Similar to the previous commit, the net_helper.sh file from the parent
directory is used by the MPTCP selftests and it needs to be present when
running the tests.
This file then needs to be listed in the Makefile to be included when
exporting or installing the tests, e.g. with:
make -C tools/testing/selftests \
TARGETS=net/mptcp \
install INSTALL_PATH=$KSFT_INSTALL_PATH
cd $KSFT_INSTALL_PATH
./run_kselftest.sh -c net/mptcp
selftests: mptcp: join: restrict fullmesh endp on 1st sf
A new endpoint using the IP of the initial subflow has been recently
added to increase the code coverage. But it breaks the test when using
old kernels not having commit 86e39e04482b ("mptcp: keep track of local
endpoint still available for each msk"), e.g. on v5.15.
Similar to commit d4c81bbb8600 ("selftests: mptcp: join: support local
endpoint being tracked or not"), it is possible to add the new endpoint
conditionally, by checking if "mptcp_pm_subflow_check_next" is present
in kallsyms: this is not directly linked to the commit introducing this
symbol but for the parent one which is linked anyway. So we can know in
advance what will be the expected behaviour, and add the new endpoint
only when it makes sense to do so.
Jakub Kicinski [Wed, 11 Sep 2024 03:15:10 +0000 (20:15 -0700)]
Merge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2024-09-09 (ice, igb)
This series contains updates to ice and igb drivers.
Martyna moves LLDP rule removal to the proper uninitialization function
for ice.
Jake corrects accounting logic for FWD_TO_VSI_LIST switch filters on
ice.
Przemek removes incorrect, explicit calls to pci_disable_device() for
ice.
Michal Schmidt stops incorrect use of VSI list for VLAN use on ice.
Sriram Yagnaraman adjusts igb_xdp_ring_update_tail() to be called under
Tx lock on igb.
* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
igb: Always call igb_xdp_ring_update_tail() under Tx lock
ice: fix VSI lists confusion when adding VLANs
ice: stop calling pci_disable_device() as we use pcim
ice: fix accounting for filters shared by multiple VSIs
ice: Fix lldp packets dropping after changing the number of channels
====================
Jakub Kicinski [Wed, 11 Sep 2024 03:11:40 +0000 (20:11 -0700)]
Merge tag 'mlx5-fixes-2024-09-09' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5 fixes 2024-09-09
This series provides bug fixes to mlx5 driver.
* tag 'mlx5-fixes-2024-09-09' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
net/mlx5: Fix bridge mode operations when there are no VFs
net/mlx5: Verify support for scheduling element and TSAR type
net/mlx5: Add missing masks and QoS bit masks for scheduling elements
net/mlx5: Explicitly set scheduling element and TSAR type
net/mlx5e: Add missing link mode to ptys2ext_ethtool_map
net/mlx5e: Add missing link modes to ptys2ethtool_map
net/mlx5: Update the list of the PCI supported devices
====================
As Rob pointed in another mail thread [1], the binding of tja11xx PHY
is completely broken, the schema cannot catch the error in the DTS. A
compatiable string must be needed if we want to add a custom propety.
So extract known PHY IDs from the tja11xx PHY drivers and convert them
into supported compatible string list to fix the broken binding issue.
Sean Anderson [Fri, 6 Sep 2024 21:07:43 +0000 (17:07 -0400)]
selftests: net: csum: Fix checksums for packets with non-zero padding
Padding is not included in UDP and TCP checksums. Therefore, reduce the
length of the checksummed data to include only the data in the IP
payload. This fixes spurious reported checksum failures like
rx: pkt: sport=33000 len=26 csum=0xc850 verify=0xf9fe
pkt: bad csum
Technically it is possible for there to be trailing bytes after the UDP
data but before the Ethernet padding (e.g. if sizeof(ip) + sizeof(udp) +
udp.len < ip.len). However, we don't generate such packets.
Tomas Paukrt [Fri, 6 Sep 2024 10:52:40 +0000 (12:52 +0200)]
net: phy: dp83822: Fix NULL pointer dereference on DP83825 devices
The probe() function is only used for DP83822 and DP83826 PHY,
leaving the private data pointer uninitialized for the DP83825 models
which causes a NULL pointer dereference in the recently introduced/changed
functions dp8382x_config_init() and dp83822_set_wol().
Add the dp8382x_probe() function, so all PHY models will have a valid
private data pointer to fix this issue and also prevent similar issues
in the future.
net: ftgmac100: Enable TX interrupt to avoid TX timeout
Currently, the driver only enables RX interrupt to handle RX
packets and TX resources. Sometimes there is not RX traffic,
so the TX resource needs to wait for RX interrupt to free.
This situation will toggle the TX timeout watchdog when the MAC
TX ring has no more resources to transmit packets.
Therefore, enable TX interrupt to release TX resources at any time.
When I am verifying iperf3 over UDP, the network hangs.
Like the log below.
The network topology is FTGMAC connects directly to a PC.
UDP does not need to wait for ACK, unlike TCP.
Therefore, FTGMAC needs to enable TX interrupt to release TX resources instead
of waiting for the RX interrupt.
octeontx2-af: Modify SMQ flush sequence to drop packets
The current implementation of SMQ flush sequence waits for the packets
in the TM pipeline to be transmitted out of the link. This sequence
doesn't succeed in HW when there is any issue with link such as lack of
link credits, link down or any other traffic that is fully occupying the
link bandwidth (QoS). This patch modifies the SMQ flush sequence to
drop the packets after TL1 level (SQM) instead of polling for the packets
to be sent out of RPM/CGX link.
The bridge mode is only relevant when there are multiple functions per
port. Therefore, prevent setting and getting this setting when there are no
VFs.
Note that after this change, there are no settings to change on the PF
interface using `bridge link` when there are no VFs, so the interface no
longer appears in the `bridge link` output.
Fixes: 4b89251de024 ("net/mlx5: Support ndo bridge_setlink and getlink") Signed-off-by: Benjamin Poirier <[email protected]> Reviewed-by: Cosmin Ratiu <[email protected]> Signed-off-by: Saeed Mahameed <[email protected]>
Carolina Jubran [Mon, 5 Aug 2024 10:13:03 +0000 (13:13 +0300)]
net/mlx5: Verify support for scheduling element and TSAR type
Before creating a scheduling element in a NIC or E-Switch scheduler,
ensure that the requested element type is supported. If the element is
of type Transmit Scheduling Arbiter (TSAR), also verify that the
specific TSAR type is supported.
Fixes: 214baf22870c ("net/mlx5e: Support HTB offload") Fixes: 85c5f7c9200e ("net/mlx5: E-switch, Create QoS on demand") Fixes: 0fe132eac38c ("net/mlx5: E-switch, Allow to add vports to rate groups") Signed-off-by: Carolina Jubran <[email protected]> Reviewed-by: Cosmin Ratiu <[email protected]> Signed-off-by: Saeed Mahameed <[email protected]>
igb: Always call igb_xdp_ring_update_tail() under Tx lock
Always call igb_xdp_ring_update_tail() under __netif_tx_lock, add a comment
and lockdep assert to indicate that. This is needed to share the same TX
ring between XDP, XSK and slow paths. Furthermore, the current XDP
implementation is racy on tail updates.
Michal Schmidt [Wed, 4 Sep 2024 09:39:22 +0000 (11:39 +0200)]
ice: fix VSI lists confusion when adding VLANs
The description of function ice_find_vsi_list_entry says:
Search VSI list map with VSI count 1
However, since the blamed commit (see Fixes below), the function no
longer checks vsi_count. This causes a problem in ice_add_vlan_internal,
where the decision to share VSI lists between filter rules relies on the
vsi_count of the found existing VSI list being 1.
The reproducing steps:
1. Have a PF and two VFs.
There will be a filter rule for VLAN 0, referring to a VSI list
containing VSIs: 0 (PF), 2 (VF#0), 3 (VF#1).
2. Add VLAN 1234 to VF#0.
ice will make the wrong decision to share the VSI list with the new
rule. The wrong behavior may not be immediately apparent, but it can
be observed with debug prints.
3. Add VLAN 1234 to VF#1.
ice will unshare the VSI list for the VLAN 1234 rule. Due to the
earlier bad decision, the newly created VSI list will contain
VSIs 0 (PF) and 3 (VF#1), instead of expected 2 (VF#0) and 3 (VF#1).
4. Try pinging a network peer over the VLAN interface on VF#0.
This fails.
Reproducer script at:
https://gitlab.com/mschmidt2/repro/-/blob/master/RHEL-46814/test-vlan-vsi-list-confusion.sh
Commented debug trace:
https://gitlab.com/mschmidt2/repro/-/blob/master/RHEL-46814/ice-vlan-vsi-lists-debug.txt
Patch adding the debug prints:
https://gitlab.com/mschmidt2/linux/-/commit/f8a8814623944a45091a77c6094c40bfe726bfdb
(Unsafe, by the way. Lacks rule_lock when dumping in ice_remove_vlan.)
Michal Swiatkowski added to the explanation that the bug is caused by
reusing a VSI list created for VLAN 0. All created VFs' VSIs are added
to VLAN 0 filter. When a non-zero VLAN is created on a VF which is already
in VLAN 0 (normal case), the VSI list from VLAN 0 is reused.
It leads to a problem because all VFs (VSIs to be specific) that are
subscribed to VLAN 0 will now receive a new VLAN tag traffic. This is
one bug, another is the bug described above. Removing filters from
one VF will remove VLAN filter from the previous VF. It happens a VF is
reset. Example:
- creation of 3 VFs
- we have VSI list (used for VLAN 0) [0 (pf), 2 (vf1), 3 (vf2), 4 (vf3)]
- we are adding VLAN 100 on VF1, we are reusing the previous list
because 2 is there
- VLAN traffic works fine, but VLAN 100 tagged traffic can be received
on all VSIs from the list (for example broadcast or unicast)
- trust is turning on VF2, VF2 is resetting, all filters from VF2 are
removed; the VLAN 100 filter is also removed because 3 is on the list
- VLAN traffic to VF1 isn't working anymore, there is a need to recreate
VLAN interface to readd VLAN filter
One thing I'm not certain about is the implications for the LAG feature,
which is another caller of ice_find_vsi_list_entry. I don't have a
LAG-capable card at hand to test.
Przemek Kitszel [Fri, 30 Aug 2024 13:44:11 +0000 (15:44 +0200)]
ice: stop calling pci_disable_device() as we use pcim
Our driver uses devres to manage resources, in particular we call
pcim_enable_device(), what also means we express the intent to get
automatic pci_disable_device() call at driver removal. Manual calls to
pci_disable_device() misuse the API.
Recent commit (see "Fixes" tag) has changed the removal action from
conditional (silent ignore of double call to pci_disable_device()) to
unconditional, but able to catch unwanted redundant calls; see cited
"Fixes" commit for details.
Since that, unloading the driver yields following warn+splat:
Jacob Keller [Wed, 31 Jul 2024 16:55:55 +0000 (09:55 -0700)]
ice: fix accounting for filters shared by multiple VSIs
When adding a switch filter (such as a MAC or VLAN filter), it is expected
that the driver will detect the case where the filter already exists, and
return -EEXIST. This is used by calling code such as ice_vc_add_mac_addr,
and ice_vsi_add_vlan to avoid incrementing the accounting fields such as
vsi->num_vlan or vf->num_mac.
This logic works correctly for the case where only a single VSI has added a
given switch filter.
When a second VSI adds the same switch filter, the driver converts the
existing filter from an ICE_FWD_TO_VSI filter into an ICE_FWD_TO_VSI_LIST
filter. This saves switch resources, by ensuring that multiple VSIs can
re-use the same filter.
The ice_add_update_vsi_list() function is responsible for doing this
conversion. When first converting a filter from the FWD_TO_VSI into
FWD_TO_VSI_LIST, it checks if the VSI being added is the same as the
existing rule's VSI. In such a case it returns -EEXIST.
However, when the switch rule has already been converted to a
FWD_TO_VSI_LIST, the logic is different. Adding a new VSI in this case just
requires extending the VSI list entry. The logic for checking if the rule
already exists in this case returns 0 instead of -EEXIST.
This breaks the accounting logic mentioned above, so the counters for how
many MAC and VLAN filters exist for a given VF or VSI no longer accurately
reflect the actual count. This breaks other code which relies on these
counts.
In typical usage this primarily affects such filters generally shared by
multiple VSIs such as VLAN 0, or broadcast and multicast MAC addresses.
Fix this by correctly reporting -EEXIST in the case of adding the same VSI
to a switch rule already converted to ICE_FWD_TO_VSI_LIST.
Fixes: 9daf8208dd4d ("ice: Add support for switch filter programming") Signed-off-by: Jacob Keller <[email protected]> Tested-by: Rafal Romanowski <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
ice: Fix lldp packets dropping after changing the number of channels
After vsi setup refactor commit 6624e780a577 ("ice: split ice_vsi_setup
into smaller functions") ice_cfg_sw_lldp function which removes rx rule
directing LLDP packets to vsi is moved from ice_vsi_release to
ice_vsi_decfg function. ice_vsi_decfg is used in more cases than just in
vsi_release resulting in unnecessary removal of rx lldp packets handling
switch rule. This leads to lldp packets being dropped after a change number
of channels via ethtool.
This patch moves ice_cfg_sw_lldp function that removes rx lldp sw rule back
to ice_vsi_release function.
Eric Dumazet [Wed, 4 Sep 2024 13:37:25 +0000 (13:37 +0000)]
net: hsr: remove seqnr_lock
syzbot found a new splat [1].
Instead of adding yet another spin_lock_bh(&hsr->seqnr_lock) /
spin_unlock_bh(&hsr->seqnr_lock) pair, remove seqnr_lock
and use atomic_t for hsr->sequence_nr and hsr->sup_sequence_nr.
This also avoid a race in hsr_fill_info().
Also remove interlink_sequence_nr which is unused.
Fixes: 06afd2c31d33 ("hsr: Synchronize sending frames to have always incremented outgoing seq nr.") Fixes: f421436a591d ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)") Reported-by: syzbot <[email protected]> Signed-off-by: Eric Dumazet <[email protected]> Cc: Sebastian Andrzej Siewior <[email protected]> Reviewed-by: Simon Horman <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Merge tag 'net-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Pull networking fixes from Jakub Kicinski:
"Including fixes from can, bluetooth and wireless.
No known regressions at this point. Another calm week, but chances are
that has more to do with vacation season than the quality of our work.
Current release - new code bugs:
- smc: prevent NULL pointer dereference in txopt_get
- eth: ti: am65-cpsw: number of XDP-related fixes
Previous releases - regressions:
- Revert "Bluetooth: MGMT/SMP: Fix address type when using SMP over
BREDR/LE", it breaks existing user space
- Bluetooth: qca: if memdump doesn't work, re-enable IBS to avoid
later problems with suspend
- can: mcp251x: fix deadlock if an interrupt occurs during
mcp251x_open
- eth: r8152: fix the firmware communication error due to use of bulk
write
- ptp: ocp: fix serial port information export
- eth: igb: fix not clearing TimeSync interrupts for 82580
- Revert "wifi: ath11k: support hibernation", fix suspend on Lenovo
Previous releases - always broken:
- eth: intel: fix crashes and bugs when reconfiguration and resets
happening in parallel
- wifi: ath11k: fix NULL dereference in ath11k_mac_get_eirp_power()
Misc:
- docs: netdev: document guidance on cleanup.h"
* tag 'net-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (61 commits)
ila: call nf_unregister_net_hooks() sooner
tools/net/ynl: fix cli.py --subscribe feature
MAINTAINERS: fix ptp ocp driver maintainers address
selftests: net: enable bind tests
net: dsa: vsc73xx: fix possible subblocks range of CAPT block
sched: sch_cake: fix bulk flow accounting logic for host fairness
docs: netdev: document guidance on cleanup.h
net: xilinx: axienet: Fix race in axienet_stop
net: bridge: br_fdb_external_learn_add(): always set EXT_LEARN
r8152: fix the firmware doesn't work
fou: Fix null-ptr-deref in GRO.
bareudp: Fix device stats updates.
net: mana: Fix error handling in mana_create_txq/rxq's NAPI cleanup
bpf, net: Fix a potential race in do_sock_getsockopt()
net: dqs: Do not use extern for unused dql_group
sch/netem: fix use after free in netem_dequeue
usbnet: modern method to get random MAC
MAINTAINERS: wifi: cw1200: add net-cw1200.h
ice: do not bring the VSI up, if it was down before the XDP setup
ice: remove ICE_CFG_BUSY locking from AF_XDP code
...
Merge tag 'spi-fix-v6.11-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi
Pull spi fixes from Mark Brown:
"A few small driver specific fixes (including some of the widespread
work on fixing missing ID tables for module autoloading and the revert
of some problematic PM work in spi-rockchip), some improvements to the
MAINTAINERS information for the NXP drivers and the addition of a new
device ID to spidev"
* tag 'spi-fix-v6.11-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
MAINTAINERS: SPI: Add mailing list [email protected] for nxp spi drivers
MAINTAINERS: SPI: Add freescale lpspi maintainer information
spi: spi-fsl-lpspi: Fix off-by-one in prescale max
spi: spidev: Add missing spi_device_id for jg10309-01
spi: bcm63xx: Enable module autoloading
spi: intel: Add check devm_kasprintf() returned value
spi: spidev: Add an entry for elgin,jg10309-01
spi: rockchip: Resolve unbalanced runtime PM / system PM handling
Merge tag 'regulator-fix-v6.11-stub' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator
Pull regulator fix from Mark Brown:
"A fix from Doug Anderson for a missing stub, required to fix the build
for some newly added users of devm_regulator_bulk_get_const() in
!REGULATOR configurations"
* tag 'regulator-fix-v6.11-stub' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
regulator: core: Stub devm_regulator_bulk_get_const() if !CONFIG_REGULATOR
Merge tag 'rust-fixes-6.11-2' of https://github.com/Rust-for-Linux/linux
Pull Rust fixes from Miguel Ojeda:
"Toolchain and infrastructure:
- Fix builds for nightly compiler users now that 'new_uninit' was
split into new features by using an alternative approach for the
code that used what is now called the 'box_uninit_write' feature
- Allow the 'stable_features' lint to preempt upcoming warnings about
them, since soon there will be unstable features that will become
stable in nightly compilers
- Export bss symbols too
'kernel' crate:
- 'block' module: fix wrong usage of lockdep API
'macros' crate:
- Provide correct provenance when constructing 'THIS_MODULE'
Documentation:
- Remove unintended indentation (blockquotes) in generated output
- Fix a couple typos
MAINTAINERS:
- Remove Wedson as Rust maintainer
- Update Andreas' email"
* tag 'rust-fixes-6.11-2' of https://github.com/Rust-for-Linux/linux:
MAINTAINERS: update Andreas Hindborg's email address
MAINTAINERS: Remove Wedson as Rust maintainer
rust: macros: provide correct provenance when constructing THIS_MODULE
rust: allow `stable_features` lint
docs: rust: remove unintended blockquote in Quick Start
rust: alloc: eschew `Box<MaybeUninit<T>>::write`
rust: kernel: fix typos in code comments
docs: rust: remove unintended blockquote in Coding Guidelines
rust: block: fix wrong usage of lockdep API
rust: kbuild: fix export of bss symbols
Merge tag 'trace-v6.11-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull tracing fixes from Steven Rostedt:
- Fix adding a new fgraph callback after function graph tracing has
already started.
If the new caller does not initialize its hash before registering the
fgraph_ops, it can cause a NULL pointer dereference. Fix this by
adding a new parameter to ftrace_graph_enable_direct() passing in the
newly added gops directly and not rely on using the fgraph_array[],
as entries in the fgraph_array[] must be initialized.
Assign the new gops to the fgraph_array[] after it goes through
ftrace_startup_subops() as that will properly initialize the
gops->ops and initialize its hashes.
- Fix a memory leak in fgraph storage memory test.
If the "multiple fgraph storage on a function" boot up selftest fails
in the registering of the function graph tracer, it will not free the
memory it allocated for the filter. Break the loop up into two where
it allocates the filters first and then registers the functions where
any errors will do the appropriate clean ups.
- Only clear the timerlat timers if it has an associated kthread.
In the rtla tool that uses timerlat, if it was killed just as it was
shutting down, the signals can free the kthread and the timer. But
the closing of the timerlat files could cause the hrtimer_cancel() to
be called on the already freed timer. As the kthread variable is is
set to NULL when the kthreads are stopped and the timers are freed it
can be used to know not to call hrtimer_cancel() on the timer if the
kthread variable is NULL.
- Use a cpumask to keep track of osnoise/timerlat kthreads
The timerlat tracer can use user space threads for its analysis. With
the killing of the rtla tool, the kernel can get confused between if
it is using a user space thread to analyze or one of its own kernel
threads. When this confusion happens, kthread_stop() can be called on
a user space thread and bad things happen. As the kernel threads are
per-cpu, a bitmask can be used to know when a kernel thread is used
or when a user space thread is used.
- Add missing interface_lock to osnoise/timerlat stop_kthread()
The stop_kthread() function in osnoise/timerlat clears the osnoise
kthread variable, and if it was a user space thread does a put_task
on it. But this can race with the closing of the timerlat files that
also does a put_task on the kthread, and if the race happens the task
will have put_task called on it twice and oops.
- Add cond_resched() to the tracing_iter_reset() loop.
The latency tracers keep writing to the ring buffer without resetting
when it issues a new "start" event (like interrupts being disabled).
When reading the buffer with an iterator, the tracing_iter_reset()
sets its pointer to that start event by walking through all the
events in the buffer until it gets to the time stamp of the start
event. In the case of a very large buffer, the loop that looks for
the start event has been reported taking a very long time with a non
preempt kernel that it can trigger a soft lock up warning. Add a
cond_resched() into that loop to make sure that doesn't happen.
- Use list_del_rcu() for eventfs ei->list variable
It was reported that running loops of creating and deleting kprobe
events could cause a crash due to the eventfs list iteration hitting
a LIST_POISON variable. This is because the list is protected by SRCU
but when an item is deleted from the list, it was using list_del()
which poisons the "next" pointer. This is what list_del_rcu() was to
prevent.
* tag 'trace-v6.11-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tracing/timerlat: Add interface_lock around clearing of kthread in stop_kthread()
tracing/timerlat: Only clear timer if a kthread exists
tracing/osnoise: Use a cpumask to know what threads are kthreads
eventfs: Use list_del_rcu() for SRCU protected list variable
tracing: Avoid possible softlockup in tracing_iter_reset()
tracing: Fix memory leak in fgraph storage selftest
tracing: fgraph: Fix to add new fgraph_ops to array after ftrace_startup_subops()
Memory state around the buggy address: ffff88806461ff00: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc ffff88806461ff80: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff888064620000: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
^ ffff888064620080: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ffff888064620100: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
Execution of command:
./tools/net/ynl/cli.py --spec Documentation/netlink/specs/dpll.yaml /
--subscribe "monitor" --sleep 10
fails with:
File "/repo/./tools/net/ynl/cli.py", line 109, in main
ynl.check_ntf()
File "/repo/tools/net/ynl/lib/ynl.py", line 924, in check_ntf
op = self.rsp_by_value[nl_msg.cmd()]
KeyError: 19
Parsing Generic Netlink notification messages performs lookup for op in
the message. The message was not yet decoded, and is not yet considered
GenlMsg, thus msg.cmd() returns Generic Netlink family id (19) instead of
proper notification command id (i.e.: DPLL_CMD_PIN_CHANGE_NTF=13).
Allow the op to be obtained within NetlinkProtocol.decode(..) itself if the
op was not passed to the decode function, thus allow parsing of Generic
Netlink notifications without causing the failure.
Merge tag 'platform-drivers-x86-v6.11-6' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86
Pull x86 platform driver fixes from Ilpo Järvinen:
- amd/pmf: ASUS GA403 quirk matching tweak
- dell-smbios: Fix to the init function rollback path
* tag 'platform-drivers-x86-v6.11-6' of git://git.kernel.org/pub/scm/linux/kernel/git/pdx86/platform-drivers-x86:
platform/x86/amd: pmf: Make ASUS GA403 quirk generic
platform/x86: dell-smbios: Fix error path in dell_smbios_init()
Merge tag 'linux_kselftest-kunit-fixes-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest
Pull kunit fix fromShuah Khan:
"One single fix to a use-after-free bug resulting from
kunit_driver_create() failing to copy the driver name leaving it on
the stack or freeing it"
* tag 'linux_kselftest-kunit-fixes-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest:
kunit: Device wrappers should also manage driver name
Steven Rostedt [Thu, 5 Sep 2024 15:33:59 +0000 (11:33 -0400)]
tracing/timerlat: Add interface_lock around clearing of kthread in stop_kthread()
The timerlat interface will get and put the task that is part of the
"kthread" field of the osn_var to keep it around until all references are
released. But here's a race in the "stop_kthread()" code that will call
put_task_struct() on the kthread if it is not a kernel thread. This can
race with the releasing of the references to that task struct and the
put_task_struct() can be called twice when it should have been called just
once.
Take the interface_lock() in stop_kthread() to synchronize this change.
But to do so, the function stop_per_cpu_kthreads() needs to change the
loop from for_each_online_cpu() to for_each_possible_cpu() and remove the
cpu_read_lock(), as the interface_lock can not be taken while the cpu
locks are held. The only side effect of this change is that it may do some
extra work, as the per_cpu variables of the offline CPUs would not be set
anyway, and would simply be skipped in the loop.
Steven Rostedt [Thu, 5 Sep 2024 12:53:30 +0000 (08:53 -0400)]
tracing/timerlat: Only clear timer if a kthread exists
The timerlat tracer can use user space threads to check for osnoise and
timer latency. If the program using this is killed via a SIGTERM, the
threads are shutdown one at a time and another tracing instance can start
up resetting the threads before they are fully closed. That causes the
hrtimer assigned to the kthread to be shutdown and freed twice when the
dying thread finally closes the file descriptors, causing a use-after-free
bug.
Only cancel the hrtimer if the associated thread is still around. Also add
the interface_lock around the resetting of the tlat_var->kthread.
Note, this is just a quick fix that can be backported to stable. A real
fix is to have a better synchronization between the shutdown of old
threads and the starting of new ones.
Steven Rostedt [Wed, 4 Sep 2024 14:34:28 +0000 (10:34 -0400)]
tracing/osnoise: Use a cpumask to know what threads are kthreads
The start_kthread() and stop_thread() code was not always called with the
interface_lock held. This means that the kthread variable could be
unexpectedly changed causing the kthread_stop() to be called on it when it
should not have been, leading to:
while true; do
rtla timerlat top -u -q & PID=$!;
sleep 5;
kill -INT $PID;
sleep 0.001;
kill -TERM $PID;
wait $PID;
done
This is because it would mistakenly call kthread_stop() on a user space
thread making it "exit" before it actually exits.
Since kthreads are created based on global behavior, use a cpumask to know
when kthreads are running and that they need to be shutdown before
proceeding to do new work.
Steven Rostedt [Wed, 4 Sep 2024 17:16:05 +0000 (13:16 -0400)]
eventfs: Use list_del_rcu() for SRCU protected list variable
Chi Zhiling reported:
We found a null pointer accessing in tracefs[1], the reason is that the
variable 'ei_child' is set to LIST_POISON1, that means the list was
removed in eventfs_remove_rec. so when access the ei_child->is_freed, the
panic triggered.
by the way, the following script can reproduce this panic
loop1 (){
while true
do
echo "p:kp submit_bio" > /sys/kernel/debug/tracing/kprobe_events
echo "" > /sys/kernel/debug/tracing/kprobe_events
done
}
loop2 (){
while true
do
tree /sys/kernel/debug/tracing/events/kprobes/
done
}
loop1 &
loop2
The issue is that list_del() is used on an SRCU protected list variable
before the synchronization occurs. This can poison the list pointers while
there is a reader iterating the list.
This is simply fixed by using list_del_rcu() that is specifically made for
this purpose.
Zheng Yejian [Tue, 27 Aug 2024 12:46:54 +0000 (20:46 +0800)]
tracing: Avoid possible softlockup in tracing_iter_reset()
In __tracing_open(), when max latency tracers took place on the cpu,
the time start of its buffer would be updated, then event entries with
timestamps being earlier than start of the buffer would be skipped
(see tracing_iter_reset()).
Softlockup will occur if the kernel is non-preemptible and too many
entries were skipped in the loop that reset every cpu buffer, so add
cond_resched() to avoid it.
Stefan Wahren [Thu, 5 Sep 2024 11:15:37 +0000 (13:15 +0200)]
spi: spi-fsl-lpspi: Fix off-by-one in prescale max
The commit 783bf5d09f86 ("spi: spi-fsl-lpspi: limit PRESCALE bit in
TCR register") doesn't implement the prescaler maximum as intended.
The maximum allowed value for i.MX93 should be 1 and for i.MX7ULP
it should be 7. So this needs also a adjustment of the comparison
in the scldiv calculation.
sched: sch_cake: fix bulk flow accounting logic for host fairness
In sch_cake, we keep track of the count of active bulk flows per host,
when running in dst/src host fairness mode, which is used as the
round-robin weight when iterating through flows. The count of active
bulk flows is updated whenever a flow changes state.
This has a peculiar interaction with the hash collision handling: when a
hash collision occurs (after the set-associative hashing), the state of
the hash bucket is simply updated to match the new packet that collided,
and if host fairness is enabled, that also means assigning new per-host
state to the flow. For this reason, the bulk flow counters of the
host(s) assigned to the flow are decremented, before new state is
assigned (and the counters, which may not belong to the same host
anymore, are incremented again).
Back when this code was introduced, the host fairness mode was always
enabled, so the decrement was unconditional. When the configuration
flags were introduced the *increment* was made conditional, but
the *decrement* was not. Which of course can lead to a spurious
decrement (and associated wrap-around to U16_MAX).
AFAICT, when host fairness is disabled, the decrement and wrap-around
happens as soon as a hash collision occurs (which is not that common in
itself, due to the set-associative hashing). However, in most cases this
is harmless, as the value is only used when host fairness mode is
enabled. So in order to trigger an array overflow, sch_cake has to first
be configured with host fairness disabled, and while running in this
mode, a hash collision has to occur to cause the overflow. Then, the
qdisc has to be reconfigured to enable host fairness, which leads to the
array out-of-bounds because the wrapped-around value is retained and
used as an array index. It seems that syzbot managed to trigger this,
which is quite impressive in its own right.
This patch fixes the issue by introducing the same conditional check on
decrement as is used on increment.
The original bug predates the upstreaming of cake, but the commit listed
in the Fixes tag touched that code, meaning that this patch won't apply
before that.
Jakub Kicinski [Fri, 30 Aug 2024 17:14:42 +0000 (10:14 -0700)]
docs: netdev: document guidance on cleanup.h
Document what was discussed multiple times on list and various
virtual / in-person conversations. guard() being okay in functions
<= 20 LoC is a bit of my own invention. If the function is trivial
it should be fine, but feel free to disagree :)
We'll obviously revisit this guidance as time passes and we and other
subsystems get more experience.
Jakub Kicinski [Thu, 5 Sep 2024 00:37:37 +0000 (17:37 -0700)]
Merge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Tony Nguyen says:
====================
ice: fix synchronization between .ndo_bpf() and reset
Larysa Zaremba says:
PF reset can be triggered asynchronously, by tx_timeout or by a user. With some
unfortunate timings both ice_vsi_rebuild() and .ndo_bpf will try to access and
modify XDP rings at the same time, causing system crash.
The first patch factors out rtnl-locked code from VSI rebuild code to avoid
deadlock. The following changes lock rebuild and .ndo_bpf() critical sections
with an internal mutex as well and provide complementary fixes.
* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
ice: do not bring the VSI up, if it was down before the XDP setup
ice: remove ICE_CFG_BUSY locking from AF_XDP code
ice: check ICE_VSI_DOWN under rtnl_lock when preparing for reset
ice: check for XDP rings instead of bpf program when unconfiguring
ice: protect XDP configuration with a mutex
ice: move netif_queue_set_napi to rtnl-protected sections
====================
Jakub Kicinski [Thu, 5 Sep 2024 00:14:11 +0000 (17:14 -0700)]
Merge tag 'wireless-2024-09-04' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless
Kalle Valo says:
====================
wireless fixes for v6.11
Hopefully final fixes for v6.11 and this time only fixes to ath11k
driver. We need to revert hibernation support due to reported
regressions and we have a fix for kernel crash introduced in
v6.11-rc1.
* tag 'wireless-2024-09-04' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless:
MAINTAINERS: wifi: cw1200: add net-cw1200.h
Revert "wifi: ath11k: support hibernation"
Revert "wifi: ath11k: restore country code during resume"
wifi: ath11k: fix NULL pointer dereference in ath11k_mac_get_eirp_power()
====================
Sean Anderson [Tue, 3 Sep 2024 17:51:41 +0000 (13:51 -0400)]
net: xilinx: axienet: Fix race in axienet_stop
axienet_dma_err_handler can race with axienet_stop in the following
manner:
CPU 1 CPU 2
====================== ==================
axienet_stop()
napi_disable()
axienet_dma_stop()
axienet_dma_err_handler()
napi_disable()
axienet_dma_stop()
axienet_dma_start()
napi_enable()
cancel_work_sync()
free_irq()
Fix this by setting a flag in axienet_stop telling
axienet_dma_err_handler not to bother doing anything. I chose not to use
disable_work_sync to allow for easier backporting.
Jonas Gorski [Tue, 3 Sep 2024 08:19:57 +0000 (10:19 +0200)]
net: bridge: br_fdb_external_learn_add(): always set EXT_LEARN
When userspace wants to take over a fdb entry by setting it as
EXTERN_LEARNED, we set both flags BR_FDB_ADDED_BY_EXT_LEARN and
BR_FDB_ADDED_BY_USER in br_fdb_external_learn_add().
If the bridge updates the entry later because its port changed, we clear
the BR_FDB_ADDED_BY_EXT_LEARN flag, but leave the BR_FDB_ADDED_BY_USER
flag set.
If userspace then wants to take over the entry again,
br_fdb_external_learn_add() sees that BR_FDB_ADDED_BY_USER and skips
setting the BR_FDB_ADDED_BY_EXT_LEARN flags, thus silently ignores the
update.
Fix this by always allowing to set BR_FDB_ADDED_BY_EXT_LEARN regardless
if this was a user fdb entry or not.
Hayes Wang [Tue, 3 Sep 2024 06:33:33 +0000 (14:33 +0800)]
r8152: fix the firmware doesn't work
generic_ocp_write() asks the parameter "size" must be 4 bytes align.
Therefore, write the bp would fail, if the mac->bp_num is odd. Align the
size to 4 for fixing it. The way may write an extra bp, but the
rtl8152_is_fw_mac_ok() makes sure the value must be 0 for the bp whose
index is more than mac->bp_num. That is, there is no influence for the
firmware.
Besides, I check the return value of generic_ocp_write() to make sure
everything is correct.
We observed a null-ptr-deref in fou_gro_receive() while shutting down
a host. [0]
The NULL pointer is sk->sk_user_data, and the offset 8 is of protocol
in struct fou.
When fou_release() is called due to netns dismantle or explicit tunnel
teardown, udp_tunnel_sock_release() sets NULL to sk->sk_user_data.
Then, the tunnel socket is destroyed after a single RCU grace period.
So, in-flight udp4_gro_receive() could find the socket and execute the
FOU GRO handler, where sk->sk_user_data could be NULL.
Let's use rcu_dereference_sk_user_data() in fou_from_sock() and add NULL
checks in FOU GRO handlers.
Merge tag 'bcachefs-2024-09-04' of git://evilpiepirate.org/bcachefs
Pull bcachefs fixes from Kent Overstreet:
- Fix a typo in the rebalance accounting changes
- BCH_SB_MEMBER_INVALID: small on disk format feature which will be
needed for full erasure coding support; this is only the minimum so
that 6.11 can handle future versions without barfing.
* tag 'bcachefs-2024-09-04' of git://evilpiepirate.org/bcachefs:
bcachefs: BCH_SB_MEMBER_INVALID
bcachefs: fix rebalance accounting
Merge tag 'perf-tools-fixes-for-v6.11-2024-09-04' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools
Pull perf tools fixes from Namhyung Kim:
"A number of small fixes for the late cycle:
- Two more build fixes on 32-bit archs
- Fixed a segfault during perf test
- Fixed spinlock/rwlock accounting bug in perf lock contention"
* tag 'perf-tools-fixes-for-v6.11-2024-09-04' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools:
perf daemon: Fix the build on more 32-bit architectures
perf python: include "util/sample.h"
perf lock contention: Fix spinlock and rwlock accounting
perf test pmu: Set uninitialized PMU alias to null
Merge tag 'hwmon-for-v6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging
Pull hwmon fixes from Guenter Roeck:
- hp-wmi-sensors: Check if WMI event data exists before accessing it
- ltc2991: fix register bits defines
* tag 'hwmon-for-v6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
hwmon: (hp-wmi-sensors) Check if WMI event data exists
hwmon: ltc2991: fix register bits defines
Merge tag 'for-6.11-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux
Pull btrfs fixes from David Sterba:
- followup fix for direct io and fsync under some conditions, reported
by QEMU users
- fix a potential leak when disabling quotas while some extent tracking
work can still happen
- in zoned mode handle unexpected change of zone write pointer in
RAID1-like block groups, turn the zones to read-only
* tag 'for-6.11-rc6-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
btrfs: fix race between direct IO write and fsync when using same fd
btrfs: zoned: handle broken write pointer on zones
btrfs: qgroup: don't use extent changeset when not needed
Merge tag 'v6.11-rc6-server-fixes' of git://git.samba.org/ksmbd
Pull smb server fixes from Steve French:
- Fix crash in session setup
- Fix locking bug
- Improve access bounds checking
* tag 'v6.11-rc6-server-fixes' of git://git.samba.org/ksmbd:
ksmbd: Unlock on in ksmbd_tcp_set_interfaces()
ksmbd: unset the binding mark of a reused connection
smb: Annotate struct xattr_smb_acl with __counted_by()
Merge tag 'vfs-6.11-rc7.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs
Pull vfs fixes from Christian Brauner:
"Two netfs fixes for this merge window:
- Ensure that fscache_cookie_lru_time is deleted when the fscache
module is removed to prevent UAF
- Fix filemap_invalidate_inode() to use invalidate_inode_pages2_range()
Before it used truncate_inode_pages_partial() which causes
copy_file_range() to fail on cifs"
* tag 'vfs-6.11-rc7.fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/vfs/vfs:
fscache: delete fscache_cookie_lru_timer when fscache exits to avoid UAF
mm: Fix filemap_invalidate_inode() to use invalidate_inode_pages2_range()
Merge tag 'parisc-for-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux
Pull parisc architecture fix from Helge Deller:
- Fix boot issue where boot memory is marked read-only too early
* tag 'parisc-for-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/deller/parisc-linux:
parisc: Delay write-protection until mark_rodata_ro() call
Merge tag 'mm-hotfixes-stable-2024-09-03-20-19' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
Pull misc fixes from Andrew Morton:
"17 hotfixes, 15 of which are cc:stable.
Mostly MM, no identifiable theme. And a few nilfs2 fixups"
* tag 'mm-hotfixes-stable-2024-09-03-20-19' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
alloc_tag: fix allocation tag reporting when CONFIG_MODULES=n
mm: vmalloc: optimize vmap_lazy_nr arithmetic when purging each vmap_area
mailmap: update entry for Jan Kuliga
codetag: debug: mark codetags for poisoned page as empty
mm/memcontrol: respect zswap.writeback setting from parent cg too
scripts: fix gfp-translate after ___GFP_*_BITS conversion to an enum
Revert "mm: skip CMA pages when they are not available"
maple_tree: remove rcu_read_lock() from mt_validate()
kexec_file: fix elfcorehdr digest exclusion when CONFIG_CRASH_HOTPLUG=y
mm/slub: add check for s->flags in the alloc_tagging_slab_free_hook
nilfs2: fix state management in error path of log writing function
nilfs2: fix missing cleanup on rollforward recovery error
nilfs2: protect references to superblock parameters exposed in sysfs
userfaultfd: don't BUG_ON() if khugepaged yanks our page table
userfaultfd: fix checks for huge PMDs
mm: vmalloc: ensure vmap_block is initialised before adding to queue
selftests: mm: fix build errors on armhf
The error occurs in older versions of the GNU ld with version earlier
than 2.36. It makes most sense to have a minimum LD version as
a dependency for HAVE_LD_DEAD_CODE_DATA_ELIMINATION and eliminate
the impact of ".reloc .text, R_ARM_NONE, ." when
CONFIG_LD_DEAD_CODE_DATA_ELIMINATION is not enabled.
Kent Overstreet [Sun, 1 Sep 2024 22:09:18 +0000 (18:09 -0400)]
bcachefs: BCH_SB_MEMBER_INVALID
Create a sentinal value for "invalid device".
This is needed for removing devices that have stripes on them (force
removing, without evacuating); we need a sentinal value for the stripe
pointers to the device being removed.
Merge tag 'fuse-fixes-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse
Pull fuse fixes from Miklos Szeredi:
- Fix EIO if splice and page stealing are enabled on the fuse device
- Disable problematic combination of passthrough and writeback-cache
- Other bug fixes found by code review
* tag 'fuse-fixes-6.11-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse:
fuse: disable the combination of passthrough and writeback cache
fuse: update stats for pages in dropped aux writeback list
fuse: clear PG_uptodate when using a stolen page
fuse: fix memory leak in fuse_create_open
fuse: check aborted connection before adding requests to pending list for resending
fuse: use unsigned type for getxattr/listxattr size truncation
Tze-nan Wu [Fri, 30 Aug 2024 08:25:17 +0000 (16:25 +0800)]
bpf, net: Fix a potential race in do_sock_getsockopt()
There's a potential race when `cgroup_bpf_enabled(CGROUP_GETSOCKOPT)` is
false during the execution of `BPF_CGROUP_GETSOCKOPT_MAX_OPTLEN`, but
becomes true when `BPF_CGROUP_RUN_PROG_GETSOCKOPT` is called.
This inconsistency can lead to `BPF_CGROUP_RUN_PROG_GETSOCKOPT` receiving
an "-EFAULT" from `__cgroup_bpf_run_filter_getsockopt(max_optlen=0)`.
Scenario shown as below:
To resolve this, remove the `BPF_CGROUP_GETSOCKOPT_MAX_OPTLEN` macro and
directly uses `copy_from_sockptr` to ensure that `max_optlen` is always
set before `BPF_CGROUP_RUN_PROG_GETSOCKOPT` is invoked.
When CONFIG_DQL is not enabled, dql_group should be treated as a dead
declaration. However, its current extern declaration assumes the linker
will ignore it, which is generally true across most compiler and
architecture combinations.
But in certain cases, the linker still attempts to resolve the extern
struct, even when the associated code is dead, resulting in a linking
error. For instance the following error in loongarch64:
>> loongarch64-linux-ld: net-sysfs.c:(.text+0x589c): undefined reference to `dql_group'
Modify the declaration of the dead object to be an empty declaration
instead of an extern. This change will prevent the linker from
attempting to resolve an undefined reference.
If netem_dequeue() enqueues packet to inner qdisc and that qdisc
returns __NET_XMIT_STOLEN. The packet is dropped but
qdisc_tree_reduce_backlog() is not called to update the parent's
q.qlen, leading to the similar use-after-free as Commit e04991a48dbaf382 ("netem: fix return value if duplicate enqueue
fails")
Commands to trigger KASAN UaF:
ip link add type dummy
ip link set lo up
ip link set dummy0 up
tc qdisc add dev lo parent root handle 1: drr
tc filter add dev lo parent 1: basic classid 1:1
tc class add dev lo classid 1:1 drr
tc qdisc add dev lo parent 1:1 handle 2: netem
tc qdisc add dev lo parent 2: handle 3: drr
tc filter add dev lo parent 3: basic classid 3:1 action mirred egress
redirect dev dummy0
tc class add dev lo classid 3:1 drr
ping -c1 -W0.01 localhost # Trigger bug
tc class del dev lo classid 1:1
tc class add dev lo classid 1:1 drr
ping -c1 -W0.01 localhost # UaF
Simon Horman [Mon, 2 Sep 2024 09:09:37 +0000 (10:09 +0100)]
MAINTAINERS: wifi: cw1200: add net-cw1200.h
This is part of an effort [1] to assign a section in MAINTAINERS to header
files that relate to Networking. In this case the files with "net" in
their name.
Filipe Manana [Thu, 29 Aug 2024 17:25:49 +0000 (18:25 +0100)]
btrfs: fix race between direct IO write and fsync when using same fd
If we have 2 threads that are using the same file descriptor and one of
them is doing direct IO writes while the other is doing fsync, we have a
race where we can end up either:
1) Attempt a fsync without holding the inode's lock, triggering an
assertion failures when assertions are enabled;
2) Do an invalid memory access from the fsync task because the file private
points to memory allocated on stack by the direct IO task and it may be
used by the fsync task after the stack was destroyed.
The race happens like this:
1) A user space program opens a file descriptor with O_DIRECT;
2) The program spawns 2 threads using libpthread for example;
3) One of the threads uses the file descriptor to do direct IO writes,
while the other calls fsync using the same file descriptor.
4) Call task A the thread doing direct IO writes and task B the thread
doing fsyncs;
5) Task A does a direct IO write, and at btrfs_direct_write() sets the
file's private to an on stack allocated private with the member
'fsync_skip_inode_lock' set to true;
6) Task B enters btrfs_sync_file() and sees that there's a private
structure associated to the file which has 'fsync_skip_inode_lock' set
to true, so it skips locking the inode's VFS lock;
7) Task A completes the direct IO write, and resets the file's private to
NULL since it had no prior private and our private was stack allocated.
Then it unlocks the inode's VFS lock;
8) Task B enters btrfs_get_ordered_extents_for_logging(), then the
assertion that checks the inode's VFS lock is held fails, since task B
never locked it and task A has already unlocked it.
Another problem here is if task B grabs the private pointer and then uses
it after task A has finished, since the private was allocated in the stack
of task A, it results in some invalid memory access with a hard to predict
result.
This issue, triggering the assertion, was observed with QEMU workloads by
two users in the Link tags below.
Fix this by not relying on a file's private to pass information to fsync
that it should skip locking the inode and instead pass this information
through a special value stored in current->journal_info. This is safe
because in the relevant section of the direct IO write path we are not
holding a transaction handle, so current->journal_info is NULL.
The following C program triggers the issue:
$ cat repro.c
/* Get the O_DIRECT definition. */
#ifndef _GNU_SOURCE
#define _GNU_SOURCE
#endif
Kalle Valo [Tue, 3 Sep 2024 16:54:15 +0000 (19:54 +0300)]
Merge tag 'ath-current-20240903' of git://git.kernel.org/pub/scm/linux/kernel/git/ath/ath
ath.git patches for v6.11-rc7
We have three patch which address two issues in the ath11k driver
which should be addressed for 6.11-rc7:
One patch fixes a NULL pointer dereference while parsing transmit
power envelope (TPE) information, and the other two patches revert the
hibernation support since it is interfering with suspend on some
platforms. Note the cause of the suspend wakeups is still being
investigated, and it is hoped this can be addressed and hibernation
support can be restored in the near future.
Larysa Zaremba [Fri, 23 Aug 2024 09:59:30 +0000 (11:59 +0200)]
ice: remove ICE_CFG_BUSY locking from AF_XDP code
Locking used in ice_qp_ena() and ice_qp_dis() does pretty much nothing,
because ICE_CFG_BUSY is a state flag that is supposed to be set in a PF
state, not VSI one. Therefore it does not protect the queue pair from
e.g. reset.
Remove ICE_CFG_BUSY locking from ice_qp_dis() and ice_qp_ena().
Larysa Zaremba [Fri, 23 Aug 2024 09:59:28 +0000 (11:59 +0200)]
ice: check for XDP rings instead of bpf program when unconfiguring
If VSI rebuild is pending, .ndo_bpf() can attach/detach the XDP program on
VSI without applying new ring configuration. When unconfiguring the VSI, we
can encounter the state in which there is an XDP program but no XDP rings
to destroy or there will be XDP rings that need to be destroyed, but no XDP
program to indicate their presence.
When unconfiguring, rely on the presence of XDP rings rather then XDP
program, as they better represent the current state that has to be
destroyed.
Larysa Zaremba [Fri, 23 Aug 2024 09:59:27 +0000 (11:59 +0200)]
ice: protect XDP configuration with a mutex
The main threat to data consistency in ice_xdp() is a possible asynchronous
PF reset. It can be triggered by a user or by TX timeout handler.
XDP setup and PF reset code access the same resources in the following
sections:
* ice_vsi_close() in ice_prepare_for_reset() - already rtnl-locked
* ice_vsi_rebuild() for the PF VSI - not protected
* ice_vsi_open() - already rtnl-locked
With an unfortunate timing, such accesses can result in a crash such as the
one below:
The previous way of handling this through returning -EBUSY is not viable,
particularly when destroying AF_XDP socket, because the kernel proceeds
with removal anyway.
There is plenty of code between those calls and there is no need to create
a large critical section that covers all of them, same as there is no need
to protect ice_vsi_rebuild() with rtnl_lock().
Add xdp_state_lock mutex to protect ice_vsi_rebuild() and ice_xdp().
Leaving unprotected sections in between would result in two states that
have to be considered:
1. when the VSI is closed, but not yet rebuild
2. when VSI is already rebuild, but not yet open
The latter case is actually already handled through !netif_running() case,
we just need to adjust flag checking a little. The former one is not as
trivial, because between ice_vsi_close() and ice_vsi_rebuild(), a lot of
hardware interaction happens, this can make adding/deleting rings exit
with an error. Luckily, VSI rebuild is pending and can apply new
configuration for us in a managed fashion.
Therefore, add an additional VSI state flag ICE_VSI_REBUILD_PENDING to
indicate that ice_xdp() can just hot-swap the program.
Also, as ice_vsi_rebuild() flow is touched in this patch, make it more
consistent by deconfiguring VSI when coalesce allocation fails.
Larysa Zaremba [Fri, 23 Aug 2024 09:59:26 +0000 (11:59 +0200)]
ice: move netif_queue_set_napi to rtnl-protected sections
Currently, netif_queue_set_napi() is called from ice_vsi_rebuild() that is
not rtnl-locked when called from the reset. This creates the need to take
the rtnl_lock just for a single function and complicates the
synchronization with .ndo_bpf. At the same time, there no actual need to
fill napi-to-queue information at this exact point.
Fill napi-to-queue information when opening the VSI and clear it when the
VSI is being closed. Those routines are already rtnl-locked.
Also, rewrite napi-to-queue assignment in a way that prevents inclusion of
XDP queues, as this leads to out-of-bounds writes, such as one below.
Fixes: 080b0c8d6d26 ("ice: Fix ASSERT_RTNL() warning during certain scenarios") Fixes: 91fdbce7e8d6 ("ice: Add support in the driver for associating queue with napi") Reviewed-by: Wojciech Drewek <[email protected]> Reviewed-by: Jacob Keller <[email protected]> Reviewed-by: Amritha Nambiar <[email protected]> Signed-off-by: Larysa Zaremba <[email protected]> Reviewed-by: Maciej Fijalkowski <[email protected]> Tested-by: George Kuruvinakunnel <[email protected]> Signed-off-by: Tony Nguyen <[email protected]>
====================
ptp: ocp: fix serial port information export
Starting v6.8 the serial port subsystem changed the hierarchy of devices
and symlinks are not working anymore. Previous discussion made it clear
that the idea of symlinks for tty devices was wrong by design [1].
This series implements additional attributes to expose the information
and removes symlinks for tty devices.
v6 -> v7:
- fix issues with applying patches
v5 -> v6:
- split conversion to array to separate patch per Jiri's feedback
- move changelog to cover letter
v4 -> v5:
- remove unused variable in ptp_ocp_tty_show
v3 -> v4:
- re-organize info printing to use ptp_ocp_tty_port_name()
- keep uintptr_t to be consistent with other code
v2 -> v3:
- replace serial ports definitions with array and enum for index
- replace pointer math with direct array access
- nit in documentation spelling
v1 -> v2:
- add Documentation/ABI changes
====================
Vadim Fedorenko [Thu, 29 Aug 2024 18:36:03 +0000 (11:36 -0700)]
docs: ABI: update OCP TimeCard sysfs entries
Update documentation according to the changes in the driver.
New attributes group tty is exposed and ttyGNSS, ttyGNSS2, ttyMAC and
ttyNMEA are moved to this group. Also, these attributes are no more
links to the devices but rather simple text files containing names of
tty devices.
Vadim Fedorenko [Thu, 29 Aug 2024 18:36:02 +0000 (11:36 -0700)]
ptp: ocp: adjust sysfs entries to expose tty information
Implement additional attribute group to expose serial port information.
Fixes tag points to the commit which introduced the change in serial
port subsystem and made it impossible to use symlinks.
Fixes: b286f4e87e32 ("serial: core: Move tty and serdev to be children of serial core port device") Signed-off-by: Vadim Fedorenko <[email protected]> Reviewed-by: Greg Kroah-Hartman <[email protected]> Signed-off-by: Paolo Abeni <[email protected]>
Vadim Fedorenko [Thu, 29 Aug 2024 18:36:01 +0000 (11:36 -0700)]
ptp: ocp: convert serial ports to array
Simplify serial port management code by using array of ports and helpers
to get the name of the port. This change is needed to make the next
patch simplier.
Helge Deller [Sat, 31 Aug 2024 12:02:06 +0000 (14:02 +0200)]
parisc: Delay write-protection until mark_rodata_ro() call
Do not write-protect the kernel read-only and __ro_after_init sections
earlier than before mark_rodata_ro() is called. This fixes a boot issue on
parisc which is triggered by commit 91a1d97ef482 ("jump_label,module: Don't
alloc static_key_mod for __ro_after_init keys"). That commit may modify
static key contents in the __ro_after_init section at bootup, so this
section needs to be writable at least until mark_rodata_ro() is called.
The XDP implementation on am65-cpsw driver is broken in many ways
and this series fixes it.
Below are the current issues that are being fixed:
1) The following XDP_DROP test from [1] stalls the interface after
250 packets.
~# xdb-bench drop -m native eth0
This is because new RX requests are never queued. Fix that.
2) The below XDP_TX test from [1] fails with a warning
[ 499.947381] XDP_WARN: xdp_update_frame_from_buff(line:277): Driver BUG: missing reserved tailroom
~# xdb-bench tx -m native eth0
Fix that by using PAGE_SIZE during xdp_init_buf().
3) In XDP_REDIRECT case only 1 packet was processed in rx_poll.
Fix it to process up to budget packets.
~# ./xdp-bench redirect -m native eth0 eth0
4) If number of TX queues are set to 1 we get a NULL pointer
dereference during XDP_TX.
~# ethtool -L eth0 tx 1
~# ./xdp-trafficgen udp -A <ipv6-src> -a <ipv6-dst> eth0 -t 2
Transmitting on eth0 (ifindex 2)
[ 241.135257] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000030
5) Net statistics is broken for XDP_TX and XDP_REDIRECT
[1] xdp-tools suite https://github.com/xdp-project/xdp-tools
Roger Quadros [Thu, 29 Aug 2024 12:03:19 +0000 (15:03 +0300)]
net: ethernet: ti: am65-cpsw: fix XDP_DROP, XDP_TX and XDP_REDIRECT
The following XDP_DROP test from [1] stalls the interface after
250 packets.
~# xdb-bench drop -m native eth0
This is because new RX requests are never queued. Fix that.
The below XDP_TX test from [1] fails with a warning
[ 499.947381] XDP_WARN: xdp_update_frame_from_buff(line:277): Driver BUG: missing reserved tailroom
~# xdb-bench tx -m native eth0
Fix that by using PAGE_SIZE during xdp_init_buf().
In XDP_REDIRECT case only 1 packet was processed in rx_poll.
Fix it to process up to budget packets.
Fix all XDP error cases to call trace_xdp_exception() and drop the packet
in am65_cpsw_run_xdp().
[1] xdp-tools suite https://github.com/xdp-project/xdp-tools
Jakub Kicinski [Tue, 3 Sep 2024 02:08:51 +0000 (19:08 -0700)]
Merge tag 'for-net-2024-08-30' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth
Luiz Augusto von Dentz says:
====================
bluetooth pull request for net:
- qca: If memdump doesn't work, re-enable IBS
- MGMT: Fix not generating command complete for MGMT_OP_DISCONNECT
- Revert "Bluetooth: MGMT/SMP: Fix address type when using SMP over BREDR/LE"
- MGMT: Ignore keys being loaded with invalid type
* tag 'for-net-2024-08-30' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth:
Bluetooth: MGMT: Ignore keys being loaded with invalid type
Revert "Bluetooth: MGMT/SMP: Fix address type when using SMP over BREDR/LE"
Bluetooth: MGMT: Fix not generating command complete for MGMT_OP_DISCONNECT
Bluetooth: hci_sync: Introduce hci_cmd_sync_run/hci_cmd_sync_run_once
Bluetooth: qca: If memdump doesn't work, re-enable IBS
====================
Jakub Kicinski [Tue, 3 Sep 2024 02:06:05 +0000 (19:06 -0700)]
Merge tag 'linux-can-fixes-for-6.11-20240830' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can
Marc Kleine-Budde says:
====================
pull-request: can 2024-08-30
The first patch is by Kuniyuki Iwashima for the CAN BCM protocol that
adds a missing proc entry removal when a device unregistered.
Simon Horman fixes the cleanup in the error cleanup path of the m_can
driver's open function.
Markus Schneider-Pargmann contributes 7 fixes for the m_can driver,
all related to the recently added IRQ coalescing support.
The next 2 patches are by me, target the mcp251xfd driver and fix ring
and coalescing configuration problems when switching from CAN-CC to
CAN-FD mode.
Simon Arlott's patch fixes a possible deadlock in the mcp251x driver.
The last patch is by Martin Jocic for the kvaser_pciefd driver and
fixes a problem with lost IRQs, which result in starvation, under high
load situations.
* tag 'linux-can-fixes-for-6.11-20240830' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can:
can: kvaser_pciefd: Use a single write when releasing RX buffers
can: mcp251x: fix deadlock if an interrupt occurs during mcp251x_open
can: mcp251xfd: mcp251xfd_ring_init(): check TX-coalescing configuration
can: mcp251xfd: fix ring configuration when switching from CAN-CC to CAN-FD mode
can: m_can: Limit coalescing to peripheral instances
can: m_can: Reset cached active_interrupts on start
can: m_can: disable_all_interrupts, not clear active_interrupts
can: m_can: Do not cancel timer from within timer
can: m_can: Remove m_can_rx_peripheral indirection
can: m_can: Remove coalesing disable in isr during suspend
can: m_can: Reset coalescing during suspend/resume
can: m_can: Release irq on error in m_can_open
can: bcm: Remove proc entry when dev is unregistered.
====================
Naohiro Aota [Fri, 30 Aug 2024 16:32:49 +0000 (01:32 +0900)]
btrfs: zoned: handle broken write pointer on zones
Btrfs rejects to mount a FS if it finds a block group with a broken write
pointer (e.g, unequal write pointers on two zones of RAID1 block group).
Since such case can happen easily with a power-loss or crash of a system,
we need to handle the case more gently.
Handle such block group by making it unallocatable, so that there will be
no writes into it. That can be done by setting the allocation pointer at
the end of allocating region (= block_group->zone_capacity). Then, existing
code handle zone_unusable properly.
Having proper zone_capacity is necessary for the change. So, set it as fast
as possible.
We cannot handle RAID0 and RAID10 case like this. But, they are anyway
unable to read because of a missing stripe.
Fixes: 265f7237dd25 ("btrfs: zoned: allow DUP on meta-data block groups") Fixes: 568220fa9657 ("btrfs: zoned: support RAID0/1/10 on top of raid stripe tree") CC: [email protected] # 6.1+ Reported-by: HAN Yuwei <[email protected]> Cc: Xuefer <[email protected]> Signed-off-by: Naohiro Aota <[email protected]> Signed-off-by: David Sterba <[email protected]>
projects / linux.git / log