linux-2.6-microblaze.git
2 years agonet: hns3: add some required spaces
Hao Chen [Mon, 30 Aug 2021 13:51:08 +0000 (21:51 +0800)]
net: hns3: add some required spaces

Add some required spaces to improve readability.

Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: clean up a type mismatch warning
Guojia Liao [Mon, 30 Aug 2021 13:51:07 +0000 (21:51 +0800)]
net: hns3: clean up a type mismatch warning

abs() returns signed long, which could not convert the type
as unsigned, and it may cause a mismatch type warning from
static tools. To fix it, this patch uses an variable to save
the abs()'s result and does a explicit conversion.

Signed-off-by: Guojia Liao <liaoguojia@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: refine function hns3_set_default_feature()
Jian Shen [Mon, 30 Aug 2021 13:51:06 +0000 (21:51 +0800)]
net: hns3: refine function hns3_set_default_feature()

Currently, the driver sets default feature for netdev->features,
netdev->hw_features, netdev->vlan_features and
netdev->hw_enc_features separately. It's fussy, because most
of the feature bits are same. So refine it by copy value from
netdev->features.

Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoipv6: remove duplicated 'net/lwtunnel.h' include
Lv Ruyi [Tue, 31 Aug 2021 09:09:06 +0000 (02:09 -0700)]
ipv6: remove duplicated 'net/lwtunnel.h' include

Remove all but the first include of net/lwtunnel.h from seg6_iptunnel.c.

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Lv Ruyi <lv.ruyi@zte.com.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: w5100: check return value after calling platform_get_resource()
Yang Yingliang [Tue, 31 Aug 2021 08:40:18 +0000 (16:40 +0800)]
net: w5100: check return value after calling platform_get_resource()

It will cause null-ptr-deref if platform_get_resource() returns NULL,
we need check the return value.

Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet/mlxbf_gige: Make use of devm_platform_ioremap_resourcexxx()
Cai Huoqing [Tue, 31 Aug 2021 08:02:31 +0000 (16:02 +0800)]
net/mlxbf_gige: Make use of devm_platform_ioremap_resourcexxx()

Use the devm_platform_ioremap_resource_byname() helper instead of
calling platform_get_resource_byname() and devm_ioremap_resource()
separately

Use the devm_platform_ioremap_resource() helper instead of
calling platform_get_resource() and devm_ioremap_resource()
separately

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: mdio: mscc-miim: Make use of the helper function devm_platform_ioremap_resource()
Cai Huoqing [Tue, 31 Aug 2021 07:58:18 +0000 (15:58 +0800)]
net: mdio: mscc-miim: Make use of the helper function devm_platform_ioremap_resource()

Use the devm_platform_ioremap_resource() helper instead of
calling platform_get_resource() and devm_ioremap_resource()
separately

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: mdio-ipq4019: Make use of devm_platform_ioremap_resource()
Cai Huoqing [Tue, 31 Aug 2021 07:56:58 +0000 (15:56 +0800)]
net: mdio-ipq4019: Make use of devm_platform_ioremap_resource()

Use the devm_platform_ioremap_resource() helper instead of
calling platform_get_resource() and devm_ioremap_resource()
separately

Signed-off-by: Cai Huoqing <caihuoqing@baidu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agofou: remove sparse errors
Eric Dumazet [Tue, 31 Aug 2021 03:26:08 +0000 (20:26 -0700)]
fou: remove sparse errors

We need to add __rcu qualifier to avoid these errors:

net/ipv4/fou.c:250:18: warning: incorrect type in assignment (different address spaces)
net/ipv4/fou.c:250:18:    expected struct net_offload const **offloads
net/ipv4/fou.c:250:18:    got struct net_offload const [noderef] __rcu **
net/ipv4/fou.c:251:15: error: incompatible types in comparison expression (different address spaces):
net/ipv4/fou.c:251:15:    struct net_offload const [noderef] __rcu *
net/ipv4/fou.c:251:15:    struct net_offload const *
net/ipv4/fou.c:272:18: warning: incorrect type in assignment (different address spaces)
net/ipv4/fou.c:272:18:    expected struct net_offload const **offloads
net/ipv4/fou.c:272:18:    got struct net_offload const [noderef] __rcu **
net/ipv4/fou.c:273:15: error: incompatible types in comparison expression (different address spaces):
net/ipv4/fou.c:273:15:    struct net_offload const [noderef] __rcu *
net/ipv4/fou.c:273:15:    struct net_offload const *
net/ipv4/fou.c:442:18: warning: incorrect type in assignment (different address spaces)
net/ipv4/fou.c:442:18:    expected struct net_offload const **offloads
net/ipv4/fou.c:442:18:    got struct net_offload const [noderef] __rcu **
net/ipv4/fou.c:443:15: error: incompatible types in comparison expression (different address spaces):
net/ipv4/fou.c:443:15:    struct net_offload const [noderef] __rcu *
net/ipv4/fou.c:443:15:    struct net_offload const *
net/ipv4/fou.c:489:18: warning: incorrect type in assignment (different address spaces)
net/ipv4/fou.c:489:18:    expected struct net_offload const **offloads
net/ipv4/fou.c:489:18:    got struct net_offload const [noderef] __rcu **
net/ipv4/fou.c:490:15: error: incompatible types in comparison expression (different address spaces):
net/ipv4/fou.c:490:15:    struct net_offload const [noderef] __rcu *
net/ipv4/fou.c:490:15:    struct net_offload const *
net/ipv4/udp_offload.c:170:26: warning: incorrect type in assignment (different address spaces)
net/ipv4/udp_offload.c:170:26:    expected struct net_offload const **offloads
net/ipv4/udp_offload.c:170:26:    got struct net_offload const [noderef] __rcu **
net/ipv4/udp_offload.c:171:23: error: incompatible types in comparison expression (different address spaces):
net/ipv4/udp_offload.c:171:23:    struct net_offload const [noderef] __rcu *
net/ipv4/udp_offload.c:171:23:    struct net_offload const *

Fixes: efc98d08e1ec ("fou: eliminate IPv4,v6 specific GRO functions")
Fixes: 8bce6d7d0d1e ("udp: Generalize skb_udp_segment")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoipv4: fix endianness issue in inet_rtm_getroute_build_skb()
Eric Dumazet [Tue, 31 Aug 2021 02:02:10 +0000 (19:02 -0700)]
ipv4: fix endianness issue in inet_rtm_getroute_build_skb()

The UDP length field should be in network order.
This removes the following sparse error:

net/ipv4/route.c:3173:27: warning: incorrect type in assignment (different base types)
net/ipv4/route.c:3173:27:    expected restricted __be16 [usertype] len
net/ipv4/route.c:3173:27:    got unsigned long

Fixes: 404eb77ea766 ("ipv4: support sport, dport and ip_proto in RTM_GETROUTE")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Roopa Prabhu <roopa@nvidia.com>
Cc: David Ahern <dsahern@kernel.org>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'octeon-npc-fixes'
David S. Miller [Tue, 31 Aug 2021 10:35:26 +0000 (11:35 +0100)]
Merge branch 'octeon-npc-fixes'

Subbaraya Sundeep says:

====================
octeontx2-af: Miscellaneous fixes in npc

This patchset consists of consolidated fixes in
rvu_npc file. Two of the patches prevent infinite
loop which can happen in corner cases. One patch
is for fixing static code analyzer reported issues.
And the last patch is a minor change in npc protocol
checker hardware block to report ipv4 checksum errors
as a known error codes.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-af: Set proper errorcode for IPv4 checksum errors
Sunil Goutham [Mon, 30 Aug 2021 18:00:46 +0000 (23:30 +0530)]
octeontx2-af: Set proper errorcode for IPv4 checksum errors

With current config, for packets with IPv4 checksum errors,
errorcode is being set to UNKNOWN. Hence added a separate
errorcodes for outer and inner IPv4 checksum and changed
NPC configuration accordingly.

Also turn on L2 multicast address check in NPC protocol check block.

Fixes: 6b3321bacc5a ("octeontx2-af: Enable packet length and csum validation")
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-af: Fix static code analyzer reported issues
Subbaraya Sundeep [Mon, 30 Aug 2021 18:00:45 +0000 (23:30 +0530)]
octeontx2-af: Fix static code analyzer reported issues

This patch fixes the static code analyzer reported issues
in rvu_npc.c. The reported errors are different sizes of
operands in bitops and returning uninitialized values.

Fixes: 651cd2652339 ("octeontx2-af: MCAM entry installation support")
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-af: Fix mailbox errors in nix_rss_flowkey_cfg
Subbaraya Sundeep [Mon, 30 Aug 2021 18:00:44 +0000 (23:30 +0530)]
octeontx2-af: Fix mailbox errors in nix_rss_flowkey_cfg

In npc_update_vf_flow_entry function the loop cursor
'index' is being changed inside the loop causing
the loop to spin forever. This in turn hogs the kworker
thread forever and no other mbox message is processed
by AF driver after that. Fix this by using
another variable in the loop.

Fixes: 55307fcb9258 ("octeontx2-af: Add mbox messages to install and delete MCAM rules")
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-af: Fix loop in free and unmap counter
Subbaraya Sundeep [Mon, 30 Aug 2021 18:00:43 +0000 (23:30 +0530)]
octeontx2-af: Fix loop in free and unmap counter

When the given counter does not belong to the entry
then code ends up in infinite loop because the loop
cursor, entry is not getting updated further. This
patch fixes that by updating entry for every iteration.

Fixes: a958dd59f9ce ("octeontx2-af: Map or unmap NPC MCAM entry and counter")
Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoaf_unix: fix potential NULL deref in unix_dgram_connect()
Eric Dumazet [Mon, 30 Aug 2021 17:21:37 +0000 (10:21 -0700)]
af_unix: fix potential NULL deref in unix_dgram_connect()

syzbot was able to trigger NULL deref in unix_dgram_connect() [1]

This happens in

if (unix_peer(sk))
sk->sk_state = other->sk_state = TCP_ESTABLISHED; // crash because @other is NULL

Because locks have been dropped, unix_peer() might be non NULL,
while @other is NULL (AF_UNSPEC case)

We need to move code around, so that we no longer access
unix_peer() and sk_state while locks have been released.

[1]
general protection fault, probably for non-canonical address 0xdffffc0000000002: 0000 [#1] PREEMPT SMP KASAN
KASAN: null-ptr-deref in range [0x0000000000000010-0x0000000000000017]
CPU: 0 PID: 10341 Comm: syz-executor239 Not tainted 5.14.0-rc7-syzkaller #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
RIP: 0010:unix_dgram_connect+0x32a/0xc60 net/unix/af_unix.c:1226
Code: 00 00 45 31 ed 49 83 bc 24 f8 05 00 00 00 74 69 e8 eb 5b a6 f9 48 8d 7d 12 48 b8 00 00 00 00 00 fc ff df 48 89 fa 48 c1 ea 03 <0f> b6 04 02 48 89 fa 83 e2 07 38 d0 7f 08 84 c0 0f 85 e0 07 00 00
RSP: 0018:ffffc9000a89fcd8 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: 0000000000000004 RCX: 0000000000000000
RDX: 0000000000000002 RSI: ffffffff87cf4ef5 RDI: 0000000000000012
RBP: 0000000000000000 R08: 0000000000000000 R09: ffff88802e1917c3
R10: ffffffff87cf4eba R11: 0000000000000001 R12: ffff88802e191740
R13: 0000000000000000 R14: ffff88802e191d38 R15: ffff88802e1917c0
FS:  00007f3eb0052700(0000) GS:ffff8880b9c00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000004787d0 CR3: 0000000029c0a000 CR4: 00000000001506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
 __sys_connect_file+0x155/0x1a0 net/socket.c:1890
 __sys_connect+0x161/0x190 net/socket.c:1907
 __do_sys_connect net/socket.c:1917 [inline]
 __se_sys_connect net/socket.c:1914 [inline]
 __x64_sys_connect+0x6f/0xb0 net/socket.c:1914
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x446a89
Code: 28 00 00 00 75 05 48 83 c4 28 c3 e8 a1 15 00 00 90 48 89 f8 48 89 f7 48 89 d6 48 89 ca 4d 89 c2 4d 89 c8 4c 8b 4c 24 08 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 bc ff ff ff f7 d8 64 89 01 48
RSP: 002b:00007f3eb0052208 EFLAGS: 00000246 ORIG_RAX: 000000000000002a
RAX: ffffffffffffffda RBX: 00000000004cc4d8 RCX: 0000000000446a89
RDX: 000000000000006e RSI: 0000000020000180 RDI: 0000000000000003
RBP: 00000000004cc4d0 R08: 00007f3eb0052700 R09: 0000000000000000
R10: 00007f3eb0052700 R11: 0000000000000246 R12: 00000000004cc4dc
R13: 00007ffd791e79cf R14: 00007f3eb0052300 R15: 0000000000022000
Modules linked in:
---[ end trace 4eb809357514968c ]---
RIP: 0010:unix_dgram_connect+0x32a/0xc60 net/unix/af_unix.c:1226
Code: 00 00 45 31 ed 49 83 bc 24 f8 05 00 00 00 74 69 e8 eb 5b a6 f9 48 8d 7d 12 48 b8 00 00 00 00 00 fc ff df 48 89 fa 48 c1 ea 03 <0f> b6 04 02 48 89 fa 83 e2 07 38 d0 7f 08 84 c0 0f 85 e0 07 00 00
RSP: 0018:ffffc9000a89fcd8 EFLAGS: 00010202
RAX: dffffc0000000000 RBX: 0000000000000004 RCX: 0000000000000000
RDX: 0000000000000002 RSI: ffffffff87cf4ef5 RDI: 0000000000000012
RBP: 0000000000000000 R08: 0000000000000000 R09: ffff88802e1917c3
R10: ffffffff87cf4eba R11: 0000000000000001 R12: ffff88802e191740
R13: 0000000000000000 R14: ffff88802e191d38 R15: ffff88802e1917c0
FS:  00007f3eb0052700(0000) GS:ffff8880b9d00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00007ffd791fe960 CR3: 0000000029c0a000 CR4: 00000000001506e0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400

Fixes: 83301b5367a9 ("af_unix: Set TCP_ESTABLISHED for datagram sockets too")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Cong Wang <cong.wang@bytedance.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agodpaa2-eth: Replace strlcpy with strscpy
Jason Wang [Mon, 30 Aug 2021 13:05:31 +0000 (21:05 +0800)]
dpaa2-eth: Replace strlcpy with strscpy

The strlcpy should not be used because it doesn't limit the source
length. As linus says, it's a completely useless function if you
can't implicitly trust the source string - but that is almost always
why people think they should use it! All in all the BSD function
will lead some potential bugs.

But the strscpy doesn't require reading memory from the src string
beyond the specified "count" bytes, and since the return value is
easier to error-check than strlcpy()'s. In addition, the implementation
is robust to the string changing out from underneath it, unlike the
current strlcpy() implementation.

Thus, We prefer using strscpy instead of strlcpy.

Signed-off-by: Jason Wang <wangborong@cdjrlc.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-af: Use NDC TX for transmit packet data
Geetha sowjanya [Mon, 30 Aug 2021 12:55:18 +0000 (18:25 +0530)]
octeontx2-af: Use NDC TX for transmit packet data

For better performance set hardware to use NDC TX for reading packet
data specified NIX_SEND_SG_S.

Signed-off-by: Geetha sowjanya <gakula@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: bridge: use mld2r_ngrec instead of icmpv6_dataun
MichelleJin [Sun, 29 Aug 2021 04:32:29 +0000 (04:32 +0000)]
net: bridge: use mld2r_ngrec instead of icmpv6_dataun

br_ip6_multicast_mld2_report function uses icmp6h
to parse mld2_report packet.

mld2r_ngrec defines mld2r_hdr.icmp6_dataun.un_data16[1]
in include/net/mld.h.

So, it is more compact to use mld2r rather than icmp6h.

By doing printk test, it is confirmed that
icmp6h->icmp6_dataun.un_data16[1] and mld2r->mld2r_ngrec are
indeed equivalent.

Also, sizeof(*mld2r) and sizeof(*icmp6h) are equivalent, too.

Signed-off-by: MichelleJin <shjy180909@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: qualcomm: fix QCA7000 checksum handling
Stefan Wahren [Sat, 28 Aug 2021 14:23:15 +0000 (16:23 +0200)]
net: qualcomm: fix QCA7000 checksum handling

Based on tests the QCA7000 doesn't support checksum offloading. So assume
ip_summed is CHECKSUM_NONE and let the kernel take care of the checksum
handling. This fixes data transfer issues in noisy environments.

Reported-by: Michael Heimpold <michael.heimpold@in-tech.com>
Fixes: 291ab06ecf67 ("net: qualcomm: new Ethernet over SPI driver for QCA7000")
Signed-off-by: Stefan Wahren <stefan.wahren@i2se.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: pasemi: Remove usage of the deprecated "pci-dma-compat.h" API
Christophe JAILLET [Sat, 28 Aug 2021 11:28:48 +0000 (13:28 +0200)]
net: pasemi: Remove usage of the deprecated "pci-dma-compat.h" API

In [1], Christoph Hellwig has proposed to remove the wrappers in
include/linux/pci-dma-compat.h.

Some reasons why this API should be removed have been given by Julia
Lawall in [2].

A coccinelle script has been used to perform the needed transformation
Only relevant parts are given below.

An 'unlikely()' has been removed when calling 'dma_mapping_error()' because
this function, which is inlined, already has such an annotation.

@@ @@
-    PCI_DMA_TODEVICE
+    DMA_TO_DEVICE

@@ @@
-    PCI_DMA_FROMDEVICE
+    DMA_FROM_DEVICE

@@
expression e1, e2, e3, e4;
@@
-    pci_map_single(e1, e2, e3, e4)
+    dma_map_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_single(e1, e2, e3, e4)
+    dma_unmap_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4, e5;
@@
-    pci_map_page(e1, e2, e3, e4, e5)
+    dma_map_page(&e1->dev, e2, e3, e4, e5)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_page(e1, e2, e3, e4)
+    dma_unmap_page(&e1->dev, e2, e3, e4)

@@
expression e1, e2;
@@
-    pci_dma_mapping_error(e1, e2)
+    dma_mapping_error(&e1->dev, e2)

[1]: https://lore.kernel.org/kernel-janitors/20200421081257.GA131897@infradead.org/
[2]: https://lore.kernel.org/kernel-janitors/alpine.DEB.2.22.394.2007120902170.2424@hadrien/

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Link: https://lore.kernel.org/r/bc6cd281eae024b26fd9c7ef6678d2d1dc9d74fd.1630150008.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agonet: sched: Fix qdisc_rate_table refcount leak when get tcf_block failed
Xiyu Yang [Sun, 29 Aug 2021 15:58:01 +0000 (23:58 +0800)]
net: sched: Fix qdisc_rate_table refcount leak when get tcf_block failed

The reference counting issue happens in one exception handling path of
cbq_change_class(). When failing to get tcf_block, the function forgets
to decrease the refcount of "rtab" increased by qdisc_put_rtab(),
causing a refcount leak.

Fix this issue by jumping to "failure" label when get tcf_block failed.

Fixes: 6529eaba33f0 ("net: sched: introduce tcf block infractructure")
Signed-off-by: Xiyu Yang <xiyuyang19@fudan.edu.cn>
Reviewed-by: Cong Wang <cong.wang@bytedance.com>
Link: https://lore.kernel.org/r/1630252681-71588-1-git-send-email-xiyuyang19@fudan.edu.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMerge https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next
Jakub Kicinski [Mon, 30 Aug 2021 23:42:46 +0000 (16:42 -0700)]
Merge https://git./linux/kernel/git/bpf/bpf-next

Daniel Borkmann says:

====================
bpf-next 2021-08-31

We've added 116 non-merge commits during the last 17 day(s) which contain
a total of 126 files changed, 6813 insertions(+), 4027 deletions(-).

The main changes are:

1) Add opaque bpf_cookie to perf link which the program can read out again,
   to be used in libbpf-based USDT library, from Andrii Nakryiko.

2) Add bpf_task_pt_regs() helper to access userspace pt_regs, from Daniel Xu.

3) Add support for UNIX stream type sockets for BPF sockmap, from Jiang Wang.

4) Allow BPF TCP congestion control progs to call bpf_setsockopt() e.g. to switch
   to another congestion control algorithm during init, from Martin KaFai Lau.

5) Extend BPF iterator support for UNIX domain sockets, from Kuniyuki Iwashima.

6) Allow bpf_{set,get}sockopt() calls from setsockopt progs, from Prankur Gupta.

7) Add bpf_get_netns_cookie() helper for BPF_PROG_TYPE_{SOCK_OPS,CGROUP_SOCKOPT}
   progs, from Xu Liu and Stanislav Fomichev.

8) Support for __weak typed ksyms in libbpf, from Hao Luo.

9) Shrink struct cgroup_bpf by 504 bytes through refactoring, from Dave Marchevsky.

10) Fix a smatch complaint in verifier's narrow load handling, from Andrey Ignatov.

11) Fix BPF interpreter's tail call count limit, from Daniel Borkmann.

12) Big batch of improvements to BPF selftests, from Magnus Karlsson, Li Zhijian,
    Yucong Sun, Yonghong Song, Ilya Leoshkevich, Jussi Maki, Ilya Leoshkevich, others.

13) Another big batch to revamp XDP samples in order to give them consistent look
    and feel, from Kumar Kartikeya Dwivedi.

* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next: (116 commits)
  MAINTAINERS: Remove self from powerpc BPF JIT
  selftests/bpf: Fix potential unreleased lock
  samples: bpf: Fix uninitialized variable in xdp_redirect_cpu
  selftests/bpf: Reduce more flakyness in sockmap_listen
  bpf: Fix bpf-next builds without CONFIG_BPF_EVENTS
  bpf: selftests: Add dctcp fallback test
  bpf: selftests: Add connect_to_fd_opts to network_helpers
  bpf: selftests: Add sk_state to bpf_tcp_helpers.h
  bpf: tcp: Allow bpf-tcp-cc to call bpf_(get|set)sockopt
  selftests: xsk: Preface options with opt
  selftests: xsk: Make enums lower case
  selftests: xsk: Generate packets from specification
  selftests: xsk: Generate packet directly in umem
  selftests: xsk: Simplify cleanup of ifobjects
  selftests: xsk: Decrease sending speed
  selftests: xsk: Validate tx stats on tx thread
  selftests: xsk: Simplify packet validation in xsk tests
  selftests: xsk: Rename worker_* functions that are not thread entry points
  selftests: xsk: Disassociate umem size with packets sent
  selftests: xsk: Remove end-of-test packet
  ...
====================

Link: https://lore.kernel.org/r/20210830225618.11634-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agosch_htb: Fix inconsistency when leaf qdisc creation fails
Maxim Mikityanskiy [Thu, 26 Aug 2021 11:54:25 +0000 (14:54 +0300)]
sch_htb: Fix inconsistency when leaf qdisc creation fails

In HTB offload mode, qdiscs of leaf classes are grafted to netdev
queues. sch_htb expects the dev_queue field of these qdiscs to point to
the corresponding queues. However, qdisc creation may fail, and in that
case noop_qdisc is used instead. Its dev_queue doesn't point to the
right queue, so sch_htb can lose track of used netdev queues, which will
cause internal inconsistencies.

This commit fixes this bug by keeping track of the netdev queue inside
struct htb_class. All reads of cl->leaf.q->dev_queue are replaced by the
new field, the two values are synced on writes, and WARNs are added to
assert equality of the two values.

The driver API has changed: when TC_HTB_LEAF_DEL needs to move a queue,
the driver used to pass the old and new queue IDs to sch_htb. Now that
there is a new field (offload_queue) in struct htb_class that needs to
be updated on this operation, the driver will pass the old class ID to
sch_htb instead (it already knows the new class ID).

Fixes: d03b195b5aa0 ("sch_htb: Hierarchical QoS hardware offload")
Signed-off-by: Maxim Mikityanskiy <maximmi@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://lore.kernel.org/r/20210826115425.1744053-1-maximmi@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
2 years agoMAINTAINERS: Remove self from powerpc BPF JIT
Sandipan Das [Fri, 27 Aug 2021 11:19:05 +0000 (16:49 +0530)]
MAINTAINERS: Remove self from powerpc BPF JIT

Stepping down as I haven't had a chance to look into the powerpc
BPF JIT compilers for a while.

Signed-off-by: Sandipan Das <sandipan@linux.ibm.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20210827111905.396145-1-sandipan@linux.ibm.com
2 years agonet: ipv4: Fix the warning for dereference
Yajun Deng [Mon, 30 Aug 2021 09:16:40 +0000 (17:16 +0800)]
net: ipv4: Fix the warning for dereference

Add a if statements to avoid the warning.

Dan Carpenter report:
The patch faf482ca196a: "net: ipv4: Move ip_options_fragment() out of
loop" from Aug 23, 2021, leads to the following Smatch complaint:

    net/ipv4/ip_output.c:833 ip_do_fragment()
    warn: variable dereferenced before check 'iter.frag' (see line 828)

Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Fixes: faf482ca196a ("net: ipv4: Move ip_options_fragment() out of loop")
Link: https://lore.kernel.org/netdev/20210830073802.GR7722@kadam/T/#t
Signed-off-by: Yajun Deng <yajun.deng@linux.dev>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: qrtr: make checks in qrtr_endpoint_post() stricter
Dan Carpenter [Mon, 30 Aug 2021 08:37:17 +0000 (11:37 +0300)]
net: qrtr: make checks in qrtr_endpoint_post() stricter

These checks are still not strict enough.  The main problem is that if
"cb->type == QRTR_TYPE_NEW_SERVER" is true then "len - hdrlen" is
guaranteed to be 4 but we need to be at least 16 bytes.  In fact, we
can reject everything smaller than sizeof(*pkt) which is 20 bytes.

Also I don't like the ALIGN(size, 4).  It's better to just insist that
data is needs to be aligned at the start.

Fixes: 0baa99ee353c ("net: qrtr: Allow non-immediate node routing")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agofix array-index-out-of-bounds in taprio_change
Haimin Zhang [Mon, 30 Aug 2021 03:47:01 +0000 (11:47 +0800)]
fix array-index-out-of-bounds in taprio_change

syzbot report an array-index-out-of-bounds in taprio_change
index 16 is out of range for type '__u16 [16]'
that's because mqprio->num_tc is lager than TC_MAX_QUEUE,so we check
the return value of netdev_set_num_tc.

Reported-by: syzbot+2b3e5fb6c7ef285a94f6@syzkaller.appspotmail.com
Signed-off-by: Haimin Zhang <tcs_kernel@tencent.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: fix NULL pointer reference in cipso_v4_doi_free
王贇 [Mon, 30 Aug 2021 10:28:01 +0000 (18:28 +0800)]
net: fix NULL pointer reference in cipso_v4_doi_free

In netlbl_cipsov4_add_std() when 'doi_def->map.std' alloc
failed, we sometime observe panic:

  BUG: kernel NULL pointer dereference, address:
  ...
  RIP: 0010:cipso_v4_doi_free+0x3a/0x80
  ...
  Call Trace:
   netlbl_cipsov4_add_std+0xf4/0x8c0
   netlbl_cipsov4_add+0x13f/0x1b0
   genl_family_rcv_msg_doit.isra.15+0x132/0x170
   genl_rcv_msg+0x125/0x240

This is because in cipso_v4_doi_free() there is no check
on 'doi_def->map.std' when doi_def->type got value 1, which
is possibe, since netlbl_cipsov4_add_std() haven't initialize
it before alloc 'doi_def->map.std'.

This patch just add the check to prevent panic happen in similar
cases.

Reported-by: Abaci <abaci@linux.alibaba.com>
Signed-off-by: Michael Wang <yun.wang@linux.alibaba.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'inet-exceptions-less-predictable'
David S. Miller [Mon, 30 Aug 2021 11:21:38 +0000 (12:21 +0100)]
Merge branch 'inet-exceptions-less-predictable'

Eric Dumazet says:

====================
inet: make exception handling less predictible

This second round of patches is addressing Keyu Man recommendations
to make linux hosts more robust against a class of brute force attacks.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoipv4: make exception cache less predictible
Eric Dumazet [Sun, 29 Aug 2021 22:16:15 +0000 (15:16 -0700)]
ipv4: make exception cache less predictible

Even after commit 6457378fe796 ("ipv4: use siphash instead of Jenkins in
fnhe_hashfun()"), an attacker can still use brute force to learn
some secrets from a victim linux host.

One way to defeat these attacks is to make the max depth of the hash
table bucket a random value.

Before this patch, each bucket of the hash table used to store exceptions
could contain 6 items under attack.

After the patch, each bucket would contains a random number of items,
between 6 and 10. The attacker can no longer infer secrets.

This is slightly increasing memory size used by the hash table,
by 50% in average, we do not expect this to be a problem.

This patch is more complex than the prior one (IPv6 equivalent),
because IPv4 was reusing the oldest entry.
Since we need to be able to evict more than one entry per
update_or_create_fnhe() call, I had to replace
fnhe_oldest() with fnhe_remove_oldest().

Also note that we will queue extra kfree_rcu() calls under stress,
which hopefully wont be a too big issue.

Fixes: 4895c771c7f0 ("ipv4: Add FIB nexthop exceptions.")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Keyu Man <kman001@ucr.edu>
Cc: Willy Tarreau <w@1wt.eu>
Signed-off-by: David S. Miller <davem@davemloft.net>
Reviewed-by: David Ahern <dsahern@kernel.org>
Tested-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoipv6: make exception cache less predictible
Eric Dumazet [Sun, 29 Aug 2021 22:16:14 +0000 (15:16 -0700)]
ipv6: make exception cache less predictible

Even after commit 4785305c05b2 ("ipv6: use siphash in rt6_exception_hash()"),
an attacker can still use brute force to learn some secrets from a victim
linux host.

One way to defeat these attacks is to make the max depth of the hash
table bucket a random value.

Before this patch, each bucket of the hash table used to store exceptions
could contain 6 items under attack.

After the patch, each bucket would contains a random number of items,
between 6 and 10. The attacker can no longer infer secrets.

This is slightly increasing memory size used by the hash table,
we do not expect this to be a problem.

Following patch is dealing with the same issue in IPv4.

Fixes: 35732d01fe31 ("ipv6: introduce a hash table to store dst cache")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Keyu Man <kman001@ucr.edu>
Cc: Wei Wang <weiwan@google.com>
Cc: Martin KaFai Lau <kafai@fb.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/pablo/nf-next
David S. Miller [Mon, 30 Aug 2021 09:57:54 +0000 (10:57 +0100)]
Merge git://git./linux/kernel/git/pablo/nf-next

Pablo Neira Ayuso says:

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

The following patchset contains Netfilter updates for net-next:

1) Clean up and consolidate ct ecache infrastructure by merging ct and
   expect notifiers, from Florian Westphal.

2) Missing counters and timestamp in nfnetlink_queue and _log conntrack
   information.

3) Missing error check for xt_register_template() in iptables mangle,
   as a incremental fix for the previous pull request, also from
   Florian Westphal.

4) Add netfilter hooks for the SRv6 lightweigh tunnel driver, from
   Ryoga Sato. The hooks are enabled via nf_hooks_lwtunnel sysctl
   to make sure existing netfilter rulesets do not break. There is
   a static key to disable the hooks by default.

   The pktgen_bench_xmit_mode_netif_receive.sh shows no noticeable
   impact in the seg6_input path for non-netfilter users: similar
   numbers with and without this patch.

   This is a sample of the perf report output:

    11.67%  kpktgend_0       [ipv6]                    [k] ipv6_get_saddr_eval
     7.89%  kpktgend_0       [ipv6]                    [k] __ipv6_addr_label
     7.52%  kpktgend_0       [ipv6]                    [k] __ipv6_dev_get_saddr
     6.63%  kpktgend_0       [kernel.vmlinux]          [k] asm_exc_nmi
     4.74%  kpktgend_0       [ipv6]                    [k] fib6_node_lookup_1
     3.48%  kpktgend_0       [kernel.vmlinux]          [k] pskb_expand_head
     3.33%  kpktgend_0       [ipv6]                    [k] ip6_rcv_core.isra.29
     3.33%  kpktgend_0       [ipv6]                    [k] seg6_do_srh_encap
     2.53%  kpktgend_0       [ipv6]                    [k] ipv6_dev_get_saddr
     2.45%  kpktgend_0       [ipv6]                    [k] fib6_table_lookup
     2.24%  kpktgend_0       [kernel.vmlinux]          [k] ___cache_free
     2.16%  kpktgend_0       [ipv6]                    [k] ip6_pol_route
     2.11%  kpktgend_0       [kernel.vmlinux]          [k] __ipv6_addr_type
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'IXP46x-PTP-Timer'
David S. Miller [Mon, 30 Aug 2021 08:59:11 +0000 (09:59 +0100)]
Merge branch 'IXP46x-PTP-Timer'

Linus Walleij says:

====================
IXP46x PTP Timer clean-up and DT

ChangeLog v2->v3:

- Dropped the patch enabling compile tests: we are still dependent
  on some machine-specific headers. The plan is to get rid of this
  after device tree conversion. We include one of the compile testing
  fixes anyway, because it is nice to have fixed.

- Rebased on the latest net-next
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoixp4xx_eth: Probe the PTP module from the device tree
Linus Walleij [Sat, 28 Aug 2021 17:15:48 +0000 (19:15 +0200)]
ixp4xx_eth: Probe the PTP module from the device tree

This adds device tree probing support for the PTP module
adjacent to the ethernet module. It is pretty straight
forward, all resources are in the device tree as they
come to the platform device.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoixp4xx_eth: Add devicetree bindings
Linus Walleij [Sat, 28 Aug 2021 17:15:47 +0000 (19:15 +0200)]
ixp4xx_eth: Add devicetree bindings

This adds device tree bindings for the IXP46x PTP Timer, a companion
to the IXP4xx ethernet in newer platforms.

Cc: devicetree@vger.kernel.org
Cc: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoixp4xx_eth: Stop referring to GPIOs
Linus Walleij [Sat, 28 Aug 2021 17:15:46 +0000 (19:15 +0200)]
ixp4xx_eth: Stop referring to GPIOs

The driver is being passed interrupts, then looking up the
same interrupts as GPIOs a second time to convert them into
interrupts and set properties on them.

This is pointless: the GPIO and irqchip APIs of a GPIO chip
are orthogonal. Just request the interrupts and be done
with it, drop reliance on any GPIO functions or definitions.

Use devres-managed functions and add a small devress quirk
to unregister the clock as well and we can rely on devres
to handle all the resources and cut down a bunch of
boilerplate in the process.

Cc: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoixp4xx_eth: fix compile-testing
Arnd Bergmann [Sat, 28 Aug 2021 17:15:45 +0000 (19:15 +0200)]
ixp4xx_eth: fix compile-testing

Change the driver to use portable integer types to avoid warnings
during compile testing, including:

drivers/net/ethernet/xscale/ixp4xx_eth.c:721:21: error: cast to 'u32 *' (aka 'unsigned int *') from smaller integer type 'int' [-Werror,-Wint-to-pointer-cast]
        memcpy_swab32(mem, (u32 *)((int)skb->data & ~3), bytes / 4);
                           ^
drivers/net/ethernet/xscale/ixp4xx_eth.c:963:12: error: incompatible pointer types passing 'u32 *' (aka 'unsigned int *') to parameter of type 'dma_addr_t *' (aka 'unsigned long long *') [-Werror,-Wincompatible-pointer-types]
                                              &port->desc_tab_phys)))
                                              ^~~~~~~~~~~~~~~~~~~~
include/linux/dmapool.h:27:20: note: passing argument to parameter 'handle' here
                     dma_addr_t *handle);
                                 ^

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoixp4xx_eth: make ptp support a platform driver
Arnd Bergmann [Sat, 28 Aug 2021 17:15:44 +0000 (19:15 +0200)]
ixp4xx_eth: make ptp support a platform driver

After the recent ixp4xx cleanups, the ptp driver has gained a
build failure in some configurations:

drivers/net/ethernet/xscale/ptp_ixp46x.c: In function 'ptp_ixp_init':
drivers/net/ethernet/xscale/ptp_ixp46x.c:290:51: error: 'IXP4XX_TIMESYNC_BASE_VIRT' undeclared (first use in this function)

Avoid the last bit of hardcoded constants from platform headers
by turning the ptp driver bit into a platform driver and passing
the IRQ and MMIO address as resources.

This is a bit tricky:

- The interface between the two drivers is now the new
  ixp46x_ptp_find() function, replacing the global
  ixp46x_phc_index variable. The call is done as late
  as possible, in hwtstamp_set(), to ensure that the
  ptp device is fully probed.

- As the ptp driver is now called by the network driver, the
  link dependency is reversed, which in turn requires a small
  Makefile hack

- The GPIO number is still left hardcoded. This is clearly not
  great, but it can be addressed later. Note that commit 98ac0cc270b7
  ("ARM: ixp4xx: Convert to MULTI_IRQ_HANDLER") changed the
  IRQ number to something meaningless. Passing the correct IRQ
  in a resource fixes this.

- When the PTP driver is disabled, ethtool .get_ts_info()
  now correctly lists only software timestamping regardless
  of the hardware.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
[Fix a missing include]
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'hns3-cleanups'
David S. Miller [Mon, 30 Aug 2021 08:43:19 +0000 (09:43 +0100)]
Merge branch 'hns3-cleanups'

Guangbin Huang says:

====================
net: hns3: add some cleanups

This series includes some cleanups for the HNS3 ethernet driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: uniform parameter name of hclge_ptp_clean_tx_hwts()
Hao Chen [Mon, 30 Aug 2021 06:06:42 +0000 (14:06 +0800)]
net: hns3: uniform parameter name of hclge_ptp_clean_tx_hwts()

The parameter name of hclge_ptp_clean_tx_hwts() in declaration is "dev",
but the definition of this function is used the common name "hdev" as
other functions, so modify it.

Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hnss3: use max() to simplify code
Hao Chen [Mon, 30 Aug 2021 06:06:41 +0000 (14:06 +0800)]
net: hnss3: use max() to simplify code

Replace the "? :" statement wich max() to simplify code.

Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: modify a print format of hns3_dbg_queue_map()
Hao Chen [Mon, 30 Aug 2021 06:06:40 +0000 (14:06 +0800)]
net: hns3: modify a print format of hns3_dbg_queue_map()

The type of tqp_vector->vector_irq is int, so modify its print format
to "%d".

Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: refine function hclge_dbg_dump_tm_pri()
Guangbin Huang [Mon, 30 Aug 2021 06:06:39 +0000 (14:06 +0800)]
net: hns3: refine function hclge_dbg_dump_tm_pri()

To improve flexibility, simplicity and maintainability to dump info of
every element of tm priority, add a struct hclge_dbg_item array of tm
priority and fill string of every data according to this array.

Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: reconstruct function hclge_ets_validate()
Guangbin Huang [Mon, 30 Aug 2021 06:06:38 +0000 (14:06 +0800)]
net: hns3: reconstruct function hclge_ets_validate()

This patch reconstructs function hclge_ets_validate() to reduce the code
cycle complexity and make code more concise.

Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: reconstruct function hns3_self_test
Peng Li [Mon, 30 Aug 2021 06:06:37 +0000 (14:06 +0800)]
net: hns3: reconstruct function hns3_self_test

This patch reconstructs function hns3_self_test to reduce the code
cycle complexity and make code more concise.

Signed-off-by: Peng Li <lipeng321@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: initialize each member of structure array on a separate line
Jiaran Zhang [Mon, 30 Aug 2021 06:06:36 +0000 (14:06 +0800)]
net: hns3: initialize each member of structure array on a separate line

To make the format of each member initialization of structure array
clearer, initialize each member on a separate line.

Signed-off-by: Jiaran Zhang <zhangjiaran@huawei.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'bnxt_en-fw-messages'
David S. Miller [Mon, 30 Aug 2021 08:35:05 +0000 (09:35 +0100)]
Merge branch 'bnxt_en-fw-messages'

Michael Chan says:

====================
bnxt_en: Implement new driver APIs to send FW messages

The current driver APIs to send messages to the firmware allow only one
outstanding message in flight.  There is only one buffer for the firmware
response for each firmware channel.  To send a firmware message, all
callers must take a mutex and it is released after the firmware response
has been read.  This scheme does not allow multiple firmware messages
in flight.  Firmware may take a long time to respond to some messages
(e.g. NVRAM related ones) and this causes the mutex to be held for
a long time, blocking other callers.

This patchset intoduces the new driver APIs to address the above
shortcomings.  The new APIs are compatible with new and old firmware.
But the new deferred firmware response mechanism will require newer
firmware in order to allow multiple outstanding firmware commands.

All callers are updated to use the new APIs.

v2: Patch 4 and patch 9 updated to fix issues reported by test robot
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: support multiple HWRM commands in flight
Edwin Peer [Sun, 29 Aug 2021 07:35:06 +0000 (03:35 -0400)]
bnxt_en: support multiple HWRM commands in flight

Add infrastructure to maintain a pending list of HWRM commands awaiting
completion and reduce the scope of the hwrm_cmd_lock mutex so that it
protects only the request mailbox. The mailbox is free to use for one
or more concurrent commands after receiving deferred response events.

For uniformity and completeness, use the same pending list for
collecting completions for commands that respond via a completion ring.
These commands are only used for freeing rings and for IRQ test and
we only support one such command in flight.

Note deferred responses are also only supported on the main channel.
The secondary channel (KONG) does not support deferred responses.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: remove legacy HWRM interface
Edwin Peer [Sun, 29 Aug 2021 07:35:05 +0000 (03:35 -0400)]
bnxt_en: remove legacy HWRM interface

There are no longer any callers relying on the old API.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: update all firmware calls to use the new APIs
Edwin Peer [Sun, 29 Aug 2021 07:35:04 +0000 (03:35 -0400)]
bnxt_en: update all firmware calls to use the new APIs

The conversion follows this general pattern for most of the calls:

1. The input message is changed from a stack variable initialized
using bnxt_hwrm_cmd_hdr_init() to a pointer allocated and intialized
using hwrm_req_init().

2. If we don't need to read the firmware response, the hwrm_send_message()
call is replaced with hwrm_req_send().

3. If we need to read the firmware response, the mutex lock is replaced
by hwrm_req_hold() to hold the response.  When the response is read, the
mutex unlock is replaced by hwrm_req_drop().

If additional DMA buffers are needed for firmware response data, the
hwrm_req_dma_slice() is used instead of calling dma_alloc_coherent().

Some minor refactoring is also done while doing these conversions.

v2: Fix unintialized variable warnings in __bnxt_hwrm_get_tx_rings()
and bnxt_approve_mac()

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: use link_lock instead of hwrm_cmd_lock to protect link_info
Edwin Peer [Sun, 29 Aug 2021 07:35:03 +0000 (03:35 -0400)]
bnxt_en: use link_lock instead of hwrm_cmd_lock to protect link_info

We currently use the hwrm_cmd_lock to serialize the update of the
firmware's link status response data and the copying of link status data
to the VF.  This won't work when we update the firmware message APIs, so
we use the link_lock mutex instead.  All link_info data should be
updated under the link_lock mutex.  Also add link_lock to functions that
touch link_info in __bnxt_open_nic() and bnxt_probe_phy(). The locking
is probably not strictly necessary during probe, but it's more consistent.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: add support for HWRM request slices
Edwin Peer [Sun, 29 Aug 2021 07:35:02 +0000 (03:35 -0400)]
bnxt_en: add support for HWRM request slices

Slices are a mechanism for suballocating DMA mapped regions from the
request buffer. Such regions can be used for indirect command data
instead of creating new mappings with dma_alloc_coherent().

The advantage of using a slice is that the lifetime of the slice is
bound to the request and will be automatically unmapped when the
request is consumed.

A single external region is also supported. This allows for regions
that will not fit inside the spare request buffer space such that
the same API can be used consistently even for larger mappings.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: add HWRM request assignment API
Edwin Peer [Sun, 29 Aug 2021 07:35:01 +0000 (03:35 -0400)]
bnxt_en: add HWRM request assignment API

hwrm_req_replace() provides an assignment like operation to replace a
managed HWRM request object with data from a pre-built source. This is
useful for handling request data provided by higher layer HWRM clients.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: discard out of sequence HWRM responses
Edwin Peer [Sun, 29 Aug 2021 07:35:00 +0000 (03:35 -0400)]
bnxt_en: discard out of sequence HWRM responses

During firmware crash recovery, it is possible for firmware to respond
to stale HWRM commands that have already timed out. Because response
buffers may be reused, any out of sequence responses need to be ignored
and only the matching seq_id should be accepted.

Also, READ_ONCE should be used for the reads from the DMA buffer to
ensure that the necessary loads are scheduled.

Reviewed-by: Scott Branden <scott.branden@broadcom.com>
Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: introduce new firmware message API based on DMA pools
Edwin Peer [Sun, 29 Aug 2021 07:34:59 +0000 (03:34 -0400)]
bnxt_en: introduce new firmware message API based on DMA pools

This change constitutes a major step towards supporting multiple
firmware commands in flight by maintaining a separate response buffer
for the duration of each request. These firmware commands are also
known as Hardware Resource Manager (HWRM) commands.  Using separate
response buffers requires an API change in order for callers to be
able to free the buffer when done.

It is impossible to keep the existing APIs unchanged.  The existing
usage for a simple HWRM message request such as the following:

        struct input req = {0};
        bnxt_hwrm_cmd_hdr_init(bp, &req, REQ_TYPE, -1, -1);
        rc = hwrm_send_message(bp, &req, sizeof(req), HWRM_CMD_TIMEOUT);
        if (rc)
                /* error */

changes to:

         struct input *req;
         rc = hwrm_req_init(bp, req, REQ_TYPE);
         if (rc)
                 /* error */
         rc = hwrm_req_send(bp, req); /* consumes req */
         if (rc)
                 /* error */

The key changes are:

1. The req is no longer allocated on the stack.
2. The caller must call hwrm_req_init() to allocate a req buffer and
   check for a valid buffer.
3. The req buffer is automatically released when hwrm_req_send() returns.
4. If the caller wants to check the firmware response, the caller must
   call hwrm_req_hold() to take ownership of the response buffer and
   release it afterwards using hwrm_req_drop().  The caller is no longer
   required to explicitly hold the hwrm_cmd_lock mutex to read the
   response.
5. Because the firmware commands and responses all have different sizes,
   some safeguards are added to the code.

This patch maintains legacy API compatibiltiy, implementing the old
API in terms of the new.  The follow-on patches will convert all
callers to use the new APIs.

v2: Fix redefined writeq with parisc .config
    Fix "cast from pointer to integer of different size" warning in
hwrm_calc_sentinel()

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: move HWRM API implementation into separate file
Edwin Peer [Sun, 29 Aug 2021 07:34:58 +0000 (03:34 -0400)]
bnxt_en: move HWRM API implementation into separate file

Move all firmware messaging functions and definitions to new
bnxt_hwrm.[ch].  The follow-on patches will make major modifications
to these APIs.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: Refactor the HWRM_VER_GET firmware calls
Edwin Peer [Sun, 29 Aug 2021 07:34:57 +0000 (03:34 -0400)]
bnxt_en: Refactor the HWRM_VER_GET firmware calls

Refactor the code so that __bnxt_hwrm_ver_get() does not call
bnxt_hwrm_do_send_msg() directly.  The new APIs will not expose this
internal call.  Add a new bnxt_hwrm_poll() to poll the HWRM_VER_GET
firmware call silently.  The other bnxt_hwrm_ver_get() function will
send the HWRM_VER_GET message directly with error logs enabled.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agobnxt_en: remove DMA mapping for KONG response
Edwin Peer [Sun, 29 Aug 2021 07:34:56 +0000 (03:34 -0400)]
bnxt_en: remove DMA mapping for KONG response

The additional response buffer serves no useful purpose. There can
be only one firmware command in flight due to the hwrm_cmd_lock mutex,
which is taken for the entire duration of any command completion,
KONG or otherwise. It is thus safe to share a single DMA buffer.

Removing the code associated with the additional mapping will simplify
matters in the next patch, which allocates response buffers from DMA
pools on a per request basis.

Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonetfilter: add netfilter hooks to SRv6 data plane
Ryoga Saito [Tue, 17 Aug 2021 08:39:37 +0000 (08:39 +0000)]
netfilter: add netfilter hooks to SRv6 data plane

This patch introduces netfilter hooks for solving the problem that
conntrack couldn't record both inner flows and outer flows.

This patch also introduces a new sysctl toggle for enabling lightweight
tunnel netfilter hooks.

Signed-off-by: Ryoga Saito <contact@proelbtn.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
2 years agoMerge tag 'wireless-drivers-next-2021-08-29' of git://git.kernel.org/pub/scm/linux...
David S. Miller [Sun, 29 Aug 2021 14:11:16 +0000 (15:11 +0100)]
Merge tag 'wireless-drivers-next-2021-08-29' of git://git./linux/kernel/git/kvalo/wireless-drivers-next

Kalle Valo says:

====================
pull-request: wireless-drivers-next-2021-08-29

here's a pull request to net-next tree, more info below. Please let me know if
there are any problems.

wireless-drivers-next patches for v5.15

Second, and most likely last, set of patches for v5.15. Lots of
iwlwifi patches this time, but smaller changes to other drivers as
well. Nothing special standing out.

Major changes:

rtw88

* add quirk to disable pci caps on HP Pavilion 14-ce0xxx

brcmfmac

* Add WPA3 Personal with FT to supported cipher suites

wcn36xx

* allow firmware name to be overridden by DT

iwlwifi

* support scanning hidden 6GHz networks

* support for a new hardware family (Bz)

* support for new firmware API versions

mwifiex

* add reset_d3cold quirk for Surface gen4+ devices
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agossb: fix boolreturn.cocci warning
Jing Yangyang [Tue, 24 Aug 2021 06:13:41 +0000 (23:13 -0700)]
ssb: fix boolreturn.cocci warning

./include/linux/ssb/ssb_driver_extif.h:200:8-9:WARNING: return of 0/1 in
function 'ssb_extif_available' with return type bool

Return statements in functions returning bool should use true/false
instead of 1/0.

Generated by: scripts/coccinelle/misc/boolreturn.cocci

Reported-by: Zeal Robot <zealci@zte.com.cn>
Signed-off-by: Jing Yangyang <jing.yangyang@zte.com.cn>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210824061341.59255-1-deng.changcheng@zte.com.cn
2 years agointel: switch from 'pci_' to 'dma_' API
Christophe JAILLET [Sun, 22 Aug 2021 08:03:50 +0000 (10:03 +0200)]
intel: switch from 'pci_' to 'dma_' API

The wrappers in include/linux/pci-dma-compat.h should go away.

The patch has been generated with the coccinelle script below.

It has been hand modified to use 'dma_set_mask_and_coherent()' instead of
'pci_set_dma_mask()/pci_set_consistent_dma_mask()' when applicable.
This is less verbose.

It has been compile tested.

@@
@@
-    PCI_DMA_BIDIRECTIONAL
+    DMA_BIDIRECTIONAL

@@
@@
-    PCI_DMA_TODEVICE
+    DMA_TO_DEVICE

@@
@@
-    PCI_DMA_FROMDEVICE
+    DMA_FROM_DEVICE

@@
@@
-    PCI_DMA_NONE
+    DMA_NONE

@@
expression e1, e2, e3;
@@
-    pci_alloc_consistent(e1, e2, e3)
+    dma_alloc_coherent(&e1->dev, e2, e3, GFP_)

@@
expression e1, e2, e3;
@@
-    pci_zalloc_consistent(e1, e2, e3)
+    dma_alloc_coherent(&e1->dev, e2, e3, GFP_)

@@
expression e1, e2, e3, e4;
@@
-    pci_free_consistent(e1, e2, e3, e4)
+    dma_free_coherent(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_map_single(e1, e2, e3, e4)
+    dma_map_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_single(e1, e2, e3, e4)
+    dma_unmap_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4, e5;
@@
-    pci_map_page(e1, e2, e3, e4, e5)
+    dma_map_page(&e1->dev, e2, e3, e4, e5)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_page(e1, e2, e3, e4)
+    dma_unmap_page(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_map_sg(e1, e2, e3, e4)
+    dma_map_sg(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_sg(e1, e2, e3, e4)
+    dma_unmap_sg(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_single_for_cpu(e1, e2, e3, e4)
+    dma_sync_single_for_cpu(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_single_for_device(e1, e2, e3, e4)
+    dma_sync_single_for_device(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_sg_for_cpu(e1, e2, e3, e4)
+    dma_sync_sg_for_cpu(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_dma_sync_sg_for_device(e1, e2, e3, e4)
+    dma_sync_sg_for_device(&e1->dev, e2, e3, e4)

@@
expression e1, e2;
@@
-    pci_dma_mapping_error(e1, e2)
+    dma_mapping_error(&e1->dev, e2)

@@
expression e1, e2;
@@
-    pci_set_dma_mask(e1, e2)
+    dma_set_mask(&e1->dev, e2)

@@
expression e1, e2;
@@
-    pci_set_consistent_dma_mask(e1, e2)
+    dma_set_coherent_mask(&e1->dev, e2)

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/f55043d0c847bfae60087707778563cf732a7bf9.1629619229.git.christophe.jaillet@wanadoo.fr
2 years agomwifiex: pcie: add reset_d3cold quirk for Surface gen4+ devices
Tsuchiya Yuto [Fri, 20 Aug 2021 14:20:50 +0000 (16:20 +0200)]
mwifiex: pcie: add reset_d3cold quirk for Surface gen4+ devices

To reset mwifiex on Surface gen4+ (Pro 4 or later gen) devices, it
seems that putting the wifi device into D3cold is required according
to errata.inf file on Windows installation (Windows/INF/errata.inf).

This patch adds a function that performs power-cycle (put into D3cold
then D0) and call the function at the end of reset_prepare().

Note: Need to also reset the parent device (bridge) of wifi on SB1;
it might be because the bridge of wifi always reports it's in D3hot.
When I tried to reset only the wifi device (not touching parent), it gave
the following error and the reset failed:

    acpi device:4b: Cannot transition to power state D0 for parent in D3hot
    mwifiex_pcie 0000:03:00.0: can't change power state from D3cold to D0 (config space inaccessible)

Signed-off-by: Tsuchiya Yuto <kitakar@gmail.com>
Signed-off-by: Jonas Dreßler <verdre@v0yd.nl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210820142050.35741-3-verdre@v0yd.nl
2 years agomwifiex: pcie: add DMI-based quirk implementation for Surface devices
Jonas Dreßler [Fri, 20 Aug 2021 14:20:49 +0000 (16:20 +0200)]
mwifiex: pcie: add DMI-based quirk implementation for Surface devices

This commit adds the ability to apply device-specific quirks to the
mwifiex driver. It uses DMI matching similar to the quirks brcmfmac uses
with dmi.c. We'll add identifiers to match various MS Surface devices,
which this is primarily meant for, later.

This commit is a slightly modified version of a previous patch sent in
by Tsuchiya Yuto.

Co-developed-by: Tsuchiya Yuto <kitakar@gmail.com>
Signed-off-by: Tsuchiya Yuto <kitakar@gmail.com>
Signed-off-by: Jonas Dreßler <verdre@v0yd.nl>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210820142050.35741-2-verdre@v0yd.nl
2 years agobrcmfmac: pcie: fix oops on failure to resume and reprobe
Ahmad Fatoum [Tue, 17 Aug 2021 06:35:22 +0000 (08:35 +0200)]
brcmfmac: pcie: fix oops on failure to resume and reprobe

When resuming from suspend, brcmf_pcie_pm_leave_D3 will first attempt a
hot resume and then fall back to removing the PCI device and then
reprobing. If this probe fails, the kernel will oops, because brcmf_err,
which is called to report the failure will dereference the stale bus
pointer. Open code and use the default bus-less brcmf_err to avoid this.

Fixes: 8602e62441ab ("brcmfmac: pass bus to the __brcmf_err() in pcie.c")
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210817063521.22450-1-a.fatoum@pengutronix.de
2 years agobcma: Drop the unused parameter of bcma_scan_read32()
Zenghui Yu [Tue, 27 Jul 2021 02:52:32 +0000 (10:52 +0800)]
bcma: Drop the unused parameter of bcma_scan_read32()

As it had never been used since the initial commit 8369ae33b705 ("bcma: add
Broadcom specific AMBA bus driver").

Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210727025232.663-3-yuzenghui@huawei.com
2 years agobcma: Fix memory leak for internally-handled cores
Zenghui Yu [Tue, 27 Jul 2021 02:52:31 +0000 (10:52 +0800)]
bcma: Fix memory leak for internally-handled cores

kmemleak reported that dev_name() of internally-handled cores were leaked
on driver unbinding. Let's use device_initialize() to take refcounts for
them and put_device() to properly free the related stuff.

While looking at it, there's another potential issue for those which should
be *registered* into driver core. If device_register() failed, we put
device once and freed bcma_device structures. In bcma_unregister_cores(),
they're treated as unregistered and we hit both UAF and double-free. That
smells not good and has also been fixed now.

Fixes: ab54bc8460b5 ("bcma: fill core details for every device")
Signed-off-by: Zenghui Yu <yuzenghui@huawei.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210727025232.663-2-yuzenghui@huawei.com
2 years agoMerge commit 'e257d969f36503b8eb1240f32653a1afb3109f86' of git://git.kernel.org/pub...
Kalle Valo [Sun, 29 Aug 2021 10:20:48 +0000 (13:20 +0300)]
Merge commit 'e257d969f36503b8eb1240f32653a1afb3109f86' of git://git./linux/kernel/git/iwlwifi/iwlwifi-next

I think last commits in tag iwlwifi-next-for-kalle-2021-08-26 are not
ready yet so I'm skipping those and pulling an earlier commit. I
modified Luca's description below to not include the skipped commits.

iwlwifi patches for v5.15

* Support scanning hidden 6GHz networks;
* Some improvements in the FW error dumps;
* Add some HE capability flags
* A bunch of janitorial clean-ups;
* Clean-ups in the TX code;
* Small fix for SMPS;
* Support for a new hardware family (Bz);
* Small fix in the scan code;
* A bunch of changes in the D3 code, including new FW API;
* Finalize the refactoring of 6GHz scan;
* Initial changes in the SAR profile code;
* Fix reading one of our ACPI tables (WTAS);
* Support some new ACPI table revisions;
* Support new API of the WoWlan status FW notification;
* Fixes in SAR ACPI tables handling;
* Some debugging improvements;
* Fix in ROC;
* Support for new FW API versions;
* Support new FW command versions;
* Some other small fixes, clean-ups and improvements.

2 years agovirtio_net: reduce raw_smp_processor_id() calling in virtnet_xdp_get_sq
Li RongQing [Thu, 26 Aug 2021 08:21:35 +0000 (16:21 +0800)]
virtio_net: reduce raw_smp_processor_id() calling in virtnet_xdp_get_sq

smp_processor_id()/raw* will be called once each when not
more queues in virtnet_xdp_get_sq() which is called in
non-preemptible context, so it's safe to call the function
smp_processor_id() once.

Signed-off-by: Li RongQing <lirongqing@baidu.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoniu: switch from 'pci_' to 'dma_' API
Christophe JAILLET [Fri, 27 Aug 2021 20:06:37 +0000 (22:06 +0200)]
niu: switch from 'pci_' to 'dma_' API

In [1], Christoph Hellwig has proposed to remove the wrappers in
include/linux/pci-dma-compat.h.

Some reasons why this API should be removed have been given by Julia
Lawall in [2].

A coccinelle script has been used to perform the needed transformation
Only relevant parts are given below.

It has been hand modified to use 'dma_set_mask_and_coherent()' instead of
'pci_set_dma_mask()/pci_set_consistent_dma_mask()' when applicable.
This is less verbose.

Finally, the now useless 'dma_mask' variable has been removed.

It has been compile tested.

@@
expression e1, e2;
@@
-    pci_set_dma_mask(e1, e2)
+    dma_set_mask(&e1->dev, e2)

@@
expression e1, e2;
@@
-    pci_set_consistent_dma_mask(e1, e2)
+    dma_set_coherent_mask(&e1->dev, e2)

[1]: https://lore.kernel.org/kernel-janitors/20200421081257.GA131897@infradead.org/
[2]: https://lore.kernel.org/kernel-janitors/alpine.DEB.2.22.394.2007120902170.2424@hadrien/

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agofddi: switch from 'pci_' to 'dma_' API
Christophe JAILLET [Fri, 27 Aug 2021 20:00:57 +0000 (22:00 +0200)]
fddi: switch from 'pci_' to 'dma_' API

In [1], Christoph Hellwig has proposed to remove the wrappers in
include/linux/pci-dma-compat.h.

Some reasons why this API should be removed have been given by Julia
Lawall in [2].

A coccinelle script has been used to perform the needed transformation
Only relevant parts are given below.

It has been compile tested.

@@ @@
-    PCI_DMA_TODEVICE
+    DMA_TO_DEVICE

@@ @@
-    PCI_DMA_FROMDEVICE
+    DMA_FROM_DEVICE

@@
expression e1, e2, e3, e4;
@@
-    pci_map_single(e1, e2, e3, e4)
+    dma_map_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_single(e1, e2, e3, e4)
+    dma_unmap_single(&e1->dev, e2, e3, e4)

[1]: https://lore.kernel.org/kernel-janitors/20200421081257.GA131897@infradead.org/
[2]: https://lore.kernel.org/kernel-janitors/alpine.DEB.2.22.394.2007120902170.2424@hadrien/

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: spider_net: switch from 'pci_' to 'dma_' API
Christophe JAILLET [Fri, 27 Aug 2021 19:56:28 +0000 (21:56 +0200)]
net: spider_net: switch from 'pci_' to 'dma_' API

In [1], Christoph Hellwig has proposed to remove the wrappers in
include/linux/pci-dma-compat.h.

Some reasons why this API should be removed have been given by Julia
Lawall in [2].

A coccinelle script has been used to perform the needed transformation
Only relevant parts are given below.

@@ @@
-    PCI_DMA_BIDIRECTIONAL
+    DMA_BIDIRECTIONAL

@@ @@
-    PCI_DMA_TODEVICE
+    DMA_TO_DEVICE

@@ @@
-    PCI_DMA_FROMDEVICE
+    DMA_FROM_DEVICE

@@
expression e1, e2, e3, e4;
@@
-    pci_map_single(e1, e2, e3, e4)
+    dma_map_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2, e3, e4;
@@
-    pci_unmap_single(e1, e2, e3, e4)
+    dma_unmap_single(&e1->dev, e2, e3, e4)

@@
expression e1, e2;
@@
-    pci_dma_mapping_error(e1, e2)
+    dma_mapping_error(&e1->dev, e2)

[1]: https://lore.kernel.org/kernel-janitors/20200421081257.GA131897@infradead.org/
[2]: https://lore.kernel.org/kernel-janitors/alpine.DEB.2.22.394.2007120902170.2424@hadrien/

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Acked-by: Geoff Levand <geoff@infradead.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-pf: Add vlan-etype to ntuple filters
Subbaraya Sundeep [Fri, 27 Aug 2021 13:30:55 +0000 (19:00 +0530)]
octeontx2-pf: Add vlan-etype to ntuple filters

NPC extraction profile marks layer types
NPC_LT_LB_CTAG for CTAG and NPC_LT_LB_STAG_QINQ for
STAG after parsing input packet. Those layer types
can be used to install ntuple filters using
vlan-etype option. Below are the commands and
corresponding behavior with this patch in place.

> alias nt "ethtool -U eth0 flow-type ether"

> nt vlan 5 m 0xf000 action 0
Input packets with outer VLAN id as 5 i.e,
stag packets with VLAN id 5 and ctag packets with
VLAN id as 5 are hit.

> nt vlan-etype 0x8100 action 0
All input ctag packets with any VLAN id are hit.

> nt vlan-etype 0x88A8 action 0
All input stag packets with any VLAN id are hit.

> nt vlan-etype 0x8100 vlan 5 m 0xf000 action 0
All input ctag packets with VLAN id 5 are hit.

> nt vlan-etype 0x88A8 vlan 5 m 0xf000 action 0
All input stag packets with VLAN id 5 are hit.

Signed-off-by: Subbaraya Sundeep <sbhatta@marvell.com>
Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-af: Fix inconsistent license text
Sunil Goutham [Fri, 27 Aug 2021 11:46:44 +0000 (17:16 +0530)]
octeontx2-af: Fix inconsistent license text

Fixed inconsistent license text across the RVU admin
function driver.

Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoocteontx2-pf: Fix inconsistent license text
Sunil Goutham [Fri, 27 Aug 2021 11:45:07 +0000 (17:15 +0530)]
octeontx2-pf: Fix inconsistent license text

Fixed inconsistent license text across the netdev
drivers.

Signed-off-by: Sunil Goutham <sgoutham@marvell.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch '1GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/next-
David S. Miller [Sun, 29 Aug 2021 09:44:31 +0000 (10:44 +0100)]
Merge branch '1GbE' of git://git./linux/kernel/git/tnguy/next-
queue

Tony Nguyen says:

====================
1GbE Intel Wired LAN Driver Updates 2021-08-27

ravindhan Gunasekaran says:

This adds support for Credit-based shaper qdisc offload from
Traffic Control system. It enables traffic prioritization and
bandwidth reservation via the Credit-Based Shaper which is
implemented in hardware by i225 controller.

Patch 1/3 adds a default cycle-time for TSN mode to be configured.

Patch 2/3 helps to separate TSN mode programming on the fly and
during reset sequence. It also simplifies handling features flags
for various TSN modes supported by i225 in the driver.

Patch 3/3 adds support for IEEE802.1Qav(CBS) standard
implemented in i225 HW. Two sets of CBS HW shapers are present
in i225 and driver enables them in the two high priority queues.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge ath-next from git://git.kernel.org/pub/scm/linux/kernel/git/kvalo/ath.git
Kalle Valo [Sun, 29 Aug 2021 08:56:22 +0000 (11:56 +0300)]
Merge ath-next from git://git./linux/kernel/git/kvalo/ath.git

ath.git patches for v5.15. Major changes:

wcn36xx

* allow firmware name to be overridden by DT

2 years agowilc1000: Convert module-global "isinit" to device-specific variable
David Mosberger-Tang [Sat, 28 Aug 2021 15:19:11 +0000 (15:19 +0000)]
wilc1000: Convert module-global "isinit" to device-specific variable

Move static "isinit" variable to wilc_spi structure to allow multiple
WILC1000 devices per machine.

Signed-off-by: David Mosberger-Tang <davidm@egauge.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210828151346.3012697-1-davidm@egauge.net
2 years agobrcmfmac: Add WPA3 Personal with FT to supported cipher suites
Paweł Drewniak [Tue, 24 Aug 2021 22:13:30 +0000 (23:13 +0100)]
brcmfmac: Add WPA3 Personal with FT to supported cipher suites

This allows the driver to connect to BSSIDs supporting SAE with 802.11r.
Tested on Raspberry Pi 4 Model B (STA) and UniFi 6LR/OpenWRT 21.02.0-rc2.
AP was set to 'sae-mixed' (WPA2/3 Personal).

Signed-off-by: Paweł Drewniak <czajernia@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210824221330.3847139-1-czajernia@gmail.com
2 years agossb: Drop legacy header include
Linus Walleij [Sat, 21 Aug 2021 23:58:00 +0000 (01:58 +0200)]
ssb: Drop legacy header include

The SSB header only uses the legacy <linux/gpio.h> header to get
struct gpio_chip so inluce <linux/gpio/driver.h> which is the right
include to deal with gpio_chip.

Cc: Michael Buesch <m@bues.ch>
Cc: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210821235800.138817-1-linus.walleij@linaro.org
2 years agorsi: make array fsm_state static const, makes object smaller
Colin Ian King [Thu, 19 Aug 2021 12:50:18 +0000 (13:50 +0100)]
rsi: make array fsm_state static const, makes object smaller

Don't populate the array fsm_state on the stack but instead it
static const. Makes the object code smaller by 154 bytes:

Before:
   text    data     bss     dec     hex filename
   9213    3904       0   13117    333d .../wireless/rsi/rsi_91x_debugfs.o

After:
   text    data     bss     dec     hex filename
   8995    3968       0   12963    32a3 .../wireless/rsi/rsi_91x_debugfs.o

(gcc version 10.3.0)

Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210819125018.8577-1-colin.king@canonical.com
2 years agortlwifi: rtl8192de: Fix initialization of place in _rtl92c_phy_get_rightchnlplace()
Nathan Chancellor [Mon, 23 Aug 2021 22:20:14 +0000 (15:20 -0700)]
rtlwifi: rtl8192de: Fix initialization of place in _rtl92c_phy_get_rightchnlplace()

Clang warns:

drivers/net/wireless/realtek/rtlwifi/rtl8192de/phy.c:901:6: warning:
variable 'place' is used uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]
        if (chnl > 14) {
            ^~~~~~~~~
drivers/net/wireless/realtek/rtlwifi/rtl8192de/phy.c:909:9: note:
uninitialized use occurs here
        return place;
               ^~~~~
drivers/net/wireless/realtek/rtlwifi/rtl8192de/phy.c:901:2: note: remove
the 'if' if its condition is always true
        if (chnl > 14) {
        ^~~~~~~~~~~~~~~
drivers/net/wireless/realtek/rtlwifi/rtl8192de/phy.c:899:10: note:
initialize the variable 'place' to silence this warning
        u8 place;
                ^
                 = '\0'
1 warning generated.

Commit 369956ae5720 ("rtlwifi: rtl8192de: Remove redundant variable
initializations") removed the initialization of place but it appears
that this removal was in the wrong function.

_rtl92c_phy_get_rightchnlplace() returns place's value at the end of the
function so now if the if statement is false, place never gets
initialized. Add that initialization back to address the warning.

place's initialization is not necessary in
rtl92d_get_rightchnlplace_for_iqk() as place is only used within the if
statement so it can be removed, which is likely what was intended in the
first place.

Fixes: 369956ae5720 ("rtlwifi: rtl8192de: Remove redundant variable initializations")
Signed-off-by: Nathan Chancellor <nathan@kernel.org>
Acked-by: Ping-Ke Shih <pkshih@realtek.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210823222014.764557-1-nathan@kernel.org
2 years agortw88: add quirk to disable pci caps on HP Pavilion 14-ce0xxx
Ugo Rémery [Sun, 22 Aug 2021 19:49:32 +0000 (14:49 -0500)]
rtw88: add quirk to disable pci caps on HP Pavilion 14-ce0xxx

8821CE causes random freezes on HP Pavilion 14-ce0019nf. Add a quirk
to disable pci ASPM capability.

Signed-off-by: Ugo Rémery <ugo.remery@gmail.com>
Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210822194932.29630-1-Larry.Finger@lwfinger.net
2 years agoath6kl: wmi: fix an error code in ath6kl_wmi_sync_point()
Dan Carpenter [Fri, 13 Aug 2021 11:34:38 +0000 (14:34 +0300)]
ath6kl: wmi: fix an error code in ath6kl_wmi_sync_point()

This error path is unlikely because of it checked for NULL and
returned -ENOMEM earlier in the function.  But it should return
an error code here as well if we ever do hit it because of a
race condition or something.

Fixes: bdcd81707973 ("Add ath6kl cleaned up driver")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210813113438.GB30697@kili
2 years agoath9k: fix sleeping in atomic context
Miaoqing Pan [Mon, 9 Aug 2021 04:05:16 +0000 (12:05 +0800)]
ath9k: fix sleeping in atomic context

The problem is that gpio_free() can sleep and the cfg_soc() can be
called with spinlocks held. One problematic call tree is:

--> ath_reset_internal() takes &sc->sc_pcu_lock spin lock
   --> ath9k_hw_reset()
      --> ath9k_hw_gpio_request_in()
         --> ath9k_hw_gpio_request()
            --> ath9k_hw_gpio_cfg_soc()

Remove gpio_free(), use error message instead, so we should make sure
there is no GPIO conflict.

Also remove ath9k_hw_gpio_free() from ath9k_hw_apply_gpio_override(),
as gpio_mask will never be set for SOC chips.

Signed-off-by: Miaoqing Pan <miaoqing@codeaurora.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1628481916-15030-1-git-send-email-miaoqing@codeaurora.org
2 years agoath9k: fix OOB read ar9300_eeprom_restore_internal
Zekun Shen [Sat, 19 Jun 2021 13:29:14 +0000 (09:29 -0400)]
ath9k: fix OOB read ar9300_eeprom_restore_internal

Bad header can have large length field which can cause OOB.
cptr is the last bytes for read, and the eeprom is parsed
from high to low address. The OOB, triggered by the condition
length > cptr could cause memory error with a read on
negative index.

There are some sanity check around length, but it is not
compared with cptr (the remaining bytes). Here, the
corrupted/bad EEPROM can cause panic.

I was able to reproduce the crash, but I cannot find the
log and the reproducer now. After I applied the patch, the
bug is no longer reproducible.

Signed-off-by: Zekun Shen <bruceshenzk@gmail.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/YM3xKsQJ0Hw2hjrc@Zekuns-MBP-16.fios-router.home
2 years agowcn36xx: Fix missing frame timestamp for beacon/probe-resp
Loic Poulain [Thu, 26 Aug 2021 15:46:08 +0000 (17:46 +0200)]
wcn36xx: Fix missing frame timestamp for beacon/probe-resp

When receiving a beacon or probe response, we should update the
boottime_ns field which is the timestamp the frame was received at.
(cf mac80211.h)

This fixes a scanning issue with Android since it relies on this
timestamp to determine when the AP has been seen for the last time
(via the nl80211 BSS_LAST_SEEN_BOOTTIME parameter).

Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Reviewed-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1629992768-23785-1-git-send-email-loic.poulain@linaro.org
2 years agowcn36xx: Allow firmware name to be overridden by DT
Bjorn Andersson [Tue, 24 Aug 2021 17:12:25 +0000 (10:12 -0700)]
wcn36xx: Allow firmware name to be overridden by DT

The WLAN NV firmware blob differs between platforms, and possibly
devices, so add support in the wcn36xx driver for reading the path of
this file from DT in order to allow these files to live in a generic
file system (or linux-firmware).

For some reason the parent (wcnss_ctrl) also needs to upload this blob,
so rather than specifying the same information in both nodes wcn36xx
reads the string from the parent's of_node.

Signed-off-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Tested-by: Aníbal Limón <anibal.limon@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20210824171225.686683-1-bjorn.andersson@linaro.org
2 years agowcn36xx: Ensure finish scan is not requested before start scan
Joseph Gates [Wed, 18 Aug 2021 11:31:43 +0000 (13:31 +0200)]
wcn36xx: Ensure finish scan is not requested before start scan

If the operating channel is the first in the scan list, it was seen that
a finish scan request would be sent before a start scan request was
sent, causing the firmware to fail all future scans. Track the current
channel being scanned to avoid requesting the scan finish before it
starts.

Cc: <stable@vger.kernel.org>
Fixes: 5973a2947430 ("wcn36xx: Fix software-driven scan")
Signed-off-by: Joseph Gates <jgates@squareup.com>
Signed-off-by: Loic Poulain <loic.poulain@linaro.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/1629286303-13179-1-git-send-email-loic.poulain@linaro.org
2 years agoMerge branch 'ionic-queue-mgmt'
David S. Miller [Sat, 28 Aug 2021 10:23:09 +0000 (11:23 +0100)]
Merge branch 'ionic-queue-mgmt'

Shannon Nelson says:

====================
ionic: queue mgmt updates

The first pair of patches help smooth the driver's response when
the firmware has gone through a recovery/reboot cycle.

The next four patches take care of a couple things seen when
changing the interface status.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoionic: recreate hwstamp queues on ifup
Shannon Nelson [Fri, 27 Aug 2021 18:55:12 +0000 (11:55 -0700)]
ionic: recreate hwstamp queues on ifup

The queues can be freed in ionic_close().  They need to be recreated
after ionic_open().  It doesn't need to replay the whole config.  It
only needs to create the timestamping queues again.

Signed-off-by: Allen Hubbe <allenbh@pensando.io>
Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoionic: pull hwstamp queue_lock up a level
Shannon Nelson [Fri, 27 Aug 2021 18:55:11 +0000 (11:55 -0700)]
ionic: pull hwstamp queue_lock up a level

Move the hwstamp configuration use of queue_lock up
a level to simplify use and error handling.

Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoionic: add queue lock around open and stop
Shannon Nelson [Fri, 27 Aug 2021 18:55:10 +0000 (11:55 -0700)]
ionic: add queue lock around open and stop

Add the queue configuration lock to ionic_open() and
ionic_stop() so that they don't collide with other in parallel
queue configuration actions such as MTU changes as can be
demonstrated with a tight loop of ifup/change-mtu/ifdown.

Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoionic: fill mac addr earlier in add_addr
Shannon Nelson [Fri, 27 Aug 2021 18:55:09 +0000 (11:55 -0700)]
ionic: fill mac addr earlier in add_addr

Make sure the ctx struct has the new mac address before
any save operations happen.

Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoionic: squelch unnecessary fw halted message
Shannon Nelson [Fri, 27 Aug 2021 18:55:08 +0000 (11:55 -0700)]
ionic: squelch unnecessary fw halted message

Since the heartbeat check will already have complained about
the firmware status, don't bother complaining about the
DEVCMD failing.  We'll keep the print message but demote it
to a debug messages so that we normally no longer see it.

Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoionic: fire watchdog again after fw_down
Shannon Nelson [Fri, 27 Aug 2021 18:55:07 +0000 (11:55 -0700)]
ionic: fire watchdog again after fw_down

In some cases of fw_down it was called because there was a
fw_generation change, and the firmware is already back up.
In order to keep the down time to a minimum, don't wait for
the next watchdog polling cycle, fire another watchdog off
as soon as we can - an out-of-cycle check won't hurt, and
may well speed up the recovery.

Signed-off-by: Shannon Nelson <snelson@pensando.io>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agoMerge branch 'hns3-next'
David S. Miller [Sat, 28 Aug 2021 10:20:05 +0000 (11:20 +0100)]
Merge branch 'hns3-next'

Guangbin Huang says:

====================
net: hns3: updates for -next

This series includes some updates for the HNS3 ethernet driver.

This series includes some optimizations, cleanups and one
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: add required space in comment
Hao Chen [Sat, 28 Aug 2021 06:55:21 +0000 (14:55 +0800)]
net: hns3: add required space in comment

Add some required spaces in comment for cleanup.

Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
2 years agonet: hns3: remove unnecessary "static" of local variables in function
Hao Chen [Sat, 28 Aug 2021 06:55:20 +0000 (14:55 +0800)]
net: hns3: remove unnecessary "static" of local variables in function

Some local variable declarations are no need to add "static", so remove it.

Signed-off-by: Hao Chen <chenhao288@hisilicon.com>
Signed-off-by: Guangbin Huang <huangguangbin2@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>