Change ndisc_redirect_rcv() to return a drop reason.
For the moment, return PKT_TOO_SMALL, NOT_SPECIFIED
and values from icmpv6_notify().
More reasons are added later.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change ndisc_router_discovery() to return a drop reason.
For the moment, return PKT_TOO_SMALL, NOT_SPECIFIED
and SKB_CONSUMED.
More reasons are added later.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change ndisc_recv_rs() to return a drop reason.
For the moment, return PKT_TOO_SMALL, NOT_SPECIFIED
or SKB_CONSUMED. More reasons are added later.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change ndisc_recv_na() to return a drop reason.
For the moment, return PKT_TOO_SMALL, NOT_SPECIFIED
or SKB_CONSUMED. More reasons are added later.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change ndisc_recv_ns() to return a drop reason.
For the moment, return PKT_TOO_SMALL, NOT_SPECIFIED
or SKB_CONSUMED.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we try to start AF_XDP on some machines with long running time, due
to the machine's memory fragmentation problem, there is no sufficient
contiguous physical memory that will cause the start failure.
If the size of the queue is 8 * 1024, then the size of the desc[] is
8 * 1024 * 8 = 16 * PAGE, but we also add struct xdp_ring size, so it is
16page+. This is necessary to apply for a 4-order memory. If there are a
lot of queues, it is difficult to these machine with long running time.
Here, that we actually waste 15 pages. 4-Order memory is 32 pages, but
we only use 17 pages.
This patch replaces __get_free_pages() by vmalloc() to allocate memory
to solve these problems.
Signed-off-by: Xuan Zhuo <xuanzhuo@linux.alibaba.com>
Acked-by: Magnus Karlsson <magnus.karlsson@intel.com>
Reviewed-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean says:
====================
taprio queueMaxSDU fixes
This fixes 3 issues noticed while attempting to reoffload the
dynamically calculated queueMaxSDU values. These are:
- Dynamic queueMaxSDU is not calculated correctly due to a lost patch
- Dynamically calculated queueMaxSDU needs to be clamped on the low end
- Dynamically calculated queueMaxSDU needs to be clamped on the high end
====================
Link: https://lore.kernel.org/r/20230215224632.2532685-1-vladimir.oltean@nxp.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
It makes no sense to keep randomly large max_sdu values, especially if
larger than the device's max_mtu. These are visible in "tc qdisc show".
Such a max_sdu is practically unlimited and will cause no packets for
that traffic class to be dropped on enqueue.
Just set max_sdu_dynamic to U32_MAX, which in the logic below causes
taprio to save a max_frm_len of U32_MAX and a max_sdu presented to user
space of 0 (unlimited).
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The overhead specified in the size table comes from the user. With small
time intervals (or gates always closed), the overhead can be larger than
the max interval for that traffic class, and their difference is
negative.
What we want to happen is for max_sdu_dynamic to have the smallest
non-zero value possible (1) which means that all packets on that traffic
class are dropped on enqueue. However, since max_sdu_dynamic is u32, a
negative is represented as a large value and oversized dropping never
happens.
Use max_t with int to force a truncation of max_frm_len to no smaller
than dev->hard_header_len + 1, which in turn makes max_sdu_dynamic no
smaller than 1.
Fixes: fed87cc671 ("net/sched: taprio: automatically calculate queueMaxSDU based on TC gate durations")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
taprio_calculate_gate_durations() depends on netdev_get_num_tc() and
this returns 0. So it calculates the maximum gate durations for no
traffic class.
I had tested the blamed commit only with another patch in my tree, one
which in the end I decided isn't valuable enough to submit ("net/sched:
taprio: mask off bits in gate mask that exceed number of TCs").
The problem is that having this patch threw off my testing. By moving
the netdev_set_num_tc() call earlier, we implicitly gave to
taprio_calculate_gate_durations() the information it needed.
Extract only the portion from the unsubmitted change which applies the
mqprio configuration to the netdev earlier.
Link: https://patchwork.kernel.org/project/netdevbpf/patch/20230130173145.475943-15-vladimir.oltean@nxp.com/
Fixes: a306a90c8f ("net/sched: taprio: calculate tc gate durations")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Kurt Kanzenbach <kurt@linutronix.de>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Fix three cases of overproduction of wakeups:
(1) rxrpc_input_split_jumbo() conditionally notifies the app that there's
data for recvmsg() to collect if it queues some data - and then its
only caller, rxrpc_input_data(), goes and wakes up recvmsg() anyway.
Fix the rxrpc_input_data() to only do the wakeup in failure cases.
(2) If a DATA packet is received for a call by the I/O thread whilst
recvmsg() is busy draining the call's rx queue in the app thread, the
call will left on the recvmsg() queue for recvmsg() to pick up, even
though there isn't any data on it.
This can cause an unexpected recvmsg() with a 0 return and no MSG_EOR
set after the reply has been posted to a service call.
Fix this by discarding pending calls from the recvmsg() queue that
don't need servicing yet.
(3) Not-yet-completed calls get requeued after having data read from them,
even if they have no data to read.
Fix this by only requeuing them if they have data waiting on them; if
they don't, the I/O thread will requeue them when data arrives or they
fail.
Signed-off-by: David Howells <dhowells@redhat.com>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: linux-afs@lists.infradead.org
Link: https://lore.kernel.org/r/3386149.1676497685@warthog.procyon.org.uk
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Alex Elder says:
====================
net: final GSI register updates
I believe this is the last set of changes required to allow IPA v5.0
to be supported. There is a little cleanup work remaining, but that
can happen in the next Linux release cycle. Otherwise we just need
config data and register definitions for IPA v5.0 (and DTS updates).
These are ready but won't be posted without further testing.
The first patch in this series fixes a minor bug in a patch just
posted, which I found too late. The second eliminates the GSI
memory "adjustment"; this was done previously to avoid/delay the
need to implement a more general way to define GSI register offsets.
Note that this patch causes "checkpatch" warnings due to indentation
that aligns with an open parenthesis.
The third patch makes use of the newly-defined register offsets, to
eliminate the need for a function that hid a few details. The next
modifies a different helper function to work properly for IPA v5.0+.
The fifth patch changes the way the event ring size is specified
based on how it's now done for IPA v5.0+. And the last defines a
new register required for IPA v5.0+.
====================
Link: https://lore.kernel.org/r/20230215195352.755744-1-elder@linaro.org
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Starting at IPA v5.0, the number of event rings per EE is defined
in a field in a new HW_PARAM_4 GSI register rather than HW_PARAM_2.
Define this new register and its fields, and update the code that
checks the number of rings supported by hardware to use the proper
field based on IPA version.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Starting with IPA v5.0, a channel's event ring index is encoded in
a field in the CH_C_CNTXT_1 GSI register rather than CH_C_CNTXT_0.
Define a new field ID for the former register and encode the event
ring in the appropriate register.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The GSI channel protocol field in the CH_C_CNTXT_0 GSI register is
widened starting IPA v5.0, making the CHTYPE_PROTOCOL_MSB field
added in IPA v4.5 unnecessary. Update the code to reflect this.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Now that we explicitly define each register field width there is no
need to have a special encoding function for the event ring length.
Add a field for this to the EV_CH_E_CNTXT_1 GSI register, and use it
in place of ev_ch_e_cntxt_1_length_encode() (which can be removed).
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Starting at IPA v4.5, almost all GSI registers had their offsets
changed by a fixed amount (shifted downward by 0xd000). Rather than
defining offsets for all those registers dependent on version, an
adjustment was applied for most register accesses. This was
implemented in commit cdeee49f3e ("net: ipa: adjust GSI register
addresses"). It was later modified to be a bit more obvious about
the adjusment, in commit 571b1e7e58 ("net: ipa: use a separate
pointer for adjusted GSI memory").
We now are able to define every GSI register with its own offset, so
there's no need to implement this special adjustment.
So get rid of the "virt_raw" pointer, and just maintain "virt" as
the (non-adjusted) base address of I/O mapped GSI register memory.
Redefine the offsets of all GSI registers (other than the INTER_EE
ones, which were not subject to the adjustment) for IPA v4.5+,
subtracting 0xd000 from their defined offsets instead.
Move the ERROR_LOG and ERROR_LOG_CLR definitions further down in the
register definition files so all registers are defined in order of
their offset.
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
I spotted an error in a patch posted this week, unfortunately just
after it got accepted. The effect of the bug is that time-based
interrupt moderation is disabled. This is not technically a bug,
but it is not what is intended. The problem is that a |= assignment
got implemented as a simple assignment, so the previously assigned
value was ignored.
Fixes: edc6158b18 ("net: ipa: define fields for event-ring related registers")
Signed-off-by: Alex Elder <elder@linaro.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Pull drm fixes from Dave Airlie:
"Just a final collection of misc fixes, the biggest disables the
recently added dynamic debugging support, it has a regression that
needs some bigger fixes.
Otherwise a bunch of fixes across the board, vc4, amdgpu and vmwgfx
mostly, with some smaller i915 and ast fixes.
drm:
- dynamic debug disable for now
fbdev:
- deferred i/o device close fix
amdgpu:
- Fix GC11.x suspend warning
- Fix display warning
vc4:
- YUV planes fix
- hdmi display fix
- crtc reduced blanking fix
ast:
- fix start address computation
vmwgfx:
- fix bo/handle races
i915:
- gen11 WA fix"
* tag 'drm-fixes-2023-02-17' of git://anongit.freedesktop.org/drm/drm:
drm/amd/display: Fail atomic_check early on normalize_zpos error
drm/amd/amdgpu: fix warning during suspend
drm/vmwgfx: Do not drop the reference to the handle too soon
drm/vmwgfx: Stop accessing buffer objects which failed init
drm/i915/gen11: Wa_1408615072/Wa_1407596294 should be on GT list
drm: Disable dynamic debug as broken
drm/ast: Fix start address computation
fbdev: Fix invalid page access after closing deferred I/O devices
drm/vc4: crtc: Increase setup cost in core clock calculation to handle extreme reduced blanking
drm/vc4: hdmi: Always enable GCP with AVMUTE cleared
drm/vc4: Fix YUV plane handling when planes are in different buffers
Pull networking fixes from Jakub Kicinski:
"Fixes from the main networking tree only, probably because all
sub-trees have backed off and haven't submitted their changes.
None of the fixes here are particularly scary and no outstanding
regressions. In an ideal world the "current release" sections would be
empty at this stage but that never happens.
Current release - regressions:
- fix unwanted sign extension in netdev_stats_to_stats64()
Current release - new code bugs:
- initialize net->notrefcnt_tracker earlier
- devlink: fix netdev notifier chain corruption
- nfp: make sure mbox accesses in IPsec code are atomic
- ice: fix check for weight and priority of a scheduling node
Previous releases - regressions:
- ice: xsk: fix cleaning of XDP_TX frame, prevent inf loop
- igb: fix I2C bit banging config with external thermal sensor
Previous releases - always broken:
- sched: tcindex: update imperfect hash filters respecting rcu
- mpls: fix stale pointer if allocation fails during device rename
- dccp/tcp: avoid negative sk_forward_alloc by ipv6_pinfo.pktoptions
- remove WARN_ON_ONCE(sk->sk_forward_alloc) from
sk_stream_kill_queues()
- af_key: fix heap information leak
- ipv6: fix socket connection with DSCP (correct interpretation of
the tclass field vs fib rule matching)
- tipc: fix kernel warning when sending SYN message
- vmxnet3: read RSS information from the correct descriptor (eop)"
* tag 'net-6.2-final' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (35 commits)
devlink: Fix netdev notifier chain corruption
igb: conditionalize I2C bit banging on external thermal sensor support
net: mpls: fix stale pointer if allocation fails during device rename
net/sched: tcindex: search key must be 16 bits
tipc: fix kernel warning when sending SYN message
igb: Fix PPS input and output using 3rd and 4th SDP
net: use a bounce buffer for copying skb->mark
ixgbe: add double of VLAN header when computing the max MTU
i40e: add double of VLAN header when computing the max MTU
ixgbe: allow to increase MTU to 3K with XDP enabled
net: stmmac: Restrict warning on disabling DMA store and fwd mode
net/sched: act_ctinfo: use percpu stats
net: stmmac: fix order of dwmac5 FlexPPS parametrization sequence
ice: fix lost multicast packets in promisc mode
ice: Fix check for weight and priority of a scheduling node
bnxt_en: Fix mqprio and XDP ring checking logic
net: Fix unwanted sign extension in netdev_stats_to_stats64()
net/usb: kalmia: Don't pass act_len in usb_bulk_msg error path
net: openvswitch: fix possible memory leak in ovs_meter_cmd_set()
af_key: Fix heap information leak
...
Pull block fixes from Jens Axboe:
"Just a few NVMe fixes that should go into the 6.2 release, adding a
quirk and fixing two issues introduced in this release:
- NVMe fixes via Christoph:
- Always return an ERR_PTR from nvme_pci_alloc_dev (Irvin Cote)
- Add bogus ID quirk for ADATA SX6000PNP (Daniel Wagner)
- Set the DMA mask earlier (Christoph Hellwig)"
* tag 'block-6.2-2023-02-16' of git://git.kernel.dk/linux:
nvme-pci: always return an ERR_PTR from nvme_pci_alloc_dev
nvme-pci: set the DMA mask earlier
nvme-pci: add bogus ID quirk for ADATA SX6000PNP
Pull spi fix from Mark Brown:
"One more last minute patch for v6.2 updating the parsing of the newly
added spi-cs-setup-delay-ns.
It's been pointed out that due to the way DT parsing works the change
in property size is ABI visible so let's not let a release go out
without it being fixed. The change got split from some earlier ABI
related fixes to the property since the first version sent had a build
error"
* tag 'spi-v6.2-rc8-abi' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
spi: Use a 32-bit DT property for spi-cs-setup-delay-ns
Pull gpio fixes from Bartosz Golaszewski:
- fix a potential Kconfig issue with gpio-mlxbf2 not selecting
GPIOLIB_IRQCHIP
- another immutable irqchip conversion, this time for gpio-vf610
- fix a wakeup issue on Clevo NH5xAx
* tag 'gpio-fixes-for-v6.2' of git://git.kernel.org/pub/scm/linux/kernel/git/brgl/linux:
gpio: mlxbf2: select GPIOLIB_IRQCHIP
gpiolib: acpi: Add a ignore wakeup quirk for Clevo NH5xAx
gpio: vf610: make irq_chip immutable
gpiolib: acpi: remove redundant declaration
The uuid code is very low maintainance now that the major overhaul
has completed, and doesn't need it's own tree. All the recent work
has been done by Andy who'd like to stay on as a reviewer without an
explicit tree.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Acked-by: Andy Shevchenko <andy@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This code has been stale for years and I have no way to test it.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Johannes Berg says:
====================
Major stack changes:
* EHT channel puncturing support (client & AP)
* some support for AP MLD without mac80211
* fixes for A-MSDU on mesh connections
Major driver changes:
iwlwifi
* EHT rate reporting
* Bump FW API to 74 for AX devices
* STEP equalizer support: transfer some STEP (connection to radio
on platforms with integrated wifi) related parameters from the
BIOS to the firmware
mt76
* switch to using page pool allocator
* mt7996 EHT (Wi-Fi 7) support
* Wireless Ethernet Dispatch (WED) reset support
libertas
* WPS enrollee support
brcmfmac
* Rename Cypress 89459 to BCM4355
* BCM4355 and BCM4377 support
mwifiex
* SD8978 chipset support
rtl8xxxu
* LED support
ath12k
* new driver for Qualcomm Wi-Fi 7 devices
ath11k
* IPQ5018 support
* Fine Timing Measurement (FTM) responder role support
* channel 177 support
ath10k
* store WLAN firmware version in SMEM image table
* tag 'wireless-next-2023-03-16' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (207 commits)
wifi: brcmfmac: p2p: Introduce generic flexible array frame member
wifi: mac80211: add documentation for amsdu_mesh_control
wifi: cfg80211: remove gfp parameter from cfg80211_obss_color_collision_notify description
wifi: mac80211: always initialize link_sta with sta
wifi: mac80211: pass 'sta' to ieee80211_rx_data_set_sta()
wifi: cfg80211: Set SSID if it is not already set
wifi: rtw89: move H2C of del_pkt_offload before polling FW status ready
wifi: rtw89: use readable return 0 in rtw89_mac_cfg_ppdu_status()
wifi: rtw88: usb: drop now unnecessary URB size check
wifi: rtw88: usb: send Zero length packets if necessary
wifi: rtw88: usb: Set qsel correctly
wifi: mac80211: fix off-by-one link setting
wifi: mac80211: Fix for Rx fragmented action frames
wifi: mac80211: avoid u32_encode_bits() warning
wifi: mac80211: Don't translate MLD addresses for multicast
wifi: cfg80211: call reg_notifier for self managed wiphy from driver hint
wifi: cfg80211: get rid of gfp in cfg80211_bss_color_notify
wifi: nl80211: Allow authentication frames and set keys on NAN interface
wifi: mac80211: fix non-MLO station association
wifi: mac80211: Allow NSS change only up to capability
...
====================
Link: https://lore.kernel.org/r/20230216105406.208416-1-johannes@sipsolutions.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Andrea Mayer says:
====================
seg6: add PSP flavor support for SRv6 End behavior
Segment Routing for IPv6 (SRv6 in short) [1] is the instantiation of the
Segment Routing (SR) [2] architecture on the IPv6 dataplane.
In SRv6, the segment identifiers (SID) are IPv6 addresses and the segment list
(SID List) is carried in the Segment Routing Header (SRH). A segment may be
bound to a specific packet processing operation called "behavior". The RFC8986
[3] defines and standardizes the most common/relevant behaviors for network
operators, e.g., End, End.X and End.T and so on.
The RFC8986 also introduces the "flavors" framework aiming to modify or extend
the capabilities of SRv6 End, End.X and End.T behaviors. Specifically, these
behaviors support the following flavors (either individually or in
combinations):
- Penultimate Segment Pop (PSP);
- Ultimate Segment Pop (USP);
- Ultimate Segment Decapsulation (USD).
Such flavors enable an End/End.X/End.T behavior to pop the SRH on the
penultimate/ultimate SR endpoint node listed in the SID List or to perform a
full decapsulation.
Currently, the Linux kernel supports a large subset of behaviors described in
RFC8986, including the End, End.X and End.T. However, PSP, USP and USD flavors
have not yet been implemented.
In this patchset, we extend the SRv6 subsystem to implement the PSP flavor in
the SRv6 End behavior. To accomplish this task, we leverage the flavor
framework previously introduced by another patchset required for supporting the
efficient representation of the SID List through the NEXT-C-SID mechanism [4].
In details, the patchset is made of:
- patch 1/3: seg6: factor out End lookup nexthop processing to a dedicated
function
- patch 2/3: seg6: add PSP flavor support for SRv6 End behavior
- patch 3/3: selftests: seg6: add selftest for PSP flavor in SRv6 End
behavior
From the user space perspective, we do not need to change the iproute2 code to
support the PSP flavor. However, we provide the man page for the PSP flavor in
a separate patch.
Comments, improvements and suggestions are always appreciated.
[1] - RFC8754: https://datatracker.ietf.org/doc/html/rfc8754
[2] - RFC8402: https://datatracker.ietf.org/doc/html/rfc8402
[3] - RFC8986: https://datatracker.ietf.org/doc/html/rfc8986
[4] - https://datatracker.ietf.org/doc/html/draft-ietf-spring-srv6-srh-compression
====================
Link: https://lore.kernel.org/r/20230215134659.7613-1-andrea.mayer@uniroma2.it
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
This selftest is designed for testing the PSP flavor in SRv6 End behavior.
It instantiates a virtual network composed of several nodes: hosts and
SRv6 routers. Each node is realized using a network namespace that is
properly interconnected to others through veth pairs.
The test makes use of the SRv6 End behavior and of the PSP flavor needed
for removing the SRH from the IPv6 header at the penultimate node.
The correct execution of the behavior is verified through reachability
tests carried out between hosts.
Signed-off-by: Andrea Mayer <andrea.mayer@uniroma2.it>
Signed-off-by: Paolo Lungaroni <paolo.lungaroni@uniroma2.it>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The "flavors" framework defined in RFC8986 [1] represents additional
operations that can modify or extend a subset of existing behaviors such as
SRv6 End, End.X and End.T. We report these flavors hereafter:
- Penultimate Segment Pop (PSP);
- Ultimate Segment Pop (USP);
- Ultimate Segment Decapsulation (USD).
Depending on how the Segment Routing Header (SRH) has to be handled, an
SRv6 End* behavior can support these flavors either individually or in
combinations.
In this patch, we only consider the PSP flavor for the SRv6 End behavior.
A PSP enabled SRv6 End behavior is used by the Source/Ingress SR node
(i.e., the one applying the SRv6 Policy) when it needs to instruct the
penultimate SR Endpoint node listed in the SID List (carried by the SRH) to
remove the SRH from the IPv6 header.
Specifically, a PSP enabled SRv6 End behavior processes the SRH by:
i) decreasing the Segment Left (SL) from 1 to 0;
ii) copying the Last Segment IDentifier (SID) into the IPv6 Destination
Address (DA);
iii) removing (i.e., popping) the outer SRH from the extension headers
following the IPv6 header.
It is important to note that PSP operation (steps i, ii, iii) takes place
only at a penultimate SR Segment Endpoint node (i.e., when the SL=1) and
does not happen at non-penultimate Endpoint nodes. Indeed, when a SID of
PSP flavor is processed at a non-penultimate SR Segment Endpoint node, the
PSP operation is not performed because it would not be possible to decrease
the SL from 1 to 0.
SL=2 SL=1 SL=0
| | |
For example, given the SRv6 policy (SID List := < X, Y, Z >):
- a PSP enabled SRv6 End behavior bound to SID "Y" will apply the PSP
operation as Segment Left (SL) is 1, corresponding to the Penultimate
Segment of the SID List;
- a PSP enabled SRv6 End behavior bound to SID "X" will *NOT* apply the
PSP operation as the Segment Left is 2. This behavior instance will
apply the "standard" End packet processing, ignoring the configured PSP
flavor at all.
[1] - RFC8986: https://datatracker.ietf.org/doc/html/rfc8986
Signed-off-by: Andrea Mayer <andrea.mayer@uniroma2.it>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
The End nexthop lookup/input operations are moved into a new helper
function named input_action_end_finish(). This avoids duplicating the
code needed to compute the nexthop in the different flavors of the End
behavior.
Signed-off-by: Andrea Mayer <andrea.mayer@uniroma2.it>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Alejandro Lucero says:
====================
sfc: devlink support for ef100
This patchset adds devlink port support for ef100 allowing setting VFs
mac addresses through the VF representor devlink ports.
Basic devlink infrastructure is first introduced, then support for info
command. Next changes for enumerating MAE ports which will be used for
devlink port creation when netdevs are registered.
Adding support for devlink port_function_hw_addr_get requires changes in
the ef100 driver for getting the mac address based on a client handle.
This allows to obtain VFs mac addresses during netdev initialization as
well what is included in patch 6.
Such client handle is used in patches 7 and 8 for getting and setting
devlink port addresses.
====================
Link: https://lore.kernel.org/r/20230215090828.11697-1-alejandro.lucero-palau@amd.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Using the builtin client handle id infrastructure, add support for
setting the mac address linked to mports in ef100. This implies to
execute an MCDI command for giving the address to the firmware for
the specific devlink port.
Signed-off-by: Alejandro Lucero <alejandro.lucero-palau@amd.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Using the builtin client handle id infrastructure, add support for
obtaining the mac address linked to mports in ef100. This implies
to execute an MCDI command for getting the data from the firmware
for each devlink port.
Signed-off-by: Alejandro Lucero <alejandro.lucero-palau@amd.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Getting device mac address is currently based on a specific MCDI command
only available for the PF. This patch changes the MCDI command to a
generic one for PFs and VFs based on a client handle. This allows both
PFs and VFs to ask for their mac address during initialization using the
CLIENT_HANDLE_SELF.
Moreover, the patch allows other client handles which will be used by
the PF to ask for mac addresses linked to VFs. This is necessary for
suporting the port_function_hw_addr_get devlink function in further
patches.
Signed-off-by: Alejandro Lucero <alejandro.lucero-palau@amd.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Using the data when enumerating mports, create devlink ports just before
netdevs are registered and remove those devlink ports after netdev has
been unregistered.
Signed-off-by: Alejandro Lucero <alejandro.lucero-palau@amd.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Obtaining mport id is based on asking the firmware about it. This is
still needed for mport initialization itself, but once the mport data is
now kept by the driver, further mport id request can be satisfied
internally without firmware interaction.
Previous function is just modified in name making clear the firmware
interaction. The new function uses the old name and looks for the data
in the mport data structure.
Signed-off-by: Alejandro Lucero <alejandro.lucero-palau@amd.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
MAE ports (mports) are the ports on the EF100 embedded switch such
as networking PCIe functions, the physical port, and potentially
others.
Signed-off-by: Alejandro Lucero <alejandro.lucero-palau@amd.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Add devlink info support for ef100. The information reported is obtained
through the MCDI interface with the specific meaning defined in new
documentation file.
Signed-off-by: Alejandro Lucero <alejandro.lucero-palau@amd.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Acked-by: Martin Habets <habetsm.xilinx@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Cited commit changed devlink to register its netdev notifier block on
the global netdev notifier chain instead of on the per network namespace
one.
However, when changing the network namespace of the devlink instance,
devlink still tries to unregister its notifier block from the chain of
the old namespace and register it on the chain of the new namespace.
This results in corruption of the notifier chains, as the same notifier
block is registered on two different chains: The global one and the per
network namespace one. In turn, this causes other problems such as the
inability to dismantle namespaces due to netdev reference count issues.
Fix by preventing devlink from moving its notifier block between
namespaces.
Reproducer:
# echo "10 1" > /sys/bus/netdevsim/new_device
# ip netns add test123
# devlink dev reload netdevsim/netdevsim10 netns test123
# ip netns del test123
[ 71.935619] unregister_netdevice: waiting for lo to become free. Usage count = 2
[ 71.938348] leaked reference.
Fixes: 565b4824c3 ("devlink: change port event netdev notifier from per-net to global")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230215073139.1360108-1-idosch@nvidia.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Pedro Tammela says:
====================
net/sched: transition actions to pcpu stats and rcu
Following the work done for act_pedit[0], transition the remaining tc
actions to percpu stats and rcu, whenever possible.
Percpu stats make updating the action stats very cheap, while combining
it with rcu action parameters makes it possible to get rid of the per
action lock in the datapath.
For act_connmark and act_nat we run the following tests:
- tc filter add dev ens2f0 ingress matchall action connmark
- tc filter add dev ens2f0 ingress matchall action nat ingress any 10.10.10.10
Our setup consists of a 26 cores Intel CPU and a 25G NIC.
We use TRex to shoot 10mpps TCP packets and take perf measurements.
Both actions improved performance as expected since the datapath lock disappeared.
For act_pedit we move the drop counter to percpu, when available.
For act_gate we move the counters to percpu, when available.
[0] https://lore.kernel.org/all/20230131145149.3776656-1-pctammela@mojatatu.com/
====================
Link: https://lore.kernel.org/r/20230214211534.735718-1-pctammela@mojatatu.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
