Whether VLAN-aware or not, on every VID VLAN table entry that has the CPU
port as a member of it, frames are set to egress the CPU port with the VLAN
tag stacked. This is so that VLAN tags can be appended after hardware
special tag (called DSA tag in the context of Linux drivers).
For user ports on a VLAN-unaware bridge, frame ingressing the user port
egresses CPU port with only the special tag.
For user ports on a VLAN-aware bridge, frame ingressing the user port
egresses CPU port with the special tag and the VLAN tag.
This causes issues with link-local frames, specifically BPDUs, because the
software expects to receive them VLAN-untagged.
There are two options to make link-local frames egress untagged. Setting
CONSISTENT or UNTAGGED on the EG_TAG bits on the relevant register.
CONSISTENT means frames egress exactly as they ingress. That means
egressing with the VLAN tag they had at ingress or egressing untagged if
they ingressed untagged. Although link-local frames are not supposed to be
transmitted VLAN-tagged, if they are done so, when egressing through a CPU
port, the special tag field will be broken.
BPDU egresses CPU port with VLAN tag egressing stacked, received on
software:
To prevent confusing the software, force the frame to egress UNTAGGED
instead of CONSISTENT. This way, frames can't possibly be received TAGGED
by software which would have the special tag field broken.
VLAN Tag Egress Procedure
For all frames, one of these options set the earliest in this order will
apply to the frame:
- EG_TAG in certain registers for certain frames.
This will apply to frame with matching MAC DA or EtherType.
- EG_TAG in the address table.
This will apply to frame at its incoming port.
- EG_TAG in the PVC register.
This will apply to frame at its incoming port.
- EG_CON and [EG_TAG per port] in the VLAN table.
This will apply to frame at its outgoing port.
- EG_TAG in the PCR register.
This will apply to frame at its outgoing port.
EG_TAG in certain registers for certain frames:
PPPoE Discovery_ARP/RARP: PPP_EG_TAG and ARP_EG_TAG in the APC register.
IGMP_MLD: IGMP_EG_TAG and MLD_EG_TAG in the IMC register.
BPDU and PAE: BPDU_EG_TAG and PAE_EG_TAG in the BPC register.
REV_01 and REV_02: R01_EG_TAG and R02_EG_TAG in the RGAC1 register.
REV_03 and REV_0E: R03_EG_TAG and R0E_EG_TAG in the RGAC2 register.
REV_10 and REV_20: R10_EG_TAG and R20_EG_TAG in the RGAC3 register.
REV_21 and REV_UN: R21_EG_TAG and RUN_EG_TAG in the RGAC4 register.
With this change, it can be observed that a bridge interface with stp_state
and vlan_filtering enabled will properly block ports now.
Fixes: b8f126a8d543 ("net-next: dsa: add dsa support for Mediatek MT7530 switch") Signed-off-by: Arınç ÜNAL <[email protected]> Signed-off-by: Paolo Abeni <[email protected]>
====================
Report RCU QS for busy network kthreads
This changeset fixes a common problem for busy networking kthreads.
These threads, e.g. NAPI threads, typically will do:
* polling a batch of packets
* if there are more work, call cond_resched() to allow scheduling
* continue to poll more packets when rx queue is not empty
We observed this being a problem in production, since it can block RCU
tasks from making progress under heavy load. Investigation indicates
that just calling cond_resched() is insufficient for RCU tasks to reach
quiescent states. This also has the side effect of frequently clearing
the TIF_NEED_RESCHED flag on voluntary preempt kernels. As a result,
schedule() will not be called in these circumstances, despite schedule()
in fact provides required quiescent states. This at least affects NAPI
threads, napi_busy_loop, and also cpumap kthread.
By reporting RCU QSes in these kthreads periodically before cond_resched, the
blocked RCU waiters can correctly progress. Instead of just reporting QS for
RCU tasks, these code share the same concern as noted in the commit d28139c4e967 ("rcu: Apply RCU-bh QSes to RCU-sched and RCU-preempt when safe").
So report a consolidated QS for safety.
It is worth noting that, although this problem is reproducible in
napi_busy_loop, it only shows up when setting the polling interval to as high
as 2ms, which is far larger than recommended 50us-100us in the documentation.
So napi_busy_loop is left untouched.
Lastly, this does not affect RT kernels, which does not enter the scheduler
through cond_resched(). Without the mentioned side effect, schedule() will
be called time by time, and clear the RCU task holdouts.
Yan Zhai [Tue, 19 Mar 2024 20:44:40 +0000 (13:44 -0700)]
bpf: report RCU QS in cpumap kthread
When there are heavy load, cpumap kernel threads can be busy polling
packets from redirect queues and block out RCU tasks from reaching
quiescent states. It is insufficient to just call cond_resched() in such
context. Periodically raise a consolidated RCU QS before cond_resched
fixes the problem.
Yan Zhai [Tue, 19 Mar 2024 20:44:37 +0000 (13:44 -0700)]
net: report RCU QS on threaded NAPI repolling
NAPI threads can keep polling packets under load. Currently it is only
calling cond_resched() before repolling, but it is not sufficient to
clear out the holdout of RCU tasks, which prevent BPF tracing programs
from detaching for long period. This can be reproduced easily with
following set up:
ip netns add test1
ip netns add test2
ip -n test1 link add veth1 type veth peer name veth2 netns test2
ip -n test1 link set veth1 up
ip -n test1 link set lo up
ip -n test2 link set veth2 up
ip -n test2 link set lo up
ip -n test1 addr add 192.168.1.2/31 dev veth1
ip -n test1 addr add 1.1.1.1/32 dev lo
ip -n test2 addr add 192.168.1.3/31 dev veth2
ip -n test2 addr add 2.2.2.2/31 dev lo
ip -n test1 route add default via 192.168.1.3
ip -n test2 route add default via 192.168.1.2
for i in `seq 10 210`; do
for j in `seq 10 210`; do
ip netns exec test2 iptables -I INPUT -s 3.3.$i.$j -p udp --dport 5201
done
done
ip netns exec test2 ethtool -K veth2 gro on
ip netns exec test2 bash -c 'echo 1 > /sys/class/net/veth2/threaded'
ip netns exec test1 ethtool -K veth1 tso off
Then run an iperf3 client/server and a bpftrace script can trigger it:
Yan Zhai [Tue, 19 Mar 2024 20:44:34 +0000 (13:44 -0700)]
rcu: add a helper to report consolidated flavor QS
When under heavy load, network processing can run CPU-bound for many
tens of seconds. Even in preemptible kernels (non-RT kernel), this can
block RCU Tasks grace periods, which can cause trace-event removal to
take more than a minute, which is unacceptably long.
This commit therefore creates a new helper function that passes through
both RCU and RCU-Tasks quiescent states every 100 milliseconds. This
hard-coded value suffices for current workloads.
Herve Codina [Thu, 14 Mar 2024 12:00:06 +0000 (13:00 +0100)]
lib/bitmap: Fix bitmap_scatter() and bitmap_gather() kernel doc
The make htmldoc command failed with the following error
... include/linux/bitmap.h:524: ERROR: Unexpected indentation.
... include/linux/bitmap.h:524: CRITICAL: Unexpected section title or transition.
Move the visual representation to a literal block.
Linus Torvalds [Wed, 20 Mar 2024 23:42:47 +0000 (16:42 -0700)]
Merge tag 'v6.9-rc-smb3-server-fixes' of git://git.samba.org/ksmbd
Pull smb server updates from Steve French:
- add support for durable file handles (an important data integrity
feature)
- fixes for potential out of bounds issues
- fix possible null dereference in close
- getattr fixes
- trivial typo fix and minor cleanup
* tag 'v6.9-rc-smb3-server-fixes' of git://git.samba.org/ksmbd:
ksmbd: remove module version
ksmbd: fix potencial out-of-bounds when buffer offset is invalid
ksmbd: fix slab-out-of-bounds in smb_strndup_from_utf16()
ksmbd: Fix spelling mistake "connction" -> "connection"
ksmbd: fix possible null-deref in smb_lazy_parent_lease_break_close
ksmbd: add support for durable handles v1/v2
ksmbd: mark SMB2_SESSION_EXPIRED to session when destroying previous session
ksmbd: retrieve number of blocks using vfs_getattr in set_file_allocation_info
ksmbd: replace generic_fillattr with vfs_getattr
Linus Torvalds [Wed, 20 Mar 2024 23:37:07 +0000 (16:37 -0700)]
Merge tag 'trace-tools-v6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace
Pull trace tool updates from Steven Rostedt:
"Tracing:
- Update makefiles for latency-collector and RTLA, using tools/build/
makefiles like perf does, inheriting its benefits. For example,
having a proper way to handle library dependencies.
- The timerlat tracer has an interface for any tool to use. rtla
timerlat tool uses this interface dispatching its own threads as
workload. But, rtla timerlat could also be used for any other
process. So, add 'rtla timerlat -U' option, allowing the timerlat
tool to measure the latency of any task using the timerlat tracer
interface.
Verification:
- Update makefiles for verification/rv, using tools/build/ makefiles
like perf does, inheriting its benefits. For example, having a
proper way to handle dependencies"
* tag 'trace-tools-v6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
tools/rtla: Add -U/--user-load option to timerlat
tools/verification: Use tools/build makefiles on rv
tools/rtla: Use tools/build makefiles to build rtla
tools/tracing: Use tools/build makefiles on latency-collector
Sami Tolvanen [Mon, 11 Mar 2024 19:31:44 +0000 (19:31 +0000)]
riscv: Fix syscall wrapper for >word-size arguments
The current syscall wrapper macros break 64-bit arguments on
rv32 because they only guarantee the first N input registers are
passed to syscalls that accept N arguments. According to the
calling convention, values twice the word size reside in register
pairs and as a result, syscall arguments don't always have a
direct register mapping on rv32.
Instead of using `__MAP(x,__SC_LONG,__VA_ARGS__)` to declare the
type of the `__se(_compat)_sys_*` functions on rv32, change the
function declarations to accept `ulong` arguments and alias them
to the actual syscall implementations, similarly to the existing
macros in include/linux/syscalls.h. This matches previous
behavior and ensures registers are passed to syscalls as-is, no
matter which argument types they expect.
Masahiro Yamada [Wed, 20 Mar 2024 16:52:11 +0000 (01:52 +0900)]
kconfig: tests: test dependency after shuffling choices
Commit c8fb7d7e48d1 ("kconfig: fix broken dependency in randconfig-
generated .config") fixed the issue, but I did not add a test case.
This commit adds a test case that emulates the reported situation.
The test would fail without c8fb7d7e48d1.
To handle the choice "choose X", FOO must be calculated beforehand.
FOO depends on A, which is a member of another choice "choose A or B".
Kconfig _temporarily_ assumes the value of A to proceed. The choice
"choose A or B" will be shuffled later, but the result may or may not
meet "FOO depends on A". Kconfig should invalidate the symbol values
and recompute them.
In the real example for ARCH=arm64, the choice "Instrumentation type"
needs the value of CPU_BIG_ENDIAN. The choice "Endianness" will be
shuffled later.
Masahiro Yamada [Wed, 20 Mar 2024 16:52:10 +0000 (01:52 +0900)]
kconfig: tests: add a test for randconfig with dependent choices
Since commit 3b9a19e08960 ("kconfig: loop as long as we changed some
symbols in randconfig"), conf_set_all_new_symbols() is repeated until
there is no more choice left to be shuffled. The motivation was to
shuffle a choice nested in another choice.
Although commit 09d5873e4d1f ("kconfig: allow only 'config', 'comment',
and 'if' inside 'choice'") disallowed the nested choice structure,
we must still keep 3b9a19e08960 because there are still cases where
conf_set_all_new_symbols() must iterate.
scripts/kconfig/tests/choice_randomize/Kconfig is the test case.
The second choice depends on 'B', which is the member of the first
choice.
With 3b9a19e08960 reverted, we would never get the pattern specified by
scripts/kconfig/tests/choice_randomize/expected_config2.
A real example can be found in lib/Kconfig.debug. Without 3b9a19e08960,
the randconfig would not shuffle the "Compressed Debug information"
choice, which depends on DEBUG_INFO, which is derived from another
choice "Debug information".
My goal is to refactor Kconfig so that randconfig will work more
simply, without using the loop.
For now, let's add a test case to ensure all dependent choices are
shuffled, as it is a somewhat tricky case for the current Kconfig.
Le Ma [Fri, 15 Mar 2024 08:55:39 +0000 (16:55 +0800)]
drm/amdgpu: drop setting buffer funcs in sdma442
To fix the entity rq NULL issue. This setting has been moved
to upper level.
Fixes: b70438004a14 ("drm/amdgpu: move buffer funcs setting up a level") Signed-off-by: Le Ma <[email protected]> Reviewed-by: Hawking Zhang <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
Wenjing Liu [Mon, 4 Mar 2024 16:20:27 +0000 (11:20 -0500)]
drm/amd/display: Revert Remove pixle rate limit for subvp
This reverts commit 340383c734f8 ("drm/amd/display: Remove pixle rate
limit for subvp")
[why]
The original commit causes a regression when subvp is applied
on ODM required 8k60hz timing. The display shows black screen
on boot. The issue can be recovered with hotplug. It also causes
MPO to fail. We will temprarily revert this commit and investigate
the root cause further.
[WHY]
Odm update is doubled buffered. We need to wait for ODM update to be
completed before optimizing bandwidth or programming new udpates.
[HOW]
implement wait_for_odm_update_pending_complete function to wait for:
1. odm configuration update is no longer pending in timing generator.
2. no pending dpg pattern update for each active OPP.
Wenjing Liu [Fri, 23 Feb 2024 20:17:39 +0000 (15:17 -0500)]
drm/amd/display: Lock all enabled otg pipes even with no planes
[WHY]
On DCN32 we support dynamic ODM even when OTG is blanked. When ODM
configuration is dynamically changed and the OTG is on blank pattern,
we will need to reprogram OPP's test pattern based on new ODM
configuration. Therefore we need to lock the OTG pipe to avoid temporary
corruption when we are reprogramming OPP blank patterns.
[HOW]
Add a new interdependent update lock implementation to lock all enabled
OTG pipes even when there is no plane on the OTG for DCN32.
Wenjing Liu [Wed, 21 Feb 2024 21:55:04 +0000 (16:55 -0500)]
drm/amd/display: Update odm when ODM combine is changed on an otg master pipe with no plane
[WHY]
When committing an update with ODM combine change when the plane is
removing or already removed, we fail to detect odm change in pipe
update flags. This has caused mismatch between new dc state and the
actual hardware state, because we missed odm programming.
[HOW]
- Detect odm change even for otg master pipe without a plane.
- Update odm config before calling program pipes for pipe with planes.
The commit also updates blank pattern programming when odm is changed
without plane. This is because number of OPP is changed when ODM
combine is changed. Blank pattern is per OPP so we will need to
reprogram OPP based on the new pipe topology.
Ryan Lin [Wed, 21 Feb 2024 11:10:27 +0000 (19:10 +0800)]
drm/amd/display: Add monitor patch for specific eDP
[WHY]
Some eDP panels' ext caps don't write initial values. The value of
dpcd_addr (0x317) can be random and the backlight control interface
will be incorrect.
[HOW]
Add new panel patches to remove sink ext caps.
Ahmad Rehman [Mon, 4 Mar 2024 21:56:00 +0000 (15:56 -0600)]
drm/amdgpu: Init zone device and drm client after mode-1 reset on reload
In passthrough environment, when amdgpu is reloaded after unload, mode-1
is triggered after initializing the necessary IPs, That init does not
include KFD, and KFD init waits until the reset is completed. KFD init
is called in the reset handler, but in this case, the zone device and
drm client is not initialized, causing app to create kernel panic.
v2: Removing the init KFD condition from amdgpu_amdkfd_drm_client_create.
As the previous version has the potential of creating DRM client twice.
v3: v2 patch results in SDMA engine hung as DRM open causes VM clear to SDMA
before SDMA init. Adding the condition to in drm client creation, on top of v1,
to guard against drm client creation call multiple times.
Philip Yang [Mon, 11 Mar 2024 22:07:34 +0000 (18:07 -0400)]
drm/amdgpu: amdgpu_ttm_gart_bind set gtt bound flag
Otherwise after the GTT bo is released, the GTT and gart space is freed
but amdgpu_ttm_backend_unbind will not clear the gart page table entry
and leave valid mapping entry pointing to the stale system page. Then
if GPU access the gart address mistakely, it will read undefined value
instead page fault, harder to debug and reproduce the real issue.
Swapnil Patel [Tue, 13 Feb 2024 13:09:48 +0000 (08:09 -0500)]
drm/amd/display: Change default size for dummy plane in DML2
[WHY & HOW]
Currently, to map dc states into dml_display_cfg,
We create a dummy plane if the stream doesn't have any planes
attached to it. This dummy plane uses max addersable width height.
This results in certain mode validations failing when they shouldn't.
Vitaly Prosyak [Wed, 6 Mar 2024 19:57:48 +0000 (14:57 -0500)]
drm/amdgpu: fix use-after-free bug
The bug can be triggered by sending a single amdgpu_gem_userptr_ioctl
to the AMDGPU DRM driver on any ASICs with an invalid address and size.
The bug was reported by Joonkyo Jung <[email protected]>.
For example the following code:
static void Syzkaller1(int fd)
{
struct drm_amdgpu_gem_userptr arg;
int ret;
Due to the address and size are not valid there is a failure in
amdgpu_hmm_register->mmu_interval_notifier_insert->__mmu_interval_notifier_insert->
check_shl_overflow, but we even the amdgpu_hmm_register failure we still call
amdgpu_hmm_unregister into amdgpu_gem_object_free which causes access to a bad address.
The following stack is below when the issue is reproduced when Kazan is enabled:
Mukul Joshi [Fri, 8 Mar 2024 16:11:03 +0000 (11:11 -0500)]
drm/amdgpu: Handle duplicate BOs during process restore
In certain situations, some apps can import a BO multiple times
(through IPC for example). To restore such processes successfully,
we need to tell drm to ignore duplicate BOs.
While at it, also add additional logging to prevent silent failures
when process restore fails.
drm/amd/display: Use freesync when `DRM_EDID_FEATURE_CONTINUOUS_FREQ` found
The monitor shipped with the Framework 16 supports VRR [1], but it's not
being advertised.
This is because the detailed timing block doesn't contain
`EDID_DETAIL_MONITOR_RANGE` which amdgpu looks for to find min and max
frequencies. This check however is superfluous for this case because
update_display_info() calls drm_get_monitor_range() to get these ranges
already.
So if the `DRM_EDID_FEATURE_CONTINUOUS_FREQ` EDID feature is found then
turn on freesync without extra checks.
v2: squash in fix from Harry
Closes: https://www.reddit.com/r/framework/comments/1b4y2i5/no_variable_refresh_rate_on_the_framework_16_on/ Closes: https://www.reddit.com/r/framework/comments/1b6vzcy/framework_16_variable_refresh_rate/ Closes: https://community.frame.work/t/resolved-no-vrr-freesync-with-amd-version/42338 Link: https://gist.github.com/superm1/e8fbacfa4d0f53150231d3a3e0a13faf Signed-off-by: Mario Limonciello <[email protected]> Reviewed-by: Harry Wentland <[email protected]> Signed-off-by: Alex Deucher <[email protected]>
On riscv, mmap currently returns an address from the largest address
space that can fit entirely inside of the hint address. This makes it
such that the hint address is almost never returned. This patch raises
the mappable area up to and including the hint address. This allows mmap
to often return the hint address, which allows a performance improvement
over searching for a valid address as well as making the behavior more
similar to other architectures.
Note that a previous patch introduced stronger semantics compared to
other architectures for riscv mmap. On riscv, mmap will not use bits in
the upper bits of the virtual address depending on the hint address. On
other architectures, a random address is returned in the address space
requested. On all architectures the hint address will be returned if it
is available. This allows riscv applications to configure how many bits
in the virtual address should be left empty. This has the two benefits
of being able to request address spaces that are smaller than the
default and doesn't require the application to know the page table
layout of riscv.
* b4-shazam-merge:
docs: riscv: Define behavior of mmap
selftests: riscv: Generalize mm selftests
riscv: mm: Use hint address in mmap if available
This series makes barrier-related macro more neat and clear.
This is a follow-up to [0-3], change to multiple patches,
for readability, create new message thread.
Add an implementation of cts(cbc(aes)) accelerated using the Zvkned
RISC-V vector crypto extension. This is mainly useful for fscrypt,
where cts(cbc(aes)) is the "default" filenames encryption algorithm. In
that use case, typically most messages are short and are block-aligned.
The CBC-CTS variant implemented is CS3; this is the variant Linux uses.
To perform well on short messages, the new implementation processes the
full message in one call to the assembly function if the data is
contiguous. Otherwise it falls back to CBC operations followed by CTS
at the end. For decryption, to further improve performance on short
messages, especially block-aligned messages, the CBC-CTS assembly
function parallelizes the AES decryption of all full blocks. This
improves on the arm64 implementation of cts(cbc(aes)), which always
splits the CBC part(s) from the CTS part, doing the AES decryptions for
the last two blocks serially and usually loading the round keys twice.
Tested in QEMU with CONFIG_CRYPTO_MANAGER_EXTRA_TESTS=y.
Eric Biggers [Thu, 8 Feb 2024 06:08:51 +0000 (22:08 -0800)]
crypto: riscv - parallelize AES-CBC decryption
Since CBC decryption is parallelizable, make the RISC-V implementation
of AES-CBC decryption process multiple blocks at a time, instead of
processing the blocks one by one. This should improve performance.
This series enables the support for "Collaborative Processor Performance
Control (CPPC) on ACPI based RISC-V platforms. It depends on the
encoding of CPPC registers as defined in RISC-V FFH spec [2].
CPPC is described in the ACPI spec [1]. RISC-V FFH spec required to
enable this, is available at [2].
Alexandre Ghiti [Fri, 2 Feb 2024 12:47:11 +0000 (13:47 +0100)]
riscv: Only flush the mm icache when setting an exec pte
We used to emit a flush_icache_all() whenever a dirty executable
mapping is set in the page table but we can instead call
flush_icache_mm() which will only send IPIs to cores that currently run
this mm and add a deferred icache flush to the others.
The number of calls to sbi_remote_fence_i() (tested without IPI
support):
With a simple buildroot rootfs:
* Before: ~5k
* After : 4 (!)
Tested on HW, the boot to login is ~4.5% faster.
With an ubuntu rootfs:
* Before: ~24k
* After : ~13k
Erick Archer [Sat, 20 Jan 2024 13:54:00 +0000 (14:54 +0100)]
riscv: Use kcalloc() instead of kzalloc()
As noted in the "Deprecated Interfaces, Language Features, Attributes,
and Conventions" documentation [1], size calculations (especially
multiplication) should not be performed in memory allocator (or similar)
function arguments due to the risk of them overflowing. This could lead
to values wrapping around and a smaller allocation being made than the
caller was expecting. Using those allocations could lead to linear
overflows of heap memory and other misbehaviors.
So, use the purpose specific kcalloc() function instead of the argument
count * size in the kzalloc() function.
Also, it is preferred to use sizeof(*pointer) instead of sizeof(type)
due to the type of the variable can change and one needs not change the
former (unlike the latter).
I just saw the opportunity of optimizing the helper is_compat_task() by
introducing a compile-time test, and it made possible to remove some
#ifdef's without any loss of performance.
I also saw the possibility of removing the direct check of task flags from
general code, and concentrated it in asm/compat.h by creating a few more
helpers, which in the end helped optimize code.
arch_get_mmap_end() just got a simple improvement and some extra docs.
* b4-shazam-merge:
riscv: Introduce set_compat_task() in asm/compat.h
riscv: Introduce is_compat_thread() into compat.h
riscv: add compile-time test into is_compat_task()
riscv: Replace direct thread flag check with is_compat_task()
riscv: Improve arch_get_mmap_end() macro
This patchset fixes the problems related to RVU mailbox.
During long run tests some times VF commands like setting
MTU or toggling interface fails because VF mailbox is timedout
waiting for response from PF.
Below are the fixes
Patch 1: There are two types of messages in RVU mailbox namely up and down
messages. Down messages are synchronous messages where a PF/VF sends
a message to AF and AF replies back with response. UP messages are
notifications and are asynchronous like AF sending link events to
PF. When VF sends a down message to PF, PF forwards to AF and sends
the response from AF back to VF. PF has to forward VF messages since
there is no path in hardware for VF to send directly to AF.
There is one mailbox interrupt from AF to PF when raised could mean
two scenarios one is where AF sending reply to PF for a down message
sent by PF and another one is AF sending up message asynchronously
when link changed for that PF. Receiving the up message interrupt while
PF is in middle of forwarding down message causes mailbox errors.
Fix this by receiver detecting the type of message from the mbox data register
set by sender.
Patch 2:
During VF driver remove, VF has to wait until last message is
completed and then turn off mailbox interrupts from PF.
Patch 3:
Do not use ordered workqueue for message processing since multiple works are
queued simultaneously by all the VFs and PF link UP messages.
Patch 4:
When sending link event to VF by PF check whether VF is really up to
receive this message.
Patch 5:
In AF driver, use separate interrupt handlers for the AF-VF interrupt and
AF-PF interrupt. Sometimes both interrupts are raised to two CPUs at same
time and both CPUs execute same function at same time corrupting the data.
v2 changes:
Added missing mutex unlock in error path in patch 1
Refactored if else logic in patch 1 as suggested by Paolo Abeni
====================
octeontx2-af: Use separate handlers for interrupts
For PF to AF interrupt vector and VF to AF vector same
interrupt handler is registered which is causing race condition.
When two interrupts are raised to two CPUs at same time
then two cores serve same event corrupting the data.
Fixes: 7304ac4567bc ("octeontx2-af: Add mailbox IRQ and msg handlers") Signed-off-by: Subbaraya Sundeep <[email protected]> Signed-off-by: David S. Miller <[email protected]>
octeontx2-pf: Send UP messages to VF only when VF is up.
When PF sending link status messages to VF, it is possible
that by the time link_event_task work function is executed
VF might have brought down. Hence before sending VF link
status message check whether VF is up to receive it.
Fixes: ad513ed938c9 ("octeontx2-vf: Link event notification support") Signed-off-by: Subbaraya Sundeep <[email protected]> Signed-off-by: David S. Miller <[email protected]>
octeontx2-pf: Use default max_active works instead of one
Only one execution context for the workqueue used for PF and
VFs mailbox communication is incorrect since multiple works are
queued simultaneously by all the VFs and PF link UP messages.
Hence use default number of execution contexts by passing zero
as max_active to alloc_workqueue function. With this fix in place,
modify UP messages also to wait until completion.
Fixes: d424b6c02415 ("octeontx2-pf: Enable SRIOV and added VF mbox handling") Signed-off-by: Subbaraya Sundeep <[email protected]> Signed-off-by: David S. Miller <[email protected]>
octeontx2-pf: Wait till detach_resources msg is complete
During VF driver remove, a message is sent to detach VF
resources to PF but VF is not waiting until message is
complete. Also mailbox interrupts need to be turned off
after the detach resource message is complete. This patch
fixes that problem.
Fixes: 05fcc9e08955 ("octeontx2-pf: Attach NIX and NPA block LFs") Signed-off-by: Subbaraya Sundeep <[email protected]> Signed-off-by: David S. Miller <[email protected]>
octeontx2: Detect the mbox up or down message via register
A single line of interrupt is used to receive up notifications
and down reply messages from AF to PF (similarly from PF to its VF).
PF acts as bridge and forwards VF messages to AF and sends respsones
back from AF to VF. When an async event like link event is received
by up message when PF is in middle of forwarding VF message then
mailbox errors occur because PF state machine is corrupted.
Since VF is a separate driver or VF driver can be in a VM it is
not possible to serialize from the start of communication at VF.
Hence to differentiate between type of messages at PF this patch makes
sender to set mbox data register with distinct values for up and down
messages. Sender also checks whether previous interrupt is received
before triggering current interrupt by waiting for mailbox data register
to become zero.
Fixes: 5a6d7c9daef3 ("octeontx2-pf: Mailbox communication with AF") Signed-off-by: Subbaraya Sundeep <[email protected]> Signed-off-by: David S. Miller <[email protected]>
Chris Packham [Tue, 12 Mar 2024 20:14:00 +0000 (09:14 +1300)]
i2c: muxes: pca954x: Allow sharing reset GPIO
Some hardware designs with multiple PCA954x devices use a reset GPIO
connected to all the muxes. Support this configuration by making use of
the reset controller framework which can deal with the shared reset
GPIOs. Fall back to the old GPIO descriptor method if the reset
controller framework is not enabled.
Wolfram Sang [Wed, 20 Mar 2024 08:28:51 +0000 (09:28 +0100)]
Merge tag 'i2c-host-6.9-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/andi.shyti/linux into i2c/for-mergewindow
Théo adds support for the Mobileye EyeQ5-I2C in the bindings.
This patch is followed by eight commits featuring improvements to
the Nomadik controller, such as simplification of the IRQ logic,
renaming of the private data structure, more efficient use of
FIELD_PREP/GET, GENMASK, etc., better time measurement with
ktime, and more.
The timerlat tracer provides an interface for any application to wait
for the timerlat's periodic wakeup. Currently, rtla timerlat uses it
to dispatch its user-space workload (-u option).
But as the tracer interface is generic, rtla timerlat can also be used
to monitor any workload that uses it. For example, a user might
place their own workload to wait on the tracer interface, and
monitor the results with rtla timerlat.
Add the -U option to rtla timerlat top and hist. With this option, rtla
timerlat will not dispatch its workload but only setting up the
system, waiting for a user to dispatch its workload.
The sample code in this patch is an example of python application
that loops in the timerlat tracer fd.
To use it, dispatch:
# rtla timerlat -U
In a terminal, then run the python program on another terminal,
specifying the CPU to run it. For example, setting on CPU 1:
#./timerlat_load.py 1
Then rtla timerlat will start printing the statistics of the
./timerlat_load.py app.
An interesting point is that the "Ret user Timer Latency" value
is the overall response time of the load. The sample load does
a memory copy to exemplify that.
The stop tracing options on rtla timerlat works in this setup
as well, including auto analysis.
tools/rtla: Use tools/build makefiles to build rtla
Use tools/build/ makefiles to build rtla, inheriting the benefits of
it. For example, having a proper way to handle dependencies.
rtla is built using perf infra-structure when building inside the
kernel tree.
At this point, rtla diverges from perf in two points: Documentation
and tarball generation/build.
At the documentation level, rtla is one step ahead, placing the
documentation at Documentation/tools/rtla/, using the same build
tools as kernel documentation. The idea is to move perf
documentation to the same scheme and then share the same makefiles.
rtla has a tarball target that the (old) RHEL8 uses. The tarball was
kept using a simple standalone makefile for compatibility. The
standalone makefile shares most of the code, e.g., flags, with
regular buildings.
The tarball method was set as deprecated. If necessary, we can make
a rtla tarball like perf, which includes the entire tools/build.
But this would also require changes in the user side (the directory
structure changes, and probably the deps to build the package).
tools/tracing: Use tools/build makefiles on latency-collector
Use tools/build/ makefiles to build latency-collector, inheriting
the benefits of it. For example: Before this patch, a missing
tracefs/traceevents headers will result in fail like this:
~/linux/tools/tracing/latency $ make
cc -Wall -Wextra -g -O2 -o latency-collector latency-collector.c -lpthread
latency-collector.c:26:10: fatal error: tracefs.h: No such file or directory
26 | #include <tracefs.h>
| ^~~~~~~~~~~
compilation terminated.
make: *** [Makefile:14: latency-collector] Error 1
Which is not that helpful. After this change it reports:
~/linux/tools/tracing/latency# make
Auto-detecting system features:
... libtraceevent: [ OFF ]
... libtracefs: [ OFF ]
libtraceevent is missing. Please install libtraceevent-dev/libtraceevent-devel
libtracefs is missing. Please install libtracefs-dev/libtracefs-devel
Makefile.config:29: *** Please, check the errors above.. Stop.
This type of output is common across other tools in tools/ like perf
and objtool.
1) Fix possible page_pool leak triggered by esp_output.
From Dragos Tatulea.
2) Fix UDP encapsulation in software GSO path.
From Leon Romanovsky.
* tag 'ipsec-2024-03-19' of git://git.kernel.org/pub/scm/linux/kernel/git/klassert/ipsec:
xfrm: Allow UDP encapsulation only in offload modes
net: esp: fix bad handling of pages from page_pool
====================
syzkaller reported a read of uninit req->syncookie. [0]
Originally, req->syncookie was used only in tcp_conn_request()
to indicate if we need to encode SYN cookie in SYN+ACK, so the
field remains uninitialised in other places.
The commit 695751e31a63 ("bpf: tcp: Handle BPF SYN Cookie in
cookie_v[46]_check().") added another meaning in ACK path;
req->syncookie is set true if SYN cookie is validated by BPF
kfunc.
After the change, cookie_v[46]_check() always read req->syncookie,
but it is not initialised in the normal SYN cookie case as reported
by KMSAN.
Let's make sure we always initialise req->syncookie in reqsk_alloc().
Thinh Tran [Fri, 15 Mar 2024 20:55:35 +0000 (15:55 -0500)]
net/bnx2x: Prevent access to a freed page in page_pool
Fix race condition leading to system crash during EEH error handling
During EEH error recovery, the bnx2x driver's transmit timeout logic
could cause a race condition when handling reset tasks. The
bnx2x_tx_timeout() schedules reset tasks via bnx2x_sp_rtnl_task(),
which ultimately leads to bnx2x_nic_unload(). In bnx2x_nic_unload()
SGEs are freed using bnx2x_free_rx_sge_range(). However, this could
overlap with the EEH driver's attempt to reset the device using
bnx2x_io_slot_reset(), which also tries to free SGEs. This race
condition can result in system crashes due to accessing freed memory
locations in bnx2x_free_rx_sge()
799 static inline void bnx2x_free_rx_sge(struct bnx2x *bp,
800 struct bnx2x_fastpath *fp, u16 index)
801 {
802 struct sw_rx_page *sw_buf = &fp->rx_page_ring[index];
803 struct page *page = sw_buf->page;
....
where sw_buf was set to NULL after the call to dma_unmap_page()
by the preceding thread.
Eric Chan [Sat, 17 Feb 2024 13:13:28 +0000 (13:13 +0000)]
riscv/barrier: Add missing space after ','
The past form of RISCV_FENCE would cause checkpatch.pl to issue
error messages, the example is as follows:
ERROR: space required after that ',' (ctx:VxV)
26: FILE: arch/riscv/include/asm/barrier.h:27:
+#define __smp_mb() RISCV_FENCE(rw,rw)
^
fix the remaining of RISCV_FENCE.
Eric Chan [Sat, 17 Feb 2024 13:13:02 +0000 (13:13 +0000)]
riscv/barrier: Define RISCV_FULL_BARRIER
Introduce RISCV_FULL_BARRIER and use in arch_atomic* function.
like RISCV_ACQUIRE_BARRIER and RISCV_RELEASE_BARRIER, the fence
instruction can be eliminated When SMP is not enabled.
Eric Chan [Sat, 17 Feb 2024 13:12:49 +0000 (13:12 +0000)]
riscv/barrier: Define __{mb,rmb,wmb}
Introduce __{mb,rmb,wmb}, and rely on the generic definitions for
{mb,rmb,wmb}. Although KCSAN is not supported yet, the definitions can
be made more consistent with generic instrumentation. Also add a space
to make the changes pass check by checkpatch.pl.
Without the space, the error message is as below:
ERROR: space required after that ',' (ctx:VxV)
26: FILE: arch/riscv/include/asm/barrier.h:23:
+#define __mb() RISCV_FENCE(iorw,iorw)
^
Sunil V L [Thu, 8 Feb 2024 03:44:13 +0000 (09:14 +0530)]
cpufreq: Move CPPC configs to common Kconfig and add RISC-V
CPPC related config options are currently defined only in ARM specific
file. However, they are required for RISC-V as well. Instead of creating
a new Kconfig.riscv file and duplicating them, move them to the common
Kconfig file and enable RISC-V too.
Sunil V L [Thu, 8 Feb 2024 03:44:12 +0000 (09:14 +0530)]
ACPI: RISC-V: Add CPPC driver
Add cpufreq driver based on ACPI CPPC for RISC-V. The driver uses either
SBI CPPC interfaces or the CSRs to access the CPPC registers as defined
by the RISC-V FFH spec.
Sunil V L [Thu, 18 Jan 2024 06:29:30 +0000 (11:59 +0530)]
ACPI: Enable ACPI_PROCESSOR for RISC-V
The ACPI processor driver is not currently enabled for RISC-V.
This is required to enable CPU related functionalities like
LPI and CPPC. Hence, enable ACPI_PROCESSOR for RISC-V.
Sunil V L [Thu, 18 Jan 2024 06:29:28 +0000 (11:59 +0530)]
cpuidle: RISC-V: Move few functions to arch/riscv
To support ACPI Low Power Idle (LPI), few functions are required which
are currently static functions in the DT based cpuidle driver. Hence,
move them under arch/riscv so that ACPI driver also can use them. Since
they are no longer static functions, append "riscv_" prefix to the
function name.
Linus Torvalds [Wed, 20 Mar 2024 00:27:25 +0000 (17:27 -0700)]
Merge tag 'bcachefs-2024-03-19' of https://evilpiepirate.org/git/bcachefs
Pull bcachefs fixes from Kent Overstreet:
"Assorted bugfixes.
Most are fixes for simple assertion pops; the most significant fix is
for a deadlock in recovery when we have to rewrite large numbers of
btree nodes to fix errors. This was incorrectly running out of the
same workqueue as the core interior btree update path - we now give it
its own single threaded workqueue.
This was visible to users as "bch2_btree_update_start(): error:
BCH_ERR_journal_reclaim_would_deadlock" - and then recovery hanging"
* tag 'bcachefs-2024-03-19' of https://evilpiepirate.org/git/bcachefs:
bcachefs: Fix lost wakeup on journal shutdown
bcachefs; Fix deadlock in bch2_btree_update_start()
bcachefs: ratelimit errors from async_btree_node_rewrite
bcachefs: Run check_topology() first
bcachefs: Improve bch2_fatal_error()
bcachefs: Fix lost transaction restart error
bcachefs: Don't corrupt journal keys gap buffer when dropping alloc info
bcachefs: fix for building in userspace
bcachefs: bch2_snapshot_is_ancestor() now safe to call in early recovery
bcachefs: Fix nested transaction restart handling in bch2_bucket_gens_init()
bcachefs: Improve sysfs internal/btree_updates
bcachefs: Split out btree_node_rewrite_worker
bcachefs: Fix locking in bch2_alloc_write_key()
bcachefs: Avoid extent entry type assertions in .invalid()
bcachefs: Fix spurious -BCH_ERR_transaction_restart_nested
bcachefs: Fix check_key_has_snapshot() call
bcachefs: Change "accounting overran journal reservation" to a warning
Leonardo Bras [Wed, 3 Jan 2024 16:00:21 +0000 (13:00 -0300)]
riscv: add compile-time test into is_compat_task()
Currently several places will test for CONFIG_COMPAT before testing
is_compat_task(), probably in order to avoid a run-time test into the task
structure.
Since is_compat_task() is an inlined function, it would be helpful to add a
compile-time test of CONFIG_COMPAT, making sure it always returns zero when
the option is not enabled during the kernel build.
With this, the compiler is able to understand in build-time that
is_compat_task() will always return 0, and optimize-out some of the extra
code introduced by the option.
This will also allow removing a lot #ifdefs that were introduced, and make
the code more clean.
Leonardo Bras [Wed, 3 Jan 2024 16:00:19 +0000 (13:00 -0300)]
riscv: Improve arch_get_mmap_end() macro
This macro caused me some confusion, which took some reviewer's time to
make it clear, so I propose adding a short comment in code to avoid
confusion in the future.
Also, added some improvements to the macro, such as removing the
assumption of VA_USER_SV57 being the largest address space.
Conor Dooley [Mon, 18 Mar 2024 15:35:04 +0000 (15:35 +0000)]
perf: starfive: fix 64-bit only COMPILE_TEST condition
ARCH_STARFIVE is not restricted to 64-bit platforms, so while Will's
addition of a 64-bit only condition satisfied the build robots doing
COMPILE_TEST builds, Palmer ran into the same problems with writeq()
being undefined during regular rv32 builds.
Promote the dependency on 64-bit to its own `depends on` so that the
driver can never be included in 32-bit builds.
Linus Torvalds [Tue, 19 Mar 2024 18:57:26 +0000 (11:57 -0700)]
Merge tag 'soc-late-6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Pull more ARM SoC updates from Arnd Bergmann:
"These are changes that for some reason ended up not making it into the
first four branches but that should still make it into 6.9:
- A rework of the omap clock support that touches both drivers and
device tree files
- The reset controller branch changes that had a dependency on late
bugfixes. Merging them here avoids a backmerge of 6.8-rc5 into the
drivers branch
- The RISC-V/starfive, RISC-V/microchip and ARM/Broadcom devicetree
changes that got delayed and needed some extra time in linux-next
for wider testing"
* tag 'soc-late-6.9' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc: (31 commits)
soc: fsl: dpio: fix kcalloc() argument order
bus: ts-nbus: Improve error reporting
bus: ts-nbus: Convert to atomic pwm API
riscv: dts: starfive: jh7110: Add camera subsystem nodes
ARM: bcm: stop selecing CONFIG_TICK_ONESHOT
ARM: dts: omap3: Update clksel clocks to use reg instead of ti,bit-shift
ARM: dts: am3: Update clksel clocks to use reg instead of ti,bit-shift
clk: ti: Improve clksel clock bit parsing for reg property
clk: ti: Handle possible address in the node name
dt-bindings: pwm: opencores: Add compatible for StarFive JH8100
dt-bindings: riscv: cpus: reg matches hart ID
reset: Instantiate reset GPIO controller for shared reset-gpios
reset: gpio: Add GPIO-based reset controller
cpufreq: do not open-code of_phandle_args_equal()
of: Add of_phandle_args_equal() helper
reset: simple: add support for Sophgo SG2042
dt-bindings: reset: sophgo: support SG2042
riscv: dts: microchip: add specific compatible for mpfs pdma
riscv: dts: microchip: add missing CAN bus clocks
ARM: brcmstb: Add debug UART entry for 74165
...
projects / linux.git / log