This patch uses a helper function for assignment of xdp_features.
This change simplifies backports.
Signed-off-by: Konstantin Taranov <kotaranov@microsoft.com>
Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Link: https://lore.kernel.org/r/1698430011-21562-1-git-send-email-haiyangz@microsoft.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This patch is basically a followup to commit 4e4b1798cc ("vxlan: Add
missing entries to vxlan_get_size()"). All of the attributes in
vxlan_get_size() appear in the same order that they are filled in
vxlan_fill_info() except for IFLA_VXLAN_PORT_RANGE. For consistency, move
that entry to match its order and add a comment, like for all other
entries.
Signed-off-by: Benjamin Poirier <bpoirier@nvidia.com>
Link: https://lore.kernel.org/r/20231027184410.236671-1-bpoirier@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jacob Keller says:
====================
Intel Wired LAN Driver Updates for 2023-10-23 (iavf)
This series includes iAVF driver cleanups from Michal Schmidt.
Michal removes and updates stale comments, fixes some locking anti-patterns,
improves handling of resets when the PF is slow, avoids unnecessary
duplication of netdev state, refactors away some duplicate code, and finally
removes the never-actually-used client interface.
Changes since v1:
* Dropped patch ("iavf: in iavf_down, disable queues when removing the
driver") which was applied directly to net.
* Fixed a merge conflict due to 7db3111043 ("iavf: initialize waitqueues
before starting watchdog_task").
V1 was originally posted at:
https://lore.kernel.org/netdev/20231027104109.4f536f51@kernel.org/T/#mfadbdb39313eeccc616fdee80a4fdd6bda7e2822
====================
Link: https://lore.kernel.org/r/20231027175941.1340255-1-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The iavf client interface was added in 2017 by commit ed0e894de7
("i40evf: add client interface"), but there have never been any in-tree
callers.
It's not useful for future development either. The Intel out-of-tree
iavf and irdma drivers instead use an auxiliary bus, which is a better
solution.
Remove the iavf client interface code. Also gone are the client_task
work and the client_lock mutex.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-9-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add a new function iavf_free_interrupt_scheme that does the inverse of
iavf_init_interrupt_scheme. Symmetry is nice. And there will be three
callers already.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-8-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Use unregister_netdev, which takes rtnl_lock for us. We don't have to
check the reg_state under rtnl_lock. There's nothing to race with. We
have just cancelled the finish_config work.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-7-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The information whether a netdev has been registered is already present
in the netdev itself. There's no need for a driver flag with the same
meaning.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-6-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Every time I create VFs on ice, I receive at least one "Device is still
in reset (-16), retrying" message per VF. It recovers fine, but typical
usecases should not trigger scary-looking messages.
The waiting for reset is too short. It makes no sense to check every 10
microseconds. Typical reset waiting times are at least tens of
milliseconds and can be several seconds. I suspect the polling interval
was meant to be 10 milliseconds all along.
IAVF_RESET_WAIT_COMPLETE_COUNT is defined as 2000, so the total waiting
time could be over 20 seconds. I have seen resets take 5 seconds (with
128 VFs on ice).
The added benefit of not triggering the "Device is still in reset" path
is that we avoid going through the __IAVF_INIT_FAILED state, which would
take a full second before retrying.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-5-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The reason for queueing watchdog_task is to have it process the
aq_required flags that are being set here. If comms failed, there's
nothing to do, so return early.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-4-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
This pattern appears in two places in the iavf source code:
while (!mutex_trylock(...))
usleep_range(...);
That's just mutex_lock with extra steps.
The pattern is a leftover from when iavf used bit flags instead of
mutexes for locking. Commit 5ac49f3c27 ("iavf: use mutexes for locking
of critical sections") replaced test_and_set_bit with !mutex_trylock,
preserving the pattern.
Simplify it to mutex_lock.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-3-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Bit lock __IAVF_IN_CRITICAL_TASK does not exist anymore since commit
5ac49f3c27 ("iavf: use mutexes for locking of critical sections").
Adjust the comments accordingly.
Signed-off-by: Michal Schmidt <mschmidt@redhat.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://lore.kernel.org/r/20231027175941.1340255-2-jacob.e.keller@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
allow specifying cmd-cnt-name and cmd-max-name in netlink specs, in
accordance with Documentation/userspace-api/netlink/c-code-gen.rst.
Use cmd-cnt-name and attr-cnt-name in the mptcp yaml spec and in the
corresponding uAPI headers, to preserve the #defines we had in the past
and avoid adding new ones.
v2:
- squash modification in mptcp.yaml and MPTCP uAPI headers
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Link: https://lore.kernel.org/r/12d4ed0116d8883cf4b533b856f3125a34e56749.1698415310.git.dcaratti@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In case the kernel sends message back containing attribute not defined
in family spec, following exception is raised to the user:
$ sudo ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/devlink.yaml --do trap-get --json '{"bus-name": "netdevsim", "dev-name": "netdevsim1", "trap-name": "source_mac_is_multicast"}'
Traceback (most recent call last):
File "/home/jiri/work/linux/tools/net/ynl/lib/ynl.py", line 521, in _decode
attr_spec = attr_space.attrs_by_val[attr.type]
~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^
KeyError: 132
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "/home/jiri/work/linux/./tools/net/ynl/cli.py", line 61, in <module>
main()
File "/home/jiri/work/linux/./tools/net/ynl/cli.py", line 49, in main
reply = ynl.do(args.do, attrs, args.flags)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/jiri/work/linux/tools/net/ynl/lib/ynl.py", line 731, in do
return self._op(method, vals, flags)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/jiri/work/linux/tools/net/ynl/lib/ynl.py", line 719, in _op
rsp_msg = self._decode(decoded.raw_attrs, op.attr_set.name)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/jiri/work/linux/tools/net/ynl/lib/ynl.py", line 525, in _decode
raise Exception(f"Space '{space}' has no attribute with value '{attr.type}'")
Exception: Space 'devlink' has no attribute with value '132'
Introduce a command line option "process-unknown" and pass it down to
YnlFamily class constructor to allow user to process unknown
attributes and types and print them as binaries.
$ sudo ./tools/net/ynl/cli.py --spec Documentation/netlink/specs/devlink.yaml --do trap-get --json '{"bus-name": "netdevsim", "dev-name": "netdevsim1", "trap-name": "source_mac_is_multicast"}' --process-unknown
{'UnknownAttr(129)': {'UnknownAttr(0)': b'\x00\x00\x00\x00\x00\x00\x00\x00',
'UnknownAttr(1)': b'\x00\x00\x00\x00\x00\x00\x00\x00',
'UnknownAttr(2)': b'\x0e\x00\x00\x00\x00\x00\x00\x00'},
'UnknownAttr(132)': b'\x00',
'UnknownAttr(133)': b'',
'UnknownAttr(134)': {'UnknownAttr(0)': b''},
'bus-name': 'netdevsim',
'dev-name': 'netdevsim1',
'trap-action': 'drop',
'trap-group-name': 'l2_drops',
'trap-name': 'source_mac_is_multicast'}
Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20231027092525.956172-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Both ipvlan_process_v4_outbound() and ipvlan_process_v6_outbound()
increment dev->stats.tx_errors in case of errors.
Unfortunately there are two issues :
1) ipvlan_get_stats64() does not propagate dev->stats.tx_errors to user.
2) Increments are not atomic. KCSAN would complain eventually.
Use DEV_STATS_INC() to not miss an update, and change ipvlan_get_stats64()
to copy the value back to user.
Fixes: 2ad7bf3638 ("ipvlan: Initial check-in of the IPVLAN driver.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Mahesh Bandewar <maheshb@google.com>
Link: https://lore.kernel.org/r/20231026131446.3933175-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Like other buses, devices on the netdevsim bus have a release callback
that is invoked when the reference count of the device drops to zero.
However, unlike other buses such as PCI, the release callback is not
necessarily built into the kernel, as netdevsim can be built as a
module.
The above is problematic as nothing prevents the module from being
unloaded before the release callback has been invoked, which can happen
asynchronously. One such example can be found in commit a380687200
("devlink: take device reference for devlink object") where devlink
calls put_device() from an RCU callback.
The issue is not theoretical and the reproducer in [1] can reliably
crash the kernel. The conclusion of this discussion was that the issue
should be solved in netdevsim, which is what this patch is trying to do.
Add a reference count that is increased when a device is added to the
bus and decreased when a device is released. Signal a completion when
the reference count drops to zero and wait for the completion when
unloading the module so that the module will not be unloaded before all
the devices were released. The reference count is initialized to one so
that completion is only signaled when unloading the module.
With this patch, the reproducer in [1] no longer crashes the kernel.
[1] https://lore.kernel.org/netdev/20230619125015.1541143-2-idosch@nvidia.com/
Fixes: a380687200 ("devlink: take device reference for devlink object")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20231026083343.890689-1-idosch@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The napi_build_skb() can reuse the skb in skb cache per CPU or
can allocate skbs in bulk, which helps improve the performance.
Signed-off-by: Fei Qin <fei.qin@corigine.com>
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Reviewed-by: Wojciech Drewek <wojciech.drewek@intel.com>
Link: https://lore.kernel.org/r/20231026080058.22810-1-louis.peens@corigine.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
There is a spelling mistake in a dev_dbg message. Fix it.
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/all/20231026065408.1087824-1-colin.i.king@gmail.com/
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Oleksij Rempel says:
====================
net: dsa: microchip: provide Wake on LAN support (part 2)
This patch series introduces extensive Wake on LAN (WoL) support for the
Microchip KSZ9477 family of switches, coupled with some code refactoring
and error handling enhancements. The principal aim is to enable and
manage Wake on Magic Packet and other PHY event triggers for waking up
the system, whilst ensuring that the switch isn't reset during a
shutdown if WoL is active.
The Wake on LAN functionality is optional and is particularly beneficial
if the PME pins are connected to the SoC as a wake source or to a PMIC
that can enable or wake the SoC.
====================
Link: https://lore.kernel.org/r/20231026051051.2316937-1-o.rempel@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Ensures a stable PME (Power Management Event) pin state by disabling PME
on system start and enabling it on shutdown only if WoL (Wake-on-LAN) is
configured. This is needed to avoid issues with some PMICs (Power
Management ICs).
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231026051051.2316937-6-o.rempel@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Centralize the switch shutdown routine in a dedicated function,
ksz_switch_shutdown(), to enhance code maintainability and reduce
redundancy. This change abstracts the common shutdown operations
previously duplicated in ksz9477_i2c_shutdown() and ksz_spi_shutdown().
This refactoring is a preparatory step for an upcoming patch to avoid
reset on shutdown if Wake-on-LAN (WoL) is enabled.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231026051051.2316937-5-o.rempel@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Enhance the ksz_switch_macaddr_get() function to handle errors that may
occur during the call to ksz_write8(). Specifically, this update checks
the return value of ksz_write8(), which may fail if regmap ranges
validation is not passed and returns the error code.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231026051051.2316937-4-o.rempel@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Introduce Wake on Magic Packet (WoL) functionality to the ksz9477
driver.
Major changes include:
1. Extending the `ksz9477_handle_wake_reason` function to identify Magic
Packet wake events alongside existing wake reasons.
2. Updating the `ksz9477_get_wol` and `ksz9477_set_wol` functions to
handle WAKE_MAGIC alongside the existing WAKE_PHY option, and to
program the switch's MAC address register accordingly when Magic
Packet wake-up is enabled. This change will prevent WAKE_MAGIC
activation if the related port has a different MAC address compared
to a MAC address already used by HSR or an already active WAKE_MAGIC
on another port.
3. Adding a restriction in `ksz_port_set_mac_address` to prevent MAC
address changes on ports with active Wake on Magic Packet, as the
switch's MAC address register is utilized for this feature.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20231026051051.2316937-2-o.rempel@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
On RISC-V, the return address is before the current frame pointer,
unlike on most other architectures. Use the correct offset on RISC-V
to fix the CFI_BACKWARD test.
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Acked-by: Kees Cook <keescook@chromium.org>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20230927224757.1154247-14-samitolvanen@google.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
When both CONFIG_IRQ_STACKS and SCS are enabled, also use a separate
per-CPU shadow call stack.
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20230927224757.1154247-13-samitolvanen@google.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
Implement CONFIG_SHADOW_CALL_STACK for RISC-V. When enabled, the
compiler injects instructions to all non-leaf C functions to
store the return address to the shadow stack and unconditionally
load it again before returning, which makes it harder to corrupt
the return address through a stack overflow, for example.
The active shadow call stack pointer is stored in the gp
register, which makes SCS incompatible with gp relaxation. Use
--no-relax-gp to ensure gp relaxation is disabled and disable
global pointer loading. Add SCS pointers to struct thread_info,
implement SCS initialization, and task switching
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20230927224757.1154247-12-samitolvanen@google.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
In Clang 17, -fsanitize=shadow-call-stack uses the newly declared
platform register gp for storing shadow call stack pointers. As
this is obviously incompatible with gp relaxation, in preparation
for CONFIG_SHADOW_CALL_STACK support, move global pointer loading
to a single macro, which we can cleanly disable when SCS is used
instead.
Link: https://reviews.llvm.org/rGaa1d2693c256
Link: a484e843e6
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Link: https://lore.kernel.org/r/20230927224757.1154247-11-samitolvanen@google.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
With CONFIG_IRQ_STACKS, we switch to a separate per-CPU IRQ stack
before calling handle_riscv_irq or __do_softirq. We currently
have duplicate inline assembly snippets for stack switching in
both code paths. Now that we can access per-CPU variables in
assembly, implement call_on_irq_stack in assembly, and use that
instead of redundant inline assembly.
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Guo Ren <guoren@kernel.org>
Link: https://lore.kernel.org/r/20230927224757.1154247-10-samitolvanen@google.com
Signed-off-by: Palmer Dabbelt <palmer@rivosinc.com>
The bus .map() functions vary only by checking the flag cells values
and skipping over any flag cells to read the addresses. Otherwise they
all do the same reading 'ranges' address and size and returning the
address's offset if it is within the 'ranges' entry.
Refactor all the .map() functions to pass in the flag cell size so that
each bus can check the bus specific flags and then call a common
function to do everything else.
Acked-by: Herve Codina <herve.codina@bootlin.com>
Link: https://lore.kernel.org/r/20231026135358.3564307-3-robh@kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
It is more useful to know how many flags cells a bus has rather than
whether a bus has flags or not as ultimately the number of cells is the
information used. Replace 'has_flags' boolean with 'flag_cells' count.
Acked-by: Herve Codina <herve.codina@bootlin.com>
Link: https://lore.kernel.org/r/20231026135358.3564307-2-robh@kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
Add tests to exercise address translations based on ranges properties.
Tests added cover "default" (2cell) address translations, "default
flags" (3cell) address translations and PCI address translations.
They also cover PCI BAR translations introduced in commit 407d1a5192
("PCI: Create device tree node for bridge").
Signed-off-by: Herve Codina <herve.codina@bootlin.com>
Link: https://lore.kernel.org/r/20231017110221.189299-4-herve.codina@bootlin.com
Signed-off-by: Rob Herring <robh@kernel.org>
Improve readability and maintainability by replacing a hardcoded string
allocation and formatting by the use of the kasprintf() helper.
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Dipen Patel <dipenp@nvidia.com>
Signed-off-by: Dipen Patel <dipenp@nvidia.com>
The value of 'ret' is zero when of_hte_req_count() fails to get number
of entitties to timestamp. And returning success(zero) on this failure
path is incorrect.
Fixes: 9a75a7cd03 ("hte: Add Tegra HTE test driver")
Signed-off-by: Harshit Mogalapalli <harshit.m.mogalapalli@oracle.com>
Reviewed-by: Dipen Patel <dipenp@nvidia.com>
Signed-off-by: Dipen Patel <dipenp@nvidia.com>
Service NMI and SMI requests after PMI requests in vcpu_enter_guest() so
that KVM does not need to cancel and redo the VM-Enter if the guest
configures its PMIs to be delivered as NMIs (likely) or SMIs (unlikely).
Because APIC emulation "injects" NMIs via KVM_REQ_NMI, handling PMI
requests after NMI requests (the likely case) means KVM won't detect the
pending NMI request until the final check for outstanding requests.
Detecting requests at the final stage is costly as KVM has already loaded
guest state, potentially queued events for injection, disabled IRQs,
dropped SRCU, etc., most of which needs to be unwound.
Note that changing the order of request processing doesn't change the end
result, as KVM's final check for outstanding requests prevents entering
the guest until all requests are serviced. I.e. KVM will ultimately
coalesce events (or not) regardless of the ordering.
Using SPEC2017 benchmark programs running along with Intel vtune in a VM
demonstrates that the following code change reduces 800~1500 canceled
VM-Enters per second.
Some glory details:
Probe the invocation to vmx_cancel_injection():
$ perf probe -a vmx_cancel_injection
$ perf stat -a -e probe:vmx_cancel_injection -I 10000 # per 10 seconds
Partial results when SPEC2017 with Intel vtune are running in the VM:
On kernel without the change:
10.010018010 14254 probe:vmx_cancel_injection
20.037646388 15207 probe:vmx_cancel_injection
30.078739816 15261 probe:vmx_cancel_injection
40.114033258 15085 probe:vmx_cancel_injection
50.149297460 15112 probe:vmx_cancel_injection
60.185103088 15104 probe:vmx_cancel_injection
On kernel with the change:
10.003595390 40 probe:vmx_cancel_injection
20.017855682 31 probe:vmx_cancel_injection
30.028355883 34 probe:vmx_cancel_injection
40.038686298 31 probe:vmx_cancel_injection
50.048795162 20 probe:vmx_cancel_injection
60.069057747 19 probe:vmx_cancel_injection
Suggested-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Mingwei Zhang <mizhang@google.com>
Link: https://lore.kernel.org/r/20231002040839.2630027-1-mizhang@google.com
[sean: hoist PMU/PMI above SMI too, massage changelog]
Signed-off-by: Sean Christopherson <seanjc@google.com>
The cxl-cli unit test for firmware update does operations like starting
an asynchronous firmware update, making sure it is in progress, and
attempting to cancel it. In some cases, such as with no or minimal
dynamic debugging turned on, the firmware update completes too quickly,
not allowing the test to have a chance to verify it was in progress.
This caused a failure of the signature:
expected fw_update_in_progress:true
test/cxl-update-firmware.sh: failed at line 88
Fix this by adding a delay (~1.5 - 2 ms) to each firmware transfer
request handled by the mocked interface.
Reported-by: Dan Williams <dan.j.williams@intel.com>
Tested-by: Dan Williams <dan.j.williams@intel.com>
Signed-off-by: Vishal Verma <vishal.l.verma@intel.com>
Link: https://lore.kernel.org/r/20231026-vv-fw_upd_test_fix-v2-1-5282fd193883@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Root decoder granularity must match value from CFWMS, which may not
be the region's granularity for non-interleaved root decoders.
So when calculating granularities for host bridge decoders, use the
region's granularity instead of the root decoder's granularity to ensure
the correct granularities are set for the host bridge decoders and any
downstream switch decoders.
Test configuration is 1 host bridge * 2 switches * 2 endpoints per switch.
Region created with 2048 granularity using following command line:
cxl create-region -m -d decoder0.0 -w 4 mem0 mem2 mem1 mem3 \
-g 2048 -s 2048M
Use "cxl list -PDE | grep granularity" to get a view of the granularity
set at each level of the topology.
Before this patch:
"interleave_granularity":2048,
"interleave_granularity":2048,
"interleave_granularity":512,
"interleave_granularity":2048,
"interleave_granularity":2048,
"interleave_granularity":512,
"interleave_granularity":256,
After:
"interleave_granularity":2048,
"interleave_granularity":2048,
"interleave_granularity":4096,
"interleave_granularity":2048,
"interleave_granularity":2048,
"interleave_granularity":4096,
"interleave_granularity":2048,
Fixes: 27b3f8d138 ("cxl/region: Program target lists")
Cc: <stable@vger.kernel.org>
Signed-off-by: Jim Harris <jim.harris@samsung.com>
Link: https://lore.kernel.org/r/169824893473.1403938.16110924262989774582.stgit@bgt-140510-bm03.eng.stellus.in
[djbw: fixup the prebuilt cxl_test region]
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Fix a missed "goto out" to unlock on error to cleanup this splat:
WARNING: lock held when returning to user space!
6.6.0-rc3-lizhijian+ #213 Not tainted
------------------------------------------------
cxl/673 is leaving the kernel with locks still held!
1 lock held by cxl/673:
#0: ffffffffa013b9d0 (cxl_region_rwsem){++++}-{3:3}, at: commit_store+0x7d/0x3e0 [cxl_core]
In terms of user visible impact of this bug for backports:
cxl_region_invalidate_memregion() on x86 invokes wbinvd which is a
problematic instruction for virtualized environments. So, on virtualized
x86, cxl_region_invalidate_memregion() returns an error. This failure
case got missed because CXL memory-expander device passthrough is not a
production use case, and emulation of CXL devices is typically limited
to kernel development builds with CONFIG_CXL_REGION_INVALIDATION_TEST=y,
that makes cxl_region_invalidate_memregion() succeed.
In other words, the expected exposure of this bug is limited to CXL
subsystem development environments using QEMU that neglected
CONFIG_CXL_REGION_INVALIDATION_TEST=y.
Fixes: d1257d098a ("cxl/region: Move cache invalidation before region teardown, and before setup")
Signed-off-by: Li Zhijian <lizhijian@fujitsu.com>
Reviewed-by: Ira Weiny <ira.weiny@intel.com>
Link: https://lore.kernel.org/r/20231025085450.2514906-1-lizhijian@fujitsu.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
For auto-discovered regions the driver must assign each target to
a valid position in the region interleave set based on the decoder
topology.
The current implementation fails to parse valid decode topologies,
as it does not consider the child offset into a parent port. The sort
put all targets of one port ahead of another port when an interleave
was expected, causing the region assembly to fail.
Replace the existing relative sort with cxl_calc_interleave_pos() that
finds the exact position in a region interleave for an endpoint based
on a walk up the ancestral tree from endpoint to root decoder.
cxl_calc_interleave_pos() was introduced in a prior patch, so the work
here is to use it in cxl_region_sort_targets().
Remove the obsoleted helper functions from the prior sort.
Testing passes on pre-production hardware with BIOS defined regions
that natively trigger this autodiscovery path of the region driver.
Testing passes a CXL unit test using the dev_dbg() calculation test
(see cxl_region_attach()) across an expanded set of region configs:
1, 1, 1+1, 1+1+1, 2, 2+2, 2+2+2, 2+2+2+2, 4, 4+4, where each number
represents the count of endpoints per host bridge.
Fixes: a32320b71f ("cxl/region: Add region autodiscovery")
Reported-by: Dmytro Adamenko <dmytro.adamenko@intel.com>
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Reviewed-by: Jim Harris <jim.harris@samsung.com>
Link: https://lore.kernel.org/r/3946cc55ddc19678733eddc9de2c317749f43f3b.1698263080.git.alison.schofield@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Introduce a calculation to find a target's position in a region
interleave. Perform a self-test of the calculation on user-defined
regions.
The region driver uses the kernel sort() function to put region
targets in relative order. Positions are assigned based on each
target's index in that sorted list. That relative sort doesn't
consider the offset of a port into its parent port which causes
some auto-discovered regions to fail creation. In one failure case,
a 2 + 2 config (2 host bridges each with 2 endpoints), the sort
puts all the targets of one port ahead of another port when they
were expected to be interleaved.
In preparation for repairing the autodiscovery region assembly,
introduce a new method for discovering a target position in the
region interleave.
cxl_calc_interleave_pos() adds a method to find the target position by
ascending from an endpoint to a root decoder. The calculation starts
with the endpoint's local position and position in the parent port. It
traverses towards the root decoder and examines both position and ways
in order to allow the position to be refined all the way to the root
decoder.
This calculation: position = position * parent_ways + parent_pos;
applied iteratively yields the correct position.
Include a self-test that exercises this new position calculation against
every successfully configured user-defined region.
Signed-off-by: Alison Schofield <alison.schofield@intel.com>
Link: https://lore.kernel.org/r/0ac32c75cf81dd8b86bf07d70ff139d33c2300bc.1698263080.git.alison.schofield@intel.com
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Initial support for the SM4450 platform is introduced, with the Global
Clock Controller and RPMh clock controller additions.
CLK_SET_RATE_PARENT is dropped for clocks with fixed-rate GPLLs, across
a variety of IPQ platforms. On IPQ6018, GPLL0 is a missing parent of
APCS PLL, so this is corrected.
For IPQ6018 the I2C clock for QUP6 was previously omitted, as disabling
it is reported to cause problems for RPM. It's now added, but marked as
critical.
Stromer Plus is introduced, and safe source switching of the a53pll in
IPQ5332 is introduced.
SM8550 Video and GPU clock controllers are switched to use the OLE PLL
configure method, instead of manually specifying additional components
of the l-value.
A couple of fixes related to halt bit checks and SMMU GDSC are
introduced for MSM8998.
A possible integer overflow in the frequency calculation in the RCG code
is addressed.
Clocked managed through RPM are removed from the MSM8996 Global Clock
Controller.
Support for the Camera Clock Controller on SM8550 is added.
PLL configuration for the three HFPLLs in MSM8976 are added.
The MSM8996 CBF clock driver's remove function is transitioned to the
void-returning variant.
-----BEGIN PGP SIGNATURE-----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=6vlw
-----END PGP SIGNATURE-----
Merge tag 'qcom-clk-for-6.7' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux into clk-qcom
Pull Qualcomm clk driver updates from Bjorn Andersson:
- Initial support for the SM4450 Global Clock Controller and RPMh clock controllers
- Drop CLK_SET_RATE_PARENT for clocks with fixed-rate GPLLs across a variety of IPQ platforms
- Add missing parent of APCS PLL on IPQ6018
- Add I2C QUP6 clk on IPQ6018 but mark it critical to avoid problems with RPM
- Implement safe source switching for a53pll and use on IPQ5332
- Add support for Stromer Plus PLLs
- Switch SM8550 Video and GPU clock controllers to use OLE PLL configure method
- Non critical fixes to halt bit checks
- Add SMMU GDSC for MSM8998
- Fix possible integer overflow in RCG frequency calculation code
- Remove RPM managed clks from MSM8996 GCC driver
- Add Camera Clock Controller on SM8550
- Add HFPLL configuration for the three HFPLLs in MSM8976
- Switch MSM8996 CBF clock driver's remove function to return void
* tag 'qcom-clk-for-6.7' of https://git.kernel.org/pub/scm/linux/kernel/git/qcom/linux: (36 commits)
clk: qcom: apss-ipq6018: add the GPLL0 clock also as clock provider
clk: qcom: ipq5332: drop the CLK_SET_RATE_PARENT flag from GPLL clocks
clk: qcom: ipq9574: drop the CLK_SET_RATE_PARENT flag from GPLL clocks
clk: qcom: ipq5018: drop the CLK_SET_RATE_PARENT flag from GPLL clocks
clk: qcom: ipq6018: drop the CLK_SET_RATE_PARENT flag from PLL clocks
clk: qcom: ipq8074: drop the CLK_SET_RATE_PARENT flag from PLL clocks
clk: qcom: gcc-ipq6018: add QUP6 I2C clock
clk: qcom: apss-ipq6018: ipq5332: add safe source switch for a53pll
clk: qcom: apss-ipq-pll: Fix 'l' value for ipq5332_pll_config
clk: qcom: apss-ipq-pll: Use stromer plus ops for stromer plus pll
clk: qcom: clk-alpha-pll: introduce stromer plus ops
clk: qcom: config IPQ_APSS_6018 should depend on QCOM_SMEM
clk: qcom: videocc-sm8550: switch to clk_lucid_ole_pll_configure
clk: qcom: gpucc-sm8550: switch to clk_lucid_ole_pll_configure
clk: qcom: Replace of_device.h with explicit includes
clk: qcom: smd-rpm: Move CPUSS_GNoC clock to interconnect
clk: qcom: cbf-msm8996: Convert to platform remove callback returning void
clk: qcom: gcc-sm8150: Fix gcc_sdcc2_apps_clk_src
clk: qcom: Add GCC driver support for SM4450
dt-bindings: clock: qcom: Add GCC clocks for SM4450
...
Increase the size of temporary print buffer on stack to fix the
following warnings reported by LKP.
Since all the input parameters of snprintf() are under control
of this driver, it is not possible to trigger and overflow here,
but since the print buffer is on stack and discarded once driver
probe() finishes, it is not an issue to increase it by 10 bytes
and fix the warning in the process. Make it so.
"
drivers/clk/clk-si521xx.c: In function 'si521xx_probe':
>> drivers/clk/clk-si521xx.c:318:26: warning: '%d' directive output may be truncated writing between 1 and 10 bytes into a region of size 2 [-Wformat-truncation=]
snprintf(name, 6, "DIFF%d", i);
^~
drivers/clk/clk-si521xx.c:318:21: note: directive argument in the range [0, 2147483647]
snprintf(name, 6, "DIFF%d", i);
^~~~~~~~
drivers/clk/clk-si521xx.c:318:3: note: 'snprintf' output between 6 and 15 bytes into a destination of size 6
snprintf(name, 6, "DIFF%d", i);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
"
Fixes: edc12763a3 ("clk: si521xx: Clock driver for Skyworks Si521xx I2C PCIe clock generators")
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202310260412.AGASjFN4-lkp@intel.com/
Signed-off-by: Marek Vasut <marex@denx.de>
Link: https://lore.kernel.org/r/20231027085840.30098-1-marex@denx.de
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Tetsuo reported the following lockdep splat when the TSC synchronization
fails during CPU hotplug:
tsc: Marking TSC unstable due to check_tsc_sync_source failed
WARNING: inconsistent lock state
inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage.
ffffffff8cfa1c78 (watchdog_lock){?.-.}-{2:2}, at: clocksource_watchdog+0x23/0x5a0
{IN-HARDIRQ-W} state was registered at:
_raw_spin_lock_irqsave+0x3f/0x60
clocksource_mark_unstable+0x1b/0x90
mark_tsc_unstable+0x41/0x50
check_tsc_sync_source+0x14f/0x180
sysvec_call_function_single+0x69/0x90
Possible unsafe locking scenario:
lock(watchdog_lock);
<Interrupt>
lock(watchdog_lock);
stack backtrace:
_raw_spin_lock+0x30/0x40
clocksource_watchdog+0x23/0x5a0
run_timer_softirq+0x2a/0x50
sysvec_apic_timer_interrupt+0x6e/0x90
The reason is the recent conversion of the TSC synchronization function
during CPU hotplug on the control CPU to a SMP function call. In case
that the synchronization with the upcoming CPU fails, the TSC has to be
marked unstable via clocksource_mark_unstable().
clocksource_mark_unstable() acquires 'watchdog_lock', but that lock is
taken with interrupts enabled in the watchdog timer callback to minimize
interrupt disabled time. That's obviously a possible deadlock scenario,
Before that change the synchronization function was invoked in thread
context so this could not happen.
As it is not crucical whether the unstable marking happens slightly
delayed, defer the call to a worker thread which avoids the lock context
problem.
Fixes: 9d349d47f0 ("x86/smpboot: Make TSC synchronization function call based")
Reported-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp>
Cc: stable@vger.kernel.org
Link: https://lore.kernel.org/r/87zg064ceg.ffs@tglx
David and a few others reported that on certain newer systems some legacy
interrupts fail to work correctly.
Debugging revealed that the BIOS of these systems leaves the legacy PIC in
uninitialized state which makes the PIC detection fail and the kernel
switches to a dummy implementation.
Unfortunately this fallback causes quite some code to fail as it depends on
checks for the number of legacy PIC interrupts or the availability of the
real PIC.
In theory there is no reason to use the PIC on any modern system when
IO/APIC is available, but the dependencies on the related checks cannot be
resolved trivially and on short notice. This needs lots of analysis and
rework.
The PIC detection has been added to avoid quirky checks and force selection
of the dummy implementation all over the place, especially in VM guest
scenarios. So it's not an option to revert the relevant commit as that
would break a lot of other scenarios.
One solution would be to try to initialize the PIC on detection fail and
retry the detection, but that puts the burden on everything which does not
have a PIC.
Fortunately the ACPI/MADT table header has a flag field, which advertises
in bit 0 that the system is PCAT compatible, which means it has a legacy
8259 PIC.
Evaluate that bit and if set avoid the detection routine and keep the real
PIC installed, which then gets initialized (for nothing) and makes the rest
of the code with all the dependencies work again.
Fixes: e179f69141 ("x86, irq, pic: Probe for legacy PIC and set legacy_pic appropriately")
Reported-by: David Lazar <dlazar@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: David Lazar <dlazar@gmail.com>
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Mario Limonciello <mario.limonciello@amd.com>
Cc: stable@vger.kernel.org
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=218003
Link: https://lore.kernel.org/r/875y2u5s8g.ffs@tglx
Remove unused variables from amdgpu_show_fdinfo
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Umio Yasuno <coelacanth_dream@protonmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Some of the fields that are handled by drm_show_fdinfo() crept back in
when rebasing the patch. Remove them again.
Fixes: 376c25f8ca ("drm/amdgpu: Switch to fdinfo helper")
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Rob Clark <robdclark@chromium.org>
Reviewed-by: <alexander.deucher@amd.com>
Co-developed-by: Umio Yasuno <coelacanth_dream@protonmail.com>
Signed-off-by: Umio Yasuno <coelacanth_dream@protonmail.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
