Marek reported seeing a NULL pointer fault in the xenbus_thread
callstack:
BUG: kernel NULL pointer dereference, address: 0000000000000000
RIP: e030:__wake_up_common+0x4c/0x180
Call Trace:
<TASK>
__wake_up_common_lock+0x82/0xd0
process_msg+0x18e/0x2f0
xenbus_thread+0x165/0x1c0
process_msg+0x18e is req->cb(req). req->cb is set to xs_wake_up(), a
thin wrapper around wake_up(), or xenbus_dev_queue_reply(). It seems
like it was xs_wake_up() in this case.
It seems like req may have woken up the xs_wait_for_reply(), which
kfree()ed the req. When xenbus_thread resumes, it faults on the zero-ed
data.
Linux Device Drivers 2nd edition states:
"Normally, a wake_up call can cause an immediate reschedule to happen,
meaning that other processes might run before wake_up returns."
... which would match the behaviour observed.
Change to keeping two krefs on each request. One for the caller, and
one for xenbus_thread. Each will kref_put() when finished, and the last
will free it.
This use of kref matches the description in
Documentation/core-api/kref.rst
Device tree bindings state that the clock is optional for UHCI platform
controllers, and some existing device trees don't provide those - such
as those for VIA/WonderMedia devices.
The driver however fails to probe now if no clock is provided, because
devm_clk_get returns an error pointer in such case.
Switch to devm_clk_get_optional instead, so that it could probe again
on those platforms where no clocks are given.
[Why]
We incorrectly ack all bytes get written when the reply actually is defer.
When it's defer, means sink is not ready for the request. We should
retry the request.
[How]
Only reply all data get written when receive I2C_ACK|AUX_ACK. Otherwise,
reply the number of actual written bytes received from the sink.
Add some messages to facilitate debugging as well.
Fixes: ad6756b4d773 ("drm/amd/display: Shift dc link aux to aux_payload") Cc: Mario Limonciello <mario.limonciello@amd.com> Cc: Alex Deucher <alexander.deucher@amd.com> Reviewed-by: Ray Wu <ray.wu@amd.com> Signed-off-by: Wayne Lin <Wayne.Lin@amd.com> Signed-off-by: Ray Wu <ray.wu@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
(cherry picked from commit 3637e457eb0000bc37d8bbbec95964aad2fb29fd) Cc: stable@vger.kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The inclinometer channels were previously defined with 14 realbits.
However, the ADIS16201 datasheet states the resolution for these output
channels is 12 bits (Page 14, text description; Page 15, table 7).
Correct the realbits value to 12 to accurately reflect the hardware.
Fixes: f7fe1d1dd5a5 ("staging: iio: new adis16201 driver") Cc: stable@vger.kernel.org Signed-off-by: Gabriel Shahrouzi <gshahrouzi@gmail.com> Reviewed-by: Marcelo Schmitt <marcelo.schmitt1@gmail.com> Link: https://patch.msgid.link/20250421131539.912966-1-gshahrouzi@gmail.com Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When reading multiple consecutive registers, only the first one is read
properly. This is due to missing chip select deassert and assert again
between first and second 16bit transfer, as shown in the datasheet AD7606C-16, rev 0, figure 110.
Fixes: f2a22e1e172f ("iio: adc: ad7606: Add support for software mode for ad7616") Reviewed-by: David Lechner <dlechner@baylibre.com> Signed-off-by: Angelo Dureghello <adureghello@baylibre.com> Link: https://patch.msgid.link/20250418-wip-bl-ad7606-fix-reg-access-v3-1-d5eeb440c738@baylibre.com Cc: <Stable@vger.kernel.org> Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
tl;dr: There is a window in the mm switching code where the new CR3 is
set and the CPU should be getting TLB flushes for the new mm. But
should_flush_tlb() has a bug and suppresses the flush. Fix it by
widening the window where should_flush_tlb() sends an IPI.
Long Version:
=== History ===
There were a few things leading up to this.
First, updating mm_cpumask() was observed to be too expensive, so it was
made lazier. But being lazy caused too many unnecessary IPIs to CPUs
due to the now-lazy mm_cpumask(). So code was added to cull
mm_cpumask() periodically[2]. But that culling was a bit too aggressive
and skipped sending TLB flushes to CPUs that need them. So here we are
again.
=== Problem ===
The too-aggressive code in should_flush_tlb() strikes in this window:
// Turn on IPIs for this CPU/mm combination, but only
// if should_flush_tlb() agrees:
cpumask_set_cpu(cpu, mm_cpumask(next));
next_tlb_gen = atomic64_read(&next->context.tlb_gen);
choose_new_asid(next, next_tlb_gen, &new_asid, &need_flush);
load_new_mm_cr3(need_flush);
// ^ After 'need_flush' is set to false, IPIs *MUST*
// be sent to this CPU and not be ignored.
this_cpu_write(cpu_tlbstate.loaded_mm, next);
// ^ Not until this point does should_flush_tlb()
// become true!
should_flush_tlb() will suppress TLB flushes between load_new_mm_cr3()
and writing to 'loaded_mm', which is a window where they should not be
suppressed. Whoops.
=== Solution ===
Thankfully, the fuzzy "just about to write CR3" window is already marked
with loaded_mm==LOADED_MM_SWITCHING. Simply checking for that state in
should_flush_tlb() is sufficient to ensure that the CPU is targeted with
an IPI.
This will cause more TLB flush IPIs. But the window is relatively small
and I do not expect this to cause any kind of measurable performance
impact.
Update the comment where LOADED_MM_SWITCHING is written since it grew
yet another user.
Peter Z also raised a concern that should_flush_tlb() might not observe
'loaded_mm' and 'is_lazy' in the same order that switch_mm_irqs_off()
writes them. Add a barrier to ensure that they are observed in the
order they are written.
Signed-off-by: Dave Hansen <dave.hansen@linux.intel.com> Acked-by: Rik van Riel <riel@surriel.com> Link: https://lore.kernel.org/oe-lkp/202411282207.6bd28eae-lkp@intel.com/ Fixes: 6db2526c1d69 ("x86/mm/tlb: Only trim the mm_cpumask once a second") [2] Reported-by: Stephen Dolan <sdolan@janestreet.com> Cc: stable@vger.kernel.org Acked-by: Ingo Molnar <mingo@kernel.org> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove erroneous subtraction of 4 from the total FIFO depth read from
device tree. The stored depth is for checking against total capacity,
not initial vacancy. This prevented writes near the FIFO's full size.
The check performed just before data transfer, which uses live reads of
the TDFV register to determine current vacancy, correctly handles the
initial Depth - 4 hardware state and subsequent FIFO fullness.
Fixes: 4a965c5f89de ("staging: add driver for Xilinx AXI-Stream FIFO v4.1 IP core") Cc: stable@vger.kernel.org Signed-off-by: Gabriel Shahrouzi <gshahrouzi@gmail.com> Link: https://lore.kernel.org/r/20250419012937.674924-1-gshahrouzi@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The axis-fifo driver performs a full hardware reset (via
reset_ip_core()) in several error paths within the read and write
functions. This reset flushes both TX and RX FIFOs and resets the
AXI-Stream links.
Allow the user to handle the error without causing hardware disruption
or data loss in other FIFO paths.
Fixes: 4a965c5f89de ("staging: add driver for Xilinx AXI-Stream FIFO v4.1 IP core") Cc: stable@vger.kernel.org Signed-off-by: Gabriel Shahrouzi <gshahrouzi@gmail.com> Link: https://lore.kernel.org/r/20250419004306.669605-1-gshahrouzi@gmail.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The mode setting logic in ad7816_store_mode was reversed due to
incorrect handling of the strcmp return value. strcmp returns 0 on
match, so the `if (strcmp(buf, "full"))` block executed when the
input was not "full".
This resulted in "full" setting the mode to AD7816_PD (power-down) and
other inputs setting it to AD7816_FULL.
Fix this by checking it against 0 to correctly check for "full" and
"power-down", mapping them to AD7816_FULL and AD7816_PD respectively.
[ 5.989588] psmouse serio1: synaptics: Your touchpad (PNP: TOS0213 PNP0f03) says it can support a different bus. If i2c-hid and hid-rmi are not used, you might want to try setting psmouse.synaptics_intertouch to 1 and report this to linux-input@vger.kernel.org.
[ 6.039923] psmouse serio1: synaptics: Touchpad model: 1, fw: 9.32, id: 0x1e2a1, caps: 0xf00223/0x840300/0x12e800/0x52d884, board id: 3322, fw id: 2658004
The board is labelled TM3322.
Present on the Toshiba / Dynabook Portege X30-D and possibly others.
Confirmed working well with psmouse.synaptics_intertouch=1 and local build.
When VLAN filtering is off, we configure the switch to forward, but not
learn on VLAN table misses. This effectively disables learning while not
filtering.
Fix this by switching to forward and learn. Setting the learning disable
register will still control whether learning actually happens.
Currently the PVID of ports are only set when adding/updating VLANs with
PVID set or removing VLANs, but not when clearing the PVID flag of a
VLAN.
E.g. the following flow
$ ip link add br0 type bridge vlan_filtering 1
$ ip link set sw1p1 master bridge
$ bridge vlan add dev sw1p1 vid 10 pvid untagged
$ bridge vlan add dev sw1p1 vid 10 untagged
Would keep the PVID set as 10, despite the flag being cleared. Fix this
by checking if we need to unset the PVID on vlan updates.
Allow reserved multicast to ignore VLAN membership so STP and other
management protocols work without a PVID VLAN configured when using a
vlan aware bridge.
Fixes: 967dd82ffc52 ("net: dsa: b53: Add support for Broadcom RoboSwitch") Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com> Tested-by: Florian Fainelli <florian.fainelli@broadcom.com> Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> Link: https://patch.msgid.link/20250429201710.330937-2-jonas.gorski@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Region locking introduced in v5.6-rc4 contained three macros to handle
the region locks: ahash_bucket_start(), ahash_bucket_end() which gave
back the start and end hash bucket values belonging to a given region
lock and ahash_region() which should give back the region lock belonging
to a given hash bucket. The latter was incorrect which can lead to a
race condition between the garbage collector and adding new elements
when a hash type of set is defined with timeouts.
Fixes: f66ee0410b1c ("netfilter: ipset: Fix "INFO: rcu detected stall in hash_xxx" reports") Reported-by: Kota Toda <kota.toda@gmo-cybersecurity.com> Signed-off-by: Jozsef Kadlecsik <kadlec@netfilter.org> Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
As reported by Sebastian Andrzej Siewior the use of local_bh_disable()
is only feasible in uni processor systems to update the modification rules.
The usual use-case to update the modification rules is to update the data
of the modifications but not the modification types (AND/OR/XOR/SET) or
the checksum functions itself.
To omit additional memory allocations to maintain fast modification
switching times, the modification description space is doubled at gw-job
creation time so that only the reference to the active modification
description is changed under rcu protection.
Rename cgw_job::mod to cf_mod and make it a RCU pointer. Allocate in
cgw_create_job() and free it together with cgw_job in
cgw_job_free_rcu(). Update all users to dereference cgw_job::cf_mod with
a RCU accessor and if possible once.
[bigeasy: Replace mod1/mod2 from the Oliver's original patch with dynamic
allocation, use RCU annotation and accessor]
Reported-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Closes: https://lore.kernel.org/linux-can/20231031112349.y0aLoBrz@linutronix.de/ Fixes: dd895d7f21b2 ("can: cangw: introduce optional uid to reference created routing jobs") Tested-by: Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Link: https://patch.msgid.link/20250429070555.cs-7b_eZ@linutronix.de Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
The kvfree_rcu() and kfree_rcu() APIs are hazardous in that if you forget
the second argument, it works, but might sleep. This sleeping can be a
correctness bug from atomic contexts, and even in non-atomic contexts
it might introduce unacceptable latencies. This commit therefore adds
kvfree_rcu_mightsleep() and kfree_rcu_mightsleep(), which will replace
the single-argument kvfree_rcu() and kfree_rcu(), respectively.
This commit enables a series of commits that switch from single-argument
kvfree_rcu() and kfree_rcu() to their _mightsleep() counterparts. Once
all of these commits land, the single-argument versions will be removed.
Signed-off-by: Uladzislau Rezki (Sony) <urezki@gmail.com> Signed-off-by: Paul E. McKenney <paulmck@kernel.org>
Stable-dep-of: 511e64e13d8c ("can: gw: fix RCU/BH usage in cgw_create_job()") Signed-off-by: Sasha Levin <sashal@kernel.org>
Commit fb8696ab14ad ("can: gw: synchronize rcu operations
before removing gw job entry") added three synchronize_rcu() calls
to make sure one rcu grace period was observed before freeing
a "struct cgw_job" (which are tiny objects).
This should be converted to call_rcu() to avoid adding delays
in device / network dismantles.
Use the rcu_head that was already in struct cgw_job,
not yet used.
Link: https://lore.kernel.org/all/20220207190706.1499190-1-eric.dumazet@gmail.com Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Oliver Hartkopp <socketcan@hartkopp.net> Tested-by: Oliver Hartkopp <socketcan@hartkopp.net> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Stable-dep-of: 511e64e13d8c ("can: gw: fix RCU/BH usage in cgw_create_job()") Signed-off-by: Sasha Levin <sashal@kernel.org>
Use addrconf_addr_gen() to generate IPv6 link-local addresses on GRE
devices in most cases and fall back to using add_v4_addrs() only in
case the GRE configuration is incompatible with addrconf_addr_gen().
GRE used to use addrconf_addr_gen() until commit e5dd729460ca ("ip/ip6_gre:
use the same logic as SIT interfaces when computing v6LL address")
restricted this use to gretap and ip6gretap devices, and created
add_v4_addrs() (borrowed from SIT) for non-Ethernet GRE ones.
The original problem came when commit 9af28511be10 ("addrconf: refuse
isatap eui64 for INADDR_ANY") made __ipv6_isatap_ifid() fail when its
addr parameter was 0. The commit says that this would create an invalid
address, however, I couldn't find any RFC saying that the generated
interface identifier would be wrong. Anyway, since gre over IPv4
devices pass their local tunnel address to __ipv6_isatap_ifid(), that
commit broke their IPv6 link-local address generation when the local
address was unspecified.
Then commit e5dd729460ca ("ip/ip6_gre: use the same logic as SIT
interfaces when computing v6LL address") tried to fix that case by
defining add_v4_addrs() and calling it to generate the IPv6 link-local
address instead of using addrconf_addr_gen() (apart for gretap and
ip6gretap devices, which would still use the regular
addrconf_addr_gen(), since they have a MAC address).
That broke several use cases because add_v4_addrs() isn't properly
integrated into the rest of IPv6 Neighbor Discovery code. Several of
these shortcomings have been fixed over time, but add_v4_addrs()
remains broken on several aspects. In particular, it doesn't send any
Router Sollicitations, so the SLAAC process doesn't start until the
interface receives a Router Advertisement. Also, add_v4_addrs() mostly
ignores the address generation mode of the interface
(/proc/sys/net/ipv6/conf/*/addr_gen_mode), thus breaking the
IN6_ADDR_GEN_MODE_RANDOM and IN6_ADDR_GEN_MODE_STABLE_PRIVACY cases.
Fix the situation by using add_v4_addrs() only in the specific scenario
where the normal method would fail. That is, for interfaces that have
all of the following characteristics:
* run over IPv4,
* transport IP packets directly, not Ethernet (that is, not gretap
interfaces),
* tunnel endpoint is INADDR_ANY (that is, 0),
* device address generation mode is EUI64.
In all other cases, revert back to the regular addrconf_addr_gen().
Also, remove the special case for ip6gre interfaces in add_v4_addrs(),
since ip6gre devices now always use addrconf_addr_gen() instead.
Note:
This patch was originally applied as commit 183185a18ff9 ("gre: Fix
IPv6 link-local address generation."). However, it was then reverted
by commit fc486c2d060f ("Revert "gre: Fix IPv6 link-local address
generation."") because it uncovered another bug that ended up
breaking net/forwarding/ip6gre_custom_multipath_hash.sh. That other
bug has now been fixed by commit 4d0ab3a6885e ("ipv6: Start path
selection from the first nexthop"). Therefore we can now revive this
GRE patch (no changes since original commit 183185a18ff9 ("gre: Fix
IPv6 link-local address generation.").
Fixes: e5dd729460ca ("ip/ip6_gre: use the same logic as SIT interfaces when computing v6LL address") Signed-off-by: Guillaume Nault <gnault@redhat.com> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Link: https://patch.msgid.link/a88cc5c4811af36007645d610c95102dccb360a6.1746225214.git.gnault@redhat.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This patch replaces the manual Netlink attribute iteration in
output_userspace() with nla_for_each_nested(), which ensures that only
well-formed attributes are processed.
If a driver is removed, the driver framework invokes the driver's
remove callback. A CAN driver's remove function calls
unregister_candev(), which calls net_device_ops::ndo_stop further down
in the call stack for interfaces which are in the "up" state.
With the mcp251xfd driver the removal of the module causes the
following warning:
| WARNING: CPU: 0 PID: 352 at net/core/dev.c:7342 __netif_napi_del_locked+0xc8/0xd8
as can_rx_offload_del() deletes the NAPI, while it is still active,
because the interface is still up.
To fix the warning, first unregister the network interface, which
calls net_device_ops::ndo_stop, which disables the NAPI, and then call
can_rx_offload_del().
If a driver is removed, the driver framework invokes the driver's
remove callback. A CAN driver's remove function calls
unregister_candev(), which calls net_device_ops::ndo_stop further down
in the call stack for interfaces which are in the "up" state.
The removal of the module causes a warning, as can_rx_offload_del()
deletes the NAPI, while it is still active, because the interface is
still up.
To fix the warning, first unregister the network interface, which
calls net_device_ops::ndo_stop, which disables the NAPI, and then call
can_rx_offload_del().
Fixes: 1be37d3b0414 ("can: m_can: fix periph RX path: use rx-offload to ensure skbs are sent from softirq context") Cc: stable@vger.kernel.org Link: https://patch.msgid.link/20250502-can-rx-offload-del-v1-3-59a9b131589d@pengutronix.de Reviewed-by: Markus Schneider-Pargmann <msp@baylibre.com> Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In newer version of the SBC specs, we have a NDOB bit that indicates there
is no data buffer that gets written out. If this bit is set using commands
like "sg_write_same --ndob" we will crash in target_core_iblock/file's
execute_write_same handlers when we go to access the se_cmd->t_data_sg
because its NULL.
This patch adds a check for the NDOB bit in the common WRITE SAME code
because we don't support it. And, it adds a check for zero SG elements in
each handler in case the initiator tries to send a normal WRITE SAME with
no data buffer.
Link: https://lore.kernel.org/r/20220628022325.14627-2-michael.christie@oracle.com Reviewed-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Mike Christie <michael.christie@oracle.com> Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The blammed commit copied to argv the size of the reallocated argv,
instead of the size of the old_argv, thus reading and copying from
past the old_argv allocated memory.
ASPEED VGA card has two built-in devices:
0008:06:00.0 PCI bridge: ASPEED Technology, Inc. AST1150 PCI-to-PCI Bridge (rev 06)
0008:07:00.0 VGA compatible controller: ASPEED Technology, Inc. ASPEED Graphics Family (rev 52)
Its toplogy looks like this:
+-[0008:00]---00.0-[01-09]--+-00.0-[02-09]--+-00.0-[03]----00.0 Sandisk Corp Device 5017
| +-01.0-[04]--
| +-02.0-[05]----00.0 NVIDIA Corporation Device
| +-03.0-[06-07]----00.0-[07]----00.0 ASPEED Technology, Inc. ASPEED Graphics Family
| +-04.0-[08]----00.0 Renesas Technology Corp. uPD720201 USB 3.0 Host Controller
| \-05.0-[09]----00.0 Realtek Semiconductor Co., Ltd. RTL8111/8168/8411 PCI Express Gigabit Ethernet Controller
\-00.1 PMC-Sierra Inc. Device 4028
The IORT logic populaties two identical IDs into the fwspec->ids array via
DMA aliasing in iort_pci_iommu_init() called by pci_for_each_dma_alias().
Though the SMMU driver had been able to handle this situation since commit 563b5cbe334e ("iommu/arm-smmu-v3: Cope with duplicated Stream IDs"), that
got broken by the later commit cdf315f907d4 ("iommu/arm-smmu-v3: Maintain
a SID->device structure"), which ended up with allocating separate streams
with the same stuffing.
On a kernel prior to v6.15-rc1, there has been an overlooked warning:
pci 0008:07:00.0: vgaarb: setting as boot VGA device
pci 0008:07:00.0: vgaarb: bridge control possible
pci 0008:07:00.0: vgaarb: VGA device added: decodes=io+mem,owns=none,locks=none
pcieport 0008:06:00.0: Adding to iommu group 14
ast 0008:07:00.0: stream 67328 already in tree <===== WARNING
ast 0008:07:00.0: enabling device (0002 -> 0003)
ast 0008:07:00.0: Using default configuration
ast 0008:07:00.0: AST 2600 detected
ast 0008:07:00.0: [drm] Using analog VGA
ast 0008:07:00.0: [drm] dram MCLK=396 Mhz type=1 bus_width=16
[drm] Initialized ast 0.1.0 for 0008:07:00.0 on minor 0
ast 0008:07:00.0: [drm] fb0: astdrmfb frame buffer device
With v6.15-rc, since the commit bcb81ac6ae3c ("iommu: Get DT/ACPI parsing
into the proper probe path"), the error returned with the warning is moved
to the SMMU device probe flow:
arm_smmu_probe_device+0x15c/0x4c0
__iommu_probe_device+0x150/0x4f8
probe_iommu_group+0x44/0x80
bus_for_each_dev+0x7c/0x100
bus_iommu_probe+0x48/0x1a8
iommu_device_register+0xb8/0x178
arm_smmu_device_probe+0x1350/0x1db0
which then fails the entire SMMU driver probe:
pci 0008:06:00.0: Adding to iommu group 21
pci 0008:07:00.0: stream 67328 already in tree
arm-smmu-v3 arm-smmu-v3.9.auto: Failed to register iommu
arm-smmu-v3 arm-smmu-v3.9.auto: probe with driver arm-smmu-v3 failed with error -22
Since SMMU driver had been already expecting a potential duplicated Stream
ID in arm_smmu_install_ste_for_dev(), change the arm_smmu_insert_master()
routine to ignore a duplicated ID from the fwspec->sids array as well.
Note: this has been failing the iommu_device_probe() since 2021, although a
recent iommu commit in v6.15-rc1 that moves iommu_device_probe() started to
fail the SMMU driver probe. Since nobody has cared about DMA Alias support,
leave that as it was but fix the fundamental iommu_device_probe() breakage.
Fixes: cdf315f907d4 ("iommu/arm-smmu-v3: Maintain a SID->device structure") Cc: stable@vger.kernel.org Suggested-by: Jason Gunthorpe <jgg@nvidia.com> Reviewed-by: Jason Gunthorpe <jgg@nvidia.com> Signed-off-by: Nicolin Chen <nicolinc@nvidia.com> Link: https://lore.kernel.org/r/20250415185620.504299-1-nicolinc@nvidia.com Signed-off-by: Will Deacon <will@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Since v5.12 the rbtree has gained some simplifying helpers aimed at making
rb tree users write less convoluted boiler plate code. Instead the caller
provides a single comparison function and the helpers generate the prior
open-coded stuff.
Update smmu->streams to use rb_find_add() and rb_find().
Tested-by: Nicolin Chen <nicolinc@nvidia.com> Reviewed-by: Mostafa Saleh <smostafa@google.com> Signed-off-by: Jason Gunthorpe <jgg@nvidia.com> Link: https://lore.kernel.org/r/1-v3-9fef8cdc2ff6+150d1-smmuv3_tidy_jgg@nvidia.com Signed-off-by: Will Deacon <will@kernel.org>
Stable-dep-of: b00d24997a11 ("iommu/arm-smmu-v3: Fix iommu_device_probe bug due to duplicated stream ids") Signed-off-by: Sasha Levin <sashal@kernel.org>
With ACPI in place, gicv2m_get_fwnode() is registered with the pci
subsystem as pci_msi_get_fwnode_cb(), which may get invoked at runtime
during a PCI host bridge probe. But, the call back is wrongly marked as
__init, causing it to be freed, while being registered with the PCI
subsystem and could trigger:
Unable to handle kernel paging request at virtual address ffff8000816c0400
gicv2m_get_fwnode+0x0/0x58 (P)
pci_set_bus_msi_domain+0x74/0x88
pci_register_host_bridge+0x194/0x548
This is easily reproducible on a Juno board with ACPI boot.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Marc Zyngier <maz@kernel.org> Link: https://lore.kernel.org/r/20221121140048.534395323@linutronix.de
Stable-dep-of: 3318dc299b07 ("irqchip/gic-v2m: Prevent use after free of gicv2m_get_fwnode()") Signed-off-by: Sasha Levin <sashal@kernel.org>
The patch does permit the offending YUV420 @ 59.94 phy_freq and
vclk_freq mode to match in calculations. It also results in all
fractional rates being unavailable for use. This was unintended
and requires the patch to be reverted.
With lan88xx based devices the lan78xx driver can get stuck in an
interrupt loop while bringing the device up, flooding the kernel log
with messages like the following:
lan78xx 2-3:1.0 enp1s0u3: kevent 4 may have been dropped
Removing interrupt support from the lan88xx PHY driver forces the
driver to use polling instead, which avoids the problem.
The issue has been observed with Raspberry Pi devices at least since
4.14 (see [1], bug report for their downstream kernel), as well as
with Nvidia devices [2] in 2020, where disabling interrupts was the
vendor-suggested workaround (together with the claim that phylib
changes in 4.9 made the interrupt handling in lan78xx incompatible).
Iperf reports well over 900Mbits/sec per direction with client in
--dualtest mode, so there does not seem to be a significant impact on
throughput (lan88xx device connected via switch to the peer).
Commit c7e73b5051d6 ("ARM: imx: mach-imx6ul: remove 14x14 EVK specific
PHY fixup") removed a PHY fixup that setted the clock mode and the LED
mode.
Make the Ethernet interface work again by doing as advised in the
commit's log, set clock mode and the LED mode in the device tree.
Using device_find_child() to lookup the proper SCMI device to destroy
causes an unbalance in device refcount, since device_find_child() calls an
implicit get_device(): this, in turns, inhibits the call of the provided
release methods upon devices destruction.
As a consequence, one of the structures that is not freed properly upon
destruction is the internal struct device_private dev->p populated by the
drivers subsystem core.
KMemleak detects this situation since loading/unloding some SCMI driver
causes related devices to be created/destroyed without calling any
device_release method.
In externel_lb process, the hns3 driver call napi_disable()
first, then the reset happen, then the restore process of the
externel_lb will fail, and will not call napi_enable(). When
doing externel_lb again, napi_disable() will be double call,
cause a deadlock of rtnl_lock().
This patch use the HNS3_NIC_STATE_DOWN state to protect the
calling of napi_disable() and napi_enable() in externel_lb
process, just as the usage in ndo_stop() and ndo_start().
Fixes: 04b6ba143521 ("net: hns3: add support for external loopback test") Signed-off-by: Yonglong Liu <liuyonglong@huawei.com> Signed-off-by: Jijie Shao <shaojijie@huawei.com> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Link: https://lore.kernel.org/r/20230807113452.474224-5-shaojijie@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In of_modalias(), if the buffer happens to be too small even for the 1st
snprintf() call, the len parameter will become negative and str parameter
(if not NULL initially) will point beyond the buffer's end. Add the buffer
overflow check after the 1st snprintf() call and fix such check after the
strlen() call (accounting for the terminating NUL char).
Fixes: bc575064d688 ("of/device: use of_property_for_each_string to parse compatible strings") Signed-off-by: Sergey Shtylyov <s.shtylyov@omp.ru> Link: https://lore.kernel.org/r/bbfc6be0-c687-62b6-d015-5141b93f313e@omp.ru Signed-off-by: Rob Herring <robh@kernel.org> Signed-off-by: "Uwe Kleine-König" <ukleinek@debian.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The i.MX7D only has one PCIe controller, so controller_id should always be
0. The previous code is incorrect although yielding the correct result.
Fix by removing "IMX7D" from the switch case branch.
Fixes: 2d8ed461dbc9 ("PCI: imx6: Add support for i.MX8MQ") Link: https://lore.kernel.org/r/20241126075702.4099164-5-hongxing.zhu@nxp.com Signed-off-by: Richard Zhu <hongxing.zhu@nxp.com> Signed-off-by: Krzysztof Wilczyński <kwilczynski@kernel.org> Signed-off-by: Bjorn Helgaas <bhelgaas@google.com> Reviewed-by: Manivannan Sadhasivam <manivannan.sadhasivam@linaro.org> Reviewed-by: Frank Li <Frank.Li@nxp.com>
[Because this switch case does more than just controller_id
logic, move the "IMX7D" case label instead of removing it entirely.] Signed-off-by: Ryan Matthews <ryanmatthews@fastmail.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the ptp_clock_register() is called before relative
ptp resource ready. It may cause unexpected result when upper
layer called the ptp API during the timewindow. Fix it by
moving the ptp_clock_register() to the function end.
Fixes: 0bf5eb788512 ("net: hns3: add support for PTP") Signed-off-by: Jian Shen <shenjian15@huawei.com> Signed-off-by: Jijie Shao <shaojijie@huawei.com> Reviewed-by: Vadim Fedorenko <vadim.fedorenko@linux.dev> Link: https://patch.msgid.link/20250430093052.2400464-5-shaojijie@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
The size of the tm_qset file of debugfs is limited to 64 KB,
which is too small in the scenario with 1280 qsets.
The size needs to be expanded to 1 MB.
Fixes: 5e69ea7ee2a6 ("net: hns3: refactor the debugfs process") Signed-off-by: Hao Lan <lanhao@huawei.com> Signed-off-by: Peiyang Wang <wangpeiyang1@huawei.com> Signed-off-by: Jijie Shao <shaojijie@huawei.com> Link: https://patch.msgid.link/20250430093052.2400464-4-shaojijie@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
When a VF is passthrough to a VM, and the VM is killed, the reported
interrupt may not been handled, it will remain, and won't be clear by
the nic engine even with a flr or tqp reset. When the VM restart, the
interrupt of the first vector may be dropped by the second enable_irq
in vfio, see the issue below:
https://gitlab.com/qemu-project/qemu/-/issues/2884#note_2423361621
We notice that the vfio has always behaved this way, and the interrupt
is a residue of the nic engine, so we fix the problem by moving the
vector enable process out of the enable_irq loop.
This patch add support for external loopback test.
The successful test need the link is up with duplex full. The
driver do external loopback first, and then the whole offline
test.
Signed-off-by: Yonglong Liu <liuyonglong@huawei.com> Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com> Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Stable-dep-of: 8e6b9c6ea5a5 ("net: hns3: fix an interrupt residual problem") Signed-off-by: Sasha Levin <sashal@kernel.org>
The VF driver missed to store the rx VLAN tag strip state when
user change the rx VLAN tag offload state. And it will default
to enable the rx vlan tag strip when re-init VF device after
reset. So if user disable rx VLAN tag offload, and trig reset,
then the HW will still strip the VLAN tag from packet nad fill
into RX BD, but the VF driver will ignore it for rx VLAN tag
offload disabled. It may cause the rx VLAN tag dropped.
Fixes: b2641e2ad456 ("net: hns3: Add support of hardware rx-vlan-offload to HNS3 VF driver") Signed-off-by: Jian Shen <shenjian15@huawei.com> Signed-off-by: Jijie Shao <shaojijie@huawei.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20250430093052.2400464-2-shaojijie@huawei.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Activate TX hang workaround also in
fec_enet_txq_submit_skb() when TSO is not enabled.
Errata: ERR007885
Symptoms: NETDEV WATCHDOG: eth0 (fec): transmit queue 0 timed out
commit 37d6017b84f7 ("net: fec: Workaround for imx6sx enet tx hang when enable three queues")
There is a TDAR race condition for mutliQ when the software sets TDAR
and the UDMA clears TDAR simultaneously or in a small window (2-4 cycles).
This will cause the udma_tx and udma_tx_arbiter state machines to hang.
So, the Workaround is checking TDAR status four time, if TDAR cleared by
hardware and then write TDAR, otherwise don't set TDAR.
Fixes: 53bb20d1faba ("net: fec: add variable reg_desc_active to speed things up") Signed-off-by: Mattias Barthel <mattias.barthel@atlascopco.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Link: https://patch.msgid.link/20250429090826.3101258-1-mattiasbarthel@gmail.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Always map the `skb` to the LS descriptor. Previously skb was
mapped to EXT descriptor when the number of fragments is zero with
GSO enabled. Mapping the skb to EXT descriptor prevents it from
being freed, leading to a memory leak
Fixes: 23f0703c125b ("lan743x: Add main source files for new lan743x driver") Signed-off-by: Thangaraj Samynathan <thangaraj.s@microchip.com> Reviewed-by: Jacob Keller <jacob.e.keller@intel.com> Link: https://patch.msgid.link/20250429052527.10031-1-thangaraj.s@microchip.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
This patch addresses a data corruption issue observed in nvme-tcp during
testing.
In an NVMe native multipath setup, when an I/O timeout occurs, all
inflight I/Os are canceled almost immediately after the kernel socket is
shut down. These canceled I/Os are reported as host path errors,
triggering a failover that succeeds on a different path.
However, at this point, the original I/O may still be outstanding in the
host's network transmission path (e.g., the NIC’s TX queue). From the
user-space app's perspective, the buffer associated with the I/O is
considered completed since they're acked on the different path and may
be reused for new I/O requests.
Because nvme-tcp enables zero-copy by default in the transmission path,
this can lead to corrupted data being sent to the original target,
ultimately causing data corruption.
We can reproduce this data corruption by injecting delay on one path and
triggering i/o timeout.
To prevent this issue, this change ensures that all inflight
transmissions are fully completed from host's perspective before
returning from queue stop. To handle concurrent I/O timeout from multiple
namespaces under the same controller, always wait in queue stop
regardless of queue's state.
This aligns with the behavior of queue stopping in other NVMe fabric
transports.
Fixes: 3f2304f8c6d6 ("nvme-tcp: add NVMe over TCP host driver") Signed-off-by: Michael Liang <mliang@purestorage.com> Reviewed-by: Mohamed Khalfella <mkhalfella@purestorage.com> Reviewed-by: Randy Jennings <randyj@purestorage.com> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Christoph Hellwig <hch@lst.de> Signed-off-by: Sasha Levin <sashal@kernel.org>
For version 1 register dump that includes the PCIe stats, the existing
code incorrectly assumes that all PCIe stats are 64-bit values. Fix it
by using an array containing the starting and ending index of the 32-bit
values. The loop in bnxt_get_regs() will use the array to do proper
endian swap for the 32-bit values.
Fixes: b5d600b027eb ("bnxt_en: Add support for 'ethtool -d'") Reviewed-by: Shruti Parab <shruti.parab@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Reviewed-by: Andy Gospodarek <andrew.gospodarek@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
This happens when copying the coredump segment list in
bnxt_hwrm_dbg_dma_data() with the HWRM_DBG_COREDUMP_LIST FW command.
The info->dest_buf buffer is allocated based on the number of coredump
segments returned by the FW. The segment list is then DMA'ed by
the FW and the length of the DMA is returned by FW. The driver then
copies this DMA'ed segment list to info->dest_buf.
In some cases, this DMA length may exceed the info->dest_buf length
and cause the above BUG condition. Fix it by capping the copy
length to not exceed the length of info->dest_buf. The extra
DMA data contains no useful information.
This code path is shared for the HWRM_DBG_COREDUMP_LIST and the
HWRM_DBG_COREDUMP_RETRIEVE FW commands. The buffering is different
for these 2 FW commands. To simplify the logic, we need to move
the line to adjust the buffer length for HWRM_DBG_COREDUMP_RETRIEVE
up, so that the new check to cap the copy length will work for both
commands.
Fixes: c74751f4c392 ("bnxt_en: Return error if FW returns more data than dump length") Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Shruti Parab <shruti.parab@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
When handling HWRM_DBG_COREDUMP_LIST FW command in
bnxt_hwrm_dbg_dma_data(), the allocated buffer info->dest_buf is
not freed in the error path. In the normal path, info->dest_buf
is assigned to coredump->data and it will eventually be freed after
the coredump is collected.
Free info->dest_buf immediately inside bnxt_hwrm_dbg_dma_data() in
the error path.
Fixes: c74751f4c392 ("bnxt_en: Return error if FW returns more data than dump length") Reported-by: Michael Chan <michael.chan@broadcom.com> Reviewed-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com> Signed-off-by: Shruti Parab <shruti.parab@broadcom.com> Signed-off-by: Michael Chan <michael.chan@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net> Signed-off-by: Sasha Levin <sashal@kernel.org>
As it's name suggests, parse_eeprom() parses EEPROM data.
This is done by reading data, 16 bits at a time as follows:
for (i = 0; i < 128; i++)
((__le16 *) sromdata)[i] = cpu_to_le16(read_eeprom(np, i));
sromdata is at the same memory location as psrom.
And the type of psrom is a pointer to struct t_SROM.
As can be seen in the loop above, data is stored in sromdata, and thus psrom,
as 16-bit little-endian values.
However, the integer fields of t_SROM are host byte order integers.
And in the case of led_mode this leads to a little endian value
being incorrectly treated as host byte order.
Looking at rio_set_led_mode, this does appear to be a bug as that code
masks led_mode with 0x1, 0x2 and 0x8. Logic that would be effected by a
reversed byte order.
This problem would only manifest on big endian hosts.
Found by inspection while investigating a sparse warning
regarding the crc field of t_SROM.
I believe that warning is a false positive. And although I plan
to send a follow-up to use little-endian types for other the integer
fields of PSROM_t I do not believe that will involve any bug fixes.
As mentioned in the commit baeb705fd6a7 ("ice: always check VF VSI
pointer values"), we need to perform a null pointer check on the return
value of ice_get_vf_vsi() before using it.
Fixes: 6ebbe97a4881 ("ice: Add a per-VF limit on number of FDIR filters") Signed-off-by: Xuanqiang Luo <luoxuanqiang@kylinos.cn> Reviewed-by: Przemek Kitszel <przemyslaw.kitszel@intel.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com> Link: https://patch.msgid.link/20250425222636.3188441-3-anthony.l.nguyen@intel.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Some of the promiscuous mode functions take a boolean to indicate
set/clear, which affects readability. Refactor and provide an
interface for the promiscuous mode code with explicit set and clear
promiscuous mode operations.
Signed-off-by: Brett Creeley <brett.creeley@intel.com> Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com> Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Stable-dep-of: 425c5f266b2e ("ice: Check VF VSI Pointer Value in ice_vc_add_fdir_fltr()") Signed-off-by: Sasha Levin <sashal@kernel.org>
As described in Gerrard's report [1], there are use cases where a netem
child qdisc will make the parent qdisc's enqueue callback reentrant.
In the case of qfq, there won't be a UAF, but the code will add the same
classifier to the list twice, which will cause memory corruption.
This patch checks whether the class was already added to the agg->active
list (cl_is_active) before doing the addition to cater for the reentrant
case.
As described in Gerrard's report [1], there are use cases where a netem
child qdisc will make the parent qdisc's enqueue callback reentrant.
In the case of ets, there won't be a UAF, but the code will add the same
classifier to the list twice, which will cause memory corruption.
In addition to checking for qlen being zero, this patch checks whether
the class was already added to the active_list (cl_is_active) before
doing the addition to cater for the reentrant case.
As described in Gerrard's report [1], we have a UAF case when an hfsc class
has a netem child qdisc. The crux of the issue is that hfsc is assuming
that checking for cl->qdisc->q.qlen == 0 guarantees that it hasn't inserted
the class in the vttree or eltree (which is not true for the netem
duplicate case).
This patch checks the n_active class variable to make sure that the code
won't insert the class in the vttree or eltree twice, catering for the
reentrant case.
As described in Gerrard's report [1], there are use cases where a netem
child qdisc will make the parent qdisc's enqueue callback reentrant.
In the case of drr, there won't be a UAF, but the code will add the same
classifier to the list twice, which will cause memory corruption.
In addition to checking for qlen being zero, this patch checks whether the
class was already added to the active_list (cl_is_active) before adding
to the list to cover for the reentrant case.
In mtk_star_rx_poll function, on event processing completion, the
mtk_star_emac driver calls napi_complete_done but ignores its return
code and enable RX DMA interrupts inconditionally. This return code
gives the info if a device should avoid rearming its interrupts or not,
so fix this behaviour by taking it into account.
Use spin_lock_irqsave and spin_unlock_irqrestore instead of spin_lock
and spin_unlock in mtk_star_emac driver to avoid spinlock recursion
occurrence that can happen when enabling the DMA interrupts again in
rx/tx poll.
Current driver may lost tx interrupts under bidirectional test with iperf3,
which leads to some unexpected issues.
This patch let rx/tx interrupt enable/disable separately, and rx/tx are
handled in different NAPIs.
Signed-off-by: Biao Huang <biao.huang@mediatek.com> Signed-off-by: Yinghua Pan <ot_yinghua.pan@mediatek.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Stable-dep-of: e54b4db35e20 ("net: ethernet: mtk-star-emac: rearm interrupts in rx_poll only when advised") Signed-off-by: Sasha Levin <sashal@kernel.org>
The cited commit assumes enabling roce always succeeds. But it is
not true. Add error handling for it.
Fixes: 80f09dfc237f ("net/mlx5: Eswitch, enable RoCE loopback traffic") Signed-off-by: Chris Mi <cmi@nvidia.com> Reviewed-by: Roi Dayan <roid@nvidia.com> Reviewed-by: Maor Gottlieb <maorg@nvidia.com> Signed-off-by: Mark Bloch <mbloch@nvidia.com> Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com> Link: https://patch.msgid.link/20250423083611.324567-6-mbloch@nvidia.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org>
Initialize the source MAC address when creating the default GID entry.
Since this entry is used only for loopback traffic, it only needs to
be a unicast address. A zeroed-out MAC address is sufficient for this
purpose.
Without this fix, random bits would be assigned as the source address.
If these bits formed a multicast address, the firmware would return an
error, preventing the user from switching to switchdev mode:
If we're redirecting the skb, and haven't called tcf_mirred_forward(),
yet, we need to tell the core to drop the skb by setting the retcode
to SHOT. If we have called tcf_mirred_forward(), however, the skb
is out of our hands and returning SHOT will lead to UaF.
Move the retval override to the error path which actually need it.
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com> Fixes: e5cf1baf92cb ("act_mirred: use TC_ACT_REINSERT when possible") Signed-off-by: Jakub Kicinski <kuba@kernel.org> Acked-by: Jamal Hadi Salim <jhs@mojatatu.com> Signed-off-by: David S. Miller <davem@davemloft.net>
[Minor conflict resolved due to code context change.] Signed-off-by: Jianqi Ren <jianqi.ren.cn@windriver.com> Signed-off-by: He Zhe <zhe.he@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move the conditional loading of hardware DR6 with the guest's DR6 value
out of the core .vcpu_run() loop to fix a bug where KVM can load hardware
with a stale vcpu->arch.dr6.
When the guest accesses a DR and host userspace isn't debugging the guest,
KVM disables DR interception and loads the guest's values into hardware on
VM-Enter and saves them on VM-Exit. This allows the guest to access DRs
at will, e.g. so that a sequence of DR accesses to configure a breakpoint
only generates one VM-Exit.
For DR0-DR3, the logic/behavior is identical between VMX and SVM, and also
identical between KVM_DEBUGREG_BP_ENABLED (userspace debugging the guest)
and KVM_DEBUGREG_WONT_EXIT (guest using DRs), and so KVM handles loading
DR0-DR3 in common code, _outside_ of the core kvm_x86_ops.vcpu_run() loop.
But for DR6, the guest's value doesn't need to be loaded into hardware for
KVM_DEBUGREG_BP_ENABLED, and SVM provides a dedicated VMCB field whereas
VMX requires software to manually load the guest value, and so loading the
guest's value into DR6 is handled by {svm,vmx}_vcpu_run(), i.e. is done
_inside_ the core run loop.
Unfortunately, saving the guest values on VM-Exit is initiated by common
x86, again outside of the core run loop. If the guest modifies DR6 (in
hardware, when DR interception is disabled), and then the next VM-Exit is
a fastpath VM-Exit, KVM will reload hardware DR6 with vcpu->arch.dr6 and
clobber the guest's actual value.
The bug shows up primarily with nested VMX because KVM handles the VMX
preemption timer in the fastpath, and the window between hardware DR6
being modified (in guest context) and DR6 being read by guest software is
orders of magnitude larger in a nested setup. E.g. in non-nested, the
VMX preemption timer would need to fire precisely between #DB injection
and the #DB handler's read of DR6, whereas with a KVM-on-KVM setup, the
window where hardware DR6 is "dirty" extends all the way from L1 writing
DR6 to VMRESUME (in L1).
L1's view:
==========
<L1 disables DR interception>
CPU 0/KVM-7289 [023] d.... 2925.640961: kvm_entry: vcpu 0
A: L1 Writes DR6
CPU 0/KVM-7289 [023] d.... 2925.640963: <hack>: Set DRs, DR6 = 0xffff0ff1
Reported-by: John Stultz <jstultz@google.com> Closes: https://lkml.kernel.org/r/CANDhNCq5_F3HfFYABqFGCA1bPd_%2BxgNj-iDQhH4tDk%2Bwi8iZZg%40mail.gmail.com Fixes: 375e28ffc0cf ("KVM: X86: Set host DR6 only on VMX and for KVM_DEBUGREG_WONT_EXIT") Fixes: d67668e9dd76 ("KVM: x86, SVM: isolate vcpu->arch.dr6 from vmcb->save.dr6") Cc: stable@vger.kernel.org Cc: Jim Mattson <jmattson@google.com> Tested-by: John Stultz <jstultz@google.com> Link: https://lore.kernel.org/r/20250125011833.3644371-1-seanjc@google.com Signed-off-by: Sean Christopherson <seanjc@google.com>
[jth: Handled conflicts with kvm_x86_ops reshuffle] Signed-off-by: James Houghton <jthoughton@google.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
syzbot reported this bug:
==================================================================
BUG: KASAN: slab-out-of-bounds in trace_seq_to_buffer kernel/trace/trace.c:1830 [inline]
BUG: KASAN: slab-out-of-bounds in tracing_splice_read_pipe+0x6be/0xdd0 kernel/trace/trace.c:6822
Write of size 4507 at addr ffff888032b6b000 by task syz.2.320/7260
It has been reported that trace_seq_to_buffer() tries to copy more data
than PAGE_SIZE to buf. Therefore, to prevent this, we should use the
smaller of trace_seq_used(&iter->seq) and PAGE_SIZE as an argument.
Link: https://lore.kernel.org/20250422113026.13308-1-aha310510@gmail.com Reported-by: syzbot+c8cd2d2c412b868263fb@syzkaller.appspotmail.com Fixes: 3c56819b14b0 ("tracing: splice support for tracing_pipe") Suggested-by: Steven Rostedt <rostedt@goodmis.org> Signed-off-by: Jeongjun Park <aha310510@gmail.com> Signed-off-by: Steven Rostedt (Google) <rostedt@goodmis.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
On the Lenovo ThinkPad X201, when Intel VT-d is enabled in the BIOS, the
kernel boots with errors related to DMAR, the graphical interface appeared
quite choppy, and the system resets erratically within a minute after it
booted:
DMAR: DRHD: handling fault status reg 3
DMAR: [DMA Write NO_PASID] Request device [00:02.0] fault addr 0xb97ff000
[fault reason 0x05] PTE Write access is not set
Upon comparing boot logs with VT-d on/off, I found that the Intel Calpella
quirk (`quirk_calpella_no_shadow_gtt()') correctly applied the igfx IOMMU
disable/quirk correctly:
pci 0000:00:00.0: DMAR: BIOS has allocated no shadow GTT; disabling IOMMU
for graphics
Whereas with VT-d on, it went into the "else" branch, which then
triggered the DMAR handling fault above:
... else if (!disable_igfx_iommu) {
/* we have to ensure the gfx device is idle before we flush */
pci_info(dev, "Disabling batched IOTLB flush on Ironlake\n");
iommu_set_dma_strict();
}
Now, this is not exactly scientific, but moving 0x0044 to quirk_iommu_igfx
seems to have fixed the aforementioned issue. Running a few `git blame'
runs on the function, I have found that the quirk was originally
introduced as a fix specific to ThinkPad X201:
commit 9eecabcb9a92 ("intel-iommu: Abort IOMMU setup for igfx if BIOS gave
no shadow GTT space")
Which was later revised twice to the "else" branch we saw above:
- 2011: commit 6fbcfb3e467a ("intel-iommu: Workaround IOTLB hang on
Ironlake GPU")
- 2024: commit ba00196ca41c ("iommu/vt-d: Decouple igfx_off from graphic
identity mapping")
I'm uncertain whether further testings on this particular laptops were
done in 2011 and (honestly I'm not sure) 2024, but I would be happy to do
some distro-specific testing if that's what would be required to verify
this patch.
P.S., I also see IDs 0x0040, 0x0062, and 0x006a listed under the same
`quirk_calpella_no_shadow_gtt()' quirk, but I'm not sure how similar these
chipsets are (if they share the same issue with VT-d or even, indeed, if
this issue is specific to a bug in the Lenovo BIOS). With regards to
0x0062, it seems to be a Centrino wireless card, but not a chipset?
I have also listed a couple (distro and kernel) bug reports below as
references (some of them are from 7-8 years ago!), as they seem to be
similar issue found on different Westmere/Ironlake, Haswell, and Broadwell
hardware setups.
There is a string parsing logic error which can lead to an overflow of hid
or uid buffers. Comparing ACPIID_LEN against a total string length doesn't
take into account the lengths of individual hid and uid buffers so the
check is insufficient in some cases. For example if the length of hid
string is 4 and the length of the uid string is 260, the length of str
will be equal to ACPIID_LEN + 1 but uid string will overflow uid buffer
which size is 256.
The same applies to the hid string with length 13 and uid string with
length 250.
Check the length of hid and uid strings separately to prevent
buffer overflow.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
realloc_argv() was only updating the array size if it was called with
old_argv already allocated. The first time it was called to create an
argv array, it would allocate the array but return the array size as
zero. dm_split_args() would think that it couldn't store any arguments
in the array and would call realloc_argv() again, causing it to
reallocate the initial slots (this time using GPF_KERNEL) and finally
return a size. Aside from being wasteful, this could cause deadlocks on
targets that need to process messages without starting new IO. Instead,
realloc_argv should always update the allocated array size on success.
Fixes: a0651926553c ("dm table: don't copy from a NULL pointer in realloc_argv()") Cc: stable@vger.kernel.org Signed-off-by: Benjamin Marzinski <bmarzins@redhat.com> Signed-off-by: Mikulas Patocka <mpatocka@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The function brcmf_usb_dl_writeimage() calls the function
brcmf_usb_dl_cmd() but dose not check its return value. The
'state.state' and the 'state.bytes' are uninitialized if the
function brcmf_usb_dl_cmd() fails. It is dangerous to use
uninitialized variables in the conditions.
Add error handling for brcmf_usb_dl_cmd() to jump to error
handling path if the brcmf_usb_dl_cmd() fails and the
'state.state' and the 'state.bytes' are uninitialized.
Improve the error message to report more detailed error
information.
Fixes: 71bb244ba2fd ("brcm80211: fmac: add USB support for bcm43235/6/8 chipsets") Cc: stable@vger.kernel.org # v3.4+ Signed-off-by: Wentao Liang <vulab@iscas.ac.cn> Acked-by: Arend van Spriel <arend.vanspriel@broadcom.com> Link: https://patch.msgid.link/20250422042203.2259-1-vulab@iscas.ac.cn Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
According to the XGMAC specification, enabling features such as Layer 3
and Layer 4 Packet Filtering, Split Header and Virtualized Network support
automatically selects the IPC Full Checksum Offload Engine on the receive
side.
When RX checksum offload is disabled, these dependent features must also
be disabled to prevent abnormal behavior caused by mismatched feature
dependencies.
Ensure that toggling RX checksum offload (disabling or enabling) properly
disables or enables all dependent features, maintaining consistent and
expected behavior in the network device.
Cc: stable@vger.kernel.org Fixes: 1a510ccf5869 ("amd-xgbe: Add support for VXLAN offload capabilities") Signed-off-by: Vishal Badole <Vishal.Badole@amd.com> Reviewed-by: Simon Horman <horms@kernel.org> Link: https://patch.msgid.link/20250424130248.428865-1-Vishal.Badole@amd.com Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Camm noticed that on parisc a SIGFPE exception will crash an application with
a second SIGFPE in the signal handler. Dave analyzed it, and it happens
because glibc uses a double-word floating-point store to atomically update
function descriptors. As a result of lazy binding, we hit a floating-point
store in fpe_func almost immediately.
When the T bit is set, an assist exception trap occurs when when the
co-processor encounters *any* floating-point instruction except for a double
store of register %fr0. The latter cancels all pending traps. Let's fix this
by clearing the Trap (T) bit in the FP status register before returning to the
signal handler in userspace.
The issue can be reproduced with this test program:
Commit a5951389e58d ("arm64: errata: Add newer ARM cores to the
spectre_bhb_loop_affected() lists") added some additional CPUs to the
Spectre-BHB workaround, including some new arrays for designs that
require new 'k' values for the workaround to be effective.
Unfortunately, the new arrays omitted the sentinel entry and so
is_midr_in_range_list() will walk off the end when it doesn't find a
match. With UBSAN enabled, this leads to a crash during boot when
is_midr_in_range_list() is inlined (which was more common prior to c8c2647e69be ("arm64: Make _midr_in_range_list() an exported
function")):
Cc: Lee Jones <lee@kernel.org> Cc: James Morse <james.morse@arm.com> Cc: Doug Anderson <dianders@chromium.org> Cc: Shameer Kolothum <shameerali.kolothum.thodi@huawei.com> Cc: <stable@vger.kernel.org> Reported-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Fixes: a5951389e58d ("arm64: errata: Add newer ARM cores to the spectre_bhb_loop_affected() lists") Signed-off-by: Will Deacon <will@kernel.org> Reviewed-by: Lee Jones <lee@kernel.org> Reviewed-by: Douglas Anderson <dianders@chromium.org> Reviewed-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Link: https://lore.kernel.org/r/20250501104747.28431-1-will@kernel.org Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Deferred probe with pm_runtime_put() may delay clock disable, causing
incorrect clock usage count. Use pm_runtime_put_sync() to ensure the
clock is disabled immediately.
Fixes: 13d6eb20fc79 ("i2c: imx-lpi2c: add runtime pm support") Signed-off-by: Clark Wang <xiaoning.wang@nxp.com> Signed-off-by: Carlos Song <carlos.song@nxp.com> Cc: <stable@vger.kernel.org> # v4.16+ Link: https://lore.kernel.org/r/20250421062341.2471922-1-carlos.song@nxp.com Signed-off-by: Andi Shyti <andi.shyti@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Nouveau is mostly designed in a way that it's expected that fences only
ever get signaled through nouveau_fence_signal(). However, in at least
one other place, nouveau_fence_done(), can signal fences, too. If that
happens (race) a signaled fence remains in the pending list for a while,
until it gets removed by nouveau_fence_update().
Should nouveau_fence_context_kill() run in the meantime, this would be
a bug because the function would attempt to set an error code on an
already signaled fence.
Have nouveau_fence_context_kill() check for a fence being signaled.
Cc: stable@vger.kernel.org # v5.10+ Fixes: ea13e5abf807 ("drm/nouveau: signal pending fences when channel has been killed") Suggested-by: Christian König <christian.koenig@amd.com> Signed-off-by: Philipp Stanner <phasta@kernel.org> Link: https://lore.kernel.org/r/20250415121900.55719-3-phasta@kernel.org Signed-off-by: Danilo Krummrich <dakr@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
There seem to be multiple USB device IDs used for these;
the one I have reports as 0b0e:030c when powered on.
(When powered off, it reports as 0b0e:0311.)
When contiguous windows are coalesced by pci_register_host_bridge(), the
second resource is expanded to include the first, and the first is
invalidated and consequently not added to the bus. However, it remains in
the resource hierarchy. For example, these windows:
In some cases (e.g. the Xen scratch region), this causes future calls to
allocate_resource() to choose an inappropriate location which the caller
cannot handle.
Fix by releasing the zeroed-out resource and removing it from the resource
hierarchy.
When coalescing two resources for a contiguous aperture, the second
resource is enlarged to cover the full contiguous range, while the first
resource is marked invalid. This invalidation is done by clearing the
flags, start, and end members.
When adding the initial resources to the bus later, invalid resources are
skipped. Unfortunately, the check for an invalid resource considers only
the end member, causing false positives.
E.g. on r8a7791/koelsch, root bus resource 0 ("bus 00") is skipped, and no
longer registered with pci_bus_insert_busn_res() (causing the memory leak),
nor printed:
pci-rcar-gen2 ee090000.pci: host bridge /soc/pci@ee090000 ranges:
pci-rcar-gen2 ee090000.pci: MEM 0x00ee080000..0x00ee08ffff -> 0x00ee080000
pci-rcar-gen2 ee090000.pci: PCI: revision 11
pci-rcar-gen2 ee090000.pci: PCI host bridge to bus 0000:00
-pci_bus 0000:00: root bus resource [bus 00]
pci_bus 0000:00: root bus resource [mem 0xee080000-0xee08ffff]
Fix this by only skipping resources where all of the flags, start, and end
members are zero.
Commit c14f7ccc9f5d ("PCI: Assign PCI domain IDs by ida_alloc()")
introduced a use-after-free bug in the bus removal cleanup. The issue was
found with kfence:
[ 19.293351] BUG: KFENCE: use-after-free read in pci_bus_release_domain_nr+0x10/0x70
The stack trace is a bit misleading as dw_pcie_host_deinit() doesn't
directly call pci_bus_release_domain_nr(). The issue turns out to be in
pci_remove_root_bus() which first calls pci_remove_bus() which frees the
struct pci_bus when its struct device is released. Then
pci_bus_release_domain_nr() is called and accesses the freed struct
pci_bus. Reordering these fixes the issue.
Commit 62baf70c3274 caused the ANA log page to be re-read, even on
controllers that do not support ANA. While this should generally
harmless, some controllers hang on the unsupported log page and
never finish probing.
Fixes: 62baf70c3274 ("nvme: re-read ANA log page after ns scan completes") Signed-off-by: Hannes Reinecke <hare@kernel.org> Tested-by: Srikanth Aithal <sraithal@amd.com>
[hch: more detailed commit message] Signed-off-by: Christoph Hellwig <hch@lst.de> Reviewed-by: Sagi Grimberg <sagi@grimberg.me> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
xdp: Reset bpf_redirect_info before running a xdp's BPF prog.
Ricardo reported a KASAN discovered use after free in v6.6-stable.
The syzbot starts a BPF program via xdp_test_run_batch() which assigns
ri->tgt_value via dev_hash_map_redirect() and the return code isn't
XDP_REDIRECT it looks like nonsense. So the output in
bpf_warn_invalid_xdp_action() appears once.
Then the TUN driver runs another BPF program (on the same CPU) which
returns XDP_REDIRECT without setting ri->tgt_value first. It invokes
bpf_trace_printk() to print four characters and obtain the required
return value. This is enough to get xdp_do_redirect() invoked which
then accesses the pointer in tgt_value which might have been already
deallocated.
This problem does not affect upstream because since commit 401cb7dae8130 ("net: Reference bpf_redirect_info via task_struct on PREEMPT_RT.")
the per-CPU variable is referenced via task's task_struct and exists on
the stack during NAPI callback. Therefore it is cleared once before the
first invocation and remains valid within the RCU section of the NAPI
callback.
Instead of performing the huge backport of the commit (plus its fix ups)
here is an alternative version which only resets the variable in
question prior invoking the BPF program.
Flexible endpoints use DIGs from available inflexible endpoints,
so only the encoders of inflexible links need to be freed.
Otherwise, a double free issue may occur when unloading the
amdgpu module.
Reviewed-by: Rodrigo Siqueira <rodrigo.siqueira@amd.com> Signed-off-by: Tim Huang <tim.huang@amd.com> Reviewed-by: Roman Li <roman.li@amd.com> Signed-off-by: Roman Li <roman.li@amd.com> Tested-by: Daniel Wheeler <daniel.wheeler@amd.com> Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[ dc_link_destruct() moved from core/dc_link.c to link/link_factory.c since
commit: 54618888d1ea ("drm/amd/display: break down dc_link.c"), so modified
the path to apply on 5.15.y ] Signed-off-by: Xiangyu Chen <xiangyu.chen@windriver.com> Signed-off-by: He Zhe <zhe.he@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
When `jr3_pci_detach()` is called during device removal, it calls
`timer_delete_sync()` to stop the timer, but the timer expiry function
always reschedules the timer, so the synchronization is ineffective.
Call `timer_shutdown_sync()` instead. It does not matter that the timer
expiry function pointer is cleared, because the device is being removed.
projects / thirdparty / kernel / stable.git / log
