Most of the time during the RX is caused by the compound_head() call
done at the end of the RX loop:

       │     build_skb():
       [...]
       │     static inline struct page *compound_head(struct page *page)
       │     {
       │     unsigned long head = READ_ONCE(page->compound_head);
 65.23 │       ldr  x2, [x1, #8]

Prefetch the page struct as soon as possible, to speedup the RX path
noticeabily by a ~3-4% packet rate in a drop test.

       │     build_skb():
       [...]
       │     static inline struct page *compound_head(struct page *page)
       │     {
       │     unsigned long head = READ_ONCE(page->compound_head);
 17.92 │       ldr  x2, [x1, #8]
Signed-off-by: NMatteo Croce <mcroce@microsoft.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In the RX buffer, the received data starts after a headroom used to
align the IP header and to allow prepending headers efficiently.
The prefetch() should take this into account, and prefetch from
the very start of the received data.

We can see that ether_addr_equal_64bits(), which is the first function
to access the data, drops from the top of the perf top output.

prefetch(data):

Overhead  Shared Object     Symbol
  11.64%  [kernel]          [k] eth_type_trans

prefetch(data + MVPP2_MH_SIZE + MVPP2_SKB_HEADROOM):

Overhead  Shared Object     Symbol
  13.42%  [kernel]          [k] build_skb
  10.35%  [mvpp2]           [k] mvpp2_rx
   9.35%  [kernel]          [k] __netif_receive_skb_core
   8.24%  [kernel]          [k] kmem_cache_free
   7.97%  [kernel]          [k] dev_gro_receive
   7.68%  [kernel]          [k] page_pool_put_page
   7.32%  [kernel]          [k] kmem_cache_alloc
   7.09%  [mvpp2]           [k] mvpp2_bm_pool_put
   3.36%  [kernel]          [k] eth_type_trans

Also, move the eth_type_trans() call a bit down, to give the RAM more
time to prefetch the data.
Signed-off-by: NMatteo Croce <mcroce@microsoft.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use the devm_platform_ioremap_resource_byname() helper instead of
calling platform_get_resource_byname() and devm_ioremap_resource()
separately.
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use the devm_platform_ioremap_resource_byname() helper instead of
calling platform_get_resource_byname() and devm_ioremap_resource()
separately.
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

It will cause null-ptr-deref if platform_get_resource() returns NULL,
we need check the return value.
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This reverts commits:
- 76241154
     nvme: NVME_TCP_OFFLOAD should not default to m
- 5ff5622e:
     Merge branch 'NVMeTCP-Offload-ULP'

As requested on the mailing-list: https://lore.kernel.org/netdev/SJ0PR18MB3882C20793EA35A3E8DAE300CC379@SJ0PR18MB3882.namprd18.prod.outlook.com/
This patch will revert the nvme-tcp-offload ULP from net-next.

The nvme-tcp-offload ULP series will continue to be considered only on
linux-nvme@lists.infradead.org.
Signed-off-by: NPrabhakar Kushwaha <pkushwaha@marvell.com>
Signed-off-by: NMichal Kalderon <mkalderon@marvell.com>
Signed-off-by: NAriel Elior <aelior@marvell.com>
Signed-off-by: NShai Malin <smalin@marvell.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The comparison of the u16 priv->phy_addr < 0 is always false because
phy_addr is unsigned. Fix this by assigning the return from the call
to function asix_read_phy_addr to int ret and using this for the
less than zero error check comparison.

Fixes: 7e88b11a ("net: usb: asix: refactor asix_read_phy_addr() and handle errors on return")
Addresses-Coverity: ("Unsigned compared against 0")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The comparison of the u16 priv->phy_addr < 0 is always false because
phy_addr is unsigned. Fix this by assigning the return from the call
to function asix_read_phy_addr to int ret and using this for the
less than zero error check comparison.

Addresses-Coverity: ("Unsigned compared against 0")
Fixes: e532a096 ("net: usb: asix: ax88772: add phylib support")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Pablo Neira Ayuso says:

====================
Netfilter updates for net-next

The following patchset contains Netfilter updates for net-next:

1) Add nfgenmsg field to nfnetlink's struct nfnl_info and use it.

2) Remove nft_ctx_init_from_elemattr() and nft_ctx_init_from_setattr()
   helper functions.

3) Add the nf_ct_pernet() helper function to fetch the conntrack
   pernetns data area.

4) Expose TCP and UDP flowtable offload timeouts through sysctl,
   from Oz Shlomo.

5) Add nfnetlink_hook subsystem to fetch the netfilter hook
   pipeline configuration, from Florian Westphal. This also includes
   a new field to annotate the hook type as metadata.

6) Fix unsafe memory access to non-linear skbuff in the new SCTP
   chunk support for nft_exthdr, from Phil Sutter.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

In normal situations where the driver doesn't dereference
"nsim_node->ddir" or "nsim_node->rate_parent" itself then we are not
supposed to check the return from debugfs functions.  In the case of
debugfs_create_dir() the check was wrong as well because it doesn't
return NULL, it returns error pointers.
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The WARN_ON() macro takes a condition, it doesn't take a message.  Use
WARN() instead.

Fixes: 1897db2e ("devlink: Allow setting tx rate for devlink rate leaf objects")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This code check "reg" but "ret" was intended so the error handling will
never trigger.

Fixes: 7c9896e3 ("net: dsa: qca8k: check return value of read functions correctly")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

The "hi" variable is a u64 but the qca8k_read() writes to the top 32
bits of it.  That will work on little endian systems but it's a bit
subtle.  It's cleaner to make declare "hi" as a u32.  We will still need
to cast it when we shift it later on in the function but that's fine.

Fixes: 7c9896e3 ("net: dsa: qca8k: check return value of read functions correctly")
Signed-off-by: NDan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: NVladimir Oltean <olteanv@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Peng Li says:

====================
net: lapbether: clean up some code style issues

This patchset clean up some code style issues.
====================
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

According to the chackpatch.pl,
line length of 123 exceeds 100 columns, so fix it.
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Alignment should match open parenthesis.
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

According to the chackpatch.pl, comparison to NULL could
be written "lapbeth_get_x25_dev".
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Networking block comments don't use an empty /* line,
use /* Comment...

Block comments use * on subsequent lines.
Block comments use a trailing */ on a separate line.

This patch fixes the comments style issues.
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NGuangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch removes unnecessary out of memory message,
to fix the following checkpatch.pl warning:
"WARNING: Possible unnecessary 'out of memory' message"
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NGuangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch removes trailing whitespaces.
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NGuangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Should not use assignment in if condition.
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NGuangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch fixes the checkpatch error about missing a blank line
after declarations.
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

This patch the redundant blank line.
Signed-off-by: NPeng Li <lipeng321@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Using list_move_tail() instead of list_del() + list_add_tail() in hclge_main.c.
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: NBaokun Li <libaokun1@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Using list_move_tail() instead of list_del() + list_add_tail() in hclgevf_main.c.
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: NBaokun Li <libaokun1@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Using list_move() instead of list_del() + list_add() in nfp_cppcore.c.
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: NBaokun Li <libaokun1@huawei.com>
Reviewed-by: NSimon Horman <simon.horman@corigine.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Fix a mistake in grammar.
Signed-off-by: Ngushengxian <gushengxian@yulong.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use devm_platform_get_and_ioremap_resource() to simplify
code.
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Reviewed-by: NSergei Shtylyov <sergei.shtylyov@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Commit ca893194 ("net: dsa: b53: Keep CPU port as tagged in all
VLANs") forced the CPU port to be always tagged in any VLAN membership.
This was necessary back then because we did not support Broadcom tags
for all configurations so the only way to differentiate tagged and
untagged traffic while DSA_TAG_PROTO_NONE was used was to force the CPU
port into being always tagged.

With most configurations enabling Broadcom tags, especially after
8fab459e ("net: dsa: b53: Enable Broadcom tags for 531x5/539x
families") we do not need to apply this unconditional force tagging of
the CPU port in all VLANs.

A helper function is introduced to faciliate the encapsulation of the
specific condition requiring the CPU port to be tagged in all VLANs and
the dsa_switch_ops::untag_bridge_pvid boolean is moved to when
dsa_switch_ops::setup is called when we have already determined the
tagging protocol we will be using.
Reported-by: NMatthew Hagan <mnhagan88@gmail.com>
Signed-off-by: NFlorian Fainelli <f.fainelli@gmail.com>
Reviewed-by: NVladimir Oltean <olteanv@gmail.com>
Tested-by: NMatthew Hagan <mnhagan88@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

coverity scanner says:
2187  if (nft_is_base_chain(chain)) {
vvv   CID 1505166:  Memory - corruptions  (UNINIT)
vvv   Using uninitialized value "basechain".
2188  basechain->ops.hook_ops_type = NF_HOOK_OP_NF_TABLES;

... I don't see how nft_is_base_chain() can evaluate to true
while basechain pointer is garbage.

However, it seems better to place the NF_HOOK_OP_NF_TABLES annotation
in nft_basechain_hook_init() instead.
Reported-by: Ncoverity-bot <keescook+coverity-bot@chromium.org>
Addresses-Coverity-ID: 1505166 ("Memory - corruptions")
Fixes: 65b8b7bfc5284f ("netfilter: annotate nf_tables base hook ops")
Signed-off-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

nfnetlink_hook.c: In function 'nfnl_hook_put_nft_chain_info':
nfnetlink_hook.c:76:7: error: implicit declaration of 'nft_is_active'

This macro is only defined when NF_TABLES is enabled.
While its possible to also add an ifdef-guard, the infrastructure
is currently not useful without nf_tables.
Reported-by: Nkernel test robot <lkp@intel.com>
Fixes: 252956528caa ("netfilter: add new hook nfnl subsystem")
Suggested-by: NPablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

Currently the array net->nf.hooks_ipv6 is accessed by index hook
before hook is sanity checked. Fix this by moving the sanity check
to before the array access.

Addresses-Coverity: ("Out-of-bounds access")
Fixes: e2cf17d3 ("netfilter: add new hook nfnl subsystem")
Signed-off-by: NColin Ian King <colin.king@canonical.com>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

Fix some mistakes in grammar.
Signed-off-by: Ngushengxian <gushengxian@yulong.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

While iterating through an SCTP packet's chunks, skb_header_pointer() is
called for the minimum expected chunk header size. If (that part of) the
skbuff is non-linear, the following memcpy() may read data past
temporary buffer '_sch'. Use skb_copy_bits() instead which does the
right thing in this situation.

Fixes: 133dc203 ("netfilter: nft_exthdr: Support SCTP chunks")
Suggested-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NPhil Sutter <phil@nwl.cc>
Reviewed-by: NFlorian Westphal <fw@strlen.de>
Signed-off-by: NPablo Neira Ayuso <pablo@netfilter.org>

We are currently assuming that GMAC_AHB_RESET will already be deasserted
by the bootloader. However if this has not been done, probing of the GMAC
will fail. To remedy this we must ensure GMAC_AHB_RESET has been deasserted
prior to probing.

v2 changes:
 - remove NULL condition check for stmmac_ahb_rst in stmmac_main.c
 - unwrap dev_err() message in stmmac_main.c
 - add PTR_ERR() around plat->stmmac_ahb_rst in stmmac_platform.c

v3 changes:
 - add error pointer to dev_err() output
 - add reset_control_assert(stmmac_ahb_rst) in stmmac_dvr_remove
 - revert PTR_ERR() around plat->stmmac_ahb_rst since this is performed
   on the returned value of ret by the calling function
Signed-off-by: NMatthew Hagan <mnhagan88@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use devm_platform_get_and_ioremap_resource() to simplify
code.
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Reviewed-by: NSergei Shtylyov <sergei.shtylyov@gmail.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Use devm_platform_get_and_ioremap_resource() and
devm_platform_ioremap_resource_byname to simplify
code.
Signed-off-by: NYang Yingliang <yangyingliang@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Instead of relying on the static initialization done by ocelot_init_port()
which enables flow control unconditionally, set SYS_PAUSE_CFG_PAUSE_ENA
according to the parameters negotiated by the PHY.
Signed-off-by: NVladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Convert list_for_each() to list_for_each_entry() where
applicable. This simplifies the code.
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: NWang Hai <wanghai38@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>

Convert list_for_each() to list_for_each_entry() where
applicable. This simplifies the code.
Reported-by: NHulk Robot <hulkci@huawei.com>
Signed-off-by: NWang Hai <wanghai38@huawei.com>
Signed-off-by: NDavid S. Miller <davem@davemloft.net>