Now that USB HID++ devices can gather a serial number that matches the
one that would be gathered when connected through a Unifying receiver,
remove the last difference by dropping the product ID as devices
usually have different product IDs when connected through USB or
Unifying.
For example, on the serials on a G903 wired/wireless mouse:
- Unifying before patch: 4067-e8-ce-cd-45
- USB before patch: c086-e8-ce-cd-45
- Unifying and USB after patch: e8-ce-cd-45
For devices that support the 0x0003 feature (Device Information) version 4,
set the serial based on the output of that feature, rather than relying
on the usbhid code setting the USB serial.
This should allow the serial when connected through USB to (nearly)
match the one when connected through a unifying receiver.
For example, on the serials on a G903 wired/wireless mouse:
- Unifying: 4067-e8-ce-cd-45
- USB before patch: 017C385C3837
- USB after patch: c086-e8-ce-cd-45
conn->chan_lock isn't acquired before l2cap_get_chan_by_scid,
if l2cap_get_chan_by_scid returns NULL, then 'bad unlock balance'
is triggered.
Reported-by: syzbot+9519d6b5b79cf7787cf3@syzkaller.appspotmail.com Link: https://lore.kernel.org/all/000000000000894f5f05f95e9f4d@google.com/ Signed-off-by: Min Li <lm0963hack@gmail.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
A received TKIP key may be up to 32 bytes because it may contain
MIC rx/tx keys too. These are not used by iwl and copying these
over overflows the iwl_keyinfo.key field.
Add a check to not copy more data to iwl_keyinfo.key then will fit.
This fixes backtraces like this one:
memcpy: detected field-spanning write (size 32) of single field "sta_cmd.key.key" at drivers/net/wireless/intel/iwlwifi/dvm/sta.c:1103 (size 16)
WARNING: CPU: 1 PID: 946 at drivers/net/wireless/intel/iwlwifi/dvm/sta.c:1103 iwlagn_send_sta_key+0x375/0x390 [iwldvm]
<snip>
Hardware name: Dell Inc. Latitude E6430/0H3MT5, BIOS A21 05/08/2017
RIP: 0010:iwlagn_send_sta_key+0x375/0x390 [iwldvm]
<snip>
Call Trace:
<TASK>
iwl_set_dynamic_key+0x1f0/0x220 [iwldvm]
iwlagn_mac_set_key+0x1e4/0x280 [iwldvm]
drv_set_key+0xa4/0x1b0 [mac80211]
ieee80211_key_enable_hw_accel+0xa8/0x2d0 [mac80211]
ieee80211_key_replace+0x22d/0x8e0 [mac80211]
<snip>
When the length of best extent found is less than the length of goal extent
we need to make sure that the best extent atleast covers the start of the
original request. This is done by adjusting the ac_b_ex.fe_logical (logical
start) of the extent.
While doing so, the current logic sometimes results in the best extent's
logical range overflowing the goal extent. Since this best extent is later
added to the inode preallocation list, we have a possibility of introducing
overlapping preallocations. This is discussed in detail here [1].
As per Jan's suggestion, to fix this, replace the existing logic with the
below logic for adjusting best extent as it keeps fragmentation in check
while ensuring logical range of best extent doesn't overflow out of goal
extent:
1. Check if best extent can be kept at end of goal range and still cover
original start.
2. Else, check if best extent can be kept at start of goal range and still
cover original start.
3. Else, keep the best extent at start of original request.
Also, add a few extra BUG_ONs that might help catch errors faster.
We need to set ac_g_ex to notify the goal start used in
ext4_mb_find_by_goal. Set ac_g_ex instead of ac_f_ex in
ext4_mb_normalize_request.
Besides we should assure goal start is in range [first_data_block,
blocks_count) as ext4_mb_initialize_context does.
[ Added a check to make sure size is less than ar->pright; otherwise
we could end up passing an underflowed value of ar->pright - size to
ext4_get_group_no_and_offset(), which will trigger a BUG_ON later on.
- TYT ]
The maximum allowed height of an inode's metadata tree depends on the
filesystem block size; it is lower for bigger-block filesystems. When
reading in an inode, make sure that the height doesn't exceed the
maximum allowed height.
Arrays like sd_heightsize are sized to be big enough for any filesystem
block size; they will often be slightly bigger than what's needed for a
specific filesystem.
Reported-by: syzbot+45d4691b1ed3c48eba05@syzkaller.appspotmail.com Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
mptlan_probe() calls mpt_register_lan_device() which initializes the
&priv->post_buckets_task workqueue. A call to
mpt_lan_wake_post_buckets_task() will subsequently start the work.
During driver unload in mptlan_remove() the following race may occur:
When calling irq_set_affinity_notifier() with NULL at the notify
argument, it will cause freeing of the glue pointer in the
corresponding array entry but will leave the pointer in the array. A
subsequent call to free_irq_cpu_rmap() will try to free this entry again
leading to possible use after free.
Fix that by setting NULL to the array entry and checking that we have
non-zero at the array entry when iterating over the array in
free_irq_cpu_rmap().
The current code does not suffer from this since there are no cases
where irq_set_affinity_notifier(irq, NULL) (note the NULL passed for the
notify arg) is called, followed by a call to free_irq_cpu_rmap() so we
don't hit and issue. Subsequent patches in this series excersize this
flow, hence the required fix.
Cc: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Eli Cohen <elic@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
When setting the XPS value of a TX queue, warn the user once if the
index of the queue is greater than the number of allocated TX queues.
Previously, this scenario went uncaught. In the best case, it resulted
in unnecessary allocations. In the worst case, it resulted in
out-of-bounds memory references through calls to `netdev_get_tx_queue(
dev, index)`. Therefore, it is important to inform the user but not
worth returning an error and risk downing the netdevice.
With clang's kernel control flow integrity (kCFI, CONFIG_CFI_CLANG),
indirect call targets are validated against the expected function
pointer prototype to make sure the call target is valid to help mitigate
ROP attacks. If they are not identical, there is a failure at run time,
which manifests as either a kernel panic or thread getting killed. A
warning in clang aims to catch these at compile time, which reveals:
drivers/net/ethernet/pasemi/pasemi_mac.c:1665:21: error: incompatible function pointer types initializing 'netdev_tx_t (*)(struct sk_buff *, struct net_device *)' (aka 'enum netdev_tx (*)(struct sk_buff *, struct net_device *)') with an expression of type 'int (struct sk_buff *, struct net_device *)' [-Werror,-Wincompatible-function-pointer-types-strict]
.ndo_start_xmit = pasemi_mac_start_tx,
^~~~~~~~~~~~~~~~~~~
1 error generated.
->ndo_start_xmit() in 'struct net_device_ops' expects a return type of
'netdev_tx_t', not 'int'. Adjust the return type of
pasemi_mac_start_tx() to match the prototype's to resolve the warning.
While PowerPC does not currently implement support for kCFI, it could in
the future, which means this warning becomes a fatal CFI failure at run
time.
Check that log of block size stored in the superblock has sensible
value. Otherwise the shift computing the block size can overflow leading
to undefined behavior.
Reported-by: syzbot+4fec412f59eba8c01b77@syzkaller.appspotmail.com Signed-off-by: Jan Kara <jack@suse.cz> Signed-off-by: Sasha Levin <sashal@kernel.org>
Apparently the hex passphrase mechanism does not work on newer
chips/firmware (e.g. BCM4387). It seems there was a simple way of
passing it in binary all along, so use that and avoid the hexification.
OpenBSD has been doing it like this from the beginning, so this should
work on all chips.
Also clear the structure before setting the PMK. This was leaking
uninitialized stack contents to the device.
Reviewed-by: Linus Walleij <linus.walleij@linaro.org> Reviewed-by: Arend van Spriel <arend.vanspriel@broadcom.com> Signed-off-by: Hector Martin <marcan@marcan.st> Signed-off-by: Kalle Valo <kvalo@kernel.org> Link: https://lore.kernel.org/r/20230214092423.15175-6-marcan@marcan.st Signed-off-by: Sasha Levin <sashal@kernel.org>
ACPI_ALLOCATE_ZEROED may fails, object_info might be null and will cause
null pointer dereference later.
Link: https://github.com/acpica/acpica/commit/0d5f467d Signed-off-by: Bob Moore <robert.moore@intel.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Before this change we see the following UBSAN stack trace in Fuchsia:
#0 0x000021e4213b3302 in acpi_ds_init_aml_walk(struct acpi_walk_state*, union acpi_parse_object*, struct acpi_namespace_node*, u8*, u32, struct acpi_evaluate_info*, u8) ../../third_party/acpica/source/components/dispatcher/dswstate.c:682 <platform-bus-x86.so>+0x233302
#1.2 0x000020d0f660777f in ubsan_get_stack_trace() compiler-rt/lib/ubsan/ubsan_diag.cpp:41 <libclang_rt.asan.so>+0x3d77f
#1.1 0x000020d0f660777f in maybe_print_stack_trace() compiler-rt/lib/ubsan/ubsan_diag.cpp:51 <libclang_rt.asan.so>+0x3d77f
#1 0x000020d0f660777f in ~scoped_report() compiler-rt/lib/ubsan/ubsan_diag.cpp:387 <libclang_rt.asan.so>+0x3d77f
#2 0x000020d0f660b96d in handlepointer_overflow_impl() compiler-rt/lib/ubsan/ubsan_handlers.cpp:809 <libclang_rt.asan.so>+0x4196d
#3 0x000020d0f660b50d in compiler-rt/lib/ubsan/ubsan_handlers.cpp:815 <libclang_rt.asan.so>+0x4150d
#4 0x000021e4213b3302 in acpi_ds_init_aml_walk(struct acpi_walk_state*, union acpi_parse_object*, struct acpi_namespace_node*, u8*, u32, struct acpi_evaluate_info*, u8) ../../third_party/acpica/source/components/dispatcher/dswstate.c:682 <platform-bus-x86.so>+0x233302
#5 0x000021e4213e2369 in acpi_ds_call_control_method(struct acpi_thread_state*, struct acpi_walk_state*, union acpi_parse_object*) ../../third_party/acpica/source/components/dispatcher/dsmethod.c:605 <platform-bus-x86.so>+0x262369
#6 0x000021e421437fac in acpi_ps_parse_aml(struct acpi_walk_state*) ../../third_party/acpica/source/components/parser/psparse.c:550 <platform-bus-x86.so>+0x2b7fac
#7 0x000021e4214464d2 in acpi_ps_execute_method(struct acpi_evaluate_info*) ../../third_party/acpica/source/components/parser/psxface.c:244 <platform-bus-x86.so>+0x2c64d2
#8 0x000021e4213aa052 in acpi_ns_evaluate(struct acpi_evaluate_info*) ../../third_party/acpica/source/components/namespace/nseval.c:250 <platform-bus-x86.so>+0x22a052
#9 0x000021e421413dd8 in acpi_ns_init_one_device(acpi_handle, u32, void*, void**) ../../third_party/acpica/source/components/namespace/nsinit.c:735 <platform-bus-x86.so>+0x293dd8
#10 0x000021e421429e98 in acpi_ns_walk_namespace(acpi_object_type, acpi_handle, u32, u32, acpi_walk_callback, acpi_walk_callback, void*, void**) ../../third_party/acpica/source/components/namespace/nswalk.c:298 <platform-bus-x86.so>+0x2a9e98
#11 0x000021e4214131ac in acpi_ns_initialize_devices(u32) ../../third_party/acpica/source/components/namespace/nsinit.c:268 <platform-bus-x86.so>+0x2931ac
#12 0x000021e42147c40d in acpi_initialize_objects(u32) ../../third_party/acpica/source/components/utilities/utxfinit.c:304 <platform-bus-x86.so>+0x2fc40d
#13 0x000021e42126d603 in acpi::acpi_impl::initialize_acpi(acpi::acpi_impl*) ../../src/devices/board/lib/acpi/acpi-impl.cc:224 <platform-bus-x86.so>+0xed603
Add a simple check that avoids incrementing a pointer by zero, but
otherwise behaves as before. Note that our findings are against ACPICA 20221020, but the same code exists on master.
Link: https://github.com/acpica/acpica/commit/770653e3 Signed-off-by: Bob Moore <robert.moore@intel.com> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In tegra_sor_compute_config(), the 32-bit value mode->clock is
multiplied by 1000, and assigned to the u64 variable pclk. We can avoid
a potential 32-bit integer overflow by casting mode->clock to u64 before
we do the arithmetic and assignment.
When removing custom query handlers, the handler might still
be used inside the EC query workqueue, causing a kernel oops
if the module holding the callback function was already unloaded.
Fix this by flushing the EC query workqueue when removing
custom query handlers.
Tested on a Acer Travelmate 4002WLMi
Signed-off-by: Armin Wolf <W_Armin@gmx.de> Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
In r592_probe, dev->detect_timer was bound with r592_detect_timer.
In r592_irq function, the timer function will be invoked by mod_timer.
If we remove the module which will call hantro_release to make cleanup,
there may be a unfinished work. The possible sequence is as follows,
which will cause a typical UAF bug.
Fix it by canceling the work before cleanup in r592_remove.
There is no sense in doing a cache sync on REGCACHE_NONE regmaps.
Instead of panicking the kernel due to missing cache_ops, return an error
to client driver.
syzbot is hitting WARN_ON() in hfsplus_cat_{read,write}_inode(), for
crafted filesystem image can contain bogus length. There conditions are
not kernel bugs that can justify kernel to panic.
KCSAN found a data race of sk->sk_receive_queue->qlen where recvmsg()
updates qlen under the queue lock and sendmsg() checks qlen under
unix_state_sock(), not the queue lock, so the reader side needs
READ_ONCE().
BUG: KCSAN: data-race in __skb_try_recv_from_queue / unix_wait_for_peer
write (marked) to 0xffff888019fe7c68 of 4 bytes by task 49792 on cpu 0:
__skb_unlink include/linux/skbuff.h:2347 [inline]
__skb_try_recv_from_queue+0x3de/0x470 net/core/datagram.c:197
__skb_try_recv_datagram+0xf7/0x390 net/core/datagram.c:263
__unix_dgram_recvmsg+0x109/0x8a0 net/unix/af_unix.c:2452
unix_dgram_recvmsg+0x94/0xa0 net/unix/af_unix.c:2549
sock_recvmsg_nosec net/socket.c:1019 [inline]
____sys_recvmsg+0x3a3/0x3b0 net/socket.c:2720
___sys_recvmsg+0xc8/0x150 net/socket.c:2764
do_recvmmsg+0x182/0x560 net/socket.c:2858
__sys_recvmmsg net/socket.c:2937 [inline]
__do_sys_recvmmsg net/socket.c:2960 [inline]
__se_sys_recvmmsg net/socket.c:2953 [inline]
__x64_sys_recvmmsg+0x153/0x170 net/socket.c:2953
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x3b/0x90 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc
read to 0xffff888019fe7c68 of 4 bytes by task 49793 on cpu 1:
skb_queue_len include/linux/skbuff.h:2127 [inline]
unix_recvq_full net/unix/af_unix.c:229 [inline]
unix_wait_for_peer+0x154/0x1a0 net/unix/af_unix.c:1445
unix_dgram_sendmsg+0x13bc/0x14b0 net/unix/af_unix.c:2048
sock_sendmsg_nosec net/socket.c:724 [inline]
sock_sendmsg+0x148/0x160 net/socket.c:747
____sys_sendmsg+0x20e/0x620 net/socket.c:2503
___sys_sendmsg+0xc6/0x140 net/socket.c:2557
__sys_sendmmsg+0x11d/0x370 net/socket.c:2643
__do_sys_sendmmsg net/socket.c:2672 [inline]
__se_sys_sendmmsg net/socket.c:2669 [inline]
__x64_sys_sendmmsg+0x58/0x70 net/socket.c:2669
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x3b/0x90 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x72/0xdc
value changed: 0x0000000b -> 0x00000001
Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 49793 Comm: syz-executor.0 Not tainted 6.3.0-rc7-02330-gca6270c12e20 #2
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.16.0-0-gd239552ce722-prebuilt.qemu.org 04/01/2014
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-by: syzbot <syzkaller@googlegroups.com> Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Michal Kubiak <michal.kubiak@intel.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If skb enqueue the qdisc, fq_skb_cb(skb)->time_to_send is changed which
is actually skb->cb, and IPCB(skb_in)->opt will be used in
__ip_options_echo. It is possible that memcpy is out of bounds and lead
to stack overflow.
We should clear skb->cb before ip_local_out or ip6_local_out.
v2:
1. clean the stack info
2. use IPCB/IP6CB instead of skb->cb
To reproduce(ipvlan with IPVLAN_MODE_L3):
Env setting:
=======================================================
modprobe ipvlan ipvlan_default_mode=1
sysctl net.ipv4.conf.eth0.forwarding=1
iptables -t nat -A POSTROUTING -s 20.0.0.0/255.255.255.0 -o eth0 -j
MASQUERADE
ip link add gw link eth0 type ipvlan
ip -4 addr add 20.0.0.254/24 dev gw
ip netns add net1
ip link add ipv1 link eth0 type ipvlan
ip link set ipv1 netns net1
ip netns exec net1 ip link set ipv1 up
ip netns exec net1 ip -4 addr add 20.0.0.4/24 dev ipv1
ip netns exec net1 route add default gw 20.0.0.254
ip netns exec net1 tc qdisc add dev ipv1 root netem loss 10%
ifconfig gw up
iptables -t filter -A OUTPUT -p tcp --dport 8888 -j REJECT --reject-with
icmp-port-unreachable
=======================================================
And then excute the shell(curl any address of eth0 can reach):
for((i=1;i<=100000;i++))
do
ip netns exec net1 curl x.x.x.x:8888
done
=======================================================
Fixes: 2ad7bf363841 ("ipvlan: Initial check-in of the IPVLAN driver.") Signed-off-by: "t.feng" <fengtao40@huawei.com> Suggested-by: Florian Westphal <fw@strlen.de> Reviewed-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
do_recvmmsg() can write to sk->sk_err from multiple threads.
As said before, many other points reading or writing sk_err
need annotations.
Fixes: 34b88a68f26a ("net: Fix use after free in the recvmmsg exit path") Signed-off-by: Eric Dumazet <edumazet@google.com> Reported-by: syzbot <syzkaller@googlegroups.com> Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
read to 0xffff88813ea4db59 of 1 bytes by task 28222 on cpu 1:
netlink_recvmsg+0x3b4/0x730 net/netlink/af_netlink.c:2022
sock_recvmsg_nosec+0x4c/0x80 net/socket.c:1017
____sys_recvmsg+0x2db/0x310 net/socket.c:2718
___sys_recvmsg net/socket.c:2762 [inline]
do_recvmmsg+0x2e5/0x710 net/socket.c:2856
__sys_recvmmsg net/socket.c:2935 [inline]
__do_sys_recvmmsg net/socket.c:2958 [inline]
__se_sys_recvmmsg net/socket.c:2951 [inline]
__x64_sys_recvmmsg+0xe2/0x160 net/socket.c:2951
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x41/0xc0 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd
value changed: 0x00 -> 0x01
Fixes: 16b304f3404f ("netlink: Eliminate kmalloc in netlink dump operation.") Reported-by: syzbot <syzkaller@googlegroups.com> Signed-off-by: Eric Dumazet <edumazet@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
syzbot is reporting circular locking dependency which involves
zonelist_update_seq seqlock [1], for this lock is checked by memory
allocation requests which do not need to be retried.
One deadlock scenario is kmalloc(GFP_ATOMIC) from an interrupt handler.
CPU0
----
__build_all_zonelists() {
write_seqlock(&zonelist_update_seq); // makes zonelist_update_seq.seqcount odd
// e.g. timer interrupt handler runs at this moment
some_timer_func() {
kmalloc(GFP_ATOMIC) {
__alloc_pages_slowpath() {
read_seqbegin(&zonelist_update_seq) {
// spins forever because zonelist_update_seq.seqcount is odd
}
}
}
}
// e.g. timer interrupt handler finishes
write_sequnlock(&zonelist_update_seq); // makes zonelist_update_seq.seqcount even
}
This deadlock scenario can be easily eliminated by not calling
read_seqbegin(&zonelist_update_seq) from !__GFP_DIRECT_RECLAIM allocation
requests, for retry is applicable to only __GFP_DIRECT_RECLAIM allocation
requests. But Michal Hocko does not know whether we should go with this
approach.
Another deadlock scenario which syzbot is reporting is a race between
kmalloc(GFP_ATOMIC) from tty_insert_flip_string_and_push_buffer() with
port->lock held and printk() from __build_all_zonelists() with
zonelist_update_seq held.
preventing interrupt context from calling kmalloc(GFP_ATOMIC)
and
preventing printk() from calling console_flush_all()
while zonelist_update_seq.seqcount is odd.
Since Petr Mladek thinks that __build_all_zonelists() can become a
candidate for deferring printk() [2], let's address this problem by
disabling local interrupts in order to avoid kmalloc(GFP_ATOMIC)
and
disabling synchronous printk() in order to avoid console_flush_all()
.
As a side effect of minimizing duration of zonelist_update_seq.seqcount
being odd by disabling synchronous printk(), latency at
read_seqbegin(&zonelist_update_seq) for both !__GFP_DIRECT_RECLAIM and
__GFP_DIRECT_RECLAIM allocation requests will be reduced. Although, from
lockdep perspective, not calling read_seqbegin(&zonelist_update_seq) (i.e.
do not record unnecessary locking dependency) from interrupt context is
still preferable, even if we don't allow calling kmalloc(GFP_ATOMIC)
inside
write_seqlock(&zonelist_update_seq)/write_sequnlock(&zonelist_update_seq)
section...
Link: https://lkml.kernel.org/r/8796b95c-3da3-5885-fddd-6ef55f30e4d3@I-love.SAKURA.ne.jp Fixes: 3d36424b3b58 ("mm/page_alloc: fix race condition between build_all_zonelists and page allocation") Link: https://lkml.kernel.org/r/ZCrs+1cDqPWTDFNM@alley Reported-by: syzbot <syzbot+223c7461c58c58a4cb10@syzkaller.appspotmail.com> Link: https://syzkaller.appspot.com/bug?extid=223c7461c58c58a4cb10 Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Mel Gorman <mgorman@techsingularity.net> Cc: Petr Mladek <pmladek@suse.com> Cc: David Hildenbrand <david@redhat.com> Cc: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Cc: John Ogness <john.ogness@linutronix.de> Cc: Patrick Daly <quic_pdaly@quicinc.com> Cc: Sergey Senozhatsky <senozhatsky@chromium.org> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
[This patch implements subset of original commit 85e3e7fbbb72 ("printk:
remove NMI tracking") where commit 1007843a9190 ("mm/page_alloc: fix
potential deadlock on zonelist_update_seq seqlock") depends on, for
commit 3d36424b3b58 ("mm/page_alloc: fix race condition between
build_all_zonelists and page allocation") was backported to stable.]
All NMI contexts are handled the same as the safe context: store the
message and defer printing. There is no need to have special NMI
context tracking for this. Using in_nmi() is enough.
There are several parts of the kernel that are manually calling into
the printk NMI context tracking in order to cause general printk
deferred printing:
For arm/kernel/smp.c and powerpc/kexec/crash.c, provide a new
function pair printk_deferred_enter/exit that explicitly achieves the
same objective.
For ftrace, remove the printk context manipulation completely. It was
added in commit 03fc7f9c99c1 ("printk/nmi: Prevent deadlock when
accessing the main log buffer in NMI"). The purpose was to enforce
storing messages directly into the ring buffer even in NMI context.
It really should have only modified the behavior in NMI context.
There is no need for a special behavior any longer. All messages are
always stored directly now. The console deferring is handled
transparently in vprintk().
Signed-off-by: John Ogness <john.ogness@linutronix.de>
[pmladek@suse.com: Remove special handling in ftrace.c completely. Signed-off-by: Petr Mladek <pmladek@suse.com> Link: https://lore.kernel.org/r/20210715193359.25946-5-john.ogness@linutronix.de
[penguin-kernel: Copy only printk_deferred_{enter,safe}() definition ] Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When we receive a flush command (or "barrier" in DRBD), we currently use
a REQ_OP_FLUSH with the REQ_PREFLUSH flag set.
The correct way to submit a flush bio is by using a REQ_OP_WRITE without
any data, and set the REQ_PREFLUSH flag.
Since commit b4a6bb3a67aa ("block: add a sanity check for non-write
flush/fua bios"), this triggers a warning in the block layer, but this
has been broken for quite some time before that.
So use the correct set of flags to actually make the flush happen.
Cc: Christoph Hellwig <hch@infradead.org> Cc: stable@vger.kernel.org Fixes: f9ff0da56437 ("drbd: allow parallel flushes for multi-volume resources") Reported-by: Thomas Voegtle <tv@lio96.de> Signed-off-by: Christoph Böhmwalder <christoph.boehmwalder@linbit.com> Reviewed-by: Christoph Hellwig <hch@lst.de> Link: https://lore.kernel.org/r/20230503121937.17232-1-christoph.boehmwalder@linbit.com Signed-off-by: Jens Axboe <axboe@kernel.dk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ilpo Järvinen [Thu, 11 May 2023 12:32:44 +0000 (15:32 +0300)]
serial: 8250: Fix serial8250_tx_empty() race with DMA Tx
There's a potential race before THRE/TEMT deasserts when DMA Tx is
starting up (or the next batch of continuous Tx is being submitted).
This can lead to misdetecting Tx empty condition.
It is entirely normal for THRE/TEMT to be set for some time after the
DMA Tx had been setup in serial8250_tx_dma(). As Tx side is definitely
not empty at that point, it seems incorrect for serial8250_tx_empty()
claim Tx is empty.
Fix the race by also checking in serial8250_tx_empty() whether there's
DMA Tx active.
Note: This fix only addresses in-kernel race mainly to make using
TCSADRAIN/FLUSH robust. Userspace can still cause other races but they
seem userspace concurrency control problems.
Fixes: 9ee4b83e51f74 ("serial: 8250: Add support for dmaengine") Cc: stable@vger.kernel.org Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Link: https://lore.kernel.org/r/20230317113318.31327-3-ilpo.jarvinen@linux.intel.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 146a37e05d620cef4ad430e5d1c9c077fe6fa76f) Signed-off-by: Ilpo Järvinen <ilpo.jarvinen@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ilpo Järvinen [Thu, 11 May 2023 12:32:43 +0000 (15:32 +0300)]
tty: Prevent writing chars during tcsetattr TCSADRAIN/FLUSH
If userspace races tcsetattr() with a write, the drained condition
might not be guaranteed by the kernel. There is a race window after
checking Tx is empty before tty_set_termios() takes termios_rwsem for
write. During that race window, more characters can be queued by a
racing writer.
Any ongoing transmission might produce garbage during HW's
->set_termios() call. The intent of TCSADRAIN/FLUSH seems to be
preventing such a character corruption. If those flags are set, take
tty's write lock to stop any writer before performing the lower layer
Tx empty check and wait for the pending characters to be sent (if any).
The initial wait for all-writers-done must be placed outside of tty's
write lock to avoid deadlock which makes it impossible to use
tty_wait_until_sent(). The write lock is retried if a racing write is
detected.
Noticed with gcc 10 (fedora rawhide) that those variables were not being
declared as static, so end up with:
ld: /tmp/build/perf/bench/epoll-wait.o:/git/perf/tools/perf/bench/epoll-wait.c:93: multiple definition of `end'; /tmp/build/perf/bench/futex-hash.o:/git/perf/tools/perf/bench/futex-hash.c:40: first defined here
ld: /tmp/build/perf/bench/epoll-wait.o:/git/perf/tools/perf/bench/epoll-wait.c:93: multiple definition of `start'; /tmp/build/perf/bench/futex-hash.o:/git/perf/tools/perf/bench/futex-hash.c:40: first defined here
ld: /tmp/build/perf/bench/epoll-wait.o:/git/perf/tools/perf/bench/epoll-wait.c:93: multiple definition of `runtime'; /tmp/build/perf/bench/futex-hash.o:/git/perf/tools/perf/bench/futex-hash.c:40: first defined here
ld: /tmp/build/perf/bench/epoll-ctl.o:/git/perf/tools/perf/bench/epoll-ctl.c:38: multiple definition of `end'; /tmp/build/perf/bench/futex-hash.o:/git/perf/tools/perf/bench/futex-hash.c:40: first defined here
ld: /tmp/build/perf/bench/epoll-ctl.o:/git/perf/tools/perf/bench/epoll-ctl.c:38: multiple definition of `start'; /tmp/build/perf/bench/futex-hash.o:/git/perf/tools/perf/bench/futex-hash.c:40: first defined here
ld: /tmp/build/perf/bench/epoll-ctl.o:/git/perf/tools/perf/bench/epoll-ctl.c:38: multiple definition of `runtime'; /tmp/build/perf/bench/futex-hash.o:/git/perf/tools/perf/bench/futex-hash.c:40: first defined here
make[4]: *** [/git/perf/tools/build/Makefile.build:145: /tmp/build/perf/bench/perf-in.o] Error 1
Prefix those with bench__ and add them to bench/bench.h, so that we can
share those on the tools needing to access those variables from signal
handlers.
Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Davidlohr Bueso <dave@stgolabs.net> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Link: http://lore.kernel.org/lkml/20200303155811.GD13702@kernel.org Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In ext4_xattr_move_to_block(), the value of the extended attribute
which we need to move to an external block may be allocated by
kvmalloc() if the value is stored in an external inode. So at the end
of the function the code tried to check if this was the case by
testing entry->e_value_inum.
However, at this point, the pointer to the xattr entry is no longer
valid, because it was removed from the original location where it had
been stored. So we could end up calling kvfree() on a pointer which
was not allocated by kvmalloc(); or we could also potentially leak
memory by not freeing the buffer when it should be freed. Fix this by
storing whether it should be freed in a separate variable.
If a malicious fuzzer overwrites the ext4 superblock while it is
mounted such that the s_first_data_block is set to a very large
number, the calculation of the block group can underflow, and trigger
a BUG_ON check. Change this to be an ext4_warning so that we don't
crash the kernel.
In ext4_update_inline_data(), if ext4_xattr_ibody_get() fails for any
reason, it's best if we just fail as opposed to stumbling on,
especially if the failure is EFSCORRUPTED.
Normally the extended attributes in the inode body would have been
checked when the inode is first opened, but if someone is writing to
the block device while the file system is mounted, it's possible for
the inode table to get corrupted. Add bounds checking to avoid
reading beyond the end of allocated memory if this happens.
If there are failures while changing the mount options in
__ext4_remount(), we need to restore the old mount options.
This commit fixes two problem. The first is there is a chance that we
will free the old quota file names before a potential failure leading
to a use-after-free. The second problem addressed in this commit is
if there is a failed read/write to read-only transition, if the quota
has already been suspended, we need to renable quota handling.
Replace
le16_to_cpu(sbi->s_es->s_desc_size)
with
sbi->s_desc_size
It reduces ext4's compiled text size, and makes the code more efficient
(we remove an extra indirect reference and a potential byte
swap on big endian systems), and there is no downside. It also avoids the
potential KASAN / syzkaller failure, as a bonus.
__setup() handlers should return 1 to obsolete_checksetup() in
init/main.c to indicate that the boot option has been handled.
A return of 0 causes the boot option/value to be listed as an Unknown
kernel parameter and added to init's (limited) argument or environment
strings. Also, error return codes don't mean anything to
obsolete_checksetup() -- only non-zero (usually 1) or zero.
So return 1 from nmi_debug_setup().
Fixes: 1e1030dccb10 ("sh: nmi_debug support.") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Reported-by: Igor Zhbanov <izh1979@gmail.com>
Link: lore.kernel.org/r/64644a2f-4a20-bab3-1e15-3b2cdd0defe3@omprussia.ru Cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Cc: Rich Felker <dalias@libc.org> Cc: linux-sh@vger.kernel.org Cc: stable@vger.kernel.org Reviewed-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Link: https://lore.kernel.org/r/20230306040037.20350-3-rdunlap@infradead.org Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fix a warning that was reported by the kernel test robot:
In file included from ../include/math-emu/soft-fp.h:27,
from ../arch/sh/math-emu/math.c:22:
../arch/sh/include/asm/sfp-machine.h:17: warning: "__BYTE_ORDER" redefined
17 | #define __BYTE_ORDER __BIG_ENDIAN
In file included from ../arch/sh/math-emu/math.c:21:
../arch/sh/math-emu/sfp-util.h:71: note: this is the location of the previous definition
71 | #define __BYTE_ORDER __LITTLE_ENDIAN
Fixes: b929926f01f2 ("sh: define __BIG_ENDIAN for math-emu") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Reported-by: kernel test robot <lkp@intel.com>
Link: lore.kernel.org/r/202111121827.6v6SXtVv-lkp@intel.com Cc: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Cc: Yoshinori Sato <ysato@users.sourceforge.jp> Cc: Rich Felker <dalias@libc.org> Cc: linux-sh@vger.kernel.org Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Cc: stable@vger.kernel.org Reviewed-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Link: https://lore.kernel.org/r/20230306040037.20350-5-rdunlap@infradead.org Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change type of pcchunk->Length from u32 to u64 to match
smb2_copychunk_range arguments type. Fixes the problem where performing
server-side copy with CIFS_IOC_COPYCHUNK_FILE ioctl resulted in incomplete
copy of large files while returning -EINVAL.
Fixes: 9bf0c9cd4314 ("CIFS: Fix SMB2/SMB3 Copy offload support (refcopy) for large files") Cc: <stable@vger.kernel.org> Signed-off-by: Pawel Witek <pawel.ireneusz.witek@gmail.com> Signed-off-by: Steve French <stfrench@microsoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Check nodesize to sectorsize in alignment check in print_extent_item.
The comment states that and this is correct, similar check is done
elsewhere in the functions.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: ea57788eb76d ("btrfs: require only sector size alignment for parent eb bytenr") CC: stable@vger.kernel.org # 4.14+ Reviewed-by: Qu Wenruo <wqu@suse.com> Signed-off-by: Anastasia Belova <abelova@astralinux.ru> Reviewed-by: David Sterba <dsterba@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
A call to btrfs_prev_leaf() may end up returning a path that points to the
same item (key) again. This happens if while btrfs_prev_leaf(), after we
release the path, a concurrent insertion happens, which moves items off
from a sibling into the front of the previous leaf, and an item with the
computed previous key does not exists.
For example, suppose we have the two following leaves:
If we call btrfs_prev_leaf(), from btrfs_previous_item() for example, with
a path pointing to leaf B and slot 0 and the following happens:
1) At btrfs_prev_leaf() we compute the previous key to search as:
(300 96 19), which is a key that does not exists in the tree;
2) Then we call btrfs_release_path() at btrfs_prev_leaf();
3) Some other task inserts a key at leaf A, that sorts before the key at
slot 20, for example it has an objectid of 299. In order to make room
for the new key, the key at slot 22 is moved to the front of leaf B.
This happens at push_leaf_right(), called from split_leaf().
4) At btrfs_prev_leaf() we call btrfs_search_slot() for the computed
previous key: (300 96 19). Since the key does not exists,
btrfs_search_slot() returns 1 and with a path pointing to leaf B
and slot 1, the item with key (300 96 20);
5) This makes btrfs_prev_leaf() return a path that points to slot 1 of
leaf B, the same key as before it was called, since the key at slot 0
of leaf B (300 96 16) is less than the computed previous key, which is
(300 96 19);
6) As a consequence btrfs_previous_item() returns a path that points again
to the item with key (300 96 20).
For some users of btrfs_prev_leaf() or btrfs_previous_item() this may not
be functional a problem, despite not making sense to return a new path
pointing again to the same item/key. However for a caller such as
tree-log.c:log_dir_items(), this has a bad consequence, as it can result
in not logging some dir index deletions in case the directory is being
logged without holding the inode's VFS lock (logging triggered while
logging a child inode for example) - for the example scenario above, in
case the dir index keys 17, 18 and 19 were deleted in the current
transaction.
CC: stable@vger.kernel.org # 4.14+ Reviewed-by: Josef Bacik <josef@toxicpanda.com> Signed-off-by: Filipe Manana <fdmanana@suse.com> Signed-off-by: David Sterba <dsterba@suse.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In elf_read_build_id(), if gnu build_id is found, should return the size of
the actually copied data. If descsz is greater thanBuild_ID_SIZE,
write_buildid data access may occur.
Fixes: be96ea8ffa788dcc ("perf symbols: Fix issue with binaries using 16-bytes buildids (v2)") Reported-by: Will Ochowicz <Will.Ochowicz@genusplc.com> Signed-off-by: Yang Jihong <yangjihong1@huawei.com> Tested-by: Will Ochowicz <Will.Ochowicz@genusplc.com> Acked-by: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Leo Yan <leo.yan@linaro.org> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Stephane Eranian <eranian@google.com> Link: https://lore.kernel.org/lkml/CWLP265MB49702F7BA3D6D8F13E4B1A719C649@CWLP265MB4970.GBRP265.PROD.OUTLOOK.COM/T/ Link: https://lore.kernel.org/r/20230427012841.231729-1-yangjihong1@huawei.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Addresses of two data structure members were determined before
corresponding null pointer checks in the implementation of the function
“sort__sym_from_cmp”.
Thus avoid the risk for undefined behaviour by removing extra
initialisations for the local variables “from_l” and “from_r” (also
because they were already reassigned with the same value behind this
pointer check).
This issue was detected by using the Coccinelle software.
Fixes: 1b9e97a2a95e4941 ("perf tools: Fix report -F symbol_from for data without branch info") Signed-off-by: <elfring@users.sourceforge.net> Acked-by: Ian Rogers <irogers@google.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: German Gomez <german.gomez@arm.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Kan Liang <kan.liang@linux.intel.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Namhyung Kim <namhyung@kernel.org> Link: https://lore.kernel.org/cocci/54a21fea-64e3-de67-82ef-d61b90ffad05@web.de/ Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit 3c22ba5243040c13 ("perf vendor events powerpc: Update POWER9
events") added and updated power9 PMU JSON events. However some of the
JSON events which are part of other.json and pipeline.json files,
contains UTF-8 characters in their brief description. Having UTF-8
character could breaks the perf build on some distros.
Fix this issue by removing the UTF-8 characters from other.json and
pipeline.json files.
This patch adds a new case to handle the situation where the
device does not support any input methods in the
`snd_usb_caiaq_input_init` function. It returns an `-EINVAL` error code
to indicate that no input methods are supported on the device.
Fixes: 523f1dce3743 ("[ALSA] Add Native Instrument usb audio device support") Signed-off-by: Ruliang Lin <u202112092@hust.edu.cn> Reviewed-by: Dongliang Mu <dzm91@hust.edu.cn> Acked-by: Daniel Mack <daniel@zonque.org> Link: https://lore.kernel.org/r/20230504065054.3309-1-u202112092@hust.edu.cn Signed-off-by: Takashi Iwai <tiwai@suse.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Reported-by: syzbot <syzkaller@googlegroups.com> Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
There are cases where the device is adminstratively UP, but operationally
down. For example, we have a physical device (Nvidia ConnectX-6 Dx, 25Gbps)
who's cable was pulled out, here is its ip link output:
5: ens2f1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc mq state DOWN mode DEFAULT group default qlen 1000
link/ether b8:ce:f6:4b:68:35 brd ff:ff:ff:ff:ff:ff
altname enp179s0f1np1
As you can see, it's administratively UP but operationally down.
In this case, sending a packet to this port caused a nasty kernel hang (so
nasty that we were unable to capture it). Aborting a transmit based on
operational status (in addition to administrative status) fixes the issue.
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Acked-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: Victor Nogueira <victor@mojatatu.com>
v1->v2: Add fixes tag
v2->v3: Remove blank line between tags + add change log, suggested by Leon Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
When a tunnel device is bound with the underlying device, its
dev->needed_headroom needs to be updated properly. IPv4 tunnels
already do the same in ip_tunnel_bind_dev(). Otherwise we may
not have enough header room for skb, especially after commit b17f709a2401 ("gue: TX support for using remote checksum offload option").
Fixes: 32b8a8e59c9c ("sit: add IPv4 over IPv4 support") Reported-by: Palash Oswal <oswalpalash@gmail.com> Link: https://lore.kernel.org/netdev/CAGyP=7fDcSPKu6nttbGwt7RXzE3uyYxLjCSE97J64pRxJP8jPA@mail.gmail.com/ Cc: Kuniyuki Iwashima <kuniyu@amazon.com> Cc: Eric Dumazet <edumazet@google.com> Signed-off-by: Cong Wang <cong.wang@bytedance.com> Reviewed-by: Eric Dumazet <edumazet@google.com> Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
Toggle deleted anonymous sets as inactive in the next generation, so
users cannot perform any update on it. Clear the generation bitmask
in case the transaction is aborted.
The following KASAN splat shows a set element deletion for a bound
anonymous set that has been already removed in the same transaction.
Set deletion after flush coming in the same batch results in EBUSY. Add
set use counter to track the number of references to this set from
rules. We cannot rely on the list of bindings for this since such list
is still populated from the preparation phase.
Reported-by: Václav Zindulka <vaclav.zindulka@tlapnet.cz> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
If a rule that has already a bound anonymous set fails to be added, the
preparation phase releases the rule and the bound set. However, the
transaction object from the abort path still has a reference to the set
object that is stale, leading to a use-after-free when checking for the
set->bound field. Add a new field to the transaction that specifies if
the set is bound, so the abort path can skip releasing it since the rule
command owns it and it takes care of releasing it. After this update,
the set->bound field is removed.
Fixes: f6ac85858976 ("netfilter: nf_tables: unbind set in rule from commit path") Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Anonymous sets that are bound to rules from the same transaction trigger
a kernel splat from the abort path due to double set list removal and
double free.
This patch updates the logic to search for the transaction that is
responsible for creating the set and disable the set list removal and
release, given the rule is now responsible for this. Lookup is reverse
since the transaction that adds the set is likely to be at the tail of
the list.
Moreover, this patch adds the unbind step to deliver the event from the
commit path. This should not be done from the worker thread, since we
have no guarantees of in-order delivery to the listener.
This patch removes the assumption that both activate and deactivate
callbacks need to be provided.
Fixes: cd5125d8f518 ("netfilter: nf_tables: split set destruction in deactivate and destroy phase") Reported-by: Mikhail Morfikov <mmorfikov@gmail.com> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Fixes: 1b36c03e356936d6 ("perf record: Add support for using symbols in address filters") Signed-off-by: Adrian Hunter <adrian.hunter@intel.com> Cc: Adrian Hunter <adrian.hunter@intel.com> Cc: Ian Rogers <irogers@google.com> Cc: Jiri Olsa <jolsa@kernel.org> Cc: Namhyung Kim <namhyung@kernel.org> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/20230403154831.8651-2-adrian.hunter@intel.com Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In table_clear, it first acquires a write lock
https://elixir.bootlin.com/linux/v6.2/source/drivers/md/dm-ioctl.c#L1520
down_write(&_hash_lock);
Then before the lock is released at L1539, there is a path shown above:
table_clear -> __dev_status -> dm_get_inactive_table -> down_read
https://elixir.bootlin.com/linux/v6.2/source/drivers/md/dm-ioctl.c#L773
down_read(&_hash_lock);
It tries to acquire the same read lock again, resulting in the deadlock
problem.
Fix this by moving table_clear()'s __dev_status() call to after its
up_write(&_hash_lock);
The DASD driver does not kick the requeue list when requeuing IO requests
to the blocklayer. This might lead to hanging blockdevice when there is
no other trigger for this.
Fix by automatically kick the requeue list when requeuing DASD requests
to the blocklayer.
Since the introduction of scrub interface, the only flag that we support
is BTRFS_SCRUB_READONLY. Thus there is no sanity checks, if there are
some undefined flags passed in, we just ignore them.
This is problematic if we want to introduce new scrub flags, as we have
no way to determine if such flags are supported.
Address the problem by introducing a check for the flags, and if
unsupported flags are set, return -EOPNOTSUPP to inform the user space.
This check should be backported for all supported kernels before any new
scrub flags are introduced.
clk_cifout is derived from clk_cifout_src through an integer divider
limited to 32. clk_cifout_src is a child of either cpll, gpll or npll
without any possibility of a divider of any sort. The default clock
parent is cpll.
Let's allow clk_cifout to ask its parent clk_cifout_src to reparent in
order to find the real closest possible rate for clk_cifout and not one
derived from cpll only.
If the disk image that nilfs2 mounts is corrupted and a virtual block
address obtained by block lookup for a metadata file is invalid,
nilfs_bmap_lookup_at_level() may return the same internal return code as
-ENOENT, meaning the block does not exist in the metadata file.
This duplication of return codes confuses nilfs_mdt_get_block(), causing
it to read and create a metadata block indefinitely.
In particular, if this happens to the inode metadata file, ifile,
semaphore i_rwsem can be left held, causing task hangs in lock_mount.
Fix this issue by making nilfs_bmap_lookup_at_level() treat virtual block
address translation failures with -ENOENT as metadata corruption instead
of returning the error code.
According to syzbot's report, mark_buffer_dirty() called from
nilfs_segctor_do_construct() outputs a warning with some patterns after
nilfs2 detects metadata corruption and degrades to read-only mode.
After such read-only degeneration, page cache data may be cleared through
nilfs_clear_dirty_page() which may also clear the uptodate flag for their
buffer heads. However, even after the degeneration, log writes are still
performed by unmount processing etc., which causes mark_buffer_dirty() to
be called for buffer heads without the "uptodate" flag and causes the
warning.
Since any writes should not be done to a read-only file system in the
first place, this fixes the warning in mark_buffer_dirty() by letting
nilfs_segctor_do_construct() abort early if in read-only mode.
This also changes the retry check of nilfs_segctor_write_out() to avoid
unnecessary log write retries if it detects -EROFS that
nilfs_segctor_do_construct() returned.
Fix the argument pointer (ap) to point to real-mode memory
instead of virtual memory.
It's interesting that this issue hasn't shown up earlier, as this could
have happened with any 64-bit PDC ROM code.
I just noticed it because I suddenly faced a HPMC while trying to execute
the 64-bit STI ROM code of an Visualize-FXe graphics card for the STI
text console.
Do not global enable all the cyclic channels in at_xdmac_resume(). Instead
save the global status in at_xdmac_suspend() and re-enable the cyclic
channel only if it was active before suspend.
The tegra_xusb_port_unregister should be called when usb2_port
and ulpi_port map fails in tegra_xusb_add_usb2_port() or in
tegra_xusb_add_ulpi_port(), fix it.
Currently call_bind_status places a hard limit of 3 to the number of
retries on EACCES error. This limit was done to prevent NLM unlock
requests from being hang forever when the server keeps returning garbage.
However this change causes problem for cases when NLM service takes
longer than 9 seconds to register with the port mapper after a restart.
This patch removes this hard coded limit and let the RPC handles
the retry based on the standard hard/soft task semantics.
Fixes: 0b760113a3a1 ("NLM: Don't hang forever on NLM unlock requests") Reported-by: Helen Chao <helen.chao@oracle.com> Tested-by: Helen Chao <helen.chao@oracle.com> Signed-off-by: Dai Ngo <dai.ngo@oracle.com> Reviewed-by: Jeff Layton <jlayton@kernel.org> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com> Signed-off-by: Sasha Levin <sashal@kernel.org>
hfi1_mmu_rb_remove_unless_exact() did not move mmu_rb_node objects in
mmu_rb_handler->lru_list after getting a cache hit on an mmu_rb_node.
As a result, hfi1_mmu_rb_evict() was not guaranteed to evict truly
least-recently used nodes.
This could be a performance issue for an application when that
application:
- Uses some long-lived buffers frequently.
- Uses a large number of buffers once.
- Hits the mmu_rb_handler cache size or pinned-page limits, forcing
mmu_rb_handler cache entries to be evicted.
In this case, the one-time use buffers cause the long-lived buffer
entries to eventually filter to the end of the LRU list where
hfi1_mmu_rb_evict() will consider evicting a frequently-used long-lived
entry instead of evicting one of the one-time use entries.
Fix this by inserting new mmu_rb_node at the tail of
mmu_rb_handler->lru_list and move mmu_rb_ndoe to the tail of
mmu_rb_handler->lru_list when the mmu_rb_node is a hit in
hfi1_mmu_rb_remove_unless_exact(). Change hfi1_mmu_rb_evict() to evict
from the head of mmu_rb_handler->lru_list instead of the tail.
Fixes: 0636e9ab8355 ("IB/hfi1: Add cache evict LRU list") Signed-off-by: Brendan Cunningham <bcunningham@cornelisnetworks.com> Signed-off-by: Patrick Kelsey <pat.kelsey@cornelisnetworks.com> Signed-off-by: Dennis Dalessandro <dennis.dalessandro@cornelisnetworks.com> Link: https://lore.kernel.org/r/168088635931.3027109.10423156330761536044.stgit@252.162.96.66.static.eigbox.net Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When returning from of_parse_phandle_with_args(), the np member of the
of_phandle_args structure should be put after usage. Add missing
of_node_put() calls in both __set_clk_parents() and __set_clk_rates().
Fixes: 86be408bfbd8 ("clk: Support for clock parents and rates assigned from device tree") Signed-off-by: Clément Léger <clement.leger@bootlin.com> Link: https://lore.kernel.org/r/20230131083227.10990-1-clement.leger@bootlin.com Signed-off-by: Stephen Boyd <sboyd@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
There is no need to check 'rdi->qp_dev' for NULL. The field 'qp_dev'
is created in rvt_register_device() which will fail if the 'qp_dev'
allocation fails in rvt_driver_qp_init(). Overwise this pointer
doesn't changed and passed to rvt_qp_exit() by the next step.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Fixes: 0acb0cc7ecc1 ("IB/rdmavt: Initialize and teardown of qpn table") Signed-off-by: Natalia Petrova <n.petrova@fintech.ru> Link: https://lore.kernel.org/r/20230303124408.16685-1-n.petrova@fintech.ru Signed-off-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
commit e050e3f0a71bf ("perf: Fix broken interrupt rate throttling")
introduces a change in throttling threshold judgment. Before this,
compare hwc->interrupts and max_samples_per_tick, then increase
hwc->interrupts by 1, but this commit reverses order of these two
behaviors, causing the semantics of max_samples_per_tick to change.
In literal sense of "max_samples_per_tick", if hwc->interrupts ==
max_samples_per_tick, it should not be throttled, therefore, the judgment
condition should be changed to "hwc->interrupts > max_samples_per_tick".
In fact, this may cause the hardlockup to fail, The minimum value of
max_samples_per_tick may be 1, in this case, the return value of
__perf_event_account_interrupt function is 1.
As a result, nmi_watchdog gets throttled, which would stop PMU (Use x86
architecture as an example, see x86_pmu_handle_irq).
Using memcpy() isn't safe when buf is identical to rtas_err_buf, which
can happen during boot before slab is up. Full context which may not
be obvious from the diff:
if (altbuf) {
buf = altbuf;
} else {
buf = rtas_err_buf;
if (slab_is_available())
buf = kmalloc(RTAS_ERROR_LOG_MAX, GFP_ATOMIC);
}
if (buf)
memcpy(buf, rtas_err_buf, RTAS_ERROR_LOG_MAX);
This was found by inspection and I'm not aware of it causing problems
in practice. It appears to have been introduced by commit 033ef338b6e0 ("powerpc: Merge rtas.c into arch/powerpc/kernel"); the
old ppc64 version of this code did not have this problem.
Use memmove() instead.
Fixes: 033ef338b6e0 ("powerpc: Merge rtas.c into arch/powerpc/kernel") Signed-off-by: Nathan Lynch <nathanl@linux.ibm.com> Reviewed-by: Andrew Donnellan <ajd@linux.ibm.com> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20230220-rtas-queue-for-6-4-v1-2-010e4416f13f@linux.ibm.com Signed-off-by: Sasha Levin <sashal@kernel.org>
LEDS_TRIGGER_DISK depends on ATA, so selecting LEDS_TRIGGER_DISK
when ATA is not set/enabled causes a Kconfig warning:
WARNING: unmet direct dependencies detected for LEDS_TRIGGER_DISK
Depends on [n]: NEW_LEDS [=y] && LEDS_TRIGGERS [=y] && ATA [=n]
Selected by [y]:
- ADB_PMU_LED_DISK [=y] && MACINTOSH_DRIVERS [=y] && ADB_PMU_LED [=y] && LEDS_CLASS [=y]
Fix this by making ADB_PMU_LED_DISK depend on ATA.
Seen on both PPC32 and PPC64.
Fixes: 0e865a80c135 ("macintosh: Remove dependency on IDE_GD_ATA if ADB_PMU_LED_DISK is selected") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20230223014241.20878-1-rdunlap@infradead.org Signed-off-by: Sasha Levin <sashal@kernel.org>
Use "%pa" format specifier for resource_size_t to avoid a compiler
printk format warning.
arch/powerpc/sysdev/tsi108_pci.c: In function 'tsi108_setup_pci':
include/linux/kern_levels.h:5:25: error: format '%x' expects argument of type 'unsigned int', but argument 2 has type 'resource_size_t'
Fixes: c4342ff92bed ("[POWERPC] Update mpc7448hpc2 board irq support using device tree") Fixes: 2b9d7467a6db ("[POWERPC] Add tsi108 pci and platform device data register function") Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
[mpe: Use pr_info() and unsplit string] Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20230223070116.660-5-rdunlap@infradead.org Signed-off-by: Sasha Levin <sashal@kernel.org>
Use "%pa" format specifier for resource_size_t to avoid compiler
printk format warnings.
../arch/powerpc/platforms/embedded6xx/flipper-pic.c: In function 'flipper_pic_init':
../include/linux/kern_levels.h:5:25: error: format '%x' expects argument of type 'unsigned int', but argument 2 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
../arch/powerpc/platforms/embedded6xx/flipper-pic.c:148:9: note: in expansion of macro 'pr_info'
148 | pr_info("controller at 0x%08x mapped to 0x%p\n", res.start, io_base);
| ^~~~~~~
../arch/powerpc/platforms/embedded6xx/hlwd-pic.c: In function 'hlwd_pic_init':
../include/linux/kern_levels.h:5:25: error: format '%x' expects argument of type 'unsigned int', but argument 2 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
../arch/powerpc/platforms/embedded6xx/hlwd-pic.c:174:9: note: in expansion of macro 'pr_info'
174 | pr_info("controller at 0x%08x mapped to 0x%p\n", res.start, io_base);
| ^~~~~~~
../arch/powerpc/platforms/embedded6xx/wii.c: In function 'wii_ioremap_hw_regs':
../include/linux/kern_levels.h:5:25: error: format '%x' expects argument of type 'unsigned int', but argument 3 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
../arch/powerpc/platforms/embedded6xx/wii.c:77:17: note: in expansion of macro 'pr_info'
77 | pr_info("%s at 0x%08x mapped to 0x%p\n", name,
| ^~~~~~~
Use "%pa" format specifier for resource_size_t to avoid a compiler
printk format warning.
../arch/powerpc/platforms/512x/clock-commonclk.c: In function 'mpc5121_clk_provide_backwards_compat':
../arch/powerpc/platforms/512x/clock-commonclk.c:989:44: error: format '%x' expects argument of type 'unsigned int', but argument 4 has type 'resource_size_t' {aka 'long long unsigned int'} [-Werror=format=]
989 | snprintf(devname, sizeof(devname), "%08x.%s", res.start, np->name); \
| ^~~~~~~~~ ~~~~~~~~~
| |
| resource_size_t {aka long long unsigned int}
Prevents 24 such warnings.
Fixes: 01f25c371658 ("clk: mpc512x: add backwards compat to the CCF code") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Signed-off-by: Michael Ellerman <mpe@ellerman.id.au> Link: https://msgid.link/20230223070116.660-2-rdunlap@infradead.org Signed-off-by: Sasha Levin <sashal@kernel.org>
When removing a SPMI driver, there can be a crash due to NULL pointer
dereference if it does not have a remove callback defined. This is
one such call trace observed when removing the QCOM SPMI PMIC driver:
If a driver has all its resources allocated through devm_() APIs and
does not need any other explicit cleanup, it would not require a
remove callback to be defined. Hence, add a check for remove callback
presence before calling it when removing a SPMI driver.
When loading the driver for rtl8192e, the W_DISABLE# switch is working as
intended. But when the WLAN is turned off in software and then turned on
again the W_DISABLE# does not work anymore. Reason for this is that in
the function _rtl92e_dm_check_rf_ctrl_gpio() the bfirst_after_down is
checked and returned when true. bfirst_after_down is set true when
switching the WLAN off in software. But it is not set to false again
when WLAN is turned on again.
Add bfirst_after_down = false in _rtl92e_sta_up to reset bit and fix
above described bug.
Fixes: 94a799425eee ("From: wlanfae <wlanfae@realtek.com> [PATCH 1/8] rtl8192e: Import new version of driver from realtek") Signed-off-by: Philipp Hortmann <philipp.g.hortmann@gmail.com> Link: https://lore.kernel.org/r/20230418200201.GA17398@matrix-ESPRIMO-P710 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The Store Queue code allocates a bitmap buffer with the size of
multiple of sizeof(long) in sq_api_init(). While the buffer size
is calculated correctly, the code uses the wrong element size to
allocate the buffer which results in the allocated bitmap buffer
being too small.
Fix this by allocating the buffer with kcalloc() with element size
sizeof(long) instead of kzalloc() whose elements size defaults to
sizeof(char).
Fixes: d7c30c682a27 ("sh: Store Queue API rework.") Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de> Link: https://lore.kernel.org/r/20230419114854.528677-1-glaubitz@physik.fu-berlin.de Signed-off-by: Sasha Levin <sashal@kernel.org>
typeof is (still) a GNU extension, which means that it cannot be used when
building ISO C (e.g. -std=c99). It should therefore be avoided in uapi
headers in favour of the ISO-friendly __typeof__.
Unfortunately this issue could not be detected by
CONFIG_UAPI_HEADER_TEST=y as the __ALIGN_KERNEL() macro is not expanded in
any uapi header.
This matters from a userspace perspective, not a kernel one. uapi
headers and their contents are expected to be usable in a variety of
situations, and in particular when building ISO C applications (with
-std=c99 or similar).
This particular problem can be reproduced by trying to use the
__ALIGN_KERNEL macro directly in application code, say:
#include <linux/const.h>
int align(int x, int a)
{
return __KERNEL_ALIGN(x, a);
}
and trying to build that with -std=c99.
Link: https://lkml.kernel.org/r/20230411092747.3759032-1-kevin.brodsky@arm.com Fixes: a79ff731a1b2 ("netfilter: xtables: make XT_ALIGN() usable in exported headers by exporting __ALIGN_KERNEL()") Signed-off-by: Kevin Brodsky <kevin.brodsky@arm.com> Reported-by: Ruben Ayrapetyan <ruben.ayrapetyan@arm.com> Tested-by: Ruben Ayrapetyan <ruben.ayrapetyan@arm.com> Reviewed-by: Petr Vorel <pvorel@suse.cz> Tested-by: Petr Vorel <pvorel@suse.cz> Reviewed-by: Masahiro Yamada <masahiroy@kernel.org> Cc: Sam Ravnborg <sam@ravnborg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
alloc_per_cpu_data() is called by find_memory(), which is marked as
__init. Therefore alloc_per_cpu_data() can also be marked as __init to
remedy this modpost problem.
Link: https://lkml.kernel.org/r/20230223034258.12917-1-rdunlap@infradead.org Fixes: 4b9ddc7cf272 ("[IA64] Fix section mismatch in contig.c version of per_cpu_init()") Signed-off-by: Randy Dunlap <rdunlap@infradead.org> Cc: Christoph Hellwig <hch@lst.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The helper generating an OF based modalias (of_device_get_modalias())
works fine, but due to the use of snprintf() internally it needs a
buffer one byte longer than what should be needed just for the entire
string (excluding the '\0'). Most users of this helper are sysfs hooks
providing the modalias string to users. They all provide a PAGE_SIZE
buffer which is way above the number of bytes required to fit the
modalias string and hence do not suffer from this issue.
There is another user though, of_device_request_module(), which is only
called by drivers/usb/common/ulpi.c. This request module function is
faulty, but maybe because in most cases there is an alternative, ULPI
driver users have not noticed it.
In this function, of_device_get_modalias() is called twice. The first
time without buffer just to get the number of bytes required by the
modalias string (excluding the null byte), and a second time, after
buffer allocation, to fill the buffer. The allocation asks for an
additional byte, in order to store the trailing '\0'. However, the
buffer *length* provided to of_device_get_modalias() excludes this extra
byte. The internal use of snprintf() with a length that is exactly the
number of bytes to be written has the effect of using the last available
byte to store a '\0', which then smashes the last character of the
modalias string.
Provide the actual size of the buffer to of_device_get_modalias() to fix
this issue.
Note: the "str[size - 1] = '\0';" line is not really needed as snprintf
will anyway end the string with a null byte, but there is a possibility
that this function might be called on a struct device_node without
compatible, in this case snprintf() would not be executed. So we keep it
just to avoid possible unbounded strings.
Cc: Stephen Boyd <sboyd@kernel.org> Cc: Peter Chen <peter.chen@kernel.org> Fixes: 9c829c097f2f ("of: device: Support loading a module with OF based modalias") Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com> Reviewed-by: Rob Herring <robh@kernel.org> Signed-off-by: Srinivas Kandagatla <srinivas.kandagatla@linaro.org> Link: https://lore.kernel.org/r/20230404172148.82422-2-srinivas.kandagatla@linaro.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
CPM has the same problem as QE so for CPM also use the fix added
by commit 0398fb70940e ("spi/spi_mpc8xxx: Fix QE mode Litte Endian"):
CPM mode uses Little Endian so words > 8 bits are byte swapped.
Workaround this by always enforcing wordsize 8 for 16 and 32 bits
words. Unfortunately this will not work for LSB transfers
where wordsize is > 8 bits so disable these for now.
Also limit the workaround to 16 and 32 bits words because it can
only work for multiples of 8-bits.
So while priority inversion on the pmsg_lock is an occasional
problem that an rt_mutex would help with, in uses where logging
is writing to pmsg heavily from multiple threads, the pmsg_lock
can be heavily contended.
After this change landed, it was reported that cases where the
mutex locking overhead was commonly adding on the order of 10s
of usecs delay had suddenly jumped to ~msec delay with rtmutex.
It seems the slight differences in the locks under this level
of contention causes the normal mutexes to utilize the spinning
optimizations, while the rtmutexes end up in the sleeping
slowpath (which allows additional threads to pile on trying
to take the lock).
In this case, it devolves to a worse case senerio where the lock
acquisition and scheduling overhead dominates, and each thread
is waiting on the order of ~ms to do ~us of work.
Obviously, having tons of threads all contending on a single
lock for logging is non-optimal, so the proper fix is probably
reworking pstore pmsg to have per-cpu buffers so we don't have
contention.
Additionally, Steven Rostedt has provided some furhter
optimizations for rtmutexes that improves the rtmutex spinning
path, but at least in my testing, I still see the test tripping
into the sleeping path on rtmutexes while utilizing the spinning
path with mutexes.
But in the short term, lets revert the change to the rt_mutex
and go back to normal mutexes to avoid a potentially major
performance regression. And we can work on optimizations to both
rtmutexes and finer-grained locking for pstore pmsg in the
future.
Cc: Wei Wang <wvw@google.com> Cc: Midas Chien<midaschieh@google.com> Cc: "Chunhui Li (李春辉)" <chunhui.li@mediatek.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Kees Cook <keescook@chromium.org> Cc: Anton Vorontsov <anton@enomsg.org> Cc: "Guilherme G. Piccoli" <gpiccoli@igalia.com> Cc: Tony Luck <tony.luck@intel.com> Cc: kernel-team@android.com Fixes: 76d62f24db07 ("pstore: Switch pmsg_lock to an rt_mutex to avoid priority inversion") Reported-by: "Chunhui Li (李春辉)" <chunhui.li@mediatek.com> Signed-off-by: John Stultz <jstultz@google.com> Signed-off-by: Kees Cook <keescook@chromium.org> Link: https://lore.kernel.org/r/20230308204043.2061631-1-jstultz@google.com Signed-off-by: Sasha Levin <sashal@kernel.org>
sendmsg() calls msg_zerocopy_alloc(), which allocates a skb, sets
skb->cb->ubuf.refcnt to 1, and calls sock_hold(). Here, struct
ubuf_info_msgzc indirectly holds a refcnt of the socket. When the
skb is sent, __skb_tstamp_tx() clones it and puts the clone into
the socket's error queue with the TX timestamp.
When the original skb is received locally, skb_copy_ubufs() calls
skb_unclone(), and pskb_expand_head() increments skb->cb->ubuf.refcnt.
This additional count is decremented while freeing the skb, but struct
ubuf_info_msgzc still has a refcnt, so __msg_zerocopy_callback() is
not called.
The last refcnt is not released unless we retrieve the TX timestamped
skb by recvmsg(). Since we clear the error queue in inet_sock_destruct()
after the socket's refcnt reaches 0, there is a circular dependency.
If we close() the socket holding such skbs, we never call sock_put()
and leak the count, sk, and skb.
TCP has the same problem, and commit e0c8bccd40fc ("net: stream:
purge sk_error_queue in sk_stream_kill_queues()") tried to fix it
by calling skb_queue_purge() during close(). However, there is a
small chance that skb queued in a qdisc or device could be put
into the error queue after the skb_queue_purge() call.
In __skb_tstamp_tx(), the cloned skb should not have a reference
to the ubuf to remove the circular dependency, but skb_clone() does
not call skb_copy_ubufs() for zerocopy skb. So, we need to call
skb_orphan_frags_rx() for the cloned skb to call skb_copy_ubufs().