Jakub Kicinski [Sun, 19 Nov 2023 02:38:05 +0000 (18:38 -0800)]
net: partial revert of the "Make timestamping selectable: series
Revert following commits:
commit acec05fb78ab ("net_tstamp: Add TIMESTAMPING SOFTWARE and HARDWARE mask")
commit 11d55be06df0 ("net: ethtool: Add a command to expose current time stamping layer")
commit bb8645b00ced ("netlink: specs: Introduce new netlink command to get current timestamp")
commit d905f9c75329 ("net: ethtool: Add a command to list available time stamping layers")
commit aed5004ee7a0 ("netlink: specs: Introduce new netlink command to list available time stamping layers")
commit 51bdf3165f01 ("net: Replace hwtstamp_source by timestamping layer")
commit 0f7f463d4821 ("net: Change the API of PHY default timestamp to MAC")
commit 091fab122869 ("net: ethtool: ts: Update GET_TS to reply the current selected timestamp")
commit 152c75e1d002 ("net: ethtool: ts: Let the active time stamping layer be selectable")
commit ee60ea6be0d3 ("netlink: specs: Introduce time stamping set command")
Heiner Kallweit [Mon, 13 Nov 2023 19:13:26 +0000 (20:13 +0100)]
r8169: improve RTL8411b phy-down fixup
Mirsad proposed a patch to reduce the number of spinlock lock/unlock
operations and the function code size. This can be further improved
because the function sets a consecutive register block.
Suggested-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Reviewed-by: Simon Horman <horms@kernel.org> Reviewed-by: Mirsad Todorovac <mirsad.todorovac@alu.unizg.hr> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 18 Nov 2023 17:38:51 +0000 (17:38 +0000)]
Merge branch 'mlxsw-new-reset-flow'
Petr Machata says:
====================
mlxsw: Add support for new reset flow
Ido Schimmel writes:
This patchset changes mlxsw to issue a PCI reset during probe and
devlink reload so that the PCI firmware could be upgraded without a
reboot.
Unlike the old version of this patchset [1], in this version the driver
no longer tries to issue a PCI reset by triggering a PCI link toggle on
its own, but instead calls the PCI core to issue the reset.
The PCI APIs require the device lock to be held which is why patches
Patches #7 adds reset method quirk for NVIDIA Spectrum devices.
Patch #8 adds a debug level print in PCI core so that device ready delay
will be printed even if it is shorter than one second.
Patches #9-#11 are straightforward preparations in mlxsw.
Patch #12 finally implements the new reset flow in mlxsw.
Patch #13 adds PCI reset handlers in mlxsw to avoid user space from
resetting the device from underneath an unaware driver. Instead, the
driver is gracefully de-initialized before the PCI reset and then
initialized again after it.
Patch #14 adds a PCI reset selftest to make sure this code path does not
regress.
Ido Schimmel [Wed, 15 Nov 2023 12:17:23 +0000 (13:17 +0100)]
selftests: mlxsw: Add PCI reset test
Test that PCI reset works correctly by verifying that only the expected
reset methods are supported and that after issuing the reset the ifindex
of the port changes.
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Petr Machata <petrm@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:22 +0000 (13:17 +0100)]
mlxsw: pci: Implement PCI reset handlers
Implement reset_prepare() and reset_done() handlers that are invoked by
the PCI core before and after issuing a PCI reset, respectively.
Specifically, implement reset_prepare() by calling
mlxsw_core_bus_device_unregister() and reset_done() by calling
mlxsw_core_bus_device_register(). This is the same implementation as the
reload_{down,up}() devlink operations with the following differences:
1. The devlink instance is unregistered and then registered again after
the reset.
2. A reset via the device's command interface (using MRSR register) is
not issued during reset_done() as PCI core already issued a PCI
reset.
Tested:
# for i in $(seq 1 10); do echo 1 > /sys/bus/pci/devices/0000\:01\:00.0/reset; done
Reviewed-by: Petr Machata <petrm@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:21 +0000 (13:17 +0100)]
mlxsw: pci: Add support for new reset flow
The driver resets the device during probe and during a devlink reload.
The current reset method reloads the current firmware version or a
pending one, if one was previously flashed using devlink. However, the
current reset method does not result in a PCI hot reset, preventing the
PCI firmware from being upgraded, unless the system is rebooted.
To solve this problem, a new reset command (6) was implemented in the
firmware. Unlike the current command (1), after issuing the new command
the device will not start the reset immediately, but only after a PCI
hot reset.
Implement the new reset method by first verifying that it is supported
by the current firmware version by querying the Management Capabilities
Mask (MCAM) register. If supported, issue the new reset command (6) via
MRSR register followed by a PCI reset by calling
__pci_reset_function_locked().
Once the PCI firmware is operational, go back to the regular reset flow
and wait for the entire device to become ready. That is, repeatedly read
the "system_status" register from the BAR until a value of "FW_READY"
(0x5E) appears.
Tested:
# for i in $(seq 1 10); do devlink dev reload pci/0000:01:00.0; done
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Petr Machata <petrm@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Amit Cohen [Wed, 15 Nov 2023 12:17:20 +0000 (13:17 +0100)]
mlxsw: pci: Move software reset code to a separate function
In general, the existing flow of software reset in the driver is:
1. Wait for system ready status.
2. Send MRSR command, to start the reset.
3. Wait for system ready status.
This flow will be extended once a new reset command is supported. As a
preparation, move step #2 to a separate function.
Signed-off-by: Amit Cohen <amcohen@nvidia.com> Reviewed-by: Petr Machata <petrm@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Amit Cohen [Wed, 15 Nov 2023 12:17:19 +0000 (13:17 +0100)]
mlxsw: pci: Rename mlxsw_pci_sw_reset()
In the next patches, mlxsw_pci_sw_reset() will be extended to support
more reset types and will not necessarily issue a software reset. Rename
the function to reflect that.
Signed-off-by: Amit Cohen <amcohen@nvidia.com> Reviewed-by: Petr Machata <petrm@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Amit Cohen [Wed, 15 Nov 2023 12:17:18 +0000 (13:17 +0100)]
mlxsw: Extend MRSR pack() function to support new commands
Currently mlxsw_reg_mrsr_pack() always sets 'command=1'. As preparation for
support of new reset flow, pass the command as an argument to the
function and add an enum for this field.
For now, always pass 'command=1' to the pack() function.
Signed-off-by: Amit Cohen <amcohen@nvidia.com> Reviewed-by: Petr Machata <petrm@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:17 +0000 (13:17 +0100)]
PCI: Add debug print for device ready delay
Currently, the time it took a PCI device to become ready after reset is
only printed if it was longer than 1000ms ('PCI_RESET_WAIT'). However,
for debugging purposes it is useful to know this time even if it was
shorter. For example, with the device I am working on, hardware
engineers asked to verify that it becomes ready on the first try (no
delay).
To that end, add a debug level print that can be enabled using dynamic
debug. Example:
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Acked-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:16 +0000 (13:17 +0100)]
PCI: Add no PM reset quirk for NVIDIA Spectrum devices
Spectrum-{1,2,3,4} devices report that a D3hot->D0 transition causes a
reset (i.e., they advertise NoSoftRst-). However, this transition does
not have any effect on the device: It continues to be operational and
network ports remain up. Advertising this support makes it seem as if a
PM reset is viable for these devices. Mark it as unavailable to skip it
when testing reset methods.
Before:
# cat /sys/bus/pci/devices/0000\:03\:00.0/reset_method
pm bus
After:
# cat /sys/bus/pci/devices/0000\:03\:00.0/reset_method
bus
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Acked-by: Bjorn Helgaas <bhelgaas@google.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:15 +0000 (13:17 +0100)]
devlink: Add device lock assert in reload operation
Add an assert to verify that the device lock is always held throughout
reload operations.
Tested the following flows with netdevsim and mlxsw while lockdep is
enabled:
netdevsim:
# echo "10 1" > /sys/bus/netdevsim/new_device
# devlink dev reload netdevsim/netdevsim10
# ip netns add bla
# devlink dev reload netdevsim/netdevsim10 netns bla
# ip netns del bla
# echo 10 > /sys/bus/netdevsim/del_device
mlxsw:
# devlink dev reload pci/0000:01:00.0
# ip netns add bla
# devlink dev reload pci/0000:01:00.0 netns bla
# ip netns del bla
# echo 1 > /sys/bus/pci/devices/0000\:01\:00.0/remove
# echo 1 > /sys/bus/pci/rescan
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:14 +0000 (13:17 +0100)]
devlink: Acquire device lock during reload command
Device drivers register with devlink from their probe routines (under
the device lock) by acquiring the devlink instance lock and calling
devl_register().
Drivers that support a devlink reload usually implement the
reload_{down, up}() operations in a similar fashion to their remove and
probe routines, respectively.
However, while the remove and probe routines are invoked with the device
lock held, the reload operations are only invoked with the devlink
instance lock held. It is therefore impossible for drivers to acquire
the device lock from their reload operations, as this would result in
lock inversion.
The motivating use case for invoking the reload operations with the
device lock held is in mlxsw which needs to trigger a PCI reset as part
of the reload. The driver cannot call pci_reset_function() as this
function acquires the device lock. Instead, it needs to call
__pci_reset_function_locked which expects the device lock to be held.
To that end, adjust devlink to always acquire the device lock before the
devlink instance lock when performing a reload.
Do that when reload is explicitly triggered by user space by specifying
the 'DEVLINK_NL_FLAG_NEED_DEV_LOCK' flag in the pre_doit and post_doit
operations of the reload command.
A previous patch already handled the case where reload is invoked as
part of netns dismantle.
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:13 +0000 (13:17 +0100)]
devlink: Allow taking device lock in pre_doit operations
Introduce a new private flag ('DEVLINK_NL_FLAG_NEED_DEV_LOCK') to allow
netlink commands to specify that they need to acquire the device lock in
their pre_doit operation and release it in their post_doit operation.
The reload command will use this flag in the subsequent patch.
No functional changes intended.
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:12 +0000 (13:17 +0100)]
devlink: Enable the use of private flags in post_doit operations
Currently, private flags (e.g., 'DEVLINK_NL_FLAG_NEED_PORT') are only
used in pre_doit operations, but a subsequent patch will need to
conditionally lock and unlock the device lock in pre and post doit
operations, respectively.
As a preparation, enable the use of private flags in post_doit
operations in a similar fashion to how it is done for pre_doit
operations.
No functional changes intended.
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:11 +0000 (13:17 +0100)]
devlink: Acquire device lock during netns dismantle
Device drivers register with devlink from their probe routines (under
the device lock) by acquiring the devlink instance lock and calling
devl_register().
Drivers that support a devlink reload usually implement the
reload_{down, up}() operations in a similar fashion to their remove and
probe routines, respectively.
However, while the remove and probe routines are invoked with the device
lock held, the reload operations are only invoked with the devlink
instance lock held. It is therefore impossible for drivers to acquire
the device lock from their reload operations, as this would result in
lock inversion.
The motivating use case for invoking the reload operations with the
device lock held is in mlxsw which needs to trigger a PCI reset as part
of the reload. The driver cannot call pci_reset_function() as this
function acquires the device lock. Instead, it needs to call
__pci_reset_function_locked which expects the device lock to be held.
To that end, adjust devlink to always acquire the device lock before the
devlink instance lock when performing a reload.
For now, only do that when reload is triggered as part of netns
dismantle. Subsequent patches will handle the case where reload is
explicitly triggered by user space.
Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Ido Schimmel [Wed, 15 Nov 2023 12:17:10 +0000 (13:17 +0100)]
devlink: Move private netlink flags to C file
The flags are not used outside of the C file so move them there.
Suggested-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: Jiri Pirko <jiri@nvidia.com> Signed-off-by: Petr Machata <petrm@nvidia.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 18 Nov 2023 15:00:51 +0000 (15:00 +0000)]
Merge branch 'ncsi-mac-address-command'
Patrick Williams says:
====================
net/ncsi: Add NC-SI 1.2 Get MC MAC Address command
NC-SI 1.2 has now been published[1] and adds a new command for "Get MC
MAC Address". This is often used by BMCs to get the assigned MAC
address for the channel used by the BMC.
This change set has been tested on a Broadcomm 200G NIC with updated
firmware for NC-SI 1.2 and at least one other non-public NIC design.
It serves the exact same function as the existing OEM Get MAC Address
commands, so if a channel reports that it supports NC-SI 1.2, we prefer
to use the standard command rather than the OEM command.
Verified with an invalid MAC address and 2 valid ones:
[ 55.137072] ftgmac100 1e690000.ftgmac eth0: NCSI: Received 3 provisioned MAC addresses
[ 55.137614] ftgmac100 1e690000.ftgmac eth0: NCSI: MAC address 0: 00:00:00:00:00:00
[ 55.138026] ftgmac100 1e690000.ftgmac eth0: NCSI: MAC address 1: fa:ce:b0:0c:20:22
[ 55.138528] ftgmac100 1e690000.ftgmac eth0: NCSI: MAC address 2: fa:ce:b0:0c:20:23
[ 55.139241] ftgmac100 1e690000.ftgmac eth0: NCSI: Unable to assign 00:00:00:00:00:00 to device
[ 55.140098] ftgmac100 1e690000.ftgmac eth0: NCSI: Set MAC address to fa:ce:b0:0c:20:22
Signed-off-by: Peter Delevoryas <peter@pjd.dev> Signed-off-by: Patrick Williams <patrick@stwcx.xyz> Signed-off-by: David S. Miller <davem@davemloft.net>
Peter Delevoryas [Tue, 14 Nov 2023 16:07:34 +0000 (10:07 -0600)]
net/ncsi: Fix netlink major/minor version numbers
The netlink interface for major and minor version numbers doesn't actually
return the major and minor version numbers.
It reports a u32 that contains the (major, minor, update, alpha1)
components as the major version number, and then alpha2 as the minor
version number.
For whatever reason, the u32 byte order was reversed (ntohl): maybe it was
assumed that the encoded value was a single big-endian u32, and alpha2 was
the minor version.
The correct way to get the supported NC-SI version from the network
controller is to parse the Get Version ID response as described in 8.4.44
of the NC-SI spec[1].
The major, minor, and update fields are all binary-coded decimal (BCD)
encoded [2]. The spec provides examples below the Get Version ID response
format in section 8.4.44.1, but for practical purposes, this is an example
from a live network card:
"f" in the upper-nibble means to ignore it, contributing zero.
If both nibbles are "f", I think the whole field is supposed to be ignored.
Major and minor are "required", meaning they're not supposed to be "ff",
but the update field is "optional" so I think it can be ff. I think the
simplest thing to do is just set the major and minor to zero instead of
juggling some conditional logic or something.
bcd2bin() from "include/linux/bcd.h" seems to assume both nibbles are 0-9,
so I've provided a custom BCD decoding function.
Alpha1 and alpha2 are ISO/IEC 8859-1 encoded, which just means ASCII
characters as far as I can tell, although the full encoding table for
non-alphabetic characters is slightly different (I think).
I imagine the alpha fields are just supposed to be alphabetic characters,
but I haven't seen any network cards actually report a non-zero value for
either.
If people wrote software against this netlink behavior, and were parsing
the major and minor versions themselves from the u32, then this would
definitely break their code.
Signed-off-by: Peter Delevoryas <peter@pjd.dev> Fixes: 138635cc27c9 ("net/ncsi: NCSI response packet handler") Signed-off-by: David S. Miller <davem@davemloft.net>
Peter Delevoryas [Tue, 14 Nov 2023 16:07:33 +0000 (10:07 -0600)]
net/ncsi: Simplify Kconfig/dts control flow
Background:
1. CONFIG_NCSI_OEM_CMD_KEEP_PHY
If this is enabled, we send an extra OEM Intel command in the probe
sequence immediately after discovering a channel (e.g. after "Clear
Initial State").
2. CONFIG_NCSI_OEM_CMD_GET_MAC
If this is enabled, we send one of 3 OEM "Get MAC Address" commands from
Broadcom, Mellanox (Nvidida), and Intel in the *configuration* sequence
for a channel.
Cosmo claimed that the Nvidia ConnectX-4 and ConnectX-6 NIC's don't
respond to Get Version ID, et. al in the probe sequence unless you send
the Set MC Affinity command first.
Problem Statement:
We've been using a combination of #ifdef code blocks and IS_ENABLED()
conditions to conditionally send these OEM commands.
It makes adding any new code around these commands hard to understand.
Solution:
In this patch, I just want to remove the conditionally compiled blocks
of code, and always use IS_ENABLED(...) to do dynamic control flow.
I don't think the small amount of code this adds to non-users of the OEM
Kconfigs is a big deal.
Signed-off-by: Peter Delevoryas <peter@pjd.dev> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Sat, 18 Nov 2023 14:52:58 +0000 (14:52 +0000)]
Merge branch 'net-make-timestamping-selectable'
Kory Maincent says:
====================
net: Make timestamping selectable
Up until now, there was no way to let the user select the layer at
which time stamping occurs. The stack assumed that PHY time stamping
is always preferred, but some MAC/PHY combinations were buggy.
This series updates the default MAC/PHY default timestamping and aims to
allow the user to select the desired layer administratively.
Changes in v2:
- Move selected_timestamping_layer variable of the concerned patch.
- Use sysfs_streq instead of strmcmp.
- Use the PHY timestamp only if available.
Changes in v3:
- Expose the PTP choice to ethtool instead of sysfs.
You can test it with the ethtool source on branch feature_ptp of:
https://github.com/kmaincent/ethtool
- Added a devicetree binding to select the preferred timestamp.
Changes in v4:
- Move on to ethtool netlink instead of ioctl.
- Add a netdev notifier to allow packet trapping by the MAC in case of PHY
time stamping.
- Add a PHY whitelist to not break the old PHY default time-stamping
preference API.
Changes in v5:
- Update to ndo_hwstamp_get/set. This bring several new patches.
- Add few patches to make the glue.
- Convert macb to ndo_hwstamp_get/set.
- Add netlink specs description of new ethtool commands.
- Removed netdev notifier.
- Split the patches that expose the timestamping to userspace to separate
the core and ethtool development.
- Add description of software timestamping.
- Convert PHYs hwtstamp callback to use kernel_hwtstamp_config.
Changes in v6:
- Few fixes from the reviews.
- Replace the allowlist to default_timestamp flag to know which phy is
using old API behavior.
- Rename the timestamping layer enum values.
- Move to a simple enum instead of the mix between enum and bitfield.
- Update ts_info and ts-set in software timestamping case.
Changes in v7:
- Fix a temporary build error.
- Link to v6: https://lore.kernel.org/r/20231019-feature_ptp_netnext-v6-0-71affc27b0e5@bootlin.com
====================
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:42 +0000 (12:28 +0100)]
net: ethtool: ts: Update GET_TS to reply the current selected timestamp
As the default selected timestamp API change we have to change also the
timestamp return by ethtool. This patch return now the current selected
timestamp.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:41 +0000 (12:28 +0100)]
net: Change the API of PHY default timestamp to MAC
Change the API to select MAC default time stamping instead of the PHY.
Indeed the PHY is closer to the wire therefore theoretically it has less
delay than the MAC timestamping but the reality is different. Due to lower
time stamping clock frequency, latency in the MDIO bus and no PHC hardware
synchronization between different PHY, the PHY PTP is often less precise
than the MAC. The exception is for PHY designed specially for PTP case but
these devices are not very widespread. For not breaking the compatibility I
introduce a default_timestamp flag in phy_device that is set by the phy
driver to know we are using the old API behavior.
The phy_set_timestamp function is called at each call of phy_attach_direct.
In case of MAC driver using phylink this function is called when the
interface is turned up. Then if the interface goes down and up again the
last choice of timestamp will be overwritten by the default choice.
A solution could be to cache the timestamp status but it can bring other
issues. In case of SFP, if we change the module, it doesn't make sense to
blindly re-set the timestamp back to PHY, if the new module has a PHY with
mediocre timestamping capabilities.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:40 +0000 (12:28 +0100)]
net: Replace hwtstamp_source by timestamping layer
Replace hwtstamp_source which is only used by the kernel_hwtstamp_config
structure by the more widely use timestamp_layer structure. This is done
to prepare the support of selectable timestamping source.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:36 +0000 (12:28 +0100)]
net: ethtool: Add a command to expose current time stamping layer
Time stamping on network packets may happen either in the MAC or in
the PHY, but not both. In preparation for making the choice
selectable, expose both the current layers via ethtool.
In accordance with the kernel implementation as it stands, the current
layer will always read as "phy" when a PHY time stamping device is
present. Future patches will allow changing the current layer
administratively.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:35 +0000 (12:28 +0100)]
net_tstamp: Add TIMESTAMPING SOFTWARE and HARDWARE mask
Timestamping software or hardware flags are often used as a group,
therefore adding these masks will easier future use.
I did not use SOF_TIMESTAMPING_SYS_HARDWARE flag as it is deprecated and
not use at all.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:34 +0000 (12:28 +0100)]
net: phy: micrel: fix ts_info value in case of no phc
In case of no phc we should not return SOFTWARE TIMESTAMPING flags as we do
not know whether the netdev supports of timestamping.
Remove it from the lan8841_ts_info and simply return 0.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:33 +0000 (12:28 +0100)]
net: Make dev_set_hwtstamp_phylib accessible
Make the dev_set_hwtstamp_phylib function accessible in prevision to use
it from ethtool to reset the tstamp current configuration.
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:32 +0000 (12:28 +0100)]
net: macb: Convert to ndo_hwtstamp_get() and ndo_hwtstamp_set()
The hardware timestamping through ndo_eth_ioctl() is going away.
Convert the macb driver to the new API before that can be removed.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The vlan, macvlan and the bonding drivers call their "real" device driver
in order to report the time stamping capabilities. Provide a core
ethtool helper function to avoid copy/paste in the stack.
Signed-off-by: Richard Cochran <richardcochran@gmail.com> Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> Reviewed-by: Jay Vosburgh <jay.vosburgh@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:30 +0000 (12:28 +0100)]
net: phy: Remove the call to phy_mii_ioctl in phy_hwstamp_get/set
__phy_hwtstamp_set function were calling the phy_mii_ioctl function
which will then use the ifreq pointer to call the hwtstamp callback.
Now that ifreq has been removed from the hwstamp callback parameters
it seems more logical to not go through the phy_mii_ioctl function and pass
directly kernel_hwtstamp_config parameter to the hwtstamp callback.
Lets do the same for __phy_hwtstamp_get function and return directly
EOPNOTSUPP as SIOCGHWTSTAMP is not supported for now for the PHYs.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Kory Maincent [Tue, 14 Nov 2023 11:28:29 +0000 (12:28 +0100)]
net: Convert PHYs hwtstamp callback to use kernel_hwtstamp_config
The PHYs hwtstamp callback are still getting the timestamp config from
ifreq and using copy_from/to_user.
Get rid of these functions by using timestamp configuration in parameter.
This also allow to move on to kernel_hwtstamp_config and be similar to
net devices using the new ndo_hwstamp_get/set.
This adds the possibility to manipulate the timestamp configuration
from the kernel which was not possible with the copy_from/to_user.
Signed-off-by: Kory Maincent <kory.maincent@bootlin.com> Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Thu, 16 Nov 2023 08:57:07 +0000 (08:57 +0000)]
gve: add gve_features_check()
It is suboptimal to attempt skb linearization from ndo_start_xmit()
if a gso skb has pathological layout, or if host stack does not have
access to the payload (TCP direct). Linearization of large skbs
can also fail under memory pressure.
We should instead have an ndo_features_check() so that we can
fallback to GSO, which is supported even for TCP direct,
and generally much more efficient (no payload copy).
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: Bailey Forrest <bcf@google.com> Cc: Willem de Bruijn <willemb@google.com> Cc: Jeroen de Borst <jeroendb@google.com> Cc: Praveen Kaligineedi <pkaligineedi@google.com> Cc: Shailend Chand <shailend@google.com> Cc: Ziwei Xiao <ziweixiao@google.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Marco von Rosenberg [Thu, 16 Nov 2023 19:32:31 +0000 (20:32 +0100)]
net: phy: broadcom: Wire suspend/resume for BCM54612E
The BCM54612E ethernet PHY supports IDDQ-SR.
Therefore wire-up the suspend and resume callbacks
to point to bcm54xx_suspend() and bcm54xx_resume().
Signed-off-by: Marco von Rosenberg <marcovr@selfnet.de> Acked-by: Florian Fainelli <florian.fainelli@broadcom.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Shigeru Yoshida [Fri, 17 Nov 2023 00:37:04 +0000 (09:37 +0900)]
tipc: Remove redundant call to TLV_SPACE()
The purpose of TLV_SPACE() is to add the TLV descriptor size to the size of
the TLV value passed as argument and align the resulting size to
TLV_ALIGNTO.
tipc_tlv_alloc() calls TLV_SPACE() on its argument. In other words,
tipc_tlv_alloc() takes its argument as the size of the TLV value. So the
call to TLV_SPACE() in tipc_get_err_tlv() is redundant. Let's remove this
redundancy.
Acked-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Shigeru Yoshida <syoshida@redhat.com> Signed-off-by: David S. Miller <davem@davemloft.net>
The Gigabit Ethernet IP block on the RZ/Five SoC is identical to one
found on the RZ/G2UL SoC. "renesas,r9a07g043-gbeth" compatible string
will be used on the RZ/Five SoC so to make this clear and to keep this
file consistent, update the comment to include RZ/Five SoC.
No driver changes are required as generic compatible string
"renesas,rzg2l-gbeth" will be used as a fallback on RZ/Five SoC.
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com> Acked-by: Conor Dooley <conor.dooley@microchip.com> Reviewed-by: Sergey Shtylyov <s.shtylyov@omp.ru> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King (Oracle) [Wed, 15 Nov 2023 11:39:28 +0000 (11:39 +0000)]
net: sfp: use linkmode_*() rather than open coding
Use the linkmode_*() helpers rather than open coding the calls to the
bitmap operators.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King (Oracle) [Wed, 15 Nov 2023 11:39:23 +0000 (11:39 +0000)]
net: phylink: use linkmode_fill()
Use linkmode_fill() rather than open coding the bitmap operation.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King (Oracle) [Wed, 15 Nov 2023 11:39:18 +0000 (11:39 +0000)]
net: linkmode: add linkmode_fill() helper
Add a linkmode_fill() helper, which will allow us to convert phylink's
open coded bitmap_fill() operations.
Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 14 Nov 2023 17:23:41 +0000 (17:23 +0000)]
tcp: no longer abort SYN_SENT when receiving some ICMP
Currently, non fatal ICMP messages received on behalf
of SYN_SENT sockets do call tcp_ld_RTO_revert()
to implement RFC 6069, but immediately call tcp_done(),
thus aborting the connect() attempt.
Since these Unreachable messages indicate soft error
conditions, TCP MUST NOT abort the connection, and it
SHOULD make the information available to the
application.
This patch makes sure non 'fatal' ICMP[v6] messages do not
abort the connection attempt.
It enables RFC 6069 for SYN_SENT sockets as a result.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: David Morley <morleyd@google.com> Cc: Neal Cardwell <ncardwell@google.com> Cc: Yuchung Cheng <ycheng@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Eric Dumazet [Tue, 14 Nov 2023 17:23:40 +0000 (17:23 +0000)]
tcp: use tp->total_rto to track number of linear timeouts in SYN_SENT state
In commit ccce324dabfe ("tcp: make the first N SYN RTO backoffs linear")
David used icsk->icsk_backoff field to track the number of linear timeouts.
Since then, tp->total_rto has been added.
This commit uses tp->total_rto instead of icsk->icsk_backoff
so that tcp_ld_RTO_revert() no longer can trigger an overflow
in inet_csk_rto_backoff(). Other than the potential UBSAN
report, there was no issue because receiving an ICMP message
currently aborts the connect().
In the following patch, we want to adhere to RFC 6069
and RFC 1122 4.2.3.9.
Signed-off-by: Eric Dumazet <edumazet@google.com> Cc: David Morley <morleyd@google.com> Cc: Neal Cardwell <ncardwell@google.com> Cc: Yuchung Cheng <ycheng@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Alce Lafranque [Tue, 14 Nov 2023 17:36:57 +0000 (11:36 -0600)]
vxlan: add support for flowlabel inherit
By default, VXLAN encapsulation over IPv6 sets the flow label to 0, with
an option for a fixed value. This commits add the ability to inherit the
flow label from the inner packet, like for other tunnel implementations.
This enables devices using only L3 headers for ECMP to correctly balance
VXLAN-encapsulated IPv6 packets.
```
$ ./ip/ip link add dummy1 type dummy
$ ./ip/ip addr add 2001:db8::2/64 dev dummy1
$ ./ip/ip link set up dev dummy1
$ ./ip/ip link add vxlan1 type vxlan id 100 flowlabel inherit remote 2001:db8::1 local 2001:db8::2
$ ./ip/ip link set up dev vxlan1
$ ./ip/ip addr add 2001:db8:1::2/64 dev vxlan1
$ ./ip/ip link set arp off dev vxlan1
$ ping -q 2001:db8:1::1 &
$ tshark -d udp.port==8472,vxlan -Vpni dummy1 -c1
[...]
Internet Protocol Version 6, Src: 2001:db8::2, Dst: 2001:db8::1
0110 .... = Version: 6
.... 0000 0000 .... .... .... .... .... = Traffic Class: 0x00 (DSCP: CS0, ECN: Not-ECT)
.... 0000 00.. .... .... .... .... .... = Differentiated Services Codepoint: Default (0)
.... .... ..00 .... .... .... .... .... = Explicit Congestion Notification: Not ECN-Capable Transport (0)
.... 1011 0001 1010 1111 1011 = Flow Label: 0xb1afb
[...]
Virtual eXtensible Local Area Network
Flags: 0x0800, VXLAN Network ID (VNI)
Group Policy ID: 0
VXLAN Network Identifier (VNI): 100
[...]
Internet Protocol Version 6, Src: 2001:db8:1::2, Dst: 2001:db8:1::1
0110 .... = Version: 6
.... 0000 0000 .... .... .... .... .... = Traffic Class: 0x00 (DSCP: CS0, ECN: Not-ECT)
.... 0000 00.. .... .... .... .... .... = Differentiated Services Codepoint: Default (0)
.... .... ..00 .... .... .... .... .... = Explicit Congestion Notification: Not ECN-Capable Transport (0)
.... 1011 0001 1010 1111 1011 = Flow Label: 0xb1afb
```
Signed-off-by: Alce Lafranque <alce@lafranque.net> Co-developed-by: Vincent Bernat <vincent@bernat.ch> Signed-off-by: Vincent Bernat <vincent@bernat.ch> Reviewed-by: Ido Schimmel <idosch@nvidia.com> Reviewed-by: David Ahern <dsahern@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Lucas Karpinski [Tue, 14 Nov 2023 15:11:31 +0000 (10:11 -0500)]
selftests/net: synchronize udpgro tests' tx and rx connection
The sockets used by udpgso_bench_tx aren't always ready when
udpgso_bench_tx transmits packets. This issue is more prevalent in -rt
kernels, but can occur in both. Replace the hacky sleep calls with a
function that checks whether the ports in the namespace are ready for
use.
Suggested-by: Paolo Abeni <pabeni@redhat.com> Signed-off-by: Lucas Karpinski <lkarpins@redhat.com> Reviewed-by: Willem de Bruijn <willemb@google.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Thu, 16 Nov 2023 22:30:11 +0000 (22:30 +0000)]
Merge branch 'tc-testing-tdc-updates'
Pedro Tammela says:
====================
selftests: tc-testing: updates to tdc
- Patch 1 removes an obscure feature from tdc
- Patch 2 reworks the namespace and devices setup giving a nice speed
boost
- Patch 3 preloads all tc modules when running kselftests
- Patch 4 turns on parallel testing in kselftests
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Pedro Tammela [Tue, 14 Nov 2023 16:04:42 +0000 (13:04 -0300)]
selftests: tc-testing: use parallel tdc in kselftests
Leverage parallel tests in kselftests using all the available cpus.
We tested this in tuxsuite and locally extensively and it seems it's ready for prime time.
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Pedro Tammela [Tue, 14 Nov 2023 16:04:41 +0000 (13:04 -0300)]
selftests: tc-testing: preload all modules in kselftests
While running tdc tests in parallel it can race over the module loading
done by tc and fail the run with random errors.
So avoid this by preloading all modules before running tdc in kselftests.
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Pedro Tammela [Tue, 14 Nov 2023 16:04:40 +0000 (13:04 -0300)]
selftests: tc-testing: rework namespaces and devices setup
As mentioned in the TC Workshop 0x17, our recent changes to tdc broke
downstream CI systems like tuxsuite. The issue is the classic problem
with rcu/workqueue objects where you can miss them if not enough wall time
has passed. The latter is subjective to the system and kernel config,
in my machine could be nanoseconds while in another could be microseconds
or more.
In order to make the suite deterministic, poll for the existence
of the objects in a reasonable manner. Talking netlink directly is the
the best solution in order to avoid paying the cost of multiple
'fork()' calls, so introduce a netlink based setup routine using
pyroute2. We leave the iproute2 one as a fallback when pyroute2 is not
available.
Also rework the iproute2 side to mimic the netlink routine where it
creates DEV0 as the peer of DEV1 and moves DEV1 into the net namespace.
This way when the namespace is deleted DEV0 is also deleted
automatically, leaving no margin for resource leaks.
Another bonus of this change is that our setup time sped up by a factor
of 2 when using netlink.
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Pedro Tammela [Tue, 14 Nov 2023 16:04:39 +0000 (13:04 -0300)]
selftests: tc-testing: drop '-N' argument from nsPlugin
This argument would bypass the net namespace creation and run the test in
the root namespace, even if nsPlugin was specified.
Drop it as it's the same as commenting out the nsPlugin from a test and adds
additional complexity to the plugin code.
Signed-off-by: Pedro Tammela <pctammela@mojatatu.com> Reviewed-by: Simon Horman <horms@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Christian Marangi [Tue, 14 Nov 2023 14:08:44 +0000 (15:08 +0100)]
dt-bindings: Document Marvell Aquantia PHY
Document bindings for Marvell Aquantia PHY.
The Marvell Aquantia PHY require a firmware to work correctly and there
at least 3 way to load this firmware.
Describe all the different way and document the binding "firmware-name"
to load the PHY firmware from userspace.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Reviewed-by: Conor Dooley <conor.dooley@microchip.com> Acked-by: Rob Herring <robh@kernel.org> Signed-off-by: David S. Miller <davem@davemloft.net>
Robert Marko [Tue, 14 Nov 2023 14:08:43 +0000 (15:08 +0100)]
net: phy: aquantia: add firmware load support
Aquantia PHY-s require firmware to be loaded before they start operating.
It can be automatically loaded in case when there is a SPI-NOR connected
to Aquantia PHY-s or can be loaded from the host via MDIO.
This patch adds support for loading the firmware via MDIO as in most cases
there is no SPI-NOR being used to save on cost.
Firmware loading code itself is ported from mainline U-boot with cleanups.
The firmware has mixed values both in big and little endian.
PHY core itself is big-endian but it expects values to be in little-endian.
The firmware is little-endian but CRC-16 value for it is stored at the end
of firmware in big-endian.
It seems the PHY does the conversion internally from firmware that is
little-endian to the PHY that is big-endian on using the mailbox
but mailbox returns a big-endian CRC-16 to verify the written data
integrity.
Co-developed-by: Christian Marangi <ansuelsmth@gmail.com> Signed-off-by: Robert Marko <robimarko@gmail.com> Signed-off-by: Christian Marangi <ansuelsmth@gmail.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
Cleanup and optimizations to transmit code
Pad small packets to ETH_ZLEN before transmit, cleanup dma sync calls,
add xmit_more functionality and then further remove atomic
variable usage in the prior.
Changes:
V3:
- Stop returning NETDEV_TX_BUSY when ring is full in xmit_patch.
Change to inspect early if next packet can fit in ring instead of
current packet, and stop queue if not.
- Add smp_mb between stopping tx queue and checking if tx queue has
free entries again, in queue full check function to let reflect
IQ process completions that might have happened on other cpus.
- Update small packet padding patch changelog to give more info.
V2: https://lore.kernel.org/all/20231024145119.2366588-1-srasheed@marvell.com/
- Added patch for padding small packets to ETH_ZLEN, part of
optimization patches for transmit code missed out in V1
- Updated changelog to provide more details for dma_sync remove patch
- Updated changelog to use imperative tone in add xmit_more patch
V1: https://lore.kernel.org/all/20231023114449.2362147-1-srasheed@marvell.com/
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Shinas Rasheed [Tue, 14 Nov 2023 13:45:35 +0000 (05:45 -0800)]
octeon_ep: remove atomic variable usage in Tx data path
Replace atomic variable "instr_pending" which represents number of
posted tx instructions pending completion, with host_write_idx and
flush_index variables in the xmit and completion processing respectively.
Signed-off-by: Shinas Rasheed <srasheed@marvell.com> Signed-off-by: David S. Miller <davem@davemloft.net>
- af_unix: fix use-after-free in unix_stream_read_actor()
- tipc: fix kernel-infoleak due to uninitialized TLV value
- eth: bonding: stop the device in bond_setup_by_slave()
- eth: mlx5:
- fix double free of encap_header
- avoid referencing skb after free-ing in drop path
- eth: hns3: fix VF reset
- eth: mvneta: fix calls to page_pool_get_stats
Previous releases - always broken:
- core: set SOCK_RCU_FREE before inserting socket into hashtable
- bpf: fix control-flow graph checking in privileged mode
- eth: ppp: limit MRU to 64K
- eth: stmmac: avoid rx queue overrun
- eth: icssg-prueth: fix error cleanup on failing initialization
- eth: hns3: fix out-of-bounds access may occur when coalesce info is
read via debugfs
- eth: cortina: handle large frames
Misc:
- selftests: gso: support CONFIG_MAX_SKB_FRAGS up to 45"
* tag 'net-6.7-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (78 commits)
macvlan: Don't propagate promisc change to lower dev in passthru
net: sched: do not offload flows with a helper in act_ct
net/mlx5e: Check return value of snprintf writing to fw_version buffer for representors
net/mlx5e: Check return value of snprintf writing to fw_version buffer
net/mlx5e: Reduce the size of icosq_str
net/mlx5: Increase size of irq name buffer
net/mlx5e: Update doorbell for port timestamping CQ before the software counter
net/mlx5e: Track xmit submission to PTP WQ after populating metadata map
net/mlx5e: Avoid referencing skb after free-ing in drop path of mlx5e_sq_xmit_wqe
net/mlx5e: Don't modify the peer sent-to-vport rules for IPSec offload
net/mlx5e: Fix pedit endianness
net/mlx5e: fix double free of encap_header in update funcs
net/mlx5e: fix double free of encap_header
net/mlx5: Decouple PHC .adjtime and .adjphase implementations
net/mlx5: DR, Allow old devices to use multi destination FTE
net/mlx5: Free used cpus mask when an IRQ is released
Revert "net/mlx5: DR, Supporting inline WQE when possible"
bpf: Do not allocate percpu memory at init stage
net: Fix undefined behavior in netdev name allocation
dt-bindings: net: ethernet-controller: Fix formatting error
...
Linus Torvalds [Thu, 16 Nov 2023 12:39:37 +0000 (07:39 -0500)]
Merge tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost
Pull virtio fixes from Michael Tsirkin:
"Bugfixes all over the place"
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost:
vhost-vdpa: fix use after free in vhost_vdpa_probe()
virtio_pci: Switch away from deprecated irq_set_affinity_hint
riscv, qemu_fw_cfg: Add support for RISC-V architecture
vdpa_sim_blk: allocate the buffer zeroed
virtio_pci: move structure to a header
Paolo Abeni [Thu, 16 Nov 2023 10:02:52 +0000 (11:02 +0100)]
Merge tag 'nf-23-11-15' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Pablo Neira Ayuso says:
====================
Netfilter fixes for net
The following patchset contains Netfilter fixes for net:
1) Remove unused variable causing compilation warning in nft_set_rbtree,
from Yang Li. This unused variable is a left over from previous
merge window.
2) Possible return of uninitialized in nf_conntrack_bridge, from
Linkui Xiao. This is there since nf_conntrack_bridge is available.
3) Fix incorrect pointer math in nft_byteorder, from Dan Carpenter.
Problem has been there since 2016.
4) Fix bogus error in destroy set element command. Problem is there
since this new destroy command was added.
5) Fix race condition in ipset between swap and destroy commands and
add/del/test control plane. This problem is there since ipset was
merged.
6) Split async and sync catchall GC in two function to fix unsafe
iteration over RCU. This is a fix-for-fix that was included in
the previous pull request.
* tag 'nf-23-11-15' of git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
netfilter: nf_tables: split async and sync catchall in two functions
netfilter: ipset: fix race condition between swap/destroy and kernel side add/del/test
netfilter: nf_tables: bogus ENOENT when destroying element which does not exist
netfilter: nf_tables: fix pointer math issue in nft_byteorder_eval()
netfilter: nf_conntrack_bridge: initialize err to 0
netfilter: nft_set_rbtree: Remove unused variable nft_net
====================
Vlad Buslov [Tue, 14 Nov 2023 17:59:15 +0000 (18:59 +0100)]
macvlan: Don't propagate promisc change to lower dev in passthru
Macvlan device in passthru mode sets its lower device promiscuous mode
according to its MACVLAN_FLAG_NOPROMISC flag instead of synchronizing it to
its own promiscuity setting. However, macvlan_change_rx_flags() function
doesn't check the mode before propagating such changes to the lower device
which can cause net_device->promiscuity counter overflow as illustrated by
reproduction example [0] and resulting dmesg log [1]. Fix the issue by
first verifying the mode in macvlan_change_rx_flags() function before
propagating promiscuous mode change to the lower device.
[0]:
ip link add macvlan1 link enp8s0f0 type macvlan mode passthru
ip link set macvlan1 promisc on
ip l set dev macvlan1 up
ip link set macvlan1 promisc off
ip l set dev macvlan1 down
ip l set dev macvlan1 up
Xin Long [Mon, 13 Nov 2023 17:53:28 +0000 (12:53 -0500)]
net: sched: do not offload flows with a helper in act_ct
There is no hardware supporting ct helper offload. However, prior to this
patch, a flower filter with a helper in the ct action can be successfully
set into the HW, for example (eth1 is a bnxt NIC):
# tc qdisc add dev eth1 ingress_block 22 ingress
# tc filter add block 22 proto ip flower skip_sw ip_proto tcp \
dst_port 21 ct_state -trk action ct helper ipv4-tcp-ftp
# tc filter show dev eth1 ingress
Rahul Rameshbabu [Tue, 14 Nov 2023 21:58:46 +0000 (13:58 -0800)]
net/mlx5e: Check return value of snprintf writing to fw_version buffer for representors
Treat the operation as an error case when the return value is equivalent to
the size of the name buffer. Failed to write null terminator to the name
buffer, making the string malformed and should not be used. Provide a
string with only the firmware version when forming the string with the
board id fails. This logic for representors is identical to normal flow
with ethtool.
Without check, will trigger -Wformat-truncation with W=1.
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c: In function 'mlx5e_rep_get_drvinfo':
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c:78:31: warning: '%.16s' directive output may be truncated writing up to 16 bytes into a region of size between 13 and 22 [-Wformat-truncation=]
78 | "%d.%d.%04d (%.16s)",
| ^~~~~
drivers/net/ethernet/mellanox/mlx5/core/en_rep.c:77:9: note: 'snprintf' output between 12 and 37 bytes into a destination of size 32
77 | snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
78 | "%d.%d.%04d (%.16s)",
| ~~~~~~~~~~~~~~~~~~~~~
79 | fw_rev_maj(mdev), fw_rev_min(mdev),
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
80 | fw_rev_sub(mdev), mdev->board_id);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Rahul Rameshbabu [Tue, 14 Nov 2023 21:58:45 +0000 (13:58 -0800)]
net/mlx5e: Check return value of snprintf writing to fw_version buffer
Treat the operation as an error case when the return value is equivalent to
the size of the name buffer. Failed to write null terminator to the name
buffer, making the string malformed and should not be used. Provide a
string with only the firmware version when forming the string with the
board id fails.
Without check, will trigger -Wformat-truncation with W=1.
drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c: In function 'mlx5e_ethtool_get_drvinfo':
drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c:49:31: warning: '%.16s' directive output may be truncated writing up to 16 bytes into a region of size between 13 and 22 [-Wformat-truncation=]
49 | "%d.%d.%04d (%.16s)",
| ^~~~~
drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c:48:9: note: 'snprintf' output between 12 and 37 bytes into a destination of size 32
48 | snprintf(drvinfo->fw_version, sizeof(drvinfo->fw_version),
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
49 | "%d.%d.%04d (%.16s)",
| ~~~~~~~~~~~~~~~~~~~~~
50 | fw_rev_maj(mdev), fw_rev_min(mdev), fw_rev_sub(mdev),
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
51 | mdev->board_id);
| ~~~~~~~~~~~~~~~
Saeed Mahameed [Tue, 14 Nov 2023 21:58:44 +0000 (13:58 -0800)]
net/mlx5e: Reduce the size of icosq_str
icosq_str size is unnecessarily too long, and it causes a build warning
-Wformat-truncation with W=1. Looking closely, It doesn't need to be 255B,
hence this patch reduces the size to 32B which should be more than enough
to host the string: "ICOSQ: 0x%x, ".
While here, add a missing space in the formatted string.
This fixes the following build warning:
$ KCFLAGS='-Wall -Werror'
$ make O=/tmp/kbuild/linux W=1 -s -j12 drivers/net/ethernet/mellanox/mlx5/core/
drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c: In function 'mlx5e_reporter_rx_timeout':
drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c:718:56:
error: ', CQ: 0x' directive output may be truncated writing 8 bytes into a region of size between 0 and 255 [-Werror=format-truncation=]
718 | "RX timeout on channel: %d, %sRQ: 0x%x, CQ: 0x%x",
| ^~~~~~~~
drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c:717:9: note: 'snprintf' output between 43 and 322 bytes into a destination of size 288
717 | snprintf(err_str, sizeof(err_str),
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
718 | "RX timeout on channel: %d, %sRQ: 0x%x, CQ: 0x%x",
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
719 | rq->ix, icosq_str, rq->rqn, rq->cq.mcq.cqn);
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Rahul Rameshbabu [Tue, 14 Nov 2023 21:58:43 +0000 (13:58 -0800)]
net/mlx5: Increase size of irq name buffer
Without increased buffer size, will trigger -Wformat-truncation with W=1
for the snprintf operation writing to the buffer.
drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c: In function 'mlx5_irq_alloc':
drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c:296:7: error: '@pci:' directive output may be truncated writing 5 bytes into a region of size between 1 and 32 [-Werror=format-truncation=]
296 | "%s@pci:%s", name, pci_name(dev->pdev));
| ^~~~~
drivers/net/ethernet/mellanox/mlx5/core/pci_irq.c:295:2: note: 'snprintf' output 6 or more bytes (assuming 37) into a destination of size 32
295 | snprintf(irq->name, MLX5_MAX_IRQ_NAME,
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
296 | "%s@pci:%s", name, pci_name(dev->pdev));
| ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Rahul Rameshbabu [Tue, 14 Nov 2023 21:58:42 +0000 (13:58 -0800)]
net/mlx5e: Update doorbell for port timestamping CQ before the software counter
Previously, mlx5e_ptp_poll_ts_cq would update the device doorbell with the
incremented consumer index after the relevant software counters in the
kernel were updated. In the mlx5e_sq_xmit_wqe context, this would lead to
either overrunning the device CQ or exceeding the expected software buffer
size in the device CQ if the device CQ size was greater than the software
buffer size. Update the relevant software counter only after updating the
device CQ consumer index in the port timestamping napi_poll context.
Rahul Rameshbabu [Tue, 14 Nov 2023 21:58:41 +0000 (13:58 -0800)]
net/mlx5e: Track xmit submission to PTP WQ after populating metadata map
Ensure the skb is available in metadata mapping to skbs before tracking the
metadata index for detecting undelivered CQEs. If the metadata index is put
in the tracking list before putting the skb in the map, the metadata index
might be used for detecting undelivered CQEs before the relevant skb is
available in the map, which can lead to a null-ptr-deref.
Rahul Rameshbabu [Tue, 14 Nov 2023 21:58:40 +0000 (13:58 -0800)]
net/mlx5e: Avoid referencing skb after free-ing in drop path of mlx5e_sq_xmit_wqe
When SQ is a port timestamping SQ for PTP, do not access tx flags of skb
after free-ing the skb. Free the skb only after all references that depend
on it have been handled in the dropped WQE path.
Fixes: 3178308ad4ca ("net/mlx5e: Make tx_port_ts logic resilient to out-of-order CQEs") Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Link: https://lore.kernel.org/r/20231114215846.5902-10-saeed@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Jianbo Liu [Tue, 14 Nov 2023 21:58:39 +0000 (13:58 -0800)]
net/mlx5e: Don't modify the peer sent-to-vport rules for IPSec offload
As IPSec packet offload in switchdev mode is not supported with LAG,
it's unnecessary to modify those sent-to-vport rules to the peer eswitch.
Fixes: c6c2bf5db4ea ("net/mlx5e: Support IPsec packet offload for TX in switchdev mode") Signed-off-by: Jianbo Liu <jianbol@nvidia.com> Reviewed-by: Leon Romanovsky <leonro@nvidia.com> Reviewed-by: Roi Dayan <roid@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Link: https://lore.kernel.org/r/20231114215846.5902-9-saeed@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Vlad Buslov [Tue, 14 Nov 2023 21:58:38 +0000 (13:58 -0800)]
net/mlx5e: Fix pedit endianness
Referenced commit addressed endianness issue in mlx5 pedit implementation
in ad hoc manner instead of systematically treating integer values
according to their types which left pedit fields of sizes not equal to 4
and where the bytes being modified are not least significant ones broken on
big endian machines since wrong bits will be consumed during parsing which
leads to following example error when applying pedit to source and
destination MAC addresses:
Treat masks and values of pedit and filter match as network byte order,
refactor pointers to them to void pointers instead of confusing u32
pointers and only cast to pointer-to-integer when reading a value from
them. Treat pedit mlx5_fields->field_mask as host byte order according to
its type u32, change the constants in fields array accordingly.
Fixes: 82198d8bcdef ("net/mlx5e: Fix endianness when calculating pedit mask first bit") Signed-off-by: Vlad Buslov <vladbu@nvidia.com> Reviewed-by: Gal Pressman <gal@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Link: https://lore.kernel.org/r/20231114215846.5902-8-saeed@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Gavin Li [Tue, 14 Nov 2023 21:58:37 +0000 (13:58 -0800)]
net/mlx5e: fix double free of encap_header in update funcs
Follow up to the previous patch to fix the same issue for
mlx5e_tc_tun_update_header_ipv4{6} when mlx5_packet_reformat_alloc()
fails.
When mlx5_packet_reformat_alloc() fails, the encap_header allocated in
mlx5e_tc_tun_update_header_ipv4{6} will be released within it. However,
e->encap_header is already set to the previously freed encap_header
before mlx5_packet_reformat_alloc(). As a result, the later
mlx5e_encap_put() will free e->encap_header again, causing a double free
issue.
Dust Li [Tue, 14 Nov 2023 21:58:36 +0000 (13:58 -0800)]
net/mlx5e: fix double free of encap_header
When mlx5_packet_reformat_alloc() fails, the encap_header allocated in
mlx5e_tc_tun_create_header_ipv4{6} will be released within it. However,
e->encap_header is already set to the previously freed encap_header
before mlx5_packet_reformat_alloc(). As a result, the later
mlx5e_encap_put() will free e->encap_header again, causing a double free
issue.
Rahul Rameshbabu [Tue, 14 Nov 2023 21:58:35 +0000 (13:58 -0800)]
net/mlx5: Decouple PHC .adjtime and .adjphase implementations
When running a phase adjustment operation, the free running clock should
not be modified at all. The phase control keyword is intended to trigger an
internal servo on the device that will converge to the provided delta. A
free running counter cannot implement phase adjustment.
Fixes: 8e11a68e2e8a ("net/mlx5: Add adjphase function to support hardware-only offset control") Signed-off-by: Rahul Rameshbabu <rrameshbabu@nvidia.com> Reviewed-by: Tariq Toukan <tariqt@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Link: https://lore.kernel.org/r/20231114215846.5902-5-saeed@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Erez Shitrit [Tue, 14 Nov 2023 21:58:34 +0000 (13:58 -0800)]
net/mlx5: DR, Allow old devices to use multi destination FTE
The current check isn't aware of old devices that don't have the
relevant FW capability. This patch allows multi destination FTE
in old cards, as it was before this check.
Fixes: f6f46e7173cb ("net/mlx5: DR, Add check for multi destination FTE") Signed-off-by: Erez Shitrit <erezsh@nvidia.com> Reviewed-by: Yevgeny Kliteynik <kliteyn@nvidia.com> Signed-off-by: Saeed Mahameed <saeedm@nvidia.com> Link: https://lore.kernel.org/r/20231114215846.5902-4-saeed@kernel.org Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Maher Sanalla [Tue, 14 Nov 2023 21:58:33 +0000 (13:58 -0800)]
net/mlx5: Free used cpus mask when an IRQ is released
Each EQ table maintains a cpumask of the already used CPUs that are mapped
to IRQs to ensure that each IRQ gets mapped to a unique CPU.
However, on IRQ release, the said cpumask is not updated by clearing the
CPU from the mask to allow future IRQ request, causing the following
error when a SF is reloaded after it has utilized all CPUs for its IRQs:
mlx5_irq_affinity_request:135:(pid 306010): Didn't find a matching IRQ.
err = -28
Thus, when releasing an IRQ, clear its mapped CPU from the used CPUs
mask, to prevent the case described above.
While at it, move the used cpumask update to the EQ layer as it is more
fitting and preserves symmetricity of the IRQ request/release API.
Itamar Gozlan [Tue, 14 Nov 2023 21:58:32 +0000 (13:58 -0800)]
Revert "net/mlx5: DR, Supporting inline WQE when possible"
This reverts commit 95c337cce0e11d06a715da73e6796ade9216637f.
The revert is required due to the suspicion it cause some tests
fail and will be moved to further investigation.
We've added 7 non-merge commits during the last 6 day(s) which contain
a total of 9 files changed, 200 insertions(+), 49 deletions(-).
The main changes are:
1) Do not allocate bpf specific percpu memory unconditionally, from Yonghong.
2) Fix precision backtracking instruction iteration, from Andrii.
3) Fix control flow graph checking, from Andrii.
4) Fix xskxceiver selftest build, from Anders.
* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
bpf: Do not allocate percpu memory at init stage
selftests/bpf: add more test cases for check_cfg()
bpf: fix control-flow graph checking in privileged mode
selftests/bpf: add edge case backtracking logic test
bpf: fix precision backtracking instruction iteration
bpf: handle ldimm64 properly in check_cfg()
selftests: bpf: xskxceiver: ksft_print_msg: fix format type error
====================
Rahul Rameshbabu [Thu, 30 Mar 2023 22:05:38 +0000 (15:05 -0700)]
net/mlx5: Query maximum frequency adjustment of the PTP hardware clock
Some mlx5 devices do not support the default advertised maximum frequency
adjustment value for the PTP hardware clock that is set by the driver.
These devices need to be queried when initializing the clock functionality
in order to get the maximum supported frequency adjustment value. This
value can be greater than the minimum supported frequency adjustment across
mlx5 devices (50 million ppb).
Rahul Rameshbabu [Fri, 25 Aug 2023 03:29:50 +0000 (20:29 -0700)]
net/mlx5: Convert scaled ppm values outside the s32 range for PHC frequency adjustments
Represent scaled ppm as ppb to the device when the value in scaled ppm is
not representable as a 32-bit signed integer. mlx5 devices only support a
32-bit field for the frequency adjustment value in units of either scaled
ppm or ppb.
Since mlx5 devices only support a 32-bit field for the frequency adjustment
value independent of unit used, limit the maximum frequency adjustment to
S32_MAX ppb.
Configure the PHC inside mlx5_init_timer_clock for calling mlx5_ptp_settime
later in the function. Would previously use mlx5_ptp_clock_info instance to
invoke mlx5_ptp_settime to set the NIC real-time clock to be synchronized
with the host system clock.
net/mlx5: Refactor real time clock operation checks for PHC
Check if the MTUTC register of the NIC can be modified before attempting to
execute a real-time clock operation. Previous implementation aborted the
real-time clock operation pre-emptively when the MTUTC register used to
control the real-time clock was not modifiable, indicating real-time clock
mode was not enabled on the NIC. The original control flow was confusing
since the noop-if-RTC-disabled branch looked similar to an error handling
guard clause. The purpose of this patch is purely for improving readability
and should lead to no functional change.
projects / users / dwmw2 / linux.git / log