platform/kernel/linux-rpi.git
11 months agopage_pool: add a lockdep check for recycling in hardirq
Jakub Kicinski [Fri, 4 Aug 2023 18:05:28 +0000 (20:05 +0200)]
page_pool: add a lockdep check for recycling in hardirq

Page pool use in hardirq is prohibited, add debug checks
to catch misuses. IIRC we previously discussed using
DEBUG_NET_WARN_ON_ONCE() for this, but there were concerns
that people will have DEBUG_NET enabled in perf testing.
I don't think anyone enables lockdep in perf testing,
so use lockdep to avoid pushback and arguing :)

Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Link: https://lore.kernel.org/r/20230804180529.2483231-6-aleksander.lobakin@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: skbuff: avoid accessing page_pool if !napi_safe when returning page
Alexander Lobakin [Fri, 4 Aug 2023 18:05:27 +0000 (20:05 +0200)]
net: skbuff: avoid accessing page_pool if !napi_safe when returning page

Currently, pp->p.napi is always read, but the actual variable it gets
assigned to is read-only when @napi_safe is true. For the !napi_safe
cases, which yet is still a pack, it's an unneeded operation.
Moreover, it can lead to premature or even redundant page_pool
cacheline access. For example, when page_pool_is_last_frag() returns
false (with the recent frag improvements).
Thus, read it only when @napi_safe is true. This also allows moving
@napi inside the condition block itself. Constify it while we are
here, because why not.

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Link: https://lore.kernel.org/r/20230804180529.2483231-5-aleksander.lobakin@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agopage_pool: place frag_* fields in one cacheline
Alexander Lobakin [Fri, 4 Aug 2023 18:05:26 +0000 (20:05 +0200)]
page_pool: place frag_* fields in one cacheline

On x86_64, frag_* fields of struct page_pool are scattered across two
cachelines despite the summary size of 24 bytes. All three fields are
used in pretty much the same places, but the last field, ::frag_users,
is pushed out to the next CL, provoking unwanted false-sharing on
hotpath (frags allocation code).
There are some holes and cold members to move around. Move frag_* one
block up, placing them right after &page_pool_params perfectly at the
beginning of CL2. This doesn't do any meaningful to the second block, as
those are some destroy-path cold structures, and doesn't do anything to
::alloc_stats, which still starts at 200-byte offset, 8 bytes after CL3
(still fitting into 1 cacheline).
On my setup, this yields 1-2% of Mpps when using PP frags actively.
When it comes to 32-bit architectures with 32-byte CL: &page_pool_params
plus ::pad is 44 bytes, the block taken care of is 16 bytes within one
CL, so there should be at least no regressions from the actual change.
::pages_state_hold_cnt is not related directly to that triple, but is
paired currently with ::frags_offset and decoupling them would mean
either two 4-byte holes or more invasive layout changes.

Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Link: https://lore.kernel.org/r/20230804180529.2483231-4-aleksander.lobakin@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: skbuff: don't include <net/page_pool/types.h> to <linux/skbuff.h>
Alexander Lobakin [Fri, 4 Aug 2023 18:05:25 +0000 (20:05 +0200)]
net: skbuff: don't include <net/page_pool/types.h> to <linux/skbuff.h>

Currently, touching <net/page_pool/types.h> triggers a rebuild of more
than half of the kernel. That's because it's included in
<linux/skbuff.h>. And each new include to page_pool/types.h adds more
[useless] data for the toolchain to process per each source file from
that pile.

In commit 6a5bcd84e886 ("page_pool: Allow drivers to hint on SKB
recycling"), Matteo included it to be able to call a couple of functions
defined there. Then, in commit 57f05bc2ab24 ("page_pool: keep pp info as
long as page pool owns the page") one of the calls was removed, so only
one was left. It's the call to page_pool_return_skb_page() in
napi_frag_unref(). The function is external and doesn't have any
dependencies. Having very niche page_pool_types.h included only for that
looks like an overkill.

As %PP_SIGNATURE is not local to page_pool.c (was only in the
early submissions), nothing holds this function there. Teleport
page_pool_return_skb_page() to skbuff.c, just next to the main consumer,
skb_pp_recycle(), and rename it to napi_pp_put_page(), as it doesn't
work with skbs at all and the former name tells nothing. The #if guards
here are only to not compile and have it in the vmlinux when not needed
-- both call sites are already guarded.
Now, touching page_pool_types.h only triggers rebuilding of the drivers
using it and a couple of core networking files.

Suggested-by: Jakub Kicinski <kuba@kernel.org> # make skbuff.h less heavy
Suggested-by: Alexander Duyck <alexanderduyck@fb.com> # move to skbuff.c
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Link: https://lore.kernel.org/r/20230804180529.2483231-3-aleksander.lobakin@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agopage_pool: split types and declarations from page_pool.h
Yunsheng Lin [Fri, 4 Aug 2023 18:05:24 +0000 (20:05 +0200)]
page_pool: split types and declarations from page_pool.h

Split types and pure function declarations from page_pool.h
and add them in page_page/types.h, so that C sources can
include page_pool.h and headers should generally only include
page_pool/types.h as suggested by jakub.
Rename page_pool.h to page_pool/helpers.h to have both in
one place.

Signed-off-by: Yunsheng Lin <linyunsheng@huawei.com>
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Alexander Lobakin <aleksander.lobakin@intel.com>
Reviewed-by: Alexander Duyck <alexanderduyck@fb.com>
Link: https://lore.kernel.org/r/20230804180529.2483231-2-aleksander.lobakin@intel.com
[Jakub: change microsoft/mana, fix kdoc paths in Documentation]
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge tag 'linux-can-next-for-6.6-20230807' of git://git.kernel.org/pub/scm/linux...
Jakub Kicinski [Mon, 7 Aug 2023 19:25:46 +0000 (12:25 -0700)]
Merge tag 'linux-can-next-for-6.6-20230807' of git://git./linux/kernel/git/mkl/linux-can-next

Marc Kleine-Budde says:

====================
pull-request: can-next 2023-08-07

The patch is from me and reverts the addition of the CAN controller
nodes in the allwinner d1 SoC.

* tag 'linux-can-next-for-6.6-20230807' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can-next:
  Revert "riscv: dts: allwinner: d1: Add CAN controller nodes"
====================

Link: https://lore.kernel.org/r/20230807074222.1576119-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge branch 'net-stmmac-correct-mac-propagation-delay'
Jakub Kicinski [Mon, 7 Aug 2023 19:17:14 +0000 (12:17 -0700)]
Merge branch 'net-stmmac-correct-mac-propagation-delay'

Johannes Zink says:

====================
net: stmmac: correct MAC propagation delay

Changes in v3:
- work in Richard's review feedback. Thank you for reviewing my patch:
  - as some of the hardware may have no or invalid correction value
    registers: introduce feature switch which can be enabled in the glue
    code drivers depending on the actual hardware support
  - only enable the feature on the i.MX8MP for the time being, as the patch
    improves timing accuracy and is tested for this hardware
- Link to v2: https://lore.kernel.org/r/20230719-stmmac_correct_mac_delay-v2-1-3366f38ee9a6@pengutronix.de

Changes in v2:
- fix builds for 32bit, this was found by the kernel build bot
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202307200225.B8rmKQPN-lkp@intel.com/
- while at it also fix an overflow by shifting a u32 constant from macro by 10bits
  by casting the constant to u64
- Link to v1: https://lore.kernel.org/r/20230719-stmmac_correct_mac_delay-v1-1-768aa4d09334@pengutronix.de

Tested-by: Kurt Kanzenbach <kurt@linutronix.de> # imx8mp
====================

Link: https://lore.kernel.org/r/20230719-stmmac_correct_mac_delay-v3-0-61e63427735e@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: stmmac: dwmac-imx: enable MAC propagation delay correction for i.MX8MP
Johannes Zink [Tue, 1 Aug 2023 15:44:30 +0000 (17:44 +0200)]
net: stmmac: dwmac-imx: enable MAC propagation delay correction for i.MX8MP

As the i.MX8MP supports reading MAC propagation delay and correcting the
Hardware timestamp counter for additional delays [1], enable the feature
for this SoC.

This reduces phase error of the PPS output from the PTP Hardware Clock
from approx 150ns to 100ns.

[1] i.MX8MP Reference Manual, rev.1 Section 11.7.2.5.3 "Timestamp
correction"

Signed-off-by: Johannes Zink <j.zink@pengutronix.de>
Link: https://lore.kernel.org/r/20230719-stmmac_correct_mac_delay-v3-2-61e63427735e@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: stmmac: correct MAC propagation delay
Johannes Zink [Tue, 1 Aug 2023 15:44:29 +0000 (17:44 +0200)]
net: stmmac: correct MAC propagation delay

The IEEE1588 Standard specifies that the timestamps of Packets must be
captured when the PTP message timestamp point (leading edge of first
octet after the start of frame delimiter) crosses the boundary between
the node and the network. As the MAC latches the timestamp at an
internal point, the captured timestamp must be corrected for the
additional data transmission latency, as described in the publicly
available datasheet [1].

This patch only corrects for the MAC-Internal delay, which can be read
out from the MAC_Ingress_Timestamp_Latency register on DWMAC version 5,
since the Phy framework currently does not support querying the Phy
ingress and egress latency. The Closs Domain Crossing Circuits errors as
indicated in [1] are already being accounted in the
stmmac_get_tx_hwtstamp() function and are not corrected here.

As the Latency varies for different link speeds and MII
modes of operation, the correction value needs to be updated on each
link state change.

As the delay also causes a phase shift in the timestamp counter compared
to the rest of the network, this correction will also reduce phase error
when generating PPS outputs from the timestamp counter.

Since the correction registers may be unavailable on some hardware and
no feature bits are documented for dynamically detection of the MAC
propagation delay readout, introduce a feature bit to explicitely enable
MAC delay Correction in the gluecode driver.

[1] i.MX8MP Reference Manual, rev.1 Section 11.7.2.5.3 "Timestamp
correction"

Signed-off-by: Johannes Zink <j.zink@pengutronix.de>
Link: https://lore.kernel.org/r/20230719-stmmac_correct_mac_delay-v2-1-3366f38ee9a6@pengutronix.de
Link: https://lore.kernel.org/r/20230719-stmmac_correct_mac_delay-v3-1-61e63427735e@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoudp/udplite: Remove unused function declarations udp{,lite}_get_port()
Yue Haibing [Sat, 5 Aug 2023 11:00:09 +0000 (19:00 +0800)]
udp/udplite: Remove unused function declarations udp{,lite}_get_port()

Commit 6ba5a3c52da0 ("[UDP]: Make full use of proto.h.udp_hash innovation.")
removed these implementations but leave declarations.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: sfp: Remove unused function declaration sfp_link_configure()
Yue Haibing [Sat, 5 Aug 2023 10:57:40 +0000 (18:57 +0800)]
net: sfp: Remove unused function declaration sfp_link_configure()

Commit ce0aa27ff3f6 ("sfp: add sfp-bus to bridge between network devices and sfp cages")
declared but never implemented it.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agondisc: Remove unused ndisc_ifinfo_sysctl_strategy() declaration
Yue Haibing [Sat, 5 Aug 2023 10:53:54 +0000 (18:53 +0800)]
ndisc: Remove unused ndisc_ifinfo_sysctl_strategy() declaration

Commit f8572d8f2a2b ("sysctl net: Remove unused binary sysctl code")
left behind this declaration.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: pkt_cls: Remove unused inline helpers
Yue Haibing [Sat, 5 Aug 2023 10:52:08 +0000 (18:52 +0800)]
net: pkt_cls: Remove unused inline helpers

Commit acb674428c3d ("net: sched: introduce per-block callbacks")
implemented these but never used it.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agoneighbour: Remove unused function declaration pneigh_for_each()
Yue Haibing [Sat, 5 Aug 2023 10:50:33 +0000 (18:50 +0800)]
neighbour: Remove unused function declaration pneigh_for_each()

pneigh_for_each() is never implemented since the beginning of git history.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet/tls: Remove unused function declarations
Yue Haibing [Sat, 5 Aug 2023 10:48:11 +0000 (18:48 +0800)]
net/tls: Remove unused function declarations

Commit 3c4d7559159b ("tls: kernel TLS support") declared but never implemented
these functions.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agoRevert "riscv: dts: allwinner: d1: Add CAN controller nodes"
Marc Kleine-Budde [Mon, 7 Aug 2023 07:28:50 +0000 (09:28 +0200)]
Revert "riscv: dts: allwinner: d1: Add CAN controller nodes"

It turned out the dtsi changes were not quite ready, revert them for
now.

This reverts commit 6ea1ad888f5900953a21853e709fa499fdfcb317.

Link: https://lore.kernel.org/all/2690764.mvXUDI8C0e@jernej-laptop
Suggested-by: Jernej Škrabec <jernej.skrabec@gmail.com>
Link: https://lore.kernel.org/all/20230807-riscv-allwinner-d1-revert-can-controller-nodes-v1-1-eb3f70b435d9@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
11 months agonet: omit ndo_hwtstamp_get() call when possible in dev_set_hwtstamp_phylib()
Vladimir Oltean [Fri, 4 Aug 2023 13:49:39 +0000 (16:49 +0300)]
net: omit ndo_hwtstamp_get() call when possible in dev_set_hwtstamp_phylib()

Setting dev->priv_flags & IFF_SEE_ALL_HWTSTAMP_REQUESTS is only legal
for drivers which were converted to ndo_hwtstamp_get() and
ndo_hwtstamp_set(), and it is only there that we call ndo_hwtstamp_set()
for a request that otherwise goes to phylib (for stuff like packet traps,
which need to be undone if phylib failed, hence the old_cfg logic).

The problem is that we end up calling ndo_hwtstamp_get() when we don't
need to (even if the SIOCSHWTSTAMP wasn't intended for phylib, or if it
was, but the driver didn't set IFF_SEE_ALL_HWTSTAMP_REQUESTS). For those
unnecessary conditions, we share a code path with virtual drivers (vlan,
macvlan, bonding) where ndo_hwtstamp_get() is implemented as
generic_hwtstamp_get_lower(), and may be resolved through
generic_hwtstamp_ioctl_lower() if the lower device is unconverted.

I.e. this situation:

$ ip link add link eno0 name eno0.100 type vlan id 100
$ hwstamp_ctl -i eno0.100 -t 1

We are unprepared to deal with this, because if ndo_hwtstamp_get() is
resolved through a legacy ndo_eth_ioctl(SIOCGHWTSTAMP) lower_dev
implementation, that needs a non-NULL old_cfg.ifr pointer, and we don't
have it.

But we don't even need to deal with it either. In the general case,
drivers may not even implement SIOCGHWTSTAMP handling, only SIOCSHWTSTAMP,
so it makes sense to completely avoid a SIOCGHWTSTAMP call if we can.

The solution is to split the single "if" condition into 3 smaller ones,
thus separating the decision to call ndo_hwtstamp_get() from the
decision to call ndo_hwtstamp_set(). The third "if" condition is
identical to the first one, and both are subsets of the second one.
Thus, the "cfg" argument of kernel_hwtstamp_config_changed() is always
valid.

Reported-by: Eric Dumazet <edumazet@google.com>
Closes: https://lore.kernel.org/netdev/CANn89iLOspJsvjPj+y8jikg7erXDomWe8sqHMdfL_2LQSFrPAg@mail.gmail.com/
Fixes: fd770e856e22 ("net: remove phy_has_hwtstamp() -> phy_mii_ioctl() decision from converted drivers")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: ethernet: adi: adin1110: use eth_broadcast_addr() to assign broadcast address
Yang Yingliang [Fri, 4 Aug 2023 09:35:31 +0000 (17:35 +0800)]
net: ethernet: adi: adin1110: use eth_broadcast_addr() to assign broadcast address

Use eth_broadcast_addr() to assign broadcast address instead
of memset().

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agoibmvnic: remove unused rc variable
Yu Liao [Fri, 4 Aug 2023 09:21:43 +0000 (17:21 +0800)]
ibmvnic: remove unused rc variable

gcc with W=1 reports
drivers/net/ethernet/ibm/ibmvnic.c:194:13: warning: variable 'rc' set but not used [-Wunused-but-set-variable]
                                            ^
This variable is not used so remove it.

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202308040609.zQsSXWXI-lkp@intel.com/
Signed-off-by: Yu Liao <liaoyu15@huawei.com>
Reviewed-by: Nick Child <nnac123@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: mana: Add page pool for RX buffers
Haiyang Zhang [Fri, 4 Aug 2023 20:33:53 +0000 (13:33 -0700)]
net: mana: Add page pool for RX buffers

Add page pool for RX buffers for faster buffer cycle and reduce CPU
usage.

The standard page pool API is used.

With iperf and 128 threads test, this patch improved the throughput
by 12-15%, and decreased the IRQ associated CPU's usage from 99-100% to
10-50%.

Signed-off-by: Haiyang Zhang <haiyangz@microsoft.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agoMerge branch 'gve-desc'
David S. Miller [Sun, 6 Aug 2023 07:34:37 +0000 (08:34 +0100)]
Merge branch 'gve-desc'

Rushil Gupta says:

====================
gve: Add QPL mode for DQO descriptor format

GVE supports QPL ("queue-page-list") mode where
all data is communicated through a set of pre-registered
pages. Adding this mode to DQO.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agogve: update gve.rst
Rushil Gupta [Fri, 4 Aug 2023 21:34:44 +0000 (21:34 +0000)]
gve: update gve.rst

Add a note about QPL and RDA mode

Signed-off-by: Rushil Gupta <rushilg@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Praveen Kaligineedi <pkaligineedi@google.com>
Signed-off-by: Bailey Forrest <bcf@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agogve: RX path for DQO-QPL
Rushil Gupta [Fri, 4 Aug 2023 21:34:43 +0000 (21:34 +0000)]
gve: RX path for DQO-QPL

The RX path allocates the QPL page pool at queue creation, and
tries to reuse these pages through page recycling. This patch
ensures that on refill no non-QPL pages are posted to the device.

When the driver is running low on free buffers, an ondemand
allocation step kicks in that allocates a non-qpl page for
SKB business to free up the QPL page in use.

gve_try_recycle_buf was moved to gve_rx_append_frags so that driver does
not attempt to mark buffer as used if a non-qpl page was allocated
ondemand.

Signed-off-by: Rushil Gupta <rushilg@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Praveen Kaligineedi <pkaligineedi@google.com>
Signed-off-by: Bailey Forrest <bcf@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agogve: Tx path for DQO-QPL
Rushil Gupta [Fri, 4 Aug 2023 21:34:42 +0000 (21:34 +0000)]
gve: Tx path for DQO-QPL

Each QPL page is divided into GVE_TX_BUFS_PER_PAGE_DQO buffers.
When a packet needs to be transmitted, we break the packet into max
GVE_TX_BUF_SIZE_DQO sized chunks and transmit each chunk using a TX
descriptor.
We allocate the TX buffers from the free list in dqo_tx.
We store these TX buffer indices in an array in the pending_packet
structure.

The TX buffers are returned to the free list in dqo_compl after
receiving packet completion or when removing packets from miss
completions list.

Signed-off-by: Rushil Gupta <rushilg@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Praveen Kaligineedi <pkaligineedi@google.com>
Signed-off-by: Bailey Forrest <bcf@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agogve: Control path for DQO-QPL
Rushil Gupta [Fri, 4 Aug 2023 21:34:41 +0000 (21:34 +0000)]
gve: Control path for DQO-QPL

GVE supports QPL ("queue-page-list") mode where
all data is communicated through a set of pre-registered
pages. Adding this mode to DQO descriptor format.

Add checks, abi-changes and device options to support
QPL mode for DQO in addition to GQI. Also, use
pages-per-qpl supplied by device-option to control the
size of the "queue-page-list".

Signed-off-by: Rushil Gupta <rushilg@google.com>
Reviewed-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Praveen Kaligineedi <pkaligineedi@google.com>
Signed-off-by: Bailey Forrest <bcf@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agoMerge branch 'tcp-options-lockless'
David S. Miller [Sun, 6 Aug 2023 07:24:56 +0000 (08:24 +0100)]
Merge branch 'tcp-options-lockless'

Eric Dumazet says:

====================
tcp: set few options locklessly

This series is avoiding the socket lock for six TCP options.

They are not heavily used, but this exercise can give
ideas for other parts of TCP/IP stack :)
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agotcp: set TCP_DEFER_ACCEPT locklessly
Eric Dumazet [Fri, 4 Aug 2023 14:46:16 +0000 (14:46 +0000)]
tcp: set TCP_DEFER_ACCEPT locklessly

rskq_defer_accept field can be read/written without
the need of holding the socket lock.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agotcp: set TCP_LINGER2 locklessly
Eric Dumazet [Fri, 4 Aug 2023 14:46:15 +0000 (14:46 +0000)]
tcp: set TCP_LINGER2 locklessly

tp->linger2 can be set locklessly as long as readers
use READ_ONCE().

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agotcp: set TCP_KEEPCNT locklessly
Eric Dumazet [Fri, 4 Aug 2023 14:46:14 +0000 (14:46 +0000)]
tcp: set TCP_KEEPCNT locklessly

tp->keepalive_probes can be set locklessly, readers
are already taking care of this field being potentially
set by other threads.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agotcp: set TCP_KEEPINTVL locklessly
Eric Dumazet [Fri, 4 Aug 2023 14:46:13 +0000 (14:46 +0000)]
tcp: set TCP_KEEPINTVL locklessly

tp->keepalive_intvl can be set locklessly, readers
are already taking care of this field being potentially
set by other threads.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agotcp: set TCP_USER_TIMEOUT locklessly
Eric Dumazet [Fri, 4 Aug 2023 14:46:12 +0000 (14:46 +0000)]
tcp: set TCP_USER_TIMEOUT locklessly

icsk->icsk_user_timeout can be set locklessly,
if all read sides use READ_ONCE().

Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Soheil Hassas Yeganeh <soheil@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agotcp: set TCP_SYNCNT locklessly
Eric Dumazet [Fri, 4 Aug 2023 14:46:11 +0000 (14:46 +0000)]
tcp: set TCP_SYNCNT locklessly

icsk->icsk_syn_retries can safely be set without locking the socket.

We have to add READ_ONCE() annotations in tcp_fastopen_synack_timer()
and tcp_write_timeout().

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agoMerge tag 'wireless-next-2023-08-04' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Sat, 5 Aug 2023 01:34:25 +0000 (18:34 -0700)]
Merge tag 'wireless-next-2023-08-04' of git://git./linux/kernel/git/wireless/wireless-next

Kalle Valo says:

====================
wireless-next patches for v6.6

The first pull request for v6.6 and only driver patches this time.
Nothing special really standing out, it has been quiet most likely due
to vacations.

Major changes:

rtl8xxxu
 - enable AP mode for: RTL8192FU, RTL8710BU (RTL8188GU),
   RTL8192EU and RTL8723BU

mwifiex
 - allow moving to a different namespace

mt76
 - preparation for mt7925 support
 - mt7981 support

ath12k
 - Extremely High Throughput (EHT) PHY support for Wi-Fi 7

* tag 'wireless-next-2023-08-04' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless-next: (172 commits)
  wifi: rtw89: return failure if needed firmware elements are not recognized
  wifi: rtw89: add to parse firmware elements of BB and RF tables
  wifi: rtw89: introduce infrastructure of firmware elements
  wifi: rtw89: add firmware suit for BB MCU 0/1
  wifi: rtw89: add firmware parser for v1 format
  wifi: rtw89: introduce v1 format of firmware header
  wifi: rtw89: support firmware log with formatted text
  wifi: rtw89: recognize log format from firmware file
  wifi: ath12k: avoid deadlock by change ieee80211_queue_work for regd_update_work
  wifi: ath12k: add handler for scan event WMI_SCAN_EVENT_DEQUEUED
  wifi: ath12k: relax list iteration in ath12k_mac_vif_unref()
  wifi: ath12k: configure puncturing bitmap
  wifi: ath12k: parse WMI service ready ext2 event
  wifi: ath12k: add MLO header in peer association
  wifi: ath12k: peer assoc for 320 MHz
  wifi: ath12k: add WMI support for EHT peer
  wifi: ath12k: prepare EHT peer assoc parameters
  wifi: ath12k: add EHT PHY modes
  wifi: ath12k: propagate EHT capabilities to userspace
  wifi: ath12k: WMI support to process EHT capabilities
  ...
====================

Link: https://lore.kernel.org/r/87msz7j942.fsf@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge branch 'tcp-disable-header-prediction-for-md5'
Jakub Kicinski [Sat, 5 Aug 2023 01:28:38 +0000 (18:28 -0700)]
Merge branch 'tcp-disable-header-prediction-for-md5'

Kuniyuki Iwashima says:

====================
tcp: Disable header prediction for MD5.

The 1st patch disable header prediction for MD5 flow and the 2nd
patch updates the stale comment in tcp_parse_options().
====================

Link: https://lore.kernel.org/r/20230803224552.69398-1-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp: Update stale comment for MD5 in tcp_parse_options().
Kuniyuki Iwashima [Thu, 3 Aug 2023 22:45:52 +0000 (15:45 -0700)]
tcp: Update stale comment for MD5 in tcp_parse_options().

Since commit 9ea88a153001 ("tcp: md5: check md5 signature without socket
lock"), the MD5 option is checked in tcp_v[46]_rcv().

Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230803224552.69398-3-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp: Disable header prediction for MD5 flow.
Kuniyuki Iwashima [Thu, 3 Aug 2023 22:45:51 +0000 (15:45 -0700)]
tcp: Disable header prediction for MD5 flow.

TCP socket saves the minimum required header length in tcp_header_len
of struct tcp_sock, and later the value is used in __tcp_fast_path_on()
to generate a part of TCP header in tcp_sock(sk)->pred_flags.

In tcp_rcv_established(), if the incoming packet has the same pattern
with pred_flags, we enter the fast path and skip full option parsing.

The MD5 option is parsed in tcp_v[46]_rcv(), so we need not parse it
again later in tcp_rcv_established() unless other options exist.  We
add TCPOLEN_MD5SIG_ALIGNED to tcp_header_len in two paths to avoid the
slow path.

For passive open connections with MD5, we add TCPOLEN_MD5SIG_ALIGNED
to tcp_header_len in tcp_create_openreq_child() after 3WHS.

On the other hand, we do it in tcp_connect_init() for active open
connections.  However, the value is overwritten while processing
SYN+ACK or crossed SYN in tcp_rcv_synsent_state_process().

These two cases will have the wrong value in pred_flags and never go
into the fast path.

We could update tcp_header_len in tcp_rcv_synsent_state_process(), but
a test with slightly modified netperf which uses MD5 for each flow shows
that the slow path is actually a bit faster than the fast path.

  On c5.4xlarge EC2 instance (16 vCPU, 32 GiB mem)

  $ for i in {1..10}; do
  ./super_netperf $(nproc) -H localhost -l 10 -- -m 256 -M 256;
  done

  Avg of 10
  * 36e68eadd303  : 10.376 Gbps
  * all fast path : 10.374 Gbps (patch v2, See Link)
  * all slow path : 10.394 Gbps

The header prediction is not worth adding complexity for MD5, so let's
disable it for MD5.

Link: https://lore.kernel.org/netdev/20230803042214.38309-1-kuniyu@amazon.com/
Signed-off-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20230803224552.69398-2-kuniyu@amazon.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: phy: move marking PHY on SFP module into SFP code
Russell King (Oracle) [Thu, 3 Aug 2023 15:56:24 +0000 (16:56 +0100)]
net: phy: move marking PHY on SFP module into SFP code

Move marking the PHY as being on a SFP module into the SFP code between
getting the PHY device (and thus initialising the phy_device structure)
and registering the discovered device.

This means that PHY drivers can use phy_on_sfp() in their match and
get_features methods.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Link: https://lore.kernel.org/r/E1qRaga-001vKt-8X@rmk-PC.armlinux.org.uk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agomlxsw: spectrum: Remove unused function declarations
Yue Haibing [Thu, 3 Aug 2023 14:20:47 +0000 (22:20 +0800)]
mlxsw: spectrum: Remove unused function declarations

Commit c3d2ed93b14d ("mlxsw: Remove old parsing depth infrastructure")
left behind mlxsw_sp_nve_inc_parsing_depth_get()/mlxsw_sp_nve_inc_parsing_depth_put().
And commit 532b49e41e64 ("mlxsw: spectrum_span: Derive SBIB from maximum port speed & MTU")
remove mlxsw_sp_span_port_mtu_update()/mlxsw_sp_span_speed_update_work() but leave the
declarations.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Ido Schimmel <idosch@nvidia.com>
Link: https://lore.kernel.org/r/20230803142047.42660-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoixgbevf: Remove unused function declarations
Yue Haibing [Thu, 3 Aug 2023 14:19:04 +0000 (22:19 +0800)]
ixgbevf: Remove unused function declarations

ixgbe_napi_add_all()/ixgbe_napi_del_all() are declared but never implemented in
commit 92915f71201b ("ixgbevf: Driver main and ethool interface module and main header")

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230803141904.15316-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoaf_vsock: Remove unused declaration vsock_release_pending()/vsock_init_tap()
Yue Haibing [Thu, 3 Aug 2023 13:45:07 +0000 (21:45 +0800)]
af_vsock: Remove unused declaration vsock_release_pending()/vsock_init_tap()

Commit d021c344051a ("VSOCK: Introduce VM Sockets") declared but never implemented
vsock_release_pending(). Also vsock_init_tap() never implemented since introduction
in commit 531b374834c8 ("VSOCK: Add vsockmon tap functions").

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20230803134507.22660-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: 802: Remove unused function declarations
Yue Haibing [Thu, 3 Aug 2023 13:54:24 +0000 (21:54 +0800)]
net: 802: Remove unused function declarations

Commit d8d9ba8dc9c7 ("net: 802: remove dead leftover after ipx driver removal")
remove these implementations but leave the declarations.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230803135424.41664-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp_metrics: hash table allocation cleanup
Eric Dumazet [Thu, 3 Aug 2023 13:54:16 +0000 (13:54 +0000)]
tcp_metrics: hash table allocation cleanup

After commit 098a697b497e ("tcp_metrics: Use a single hash table
for all network namespaces.") we can avoid calling tcp_net_metrics_init()
for each new netns.

Instead, rename tcp_net_metrics_init() to tcp_metrics_hash_alloc(),
and move it to __init section.

Also move tcpmhash_entries to __initdata section.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230803135417.2716879-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: hns3: Remove unused function declarations
Yue Haibing [Thu, 3 Aug 2023 13:51:38 +0000 (21:51 +0800)]
net: hns3: Remove unused function declarations

Commit 1e6e76101fd9 ("net: hns3: configure promisc mode for VF asynchronously")
left behind hclge_inform_vf_promisc_info() declaration.
And commit 68c0a5c70614 ("net: hns3: Add HNS3 IMP(Integrated Mgmt Proc) Cmd Interface Support")
declared but never implemented hclge_cmd_mdio_write() and hclge_cmd_mdio_read().

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230803135138.37456-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: llc: Remove unused function declarations
Yue Haibing [Thu, 3 Aug 2023 13:47:47 +0000 (21:47 +0800)]
net: llc: Remove unused function declarations

llc_conn_ac_send_i_rsp_as_ack() and llc_conn_ev_sendack_tmr_exp()
are never implemented since beginning of git history.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Link: https://lore.kernel.org/r/20230803134747.41512-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge branch 'devlink-use-spec-to-generate-split-ops'
Jakub Kicinski [Fri, 4 Aug 2023 21:03:03 +0000 (14:03 -0700)]
Merge branch 'devlink-use-spec-to-generate-split-ops'

Jiri Pirko says:

====================
devlink: use spec to generate split ops

This is an outcome of the discussion in the following thread:
https://lore.kernel.org/netdev/20230720121829.566974-1-jiri@resnulli.us/
It serves as a dependency on the linked selector patchset.

There is an existing spec for devlink used for userspace part
generation. There are two commands supported there.

This patchset extends the spec so kernel split ops code could
be generated from it.
====================

Link: https://lore.kernel.org/r/20230803111340.1074067-1-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodevlink: use generated split ops and remove duplicated commands from small ops
Jiri Pirko [Thu, 3 Aug 2023 11:13:40 +0000 (13:13 +0200)]
devlink: use generated split ops and remove duplicated commands from small ops

Do the switch and use generated split ops for get and info_get commands.
Remove those from small ops array.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-13-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodevlink: include the generated netlink header
Jiri Pirko [Thu, 3 Aug 2023 11:13:39 +0000 (13:13 +0200)]
devlink: include the generated netlink header

Put the newly added generated header to the include list. Remove the
duplicated temporary function prototypes.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-12-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodevlink: add split ops generated according to spec
Jiri Pirko [Thu, 3 Aug 2023 11:13:38 +0000 (13:13 +0200)]
devlink: add split ops generated according to spec

Improve the existing devlink spec in order to serve as a source for
generation of valid devlink split ops for the existing commands.
Add the generated sources.

Node that the policies are narrowed down only to the attributes that
are actually parsed. The dont-validate-strict parsing policy makes sure
that other possibly passed garbage attributes from userspace are
ignored during validation.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-11-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonetlink: specs: devlink: add info-get dump op
Jiri Pirko [Thu, 3 Aug 2023 11:13:37 +0000 (13:13 +0200)]
netlink: specs: devlink: add info-get dump op

Add missing dump op for info-get command and re-generate related
devlink-user.[ch] code.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-10-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodevlink: un-static devlink_nl_pre/post_doit()
Jiri Pirko [Thu, 3 Aug 2023 11:13:36 +0000 (13:13 +0200)]
devlink: un-static devlink_nl_pre/post_doit()

To be prepared for the follow-up generated split ops addition,
make the functions devlink_nl_pre_doit() and devlink_nl_post_doit()
usable outside of netlink.c. Introduce temporary prototypes which are
going to be removed once the generated header will be included.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-9-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodevlink: introduce couple of dumpit callbacks for split ops
Jiri Pirko [Thu, 3 Aug 2023 11:13:35 +0000 (13:13 +0200)]
devlink: introduce couple of dumpit callbacks for split ops

Introduce couple of dumpit callbacks for generated split ops. Have them
as a thin wrapper around iteration function and allow to pass dump_one()
function pointer directly without need to store in devlink_cmd structs.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-8-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodevlink: rename couple of doit netlink callbacks to match generated names
Jiri Pirko [Thu, 3 Aug 2023 11:13:34 +0000 (13:13 +0200)]
devlink: rename couple of doit netlink callbacks to match generated names

The generated names of the doit netlink callback are missing "cmd" in
their names. Change names to be ready to switch to generated split ops
header.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-7-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodevlink: rename devlink_nl_ops to devlink_nl_small_ops
Jiri Pirko [Thu, 3 Aug 2023 11:13:33 +0000 (13:13 +0200)]
devlink: rename devlink_nl_ops to devlink_nl_small_ops

In order to avoid name collision with the generated split ops array
which is going to be introduced as a follow-up patch, rename
the existing ops array to devlink_nl_small_ops.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-6-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoynl-gen-c.py: render netlink policies static for split ops
Jiri Pirko [Thu, 3 Aug 2023 11:13:32 +0000 (13:13 +0200)]
ynl-gen-c.py: render netlink policies static for split ops

When policies are rendered for split ops, they are consumed in the same
file. No need to expose them for user outside, make them static.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-5-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoynl-gen-c.py: allow directional model for kernel mode
Jiri Pirko [Thu, 3 Aug 2023 11:13:31 +0000 (13:13 +0200)]
ynl-gen-c.py: allow directional model for kernel mode

Directional model limitation is only applicable for uapi mode.
For kernel mode, the code is generated correctly using right cmd values
for do/dump requests. Lift the limitation.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-4-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoynl-gen-c.py: filter rendering of validate field values for split ops
Jiri Pirko [Thu, 3 Aug 2023 11:13:30 +0000 (13:13 +0200)]
ynl-gen-c.py: filter rendering of validate field values for split ops

For split ops, do and dump has different meaningful values in
validate field.

Fix the rendering to allow the values per op type as follows:
do: strict
dump: dump, strict-dump

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-3-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonetlink: specs: add dump-strict flag for dont-validate property
Jiri Pirko [Thu, 3 Aug 2023 11:13:29 +0000 (13:13 +0200)]
netlink: specs: add dump-strict flag for dont-validate property

Allow user to specify GENL_DONT_VALIDATE_DUMP_STRICT flag for validation
and add this flag to netlink spec schema.

Signed-off-by: Jiri Pirko <jiri@nvidia.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Link: https://lore.kernel.org/r/20230803111340.1074067-2-jiri@resnulli.us
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agonet: lan966x: Do not check 0 for platform_get_irq_byname()
Zhu Wang [Thu, 3 Aug 2023 08:29:00 +0000 (16:29 +0800)]
net: lan966x: Do not check 0 for platform_get_irq_byname()

Since platform_get_irq_byname() never returned zero, so it need not to
check whether it returned zero, it returned -EINVAL or -ENXIO when
failed, so we replace the return error code with the result it returned.

Signed-off-by: Zhu Wang <wangzhu9@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Reviewed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: vlan: update wrong comments
Eric Dumazet [Thu, 3 Aug 2023 07:14:26 +0000 (07:14 +0000)]
net: vlan: update wrong comments

vlan_insert_tag() and friends do not allocate a new skb.
However they might allocate a new skb->head.
Update their comments to better describe their behavior.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agotcp/dccp: cache line align inet_hashinfo
Eric Dumazet [Thu, 3 Aug 2023 07:53:34 +0000 (07:53 +0000)]
tcp/dccp: cache line align inet_hashinfo

I have seen tcp_hashinfo starting at a non optimal location,
forcing input handlers to pull two cache lines instead of one,
and sharing a cache line that was dirtied more than necessary:

ffffffff83680600 b tcp_orphan_timer
ffffffff83680628 b tcp_orphan_cache
ffffffff8368062c b tcp_enable_tx_delay.__tcp_tx_delay_enabled
ffffffff83680630 B tcp_hashinfo
ffffffff83680680 b tcp_cong_list_lock

After this patch, ehash, ehash_locks, ehash_mask and ehash_locks_mask
are located in a read-only cache line.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: mana: Configure hwc timeout from hardware
Souradeep Chakrabarti [Wed, 2 Aug 2023 11:07:40 +0000 (04:07 -0700)]
net: mana: Configure hwc timeout from hardware

At present hwc timeout value is a fixed value. This patch sets the hwc
timeout from the hardware. It now uses a new hardware capability
GDMA_DRV_CAP_FLAG_1_HWC_TIMEOUT_RECONFIG to query and set the value
in hwc_timeout.

Signed-off-by: Souradeep Chakrabarti <schakrabarti@linux.microsoft.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: microchip: vcap api: Use ERR_CAST() in vcap_decode_rule()
Li Zetao [Wed, 2 Aug 2023 09:31:56 +0000 (17:31 +0800)]
net: microchip: vcap api: Use ERR_CAST() in vcap_decode_rule()

There is a warning reported by coccinelle:

./drivers/net/ethernet/microchip/vcap/vcap_api.c:2399:9-16: WARNING:
ERR_CAST can be used with ri

Use ERR_CAST instead of ERR_PTR + PTR_ERR to simplify the
conversion process.

Signed-off-by: Li Zetao <lizetao1@huawei.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agoMerge tag 'linux-can-next-for-6.6-20230803' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Fri, 4 Aug 2023 07:46:07 +0000 (08:46 +0100)]
Merge tag 'linux-can-next-for-6.6-20230803' of git://git./linux/kernel/git/mkl/linux-can-next

Marc Kleine-Budde says:

====================
pull-request: can-next 2023-08-03

This is a pull request of 9 patches for net-next/master.

The 1st patch is by Ruan Jinjie, targets the flexcan driver, and
cleans up the error handling of platform_get_irq() in the
flexcan_probe() function.

Markus Schneider-Pargmann contributes 6 patches for the tcan4x5x M_CAN
driver, consisting of some cleanups, and adding support for the
tcan4552/4553 chips.

Another patch by Ruan Jinjie, that cleans up the error path of
platform_get_irq() in the c_can_plat_probe() function of the C_CAN
platform driver.

The last patch is by Frank Jungclaus and adds support for the
CAN-USB/3 and CAN FD to the ESD USB CAN driver.
================

Signed-off-by: David S. Miller <davem@davemloft.net>
11 months agonet: Space.h: Remove unused function declarations
Yue Haibing [Wed, 2 Aug 2023 13:07:16 +0000 (21:07 +0800)]
net: Space.h: Remove unused function declarations

Commit 5aa83a4c0a15 ("  [PATCH] remove two obsolete net drivers") remove fmv18x_probe().
And commmit 01f4685797a5 ("eth: amd: remove NI6510 support (ni65)") leave ni65_probe().
Commit a10079c66290 ("staging: remove hp100 driver") remove hp100 driver and hp100_probe()
declaration is not used anymore.

sonic_probe() and iph5526_probe() are never implemented since the beginning of git history.

Signed-off-by: Yue Haibing <yuehaibing@huawei.com>
Link: https://lore.kernel.org/r/20230802130716.37308-1-yuehaibing@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoeth: dpaa: add missing net/xdp.h include
Jakub Kicinski [Thu, 3 Aug 2023 23:00:07 +0000 (16:00 -0700)]
eth: dpaa: add missing net/xdp.h include

Add missing include for DPAA (fix aarch64 build).

Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202308040620.ty8oYNOP-lkp@intel.com/
Fixes: 680ee0456a57 ("net: invert the netdevice.h vs xdp.h dependency")
Reviewed-by: Martin KaFai Lau <martin.lau@kernel.org>
Link: https://lore.kernel.org/r/20230803230008.362214-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf...
Jakub Kicinski [Thu, 3 Aug 2023 22:34:36 +0000 (15:34 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf-next

Martin KaFai Lau says:

====================
pull-request: bpf-next 2023-08-03

We've added 54 non-merge commits during the last 10 day(s) which contain
a total of 84 files changed, 4026 insertions(+), 562 deletions(-).

The main changes are:

1) Add SO_REUSEPORT support for TC bpf_sk_assign from Lorenz Bauer,
   Daniel Borkmann

2) Support new insns from cpu v4 from Yonghong Song

3) Non-atomically allocate freelist during prefill from YiFei Zhu

4) Support defragmenting IPv(4|6) packets in BPF from Daniel Xu

5) Add tracepoint to xdp attaching failure from Leon Hwang

6) struct netdev_rx_queue and xdp.h reshuffling to reduce
   rebuild time from Jakub Kicinski

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (54 commits)
  net: invert the netdevice.h vs xdp.h dependency
  net: move struct netdev_rx_queue out of netdevice.h
  eth: add missing xdp.h includes in drivers
  selftests/bpf: Add testcase for xdp attaching failure tracepoint
  bpf, xdp: Add tracepoint to xdp attaching failure
  selftests/bpf: fix static assert compilation issue for test_cls_*.c
  bpf: fix bpf_probe_read_kernel prototype mismatch
  riscv, bpf: Adapt bpf trampoline to optimized riscv ftrace framework
  libbpf: fix typos in Makefile
  tracing: bpf: use struct trace_entry in struct syscall_tp_t
  bpf, devmap: Remove unused dtab field from bpf_dtab_netdev
  bpf, cpumap: Remove unused cmap field from bpf_cpu_map_entry
  netfilter: bpf: Only define get_proto_defrag_hook() if necessary
  bpf: Fix an array-index-out-of-bounds issue in disasm.c
  net: remove duplicate INDIRECT_CALLABLE_DECLARE of udp[6]_ehashfn
  docs/bpf: Fix malformed documentation
  bpf: selftests: Add defrag selftests
  bpf: selftests: Support custom type and proto for client sockets
  bpf: selftests: Support not connecting client socket
  netfilter: bpf: Support BPF_F_NETFILTER_IP_DEFRAG in netfilter link
  ...
====================

Link: https://lore.kernel.org/r/20230803174845.825419-1-martin.lau@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Jakub Kicinski [Thu, 3 Aug 2023 21:29:50 +0000 (14:29 -0700)]
Merge git://git./linux/kernel/git/netdev/net

Cross-merge networking fixes after downstream PR.

Conflicts:

net/dsa/port.c
  9945c1fb03a3 ("net: dsa: fix older DSA drivers using phylink")
  a88dd7538461 ("net: dsa: remove legacy_pre_march2020 detection")
https://lore.kernel.org/all/20230731102254.2c9868ca@canb.auug.org.au/

net/xdp/xsk.c
  3c5b4d69c358 ("net: annotate data-races around sk->sk_mark")
  b7f72a30e9ac ("xsk: introduce wrappers and helpers for supporting multi-buffer in Tx path")
https://lore.kernel.org/all/20230731102631.39988412@canb.auug.org.au/

drivers/net/ethernet/broadcom/bnxt/bnxt.c
  37b61cda9c16 ("bnxt: don't handle XDP in netpoll")
  2b56b3d99241 ("eth: bnxt: handle invalid Tx completions more gracefully")
https://lore.kernel.org/all/20230801101708.1dc7faac@canb.auug.org.au/

Adjacent changes:

drivers/net/ethernet/mellanox/mlx5/core/en_accel/ipsec_fs.c
  62da08331f1a ("net/mlx5e: Set proper IPsec source port in L4 selector")
  fbd517549c32 ("net/mlx5e: Add function to get IPsec offload namespace")

drivers/net/ethernet/sfc/selftest.c
  55c1528f9b97 ("sfc: fix field-spanning memcpy in selftest")
  ae9d445cd41f ("sfc: Miscellaneous comment removals")

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge tag 'net-6.5-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 3 Aug 2023 21:00:02 +0000 (14:00 -0700)]
Merge tag 'net-6.5-rc5' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from bpf and wireless.

  Nothing scary here. Feels like the first wave of regressions from v6.5
  is addressed - one outstanding fix still to come in TLS for the
  sendpage rework.

  Current release - regressions:

   - udp: fix __ip_append_data()'s handling of MSG_SPLICE_PAGES

   - dsa: fix older DSA drivers using phylink

  Previous releases - regressions:

   - gro: fix misuse of CB in udp socket lookup

   - mlx5: unregister devlink params in case interface is down

   - Revert "wifi: ath11k: Enable threaded NAPI"

  Previous releases - always broken:

   - sched: cls_u32: fix match key mis-addressing

   - sched: bind logic fixes for cls_fw, cls_u32 and cls_route

   - add bound checks to a number of places which hand-parse netlink

   - bpf: disable preemption in perf_event_output helpers code

   - qed: fix scheduling in a tasklet while getting stats

   - avoid using APIs which are not hardirq-safe in couple of drivers,
     when we may be in a hard IRQ (netconsole)

   - wifi: cfg80211: fix return value in scan logic, avoid page
     allocator warning

   - wifi: mt76: mt7615: do not advertise 5 GHz on first PHY of MT7615D
     (DBDC)

  Misc:

   - drop handful of inactive maintainers, put some new in place"

* tag 'net-6.5-rc5' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (98 commits)
  MAINTAINERS: update TUN/TAP maintainers
  test/vsock: remove vsock_perf executable on `make clean`
  tcp_metrics: fix data-race in tcpm_suck_dst() vs fastopen
  tcp_metrics: annotate data-races around tm->tcpm_net
  tcp_metrics: annotate data-races around tm->tcpm_vals[]
  tcp_metrics: annotate data-races around tm->tcpm_lock
  tcp_metrics: annotate data-races around tm->tcpm_stamp
  tcp_metrics: fix addr_same() helper
  prestera: fix fallback to previous version on same major version
  udp: Fix __ip_append_data()'s handling of MSG_SPLICE_PAGES
  net/mlx5e: Set proper IPsec source port in L4 selector
  net/mlx5: fs_core: Skip the FTs in the same FS_TYPE_PRIO_CHAINS fs_prio
  net/mlx5: fs_core: Make find_closest_ft more generic
  wifi: brcmfmac: Fix field-spanning write in brcmf_scan_params_v2_to_v1()
  vxlan: Fix nexthop hash size
  ip6mr: Fix skb_under_panic in ip6mr_cache_report()
  s390/qeth: Don't call dev_close/dev_open (DOWN/UP)
  net: tap_open(): set sk_uid from current_fsuid()
  net: tun_chr_open(): set sk_uid from current_fsuid()
  net: dcb: choose correct policy to parse DCB_ATTR_BCN
  ...

11 months agoMAINTAINERS: update TUN/TAP maintainers
Jakub Kicinski [Wed, 2 Aug 2023 18:28:43 +0000 (11:28 -0700)]
MAINTAINERS: update TUN/TAP maintainers

Willem and Jason have agreed to take over the maintainer
duties for TUN/TAP, thank you!

There's an existing entry for TUN/TAP which only covers
the user mode Linux implementation.
Since we haven't heard from Maxim on the list for almost
a decade, extend that entry and take it over, rather than
adding a new one.

Acked-by: Willem de Bruijn <willemb@google.com>
Acked-by: Jason Wang <jasowang@redhat.com>
Link: https://lore.kernel.org/r/20230802182843.4193099-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 3 Aug 2023 18:22:53 +0000 (11:22 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf

Martin KaFai Lau says:

====================
pull-request: bpf 2023-08-03

We've added 5 non-merge commits during the last 7 day(s) which contain
a total of 3 files changed, 37 insertions(+), 20 deletions(-).

The main changes are:

1) Disable preemption in perf_event_output helpers code,
   from Jiri Olsa

2) Add length check for SK_DIAG_BPF_STORAGE_REQ_MAP_FD parsing,
   from Lin Ma

3) Multiple warning splat fixes in cpumap from Hou Tao

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  bpf, cpumap: Handle skb as well when clean up ptr_ring
  bpf, cpumap: Make sure kthread is running before map update returns
  bpf: Add length check for SK_DIAG_BPF_STORAGE_REQ_MAP_FD parsing
  bpf: Disable preemption in bpf_event_output
  bpf: Disable preemption in bpf_perf_event_output
====================

Link: https://lore.kernel.org/r/20230803181429.994607-1-martin.lau@linux.dev
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge ath-next from git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git
Kalle Valo [Thu, 3 Aug 2023 18:16:27 +0000 (21:16 +0300)]
Merge ath-next from git://git./linux/kernel/git/kvalo/ath.git

ath.git patches for v6.6. Major changes:

ath12k

* Extremely High Throughput (EHT) PHY support for Wi-Fi 7

11 months agoMerge tag 'wireless-2023-08-03' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Thu, 3 Aug 2023 18:05:46 +0000 (11:05 -0700)]
Merge tag 'wireless-2023-08-03' of git://git./linux/kernel/git/wireless/wireless

Kalle Valo says:

====================
wireless fixes for v6.5

We did some house cleaning in MAINTAINERS file so several patches
about that. Few regressions fixed and also fix some recently enabled
memcpy() warnings. Only small commits and nothing special standing
out.

* tag 'wireless-2023-08-03' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless:
  wifi: brcmfmac: Fix field-spanning write in brcmf_scan_params_v2_to_v1()
  wifi: ray_cs: Replace 1-element array with flexible array
  MAINTAINERS: add Jeff as ath10k, ath11k and ath12k maintainer
  MAINTAINERS: wifi: mark mlw8k as orphan
  MAINTAINERS: wifi: mark b43 as orphan
  MAINTAINERS: wifi: mark zd1211rw as orphan
  MAINTAINERS: wifi: mark wl3501 as orphan
  MAINTAINERS: wifi: mark rndis_wlan as orphan
  MAINTAINERS: wifi: mark ar5523 as orphan
  MAINTAINERS: wifi: mark cw1200 as orphan
  MAINTAINERS: wifi: atmel: mark as orphan
  MAINTAINERS: wifi: rtw88: change Ping as the maintainer
  Revert "wifi: ath6k: silence false positive -Wno-dangling-pointer warning on GCC 12"
  wifi: cfg80211: Fix return value in scan logic
  Revert "wifi: ath11k: Enable threaded NAPI"
  MAINTAINERS: Update mwifiex maintainer list
  wifi: mt76: mt7615: do not advertise 5 GHz on first phy of MT7615D (DBDC)
====================

Link: https://lore.kernel.org/r/20230803140058.57476C433C9@smtp.kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotest/vsock: remove vsock_perf executable on `make clean`
Stefano Garzarella [Thu, 3 Aug 2023 08:54:54 +0000 (10:54 +0200)]
test/vsock: remove vsock_perf executable on `make clean`

We forgot to add vsock_perf to the rm command in the `clean`
target, so now we have a left over after `make clean` in
tools/testing/vsock.

Fixes: 8abbffd27ced ("test/vsock: vsock_perf utility")
Cc: AVKrasnov@sberdevices.ru
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Tested-by: Simon Horman <horms@kernel.org> # build-tested
Link: https://lore.kernel.org/r/20230803085454.30897-1-sgarzare@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge branch 'tcp_metrics-series-of-fixes'
Jakub Kicinski [Thu, 3 Aug 2023 17:58:27 +0000 (10:58 -0700)]
Merge branch 'tcp_metrics-series-of-fixes'

Eric Dumazet says:

====================
tcp_metrics: series of fixes

This series contains a fix for addr_same() and various
data-race annotations.

We still have to address races over tm->tcpm_saddr and
tm->tcpm_daddr later.
====================

Link: https://lore.kernel.org/r/20230802131500.1478140-1-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp_metrics: fix data-race in tcpm_suck_dst() vs fastopen
Eric Dumazet [Wed, 2 Aug 2023 13:15:00 +0000 (13:15 +0000)]
tcp_metrics: fix data-race in tcpm_suck_dst() vs fastopen

Whenever tcpm_new() reclaims an old entry, tcpm_suck_dst()
would overwrite data that could be read from tcp_fastopen_cache_get()
or tcp_metrics_fill_info().

We need to acquire fastopen_seqlock to maintain consistency.

For newly allocated objects, tcpm_new() can switch to kzalloc()
to avoid an extra fastopen_seqlock acquisition.

Fixes: 1fe4c481ba63 ("net-tcp: Fast Open client - cookie cache")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230802131500.1478140-7-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp_metrics: annotate data-races around tm->tcpm_net
Eric Dumazet [Wed, 2 Aug 2023 13:14:59 +0000 (13:14 +0000)]
tcp_metrics: annotate data-races around tm->tcpm_net

tm->tcpm_net can be read or written locklessly.

Instead of changing write_pnet() and read_pnet() and potentially
hurt performance, add the needed READ_ONCE()/WRITE_ONCE()
in tm_net() and tcpm_new().

Fixes: 849e8a0ca8d5 ("tcp_metrics: Add a field tcpm_net and verify it matches on lookup")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230802131500.1478140-6-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp_metrics: annotate data-races around tm->tcpm_vals[]
Eric Dumazet [Wed, 2 Aug 2023 13:14:58 +0000 (13:14 +0000)]
tcp_metrics: annotate data-races around tm->tcpm_vals[]

tm->tcpm_vals[] values can be read or written locklessly.

Add needed READ_ONCE()/WRITE_ONCE() to document this,
and force use of tcp_metric_get() and tcp_metric_set()

Fixes: 51c5d0c4b169 ("tcp: Maintain dynamic metrics in local cache.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp_metrics: annotate data-races around tm->tcpm_lock
Eric Dumazet [Wed, 2 Aug 2023 13:14:57 +0000 (13:14 +0000)]
tcp_metrics: annotate data-races around tm->tcpm_lock

tm->tcpm_lock can be read or written locklessly.

Add needed READ_ONCE()/WRITE_ONCE() to document this.

Fixes: 51c5d0c4b169 ("tcp: Maintain dynamic metrics in local cache.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230802131500.1478140-4-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp_metrics: annotate data-races around tm->tcpm_stamp
Eric Dumazet [Wed, 2 Aug 2023 13:14:56 +0000 (13:14 +0000)]
tcp_metrics: annotate data-races around tm->tcpm_stamp

tm->tcpm_stamp can be read or written locklessly.

Add needed READ_ONCE()/WRITE_ONCE() to document this.

Also constify tcpm_check_stamp() dst argument.

Fixes: 51c5d0c4b169 ("tcp: Maintain dynamic metrics in local cache.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230802131500.1478140-3-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agotcp_metrics: fix addr_same() helper
Eric Dumazet [Wed, 2 Aug 2023 13:14:55 +0000 (13:14 +0000)]
tcp_metrics: fix addr_same() helper

Because v4 and v6 families use separate inetpeer trees (respectively
net->ipv4.peers and net->ipv6.peers), inetpeer_addr_cmp(a, b) assumes
a & b share the same family.

tcp_metrics use a common hash table, where entries can have different
families.

We must therefore make sure to not call inetpeer_addr_cmp()
if the families do not match.

Fixes: d39d14ffa24c ("net: Add helper function to compare inetpeer addresses")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Link: https://lore.kernel.org/r/20230802131500.1478140-2-edumazet@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoprestera: fix fallback to previous version on same major version
Jonas Gorski [Wed, 2 Aug 2023 09:23:56 +0000 (11:23 +0200)]
prestera: fix fallback to previous version on same major version

When both supported and previous version have the same major version,
and the firmwares are missing, the driver ends in a loop requesting the
same (previous) version over and over again:

    [   76.327413] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.1.img firmware, fall-back to previous 4.0 version
    [   76.339802] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    [   76.352162] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    [   76.364502] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    [   76.376848] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    [   76.389183] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    [   76.401522] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    [   76.413860] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    [   76.426199] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.0.img firmware, fall-back to previous 4.0 version
    ...

Fix this by inverting the check to that we aren't yet at the previous
version, and also check the minor version.

This also catches the case where both versions are the same, as it was
after commit bb5dbf2cc64d ("net: marvell: prestera: add firmware v4.0
support").

With this fix applied:

    [   88.499622] Prestera DX 0000:01:00.0: missing latest mrvl/prestera/mvsw_prestera_fw-v4.1.img firmware, fall-back to previous 4.0 version
    [   88.511995] Prestera DX 0000:01:00.0: failed to request previous firmware: mrvl/prestera/mvsw_prestera_fw-v4.0.img
    [   88.522403] Prestera DX: probe of 0000:01:00.0 failed with error -2

Fixes: 47f26018a414 ("net: marvell: prestera: try to load previous fw version")
Signed-off-by: Jonas Gorski <jonas.gorski@bisdn.de>
Acked-by: Elad Nachman <enachman@marvell.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Acked-by: Taras Chornyi <taras.chornyi@plvision.eu>
Link: https://lore.kernel.org/r/20230802092357.163944-1-jonas.gorski@bisdn.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge branch 'docs-net-page_pool-sync-dev-and-kdoc'
Jakub Kicinski [Thu, 3 Aug 2023 16:54:25 +0000 (09:54 -0700)]
Merge branch 'docs-net-page_pool-sync-dev-and-kdoc'

Jakub Kicinski says:

====================
docs: net: page_pool: sync dev and kdoc

Document PP_FLAG_DMA_SYNC_DEV based on recent conversation.
Use kdoc to document structs and functions, to avoid duplication.

Olek, this will conflict with your work, but I think that trying
to make progress in parallel is the best course of action...
Retargetting at net-next to make it a little less bad.
====================

Link: https://lore.kernel.org/r/20230802161821.3621985-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodocs: net: page_pool: use kdoc to avoid duplicating the information
Jakub Kicinski [Wed, 2 Aug 2023 16:18:21 +0000 (09:18 -0700)]
docs: net: page_pool: use kdoc to avoid duplicating the information

All struct members of the driver-facing APIs are documented twice,
in the code and under Documentation. This is a bit tedious.

I also get the feeling that a lot of developers will read the header
when coding, rather than the doc. Bring the two a little closer
together by using kdoc for structs and functions.

Using kdoc also gives us links (mentioning a function or struct
in the text gets replaced by a link to its doc).

Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Tested-by: Randy Dunlap <rdunlap@infradead.org>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://lore.kernel.org/r/20230802161821.3621985-3-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agodocs: net: page_pool: document PP_FLAG_DMA_SYNC_DEV parameters
Jakub Kicinski [Wed, 2 Aug 2023 16:18:20 +0000 (09:18 -0700)]
docs: net: page_pool: document PP_FLAG_DMA_SYNC_DEV parameters

Using PP_FLAG_DMA_SYNC_DEV is a bit confusing. It was perhaps
more obvious when it was introduced but the page pool use
has grown beyond XDP and beyond packet-per-page so now
making the heads and tails out of this feature is not
trivial.

Obviously making the API more user friendly would be
a better fix, but until someone steps up to do that
let's at least document what the parameters are.

Relevant discussion in the first Link.

Link: https://lore.kernel.org/all/20230731114427.0da1f73b@kernel.org/
Reviewed-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Link: https://lore.kernel.org/r/20230802161821.3621985-2-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
11 months agoMerge tag 'nfsd-6.5-3' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux
Linus Torvalds [Thu, 3 Aug 2023 16:26:34 +0000 (09:26 -0700)]
Merge tag 'nfsd-6.5-3' of git://git./linux/kernel/git/cel/linux

Pull nfsd fix from Chuck Lever:

 - Fix tmpfs splice read support

* tag 'nfsd-6.5-3' of git://git.kernel.org/pub/scm/linux/kernel/git/cel/linux:
  nfsd: Fix reading via splice

11 months agoMerge tag 'erofs-for-6.5-rc5-fixes' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 3 Aug 2023 16:20:50 +0000 (09:20 -0700)]
Merge tag 'erofs-for-6.5-rc5-fixes' of git://git./linux/kernel/git/xiang/erofs

Pull erofs fixes from Gao Xiang:

 - Fix data corruption caused by insufficient decompression on
   deduplicated compressed extents

 - Drop a useless s_magic checking in erofs_kill_sb()

* tag 'erofs-for-6.5-rc5-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs:
  erofs: drop unnecessary WARN_ON() in erofs_kill_sb()
  erofs: fix wrong primary bvec selection on deduplicated extents

11 months agoMerge tag 's390-6.5-4' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux
Linus Torvalds [Thu, 3 Aug 2023 16:06:38 +0000 (09:06 -0700)]
Merge tag 's390-6.5-4' of git://git./linux/kernel/git/s390/linux

Pull s390 fixes from Heiko Carstens:

 - Split kernel large page mappings into 4k mappings in case debug
   pagealloc is enabled again. This got accidentally removed by commit
   bb1520d581a3 ("s390/mm: start kernel with DAT enabled")

 - Fix error handling in KVM's sthyi handling

 - Add missing include to s390's uapi ptrace.h

 - Update defconfigs

* tag 's390-6.5-4' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux:
  s390/ptrace: add missing linux/const.h include
  KVM: s390: fix sthyi error handling
  s390: update defconfigs
  s390/vmem: split pages when debug pagealloc is enabled

11 months agoMerge branch 'net: struct netdev_rx_queue and xdp.h reshuffling'
Martin KaFai Lau [Thu, 3 Aug 2023 15:38:07 +0000 (08:38 -0700)]
Merge branch 'net: struct netdev_rx_queue and xdp.h reshuffling'

Jakub Kicinski says:

====================
While poking at struct netdev_rx_queue I got annoyed by
the huge rebuild times. I split it out from netdevice.h
and then realized that it was the main reason we included
xdp.h in there. So I removed that dependency as well.

This gives us very pleasant build times for both xdp.h
and struct netdev_rx_queue changes.

I'm sending this for bpf-next because I think it'd be easiest
if it goes in there, and then bpf-next gets flushed soon after?
I can also make a branch on merge-base for net-next and bpf-next..

v2:
 - build fix
 - reorder some includes
v1: https://lore.kernel.org/all/20230802003246.2153774-1-kuba@kernel.org/
====================

Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
11 months agonet: invert the netdevice.h vs xdp.h dependency
Jakub Kicinski [Thu, 3 Aug 2023 01:02:30 +0000 (18:02 -0700)]
net: invert the netdevice.h vs xdp.h dependency

xdp.h is far more specific and is included in only 67 other
files vs netdevice.h's 1538 include sites.
Make xdp.h include netdevice.h, instead of the other way around.
This decreases the incremental allmodconfig builds size when
xdp.h is touched from 5947 to 662 objects.

Move bpf_prog_run_xdp() to xdp.h, seems appropriate and filter.h
is a mega-header in its own right so it's nice to avoid xdp.h
getting included there as well.

The only unfortunate part is that the typedef for xdp_features_t
has to move to netdevice.h, since its embedded in struct netdevice.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://lore.kernel.org/r/20230803010230.1755386-4-kuba@kernel.org
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
11 months agonet: move struct netdev_rx_queue out of netdevice.h
Jakub Kicinski [Thu, 3 Aug 2023 01:02:29 +0000 (18:02 -0700)]
net: move struct netdev_rx_queue out of netdevice.h

struct netdev_rx_queue is touched in only a few places
and having it defined in netdevice.h brings in the dependency
on xdp.h, because struct xdp_rxq_info gets embedded in
struct netdev_rx_queue.

In prep for removal of xdp.h from netdevice.h move all
the netdev_rx_queue stuff to a new header.

We could technically break the new header up to avoid
the sysfs.h include but it's so rarely included it
doesn't seem to be worth it at this point.

Reviewed-by: Amritha Nambiar <amritha.nambiar@intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://lore.kernel.org/r/20230803010230.1755386-3-kuba@kernel.org
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
11 months agoeth: add missing xdp.h includes in drivers
Jakub Kicinski [Thu, 3 Aug 2023 01:02:28 +0000 (18:02 -0700)]
eth: add missing xdp.h includes in drivers

Handful of drivers currently expect to get xdp.h by virtue
of including netdevice.h. This will soon no longer be the case
so add explicit includes.

Reviewed-by: Wei Fang <wei.fang@nxp.com>
Reviewed-by: Gerhard Engleder <gerhard@engleder-embedded.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Jesper Dangaard Brouer <hawk@kernel.org>
Link: https://lore.kernel.org/r/20230803010230.1755386-2-kuba@kernel.org
Signed-off-by: Martin KaFai Lau <martin.lau@kernel.org>
11 months agonet/mlx4: Remove many unnecessary NULL values
Ruan Jinjie [Wed, 2 Aug 2023 04:00:26 +0000 (12:00 +0800)]
net/mlx4: Remove many unnecessary NULL values

The NULL initialization of the pointers assigned by kzalloc() first is
not necessary, because if the kzalloc() failed, the pointers will be
assigned NULL, otherwise it works as usual. so remove it.

Signed-off-by: Ruan Jinjie <ruanjinjie@huawei.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20230802040026.2588675-1-ruanjinjie@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agoMerge branch 'selftests-openvswitch-add-flow-programming-cases'
Paolo Abeni [Thu, 3 Aug 2023 13:05:43 +0000 (15:05 +0200)]
Merge branch 'selftests-openvswitch-add-flow-programming-cases'

Aaron Conole says:

====================
selftests: openvswitch: add flow programming cases

The openvswitch selftests currently contain a few cases for managing the
datapath, which includes creating datapath instances, adding interfaces,
and doing some basic feature / upcall tests.  This is useful to validate
the control path.

Add the ability to program some of the more common flows with actions. This
can be improved overtime to include regression testing, etc.
====================

Link: https://lore.kernel.org/r/20230801212226.909249-1-aconole@redhat.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agoselftests: openvswitch: add ct-nat test case with ipv4
Aaron Conole [Tue, 1 Aug 2023 21:22:26 +0000 (17:22 -0400)]
selftests: openvswitch: add ct-nat test case with ipv4

Building on the previous work, add a very simplistic NAT case
using ipv4.  This just tests dnat transformation

Signed-off-by: Aaron Conole <aconole@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agoselftests: openvswitch: add basic ct test case parsing
Aaron Conole [Tue, 1 Aug 2023 21:22:25 +0000 (17:22 -0400)]
selftests: openvswitch: add basic ct test case parsing

Forwarding via ct() action is an important use case for openvswitch, but
generally would require using a full ovs-vswitchd to get working. Add a
ct action parser for basic ct test case.

Signed-off-by: Aaron Conole <aconole@redhat.com>
Reviewed-by: Adrian Moreno <amorenoz@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agoselftests: openvswitch: add a test for ipv4 forwarding
Aaron Conole [Tue, 1 Aug 2023 21:22:24 +0000 (17:22 -0400)]
selftests: openvswitch: add a test for ipv4 forwarding

This is a simple ipv4 bidirectional connectivity test.

Signed-off-by: Aaron Conole <aconole@redhat.com>
Reviewed-by: Adrian Moreno <amorenoz@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agoselftests: openvswitch: support key masks
Adrian Moreno [Tue, 1 Aug 2023 21:22:23 +0000 (17:22 -0400)]
selftests: openvswitch: support key masks

The default value for the mask actually depends on the value (e.g: if
the value is non-null, the default is full-mask), so change the convert
functions to accept the full, possibly masked string and let them figure
out how to parse the different values.

Also, implement size-aware int parsing.

With this patch we can now express flows such as the following:
"eth(src=0a:ca:fe:ca:fe:0a/ff:ff:00:00:ff:00)"
"eth(src=0a:ca:fe:ca:fe:0a)" -> mask = ff:ff:ff:ff:ff:ff
"ipv4(src=192.168.1.1)" -> mask = 255.255.255.255
"ipv4(src=192.168.1.1/24)"
"ipv4(src=192.168.1.1/255.255.255.0)"
"tcp(src=8080)" -> mask = 0xffff
"tcp(src=8080/0xf0f0)"

Signed-off-by: Adrian Moreno <amorenoz@redhat.com>
Acked-by: Aaron Conole <aconole@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agoselftests: openvswitch: add an initial flow programming case
Aaron Conole [Tue, 1 Aug 2023 21:22:22 +0000 (17:22 -0400)]
selftests: openvswitch: add an initial flow programming case

The openvswitch self-tests can test much of the control side of
the module (ie: what a vswitchd implementation would process),
but the actual packet forwarding cases aren't supported, making
the testing of limited value.

Add some flow parsing and an initial ARP based test case using
arping utility.  This lets us display flows, add some basic
output flows with simple matches, and test against a known good
forwarding case.

Signed-off-by: Aaron Conole <aconole@redhat.com>
Reviewed-by: Adrian Moreno <amorenoz@redhat.com>
Reviewed-by: Simon Horman <horms@kernel.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agoudp6: Fix __ip6_append_data()'s handling of MSG_SPLICE_PAGES
David Howells [Wed, 2 Aug 2023 07:36:50 +0000 (08:36 +0100)]
udp6: Fix __ip6_append_data()'s handling of MSG_SPLICE_PAGES

__ip6_append_data() can has a similar problem to __ip_append_data()[1] when
asked to splice into a partially-built UDP message that has more than the
frag-limit data and up to the MTU limit, but in the ipv6 case, it errors
out with EINVAL.  This can be triggered with something like:

        pipe(pfd);
        sfd = socket(AF_INET6, SOCK_DGRAM, 0);
        connect(sfd, ...);
        send(sfd, buffer, 8137, MSG_CONFIRM|MSG_MORE);
        write(pfd[1], buffer, 8);
        splice(pfd[0], 0, sfd, 0, 0x4ffe0ul, 0);

where the amount of data given to send() is dependent on the MTU size (in
this instance an interface with an MTU of 8192).

The problem is that the calculation of the amount to copy in
__ip6_append_data() goes negative in two places, but a check has been put
in to give an error in this case.

This happens because when pagedlen > 0 (which happens for MSG_ZEROCOPY and
MSG_SPLICE_PAGES), the terms in:

        copy = datalen - transhdrlen - fraggap - pagedlen;

then mostly cancel when pagedlen is substituted for, leaving just -fraggap.

Fix this by:

 (1) Insert a note about the dodgy calculation of 'copy'.

 (2) If MSG_SPLICE_PAGES, clear copy if it is negative from the above
     equation, so that 'offset' isn't regressed and 'length' isn't
     increased, which will mean that length and thus copy should match the
     amount left in the iterator.

 (3) When handling MSG_SPLICE_PAGES, give a warning and return -EIO if
     we're asked to splice more than is in the iterator.  It might be
     better to not give the warning or even just give a 'short' write.

 (4) If MSG_SPLICE_PAGES, override the copy<0 check.

[!] Note that this should also affect MSG_ZEROCOPY, but that will return
-EINVAL for the range of send sizes that requires the skbuff to be split.

Signed-off-by: David Howells <dhowells@redhat.com>
cc: Willem de Bruijn <willemdebruijn.kernel@gmail.com>
cc: "David S. Miller" <davem@davemloft.net>
cc: Eric Dumazet <edumazet@google.com>
cc: Jakub Kicinski <kuba@kernel.org>
cc: Paolo Abeni <pabeni@redhat.com>
cc: David Ahern <dsahern@kernel.org>
cc: Jens Axboe <axboe@kernel.dk>
cc: Matthew Wilcox <willy@infradead.org>
cc: netdev@vger.kernel.org
Link: https://lore.kernel.org/r/000000000000881d0606004541d1@google.com/
Reviewed-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/1580952.1690961810@warthog.procyon.org.uk
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
11 months agonet: gemini: Do not check for 0 return after calling platform_get_irq()
Ruan Jinjie [Wed, 2 Aug 2023 08:52:16 +0000 (16:52 +0800)]
net: gemini: Do not check for 0 return after calling platform_get_irq()

It is not possible for platform_get_irq() to return 0. Use the
return value from platform_get_irq().

Signed-off-by: Ruan Jinjie <ruanjinjie@huawei.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Link: https://lore.kernel.org/r/20230802085216.659238-1-ruanjinjie@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>