gmail.com> Cc: Tom Talpey <tom@talpey.com> Cc: linux-cifs@vger.kernel.org Cc: samba-technical@lists.samba.org Signed-off-by: Stefan Metzmacher <metze@samba.org> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Event ID is only using the attr::config bit [7, 0] but we check the
event range using the whole 64bit field. It blocks the usage of the
rest field of attr::config. Relax the check by only using the
bit [7, 0].
Acked-by: Jonathan Cameron <jonathan.cameron@huawei.com> Signed-off-by: Yicong Yang <yangyicong@hisilicon.com> Signed-off-by: Yushan Wang <wangyushan12@huawei.com> Signed-off-by: Will Deacon <will@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
PAGE_KERNEL_TEXT is an old macro that is used to tell kernel whether
kernel text has to be mapped read-only or read-write based on build
time options.
But nowadays, with functionnalities like jump_labels, static links,
etc ... more only less all kernels need to be read-write at some
point, and some combinations of configs failed to work due to
innacurate setting of PAGE_KERNEL_TEXT. On the other hand, today
we have CONFIG_STRICT_KERNEL_RWX which implements a more controlled
access to kernel modifications.
Instead of trying to keep PAGE_KERNEL_TEXT accurate with all
possible options that may imply kernel text modification, always
set kernel text read-write at startup and rely on
CONFIG_STRICT_KERNEL_RWX to provide accurate protection.
Do this by passing PAGE_KERNEL_X to map_kernel_page() in
__maping_ram_chunk() instead of passing PAGE_KERNEL_TEXT. Once
this is done, the only remaining user of PAGE_KERNEL_TEXT is
mmu_mark_initmem_nx() which uses it in a call to setibat().
As setibat() ignores the RW/RO, we can seamlessly replace
PAGE_KERNEL_TEXT by PAGE_KERNEL_X here as well and get rid of
PAGE_KERNEL_TEXT completely.
In gdlm_put_lock(), there is a small window of time in which the
DFL_UNMOUNT flag has been set but the lockspace hasn't been released,
yet. In that window, dlm may still call gdlm_ast() and gdlm_bast().
To prevent it from dereferencing freed glock objects, only free the
glock if the lockspace has actually been released.
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com> Reviewed-by: Andrew Price <anprice@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The `ID_AA64MMFR4_EL1.EIESB` field, is an unsigned enumeration, but was
incorrectly defined as a `SignedEnum` when introduced in commit cfc680bb04c5 ("arm64: sysreg: Add layout for ID_AA64MMFR4_EL1"). This is
corrected to `UnsignedEnum`.
Conversely, the `ID_AA64DFR0_EL1.DoubleLock` field, is a signed
enumeration, but was incorrectly defined as an `UnsignedEnum`. This is
corrected to `SignedEnum`, which wasn't correctly set when annotated as
such in commit ad16d4cf0b4f ("arm64/sysreg: Initial unsigned annotations
for ID registers").
Signed-off-by: Fuad Tabba <tabba@google.com> Acked-by: Mark Rutland <mark.rutland@arm.com> Signed-off-by: Will Deacon <will@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
In preparation for the future commit ("bitops: Add __attribute_const__ to generic
ffs()-family implementations"), which allows GCC's value range tracker
to see past ffs(), GCC 8 on ARM thinks that it might be possible that
"ffs(rq) - 8" used here:
v = FIELD_PREP(PCI_EXP_DEVCTL_READRQ, ffs(rq) - 8);
could wrap below 0, leading to a very large value, which would be out of
range for the FIELD_PREP() usage:
drivers/pci/pci.c: In function 'pcie_set_readrq':
include/linux/compiler_types.h:572:38: error: call to '__compiletime_assert_471' declared with attribute error: FIELD_PREP: value too large for the field
...
drivers/pci/pci.c:5896:6: note: in expansion of macro 'FIELD_PREP'
v = FIELD_PREP(PCI_EXP_DEVCTL_READRQ, ffs(rq) - 8);
^~~~~~~~~~
If the result of the ffs() is bounds checked before being used in
FIELD_PREP(), the value tracker seems happy again. :)
CPU: 1 UID: 0 PID: 12609 Comm: syz.1.2692 Not tainted 6.16.0-syzkaller #0 PREEMPT(none)
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/12/2025
=====================================================
The HFS_SB(sb)->bitmap buffer is allocated in hfs_mdb_get():
HFS_SB(sb)->bitmap = kmalloc(8192, GFP_KERNEL);
Finally, it can trigger the reported issue because kmalloc()
doesn't clear the allocated memory. If allocated memory contains
only zeros, then everything will work pretty fine.
But if the allocated memory contains the "garbage", then
it can affect the bitmap operations and it triggers
the reported issue.
This patch simply exchanges the kmalloc() on kzalloc()
with the goal to guarantee the correctness of bitmap operations.
Because, newly created allocation bitmap should have all
available blocks free. Potentially, initialization bitmap's read
operation could not fill the whole allocated memory and
"garbage" in the not initialized memory will be the reason of
volume coruptions and file system driver bugs.
Reported-by: syzbot <syzbot+773fa9d79b29bd8b6831@syzkaller.appspotmail.com> Closes: https://syzkaller.appspot.com/bug?extid=773fa9d79b29bd8b6831 Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com>
cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
cc: Yangtao Li <frank.li@vivo.com>
cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20250820230636.179085-1-slava@dubeyko.com Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
The main reason of the issue that struct hfsplus_inode_info
has not been properly initialized for the case of root folder.
In the case of root folder, hfsplus_fill_super() calls
the hfsplus_iget() that implements only partial initialization of
struct hfsplus_inode_info and subfolders field is not
initialized by hfsplus_iget() logic.
This patch implements complete initialization of
struct hfsplus_inode_info in the hfsplus_iget() logic with
the goal to prevent likewise issues for the case of
root folder.
Reported-by: syzbot <syzbot+fdedff847a0e5e84c39f@syzkaller.appspotmail.com> Closes: https://syzkaller.appspot.com/bug?extid=fdedff847a0e5e84c39f Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com>
cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
cc: Yangtao Li <frank.li@vivo.com>
cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20250825225103.326401-1-slava@dubeyko.com Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When calling hfsplus_bmap_alloc to allocate a free node, this function
first retrieves the bitmap from header node and map node using node->page
together with the offset and length from hfs_brec_lenoff
```
len = hfs_brec_lenoff(node, 2, &off16);
off = off16;
off += node->page_offset;
pagep = node->page + (off >> PAGE_SHIFT);
data = kmap_local_page(*pagep);
```
However, if the retrieved offset or length is invalid(i.e. exceeds
node_size), the code may end up accessing pages outside the allocated
range for this node.
This patch adds proper validation of both offset and length before use,
preventing out-of-bounds page access. Move is_bnode_offset_valid and
check_and_correct_requested_length to hfsplus_fs.h, as they may be
required by other functions.
And if inode->i_ino could be equal to zero or any non-available CNID,
then hfs_brec_find() could not find the record in the tree. As a result,
fd->key could be compared with fd->search_key. But hfsplus_find_init()
uses kmalloc() for fd->key and fd->search_key allocation:
Finally, fd->key is still not initialized if hfs_brec_find()
has found nothing.
This patch changes kmalloc() on kzalloc() in hfs_find_init()
and intializes fd->record, fd->keyoffset, fd->keylength,
fd->entryoffset, fd->entrylength for the case if hfs_brec_find()
has been found nothing in the b-tree node.
Reported-by: syzbot <syzbot+55ad87f38795d6787521@syzkaller.appspotmail.com> Closes: https://syzkaller.appspot.com/bug?extid=55ad87f38795d6787521 Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com>
cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
cc: Yangtao Li <frank.li@vivo.com>
cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20250818225232.126402-1-slava@dubeyko.com Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Potenatially, __hfs_ext_read_extent() could operate by
not initialized values of fd->key after hfs_brec_find() call:
static inline int __hfs_ext_read_extent(struct hfs_find_data *fd, struct hfs_extent *extent,
u32 cnid, u32 block, u8 type)
{
int res;
hfs_ext_build_key(fd->search_key, cnid, block, type);
fd->key->ext.FNum = 0;
res = hfs_brec_find(fd);
if (res && res != -ENOENT)
return res;
if (fd->key->ext.FNum != fd->search_key->ext.FNum ||
fd->key->ext.FkType != fd->search_key->ext.FkType)
return -ENOENT;
if (fd->entrylength != sizeof(hfs_extent_rec))
return -EIO;
hfs_bnode_read(fd->bnode, extent, fd->entryoffset, sizeof(hfs_extent_rec));
return 0;
}
This patch changes kmalloc() on kzalloc() in hfs_find_init()
and intializes fd->record, fd->keyoffset, fd->keylength,
fd->entryoffset, fd->entrylength for the case if hfs_brec_find()
has been found nothing in the b-tree node.
Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com>
cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
cc: Yangtao Li <frank.li@vivo.com>
cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20250818225252.126427-1-slava@dubeyko.com Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Currently, hfs_brec_remove() executes moving records
towards the location of deleted record and it updates
offsets of moved records. However, the hfs_brec_remove()
logic ignores the "mess" of b-tree node's free space and
it doesn't touch the offsets out of records number.
Potentially, it could confuse fsck or driver logic or
to be a reason of potential corruption cases.
This patch reworks the logic of hfs_brec_remove()
by means of clearing freed space of b-tree node
after the records moving. And it clear the last
offset that keeping old location of free space
because now the offset before this one is keeping
the actual offset to the free space after the record
deletion.
Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com>
cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
cc: Yangtao Li <frank.li@vivo.com>
cc: linux-fsdevel@vger.kernel.org Link: https://lore.kernel.org/r/20250815194918.38165-1-slava@dubeyko.com Signed-off-by: Viacheslav Dubeyko <slava@dubeyko.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
On nios2, with CONFIG_FLATMEM set, the kernel relies on
memblock_get_current_limit() to determine the limits of mem_map, in
particular for max_low_pfn.
Unfortunately, memblock.current_limit is only default initialized to
MEMBLOCK_ALLOC_ANYWHERE at this point of the bootup, potentially leading
to situations where max_low_pfn can erroneously exceed the value of
max_pfn and, thus, the valid range of available DRAM.
This can in turn cause kernel-level paging failures, e.g.:
[ 76.900000] Unable to handle kernel paging request at virtual address 20303000
[ 76.900000] ea = c0080890, ra = c000462c, cause = 14
[ 76.900000] Kernel panic - not syncing: Oops
[ 76.900000] ---[ end Kernel panic - not syncing: Oops ]---
This patch fixes this by pre-calculating memblock.current_limit
based on the upper limits of the available memory ranges via
adjust_lowmem_bounds, a simplified version of the equivalent
implementation within the arm architecture.
Signed-off-by: Simon Schuster <schuster.simon@siemens-energy.com> Signed-off-by: Andreas Oetken <andreas.oetken@siemens-energy.com> Signed-off-by: Dinh Nguyen <dinguyen@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Guenter Roeck [Mon, 20 Oct 2025 17:04:22 +0000 (10:04 -0700)]
dmaengine: Add missing cleanup on module unload
Upstream commit b7cb9a034305 ("dmaengine: idxd: Fix refcount underflow
on module unload") fixes a refcount underflow by replacing the call to
idxd_cleanup() in the remove function with direct cleanup calls. That works
fine upstream. However, upstream removed support for IOMMU_DEV_FEAT_IOPF,
which is still supported in v6.12.y. The backport of commit b7cb9a034305
into v6.12.y misses the call to disable it. This results in a warning
backtrace when unloading and reloading the module.
Enable the workaround for Neoverse-V3AE, and document this.
Signed-off-by: Mark Rutland <mark.rutland@arm.com> Cc: James Morse <james.morse@arm.com> Cc: Will Deacon <will@kernel.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Ryan Roberts <ryan.roberts@arm.com> Signed-off-by: Will Deacon <will@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
... in section A.6.1 ("MIDR_EL1, Main ID Register").
Signed-off-by: Mark Rutland <mark.rutland@arm.com> Cc: James Morse <james.morse@arm.com> Cc: Will Deacon <will@kernel.org> Cc: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Ryan Roberts <ryan.roberts@arm.com> Signed-off-by: Will Deacon <will@kernel.org> Signed-off-by: Ryan Roberts <ryan.roberts@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Kernel panics because it detects UFFD inconsistency during
userfaultfd_release_all(). Specifically, a VMA which has a valid pointer
to vma->vm_userfaultfd_ctx, but no UFFD flags in vma->vm_flags.
The inconsistency is caused in ksm_madvise(): when user calls madvise()
with MADV_UNMEARGEABLE on a VMA that is registered for UFFD in MINOR mode,
it accidentally clears all flags stored in the upper 32 bits of
vma->vm_flags.
Assuming x86_64 kernel build, unsigned long is 64-bit and unsigned int and
int are 32-bit wide. This setup causes the following mishap during the &=
~VM_MERGEABLE assignment.
VM_MERGEABLE is a 32-bit constant of type unsigned int, 0x8000'0000.
After ~ is applied, it becomes 0x7fff'ffff unsigned int, which is then
promoted to unsigned long before the & operation. This promotion fills
upper 32 bits with leading 0s, as we're doing unsigned conversion (and
even for a signed conversion, this wouldn't help as the leading bit is 0).
& operation thus ends up AND-ing vm_flags with 0x0000'0000'7fff'ffff
instead of intended 0xffff'ffff'7fff'ffff and hence accidentally clears
the upper 32-bits of its value.
Fix it by changing `VM_MERGEABLE` constant to unsigned long, using the
BIT() macro.
Note: other VM_* flags are not affected: This only happens to the
VM_MERGEABLE flag, as the other VM_* flags are all constants of type int
and after ~ operation, they end up with leading 1 and are thus converted
to unsigned long with leading 1s.
Note 2:
After commit 31defc3b01d9 ("userfaultfd: remove (VM_)BUG_ON()s"), this is
no longer a kernel BUG, but a WARNING at the same place:
[ 45.595973] WARNING: CPU: 1 PID: 2474 at mm/userfaultfd.c:2067
but the root-cause (flag-drop) remains the same.
[akpm@linux-foundation.org: rust bindgen wasn't able to handle BIT(), from Miguel] Link: https://lore.kernel.org/oe-kbuild-all/202510030449.VfSaAjvd-lkp@intel.com/ Link: https://lkml.kernel.org/r/20251001090353.57523-2-acsjakub@amazon.de Fixes: 7677f7fd8be7 ("userfaultfd: add minor fault registration mode") Signed-off-by: Jakub Acs <acsjakub@amazon.de> Signed-off-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com> Acked-by: David Hildenbrand <david@redhat.com> Acked-by: SeongJae Park <sj@kernel.org> Tested-by: Alice Ryhl <aliceryhl@google.com> Tested-by: Miguel Ojeda <miguel.ojeda.sandonis@gmail.com> Cc: Xu Xin <xu.xin16@zte.com.cn> Cc: Chengming Zhou <chengming.zhou@linux.dev> Cc: Peter Xu <peterx@redhat.com> Cc: Axel Rasmussen <axelrasmussen@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[acsjakub@amazon.de: adjust context in bindgings_helper.h] Signed-off-by: Jakub Acs <acsjakub@amazon.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Do read-modify-write so that we re-use the characterized reset value as
specified in TRM [1] to program calibration wait time which defines number
of cycles to wait for after startup state machine is in bandgap enable
state.
This fixes PLL lock timeout error faced while using RPi DSI Panel on TI's
AM62L and J721E SoC since earlier calibration wait time was getting
overwritten to zero value thus failing the PLL to lockup and causing
timeout.
When a first MPTCP connection gets successfully established after a
blackhole period, 'active_disable_times' was supposed to be reset when
this connection was done via any non-loopback interfaces.
Unfortunately, the opposite condition was checked: only reset when the
connection was established via a loopback interface. Fixing this by
simply looking at the opposite.
This is similar to what is done with TCP FastOpen, see
tcp_fastopen_active_disable_ofo_check().
This patch is a follow-up of a previous discussion linked to commit 893c49a78d9f ("mptcp: Use __sk_dst_get() and dst_dev_rcu() in
mptcp_active_enable()."), see [1].
In ip_output() skb->dev is updated from the skb_dst(skb)->dev
this can become invalid when the interface is unregistered and freed,
Introduced new skb_dst_dev_rcu() function to be used instead of
skb_dst_dev() within rcu_locks in ip_output.This will ensure that
all the skb's associated with the dev being deregistered will
be transnmitted out first, before freeing the dev.
Given that ip_output() is called within an rcu_read_lock()
critical section or from a bottom-half context, it is safe to introduce
an RCU read-side critical section within it.
Multiple panic call stacks were observed when UL traffic was run
in concurrency with device deregistration from different functions,
pasting one sample for reference.
Changes in v3:
- Replaced WARN_ON() with WARN_ON_ONCE(), as suggested by Willem de Bruijn.
- Dropped legacy lines mistakenly pulled in from an outdated branch.
Changes in v2:
- Addressed review comments from Eric Dumazet
- Used READ_ONCE() to prevent potential load/store tearing
- Added skb_dst_dev_rcu() and used along with rcu_read_lock() in ip_output
Signed-off-by: Sharath Chandra Vurukala <quic_sharathv@quicinc.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Link: https://patch.msgid.link/20250730105118.GA26100@hu-sharathv-hyd.qualcomm.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 833d4313bc1e ("mptcp: reset blackhole on success with non-loopback ifaces") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There was backward compatibility in the terms of mailbox API. Various
drivers from various OSes supporting 10G adapters from Intel portfolio
could easily negotiate mailbox API.
This convention has been broken since introducing API 1.4.
Commit 0062e7cc955e ("ixgbevf: add VF IPsec offload code") added support
for IPSec which is specific only for the kernel ixgbe driver. None of the
rest of the Intel 10G PF/VF drivers supports it. And actually lack of
support was not included in the IPSec implementation - there were no such
code paths. No possibility to negotiate support for the feature was
introduced along with introduction of the feature itself.
Commit 339f28964147 ("ixgbevf: Add support for new mailbox communication
between PF and VF") increasing API version to 1.5 did the same - it
introduced code supported specifically by the PF ESX driver. It altered API
version for the VF driver in the same time not touching the version
defined for the PF ixgbe driver. It led to additional discrepancies,
as the code provided within API 1.6 cannot be supported for Linux ixgbe
driver as it causes crashes.
The issue was noticed some time ago and mitigated by Jake within the commit d0725312adf5 ("ixgbevf: stop attempting IPSEC offload on Mailbox API 1.5").
As a result we have regression for IPsec support and after increasing API
to version 1.6 ixgbevf driver stopped to support ESX MBX.
To fix this mess add new mailbox op asking PF driver about supported
features. Basing on a response determine whether to set support for IPSec
and ESX-specific enhanced mailbox.
New mailbox op, for compatibility purposes, must be added within new API
revision, as API version of OOT PF & VF drivers is already increased to
1.6 and doesn't incorporate features negotiate op.
Features negotiation mechanism gives possibility to be extended with new
features when needed in the future.
Reported-by: Jacob Keller <jacob.e.keller@intel.com> Closes: https://lore.kernel.org/intel-wired-lan/20241101-jk-ixgbevf-mailbox-v1-5-fixes-v1-0-f556dc9a66ed@intel.com/ Fixes: 0062e7cc955e ("ixgbevf: add VF IPsec offload code") Fixes: 339f28964147 ("ixgbevf: Add support for new mailbox communication between PF and VF") Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Jedrzej Jagielski <jedrzej.jagielski@intel.com> Tested-by: Rafal Romanowski <rafal.romanowski@intel.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-4-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
E610 adapters no longer use the VFLINKS register to read PF's link
speed and linkup state. As a result VF driver cannot get actual link
state and it incorrectly reports 10G which is the default option.
It leads to a situation where even 1G adapters print 10G as actual
link speed. The same happens when PF driver set speed different than 10G.
Add new mailbox operation to let the VF driver request a PF driver
to provide actual link data. Update the mailbox api to v1.6.
Incorporate both ways of getting link status within the legacy
ixgbe_check_mac_link_vf() function.
Fixes: 4c44b450c69b ("ixgbevf: Add support for Intel(R) E610 device") Co-developed-by: Andrzej Wilczynski <andrzejx.wilczynski@intel.com> Signed-off-by: Andrzej Wilczynski <andrzejx.wilczynski@intel.com> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Aleksandr Loktionov <aleksandr.loktionov@intel.com> Cc: stable@vger.kernel.org Signed-off-by: Jedrzej Jagielski <jedrzej.jagielski@intel.com> Tested-by: Rafal Romanowski <rafal.romanowski@intel.com> Signed-off-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20251009-jk-iwl-net-2025-10-01-v3-2-ef32a425b92a@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: a7075f501bd3 ("ixgbevf: fix mailbox API compatibility by negotiating supported features") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add support for Intel(R) E610 Series of network devices. The E610
is based on X550 but adds firmware managed link, enhanced security
capabilities and support for updated server manageability
Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Signed-off-by: Piotr Kwapulinski <piotr.kwapulinski@intel.com> Reviewed-by: Simon Horman <horms@kernel.org> Tested-by: Rafal Romanowski <rafal.romanowski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Stable-dep-of: a7075f501bd3 ("ixgbevf: fix mailbox API compatibility by negotiating supported features") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When user calls open_by_handle_at() on some inode that is not cached, we
will create disconnected dentry for it. If such dentry is a directory,
exportfs_decode_fh_raw() will then try to connect this dentry to the
dentry tree through reconnect_path(). It may happen for various reasons
(such as corrupted fs or race with rename) that the call to
lookup_one_unlocked() in reconnect_one() will fail to find the dentry we
are trying to reconnect and instead create a new dentry under the
parent. Now this dentry will not be marked as disconnected although the
parent still may well be disconnected (at least in case this
inconsistency happened because the fs is corrupted and .. doesn't point
to the real parent directory). This creates inconsistency in
disconnected flags but AFAICS it was mostly harmless. At least until
commit f1ee616214cb ("VFS: don't keep disconnected dentries on d_anon")
which removed adding of most disconnected dentries to sb->s_anon list.
Thus after this commit cleanup of disconnected dentries implicitely
relies on the fact that dput() will immediately reclaim such dentries.
However when some leaf dentry isn't marked as disconnected, as in the
scenario described above, the reclaim doesn't happen and the dentries
are "leaked". Memory reclaim can eventually reclaim them but otherwise
they stay in memory and if umount comes first, we hit infamous "Busy
inodes after unmount" bug. Make sure all dentries created under a
disconnected parent are marked as disconnected as well.
Reported-by: syzbot+1d79ebe5383fc016cf07@syzkaller.appspotmail.com Fixes: f1ee616214cb ("VFS: don't keep disconnected dentries on d_anon") CC: stable@vger.kernel.org Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Christian Brauner <brauner@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Do that before new dentry is visible anywhere. It does create
a new possible state for dentries present in ->d_children/->d_sib -
DCACHE_PAR_LOOKUP present, negative, unhashed, not in in-lookup
hash chains, refcount positive. Those are going to be skipped
by all tree-walkers (both d_walk() callbacks in fs/dcache.c and
explicit loops over children/sibling lists elsewhere) and
dput() is fine with those.
NOTE: dropping the final reference to a "normal" in-lookup dentry
(in in-lookup hash) is a bug - somebody must've forgotten to
call d_lookup_done() on it and bad things will happen. With those
it's OK; if/when we get around to making __dentry_kill() complain
about such breakage, remember that predicate to check should
*not* be just d_in_lookup(victim) but rather a combination of that
with !hlist_bl_unhashed(&victim->d_u.d_in_lookup_hash). Might
be worth considering later...
Reviewed-by: Christian Brauner <brauner@kernel.org> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Stable-dep-of: 56094ad3eaa2 ("vfs: Don't leak disconnected dentries on umount") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Users can create as many monitoring groups as the number of RMIDs supported
by the hardware. However, on AMD systems, only a limited number of RMIDs
are guaranteed to be actively tracked by the hardware. RMIDs that exceed
this limit are placed in an "Unavailable" state.
When a bandwidth counter is read for such an RMID, the hardware sets
MSR_IA32_QM_CTR.Unavailable (bit 62). When such an RMID starts being tracked
again the hardware counter is reset to zero. MSR_IA32_QM_CTR.Unavailable
remains set on first read after tracking re-starts and is clear on all
subsequent reads as long as the RMID is tracked.
resctrl miscounts the bandwidth events after an RMID transitions from the
"Unavailable" state back to being tracked. This happens because when the
hardware starts counting again after resetting the counter to zero, resctrl
in turn compares the new count against the counter value stored from the
previous time the RMID was tracked.
This results in resctrl computing an event value that is either undercounting
(when new counter is more than stored counter) or a mistaken overflow (when
new counter is less than stored counter).
Reset the stored value (arch_mbm_state::prev_msr) of MSR_IA32_QM_CTR to
zero whenever the RMID is in the "Unavailable" state to ensure accurate
counting after the RMID resets to zero when it starts to be tracked again.
Example scenario that results in mistaken overflow
==================================================
1. The resctrl filesystem is mounted, and a task is assigned to a
monitoring group.
$cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes
21323 <- Total bytes on domain 0
"Unavailable" <- Total bytes on domain 1
Task is running on domain 0. Counter on domain 1 is "Unavailable".
2. The task runs on domain 0 for a while and then moves to domain 1. The
counter starts incrementing on domain 1.
$cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 7345357 <- Total bytes on domain 0
4545 <- Total bytes on domain 1
3. At some point, the RMID in domain 0 transitions to the "Unavailable"
state because the task is no longer executing in that domain.
$cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes
"Unavailable" <- Total bytes on domain 0
434341 <- Total bytes on domain 1
4. Since the task continues to migrate between domains, it may eventually
return to domain 0.
$cat /sys/fs/resctrl/mon_groups/test1/mon_data/mon_L3_*/mbm_total_bytes 17592178699059 <- Overflow on domain 0 3232332 <- Total bytes on domain 1
In this case, the RMID on domain 0 transitions from "Unavailable" state to
active state. The hardware sets MSR_IA32_QM_CTR.Unavailable (bit 62) when
the counter is read and begins tracking the RMID counting from 0.
Subsequent reads succeed but return a value smaller than the previously
saved MSR value (7345357). Consequently, the resctrl's overflow logic is
triggered, it compares the previous value (7345357) with the new, smaller
value and incorrectly interprets this as a counter overflow, adding a large
delta.
In reality, this is a false positive: the counter did not overflow but was
simply reset when the RMID transitioned from "Unavailable" back to active
state.
Here is the text from APM [1] available from [2].
"In PQOS Version 2.0 or higher, the MBM hardware will set the U bit on the
first QM_CTR read when it begins tracking an RMID that it was not
previously tracking. The U bit will be zero for all subsequent reads from
that RMID while it is still tracked by the hardware. Therefore, a QM_CTR
read with the U bit set when that RMID is in use by a processor can be
considered 0 when calculating the difference with a subsequent read."
resctrl_arch_rmid_read() adjusts the value obtained from MSR_IA32_QM_CTR to
account for the overflow for MBM events and apply counter scaling for all the
events. This logic is common to both reading an RMID and reading a hardware
counter directly.
Refactor the hardware value adjustment logic into get_corrected_val() to
prepare for support of reading a hardware counter.
If bio is split by internal handling like chunksize or badblocks, the
corresponding trace_block_split() is missing, resulting in blktrace
inability to catch BIO split events and making it harder to analyze the
BIO sequence.
When seq_nr wraps around, the next reorder job with seq 0 is hashed to
the first CPU in padata_do_serial(). Correspondingly, need reset pd->cpu
to the first one when pd->processed wraps around. Otherwise, if the
number of used CPUs is not a power of 2, padata_find_next() will be
checking a wrong list, hence deadlock.
Fixes: 6fc4dbcf0276 ("padata: Replace delayed timer with immediate workqueue in padata_reorder") Cc: <stable@vger.kernel.org> Signed-off-by: Xiao Liang <shaw.leon@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
[ relocated fix to padata_find_next() using pd->processed and pd->cpu structure fields ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When we're removing rmap records for crosslinked blocks, use deferred
intent items so that we can try to free/unmap as many of the old data
structure's blocks as we can in the same transaction as the commit.
Cc: <stable@vger.kernel.org> # v6.6 Fixes: 1c7ce115e52106 ("xfs: reap large AG metadata extents when possible") Signed-off-by: "Darrick J. Wong" <djwong@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de>
[ adapted xfs_refcount_free_cow_extent() and xfs_rmap_free_extent() ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The value of skb_data->wait indicates whether skb is passed on to the
core mac80211 stack or released by the driver itself. Make sure that by
the time skb is added to txwd queue and becomes visible to the completing
side, it has already allocated and initialized TX wait related data (in
case it's needed).
This is found by code review and addresses a possible race scenario
described below:
Waiting thread Completing thread
rtw89_core_send_nullfunc()
rtw89_core_tx_write_link()
...
rtw89_pci_txwd_submit()
skb_data->wait = NULL
/* add skb to the queue */
skb_queue_tail(&txwd->queue, skb)
/* another thread (e.g. rtw89_ops_tx) performs TX kick off for the same queue */
rtw89_pci_napi_poll()
...
rtw89_pci_release_txwd_skb()
/* get skb from the queue */
skb_unlink(skb, &txwd->queue)
rtw89_pci_tx_status()
rtw89_core_tx_wait_complete()
/* use incorrect skb_data->wait */
rtw89_core_tx_kick_off_and_wait()
/* assign skb_data->wait but too late */
Found by Linux Verification Center (linuxtesting.org).
Fixes: 1ae5ca615285 ("wifi: rtw89: add function to wait for completion of TX skbs") Cc: stable@vger.kernel.org Signed-off-by: Fedor Pchelkin <pchelkin@ispras.ru> Acked-by: Ping-Ke Shih <pkshih@realtek.com> Signed-off-by: Ping-Ke Shih <pkshih@realtek.com> Link: https://patch.msgid.link/20250919210852.823912-3-pchelkin@ispras.ru
[ adapted rtw89_core_tx_write_link() modifications to rtw89_core_tx_write() ] Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The data type of loca_last_write_offset is newoffset4 and is switched
on a boolean value, no_newoffset, that indicates if a previous write
occurred or not. If no_newoffset is FALSE, an offset is not given.
This means that client does not try to update the file size. Thus,
server should not try to calculate new file size and check if it fits
into the segment range. See RFC 8881, section 12.5.4.2.
Sometimes the current incorrect logic may cause clients to hang when
trying to sync an inode. If layoutcommit fails, the client marks the
inode as dirty again.
Fixes: 9cf514ccfacb ("nfsd: implement pNFS operations") Cc: stable@vger.kernel.org Co-developed-by: Konstantin Evtushenko <koevtushenko@yandex.com> Signed-off-by: Konstantin Evtushenko <koevtushenko@yandex.com> Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When pNFS client in the block or scsi layout mode sends layoutcommit
to MDS, a variable length array of modified extents is supplied within
the request. This patch allows the server to accept such extent arrays
if they do not fit within single memory page.
The issue can be reproduced when writing to a 1GB file using FIO with
O_DIRECT, 4K block and large I/O depth without preallocation of the
file. In this case, the server returns NFSERR_BADXDR to the client.
Co-developed-by: Konstantin Evtushenko <koevtushenko@yandex.com> Signed-off-by: Konstantin Evtushenko <koevtushenko@yandex.com> Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Stable-dep-of: d68886bae76a ("NFSD: Fix last write offset handling in layoutcommit") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove dprintk in nfsd4_layoutcommit. These are not needed
in day to day usage, and the information is also available
in Wireshark when capturing NFS traffic.
Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Stable-dep-of: d68886bae76a ("NFSD: Fix last write offset handling in layoutcommit") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Compilers may optimize the layout of C structures, so we should not rely
on sizeof struct and memcpy to encode and decode XDR structures. The byte
order of the fields should also be taken into account.
This patch adds the correct functions to handle the deviceid4 structure
and removes the pad field, which is currently not used by NFSD, from the
runtime state. The server's byte order is preserved because the deviceid4
blob on the wire is only used as a cookie by the client.
Signed-off-by: Sergey Bashirov <sergeybashirov@gmail.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Stable-dep-of: d68886bae76a ("NFSD: Fix last write offset handling in layoutcommit") Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
projects / thirdparty / kernel / stable.git / log
