Add support for SO_MARK to the CAN_RAW protocol. This makes it
possible to add traffic control filters based on the fwmark.
Link: https://lore.kernel.org/all/20221210113653.170346-1-mkl@pengutronix.de
Acked-by: Oliver Hartkopp <socketcan@hartkopp.net>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
When we try to access the mcan message ram addresses during the probe,
hclk is gated by any other drivers or disabled, because of that probe
gets failed.
Move the mram init functionality to mcan chip config called by
m_can_start from mcan open function, by that time clocks are
enabled.
Suggested-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Vivek Yadav <vivek.2311@samsung.com>
Link: https://lore.kernel.org/all/20221207100632.96200-2-vivek.2311@samsung.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Vincent Mailhol <mailhol.vincent@wanadoo.fr> says:
The gs_can and ucan drivers keep a pointer to struct usb_interface in
their private structure. This is not needed. For gs_can the only use
is to retrieve struct usb_device, which is already available in
gs_usb::udev. For ucan, the field is set but never used.
Remove the struct usb_interface fields and clean up.
Link: https://lore.kernel.org/all/20221208081142.16936-1-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The iface field of struct gs_can is only used to retrieve the
usb_device which is already available in gs_can::udev.
Replace each occurrence of interface_to_usbdev(dev->iface) with
dev->udev. This done, remove gs_can::iface.
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221208081142.16936-3-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.
Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Link: https://lore.kernel.org/all/202212070909095189693@zte.com.cn
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Vincent Mailhol <mailhol.vincent@wanadoo.fr> says:
The goal of this series is to report the firmware version, the
bootloader version and the hardware revision of ETAS ES58x devices.
These are already reported in the kernel log but this isn't best
practice. Remove the kernel log and instead export all these through
devlink. The devlink core automatically exports the firmware and the
bootloader version to ethtool, so no need to implement the
ethtool_ops::get_drvinfo() callback anymore.
Patch one and two implement the core support for devlink (at device
level) and devlink port (at the network interface level).
Patch three export usb_cache_string() and patch four add a new info
attribute to devlink.h. Both are prerequisites for patch five.
Patch five is the actual goal: it parses the product information from
a custom usb string returned by the device and expose them through
devlink.
Patch six removes the product information from the kernel log.
Finally, patch seven add a devlink documentation page with list all
the information attributes reported by the driver.
* Sample outputs following this series *
| $ devlink dev info
| usb/1-9:1.1:
| serial_number 0108954
| versions:
| fixed:
| board.rev B012/000
| running:
| fw 04.00.01
| fw.bootloader 02.00.00
| $ devlink port show can0
| usb/1-9:1.1/0: type eth netdev can0 flavour physical port 0 splittable false
| $ ethtool -i can0
| driver: etas_es58x
| version: 6.1.0-rc7+
| firmware-version: 04.00.01 02.00.00
| expansion-rom-version:
| bus-info: 1-9:1.1
| supports-statistics: no
| supports-test: no
| supports-eeprom-access: no
| supports-register-dump: no
| supports-priv-flags: no
* Changelog *
v4 -> v5: https://lore.kernel.org/all/20221130174658.29282-1-mailhol.vincent@wanadoo.fr
* [PATH 2/7] add devlink port support. This extends devlink to the
network interface.
* thanks to devlink port, 'ethtool -i' is now able to retrieve the
firmware version from devlink. No need to implement the
ethtool_ops::get_drvinfo() callback anymore: remove one patch from
the series.
* [PATCH 4/7] A new patch to add a new info attribute for the
bootloader version in devlink.h. This patch was initially sent as
a standalone patch here:
https://lore.kernel.org/netdev/20221129031406.3849872-1-mailhol.vincent@wanadoo.fr
Merging it to this series so that it is both added and used at the
same time.
* [PATCH 5/7] use the newly info attribute defined in patch 4/7 to
report the bootloader version instead of the custom string "bl".
* [PATCH 5/7] because the series does not implement
ethtool_ops::get_drvinfo() anymore, the two helper functions
es58x_sw_version_is_set() and es58x_hw_revision_is_set() are only
used in devlink.c. Move them from es58x_core.h to es58x_devlink.c.
* [PATCH 5/7] small rework of the helper function
es58x_hw_revision_is_set(): it is OK to only check the letter (if
the letter is '\0', it will not be possible to print the next
numbers).
* [PATCH 5/7 and 6/7] add reviewed-by Andrew Lunn tag.
* [PATCH 7/7] Now, 'ethtool -i' reports both the firmware version
and the bootloader version (this is how the core export the
information from devlink to ethtool). Update the documentation to
reflect this fact.
* Reoder the patches.
v3 -> v4: https://lore.kernel.org/all/20221126162211.93322-1-mailhol.vincent@wanadoo.fr
* major rework to use devlink instead of sysfs following Andrew's
comment.
* split the series in 6 patches.
* [PATCH 1/6] add Acked-by: Greg Kroah-Hartman
v2 -> v3: https://lore.kernel.org/all/20221113040108.68249-1-mailhol.vincent@wanadoo.fr
* patch 2/3: do not spam the kernel log anymore with the product
number. Instead parse the product information string, extract the
firmware version, the bootloadar version and the hardware revision
and export them through sysfs.
* patch 2/3: rework the parsing in order not to need additional
fields in struct es58x_parameters.
* patch 3/3: only populate ethtool_drvinfo::fw_version because since
commit edaf5df22c ("ethtool: ethtool_get_drvinfo: populate
drvinfo fields even if callback exits"), there is no need to
populate ethtool_drvinfo::driver and ethtool_drvinfo::bus_info in
the driver.
v1 -> v2: https://lore.kernel.org/all/20221104171604.24052-1-mailhol.vincent@wanadoo.fr
* was a single patch. It is now a series of three patches.
* add a first new patch to export usb_cache_string().
* add a second new patch to apply usb_cache_string() to existing code.
* add missing check on product info string to prevent a buffer overflow.
* add comma on the last entry of struct es58x_parameters.
v1: https://lore.kernel.org/all/20221104073659.414147-1-mailhol.vincent@wanadoo.fr
Link: https://lore.kernel.org/all/20221130174658.29282-1-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
List all the version information reported by the etas_es58x driver
through devlink. Also, update MAINTAINERS with the newly created file.
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-8-mailhol.vincent@wanadoo.fr
[mkl: fixed version information table: "bl" -> "fw.bootloader"
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Now that the product information are available under devlink, no more
need to print them in the kernel log. Remove es58x_get_product_info().
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20221130174658.29282-7-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
ES58x devices report below product information through a custom usb
string:
* the firmware version
* the bootloader version
* the hardware revision
Parse this string, store the results in struct es58x_dev, export:
* the firmware version through devlink's "fw" name
* the bootloader version through devlink's "fw.bootloader" name
* the hardware revisionthrough devlink's "board.rev" name
Those devlink entries are not critical to use the device, if parsing
fails, print an informative log message and continue to probe the
device.
In addition to that, use usb_device::serial to report the device
serial number.
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/all/20221130174658.29282-6-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
As discussed in [1], abbreviating the bootloader to "bl" might not be
well understood. Instead, a bootloader technically being a firmware,
name it "fw.bootloader".
Add a new macro to devlink.h to formalize this new info attribute name
and update the documentation.
[1] https://lore.kernel.org/netdev/20221128142723.2f826d20@kernel.org/
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-5-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Convert platform_get_resource(), devm_ioremap_resource() to a single
call to devm_platform_get_and_ioremap_resource(), as this is exactly
what this function does.
Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Minghao Chi <chi.minghao@zte.com.cn>
Link: https://lore.kernel.org/all/202211111443005202576@zte.com.cn
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
usb_cache_string() can also be useful for the drivers so export it.
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Link: https://lore.kernel.org/all/20221130174658.29282-4-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The CANFD block on the RZ/Five SoC is identical to one found on the
RZ/G2UL SoC. "renesas,r9a07g043-canfd" compatible string will be used
on the RZ/Five SoC so to make this clear, update the comment to include
RZ/Five SoC.
No driver changes are required as generic compatible string
"renesas,rzg2l-canfd" will be used as a fallback on RZ/Five SoC.
Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Acked-by: Rob Herring <robh@kernel.org>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/all/20221115123811.1182922-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Add support for devlink port which extends the devlink support to the
network interface level. For now, the etas_es58x driver will only rely
on the default features that devlink port has to offer and not
implement additional feature ones.
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-3-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Add basic support for devlink at the device level. The callbacks of
struct devlink_ops will be implemented next.
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221130174658.29282-2-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
IMX93 do not contain a GPR to config the stop mode, it will set
the flexcan into stop mode automatically once the ARM core go
into low power mode (WFI instruct) and gate off the flexcan
related clock automatically. But to let these logic work as
expect, before ARM core go into low power mode, need to make
sure the flexcan related clock keep on.
To support stop mode and wakeup feature on imx93, this patch
add a new fsl_imx93_devtype_data to separate from imx8mp.
Signed-off-by: Haibo Chen <haibo.chen@nxp.com>
Link: https://lore.kernel.org/all/1669116752-4260-1-git-send-email-haibo.chen@nxp.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Follow the best practices, reorder the includes.
While doing so, bump up copyright year of each modified files.
Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
Link: https://lore.kernel.org/all/20221126160525.87036-1-mailhol.vincent@wanadoo.fr
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Since commit 0166dc11be ("of: make CONFIG_OF user selectable"), it
is possible to test-build any driver which depends on OF on any
architecture by explicitly selecting OF. Therefore depending on
COMPILE_TEST as an alternative is no longer needed.
It is actually better to always build such drivers with OF enabled,
so that the test builds are closer to how each driver will actually be
built on its intended target. Building them without OF may not test
much as the compiler will optimize out potentially large parts of the
code. In the worst case, this could even pop false positive warnings.
Dropping COMPILE_TEST here improves the quality of our testing and
avoids wasting time on non-existent issues.
Signed-off-by: Jean Delvare <jdelvare@suse.de>
Cc: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Cc: Ondrej Ille <ondrej.ille@gmail.com>
Acked-by: Pavel Pisa <pisa@cmp.felk.cvut.cz>
Link: https://lore.kernel.org/all/20221124141604.4265225f@endymion.delvare
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Biju Das <biju.das.jz@bp.renesas.com> says:
The CAN FD IP found on RZ/G2L SoC has some HW features different to
that of R-Car. For example, it has multiple resets, dedicated channel
tx and error interrupts, separate global rx and error interrupts
compared to shared irq for R-Car. it does not s ECC error flag
registers and clk post divider present on R-Car.
Similarly, R-Car V3U has 8 channels whereas other SoCs has only 2
channels. Currently all the HW differences are handled by comparing
with chip_id enum.
This patch series aims to replace chip_id with struct
rcar_canfd_hw_info to handle the HW feature differences and driver
data present on both IPs.
The changes are trivial and tested on RZ/G2L SMARC EVK.
This patch series depend upon [1].
[1] https://lore.kernel.org/all/20221025155657.1426948-1-biju.das.jz@bp.renesas.com
changes since v2: https://lore.kernel.org/all/20221026131732.1843105-1-biju.das.jz@bp.renesas.com
* Replaced data type of max_channels from unsigned int->u8 to save memory.
* Replaced data type of postdiv from unsigned int->u8 to save memory.
changes since v1: https://lore.kernel.org/all/20221022104357.1276740-1-biju.das.jz@bp.renesas.com
* Updated commit description for R-Car V3U SoC detection using
driver data.
* Replaced data type of max_channels from u32->unsigned int.
* Replaced multi_global_irqs->shared_global_irqs to make it
positive checks.
* Replaced clk_postdiv->postdiv driver data variable.
* Simplified the calcualtion for fcan_freq.
* Replaced info->has_gerfl to gpriv->info->has_gerfl and wrapped
the ECC error flag checks inside a single if statement.
* Added Rb tag from Geert patch#1,#2,#3 and #5
Link: https://lore.kernel.org/all/20221027082158.95895-1-biju.das.jz@bp.renesas.com
[mkl: only take patches 1...5 to avoid merge conflict]
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Xin Long says:
====================
net: eliminate the duplicate code in the ct nat functions of ovs and tc
The changes in the patchset:
"net: add helper support in tc act_ct for ovs offloading"
had moved some common ct code used by both OVS and TC into netfilter.
There are still some big functions pretty similar defined and used in
each of OVS and TC. It is not good to maintain such big function in 2
places. This patchset is to extract the functions for NAT processing
from OVS and TC to netfilter.
To make this change clear and safe, this patchset gets the common code
out of OVS and TC step by step: The patch 1-4 make some minor changes
in OVS and TC to make the NAT code of them completely the same, then
the patch 5 moves the common code to the netfilter and exports one
function called by each of OVS and TC.
v1->v2:
- Create nf_nat_ovs.c to include the nat functions, as Pablo suggested.
v2->v3:
- fix a typo in subject of patch 2/5, as Marcelo noticed.
- fix in openvswitch to keep OVS ct nat and TC ct nat consistent in
patch 3/5 instead of in tc, as Marcelo noticed.
- use BIT(var) macro instead of (1 << var) in patch 5/5, as Marcelo
suggested.
- use ifdef in netfilter/Makefile to build nf_nat_ovs only when OVS
or TC ct action is enabled in patch 5/5, as Marcelo suggested.
v3->v4:
- add NF_NAT_OVS in netfilter/Kconfig and add select NF_NAT_OVS in
OVS and TC Kconfig instead of using ifdef in netfilter/Makefile,
as Pablo suggested.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
There are two nat functions are nearly the same in both OVS and
TC code, (ovs_)ct_nat_execute() and ovs_ct_nat/tcf_ct_act_nat().
This patch creates nf_nat_ovs.c under netfilter and moves them
there then exports nf_ct_nat() so that it can be shared by both
OVS and TC, and keeps the nat (type) check and nat flag update
in OVS and TC's own place, as these parts are different between
OVS and TC.
Note that in OVS nat function it was using skb->protocol to get
the proto as it already skips vlans in key_extract(), while it
doesn't in TC, and TC has to call skb_protocol() to get proto.
So in nf_ct_nat_execute(), we keep using skb_protocol() which
works for both OVS and TC contrack.
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Aaron Conole <aconole@redhat.com>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
In ovs_ct_nat_execute(), the packet flow key nat flags are updated
when it processes ICMP(v6) error packets translation successfully.
In ct_nat_execute() when processing ICMP(v6) error packets translation
successfully, it should have done the same in ct_nat_execute() to set
post_ct_s/dnat flag, which will be used to update flow key nat flags
in OVS module later.
Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When it fails to allocate nat ext, the packet should be dropped, like
the memory allocation failures in other places in ovs_ct_nat().
This patch changes to return NF_DROP when fails to add nat ext before
doing NAT in ovs_ct_nat(), also it would keep consistent with tc
action ct' processing in tcf_ct_act_nat().
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Either OVS_CT_SRC_NAT or OVS_CT_DST_NAT is set, OVS_CT_NAT must be
set in info->nat. Thus, if OVS_CT_NAT is not set in info->nat, it
will definitely not do NAT but returns NF_ACCEPT in ovs_ct_nat().
This patch changes nothing funcational but only makes this return
earlier in ovs_ct_nat() to keep consistent with TC's processing
in tcf_ct_act_nat().
Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The calls to ovs_ct_nat_execute() are as below:
ovs_ct_execute()
ovs_ct_lookup()
__ovs_ct_lookup()
ovs_ct_nat()
ovs_ct_nat_execute()
ovs_ct_commit()
__ovs_ct_lookup()
ovs_ct_nat()
ovs_ct_nat_execute()
and since skb_pull_rcsum() and skb_push_rcsum() are already
called in ovs_ct_execute(), there's no need to do it again
in ovs_ct_nat_execute().
Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.
Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.
Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The implementation of strscpy() is more robust and safer.
That's now the recommended way to copy NUL terminated strings.
Signed-off-by: Xu Panda <xu.panda@zte.com.cn>
Signed-off-by: Yang Yang <yang.yang29@zte.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adds a boundary check to prevent negative basetime input from user
while configuring taprio.
Signed-off-by: Michael Sit Wei Hong <michael.wei.hong.sit@intel.com>
Signed-off-by: Lai Peter Jun Ann <jun.ann.lai@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Andrew Melnychenko says:
====================
TUN/VirtioNet USO features support.
Added new offloads for TUN devices TUN_F_USO4 and TUN_F_USO6.
Technically they enable NETIF_F_GSO_UDP_L4
(and only if USO4 & USO6 are set simultaneously).
It allows the transmission of large UDP packets.
UDP Segmentation Offload (USO/GSO_UDP_L4) - ability to split UDP packets
into several segments. It's similar to UFO, except it doesn't use IP
fragmentation. The drivers may push big packets and the NIC will split
them(or assemble them in case of receive), but in the case of VirtioNet
we just pass big UDP to the host. So we are freeing the driver from doing
the unnecessary job of splitting. The same thing for several guests
on one host, we can pass big packets between guests.
Different features USO4 and USO6 are required for qemu where Windows
guests can enable disable USO receives for IPv4 and IPv6 separately.
On the other side, Linux can't really differentiate USO4 and USO6, for now.
For now, to enable USO for TUN it requires enabling USO4 and USO6 together.
In the future, there would be a mechanism to control UDP_L4 GSO separately.
New types for virtio-net already in virtio-net specification:
https://github.com/oasis-tcs/virtio-spec/issues/120
Test it WIP Qemu https://github.com/daynix/qemu/tree/USOv3
Changes since v4 & RFC:
* Fixed typo and refactored.
* Tun USO offload refactored.
* Add support for guest-to-guest segmentation offload (thx Jason).
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Now, it possible to enable GSO_UDP_L4("tx-udp-segmentation") for VirtioNet.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Now, it's possible to convert USO vnet packets from/to skb.
Added support for GSO_UDP_L4 offload.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Added new GSO type for USO: VIRTIO_NET_HDR_GSO_UDP_L4.
Feature VIRTIO_NET_F_HOST_USO allows to enable NETIF_F_GSO_UDP_L4.
Separated VIRTIO_NET_F_GUEST_USO4 & VIRTIO_NET_F_GUEST_USO6 features
required for Windows guests.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Added support for USO4 and USO6.
For now, to "enable" USO, it's required to set both USO4 and USO6 simultaneously.
USO enables NETIF_F_GSO_UDP_L4.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Added 2 additional offlloads for USO(IPv4 & IPv6).
Separate offloads are required for Windows VM guests,
g.e. Windows may set USO rx only for IPv4.
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow UDP_L4 for robust packets.
Signed-off-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Andrew Melnychenko <andrew@daynix.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow the kthreads for stats to be configured for
specific cpulist (isolation) and niceness (scheduling
priority).
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Cc: yunhong-cgl jiang <xintian1976@gmail.com>
Cc: "dust.li" <dust.li@linux.alibaba.com>
Reviewed-by: Jiri Wiesner <jwiesner@suse.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Estimating all entries in single list in timer context
by single CPU causes large latency with multiple IPVS rules
as reported in [1], [2], [3].
Spread the estimator structures in multiple chains and
use kthread(s) for the estimation. The chains are processed
in multiple (50) timer ticks to ensure the 2-second interval
between estimations with some accuracy. Every chain is
processed under RCU lock.
Every kthread works over its own data structure and all
such contexts are attached to array. The contexts can be
preserved while the kthread tasks are stopped or restarted.
When estimators are removed, unused kthread contexts are
released and the slots in array are left empty.
First kthread determines parameters to use, eg. maximum
number of estimators to process per kthread based on
chain's length (chain_max), allowing sub-100us cond_resched
rate and estimation taking up to 1/8 of the CPU capacity
to avoid any problems if chain_max is not correctly
calculated.
chain_max is calculated taking into account factors
such as CPU speed and memory/cache speed where the
cache_factor (4) is selected from real tests with
current generation of CPU/NUMA configurations to
correct the difference in CPU usage between
cached (during calc phase) and non-cached (working) state
of the estimated per-cpu data.
First kthread also plays the role of distributor of
added estimators to all kthreads, keeping low the
time to add estimators. The optimization is based on
the fact that newly added estimator should be estimated
after 2 seconds, so we have the time to offload the
adding to chain from controlling process to kthread 0.
The allocated kthread context may grow from 1 to 50
allocated structures for timer ticks which saves memory for
setups with small number of estimators.
We also add delayed work est_reload_work that will
make sure the kthread tasks are properly started/stopped.
ip_vs_start_estimator() is changed to report errors
which allows to safely store the estimators in
allocated structures.
Many thanks to Jiri Wiesner for his valuable comments
and for spending a lot of time reviewing and testing
the changes on different platforms with 48-256 CPUs and
1-8 NUMA nodes under different cpufreq governors.
[1] Report from Yunhong Jiang:
https://lore.kernel.org/netdev/D25792C1-1B89-45DE-9F10-EC350DC04ADC@gmail.com/
[2]
https://marc.info/?l=linux-virtual-server&m=159679809118027&w=2
[3] Report from Dust:
https://archive.linuxvirtualserver.org/html/lvs-devel/2020-12/msg00000.html
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Cc: yunhong-cgl jiang <xintian1976@gmail.com>
Cc: "dust.li" <dust.li@linux.alibaba.com>
Reviewed-by: Jiri Wiesner <jwiesner@suse.de>
Tested-by: Jiri Wiesner <jwiesner@suse.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
In preparation to using RCU locking for the list
with estimators, make sure the struct ip_vs_stats
are released after RCU grace period by using RCU
callbacks. This affects ipvs->tot_stats where we
can not use RCU callbacks for ipvs, so we use
allocated struct ip_vs_stats_rcu. For services
and dests we force RCU callbacks for all cases.
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Cc: yunhong-cgl jiang <xintian1976@gmail.com>
Cc: "dust.li" <dust.li@linux.alibaba.com>
Reviewed-by: Jiri Wiesner <jwiesner@suse.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Eduard Zingerman says:
====================
This patch-set consists of a series of bug fixes for register ID
tracking in verifier.c:states_equal()/regsafe() functions:
- for registers of type PTR_TO_MAP_{KEY,VALUE}, PTR_TO_PACKET[_META]
the regsafe() should call check_ids() even if registers are
byte-to-byte equal;
- states_equal() must maintain idmap that covers all function frames
in the state because functions like mark_ptr_or_null_regs() operate
on all registers in the state;
- regsafe() must compare spin lock ids for PTR_TO_MAP_VALUE registers.
The last point covers issue reported by Kumar Kartikeya Dwivedi in [1],
I borrowed the test commit from there.
Note, that there is also an issue with register id tracking for
scalars described here [2], it would be addressed separately.
[1] https://lore.kernel.org/bpf/20221111202719.982118-1-memxor@gmail.com/
[2] https://lore.kernel.org/bpf/20221128163442.280187-2-eddyz87@gmail.com/
Eduard Zingerman (6):
bpf: regsafe() must not skip check_ids()
selftests/bpf: test cases for regsafe() bug skipping check_id()
bpf: states_equal() must build idmap for all function frames
selftests/bpf: verify states_equal() maintains idmap across all frames
bpf: use check_ids() for active_lock comparison
selftests/bpf: test case for relaxed prunning of active_lock.id
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Check that verifier.c:states_equal() uses check_ids() to match
consistent active_lock/map_value configurations. This allows to prune
states with active spin locks even if numerical values of
active_lock ids do not match across compared states.
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-8-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Test that when reg->id is not same for the same register of type
PTR_TO_MAP_VALUE between current and old explored state, we currently
return false from regsafe and continue exploring.
Without the fix in prior commit, the test case fails.
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-7-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
An update for verifier.c:states_equal()/regsafe() to use check_ids()
for active spin lock comparisons. This fixes the issue reported by
Kumar Kartikeya Dwivedi in [1] using technique suggested by Edward Cree.
W/o this commit the verifier might be tricked to accept the following
program working with a map containing spin locks:
0: r9 = map_lookup_elem(...) ; Returns PTR_TO_MAP_VALUE_OR_NULL id=1.
1: r8 = map_lookup_elem(...) ; Returns PTR_TO_MAP_VALUE_OR_NULL id=2.
2: if r9 == 0 goto exit ; r9 -> PTR_TO_MAP_VALUE.
3: if r8 == 0 goto exit ; r8 -> PTR_TO_MAP_VALUE.
4: r7 = ktime_get_ns() ; Unbound SCALAR_VALUE.
5: r6 = ktime_get_ns() ; Unbound SCALAR_VALUE.
6: bpf_spin_lock(r8) ; active_lock.id == 2.
7: if r6 > r7 goto +1 ; No new information about the state
; is derived from this check, thus
; produced verifier states differ only
; in 'insn_idx'.
8: r9 = r8 ; Optionally make r9.id == r8.id.
--- checkpoint --- ; Assume is_state_visisted() creates a
; checkpoint here.
9: bpf_spin_unlock(r9) ; (a,b) active_lock.id == 2.
; (a) r9.id == 2, (b) r9.id == 1.
10: exit(0)
Consider two verification paths:
(a) 0-10
(b) 0-7,9-10
The path (a) is verified first. If checkpoint is created at (8)
the (b) would assume that (8) is safe because regsafe() does not
compare register ids for registers of type PTR_TO_MAP_VALUE.
[1] https://lore.kernel.org/bpf/20221111202719.982118-1-memxor@gmail.com/
Reported-by: Kumar Kartikeya Dwivedi <memxor@gmail.com>
Suggested-by: Edward Cree <ecree.xilinx@gmail.com>
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-6-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
A test case that would erroneously pass verification if
verifier.c:states_equal() maintains separate register ID mappings for
call frames.
Signed-off-by: Eduard Zingerman <eddyz87@gmail.com>
Link: https://lore.kernel.org/r/20221209135733.28851-5-eddyz87@gmail.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
