Currently when GID is deleted, it zero out all the fields of the RoCE
address in the SET_ROCE_ADDRESS command for a specified index.
roce_version = 0 means RoCEv1 in the SET_ROCE_ADDRESS command.
This assumes that device has RoCEv1 always enabled which is not always
correct. For example Subfunction does not support RoCEv1.
Due to this assumption a previously added RoCEv2 GID is always deleted as
RoCEv1 GID. This results in a below syndrome:
mlx5_core.sf mlx5_core.sf.4: mlx5_cmd_check:777:(pid 4256): SET_ROCE_ADDRESS(0x761) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0x12822d)
Hence set the right RoCE version during GID deletion provided by the core.
Link: https://lore.kernel.org/r/d3f54129c90ca329caf438dbe31875d8ad08d91a.1618753425.git.leonro@nvidia.com
Signed-off-by: Parav Pandit <parav@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Saeed Mahameed says:
====================
This pr contains changes from mlx5-next branch,
already reviewed on netdev and rdma mailing lists, links below.
1) From Leon, Dynamically assign MSI-X vectors count
Already Acked by Bjorn Helgaas.
https://patchwork.kernel.org/project/netdevbpf/cover/20210314124256.70253-1-leon@kernel.org/
2) Cleanup series:
https://patchwork.kernel.org/project/netdevbpf/cover/20210311070915.321814-1-saeed@kernel.org/
From Mark, E-Switch cleanups and refactoring, and the addition
of single FDB mode needed HW bits.
From Mikhael, Remove unused struct field
From Saeed, Cleanup W=1 prototype warning
From Zheng, Esw related cleanup
From Tariq, User order-0 page allocation for EQs
====================
* mlx5-next:
net/mlx5: Implement sriov_get_vf_total_msix/count() callbacks
net/mlx5: Dynamically assign MSI-X vectors count
net/mlx5: Add dynamic MSI-X capabilities bits
PCI/IOV: Add sysfs MSI-X vector assignment interface
net/mlx5: Use order-0 allocations for EQs
net/mlx5: Add IFC bits needed for single FDB mode
net/mlx5: E-Switch, Refactor send to vport to be more generic
RDMA/mlx5: Use representor E-Switch when getting netdev and metadata
net/mlx5: E-Switch, Add eswitch pointer to each representor
net/mlx5: E-Switch, Add match on vhca id to default send rules
net/mlx5: Remove unused mlx5_core_health member recover_work
net/mlx5: simplify the return expression of mlx5_esw_offloads_pair()
net/mlx5: Cleanup prototype warning
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
The number of MSI-X vectors is a PCI property visible through lspci. The
field is read-only and configured by the device. The mlx5 devices work in
a static or dynamic assignment mode.
Static assignment means that all newly created VFs have a preset number of
MSI-X vectors determined by device configuration parameters. This can
result in some VFs having too many or too few MSI-X vectors. Till now this
has been the only means of fine-tuning the MSI-X vector count and it was
acceptable for small numbers of VFs.
With dynamic assignment the inefficiency of having a fixed number of MSI-X
vectors can be avoided with each VF having exactly the required
vectors. Userspace will provide this information while provisioning the VF
for use, based on the intended use. For instance if being used with a VM,
the MSI-X vector count might be matched to the CPU count of the VM.
For compatibility mlx5 continues to start up with MSI-X vector assignment,
but the kernel can now access a larger dynamic vector pool and assign more
vectors to created VFs.
Link: https://lore.kernel.org/linux-pci/20210314124256.70253-4-leon@kernel.org
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Current code uses many different types when dealing with a port of a RDMA
device: u8, unsigned int and u32. Switch to u32 to clean up the logic.
This allows us to make (at least) the core view consistent and use the
same type. Unfortunately not all places can be converted. Many uverbs
functions expect port to be u8 so keep those places in order not to break
UAPIs. HW/Spec defined values must also not be changed.
With the switch to u32 we now can support devices with more than 255
ports. U32_MAX is reserved to make control logic a bit easier to deal
with. As a device with U32_MAX ports probably isn't going to happen any
time soon this seems like a non issue.
When a device with more than 255 ports is created uverbs will report the
RDMA device as having 255 ports as this is the max currently supported.
The verbs interface is not changed yet because the IBTA spec limits the
port size in too many places to be u8 and all applications that relies in
verbs won't be able to cope with this change. At this stage, we are
extending the interfaces that are using vendor channel solely
Once the limitation is lifted mlx5 in switchdev mode will be able to have
thousands of SFs created by the device. As the only instance of an RDMA
device that reports more than 255 ports will be a representor device and
it exposes itself as a RAW Ethernet only device CM/MAD/IPoIB and other
ULPs aren't effected by this change and their sysfs/interfaces that are
exposes to userspace can remain unchanged.
While here cleanup some alignment issues and remove unneeded sanity
checks (mainly in rdmavt),
Link: https://lore.kernel.org/r/20210301070420.439400-1-leon@kernel.org
Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
Currently we are allocating high-order page for EQs. In case of
fragmented system, VF hot remove/add in VMs for example, there isn't
enough contiguous memory for EQs allocation, which results in crashing
of the VM.
Therefore, use order-0 fragments for the EQ allocations instead.
Performance tests:
ConnectX-5 100Gbps, CPU: Intel(R) Xeon(R) CPU E5-2697 v3 @ 2.60GHz
Performance tests show no sensible degradation.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Shay Drory <shayd@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Now that each representor stores a pointer to the managing E-Switch
use that information when creating the send-to-vport rules.
Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Store the managing E-Switch of each representor. This will be used
when a representor is created on eswitch manager 0 but the vport
belongs to eswitch manager 1.
Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Match on the vhca id of the E-Switch owner when creating the send-to-vport
representor rules.
Signed-off-by: Mark Bloch <mbloch@nvidia.com>
Reviewed-by: Saeed Mahameed <saeedm@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Cleanup W=1 warning:
drivers/net/ethernet/mellanox/mlx5/core/lag_mp.c:49:
warning: expecting prototype for Set lag port affinity().
Prototype was for mlx5_lag_set_port_affinity() instead
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
Pull xen fixes from Juergen Gross:
"Two security issues (XSA-367 and XSA-369)"
* tag 'for-linus-5.12b-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: fix p2m size in dom0 for disabled memory hotplug case
xen-netback: respect gnttab_map_refs()'s return value
Xen/gnttab: handle p2m update errors on a per-slot basis
Pull virtio updates from Michael Tsirkin:
- new vdpa features to allow creation and deletion of new devices
- virtio-blk support per-device queue depth
- fixes, cleanups all over the place
* tag 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mst/vhost: (31 commits)
virtio-input: add multi-touch support
virtio_mmio: fix one typo
vdpa/mlx5: fix param validation in mlx5_vdpa_get_config()
virtio_net: Fix fall-through warnings for Clang
virtio_input: Prevent EV_MSC/MSC_TIMESTAMP loop storm for MT.
virtio-blk: support per-device queue depth
virtio_vdpa: don't warn when fail to disable vq
virtio-pci: introduce modern device module
virito-pci-modern: rename map_capability() to vp_modern_map_capability()
virtio-pci-modern: introduce helper to get notification offset
virtio-pci-modern: introduce helper for getting queue nums
virtio-pci-modern: introduce helper for setting/geting queue size
virtio-pci-modern: introduce helper to set/get queue_enable
virtio-pci-modern: introduce vp_modern_queue_address()
virtio-pci-modern: introduce vp_modern_set_queue_vector()
virtio-pci-modern: introduce vp_modern_generation()
virtio-pci-modern: introduce helpers for setting and getting features
virtio-pci-modern: introduce helpers for setting and getting status
virtio-pci-modern: introduce helper to set config vector
virtio-pci-modern: introduce vp_modern_remove()
...
Pull networking fixes from Jakub Kicinski:
"Rather small batch this time.
Current release - regressions:
- bcm63xx_enet: fix sporadic kernel panic due to queue length
mis-accounting
Current release - new code bugs:
- bcm4908_enet: fix RX path possible mem leak
- bcm4908_enet: fix NAPI poll returned value
- stmmac: fix missing spin_lock_init in visconti_eth_dwmac_probe()
- sched: cls_flower: validate ct_state for invalid and reply flags
Previous releases - regressions:
- net: introduce CAN specific pointer in the struct net_device to
prevent mis-interpreting memory
- phy: micrel: set soft_reset callback to genphy_soft_reset for
KSZ8081
- psample: fix netlink skb length with tunnel info
Previous releases - always broken:
- icmp: pass zeroed opts from icmp{,v6}_ndo_send before sending
- wireguard: device: do not generate ICMP for non-IP packets
- mptcp: provide subflow aware release function to avoid a mem leak
- hsr: add support for EntryForgetTime
- r8169: fix jumbo packet handling on RTL8168e
- octeontx2-af: fix an off by one in rvu_dbg_qsize_write()
- i40e: fix flow for IPv6 next header (extension header)
- phy: icplus: call phy_restore_page() when phy_select_page() fails
- dpaa_eth: fix the access method for the dpaa_napi_portal"
* tag 'net-5.12-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (55 commits)
r8169: fix jumbo packet handling on RTL8168e
net: phy: micrel: set soft_reset callback to genphy_soft_reset for KSZ8081
net: psample: Fix netlink skb length with tunnel info
net: broadcom: bcm4908_enet: fix NAPI poll returned value
net: broadcom: bcm4908_enet: fix RX path possible mem leak
net: hsr: add support for EntryForgetTime
net: dsa: sja1105: Remove unneeded cast in sja1105_crc32()
ibmvnic: fix a race between open and reset
net: stmmac: Fix missing spin_lock_init in visconti_eth_dwmac_probe()
net: introduce CAN specific pointer in the struct net_device
net: usb: qmi_wwan: support ZTE P685M modem
wireguard: kconfig: use arm chacha even with no neon
wireguard: queueing: get rid of per-peer ring buffers
wireguard: device: do not generate ICMP for non-IP packets
wireguard: peer: put frequently used members above cache lines
wireguard: selftests: test multiple parallel streams
wireguard: socket: remove bogus __be32 annotation
wireguard: avoid double unlikely() notation when using IS_ERR()
net: qrtr: Fix memory leak in qrtr_tun_open
vxlan: move debug check after netdev unregister
...
Pull Kbuild updates from Masahiro Yamada:
- Fix false-positive build warnings for ARCH=ia64 builds
- Optimize dictionary size for module compression with xz
- Check the compiler and linker versions in Kconfig
- Fix misuse of extra-y
- Support DWARF v5 debug info
- Clamp SUBLEVEL to 255 because stable releases 4.4.x and 4.9.x
exceeded the limit
- Add generic syscall{tbl,hdr}.sh for cleanups across arches
- Minor cleanups of genksyms
- Minor cleanups of Kconfig
* tag 'kbuild-v5.12' of git://git.kernel.org/pub/scm/linux/kernel/git/masahiroy/linux-kbuild: (38 commits)
initramfs: Remove redundant dependency of RD_ZSTD on BLK_DEV_INITRD
kbuild: remove deprecated 'always' and 'hostprogs-y/m'
kbuild: parse C= and M= before changing the working directory
kbuild: reuse this-makefile to define abs_srctree
kconfig: unify rule of config, menuconfig, nconfig, gconfig, xconfig
kconfig: omit --oldaskconfig option for 'make config'
kconfig: fix 'invalid option' for help option
kconfig: remove dead code in conf_askvalue()
kconfig: clean up nested if-conditionals in check_conf()
kconfig: Remove duplicate call to sym_get_string_value()
Makefile: Remove # characters from compiler string
Makefile: reuse CC_VERSION_TEXT
kbuild: check the minimum linker version in Kconfig
kbuild: remove ld-version macro
scripts: add generic syscallhdr.sh
scripts: add generic syscalltbl.sh
arch: syscalls: remove $(srctree)/ prefix from syscall tables
arch: syscalls: add missing FORCE and fix 'targets' to make if_changed work
gen_compile_commands: prune some directories
kbuild: simplify access to the kernel's version
...
Pull PCI updates from Bjorn Helgaas:
"Enumeration:
- Remove unnecessary locking around _OSC (Bjorn Helgaas)
- Clarify message about _OSC failure (Bjorn Helgaas)
- Remove notification of PCIe bandwidth changes (Bjorn Helgaas)
- Tidy checking of syscall user config accessors (Heiner Kallweit)
Resource management:
- Decline to resize resources if boot config must be preserved (Ard
Biesheuvel)
- Fix pci_register_io_range() memory leak (Geert Uytterhoeven)
Error handling (Keith Busch):
- Clear error status from the correct device
- Retain error recovery status so drivers can use it after reset
- Log the type of Port (Root or Switch Downstream) that we reset
- Always request a reset for Downstream Ports in frozen state
Endpoint framework and NTB (Kishon Vijay Abraham I):
- Make *_get_first_free_bar() take into account 64 bit BAR
- Add helper API to get the 'next' unreserved BAR
- Make *_free_bar() return error codes on failure
- Remove unused pci_epf_match_device()
- Add support to associate secondary EPC with EPF
- Add support in configfs to associate two EPCs with EPF
- Add pci_epc_ops to map MSI IRQ
- Add pci_epf_ops to expose function-specific attrs
- Allow user to create sub-directory of 'EPF Device' directory
- Implement ->msi_map_irq() ops for cadence
- Configure LM_EP_FUNC_CFG based on epc->function_num_map for cadence
- Add EP function driver to provide NTB functionality
- Add support for EPF PCI Non-Transparent Bridge
- Add specification for PCI NTB function device
- Add PCI endpoint NTB function user guide
- Add configfs binding documentation for pci-ntb endpoint function
Broadcom STB PCIe controller driver:
- Add support for BCM4908 and external PERST# signal controller
(Rafał Miłecki)
Cadence PCIe controller driver:
- Retrain Link to work around Gen2 training defect (Nadeem Athani)
- Fix merge botch in cdns_pcie_host_map_dma_ranges() (Krzysztof
Wilczyński)
Freescale Layerscape PCIe controller driver:
- Add LX2160A rev2 EP mode support (Hou Zhiqiang)
- Convert to builtin_platform_driver() (Michael Walle)
MediaTek PCIe controller driver:
- Fix OF node reference leak (Krzysztof Wilczyński)
Microchip PolarFlare PCIe controller driver:
- Add Microchip PolarFire PCIe controller driver (Daire McNamara)
Qualcomm PCIe controller driver:
- Use PHY_REFCLK_USE_PAD only for ipq8064 (Ansuel Smith)
- Add support for ddrss_sf_tbu clock for sm8250 (Dmitry Baryshkov)
Renesas R-Car PCIe controller driver:
- Drop PCIE_RCAR config option (Lad Prabhakar)
- Always allocate MSI addresses in 32bit space (Marek Vasut)
Rockchip PCIe controller driver:
- Add FriendlyARM NanoPi M4B DT binding (Chen-Yu Tsai)
- Make 'ep-gpios' DT property optional (Chen-Yu Tsai)
Synopsys DesignWare PCIe controller driver:
- Work around ECRC configuration hardware defect (Vidya Sagar)
- Drop support for config space in DT 'ranges' (Rob Herring)
- Change size to u64 for EP outbound iATU (Shradha Todi)
- Add upper limit address for outbound iATU (Shradha Todi)
- Make dw_pcie ops optional (Jisheng Zhang)
- Remove unnecessary dw_pcie_ops from al driver (Jisheng Zhang)
Xilinx Versal CPM PCIe controller driver:
- Fix OF node reference leak (Pan Bian)
Miscellaneous:
- Remove tango host controller driver (Arnd Bergmann)
- Remove IRQ handler & data together (altera-msi, brcmstb, dwc)
(Martin Kaiser)
- Fix xgene-msi race in installing chained IRQ handler (Martin
Kaiser)
- Apply CONFIG_PCI_DEBUG to entire drivers/pci hierarchy (Junhao He)
- Fix pci-bridge-emul array overruns (Russell King)
- Remove obsolete uses of WARN_ON(in_interrupt()) (Sebastian Andrzej
Siewior)"
* tag 'pci-v5.12-changes' of git://git.kernel.org/pub/scm/linux/kernel/git/helgaas/pci: (69 commits)
PCI: qcom: Use PHY_REFCLK_USE_PAD only for ipq8064
PCI: qcom: Add support for ddrss_sf_tbu clock
dt-bindings: PCI: qcom: Document ddrss_sf_tbu clock for sm8250
PCI: al: Remove useless dw_pcie_ops
PCI: dwc: Don't assume the ops in dw_pcie always exist
PCI: dwc: Add upper limit address for outbound iATU
PCI: dwc: Change size to u64 for EP outbound iATU
PCI: dwc: Drop support for config space in 'ranges'
PCI: layerscape: Convert to builtin_platform_driver()
PCI: layerscape: Add LX2160A rev2 EP mode support
dt-bindings: PCI: layerscape: Add LX2160A rev2 compatible strings
PCI: dwc: Work around ECRC configuration issue
PCI/portdrv: Report reset for frozen channel
PCI/AER: Specify the type of Port that was reset
PCI/ERR: Retain status from error notification
PCI/AER: Clear AER status from Root Port when resetting Downstream Port
PCI/ERR: Clear status of the reporting device
dt-bindings: arm: rockchip: Add FriendlyARM NanoPi M4B
PCI: rockchip: Make 'ep-gpios' DT property optional
Documentation: PCI: Add PCI endpoint NTB function user guide
...
Following a similar reinstate for the KSZ9031.
Older kernels would use the genphy_soft_reset if the PHY did not implement
a .soft_reset.
Bluntly removing that default may expose a lot of situations where various
PHYs/board implementations won't recover on various changes.
Like with this implementation during a 4.9.x to 5.4.x LTS transition.
I think it's a good thing to remove unwanted soft resets but wonder if it
did open a can of worms?
Atleast this fixes one iMX6 FEC/RMII/8081 combo.
Fixes: 6e2d85ec05 ("net: phy: Stop with excessive soft reset")
Signed-off-by: Christian Melki <christian.melki@t2data.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20210224205536.9349-1-christian.melki@t2data.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
__ibmvnic_reset() currently reads the adapter->state before getting the
rtnl and saves that state as the "target state" for the reset. If this
read occurs when adapter is in PROBED state, the target state would be
PROBED.
Just after the target state is saved, and before the actual reset process
is started (i.e before rtnl is acquired) if we get an ibmvnic_open() call
we would move the adapter to OPEN state.
But when the reset is processed (after ibmvnic_open()) drops the rtnl),
it will leave the adapter in PROBED state even though we already moved
it to OPEN.
To fix this, use the RTNL to improve serialization when reading/updating
the adapter state. i.e determine the target state of a reset only after
getting the RTNL. And if a reset is in progress during an open, simply
set the target state of the adapter and let the reset code finish the
open (like we currently do if failover is pending).
One twist to this serialization is if the adapter state changes when we
drop the RTNL to update the link state. Account for this by checking if
there was an intervening open and update the target state for the reset
accordingly (see new comments in the code). Note that only the reset
functions and ibmvnic_open() can set the adapter to OPEN state and this
must happen under rtnl.
Fixes: 7d7195a026 ("ibmvnic: Do not process device remove during device reset")
Signed-off-by: Sukadev Bhattiprolu <sukadev@linux.ibm.com>
Reviewed-by: Dany Madden <drt@linux.ibm.com>
Link: https://lore.kernel.org/r/20210224050229.1155468-1-sukadev@linux.ibm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Since 20dd3850bc ("can: Speed up CAN frame receiption by using
ml_priv") the CAN framework uses per device specific data in the AF_CAN
protocol. For this purpose the struct net_device->ml_priv is used. Later
the ml_priv usage in CAN was extended for other users, one of them being
CAN_J1939.
Later in the kernel ml_priv was converted to an union, used by other
drivers. E.g. the tun driver started storing it's stats pointer.
Since tun devices can claim to be a CAN device, CAN specific protocols
will wrongly interpret this pointer, which will cause system crashes.
Mostly this issue is visible in the CAN_J1939 stack.
To fix this issue, we request a dedicated CAN pointer within the
net_device struct.
Reported-by: syzbot+5138c4dd15a0401bec7b@syzkaller.appspotmail.com
Fixes: 20dd3850bc ("can: Speed up CAN frame receiption by using ml_priv")
Fixes: ffd956eef6 ("can: introduce CAN midlayer private and allocate it automatically")
Fixes: 9d71dd0c70 ("can: add support of SAE J1939 protocol")
Fixes: 497a5757ce ("tun: switch to net core provided statistics counters")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Link: https://lore.kernel.org/r/20210223070127.4538-1-o.rempel@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2021-02-22
Dave corrects reporting of max TCs to use the value from hardware
capabilities and setting of DCBx capability bits when changing between
SW and FW LLDP.
Brett fixes trusted VF multicast promiscuous not receiving expected
packets and corrects VF max packet size when a port VLAN is configured.
Henry updates available RSS queues following a change in channel count
with a user defined LUT.
* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
ice: update the number of available RSS queues
ice: Fix state bits on LLDP mode switch
ice: Account for port VLAN in VF max packet size calculation
ice: Set trusted VF as default VSI when setting allmulti on
ice: report correct max number of TCs
====================
Link: https://lore.kernel.org/r/20210222235814.834282-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The condition here was incorrect: a non-neon fallback implementation is
available on arm32 when NEON is not supported.
Reported-by: Ilya Lipnitskiy <ilya.lipnitskiy@gmail.com>
Fixes: e7096c131e ("net: WireGuard secure network tunnel")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Having two ring buffers per-peer means that every peer results in two
massive ring allocations. On an 8-core x86_64 machine, this commit
reduces the per-peer allocation from 18,688 bytes to 1,856 bytes, which
is an 90% reduction. Ninety percent! With some single-machine
deployments approaching 500,000 peers, we're talking about a reduction
from 7 gigs of memory down to 700 megs of memory.
In order to get rid of these per-peer allocations, this commit switches
to using a list-based queueing approach. Currently GSO fragments are
chained together using the skb->next pointer (the skb_list_* singly
linked list approach), so we form the per-peer queue around the unused
skb->prev pointer (which sort of makes sense because the links are
pointing backwards). Use of skb_queue_* is not possible here, because
that is based on doubly linked lists and spinlocks. Multiple cores can
write into the queue at any given time, because its writes occur in the
start_xmit path or in the udp_recv path. But reads happen in a single
workqueue item per-peer, amounting to a multi-producer, single-consumer
paradigm.
The MPSC queue is implemented locklessly and never blocks. However, it
is not linearizable (though it is serializable), with a very tight and
unlikely race on writes, which, when hit (some tiny fraction of the
0.15% of partial adds on a fully loaded 16-core x86_64 system), causes
the queue reader to terminate early. However, because every packet sent
queues up the same workqueue item after it is fully added, the worker
resumes again, and stopping early isn't actually a problem, since at
that point the packet wouldn't have yet been added to the encryption
queue. These properties allow us to avoid disabling interrupts or
spinning. The design is based on Dmitry Vyukov's algorithm [1].
Performance-wise, ordinarily list-based queues aren't preferable to
ringbuffers, because of cache misses when following pointers around.
However, we *already* have to follow the adjacent pointers when working
through fragments, so there shouldn't actually be any change there. A
potential downside is that dequeueing is a bit more complicated, but the
ptr_ring structure used prior had a spinlock when dequeueing, so all and
all the difference appears to be a wash.
Actually, from profiling, the biggest performance hit, by far, of this
commit winds up being atomic_add_unless(count, 1, max) and atomic_
dec(count), which account for the majority of CPU time, according to
perf. In that sense, the previous ring buffer was superior in that it
could check if it was full by head==tail, which the list-based approach
cannot do.
But all and all, this enables us to get massive memory savings, allowing
WireGuard to scale for real world deployments, without taking much of a
performance hit.
[1] http://www.1024cores.net/home/lock-free-algorithms/queues/intrusive-mpsc-node-based-queue
Reviewed-by: Dmitry Vyukov <dvyukov@google.com>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Fixes: e7096c131e ("net: WireGuard secure network tunnel")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
If skb->protocol doesn't match the actual skb->data header, it's
probably not a good idea to pass it off to icmp{,v6}_ndo_send, which is
expecting to reply to a valid IP packet. So this commit has that early
mismatch case jump to a later error label.
Fixes: e7096c131e ("net: WireGuard secure network tunnel")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The is_dead boolean is checked for every single packet, while the
internal_id member is used basically only for pr_debug messages. So it
makes sense to hoist up is_dead into some space formerly unused by a
struct hole, while demoting internal_api to below the lowest struct
cache line.
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The endpoint->src_if4 has nothing to do with fixed-endian numbers; remove
the bogus annotation.
This was introduced in
https://git.zx2c4.com/wireguard-monolithic-historical/commit?id=14e7d0a499a676ec55176c0de2f9fcbd34074a82
in the historical WireGuard repo because the old code used to
zero-initialize multiple members as follows:
endpoint->src4.s_addr = endpoint->src_if4 = fl.saddr = 0;
Because fl.saddr is fixed-endian and an assignment returns a value with the
type of its left operand, this meant that sparse detected an assignment
between values of different endianness.
Since then, this assignment was already split up into separate statements;
just the cast survived.
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The definition of IS_ERR() already applies the unlikely() notation
when checking the error status of the passed pointer. For this
reason there is no need to have the same notation outside of
IS_ERR() itself.
Clean up code by removing redundant notation.
Signed-off-by: Antonio Quartulli <a@unstable.cc>
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Some messages are before calling register_netdev(), so replace
netif_err() with dev_err().
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Return error code if autosuspend_en, eee_get, or eee_set don't exist.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
U1/U2 shoued be enabled for USB 3.0 or later. The USB 2.0 doesn't
support it.
Signed-off-by: Hayes Wang <hayeswang@realtek.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Add support for being able to set the learning attribute on port, and
make sure that the standalone ports start up with learning disabled.
We can remove the code in bcm_sf2 that configured the ports learning
attribute because we want the standalone ports to have learning disabled
by default and port 7 cannot be bridged, so its learning attribute will
not change past its initial configuration.
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Because bcm_sf2 implements its own dsa_switch_ops we need to export the
b53_br_flags_pre(), b53_br_flags() and b53_set_mrouter so we can wire-up
them up like they used to be with the former b53_br_egress_floods().
Fixes: a8b659e7ff ("net: dsa: act as passthrough for bridge port flags")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
In ndo_stop functions, netdev_completed_queue() is called during forced
tx reclaim, after netdev_reset_queue(). This may trigger kernel panic if
there is any tx skb left.
This patch moves netdev_reset_queue() to after tx reclaim, so BQL can
complete successfully then reset.
Signed-off-by: Sieng Piaw Liew <liew.s.piaw@gmail.com>
Fixes: 4c59b0f554 ("bcm63xx_enet: add BQL support")
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20210222013530.1356-1-liew.s.piaw@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The icmp{,v6}_send functions make all sorts of use of skb->cb, casting
it with IPCB or IP6CB, assuming the skb to have come directly from the
inet layer. But when the packet comes from the ndo layer, especially
when forwarded, there's no telling what might be in skb->cb at that
point. As a result, the icmp sending code risks reading bogus memory
contents, which can result in nasty stack overflows such as this one
reported by a user:
panic+0x108/0x2ea
__stack_chk_fail+0x14/0x20
__icmp_send+0x5bd/0x5c0
icmp_ndo_send+0x148/0x160
In icmp_send, skb->cb is cast with IPCB and an ip_options struct is read
from it. The optlen parameter there is of particular note, as it can
induce writes beyond bounds. There are quite a few ways that can happen
in __ip_options_echo. For example:
// sptr/skb are attacker-controlled skb bytes
sptr = skb_network_header(skb);
// dptr/dopt points to stack memory allocated by __icmp_send
dptr = dopt->__data;
// sopt is the corrupt skb->cb in question
if (sopt->rr) {
optlen = sptr[sopt->rr+1]; // corrupt skb->cb + skb->data
soffset = sptr[sopt->rr+2]; // corrupt skb->cb + skb->data
// this now writes potentially attacker-controlled data, over
// flowing the stack:
memcpy(dptr, sptr+sopt->rr, optlen);
}
In the icmpv6_send case, the story is similar, but not as dire, as only
IP6CB(skb)->iif and IP6CB(skb)->dsthao are used. The dsthao case is
worse than the iif case, but it is passed to ipv6_find_tlv, which does
a bit of bounds checking on the value.
This is easy to simulate by doing a `memset(skb->cb, 0x41,
sizeof(skb->cb));` before calling icmp{,v6}_ndo_send, and it's only by
good fortune and the rarity of icmp sending from that context that we've
avoided reports like this until now. For example, in KASAN:
BUG: KASAN: stack-out-of-bounds in __ip_options_echo+0xa0e/0x12b0
Write of size 38 at addr ffff888006f1f80e by task ping/89
CPU: 2 PID: 89 Comm: ping Not tainted 5.10.0-rc7-debug+ #5
Call Trace:
dump_stack+0x9a/0xcc
print_address_description.constprop.0+0x1a/0x160
__kasan_report.cold+0x20/0x38
kasan_report+0x32/0x40
check_memory_region+0x145/0x1a0
memcpy+0x39/0x60
__ip_options_echo+0xa0e/0x12b0
__icmp_send+0x744/0x1700
Actually, out of the 4 drivers that do this, only gtp zeroed the cb for
the v4 case, while the rest did not. So this commit actually removes the
gtp-specific zeroing, while putting the code where it belongs in the
shared infrastructure of icmp{,v6}_ndo_send.
This commit fixes the issue by passing an empty IPCB or IP6CB along to
the functions that actually do the work. For the icmp_send, this was
already trivial, thanks to __icmp_send providing the plumbing function.
For icmpv6_send, this required a tiny bit of refactoring to make it
behave like the v4 case, after which it was straight forward.
Fixes: a2b78e9b2c ("sunvnet: generate ICMP PTMUD messages for smaller port MTUs")
Reported-by: SinYu <liuxyon@gmail.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/netdev/CAF=yD-LOF116aHub6RMe8vB8ZpnrrnoTdqhobEx+bvoA8AsP0w@mail.gmail.com/T/
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Link: https://lore.kernel.org/r/20210223131858.72082-1-Jason@zx2c4.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2021-02-19
This series contains updates to i40e driver only.
Slawomir resolves an issue with the IPv6 extension headers being
processed incorrectly.
Keita Suzuki fixes a memory leak on probe failure.
Mateusz initializes AQ command structures to zero to comply with
spec, fixes FW flow control settings being overwritten and resolves an
issue with adding VLAN filters after enabling FW LLDP. He also adds
an additional check when adding TC filter as the current check doesn't
properly distinguish between IPv4 and IPv6.
Sylwester removes setting disabled bit when syncing filters as this
prevents VFs from completing setup.
Norbert cleans up sparse warnings.
v2:
- Fix fixes tag on patch 7
* '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
i40e: Fix endianness conversions
i40e: Fix add TC filter for IPv6
i40e: Fix VFs not created
i40e: Fix addition of RX filters after enabling FW LLDP agent
i40e: Fix overwriting flow control settings during driver loading
i40e: Add zero-initialization of AQ command structures
i40e: Fix memory leak in i40e_probe
i40e: Fix flow for IPv6 next header (extension header)
====================
Link: https://lore.kernel.org/r/20210219213606.2567536-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
When link speed is not 100 Mbps, port transmit rate and speed divider
are set to 8 and 1000000 respectively. These values are incorrect for
CBS idleslope and sendslope HW values calculation if the link speed is
not 1 Gbps.
This patch adds switch statement to set the values of port transmit rate
and speed divider for 10 Gbps, 5 Gbps, 2.5 Gbps, 1 Gbps, and 100 Mbps.
Note that CBS is not supported at 10 Mbps.
Fixes: bc41a6689b ("net: stmmac: tc: Remove the speed dependency")
Fixes: 1f705bc61a ("net: stmmac: Add support for CBS QDISC")
Signed-off-by: Song, Yoong Siang <yoong.siang.song@intel.com>
Link: https://lore.kernel.org/r/1613655653-11755-1-git-send-email-yoong.siang.song@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
The comments to phy_select_page() say that "phy_restore_page() must
always be called after this, irrespective of success or failure of this
call." If we don't call phy_restore_page() then we are still holding
the phy_lock_mdio_bus() so it eventually leads to a dead lock.
Fixes: 32ab60e539 ("net: phy: icplus: add MDI/MDIX support for IP101A/G")
Fixes: f9bc51e6cc ("net: phy: icplus: fix paged register access")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Michael Walle <michael@walle.cc>
Reviewed-by: Russell King <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/YC+OpFGsDPXPnXM5@mwanda
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
