| Age | Commit message (Collapse) | Author |
|---|
|
The remove function has these in the wrong order. The switch should be
unregistered last. Simpler to use devm so that the right thing is done.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://patch.msgid.link/20250708014144.2514-3-rosenp@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
There's no architecture specific requirement for it to compile. Allows
the bots to test compilation properly.
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://patch.msgid.link/20250708014144.2514-2-rosenp@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
This patch provides a setsockopt method to let applications leverage to
adjust how many descs to be handled at most in one send syscall. It
mitigates the situation where the default value (32) that is too small
leads to higher frequency of triggering send syscall.
Considering the prosperity/complexity the applications have, there is no
absolutely ideal suggestion fitting all cases. So keep 32 as its default
value like before.
The patch does the following things:
- Add XDP_MAX_TX_SKB_BUDGET socket option.
- Set max_tx_budget to 32 by default in the initialization phase as a
per-socket granular control.
- Set the range of max_tx_budget as [32, xs->tx->nentries].
The idea behind this comes out of real workloads in production. We use a
user-level stack with xsk support to accelerate sending packets and
minimize triggering syscalls. When the packets are aggregated, it's not
hard to hit the upper bound (namely, 32). The moment user-space stack
fetches the -EAGAIN error number passed from sendto(), it will loop to try
again until all the expected descs from tx ring are sent out to the driver.
Enlarging the XDP_MAX_TX_SKB_BUDGET value contributes to less frequency of
sendto() and higher throughput/PPS.
Here is what I did in production, along with some numbers as follows:
For one application I saw lately, I suggested using 128 as max_tx_budget
because I saw two limitations without changing any default configuration:
1) XDP_MAX_TX_SKB_BUDGET, 2) socket sndbuf which is 212992 decided by
net.core.wmem_default. As to XDP_MAX_TX_SKB_BUDGET, the scenario behind
this was I counted how many descs are transmitted to the driver at one
time of sendto() based on [1] patch and then I calculated the
possibility of hitting the upper bound. Finally I chose 128 as a
suitable value because 1) it covers most of the cases, 2) a higher
number would not bring evident results. After twisting the parameters,
a stable improvement of around 4% for both PPS and throughput and less
resources consumption were found to be observed by strace -c -p xxx:
1) %time was decreased by 7.8%
2) error counter was decreased from 18367 to 572
[1]: https://lore.kernel.org/all/20250619093641.70700-1-kerneljasonxing@gmail.com/
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Acked-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Link: https://patch.msgid.link/20250704160138.48677-1-kerneljasonxing@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
In reconfig we add the virtual monitor in 2 cases:
1. If we are resuming (it was deleted on suspend)
2. If it was added after an error but before the reconfig
(due to the last non-monitor interface removal).
In the second case, the removal of the non-monitor interface will succeed
but the addition of the virtual monitor will fail, so we add it in the
reconfig.
The problem is that we mislead the driver to think that this is an existing
interface that is getting re-added - while it is actually a completely new
interface from the drivers' point of view.
Some drivers act differently when a interface is re-added. For example, it
might not initialize things because they were already initialized.
Such drivers will - in this case - be left with a partialy initialized vif.
To fix it, add the virtual monitor after reconfig_complete, so the
driver will know that this is a completely new interface.
Fixes: 3c3e21e7443b ("mac80211: destroy virtual monitor interface across suspend")
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://patch.msgid.link/20250709233451.648d39b041e8.I2e37b68375278987e303d6c00cc5f3d8334d2f96@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
This is currently not initialized for a virtual monitor, leading to a
NULL pointer dereference when - for example - iterating over all the
keys of all the vifs.
Reviewed-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Miri Korenblit <miriam.rachel.korenblit@intel.com>
Link: https://patch.msgid.link/20250709233400.8dcefe578497.I4c90a00ae3256520e063199d7f6f2580d5451acf@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
|
|
To prevent a potential crash in agg_dequeue (net/sched/sch_qfq.c)
when cl->qdisc->ops->peek(cl->qdisc) returns NULL, we check the return
value before using it, similar to the existing approach in sch_hfsc.c.
To avoid code duplication, the following changes are made:
1. Changed qdisc_warn_nonwc(include/net/pkt_sched.h) into a static
inline function.
2. Moved qdisc_peek_len from net/sched/sch_hfsc.c to
include/net/pkt_sched.h so that sch_qfq can reuse it.
3. Applied qdisc_peek_len in agg_dequeue to avoid crashing.
Signed-off-by: Xiang Mei <xmei5@asu.edu>
Reviewed-by: Cong Wang <xiyou.wangcong@gmail.com>
Link: https://patch.msgid.link/20250705212143.3982664-1-xmei5@asu.edu
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
|
|
Tariq Toukan says:
====================
net/mlx5: misc changes 2025-07-09
This series contains misc enhancements to the mlx5 driver.
====================
Link: https://patch.msgid.link/1752009387-13300-1-git-send-email-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
RQTs (Receive Queue Table) should redirect traffic to the channels' RQs
when they're active. Otherwise, redirect to the designated "drop RQ".
RQTs are created in "inactive" state, pointing to the "drop RQ".
In activate and de-activate flows, do not "deactivate" the rest of RQTs
(beyond the num of channels), as they are already inactive.
This cuts down unnecessary execution of FW commands (MODIFY_RQT), and
improves the latency of open/close channels or configuration change.
Perf:
NIC: Connect-X7.
Configuration: 1 combined channel, max num channels 248.
Measure time for "interface up + interface down".
Before: 0.313 sec
After: 0.057 sec (5.5x faster)
247 MODIFY_RQT commands saved in interface up.
247 MODIFY_RQT commands saved in interface down.
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/1752009387-13300-6-git-send-email-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Warn if write combining is not supported, as it can impact latency.
Add the warning message to be printed only when the driver actually
run the test and detect unsupported state, rather than when
inheriting parent's result for SFs.
Signed-off-by: Maor Gottlieb <maorg@nvidia.com>
Reviewed-by: Michael Guralnik <michaelgur@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/1752009387-13300-5-git-send-email-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
mlx5e_tc_act_vlan_add_push_action() uses tail-recursion to walk through
a stack of VLAN devices.
There is no need for a complicated recursion with unnecessary stack
consumption and less obvious code flow, rewrite the function so that it
uses a do while loop instead.
Signed-off-by: Gal Pressman <gal@nvidia.com>
Reviewed-by: Cosmin Ratiu <cratiu@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/1752009387-13300-4-git-send-email-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
This reduces the time the lock is held and reduces contention.
Signed-off-by: Cosmin Ratiu <cratiu@nvidia.com>
Reviewed-by: Dragos Tatulea <dtatulea@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/1752009387-13300-3-git-send-email-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The VLAN insertion capability (`wqe_vlan_insert`) was never enabled on
all mlx5 devices. When VLAN TX offload is advertised but this
capability is not supported, the driver uses inline headers to insert
the VLAN tag.
To support this, the driver used to set the
`MLX5E_SQ_STATE_VLAN_NEED_L2_INLINE` bit to enforce L2 inline mode
when `wqe_vlan_insert` was not supported. Since the capability is
disabled on all devices, this logic was always active, and the SQ flag
has become redundant. L2 inline is enforced unconditionally for
VLAN-tagged packets.
The `skb_vlan_tag_present()` check in the else-if block of
`mlx5e_sq_xmit_wqe()` is never true by this point in the TX flow,
as the VLAN tag has already been inserted by the driver using inline
headers. As a result, this code is never executed.
Remove the redundant SQ state, dead VLAN insertion code block, and
related logic.
Signed-off-by: Carolina Jubran <cjubran@nvidia.com>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/1752009387-13300-2-git-send-email-tariqt@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
David Howells says:
====================
rxrpc: Miscellaneous fixes
Here are some miscellaneous fixes for rxrpc:
(1) Fix assertion failure due to preallocation collision.
(2) Fix oops due to prealloc backlog struct not yet having been allocated
if no service calls have yet been preallocated.
====================
Link: https://patch.msgid.link/20250708211506.2699012-1-dhowells@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
If an AF_RXRPC service socket is opened and bound, but calls are
preallocated, then rxrpc_alloc_incoming_call() will oops because the
rxrpc_backlog struct doesn't get allocated until the first preallocation is
made.
Fix this by returning NULL from rxrpc_alloc_incoming_call() if there is no
backlog struct. This will cause the incoming call to be aborted.
Reported-by: Junvyyang, Tencent Zhuque Lab <zhuque@tencent.com>
Suggested-by: Junvyyang, Tencent Zhuque Lab <zhuque@tencent.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: LePremierHomme <kwqcheii@proton.me>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Willy Tarreau <w@1wt.eu>
cc: Simon Horman <horms@kernel.org>
cc: linux-afs@lists.infradead.org
Link: https://patch.msgid.link/20250708211506.2699012-3-dhowells@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
When userspace is using AF_RXRPC to provide a server, it has to preallocate
incoming calls and assign to them call IDs that will be used to thread
related recvmsg() and sendmsg() together. The preallocated call IDs will
automatically be attached to calls as they come in until the pool is empty.
To the kernel, the call IDs are just arbitrary numbers, but userspace can
use the call ID to hold a pointer to prepared structs. In any case, the
user isn't permitted to create two calls with the same call ID (call IDs
become available again when the call ends) and EBADSLT should result from
sendmsg() if an attempt is made to preallocate a call with an in-use call
ID.
However, the cleanup in the error handling will trigger both assertions in
rxrpc_cleanup_call() because the call isn't marked complete and isn't
marked as having been released.
Fix this by setting the call state in rxrpc_service_prealloc_one() and then
marking it as being released before calling the cleanup function.
Fixes: 00e907127e6f ("rxrpc: Preallocate peers, conns and calls for incoming service requests")
Reported-by: Junvyyang, Tencent Zhuque Lab <zhuque@tencent.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: LePremierHomme <kwqcheii@proton.me>
cc: Marc Dionne <marc.dionne@auristor.com>
cc: Simon Horman <horms@kernel.org>
cc: linux-afs@lists.infradead.org
Link: https://patch.msgid.link/20250708211506.2699012-2-dhowells@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
In test_stream_transport_change_client(), the client sends CONTROL_CONTINUE
on each iteration, even when connect() is unsuccessful. This causes a flood
of control messages in the server that hangs around for more than 10
seconds after the test finishes, triggering several timeouts and causing
subsequent tests to fail. This was discovered in testing a newly proposed
test that failed in this way on the client side:
...
33 - SOCK_STREAM transport change null-ptr-deref...ok
34 - SOCK_STREAM ioctl(SIOCINQ) functionality...recv timed out
The CONTROL_CONTINUE message is used only to tell to the server to call
accept() to consume successful connections, so that subsequent connect()
will not fail for finding the queue full.
Send CONTROL_CONTINUE message only when the connect() has succeeded, or
found the queue full. Note that the second connect() can also succeed if
the first one was interrupted after sending the request.
Fixes: 3a764d93385c ("vsock/test: Add test for null ptr deref when transport changes")
Cc: leonardi@redhat.com
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Luigi Leonardi <leonardi@redhat.com>
Link: https://patch.msgid.link/20250708111701.129585-1-sgarzare@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
says:
====================
net: phy: bcm54811: PHY initialization
Proper bcm54811 PHY driver initialization for MII-Lite.
The bcm54811 PHY in MLP package must be setup for MII-Lite interface
mode by software. Normally, the PHY to MAC interface is selected in
hardware by setting the bootstrap pins of the PHY. However, MII and
MII-Lite share the same hardware setup and must be distinguished by
software, setting appropriate bit in a configuration register.
The MII-Lite interface mode is non-standard one, defined by Broadcom
for some of their PHYs. The MII-Lite lightness consist in omitting
RXER, TXER, CRS and COL signals of the standard MII interface.
Absence of COL them makes half-duplex links modes impossible but
does not interfere with Broadcom's BroadR-Reach link modes, because
they are full-duplex only.
To do it in a clean way, MII-Lite must be introduced first, including
its limitation to link modes (no half-duplex), because it is a
prerequisite for the patch #3 of this series. The patch #4 does not
depend on MII-Lite directly but both #3 and #4 are necessary for
bcm54811 to work properly without additional configuration steps to be
done - for example in the bootloader, before the kernel starts.
PATCH 1 - Add MII-Lite PHY interface mode as defined by Broadcom for
their two-wire PHYs. It can be used with most Ethernet controllers
under certain limitations (no half-duplex link modes etc.).
PATCH 2 - Add MII-Lite PHY interface type
PATCH 3 - Activation of MII-Lite interface mode on Broadcom bcm5481x
PHYs
PATCH 4 - Initialize the BCM54811 PHY properly so that it conforms
to the datasheet regarding a reserved bit in the LRE Control
register, which must be written to zero after every device reset.
Ignore the LDS capability bit in LRE Status register on bcm54811.
====================
Link: https://patch.msgid.link/20250708090140.61355-1-kamilh@axis.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Reset the bit 12 in PHY's LRE Control register upon initialization.
According to the datasheet, this bit must be written to zero after
every device reset.
Signed-off-by: Kamil Horák - 2N <kamilh@axis.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Link: https://patch.msgid.link/20250708090140.61355-5-kamilh@axis.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Broadcom PHYs featuring the BroadR-Reach two-wire link mode are usually
capable to operate in simplified MII mode, without TXER, RXER, CRS and
COL signals as defined for the MII. The absence of COL signal makes
half-duplex link modes impossible, however, the BroadR-Reach modes are
all full-duplex only.
Depending on the IC encapsulation, there exist MII-Lite-only PHYs such
as bcm54811 in MLP. The PHY itself is hardware-strapped to select among
multiple RGMII and MII-Lite modes, but the MII-Lite mode must be also
activated by software.
Add MII-Lite activation for bcm5481x PHYs.
Signed-off-by: Kamil Horák - 2N <kamilh@axis.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/20250708090140.61355-4-kamilh@axis.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Some Broadcom PHYs are capable to operate in simplified MII mode,
without TXER, RXER, CRS and COL signals as defined for the MII.
The MII-Lite mode can be used on most Ethernet controllers with full
MII interface by just leaving the input signals (RXER, CRS, COL)
inactive. The absence of COL signal makes half-duplex link modes
impossible but does not interfere with BroadR-Reach link modes on
Broadcom PHYs, because they are all full-duplex only.
Add new interface type "mii-lite" to phy-connection-type enum.
Signed-off-by: Kamil Horák - 2N <kamilh@axis.com>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Acked-by: Rob Herring (Arm) <robh@kernel.org>
Link: https://patch.msgid.link/20250708090140.61355-3-kamilh@axis.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Some Broadcom PHYs are capable to operate in simplified MII mode,
without TXER, RXER, CRS and COL signals as defined for the MII.
The MII-Lite mode can be used on most Ethernet controllers with full
MII interface by just leaving the input signals (RXER, CRS, COL)
inactive. The absence of COL signal makes half-duplex link modes
impossible but does not interfere with BroadR-Reach link modes on
Broadcom PHYs, because they are all full-duplex only.
Add MII-Lite interface mode, especially for Broadcom two-wire PHYs.
Signed-off-by: Kamil Horák - 2N <kamilh@axis.com>
Reviewed-by: Maxime Chevallier <maxime.chevallier@bootlin.com>
Reviewed-by: Florian Fainelli <florian.fainelli@broadcom.com>
Reviewed-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://patch.msgid.link/20250708090140.61355-2-kamilh@axis.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
I am moving on from Corigine to different things, for the moment
slightly removed from kernel development. Right now there is nobody I
can in good conscience recommend to take over the maintainer role, but
there are still people available for review, so put the driver state to
'Odd Fixes'.
Additionally add Simon Horman as reviewer - thanks Simon.
Signed-off-by: Louis Peens <louis.peens@corigine.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Oleksij reported that:
The smsc95xx driver fails after one down/up cycle, like this:
$ nmcli device set enu1u1 managed no
$ p a a 10.10.10.1/24 dev enu1u1
$ ping -c 4 10.10.10.3
$ ip l s dev enu1u1 down
$ ip l s dev enu1u1 up
$ ping -c 4 10.10.10.3
The second ping does not reach the host. Networking also fails on other interfaces.
Enable the work by replacing the disable_work_sync() with cancel_work_sync().
[Jun Miao: completely write the commit changelog]
Fixes: 2c04d279e857 ("net: usb: Convert tasklet API to new bottom half workqueue mechanism")
Reported-by: Oleksij Rempel <o.rempel@pengutronix.de>
Tested-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: Zqiang <qiang.zhang@linux.dev>
Signed-off-by: Jun Miao <jun.miao@intel.com>
Link: https://patch.msgid.link/20250708081653.307815-1-jun.miao@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Xuewei Niu says:
====================
vsock: Introduce SIOCINQ ioctl support
Introduce SIOCINQ ioctl support for vsock, indicating the length of unread
bytes.
Similar with SIOCOUTQ ioctl, the information is transport-dependent.
The first patch adds SIOCINQ ioctl support in AF_VSOCK.
Thanks to @dexuan, the second patch is to fix the issue where hyper-v
`hvs_stream_has_data()` doesn't return the readable bytes.
The third patch wraps the ioctl into `ioctl_int()`, which implements a
retry mechanism to prevent immediate failure.
The last one adds two test cases to check the functionality. The changes
have been tested, and the results are as expected.
====================
Link: https://patch.msgid.link/20250708-siocinq-v6-0-3775f9a9e359@antgroup.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add SIOCINQ ioctl tests for both SOCK_STREAM and SOCK_SEQPACKET.
The client waits for the server to send data, and checks if the SIOCINQ
ioctl value matches the data size. After consuming the data, the client
checks if the SIOCINQ value is 0.
Signed-off-by: Xuewei Niu <niuxuewei.nxw@antgroup.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Tested-by: Luigi Leonardi <leonardi@redhat.com>
Reviewed-by: Luigi Leonardi <leonardi@redhat.com>
Link: https://patch.msgid.link/20250708-siocinq-v6-4-3775f9a9e359@antgroup.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Wrap the ioctl in `ioctl_int()`, which takes a pointer to the actual
int value and an expected int value. The function will not return until
either the ioctl returns the expected value or a timeout occurs, thus
avoiding immediate failure.
Signed-off-by: Xuewei Niu <niuxuewei.nxw@antgroup.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Luigi Leonardi <leonardi@redhat.com>
Link: https://patch.msgid.link/20250708-siocinq-v6-3-3775f9a9e359@antgroup.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add support for SIOCINQ ioctl, indicating the length of bytes unread in the
socket. The value is obtained from `vsock_stream_has_data()`.
Signed-off-by: Xuewei Niu <niuxuewei.nxw@antgroup.com>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Luigi Leonardi <leonardi@redhat.com>
Link: https://patch.msgid.link/20250708-siocinq-v6-2-3775f9a9e359@antgroup.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
When hv_sock was originally added, __vsock_stream_recvmsg() and
vsock_stream_has_data() actually only needed to know whether there
is any readable data or not, so hvs_stream_has_data() was written to
return 1 or 0 for simplicity.
However, now hvs_stream_has_data() should return the readable bytes
because vsock_data_ready() -> vsock_stream_has_data() needs to know the
actual bytes rather than a boolean value of 1 or 0.
The SIOCINQ ioctl support also needs hvs_stream_has_data() to return
the readable bytes.
Let hvs_stream_has_data() return the readable bytes of the payload in
the next host-to-guest VMBus hv_sock packet.
Note: there may be multiple incoming hv_sock packets pending in the
VMBus channel's ringbuffer, but so far there is not a VMBus API that
allows us to know all the readable bytes in total without reading and
caching the payload of the multiple packets, so let's just return the
readable bytes of the next single packet. In the future, we'll either
add a VMBus API that allows us to know the total readable bytes without
touching the data in the ringbuffer, or the hv_sock driver needs to
understand the VMBus packet format and parse the packets directly.
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Signed-off-by: Xuewei Niu <niuxuewei.nxw@antgroup.com>
Acked-by: Stefano Garzarella <sgarzare@redhat.com>
Acked-by: Wei Liu <wei.liu@kernel.org>
Link: https://patch.msgid.link/20250708-siocinq-v6-1-3775f9a9e359@antgroup.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The modified lines are mainly related to the following commits[1][2]
which remove those tests and examples. Since samples/bpf has been
deprecated, we can refer to more examples that are easily searched
in the various xdp-projects, like the following link:
https://github.com/xdp-project/bpf-examples/tree/main/AF_XDP-example
[1]
commit f36600634282 ("libbpf: move xsk.{c,h} into selftests/bpf")
[2]
commit cfb5a2dbf141 ("bpf, samples: Remove AF_XDP samples")
Signed-off-by: Jason Xing <kernelxing@tencent.com>
Reviewed-by: Toke Høiland-Jørgensen <toke@redhat.com>
Reviewed-by: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Acked-by: Stanislav Fomichev <sdf@fomichev.me>
Link: https://patch.msgid.link/20250708062907.11557-1-kerneljasonxing@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
When using sockmap for forwarding, the average latency for different packet sizes
after sending 10,000 packets is as follows:
size old(us) new(us)
512 56 55
1472 58 58
1600 106 81
3000 145 105
5000 182 125
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Feng Yang <yangfeng@kylinos.cn>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250708054053.39551-1-yangfeng59949@163.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Easwar Hariharan says:
====================
Converge on using secs_to_jiffies() part two
This is the second series (part 1*) that converts users of msecs_to_jiffies() that
either use the multiply pattern of either of:
- msecs_to_jiffies(N*1000) or
- msecs_to_jiffies(N*MSEC_PER_SEC)
where N is a constant or an expression, to avoid the multiplication.
The conversion is made with Coccinelle with the secs_to_jiffies() script
in scripts/coccinelle/misc. Attention is paid to what the best change
can be rather than restricting to what the tool provides.
v1: https://lore.kernel.org/20250219-netdev-secs-to-jiffies-part-2-v1-0-c484cc63611b@linux.microsoft.com
====================
Link: https://patch.msgid.link/20250707-netdev-secs-to-jiffies-part-2-v2-0-b7817036342f@linux.microsoft.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Commit b35108a51cf7 ("jiffies: Define secs_to_jiffies()") introduced
secs_to_jiffies(). As the value here is a multiple of 1000, use
secs_to_jiffies() instead of msecs_to_jiffies to avoid the multiplication.
This is converted using scripts/coccinelle/misc/secs_to_jiffies.cocci with
the following Coccinelle rules:
@depends on patch@
expression E;
@@
-msecs_to_jiffies(E * 1000)
+secs_to_jiffies(E)
-msecs_to_jiffies(E * MSEC_PER_SEC)
+secs_to_jiffies(E)
While here, manually convert a couple timeouts denominated in seconds
Signed-off-by: Easwar Hariharan <eahariha@linux.microsoft.com>
Link: https://patch.msgid.link/20250707-netdev-secs-to-jiffies-part-2-v2-2-b7817036342f@linux.microsoft.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Commit b35108a51cf7 ("jiffies: Define secs_to_jiffies()") introduced
secs_to_jiffies(). As the value here is a multiple of 1000, use
secs_to_jiffies() instead of msecs_to_jiffies to avoid the multiplication.
This is converted using scripts/coccinelle/misc/secs_to_jiffies.cocci with
the following Coccinelle rules:
@depends on patch@
expression E;
@@
-msecs_to_jiffies(E * 1000)
+secs_to_jiffies(E)
-msecs_to_jiffies(E * MSEC_PER_SEC)
+secs_to_jiffies(E)
Signed-off-by: Easwar Hariharan <eahariha@linux.microsoft.com>
Reviewed-by: Dust Li <dust.li@linux.alibaba.com>
Reviewed-by: Guangguan Wang <guangguan.wang@linux.alibaba.com>
Link: https://patch.msgid.link/20250707-netdev-secs-to-jiffies-part-2-v2-1-b7817036342f@linux.microsoft.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Eric Dumazet says:
====================
tcp: better memory control for not-yet-accepted sockets
Address a possible OOM condition caused by a recent change.
Add a new packetdrill test checking the expected behavior.
====================
Link: https://patch.msgid.link/20250707213900.1543248-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Test how new passive flows react to ooo incoming packets.
Their sk_rcvbuf can increase only after accept().
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250707213900.1543248-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
When a passive flow has not been accepted yet, it is
not wise to increase sk_rcvbuf when receiving ooo packets.
A very busy server might tune down tcp_rmem[1] to better
control how much memory can be used by sockets waiting
in its listeners accept queues.
Fixes: 63ad7dfedfae ("tcp: adjust rcvbuf in presence of reorders")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Link: https://patch.msgid.link/20250707213900.1543248-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Lion's patch [1] revealed an ancient bug in the qdisc API.
Whenever a user creates/modifies a qdisc specifying as a parent another
qdisc, the qdisc API will, during grafting, detect that the user is
not trying to attach to a class and reject. However grafting is
performed after qdisc_create (and thus the qdiscs' init callback) is
executed. In qdiscs that eventually call qdisc_tree_reduce_backlog
during init or change (such as fq, hhf, choke, etc), an issue
arises. For example, executing the following commands:
sudo tc qdisc add dev lo root handle a: htb default 2
sudo tc qdisc add dev lo parent a: handle beef fq
Qdiscs such as fq, hhf, choke, etc unconditionally invoke
qdisc_tree_reduce_backlog() in their control path init() or change() which
then causes a failure to find the child class; however, that does not stop
the unconditional invocation of the assumed child qdisc's qlen_notify with
a null class. All these qdiscs make the assumption that class is non-null.
The solution is ensure that qdisc_leaf() which looks up the parent
class, and is invoked prior to qdisc_create(), should return failure on
not finding the class.
In this patch, we leverage qdisc_leaf to return ERR_PTRs whenever the
parentid doesn't correspond to a class, so that we can detect it
earlier on and abort before qdisc_create is called.
[1] https://lore.kernel.org/netdev/d912cbd7-193b-4269-9857-525bee8bbb6a@gmail.com/
Fixes: 5e50da01d0ce ("[NET_SCHED]: Fix endless loops (part 2): "simple" qdiscs")
Reported-by: syzbot+d8b58d7b0ad89a678a16@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/68663c93.a70a0220.5d25f.0857.GAE@google.com/
Reported-by: syzbot+5eccb463fa89309d8bdc@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/68663c94.a70a0220.5d25f.0858.GAE@google.com/
Reported-by: syzbot+1261670bbdefc5485a06@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/686764a5.a00a0220.c7b3.0013.GAE@google.com/
Reported-by: syzbot+15b96fc3aac35468fe77@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/686764a5.a00a0220.c7b3.0014.GAE@google.com/
Reported-by: syzbot+4dadc5aecf80324d5a51@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/netdev/68679e81.a70a0220.29cf51.0016.GAE@google.com/
Acked-by: Jamal Hadi Salim <jhs@mojatatu.com>
Reviewed-by: Cong Wang <xiyou.wangcong@gmail.com>
Signed-off-by: Victor Nogueira <victor@mojatatu.com>
Link: https://patch.msgid.link/20250707210801.372995-1-victor@mojatatu.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
All memory in GVE is currently allocated without regard for the NUMA
node of the device. Because access to NUMA-local memory access is
significantly cheaper than access to a remote node, this change attempts
to ensure that page frags used in the RX path, including page pool
frags, are allocated on the NUMA node local to the gVNIC device. Note
that this attempt is best-effort. If necessary, the driver will still
allocate non-local memory, as __GFP_THISNODE is not passed. Descriptor
ring allocations are not updated, as dma_alloc_coherent handles that.
This change also modifies the IRQ affinity setting to only select CPUs
from the node local to the device, preserving the behavior that TX and
RX queues of the same index share CPU affinity.
Signed-off-by: Bailey Forrest <bcf@google.com>
Signed-off-by: Joshua Washington <joshwash@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Harshitha Ramamurthy <hramamurthy@google.com>
Signed-off-by: Jeroen de Borst <jeroendb@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://patch.msgid.link/20250707210107.2742029-1-jeroendb@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
skb_shared_info
While transitioning from netdev_alloc_ip_align() to build_skb(), memory
for the "skb_shared_info" member of an "skb" was not allocated. Fix this
by allocating "PAGE_SIZE" as the skb length, accounting for the packet
length, headroom and tailroom, thereby including the required memory space
for skb_shared_info.
Fixes: 8acacc40f733 ("net: ethernet: ti: am65-cpsw: Add minimal XDP support")
Reviewed-by: Siddharth Vadapalli <s-vadapalli@ti.com>
Signed-off-by: Chintan Vankar <c-vankar@ti.com>
Link: https://patch.msgid.link/20250707085201.1898818-1-c-vankar@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
The current logic in nicvf_change_mtu() writes the new MTU to
netdev->mtu using WRITE_ONCE() before verifying if the hardware
update succeeds. However on hardware update failure, it attempts
to revert to the original MTU using a direct assignment
(netdev->mtu = orig_mtu)
which violates the intended of WRITE_ONCE protection introduced in
commit 1eb2cded45b3 ("net: annotate writes on dev->mtu from
ndo_change_mtu()")
Additionally, WRITE_ONCE(netdev->mtu, new_mtu) is unnecessarily
performed even when the device is not running.
Fix this by:
Only writing netdev->mtu after successfully updating the hardware.
Skipping hardware update when the device is down, and setting MTU
directly. Remove unused variable orig_mtu.
This ensures that all writes to netdev->mtu are consistent with
WRITE_ONCE expectations and avoids unintended state corruption
on failure paths.
Signed-off-by: Alok Tiwari <alok.a.tiwari@oracle.com>
Reviewed-by: Jacob Keller <jacob.e.keller@intel.com>
Link: https://patch.msgid.link/20250706194327.1369390-1-alok.a.tiwari@oracle.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
DRR/NETEM/BLACKHOLE chain
Create a tdc test for the UAF scenario with DRR/NETEM/BLACKHOLE chain
shared by Lion on his report [1].
[1] https://lore.kernel.org/netdev/45876f14-cf28-4177-8ead-bb769fd9e57a@gmail.com/
Signed-off-by: Victor Nogueira <victor@mojatatu.com>
Acked-by: Cong Wang <xiyou.wangcong@gmail.com>
Link: https://patch.msgid.link/20250705203638.246350-1-victor@mojatatu.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
atmarpd_dev_ops does not implement the send method, which may cause crash
as bellow.
BUG: kernel NULL pointer dereference, address: 0000000000000000
PGD 0 P4D 0
Oops: Oops: 0010 [#1] SMP KASAN NOPTI
CPU: 0 UID: 0 PID: 5324 Comm: syz.0.0 Not tainted 6.15.0-rc6-syzkaller-00346-g5723cc3450bc #0 PREEMPT(full)
Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS 1.16.3-debian-1.16.3-2~bpo12+1 04/01/2014
RIP: 0010:0x0
Code: Unable to access opcode bytes at 0xffffffffffffffd6.
RSP: 0018:ffffc9000d3cf778 EFLAGS: 00010246
RAX: 1ffffffff1910dd1 RBX: 00000000000000c0 RCX: dffffc0000000000
RDX: ffffc9000dc82000 RSI: ffff88803e4c4640 RDI: ffff888052cd0000
RBP: ffffc9000d3cf8d0 R08: ffff888052c9143f R09: 1ffff1100a592287
R10: dffffc0000000000 R11: 0000000000000000 R12: 1ffff92001a79f00
R13: ffff888052cd0000 R14: ffff88803e4c4640 R15: ffffffff8c886e88
FS: 00007fbc762566c0(0000) GS:ffff88808d6c2000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: ffffffffffffffd6 CR3: 0000000041f1b000 CR4: 0000000000352ef0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
vcc_sendmsg+0xa10/0xc50 net/atm/common.c:644
sock_sendmsg_nosec net/socket.c:712 [inline]
__sock_sendmsg+0x219/0x270 net/socket.c:727
____sys_sendmsg+0x52d/0x830 net/socket.c:2566
___sys_sendmsg+0x21f/0x2a0 net/socket.c:2620
__sys_sendmmsg+0x227/0x430 net/socket.c:2709
__do_sys_sendmmsg net/socket.c:2736 [inline]
__se_sys_sendmmsg net/socket.c:2733 [inline]
__x64_sys_sendmmsg+0xa0/0xc0 net/socket.c:2733
do_syscall_x64 arch/x86/entry/syscall_64.c:63 [inline]
do_syscall_64+0xf6/0x210 arch/x86/entry/syscall_64.c:94
entry_SYSCALL_64_after_hwframe+0x77/0x7f
Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: syzbot+e34e5e6b5eddb0014def@syzkaller.appspotmail.com
Closes: https://lore.kernel.org/all/682f82d5.a70a0220.1765ec.0143.GAE@google.com/T
Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@google.com>
Link: https://patch.msgid.link/20250705085228.329202-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Ivan Vecera says:
====================
Add Microchip ZL3073x support (part 1)
Add support for Microchip Azurite DPLL/PTP/SyncE chip family that
provides DPLL and PTP functionality. This series bring first part
that adds the core functionality and basic DPLL support.
The next part of the series will bring additional DPLL functionality
like eSync support, phase offset and frequency offset reporting and
phase adjustments.
Testing was done by myself and by Prathosh Satish on Microchip EDS2
development board with ZL30732 DPLL chip connected over I2C bus.
====================
Link: https://patch.msgid.link/20250704182202.1641943-1-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add support to get/set frequency on pins. The frequency for input
pins (references) is computed in the device according this formula:
freq = base_freq * multiplier * (nominator / denominator)
where the base_freq comes from the list of supported base frequencies
and other parameters are arbitrary numbers. All these parameters are
16-bit unsigned integers.
The frequency for output pin is determined by the frequency of
synthesizer the output pin is connected to and divisor of the output
to which is the given pin belongs. The resulting frequency of the
P-pin and the N-pin from this output pair depends on the signal
format of this output pair.
The device supports so-called N-divided signal formats where for the
N-pin there is an additional divisor. The frequencies for both pins
from such output pair are computed:
P-pin-freq = synth_freq / output_div
N-pin-freq = synth_freq / output_div / n_div
For other signal-format types both P and N pin have the same frequency
based only synth frequency and output divisor.
Implement output pin callbacks to get and set frequency. The frequency
setting for the output non-N-divided signal format is simple as we have
to compute just new output divisor. For N-divided formats it is more
complex because by changing of output divisor we change frequency for
both P and N pins. In this case if we are changing frequency for P-pin
we have to compute also new N-divisor for N-pin to keep its current
frequency. From this and the above it follows that the frequency of
the N-pin cannot be higher than the frequency of the P-pin and the
callback must take this limitation into account.
Co-developed-by: Prathosh Satish <Prathosh.Satish@microchip.com>
Signed-off-by: Prathosh Satish <Prathosh.Satish@microchip.com>
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20250704182202.1641943-13-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Implement input pin state setting when the DPLL is running in automatic
mode. Unlike manual mode, the DPLL mode switching is not used here and
the implementation uses special priority value (15) to make the given
pin non-selectable.
When the user sets state of the pin as disconnected the driver
internally sets its priority in HW to 15 that prevents the DPLL to
choose this input pin. Conversely, if the pin status is set to
selectable, the driver sets the pin priority in HW to the original saved
value.
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20250704182202.1641943-12-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add support for getting and setting input pin priority. Implement
required callbacks and set appropriate capability for input pins.
Although the pin priority make sense only if the DPLL is running in
automatic mode we have to expose this capability unconditionally because
input pins (references) are shared between all DPLLs where one of them
can run in automatic mode while the other one not.
Co-developed-by: Prathosh Satish <Prathosh.Satish@microchip.com>
Signed-off-by: Prathosh Satish <Prathosh.Satish@microchip.com>
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20250704182202.1641943-11-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Implement input pin state setting if the DPLL is running in manual mode.
The driver indicates manual mode if the DPLL mode is one of ref-lock,
forced-holdover, freerun.
Use these modes to implement input pin state change between connected
and disconnected states. When the user set the particular pin as
connected the driver marks this input pin as forced reference and
switches the DPLL mode to ref-lock. When the use set the pin as
disconnected the driver switches the DPLL to freerun or forced holdover
mode. The switch to holdover mode is done if the DPLL has holdover
capability (e.g is currently locked with holdover acquired).
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20250704182202.1641943-10-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Enumerate all available DPLL channels and registers a DPLL device for
each of them. Check all input references and outputs and register
DPLL pins for them.
Number of registered DPLL pins depends on configuration of references
and outputs. If the reference or output is configured as differential
one then only one DPLL pin is registered. Both references and outputs
can be also disabled from firmware configuration and in this case
no DPLL pins are registered.
All registrable references are registered to all available DPLL devices
with exception of DPLLs that are configured in NCO (numerically
controlled oscillator) mode. In this mode DPLL channel acts as PHC and
cannot be locked to any reference.
Device outputs are connected to one of synthesizers and each synthesizer
is driven by some DPLL channel. So output pins belonging to given output
are registered to DPLL device that drives associated synthesizer.
Finally add kworker task to monitor async changes on all DPLL channels
and input pins and to notify about them DPLL core. Output pins are not
monitored as their parameters are not changed asynchronously by the
device.
Co-developed-by: Prathosh Satish <Prathosh.Satish@microchip.com>
Signed-off-by: Prathosh Satish <Prathosh.Satish@microchip.com>
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20250704182202.1641943-9-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Add support for reading of DPLL types and optional pin properties from
the system firmware (DT, ACPI...).
The DPLL types are stored in property 'dpll-types' as string array and
possible values 'pps' and 'eec' are mapped to DPLL enums DPLL_TYPE_PPS
and DPLL_TYPE_EEC.
The pin properties are stored under 'input-pins' and 'output-pins'
sub-nodes and the following ones are supported:
* reg
integer that specifies pin index
* label
string that is used by driver as board label
* connection-type
string that indicates pin connection type
* supported-frequencies-hz
array of u64 values what frequencies are supported / allowed for
given pin with respect to hardware wiring
Do not blindly trust system firmware and filter out frequencies that
cannot be configured/represented in device (input frequencies have to
be factorized by one of the base frequencies and output frequencies have
to divide configured synthesizer frequency).
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20250704182202.1641943-8-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
|
Several configuration parameters will remain constant at runtime,
so we can load them during probe to avoid excessive reads from
the hardware.
Read the following parameters from the device during probe and store
them for later use:
* enablement status and frequencies of the synthesizers and their
associated DPLL channels
* enablement status and type (single-ended or differential) of input pins
* associated synthesizers, signal format, and enablement status of
outputs
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://patch.msgid.link/20250704182202.1641943-7-ivecera@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
|
