platform/kernel/linux-starfive.git
22 months agonetfilter: nfnetlink_osf: fix possible bogus match in nf_osf_find()
Pablo Neira Ayuso [Wed, 7 Sep 2022 08:26:18 +0000 (10:26 +0200)]
netfilter: nfnetlink_osf: fix possible bogus match in nf_osf_find()

nf_osf_find() incorrectly returns true on mismatch, this leads to
copying uninitialized memory area in nft_osf which can be used to leak
stale kernel stack data to userspace.

Fixes: 22c7652cdaa8 ("netfilter: nft_osf: Add version option support")
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Florian Westphal <fw@strlen.de>
22 months agonetfilter: nf_conntrack_irc: Tighten matching on DCC message
David Leadbeater [Fri, 26 Aug 2022 04:56:57 +0000 (14:56 +1000)]
netfilter: nf_conntrack_irc: Tighten matching on DCC message

CTCP messages should only be at the start of an IRC message, not
anywhere within it.

While the helper only decodes packes in the ORIGINAL direction, its
possible to make a client send a CTCP message back by empedding one into
a PING request.  As-is, thats enough to make the helper believe that it
saw a CTCP message.

Fixes: 869f37d8e48f ("[NETFILTER]: nf_conntrack/nf_nat: add IRC helper port")
Signed-off-by: David Leadbeater <dgl@dgl.cx>
Signed-off-by: Florian Westphal <fw@strlen.de>
22 months agoselftests: nft_concat_range: add socat support
Florian Westphal [Wed, 31 Aug 2022 13:12:45 +0000 (15:12 +0200)]
selftests: nft_concat_range: add socat support

There are different flavors of 'nc' around, this script fails on
my test vm because 'nc' is 'nmap-ncat' which isn't 100% compatible.

Add socat support and use it if available.

Signed-off-by: Florian Westphal <fw@strlen.de>
22 months agonetfilter: nf_conntrack_sip: fix ct_sip_walk_headers
Igor Ryzhov [Wed, 5 Jun 2019 09:32:40 +0000 (12:32 +0300)]
netfilter: nf_conntrack_sip: fix ct_sip_walk_headers

ct_sip_next_header and ct_sip_get_header return an absolute
value of matchoff, not a shift from current dataoff.
So dataoff should be assigned matchoff, not incremented by it.

This issue can be seen in the scenario when there are multiple
Contact headers and the first one is using a hostname and other headers
use IP addresses. In this case, ct_sip_walk_headers will work as follows:

The first ct_sip_get_header call to will find the first Contact header
but will return -1 as the header uses a hostname. But matchoff will
be changed to the offset of this header. After that, dataoff should be
set to matchoff, so that the next ct_sip_get_header call find the next
Contact header. But instead of assigning dataoff to matchoff, it is
incremented by it, which is not correct, as matchoff is an absolute
value of the offset. So on the next call to the ct_sip_get_header,
dataoff will be incorrect, and the next Contact header may not be
found at all.

Fixes: 05e3ced297fe ("[NETFILTER]: nf_conntrack_sip: introduce SIP-URI parsing helper")
Signed-off-by: Igor Ryzhov <iryzhov@nfware.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
22 months agoMerge branch 'dsa-felix-fixes'
David S. Miller [Wed, 7 Sep 2022 12:44:04 +0000 (13:44 +0100)]
Merge branch 'dsa-felix-fixes'

Vladimir Oltean says:

====================
Fixes for Felix DSA driver calculation of tc-taprio guard bands

This series fixes some bugs which are not quite new, but date from v5.13
when static guard bands were enabled by Michael Walle to prevent
tc-taprio overruns.

The investigation started when Xiaoliang asked privately what is the
expected max SDU for a traffic class when its minimum gate interval is
10 us. The answer, as it turns out, is not an L1 size of 1250 octets,
but 1245 octets, since otherwise, the switch will not consider frames
for egress scheduling, because the static guard band is exactly as large
as the time interval. The switch needs a minimum of 33 ns outside of the
guard band to consider a frame for scheduling, and the reduction of the
max SDU by 5 provides exactly for that.

The fix for that (patch 1/3) is relatively small, but during testing, it
became apparent that cut-through forwarding prevents oversized frame
dropping from working properly. This is solved through the larger patch
2/3. Finally, patch 3/3 fixes one more tc-taprio locking problem found
through code inspection.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: dsa: felix: access QSYS_TAG_CONFIG under tas_lock in vsc9959_sched_speed_set
Vladimir Oltean [Mon, 5 Sep 2022 17:01:25 +0000 (20:01 +0300)]
net: dsa: felix: access QSYS_TAG_CONFIG under tas_lock in vsc9959_sched_speed_set

The read-modify-write of QSYS_TAG_CONFIG from vsc9959_sched_speed_set()
runs unlocked with respect to the other functions that access it, which
are vsc9959_tas_guard_bands_update(), vsc9959_qos_port_tas_set() and
vsc9959_tas_clock_adjust(). All the others are under ocelot->tas_lock,
so move the vsc9959_sched_speed_set() access under that lock as well, to
resolve the concurrency.

Fixes: 55a515b1f5a9 ("net: dsa: felix: drop oversized frames with tc-taprio instead of hanging the port")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: dsa: felix: disable cut-through forwarding for frames oversized for tc-taprio
Vladimir Oltean [Mon, 5 Sep 2022 17:01:24 +0000 (20:01 +0300)]
net: dsa: felix: disable cut-through forwarding for frames oversized for tc-taprio

Experimentally, it looks like when QSYS_QMAXSDU_CFG_7 is set to 605,
frames even way larger than 601 octets are transmitted even though these
should be considered as oversized, according to the documentation, and
dropped.

Since oversized frame dropping depends on frame size, which is only
known at the EOF stage, and therefore not at SOF when cut-through
forwarding begins, it means that the switch cannot take QSYS_QMAXSDU_CFG_*
into consideration for traffic classes that are cut-through.

Since cut-through forwarding has no UAPI to control it, and the driver
enables it based on the mantra "if we can, then why not", the strategy
is to alter vsc9959_cut_through_fwd() to take into consideration which
tc's have oversize frame dropping enabled, and disable cut-through for
them. Then, from vsc9959_tas_guard_bands_update(), we re-trigger the
cut-through determination process.

There are 2 strategies for vsc9959_cut_through_fwd() to determine
whether a tc has oversized dropping enabled or not. One is to keep a bit
mask of traffic classes per port, and the other is to read back from the
hardware registers (a non-zero value of QSYS_QMAXSDU_CFG_* means the
feature is enabled). We choose reading back from registers, because
struct ocelot_port is shared with drivers (ocelot, seville) that don't
support either cut-through nor tc-taprio, and we don't have a felix
specific extension of struct ocelot_port. Furthermore, reading registers
from the Felix hardware is quite cheap, since they are memory-mapped.

Fixes: 55a515b1f5a9 ("net: dsa: felix: drop oversized frames with tc-taprio instead of hanging the port")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: dsa: felix: tc-taprio intervals smaller than MTU should send at least one packet
Vladimir Oltean [Mon, 5 Sep 2022 17:01:23 +0000 (20:01 +0300)]
net: dsa: felix: tc-taprio intervals smaller than MTU should send at least one packet

The blamed commit broke tc-taprio schedules such as this one:

tc qdisc replace dev $swp1 root taprio \
        num_tc 8 \
        map 0 1 2 3 4 5 6 7 \
        queues 1@0 1@1 1@2 1@3 1@4 1@5 1@6 1@7 \
        base-time 0 \
        sched-entry S 0x7f 990000 \
        sched-entry S 0x80  10000 \
        flags 0x2

because the gate entry for TC 7 (S 0x80 10000 ns) now has a static guard
band added earlier than its 'gate close' event, such that packet
overruns won't occur in the worst case of the largest packet possible.

Since guard bands are statically determined based on the per-tc
QSYS_QMAXSDU_CFG_* with a fallback on the port-based QSYS_PORT_MAX_SDU,
we need to discuss what happens with TC 7 depending on kernel version,
since the driver, prior to commit 55a515b1f5a9 ("net: dsa: felix: drop
oversized frames with tc-taprio instead of hanging the port"), did not
touch QSYS_QMAXSDU_CFG_*, and therefore relied on QSYS_PORT_MAX_SDU.

1 (before vsc9959_tas_guard_bands_update): QSYS_PORT_MAX_SDU defaults to
  1518, and at gigabit this introduces a static guard band (independent
  of packet sizes) of 12144 ns, plus QSYS::HSCH_MISC_CFG.FRM_ADJ (bit
  time of 20 octets => 160 ns). But this is larger than the time window
  itself, of 10000 ns. So, the queue system never considers a frame with
  TC 7 as eligible for transmission, since the gate practically never
  opens, and these frames are forever stuck in the TX queues and hang
  the port.

2 (after vsc9959_tas_guard_bands_update): Under the sole goal of
  enabling oversized frame dropping, we make an effort to set
  QSYS_QMAXSDU_CFG_7 to 1230 bytes. But QSYS_QMAXSDU_CFG_7 plays
  one more role, which we did not take into account: per-tc static guard
  band, expressed in L2 byte time (auto-adjusted for FCS and L1 overhead).
  There is a discrepancy between what the driver thinks (that there is
  no guard band, and 100% of min_gate_len[tc] is available for egress
  scheduling) and what the hardware actually does (crops the equivalent
  of QSYS_QMAXSDU_CFG_7 ns out of min_gate_len[tc]). In practice, this
  means that the hardware thinks it has exactly 0 ns for scheduling tc 7.

In both cases, even minimum sized Ethernet frames are stuck on egress
rather than being considered for scheduling on TC 7, even if they would
fit given a proper configuration. Considering the current situation,
with vsc9959_tas_guard_bands_update(), frames between 60 octets and 1230
octets in size are not eligible for oversized dropping (because they are
smaller than QSYS_QMAXSDU_CFG_7), but won't be considered as eligible
for scheduling either, because the min_gate_len[7] (10000 ns) minus the
guard band determined by QSYS_QMAXSDU_CFG_7 (1230 octets * 8 ns per
octet == 9840 ns) minus the guard band auto-added for L1 overhead by
QSYS::HSCH_MISC_CFG.FRM_ADJ (20 octets * 8 ns per octet == 160 octets)
leaves 0 ns for scheduling in the queue system proper.

Investigating the hardware behavior, it becomes apparent that the queue
system needs precisely 33 ns of 'gate open' time in order to consider a
frame as eligible for scheduling to a tc. So the solution to this
problem is to amend vsc9959_tas_guard_bands_update(), by giving the
per-tc guard bands less space by exactly 33 ns, just enough for one
frame to be scheduled in that interval. This allows the queue system to
make forward progress for that port-tc, and prevents it from hanging.

Fixes: 297c4de6f780 ("net: dsa: felix: re-enable TAS guard band mode")
Reported-by: Xiaoliang Yang <xiaoliang.yang_1@nxp.com>
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: usb: qmi_wwan: add Quectel RM520N
jerry.meng [Mon, 5 Sep 2022 01:24:52 +0000 (09:24 +0800)]
net: usb: qmi_wwan: add Quectel RM520N

add support for Quectel RM520N which is based on Qualcomm SDX62 chip.

0x0801: DIAG + NMEA + AT + MODEM + RMNET

T:  Bus=03 Lev=01 Prnt=01 Port=01 Cnt=02 Dev#= 10 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=00(>ifc ) Sub=00 Prot=00 MxPS=64 #Cfgs=  1
P:  Vendor=2c7c ProdID=0801 Rev= 5.04
S:  Manufacturer=Quectel
S:  Product=RM520N-GL
S:  SerialNumber=384af524
C:* #Ifs= 5 Cfg#= 1 Atr=a0 MxPwr=500mA
I:* If#= 0 Alt= 0 #EPs= 2 Cls=ff(vend.) Sub=ff Prot=30 Driver=option
E:  Ad=01(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=81(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=40 Driver=option
E:  Ad=83(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
E:  Ad=82(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 2 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option
E:  Ad=85(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
E:  Ad=84(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=03(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 3 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=00 Prot=00 Driver=option
E:  Ad=87(I) Atr=03(Int.) MxPS=  10 Ivl=32ms
E:  Ad=86(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=04(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 4 Alt= 0 #EPs= 3 Cls=ff(vend.) Sub=ff Prot=ff Driver=qmi_wwan
E:  Ad=88(I) Atr=03(Int.) MxPS=   8 Ivl=32ms
E:  Ad=8e(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=0f(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms

Signed-off-by: jerry.meng <jerry-meng@foxmail.com>
Acked-by: Bjørn Mork <bjorn@mork.no>
Link: https://lore.kernel.org/r/tencent_E50CA8A206904897C2D20DDAE90731183C05@qq.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
22 months agonet: dsa: qca8k: fix NULL pointer dereference for of_device_get_match_data
Christian Marangi [Sun, 4 Sep 2022 21:53:19 +0000 (23:53 +0200)]
net: dsa: qca8k: fix NULL pointer dereference for of_device_get_match_data

of_device_get_match_data is called on priv->dev before priv->dev is
actually set. Move of_device_get_match_data after priv->dev is correctly
set to fix this kernel panic.

Fixes: 3bb0844e7bcd ("net: dsa: qca8k: cache match data to speed up access")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Link: https://lore.kernel.org/r/20220904215319.13070-1-ansuelsmth@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
22 months agotcp: fix early ETIMEDOUT after spurious non-SACK RTO
Neal Cardwell [Sat, 3 Sep 2022 12:10:23 +0000 (08:10 -0400)]
tcp: fix early ETIMEDOUT after spurious non-SACK RTO

Fix a bug reported and analyzed by Nagaraj Arankal, where the handling
of a spurious non-SACK RTO could cause a connection to fail to clear
retrans_stamp, causing a later RTO to very prematurely time out the
connection with ETIMEDOUT.

Here is the buggy scenario, expanding upon Nagaraj Arankal's excellent
report:

(*1) Send one data packet on a non-SACK connection

(*2) Because no ACK packet is received, the packet is retransmitted
     and we enter CA_Loss; but this retransmission is spurious.

(*3) The ACK for the original data is received. The transmitted packet
     is acknowledged.  The TCP timestamp is before the retrans_stamp,
     so tcp_may_undo() returns true, and tcp_try_undo_loss() returns
     true without changing state to Open (because tcp_is_sack() is
     false), and tcp_process_loss() returns without calling
     tcp_try_undo_recovery().  Normally after undoing a CA_Loss
     episode, tcp_fastretrans_alert() would see that the connection
     has returned to CA_Open and fall through and call
     tcp_try_to_open(), which would set retrans_stamp to 0.  However,
     for non-SACK connections we hold the connection in CA_Loss, so do
     not fall through to call tcp_try_to_open() and do not set
     retrans_stamp to 0. So retrans_stamp is (erroneously) still
     non-zero.

     At this point the first "retransmission event" has passed and
     been recovered from. Any future retransmission is a completely
     new "event". However, retrans_stamp is erroneously still
     set. (And we are still in CA_Loss, which is correct.)

(*4) After 16 minutes (to correspond with tcp_retries2=15), a new data
     packet is sent. Note: No data is transmitted between (*3) and
     (*4) and we disabled keep alives.

     The socket's timeout SHOULD be calculated from this point in
     time, but instead it's calculated from the prior "event" 16
     minutes ago (step (*2)).

(*5) Because no ACK packet is received, the packet is retransmitted.

(*6) At the time of the 2nd retransmission, the socket returns
     ETIMEDOUT, prematurely, because retrans_stamp is (erroneously)
     too far in the past (set at the time of (*2)).

This commit fixes this bug by ensuring that we reuse in
tcp_try_undo_loss() the same careful logic for non-SACK connections
that we have in tcp_try_undo_recovery(). To avoid duplicating logic,
we factor out that logic into a new
tcp_is_non_sack_preventing_reopen() helper and call that helper from
both undo functions.

Fixes: da34ac7626b5 ("tcp: only undo on partial ACKs in CA_Loss")
Reported-by: Nagaraj Arankal <nagaraj.p.arankal@hpe.com>
Link: https://lore.kernel.org/all/SJ0PR84MB1847BE6C24D274C46A1B9B0EB27A9@SJ0PR84MB1847.NAMPRD84.PROD.OUTLOOK.COM/
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/20220903121023.866900-1-ncardwell.kernel@gmail.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
22 months agoMerge tag 'for-net-2022-09-02' of git://git.kernel.org/pub/scm/linux/kernel/git/bluet...
David S. Miller [Mon, 5 Sep 2022 13:43:18 +0000 (14:43 +0100)]
Merge tag 'for-net-2022-09-02' of git://git./linux/kernel/git/bluetooth/bluetooth

Luiz Augusto von Dentz says:

====================
bluetooth pull request for net:

 - Fix regression preventing ACL packet transmission
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agostmmac: intel: Simplify intel_eth_pci_remove()
Christophe JAILLET [Fri, 2 Sep 2022 16:26:56 +0000 (18:26 +0200)]
stmmac: intel: Simplify intel_eth_pci_remove()

There is no point to call pcim_iounmap_regions() in the remove function,
this frees a managed resource that would be release by the framework
anyway.

Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: mvpp2: debugfs: fix memory leak when using debugfs_lookup()
Greg Kroah-Hartman [Fri, 2 Sep 2022 13:41:11 +0000 (15:41 +0200)]
net: mvpp2: debugfs: fix memory leak when using debugfs_lookup()

When calling debugfs_lookup() the result must have dput() called on it,
otherwise the memory will leak over time.  Fix this up to be much
simpler logic and only create the root debugfs directory once when the
driver is first accessed.  That resolves the memory leak and makes
things more obvious as to what the intent is.

Cc: Marcin Wojtas <mw@semihalf.com>
Cc: Russell King <linux@armlinux.org.uk>
Cc: "David S. Miller" <davem@davemloft.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: netdev@vger.kernel.org
Cc: stable <stable@kernel.org>
Fixes: 21da57a23125 ("net: mvpp2: add a debugfs interface for the Header Parser")
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoipv6: sr: fix out-of-bounds read when setting HMAC data.
David Lebrun [Fri, 2 Sep 2022 09:45:06 +0000 (10:45 +0100)]
ipv6: sr: fix out-of-bounds read when setting HMAC data.

The SRv6 layer allows defining HMAC data that can later be used to sign IPv6
Segment Routing Headers. This configuration is realised via netlink through
four attributes: SEG6_ATTR_HMACKEYID, SEG6_ATTR_SECRET, SEG6_ATTR_SECRETLEN and
SEG6_ATTR_ALGID. Because the SECRETLEN attribute is decoupled from the actual
length of the SECRET attribute, it is possible to provide invalid combinations
(e.g., secret = "", secretlen = 64). This case is not checked in the code and
with an appropriately crafted netlink message, an out-of-bounds read of up
to 64 bytes (max secret length) can occur past the skb end pointer and into
skb_shared_info:

Breakpoint 1, seg6_genl_sethmac (skb=<optimized out>, info=<optimized out>) at net/ipv6/seg6.c:208
208 memcpy(hinfo->secret, secret, slen);
(gdb) bt
 #0  seg6_genl_sethmac (skb=<optimized out>, info=<optimized out>) at net/ipv6/seg6.c:208
 #1  0xffffffff81e012e9 in genl_family_rcv_msg_doit (skb=skb@entry=0xffff88800b1f9f00, nlh=nlh@entry=0xffff88800b1b7600,
    extack=extack@entry=0xffffc90000ba7af0, ops=ops@entry=0xffffc90000ba7a80, hdrlen=4, net=0xffffffff84237580 <init_net>, family=<optimized out>,
    family=<optimized out>) at net/netlink/genetlink.c:731
 #2  0xffffffff81e01435 in genl_family_rcv_msg (extack=0xffffc90000ba7af0, nlh=0xffff88800b1b7600, skb=0xffff88800b1f9f00,
    family=0xffffffff82fef6c0 <seg6_genl_family>) at net/netlink/genetlink.c:775
 #3  genl_rcv_msg (skb=0xffff88800b1f9f00, nlh=0xffff88800b1b7600, extack=0xffffc90000ba7af0) at net/netlink/genetlink.c:792
 #4  0xffffffff81dfffc3 in netlink_rcv_skb (skb=skb@entry=0xffff88800b1f9f00, cb=cb@entry=0xffffffff81e01350 <genl_rcv_msg>)
    at net/netlink/af_netlink.c:2501
 #5  0xffffffff81e00919 in genl_rcv (skb=0xffff88800b1f9f00) at net/netlink/genetlink.c:803
 #6  0xffffffff81dff6ae in netlink_unicast_kernel (ssk=0xffff888010eec800, skb=0xffff88800b1f9f00, sk=0xffff888004aed000)
    at net/netlink/af_netlink.c:1319
 #7  netlink_unicast (ssk=ssk@entry=0xffff888010eec800, skb=skb@entry=0xffff88800b1f9f00, portid=portid@entry=0, nonblock=<optimized out>)
    at net/netlink/af_netlink.c:1345
 #8  0xffffffff81dff9a4 in netlink_sendmsg (sock=<optimized out>, msg=0xffffc90000ba7e48, len=<optimized out>) at net/netlink/af_netlink.c:1921
...
(gdb) p/x ((struct sk_buff *)0xffff88800b1f9f00)->head + ((struct sk_buff *)0xffff88800b1f9f00)->end
$1 = 0xffff88800b1b76c0
(gdb) p/x secret
$2 = 0xffff88800b1b76c0
(gdb) p slen
$3 = 64 '@'

The OOB data can then be read back from userspace by dumping HMAC state. This
commit fixes this by ensuring SECRETLEN cannot exceed the actual length of
SECRET.

Reported-by: Lucas Leong <wmliang.tw@gmail.com>
Tested: verified that EINVAL is correctly returned when secretlen > len(secret)
Fixes: 4f4853dc1c9c1 ("ipv6: sr: implement API to control SR HMAC structure")
Signed-off-by: David Lebrun <dlebrun@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoMerge branch 'bonding-fixes'
David S. Miller [Mon, 5 Sep 2022 09:07:05 +0000 (10:07 +0100)]
Merge branch 'bonding-fixes'

Hangbin Liu says:

====================
bonding: fix lladdr finding and confirmation

This patch set fixed 3 issues when setting lladdr as bonding IPv6 target.
Please see each patch for the details.

v2: separate the patch to 3 parts
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agobonding: accept unsolicited NA message
Hangbin Liu [Tue, 30 Aug 2022 09:37:22 +0000 (17:37 +0800)]
bonding: accept unsolicited NA message

The unsolicited NA message with all-nodes multicast dest address should
be valid, as this also means the link could reach the target.

Also rename bond_validate_ns() to bond_validate_na().

Reported-by: LiLiang <liali@redhat.com>
Fixes: 5e1eeef69c0f ("bonding: NS target should accept link local address")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agobonding: add all node mcast address when slave up
Hangbin Liu [Tue, 30 Aug 2022 09:37:21 +0000 (17:37 +0800)]
bonding: add all node mcast address when slave up

When a link is enslave to bond, it need to set the interface down first.
This makes the slave remove mac multicast address 33:33:00:00:00:01(The
IPv6 multicast address ff02::1 is kept even when the interface down). When
bond set the slave up, ipv6_mc_up() was not called due to commit c2edacf80e15
("bonding / ipv6: no addrconf for slaves separately from master").

This is not an issue before we adding the lladdr target feature for bonding,
as the mac multicast address will be added back when bond interface up and
join group ff02::1.

But after adding lladdr target feature for bonding. When user set a lladdr
target, the unsolicited NA message with all-nodes multicast dest will be
dropped as the slave interface never add 33:33:00:00:00:01 back.

Fix this by calling ipv6_mc_up() to add 33:33:00:00:00:01 back when
the slave interface up.

Reported-by: LiLiang <liali@redhat.com>
Fixes: 5e1eeef69c0f ("bonding: NS target should accept link local address")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agobonding: use unspecified address if no available link local address
Hangbin Liu [Tue, 30 Aug 2022 09:37:20 +0000 (17:37 +0800)]
bonding: use unspecified address if no available link local address

When ns_ip6_target was set, the ipv6_dev_get_saddr() will be called to get
available source address and send IPv6 neighbor solicit message.

If the target is global address, ipv6_dev_get_saddr() will get any
available src address. But if the target is link local address,
ipv6_dev_get_saddr() will only get available address from our interface,
i.e. the corresponding bond interface.

But before bond interface up, all the address is tentative, while
ipv6_dev_get_saddr() will ignore tentative address. This makes we can't
find available link local src address, then bond_ns_send() will not be
called and no NS message was sent. Finally bond interface will keep in
down state.

Fix this by sending NS with unspecified address if there is no available
source address.

Reported-by: LiLiang <liali@redhat.com>
Fixes: 5e1eeef69c0f ("bonding: NS target should accept link local address")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoMerge tag 'wireless-2022-09-03' of git://git.kernel.org/pub/scm/linux/kernel/git...
David S. Miller [Sun, 4 Sep 2022 10:23:11 +0000 (11:23 +0100)]
Merge tag 'wireless-2022-09-03' of git://git./linux/kernel/git/wireless/wireless

Johannes berg says:

====================
We have a handful of fixes:
 - fix DMA from stack in wilc1000 driver
 - fix crash on chip reset failure in mt7921e
 - fix for the reported warning on aggregation timer expiry
 - check packet lengths in hwsim virtio paths
 - fix compiler warnings/errors with AAD construction by
   using struct_group
 - fix Intel 4965 driver rate scale operation
 - release channel contexts correctly in mac80211 mlme code
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agowifi: use struct_group to copy addresses
Johannes Berg [Mon, 29 Aug 2022 09:46:38 +0000 (11:46 +0200)]
wifi: use struct_group to copy addresses

We sometimes copy all the addresses from the 802.11 header
for the AAD, which may cause complaints from fortify checks.
Use struct_group() to avoid the compiler warnings/errors.

Change-Id: Ic3ea389105e7813b22095b295079eecdabde5045
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
22 months agowifi: mac80211_hwsim: check length for virtio packets
Soenke Huster [Fri, 2 Sep 2022 08:19:58 +0000 (10:19 +0200)]
wifi: mac80211_hwsim: check length for virtio packets

An invalid packet with a length shorter than the specified length in the
netlink header can lead to use-after-frees and slab-out-of-bounds in the
processing of the netlink attributes, such as the following:

  BUG: KASAN: slab-out-of-bounds in __nla_validate_parse+0x1258/0x2010
  Read of size 2 at addr ffff88800ac7952c by task kworker/0:1/12

  Workqueue: events hwsim_virtio_rx_work
  Call Trace:
   <TASK>
   dump_stack_lvl+0x45/0x5d
   print_report.cold+0x5e/0x5e5
   kasan_report+0xb1/0x1c0
   __nla_validate_parse+0x1258/0x2010
   __nla_parse+0x22/0x30
   hwsim_virtio_handle_cmd.isra.0+0x13f/0x2d0
   hwsim_virtio_rx_work+0x1b2/0x370
   process_one_work+0x8df/0x1530
   worker_thread+0x575/0x11a0
   kthread+0x29d/0x340
   ret_from_fork+0x22/0x30
 </TASK>

Discarding packets with an invalid length solves this.
Therefore, skb->len must be set at reception.

Change-Id: Ieaeb9a4c62d3beede274881a7c2722c6c6f477b6
Signed-off-by: Soenke Huster <soenke.huster@eknoes.de>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
22 months agowifi: mac80211: fix locking in auth/assoc timeout
Johannes Berg [Fri, 2 Sep 2022 14:11:14 +0000 (16:11 +0200)]
wifi: mac80211: fix locking in auth/assoc timeout

If we hit an authentication or association timeout, we only
release the chanctx for the deflink, and the other link(s)
are released later by ieee80211_vif_set_links(), but we're
not locking this correctly.

Fix the locking here while releasing the channels and links.

Change-Id: I9e08c1a5434592bdc75253c1abfa6c788f9f39b1
Fixes: 81151ce462e5 ("wifi: mac80211: support MLO authentication/association with one link")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
22 months agowifi: mac80211: mlme: release deflink channel in error case
Johannes Berg [Fri, 2 Sep 2022 14:11:15 +0000 (16:11 +0200)]
wifi: mac80211: mlme: release deflink channel in error case

In the prep_channel error case we didn't release the deflink
channel leaving it to be left around. Fix that.

Change-Id: If0dfd748125ec46a31fc6045a480dc28e03723d2
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
22 months agowifi: mac80211: fix link warning in RX agg timer expiry
Mukesh Sisodiya [Fri, 2 Sep 2022 14:11:31 +0000 (16:11 +0200)]
wifi: mac80211: fix link warning in RX agg timer expiry

The rx data link pointer isn't set from the RX aggregation timer,
resulting in a later warning. Fix that by setting it to the first
valid link for now, with a FIXME to worry about statistics later,
it's not very important since it's just the timeout case.

Reported-by: Hans de Goede <hdegoede@redhat.com>
Link: https://lore.kernel.org/r/498d714c-76be-9d04-26db-a1206878de5e@redhat.com
Fixes: 56057da4569b ("wifi: mac80211: rx: track link in RX data")
Signed-off-by: Mukesh Sisodiya <mukesh.sisodiya@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
22 months agoMerge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
David S. Miller [Sat, 3 Sep 2022 09:46:24 +0000 (10:46 +0100)]
Merge branch '40GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2022-09-02 (i40e, iavf)

This series contains updates to i40e and iavf drivers.

Przemyslaw adds reset to ADQ configuration to allow for setting of rate
limit beyond TC0 for i40e.

Ivan Vecera does not free client on failure to open which could cause
NULL pointer dereference to occur on i40e. He also detaches device
during reset to prevent NDO calls with could cause races for iavf.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net...
David S. Miller [Sat, 3 Sep 2022 09:44:01 +0000 (10:44 +0100)]
Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
This series contains updates to ice driver only.

Przemyslaw fixes memory leak of DMA memory due to incorrect freeing of
rx_buf.

Michal S corrects incorrect call to free memory.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: dsa: microchip: fix kernel oops on ksz8 switches
Oleksij Rempel [Fri, 2 Sep 2022 09:27:37 +0000 (11:27 +0200)]
net: dsa: microchip: fix kernel oops on ksz8 switches

After driver refactoring we was running ksz9477 specific CPU port
configuration on ksz8 family which ended with kernel oops. So, make sure
we run this code only on ksz9477 compatible devices.

Tested on KSZ8873 and KSZ9477.

Fixes: da8cd08520f3 ("net: dsa: microchip: add support for common phylink mac link up")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoxen-netback: only remove 'hotplug-status' when the vif is actually destroyed
Paul Durrant [Thu, 1 Sep 2022 11:55:54 +0000 (12:55 +0100)]
xen-netback: only remove 'hotplug-status' when the vif is actually destroyed

Removing 'hotplug-status' in backend_disconnected() means that it will be
removed even in the case that the frontend unilaterally disconnects (which
it is free to do at any time). The consequence of this is that, when the
frontend attempts to re-connect, the backend gets stuck in 'InitWait'
rather than moving straight to 'Connected' (which it can do because the
hotplug script has already run).
Instead, the 'hotplug-status' mode should be removed in netback_remove()
i.e. when the vif really is going away.

Fixes: 0f4558ae9187 ("Revert "xen-netback: remove 'hotplug-status' once it has served its purpose"")
Signed-off-by: Paul Durrant <pdurrant@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: fec: Use a spinlock to guard `fep->ptp_clk_on`
Csókás Bence [Thu, 1 Sep 2022 14:04:03 +0000 (16:04 +0200)]
net: fec: Use a spinlock to guard `fep->ptp_clk_on`

Mutexes cannot be taken in a non-preemptible context,
causing a panic in `fec_ptp_save_state()`. Replacing
`ptp_clk_mutex` by `tmreg_lock` fixes this.

Fixes: 6a4d7234ae9a ("net: fec: ptp: avoid register access when ipg clock is disabled")
Fixes: f79959220fa5 ("fec: Restart PPS after link state change")
Reported-by: Marc Kleine-Budde <mkl@pengutronix.de>
Link: https://lore.kernel.org/all/20220827160922.642zlcd5foopozru@pengutronix.de/
Signed-off-by: Csókás Bence <csokas.bence@prolan.hu>
Tested-by: Francesco Dolcini <francesco.dolcini@toradex.com> # Toradex Apalis iMX6
Link: https://lore.kernel.org/r/20220901140402.64804-1-csokas.bence@prolan.hu
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: fec: add pm_qos support on imx6q platform
Wei Fang [Tue, 30 Aug 2022 07:01:48 +0000 (15:01 +0800)]
net: fec: add pm_qos support on imx6q platform

There is a very low probability that tx timeout will occur during
suspend and resume stress test on imx6q platform. So we add pm_qos
support to prevent system from entering low level idles which may
affect the transmission of tx.

Signed-off-by: Wei Fang <wei.fang@nxp.com>
Link: https://lore.kernel.org/r/20220830070148.2021947-1-wei.fang@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf
Jakub Kicinski [Sat, 3 Sep 2022 02:38:25 +0000 (19:38 -0700)]
Merge git://git./linux/kernel/git/netfilter/nf

Florian Westphal says:

====================
netfilter: bug fixes for net

1. Fix IP address check in irc DCC conntrack helper, this should check
   the opposite direction rather than the destination address of the
   packets' direction, from David Leadbeater.

2. bridge netfilter needs to drop dst references, from Harsh Modi.
   This was fine back in the day the code was originally written,
   but nowadays various tunnels can pre-set metadata dsts on packets.

3. Remove nf_conntrack_helper sysctl and the modparam toggle, users
   need to explicitily assign the helpers to use via nftables or
   iptables.  Conntrack helpers, by design, may be used to add dynamic
   port redirections to internal machines, so its necessary to restrict
   which hosts/peers are allowed to use them.
   It was discovered that improper checking in the irc DCC helper makes
   it possible to trigger the 'please do dynamic port forward'
   from outside by embedding a 'DCC' in a PING request; if the client
   echos that back a expectation/port forward gets added.
   The auto-assign-for-everything mechanism has been in "please don't do this"
   territory since 2012.  From Pablo.

4. Fix a memory leak in the netdev hook error unwind path, also from Pablo.

* git://git.kernel.org/pub/scm/linux/kernel/git/netfilter/nf:
  netfilter: nf_conntrack_irc: Fix forged IP logic
  netfilter: nf_tables: clean up hook list when offload flags check fails
  netfilter: br_netfilter: Drop dst references before setting.
  netfilter: remove nf_conntrack_helper sysctl and modparam toggles
====================

Link: https://lore.kernel.org/r/20220901071238.3044-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoBluetooth: hci_sync: Fix hci_read_buffer_size_sync
Luiz Augusto von Dentz [Fri, 2 Sep 2022 00:24:13 +0000 (17:24 -0700)]
Bluetooth: hci_sync: Fix hci_read_buffer_size_sync

hci_read_buffer_size_sync shall not use HCI_OP_LE_READ_BUFFER_SIZE_V2
sinze that is LE specific, instead it is hci_le_read_buffer_size_sync
version that shall use it.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=216382
Fixes: 26afbd826ee3 ("Bluetooth: Add initial implementation of CIS connections")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
22 months agoiavf: Detach device during reset task
Ivan Vecera [Tue, 30 Aug 2022 08:16:27 +0000 (10:16 +0200)]
iavf: Detach device during reset task

iavf_reset_task() takes crit_lock at the beginning and holds
it during whole call. The function subsequently calls
iavf_init_interrupt_scheme() that grabs RTNL. Problem occurs
when userspace initiates during the reset task any ndo callback
that runs under RTNL like iavf_open() because some of that
functions tries to take crit_lock. This leads to classic A-B B-A
deadlock scenario.

To resolve this situation the device should be detached in
iavf_reset_task() prior taking crit_lock to avoid subsequent
ndos running under RTNL and reattach the device at the end.

Fixes: 62fe2a865e6d ("i40evf: add missing rtnl_lock() around i40evf_set_interrupt_capability")
Cc: Jacob Keller <jacob.e.keller@intel.com>
Cc: Patryk Piotrowski <patryk.piotrowski@intel.com>
Cc: SlawomirX Laba <slawomirx.laba@intel.com>
Tested-by: Vitaly Grinberg <vgrinber@redhat.com>
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Tested-by: Konrad Jankowski <konrad0.jankowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
22 months agoi40e: Fix kernel crash during module removal
Ivan Vecera [Tue, 16 Aug 2022 16:22:30 +0000 (18:22 +0200)]
i40e: Fix kernel crash during module removal

The driver incorrectly frees client instance and subsequent
i40e module removal leads to kernel crash.

Reproducer:
1. Do ethtool offline test followed immediately by another one
host# ethtool -t eth0 offline; ethtool -t eth0 offline
2. Remove recursively irdma module that also removes i40e module
host# modprobe -r irdma

Result:
[ 8675.035651] i40e 0000:3d:00.0 eno1: offline testing starting
[ 8675.193774] i40e 0000:3d:00.0 eno1: testing finished
[ 8675.201316] i40e 0000:3d:00.0 eno1: offline testing starting
[ 8675.358921] i40e 0000:3d:00.0 eno1: testing finished
[ 8675.496921] i40e 0000:3d:00.0: IRDMA hardware initialization FAILED init_state=2 status=-110
[ 8686.188955] i40e 0000:3d:00.1: i40e_ptp_stop: removed PHC on eno2
[ 8686.943890] i40e 0000:3d:00.1: Deleted LAN device PF1 bus=0x3d dev=0x00 func=0x01
[ 8686.952669] i40e 0000:3d:00.0: i40e_ptp_stop: removed PHC on eno1
[ 8687.761787] BUG: kernel NULL pointer dereference, address: 0000000000000030
[ 8687.768755] #PF: supervisor read access in kernel mode
[ 8687.773895] #PF: error_code(0x0000) - not-present page
[ 8687.779034] PGD 0 P4D 0
[ 8687.781575] Oops: 0000 [#1] PREEMPT SMP NOPTI
[ 8687.785935] CPU: 51 PID: 172891 Comm: rmmod Kdump: loaded Tainted: G        W I        5.19.0+ #2
[ 8687.794800] Hardware name: Intel Corporation S2600WFD/S2600WFD, BIOS SE5C620.86B.0X.02.0001.051420190324 05/14/2019
[ 8687.805222] RIP: 0010:i40e_lan_del_device+0x13/0xb0 [i40e]
[ 8687.810719] Code: d4 84 c0 0f 84 b8 25 01 00 e9 9c 25 01 00 41 bc f4 ff ff ff eb 91 90 0f 1f 44 00 00 41 54 55 53 48 8b 87 58 08 00 00 48 89 fb <48> 8b 68 30 48 89 ef e8 21 8a 0f d5 48 89 ef e8 a9 78 0f d5 48 8b
[ 8687.829462] RSP: 0018:ffffa604072efce0 EFLAGS: 00010202
[ 8687.834689] RAX: 0000000000000000 RBX: ffff8f43833b2000 RCX: 0000000000000000
[ 8687.841821] RDX: 0000000000000000 RSI: ffff8f4b0545b298 RDI: ffff8f43833b2000
[ 8687.848955] RBP: ffff8f43833b2000 R08: 0000000000000001 R09: 0000000000000000
[ 8687.856086] R10: 0000000000000000 R11: 000ffffffffff000 R12: ffff8f43833b2ef0
[ 8687.863218] R13: ffff8f43833b2ef0 R14: ffff915103966000 R15: ffff8f43833b2008
[ 8687.870342] FS:  00007f79501c3740(0000) GS:ffff8f4adffc0000(0000) knlGS:0000000000000000
[ 8687.878427] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 8687.884174] CR2: 0000000000000030 CR3: 000000014276e004 CR4: 00000000007706e0
[ 8687.891306] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 8687.898441] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 8687.905572] PKRU: 55555554
[ 8687.908286] Call Trace:
[ 8687.910737]  <TASK>
[ 8687.912843]  i40e_remove+0x2c0/0x330 [i40e]
[ 8687.917040]  pci_device_remove+0x33/0xa0
[ 8687.920962]  device_release_driver_internal+0x1aa/0x230
[ 8687.926188]  driver_detach+0x44/0x90
[ 8687.929770]  bus_remove_driver+0x55/0xe0
[ 8687.933693]  pci_unregister_driver+0x2a/0xb0
[ 8687.937967]  i40e_exit_module+0xc/0xf48 [i40e]

Two offline tests cause IRDMA driver failure (ETIMEDOUT) and this
failure is indicated back to i40e_client_subtask() that calls
i40e_client_del_instance() to free client instance referenced
by pf->cinst and sets this pointer to NULL. During the module
removal i40e_remove() calls i40e_lan_del_device() that dereferences
pf->cinst that is NULL -> crash.
Do not remove client instance when client open callbacks fails and
just clear __I40E_CLIENT_INSTANCE_OPENED bit. The driver also needs
to take care about this situation (when netdev is up and client
is NOT opened) in i40e_notify_client_of_netdev_close() and
calls client close callback only when __I40E_CLIENT_INSTANCE_OPENED
is set.

Fixes: 0ef2d5afb12d ("i40e: KISS the client interface")
Signed-off-by: Ivan Vecera <ivecera@redhat.com>
Tested-by: Helena Anna Dubel <helena.anna.dubel@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
22 months agoi40e: Fix ADQ rate limiting for PF
Przemyslaw Patynowski [Tue, 9 Aug 2022 08:57:44 +0000 (10:57 +0200)]
i40e: Fix ADQ rate limiting for PF

Fix HW rate limiting for ADQ.
Fallback to kernel queue selection for ADQ, as it is network stack
that decides which queue to use for transmit with ADQ configured.
Reset PF after creation of VMDq2 VSIs required for ADQ, as to
reprogram TX queue contexts in i40e_configure_tx_ring.
Without this patch PF would limit TX rate only according to TC0.

Fixes: a9ce82f744dc ("i40e: Enable 'channel' mode in mqprio for TC configs")
Signed-off-by: Przemyslaw Patynowski <przemyslawx.patynowski@intel.com>
Signed-off-by: Jan Sokolowski <jan.sokolowski@intel.com>
Tested-by: Bharathi Sreenivas <bharathi.sreenivas@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
22 months agoice: use bitmap_free instead of devm_kfree
Michal Swiatkowski [Wed, 17 Aug 2022 08:53:20 +0000 (10:53 +0200)]
ice: use bitmap_free instead of devm_kfree

pf->avail_txqs was allocated using bitmap_zalloc, bitmap_free should be
used to free this memory.

Fixes: 78b5713ac1241 ("ice: Alloc queue management bitmaps and arrays dynamically")
Signed-off-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Tested-by: Gurucharan <gurucharanx.g@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
22 months agoice: Fix DMA mappings leak
Przemyslaw Patynowski [Thu, 11 Aug 2022 10:09:22 +0000 (12:09 +0200)]
ice: Fix DMA mappings leak

Fix leak, when user changes ring parameters.
During reallocation of RX buffers, new DMA mappings are created for
those buffers. New buffers with different RX ring count should
substitute older ones, but those buffers were freed in ice_vsi_cfg_rxq
and reallocated again with ice_alloc_rx_buf. kfree on rx_buf caused
leak of already mapped DMA.
Reallocate ZC with xdp_buf struct, when BPF program loads. Reallocate
back to rx_buf, when BPF program unloads.
If BPF program is loaded/unloaded and XSK pools are created, reallocate
RX queues accordingly in XDP_SETUP_XSK_POOL handler.

Steps for reproduction:
while :
do
for ((i=0; i<=8160; i=i+32))
do
ethtool -G enp130s0f0 rx $i tx $i
sleep 0.5
ethtool -g enp130s0f0
done
done

Fixes: 617f3e1b588c ("ice: xsk: allocate separate memory for XDP SW ring")
Signed-off-by: Przemyslaw Patynowski <przemyslawx.patynowski@intel.com>
Signed-off-by: Mateusz Palczewski <mateusz.palczewski@intel.com>
Tested-by: Chandan <chandanx.rout@intel.com> (A Contingent Worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
22 months agoMerge tag 'rxrpc-fixes-20220901' of git://git.kernel.org/pub/scm/linux/kernel/git...
David S. Miller [Fri, 2 Sep 2022 11:45:32 +0000 (12:45 +0100)]
Merge tag 'rxrpc-fixes-20220901' of git://git./linux/kernel/git/dhowells/linux-fs

David Howells says:

====================
rxrpc fixes
Here are some fixes for AF_RXRPC:

 (1) Fix the handling of ICMP/ICMP6 packets.  This is a problem due to
     rxrpc being switched to acting as a UDP tunnel, thereby allowing it to
     steal the packets before they go through the UDP Rx queue.  UDP
     tunnels can't get ICMP/ICMP6 packets, however.  This patch adds an
     additional encap hook so that they can.

 (2) Fix the encryption routines in rxkad to handle packets that have more
     than three parts correctly.  The problem is that ->nr_frags doesn't
     count the initial fragment, so the sglist ends up too short.

 (3) Fix a problem with destruction of the local endpoint potentially
     getting repeated.

 (4) Fix the calculation of the time at which to resend.
     jiffies_to_usecs() gives microseconds, not nanoseconds.

 (5) Fix AFS to work out when callback promises and locks expire based on
     the time an op was issued rather than the time the first reply packet
     arrives.  We don't know how long the server took between calculating
     the expiry interval and transmitting the reply.

 (6) Given (5), rxrpc_get_reply_time() is no longer used, so remove it.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agotcp: TX zerocopy should not sense pfmemalloc status
Eric Dumazet [Wed, 31 Aug 2022 23:38:09 +0000 (23:38 +0000)]
tcp: TX zerocopy should not sense pfmemalloc status

We got a recent syzbot report [1] showing a possible misuse
of pfmemalloc page status in TCP zerocopy paths.

Indeed, for pages coming from user space or other layers,
using page_is_pfmemalloc() is moot, and possibly could give
false positives.

There has been attempts to make page_is_pfmemalloc() more robust,
but not using it in the first place in this context is probably better,
removing cpu cycles.

Note to stable teams :

You need to backport 84ce071e38a6 ("net: introduce
__skb_fill_page_desc_noacc") as a prereq.

Race is more probable after commit c07aea3ef4d4
("mm: add a signature in struct page") because page_is_pfmemalloc()
is now using low order bit from page->lru.next, which can change
more often than page->index.

Low order bit should never be set for lru.next (when used as an anchor
in LRU list), so KCSAN report is mostly a false positive.

Backporting to older kernel versions seems not necessary.

[1]
BUG: KCSAN: data-race in lru_add_fn / tcp_build_frag

write to 0xffffea0004a1d2c8 of 8 bytes by task 18600 on cpu 0:
__list_add include/linux/list.h:73 [inline]
list_add include/linux/list.h:88 [inline]
lruvec_add_folio include/linux/mm_inline.h:105 [inline]
lru_add_fn+0x440/0x520 mm/swap.c:228
folio_batch_move_lru+0x1e1/0x2a0 mm/swap.c:246
folio_batch_add_and_move mm/swap.c:263 [inline]
folio_add_lru+0xf1/0x140 mm/swap.c:490
filemap_add_folio+0xf8/0x150 mm/filemap.c:948
__filemap_get_folio+0x510/0x6d0 mm/filemap.c:1981
pagecache_get_page+0x26/0x190 mm/folio-compat.c:104
grab_cache_page_write_begin+0x2a/0x30 mm/folio-compat.c:116
ext4_da_write_begin+0x2dd/0x5f0 fs/ext4/inode.c:2988
generic_perform_write+0x1d4/0x3f0 mm/filemap.c:3738
ext4_buffered_write_iter+0x235/0x3e0 fs/ext4/file.c:270
ext4_file_write_iter+0x2e3/0x1210
call_write_iter include/linux/fs.h:2187 [inline]
new_sync_write fs/read_write.c:491 [inline]
vfs_write+0x468/0x760 fs/read_write.c:578
ksys_write+0xe8/0x1a0 fs/read_write.c:631
__do_sys_write fs/read_write.c:643 [inline]
__se_sys_write fs/read_write.c:640 [inline]
__x64_sys_write+0x3e/0x50 fs/read_write.c:640
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x2b/0x70 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd

read to 0xffffea0004a1d2c8 of 8 bytes by task 18611 on cpu 1:
page_is_pfmemalloc include/linux/mm.h:1740 [inline]
__skb_fill_page_desc include/linux/skbuff.h:2422 [inline]
skb_fill_page_desc include/linux/skbuff.h:2443 [inline]
tcp_build_frag+0x613/0xb20 net/ipv4/tcp.c:1018
do_tcp_sendpages+0x3e8/0xaf0 net/ipv4/tcp.c:1075
tcp_sendpage_locked net/ipv4/tcp.c:1140 [inline]
tcp_sendpage+0x89/0xb0 net/ipv4/tcp.c:1150
inet_sendpage+0x7f/0xc0 net/ipv4/af_inet.c:833
kernel_sendpage+0x184/0x300 net/socket.c:3561
sock_sendpage+0x5a/0x70 net/socket.c:1054
pipe_to_sendpage+0x128/0x160 fs/splice.c:361
splice_from_pipe_feed fs/splice.c:415 [inline]
__splice_from_pipe+0x222/0x4d0 fs/splice.c:559
splice_from_pipe fs/splice.c:594 [inline]
generic_splice_sendpage+0x89/0xc0 fs/splice.c:743
do_splice_from fs/splice.c:764 [inline]
direct_splice_actor+0x80/0xa0 fs/splice.c:931
splice_direct_to_actor+0x305/0x620 fs/splice.c:886
do_splice_direct+0xfb/0x180 fs/splice.c:974
do_sendfile+0x3bf/0x910 fs/read_write.c:1249
__do_sys_sendfile64 fs/read_write.c:1317 [inline]
__se_sys_sendfile64 fs/read_write.c:1303 [inline]
__x64_sys_sendfile64+0x10c/0x150 fs/read_write.c:1303
do_syscall_x64 arch/x86/entry/common.c:50 [inline]
do_syscall_64+0x2b/0x70 arch/x86/entry/common.c:80
entry_SYSCALL_64_after_hwframe+0x63/0xcd

value changed: 0x0000000000000000 -> 0xffffea0004a1d288

Reported by Kernel Concurrency Sanitizer on:
CPU: 1 PID: 18611 Comm: syz-executor.4 Not tainted 6.0.0-rc2-syzkaller-00248-ge022620b5d05-dirty #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 07/22/2022

Fixes: c07aea3ef4d4 ("mm: add a signature in struct page")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Shakeel Butt <shakeelb@google.com>
Reviewed-by: Shakeel Butt <shakeelb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agotipc: fix shift wrapping bug in map_get()
Dan Carpenter [Wed, 31 Aug 2022 14:47:56 +0000 (17:47 +0300)]
tipc: fix shift wrapping bug in map_get()

There is a shift wrapping bug in this code so anything thing above
31 will return false.

Fixes: 35c55c9877f8 ("tipc: add neighbor monitoring framework")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agosch_sfb: Don't assume the skb is still around after enqueueing to child
Toke Høiland-Jørgensen [Wed, 31 Aug 2022 21:52:18 +0000 (23:52 +0200)]
sch_sfb: Don't assume the skb is still around after enqueueing to child

The sch_sfb enqueue() routine assumes the skb is still alive after it has
been enqueued into a child qdisc, using the data in the skb cb field in the
increment_qlen() routine after enqueue. However, the skb may in fact have
been freed, causing a use-after-free in this case. In particular, this
happens if sch_cake is used as a child of sfb, and the GSO splitting mode
of CAKE is enabled (in which case the skb will be split into segments and
the original skb freed).

Fix this by copying the sfb cb data to the stack before enqueueing the skb,
and using this stack copy in increment_qlen() instead of the skb pointer
itself.

Reported-by: zdi-disclosures@trendmicro.com # ZDI-CAN-18231
Fixes: e13e02a3c68d ("net_sched: SFB flow scheduler")
Signed-off-by: Toke Høiland-Jørgensen <toke@toke.dk>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoRevert "net: phy: meson-gxl: improve link-up behavior"
Heiner Kallweit [Wed, 31 Aug 2022 19:20:49 +0000 (21:20 +0200)]
Revert "net: phy: meson-gxl: improve link-up behavior"

This reverts commit 2c87c6f9fbddc5b84d67b2fa3f432fcac6d99d93.
Meanwhile it turned out that the following commit is the proper
workaround for the issue that 2c87c6f9fbdd tries to address.
a3a57bf07de2 ("net: stmmac: work around sporadic tx issue on link-up")
It's nor clear why the to be reverted commit helped for one user,
for others it didn't make a difference.

Fixes: 2c87c6f9fbdd ("net: phy: meson-gxl: improve link-up behavior")
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/8deeeddc-6b71-129b-1918-495a12dc11e3@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge tag 'net-6.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 1 Sep 2022 16:20:42 +0000 (09:20 -0700)]
Merge tag 'net-6.0-rc4' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Paolo Abeni:
 "Including fixes from bluetooth, bpf and wireless.

  Current release - regressions:

   - bpf:
      - fix wrong last sg check in sk_msg_recvmsg()
      - fix kernel BUG in purge_effective_progs()

   - mac80211:
      - fix possible leak in ieee80211_tx_control_port()
      - potential NULL dereference in ieee80211_tx_control_port()

  Current release - new code bugs:

   - nfp: fix the access to management firmware hanging

  Previous releases - regressions:

   - ip: fix triggering of 'icmp redirect'

   - sched: tbf: don't call qdisc_put() while holding tree lock

   - bpf: fix corrupted packets for XDP_SHARED_UMEM

   - bluetooth: hci_sync: fix suspend performance regression

   - micrel: fix probe failure

  Previous releases - always broken:

   - tcp: make global challenge ack rate limitation per net-ns and
     default disabled

   - tg3: fix potential hang-up on system reboot

   - mac802154: fix reception for no-daddr packets

  Misc:

   - r8152: add PID for the lenovo onelink+ dock"

* tag 'net-6.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (56 commits)
  net/smc: Remove redundant refcount increase
  Revert "sch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb"
  tcp: make global challenge ack rate limitation per net-ns and default disabled
  tcp: annotate data-race around challenge_timestamp
  net: dsa: hellcreek: Print warning only once
  ip: fix triggering of 'icmp redirect'
  sch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb
  selftests: net: sort .gitignore file
  Documentation: networking: correct possessive "its"
  kcm: fix strp_init() order and cleanup
  mlxbf_gige: compute MDIO period based on i1clk
  ethernet: rocker: fix sleep in atomic context bug in neigh_timer_handler
  net: lan966x: improve error handle in lan966x_fdma_rx_get_frame()
  nfp: fix the access to management firmware hanging
  net: phy: micrel: Make the GPIO to be non-exclusive
  net: virtio_net: fix notification coalescing comments
  net/sched: fix netdevice reference leaks in attach_default_qdiscs()
  net: sched: tbf: don't call qdisc_put() while holding tree lock
  net: Use u64_stats_fetch_begin_irq() for stats fetch.
  net: dsa: xrs700x: Use irqsave variant for u64 stats update
  ...

22 months agoMerge tag 'slab-for-6.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka...
Linus Torvalds [Thu, 1 Sep 2022 16:14:56 +0000 (09:14 -0700)]
Merge tag 'slab-for-6.0-rc4' of git://git./linux/kernel/git/vbabka/slab

Pull slab fix from Vlastimil Babka:

 - A fix from Waiman Long to avoid a theoretical deadlock reported by
   lockdep.

* tag 'slab-for-6.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/vbabka/slab:
  mm/slab_common: Deleting kobject in kmem_cache_destroy() without holding slab_mutex/cpu_hotplug_lock

22 months agoMerge tag 'sound-6.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai...
Linus Torvalds [Thu, 1 Sep 2022 16:05:25 +0000 (09:05 -0700)]
Merge tag 'sound-6.0-rc4' of git://git./linux/kernel/git/tiwai/sound

Pull sound fixes from Takashi Iwai:
 "Just handful changes at this time. The only major change is the
  regression fix about the x86 WC-page buffer allocation.

  The rest are trivial data-race fixes for ALSA sequencer core, the
  possible out-of-bounds access fixes in the new ALSA control hash code,
  and a few device-specific workarounds and fixes"

* tag 'sound-6.0-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound:
  ALSA: usb-audio: Add quirk for LH Labs Geek Out HD Audio 1V5
  ALSA: hda/realtek: Add speaker AMP init for Samsung laptops with ALC298
  ALSA: control: Re-order bounds checking in get_ctl_id_hash()
  ALSA: control: Fix an out-of-bounds bug in get_ctl_id_hash()
  ALSA: hda: intel-nhlt: Correct the handling of fmt_config flexible array
  ALSA: seq: Fix data-race at module auto-loading
  ALSA: seq: oss: Fix data-race for max_midi_devs access
  ALSA: memalloc: Revive x86-specific WC page allocations again

22 months agorxrpc: Remove rxrpc_get_reply_time() which is no longer used
David Howells [Wed, 31 Aug 2022 12:24:44 +0000 (13:24 +0100)]
rxrpc: Remove rxrpc_get_reply_time() which is no longer used

Remove rxrpc_get_reply_time() as that is no longer used now that the call
issue time is used instead of the reply time.

Signed-off-by: David Howells <dhowells@redhat.com>
22 months agoafs: Use the operation issue time instead of the reply time for callbacks
David Howells [Wed, 31 Aug 2022 12:16:42 +0000 (13:16 +0100)]
afs: Use the operation issue time instead of the reply time for callbacks

rxrpc and kafs between them try to use the receive timestamp on the first
data packet (ie. the one with sequence number 1) as a base from which to
calculate the time at which callback promise and lock expiration occurs.

However, we don't know how long it took for the server to send us the reply
from it having completed the basic part of the operation - it might then,
for instance, have to send a bunch of a callback breaks, depending on the
particular operation.

Fix this by using the time at which the operation is issued on the client
as a base instead.  That should never be longer than the server's idea of
the expiry time.

Fixes: 781070551c26 ("afs: Fix calculation of callback expiry time")
Fixes: 2070a3e44962 ("rxrpc: Allow the reply time to be obtained on a client call")
Suggested-by: Jeffrey E Altman <jaltman@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
22 months agorxrpc: Fix calc of resend age
David Howells [Tue, 5 Apr 2022 12:34:09 +0000 (13:34 +0100)]
rxrpc: Fix calc of resend age

Fix the calculation of the resend age to add a microsecond value as
microseconds, not nanoseconds.

Signed-off-by: David Howells <dhowells@redhat.com>
22 months agorxrpc: Fix local destruction being repeated
David Howells [Fri, 6 May 2022 22:55:21 +0000 (23:55 +0100)]
rxrpc: Fix local destruction being repeated

If the local processor work item for the rxrpc local endpoint gets requeued
by an event (such as an incoming packet) between it getting scheduled for
destruction and the UDP socket being closed, the rxrpc_local_destroyer()
function can get run twice.  The second time it can hang because it can end
up waiting for cleanup events that will never happen.

Signed-off-by: David Howells <dhowells@redhat.com>
22 months agorxrpc: Fix an insufficiently large sglist in rxkad_verify_packet_2()
David Howells [Wed, 24 Aug 2022 21:39:28 +0000 (22:39 +0100)]
rxrpc: Fix an insufficiently large sglist in rxkad_verify_packet_2()

rxkad_verify_packet_2() has a small stack-allocated sglist of 4 elements,
but if that isn't sufficient for the number of fragments in the socket
buffer, we try to allocate an sglist large enough to hold all the
fragments.

However, for large packets with a lot of fragments, this isn't sufficient
and we need at least one additional fragment.

The problem manifests as skb_to_sgvec() returning -EMSGSIZE and this then
getting returned by userspace.  Most of the time, this isn't a problem as
rxrpc sets a limit of 5692, big enough for 4 jumbo subpackets to be glued
together; occasionally, however, the server will ignore the reported limit
and give a packet that's a lot bigger - say 19852 bytes with ->nr_frags
being 7.  skb_to_sgvec() then tries to return a "zeroth" fragment that
seems to occur before the fragments counted by ->nr_frags and we hit the
end of the sglist too early.

Note that __skb_to_sgvec() also has an skb_walk_frags() loop that is
recursive up to 24 deep.  I'm not sure if I need to take account of that
too - or if there's an easy way of counting those frags too.

Fix this by counting an extra frag and allocating a larger sglist based on
that.

Fixes: d0d5c0cd1e71 ("rxrpc: Use skb_unshare() rather than skb_cow_data()")
Reported-by: Marc Dionne <marc.dionne@auristor.com>
Signed-off-by: David Howells <dhowells@redhat.com>
cc: linux-afs@lists.infradead.org

22 months agorxrpc: Fix ICMP/ICMP6 error handling
David Howells [Fri, 26 Aug 2022 14:39:28 +0000 (15:39 +0100)]
rxrpc: Fix ICMP/ICMP6 error handling

Because rxrpc pretends to be a tunnel on top of a UDP/UDP6 socket, allowing
it to siphon off UDP packets early in the handling of received UDP packets
thereby avoiding the packet going through the UDP receive queue, it doesn't
get ICMP packets through the UDP ->sk_error_report() callback.  In fact, it
doesn't appear that there's any usable option for getting hold of ICMP
packets.

Fix this by adding a new UDP encap hook to distribute error messages for
UDP tunnels.  If the hook is set, then the tunnel driver will be able to
see ICMP packets.  The hook provides the offset into the packet of the UDP
header of the original packet that caused the notification.

An alternative would be to call the ->error_handler() hook - but that
requires that the skbuff be cloned (as ip_icmp_error() or ipv6_cmp_error()
do, though isn't really necessary or desirable in rxrpc's case is we want
to parse them there and then, not queue them).

Changes
=======
ver #3)
 - Fixed an uninitialised variable.

ver #2)
 - Fixed some missing CONFIG_AF_RXRPC_IPV6 conditionals.

Fixes: 5271953cad31 ("rxrpc: Use the UDP encap_rcv hook")
Signed-off-by: David Howells <dhowells@redhat.com>
22 months agomm/slab_common: Deleting kobject in kmem_cache_destroy() without holding slab_mutex...
Waiman Long [Fri, 12 Aug 2022 18:30:33 +0000 (14:30 -0400)]
mm/slab_common: Deleting kobject in kmem_cache_destroy() without holding slab_mutex/cpu_hotplug_lock

A circular locking problem is reported by lockdep due to the following
circular locking dependency.

  +--> cpu_hotplug_lock --> slab_mutex --> kn->active --+
  |                                                     |
  +-----------------------------------------------------+

The forward cpu_hotplug_lock ==> slab_mutex ==> kn->active dependency
happens in

  kmem_cache_destroy(): cpus_read_lock(); mutex_lock(&slab_mutex);
  ==> sysfs_slab_unlink()
      ==> kobject_del()
          ==> kernfs_remove()
      ==> __kernfs_remove()
          ==> kernfs_drain(): rwsem_acquire(&kn->dep_map, ...);

The backward kn->active ==> cpu_hotplug_lock dependency happens in

  kernfs_fop_write_iter(): kernfs_get_active();
  ==> slab_attr_store()
      ==> cpu_partial_store()
          ==> flush_all(): cpus_read_lock()

One way to break this circular locking chain is to avoid holding
cpu_hotplug_lock and slab_mutex while deleting the kobject in
sysfs_slab_unlink() which should be equivalent to doing a write_lock
and write_unlock pair of the kn->active virtual lock.

Since the kobject structures are not protected by slab_mutex or the
cpu_hotplug_lock, we can certainly release those locks before doing
the delete operation.

Move sysfs_slab_unlink() and sysfs_slab_release() to the newly
created kmem_cache_release() and call it outside the slab_mutex &
cpu_hotplug_lock critical sections. There will be a slight delay
in the deletion of sysfs files if kmem_cache_release() is called
indirectly from a work function.

Fixes: 5a836bf6b09f ("mm: slub: move flush_cpu_slab() invocations __free_slab() invocations out of IRQ context")
Signed-off-by: Waiman Long <longman@redhat.com>
Reviewed-by: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Reviewed-by: Roman Gushchin <roman.gushchin@linux.dev>
Acked-by: David Rientjes <rientjes@google.com>
Link: https://lore.kernel.org/all/YwOImVd+nRUsSAga@hyeyoo/
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
22 months agonet/smc: Remove redundant refcount increase
Yacan Liu [Tue, 30 Aug 2022 15:23:14 +0000 (23:23 +0800)]
net/smc: Remove redundant refcount increase

For passive connections, the refcount increment has been done in
smc_clcsock_accept()-->smc_sock_alloc().

Fixes: 3b2dec2603d5 ("net/smc: restructure client and server code in af_smc")
Signed-off-by: Yacan Liu <liuyacan@corp.netease.com>
Reviewed-by: Tony Lu <tonylu@linux.alibaba.com>
Link: https://lore.kernel.org/r/20220830152314.838736-1-liuyacan@corp.netease.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
22 months agoRevert "sch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb"
Jakub Kicinski [Thu, 1 Sep 2022 03:01:32 +0000 (20:01 -0700)]
Revert "sch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb"

This reverts commit 90fabae8a2c225c4e4936723c38857887edde5cc.

Patch was applied hastily, revert and let the v2 be reviewed.

Fixes: 90fabae8a2c2 ("sch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb")
Link: https://lore.kernel.org/all/87wnao2ha3.fsf@toke.dk/
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge branch 'tcp-tcp-challenge-ack-fixes'
Jakub Kicinski [Thu, 1 Sep 2022 02:56:53 +0000 (19:56 -0700)]
Merge branch 'tcp-tcp-challenge-ack-fixes'

Eric Dumazet says:

====================
tcp: tcp challenge ack fixes

syzbot found a typical data-race addressed in the first patch.

While we are at it, second patch makes the global rate limit
per net-ns and disabled by default.
====================

Link: https://lore.kernel.org/r/20220830185656.268523-1-eric.dumazet@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agotcp: make global challenge ack rate limitation per net-ns and default disabled
Eric Dumazet [Tue, 30 Aug 2022 18:56:56 +0000 (11:56 -0700)]
tcp: make global challenge ack rate limitation per net-ns and default disabled

Because per host rate limiting has been proven problematic (side channel
attacks can be based on it), per host rate limiting of challenge acks ideally
should be per netns and turned off by default.

This is a long due followup of following commits:

083ae308280d ("tcp: enable per-socket rate limiting of all 'challenge acks'")
f2b2c582e824 ("tcp: mitigate ACK loops for connections as tcp_sock")
75ff39ccc1bd ("tcp: make challenge acks less predictable")

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Jason Baron <jbaron@akamai.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agotcp: annotate data-race around challenge_timestamp
Eric Dumazet [Tue, 30 Aug 2022 18:56:55 +0000 (11:56 -0700)]
tcp: annotate data-race around challenge_timestamp

challenge_timestamp can be read an written by concurrent threads.

This was expected, but we need to annotate the race to avoid potential issues.

Following patch moves challenge_timestamp and challenge_count
to per-netns storage to provide better isolation.

Fixes: 354e4aa391ed ("tcp: RFC 5961 5.2 Blind Data Injection Attack Mitigation")
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: dsa: hellcreek: Print warning only once
Kurt Kanzenbach [Tue, 30 Aug 2022 16:34:48 +0000 (18:34 +0200)]
net: dsa: hellcreek: Print warning only once

In case the source port cannot be decoded, print the warning only once. This
still brings attention to the user and does not spam the logs at the same time.

Signed-off-by: Kurt Kanzenbach <kurt@linutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Link: https://lore.kernel.org/r/20220830163448.8921-1-kurt@linutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoip: fix triggering of 'icmp redirect'
Nicolas Dichtel [Mon, 29 Aug 2022 10:01:21 +0000 (12:01 +0200)]
ip: fix triggering of 'icmp redirect'

__mkroute_input() uses fib_validate_source() to trigger an icmp redirect.
My understanding is that fib_validate_source() is used to know if the src
address and the gateway address are on the same link. For that,
fib_validate_source() returns 1 (same link) or 0 (not the same network).
__mkroute_input() is the only user of these positive values, all other
callers only look if the returned value is negative.

Since the below patch, fib_validate_source() didn't return anymore 1 when
both addresses are on the same network, because the route lookup returns
RT_SCOPE_LINK instead of RT_SCOPE_HOST. But this is, in fact, right.
Let's adapat the test to return 1 again when both addresses are on the same
link.

CC: stable@vger.kernel.org
Fixes: 747c14307214 ("ip: fix dflt addr selection for connected nexthop")
Reported-by: kernel test robot <yujie.liu@intel.com>
Reported-by: Heng Qi <hengqi@linux.alibaba.com>
Signed-off-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20220829100121.3821-1-nicolas.dichtel@6wind.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonetfilter: nf_conntrack_irc: Fix forged IP logic
David Leadbeater [Fri, 26 Aug 2022 04:56:58 +0000 (14:56 +1000)]
netfilter: nf_conntrack_irc: Fix forged IP logic

Ensure the match happens in the right direction, previously the
destination used was the server, not the NAT host, as the comment
shows the code intended.

Additionally nf_nat_irc uses port 0 as a signal and there's no valid way
it can appear in a DCC message, so consider port 0 also forged.

Fixes: 869f37d8e48f ("[NETFILTER]: nf_conntrack/nf_nat: add IRC helper port")
Signed-off-by: David Leadbeater <dgl@dgl.cx>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
22 months agomm/rmap: Fix anon_vma->degree ambiguity leading to double-reuse
Jann Horn [Wed, 31 Aug 2022 17:06:00 +0000 (19:06 +0200)]
mm/rmap: Fix anon_vma->degree ambiguity leading to double-reuse

anon_vma->degree tracks the combined number of child anon_vmas and VMAs
that use the anon_vma as their ->anon_vma.

anon_vma_clone() then assumes that for any anon_vma attached to
src->anon_vma_chain other than src->anon_vma, it is impossible for it to
be a leaf node of the VMA tree, meaning that for such VMAs ->degree is
elevated by 1 because of a child anon_vma, meaning that if ->degree
equals 1 there are no VMAs that use the anon_vma as their ->anon_vma.

This assumption is wrong because the ->degree optimization leads to leaf
nodes being abandoned on anon_vma_clone() - an existing anon_vma is
reused and no new parent-child relationship is created.  So it is
possible to reuse an anon_vma for one VMA while it is still tied to
another VMA.

This is an issue because is_mergeable_anon_vma() and its callers assume
that if two VMAs have the same ->anon_vma, the list of anon_vmas
attached to the VMAs is guaranteed to be the same.  When this assumption
is violated, vma_merge() can merge pages into a VMA that is not attached
to the corresponding anon_vma, leading to dangling page->mapping
pointers that will be dereferenced during rmap walks.

Fix it by separately tracking the number of child anon_vmas and the
number of VMAs using the anon_vma as their ->anon_vma.

Fixes: 7a3ef208e662 ("mm: prevent endless growth of anon_vma hierarchy")
Cc: stable@kernel.org
Acked-by: Michal Hocko <mhocko@suse.com>
Acked-by: Vlastimil Babka <vbabka@suse.cz>
Signed-off-by: Jann Horn <jannh@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
22 months agosch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb
Toke Høiland-Jørgensen [Wed, 31 Aug 2022 09:21:03 +0000 (11:21 +0200)]
sch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb

When the GSO splitting feature of sch_cake is enabled, GSO superpackets
will be broken up and the resulting segments enqueued in place of the
original skb. In this case, CAKE calls consume_skb() on the original skb,
but still returns NET_XMIT_SUCCESS. This can confuse parent qdiscs into
assuming the original skb still exists, when it really has been freed. Fix
this by adding the __NET_XMIT_STOLEN flag to the return value in this case.

Fixes: 0c850344d388 ("sch_cake: Conditionally split GSO segments")
Signed-off-by: Toke Høiland-Jørgensen <toke@toke.dk>
Reported-by: zdi-disclosures@trendmicro.com # ZDI-CAN-18231
Link: https://lore.kernel.org/r/20220831092103.442868-1-toke@toke.dk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoselftests: net: sort .gitignore file
Axel Rasmussen [Mon, 29 Aug 2022 18:47:48 +0000 (11:47 -0700)]
selftests: net: sort .gitignore file

This is the result of `sort tools/testing/selftests/net/.gitignore`, but
preserving the comment at the top.

Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Axel Rasmussen <axelrasmussen@google.com>
Link: https://lore.kernel.org/r/20220829184748.1535580-1-axelrasmussen@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoDocumentation: networking: correct possessive "its"
Randy Dunlap [Mon, 29 Aug 2022 23:54:14 +0000 (16:54 -0700)]
Documentation: networking: correct possessive "its"

Change occurrences of "it's" that are possessive to "its"
so that they don't read as "it is".

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20220829235414.17110-1-rdunlap@infradead.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agokcm: fix strp_init() order and cleanup
Cong Wang [Sat, 27 Aug 2022 18:13:14 +0000 (11:13 -0700)]
kcm: fix strp_init() order and cleanup

strp_init() is called just a few lines above this csk->sk_user_data
check, it also initializes strp->work etc., therefore, it is
unnecessary to call strp_done() to cancel the freshly initialized
work.

And if sk_user_data is already used by KCM, psock->strp should not be
touched, particularly strp->work state, so we need to move strp_init()
after the csk->sk_user_data check.

This also makes a lockdep warning reported by syzbot go away.

Reported-and-tested-by: syzbot+9fc084a4348493ef65d2@syzkaller.appspotmail.com
Reported-by: syzbot+e696806ef96cdd2d87cd@syzkaller.appspotmail.com
Fixes: e5571240236c ("kcm: Check if sk_user_data already set in kcm_attach")
Fixes: dff8baa26117 ("kcm: Call strp_stop before strp_done in kcm_attach")
Cc: Tom Herbert <tom@herbertland.com>
Signed-off-by: Cong Wang <cong.wang@bytedance.com>
Link: https://lore.kernel.org/r/20220827181314.193710-1-xiyou.wangcong@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agomlxbf_gige: compute MDIO period based on i1clk
David Thompson [Fri, 26 Aug 2022 15:59:16 +0000 (11:59 -0400)]
mlxbf_gige: compute MDIO period based on i1clk

This patch adds logic to compute the MDIO period based on
the i1clk, and thereafter write the MDIO period into the YU
MDIO config register. The i1clk resource from the ACPI table
is used to provide addressing to YU bootrecord PLL registers.
The values in these registers are used to compute MDIO period.
If the i1clk resource is not present in the ACPI table, then
the current default hardcorded value of 430Mhz is used.
The i1clk clock value of 430MHz is only accurate for boards
with BF2 mid bin and main bin SoCs. The BF2 high bin SoCs
have i1clk = 500MHz, but can support a slower MDIO period.

Fixes: f92e1869d74e ("Add Mellanox BlueField Gigabit Ethernet driver")
Reviewed-by: Asmaa Mnebhi <asmaa@nvidia.com>
Signed-off-by: David Thompson <davthompson@nvidia.com>
Link: https://lore.kernel.org/r/20220826155916.12491-1-davthompson@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge tag 'fscache-fixes-20220831' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 31 Aug 2022 17:13:34 +0000 (10:13 -0700)]
Merge tag 'fscache-fixes-20220831' of git://git./linux/kernel/git/dhowells/linux-fs

Pull fscache/cachefiles fixes from David Howells:

 - Fix kdoc on fscache_use/unuse_cookie().

 - Fix the error returned by cachefiles_ondemand_copen() from an upcall
   result.

 - Fix the distribution of requests in on-demand mode in cachefiles to
   be fairer by cycling through them rather than picking the one with
   the lowest ID each time (IDs being reused).

* tag 'fscache-fixes-20220831' of git://git.kernel.org/pub/scm/linux/kernel/git/dhowells/linux-fs:
  cachefiles: make on-demand request distribution fairer
  cachefiles: fix error return code in cachefiles_ondemand_copen()
  fscache: fix misdocumented parameter

22 months agoMerge tag 'for-linus-2022083101' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Wed, 31 Aug 2022 16:54:14 +0000 (09:54 -0700)]
Merge tag 'for-linus-2022083101' of git://git./linux/kernel/git/hid/hid

Pull HID fixes from Jiri Kosina:

 - NULL pointer dereference fix for Steam driver (Lee Jones)

 - memory leak fix for hidraw (Karthik Alapati)

 - regression fix for functionality of some UCLogic tables (Benjamin
   Tissoires)

 - a few new device IDs and device-specific quirks

* tag 'for-linus-2022083101' of git://git.kernel.org/pub/scm/linux/kernel/git/hid/hid:
  HID: nintendo: fix rumble worker null pointer deref
  HID: intel-ish-hid: ipc: Add Meteor Lake PCI device ID
  HID: input: fix uclogic tablets
  HID: Add Apple Touchbar on T2 Macs in hid_have_special_driver list
  HID: add Lenovo Yoga C630 battery quirk
  HID: AMD_SFH: Add a DMI quirk entry for Chromebooks
  HID: thrustmaster: Add sparco wheel and fix array length
  hid: intel-ish-hid: ishtp: Fix ishtp client sending disordered message
  HID: ishtp-hid-clientHID: ishtp-hid-client: Fix comment typo
  HID: asus: ROG NKey: Ignore portion of 0x5a report
  HID: hidraw: fix memory leak in hidraw_release()
  HID: steam: Prevent NULL pointer dereference in steam_{recv,send}_report

22 months agoMerge tag 'v6.0-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Wed, 31 Aug 2022 16:47:06 +0000 (09:47 -0700)]
Merge tag 'v6.0-p2' of git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fix from Herbert Xu:
 "Fix a boot performance regression due to an unnecessary dependency on
  XOR_BLOCKS"

* tag 'v6.0-p2' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: lib - remove unneeded selection of XOR_BLOCKS

22 months agoMerge tag 'lsm-pr-20220829' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm
Linus Torvalds [Wed, 31 Aug 2022 16:23:16 +0000 (09:23 -0700)]
Merge tag 'lsm-pr-20220829' of git://git./linux/kernel/git/pcmoore/lsm

Pull LSM support for IORING_OP_URING_CMD from Paul Moore:
 "Add SELinux and Smack controls to the io_uring IORING_OP_URING_CMD.

  These are necessary as without them the IORING_OP_URING_CMD remains
  outside the purview of the LSMs (Luis' LSM patch, Casey's Smack patch,
  and my SELinux patch). They have been discussed at length with the
  io_uring folks, and Jens has given his thumbs-up on the relevant
  patches (see the commit descriptions).

  There is one patch that is not strictly necessary, but it makes
  testing much easier and is very trivial: the /dev/null
  IORING_OP_URING_CMD patch."

* tag 'lsm-pr-20220829' of git://git.kernel.org/pub/scm/linux/kernel/git/pcmoore/lsm:
  Smack: Provide read control for io_uring_cmd
  /dev/null: add IORING_OP_URING_CMD support
  selinux: implement the security_uring_cmd() LSM hook
  lsm,io_uring: add LSM hooks for the new uring_cmd file op

22 months agocachefiles: make on-demand request distribution fairer
Xin Yin [Thu, 25 Aug 2022 02:09:45 +0000 (10:09 +0800)]
cachefiles: make on-demand request distribution fairer

For now, enqueuing and dequeuing on-demand requests all start from
idx 0, this makes request distribution unfair. In the weighty
concurrent I/O scenario, the request stored in higher idx will starve.

Searching requests cyclically in cachefiles_ondemand_daemon_read,
makes distribution fairer.

Fixes: c8383054506c ("cachefiles: notify the user daemon when looking up cookie")
Reported-by: Yongqing Li <liyongqing@bytedance.com>
Signed-off-by: Xin Yin <yinxin.x@bytedance.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Jeffle Xu <jefflexu@linux.alibaba.com>
Reviewed-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Link: https://lore.kernel.org/r/20220817065200.11543-1-yinxin.x@bytedance.com/
Link: https://lore.kernel.org/r/20220825020945.2293-1-yinxin.x@bytedance.com/
22 months agocachefiles: fix error return code in cachefiles_ondemand_copen()
Sun Ke [Fri, 26 Aug 2022 02:35:15 +0000 (10:35 +0800)]
cachefiles: fix error return code in cachefiles_ondemand_copen()

The cache_size field of copen is specified by the user daemon.
If cache_size < 0, then the OPEN request is expected to fail,
while copen itself shall succeed. However, returning 0 is indeed
unexpected when cache_size is an invalid error code.

Fix this by returning error when cache_size is an invalid error code.

Changes
=======
v4: update the code suggested by Dan
v3: update the commit log suggested by Jingbo.

Fixes: c8383054506c ("cachefiles: notify the user daemon when looking up cookie")
Signed-off-by: Sun Ke <sunke32@huawei.com>
Suggested-by: Jeffle Xu <jefflexu@linux.alibaba.com>
Suggested-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Reviewed-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Jingbo Xu <jefflexu@linux.alibaba.com>
Reviewed-by: Dan Carpenter <dan.carpenter@oracle.com>
Link: https://lore.kernel.org/r/20220818111935.1683062-1-sunke32@huawei.com/
Link: https://lore.kernel.org/r/20220818125038.2247720-1-sunke32@huawei.com/
Link: https://lore.kernel.org/r/20220826023515.3437469-1-sunke32@huawei.com/
22 months agofscache: fix misdocumented parameter
Khalid Masum [Thu, 18 Aug 2022 04:07:38 +0000 (10:07 +0600)]
fscache: fix misdocumented parameter

This patch fixes two warnings generated by make docs. The functions
fscache_use_cookie and fscache_unuse_cookie, both have a parameter
named cookie. But they are documented with the name "object" with
unclear description. Which generates the warning when creating docs.

This commit will replace the currently misdocumented parameter names
with the correct ones while adding proper descriptions.

CC: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Khalid Masum <khalid.masum.92@gmail.com>
Signed-off-by: David Howells <dhowells@redhat.com>
Link: https://lore.kernel.org/r/20220521142446.4746-1-khalid.masum.92@gmail.com/
Link: https://lore.kernel.org/r/20220818040738.12036-1-khalid.masum.92@gmail.com/
Link: https://lore.kernel.org/r/880d7d25753fb326ee17ac08005952112fcf9bdb.1657360984.git.mchehab@kernel.org/
22 months agoethernet: rocker: fix sleep in atomic context bug in neigh_timer_handler
Duoming Zhou [Sat, 27 Aug 2022 15:38:15 +0000 (23:38 +0800)]
ethernet: rocker: fix sleep in atomic context bug in neigh_timer_handler

The function neigh_timer_handler() is a timer handler that runs in an
atomic context. When used by rocker, neigh_timer_handler() calls
"kzalloc(.., GFP_KERNEL)" that may sleep. As a result, the sleep in
atomic context bug will happen. One of the processes is shown below:

ofdpa_fib4_add()
 ...
 neigh_add_timer()

(wait a timer)

neigh_timer_handler()
 neigh_release()
  neigh_destroy()
   rocker_port_neigh_destroy()
    rocker_world_port_neigh_destroy()
     ofdpa_port_neigh_destroy()
      ofdpa_port_ipv4_neigh()
       kzalloc(sizeof(.., GFP_KERNEL) //may sleep

This patch changes the gfp_t parameter of kzalloc() from GFP_KERNEL to
GFP_ATOMIC in order to mitigate the bug.

Fixes: 00fc0c51e35b ("rocker: Change world_ops API and implementation to be switchdev independant")
Signed-off-by: Duoming Zhou <duoming@zju.edu.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonetfilter: nf_tables: clean up hook list when offload flags check fails
Pablo Neira Ayuso [Wed, 31 Aug 2022 11:11:47 +0000 (13:11 +0200)]
netfilter: nf_tables: clean up hook list when offload flags check fails

splice back the hook list so nft_chain_release_hook() has a chance to
release the hooks.

BUG: memory leak
unreferenced object 0xffff88810180b100 (size 96):
  comm "syz-executor133", pid 3619, jiffies 4294945714 (age 12.690s)
  hex dump (first 32 bytes):
    28 64 23 02 81 88 ff ff 28 64 23 02 81 88 ff ff  (d#.....(d#.....
    90 a8 aa 83 ff ff ff ff 00 00 b5 0f 81 88 ff ff  ................
  backtrace:
    [<ffffffff83a8c59b>] kmalloc include/linux/slab.h:600 [inline]
    [<ffffffff83a8c59b>] nft_netdev_hook_alloc+0x3b/0xc0 net/netfilter/nf_tables_api.c:1901
    [<ffffffff83a9239a>] nft_chain_parse_netdev net/netfilter/nf_tables_api.c:1998 [inline]
    [<ffffffff83a9239a>] nft_chain_parse_hook+0x33a/0x530 net/netfilter/nf_tables_api.c:2073
    [<ffffffff83a9b14b>] nf_tables_addchain.constprop.0+0x10b/0x950 net/netfilter/nf_tables_api.c:2218
    [<ffffffff83a9c41b>] nf_tables_newchain+0xa8b/0xc60 net/netfilter/nf_tables_api.c:2593
    [<ffffffff83a3d6a6>] nfnetlink_rcv_batch+0xa46/0xd20 net/netfilter/nfnetlink.c:517
    [<ffffffff83a3db79>] nfnetlink_rcv_skb_batch net/netfilter/nfnetlink.c:638 [inline]
    [<ffffffff83a3db79>] nfnetlink_rcv+0x1f9/0x220 net/netfilter/nfnetlink.c:656
    [<ffffffff83a13b17>] netlink_unicast_kernel net/netlink/af_netlink.c:1319 [inline]
    [<ffffffff83a13b17>] netlink_unicast+0x397/0x4c0 net/netlink/af_netlink.c:1345
    [<ffffffff83a13fd6>] netlink_sendmsg+0x396/0x710 net/netlink/af_netlink.c:1921
    [<ffffffff83865ab6>] sock_sendmsg_nosec net/socket.c:714 [inline]
    [<ffffffff83865ab6>] sock_sendmsg+0x56/0x80 net/socket.c:734
    [<ffffffff8386601c>] ____sys_sendmsg+0x36c/0x390 net/socket.c:2482
    [<ffffffff8386a918>] ___sys_sendmsg+0xa8/0x110 net/socket.c:2536
    [<ffffffff8386aaa8>] __sys_sendmsg+0x88/0x100 net/socket.c:2565
    [<ffffffff845e5955>] do_syscall_x64 arch/x86/entry/common.c:50 [inline]
    [<ffffffff845e5955>] do_syscall_64+0x35/0xb0 arch/x86/entry/common.c:80
    [<ffffffff84800087>] entry_SYSCALL_64_after_hwframe+0x63/0xcd

Fixes: d54725cd11a5 ("netfilter: nf_tables: support for multiple devices per netdev hook")
Reported-by: syzbot+5fcdbfab6d6744c57418@syzkaller.appspotmail.com
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
22 months agonetfilter: br_netfilter: Drop dst references before setting.
Harsh Modi [Wed, 31 Aug 2022 05:36:03 +0000 (22:36 -0700)]
netfilter: br_netfilter: Drop dst references before setting.

The IPv6 path already drops dst in the daddr changed case, but the IPv4
path does not. This change makes the two code paths consistent.

Further, it is possible that there is already a metadata_dst allocated from
ingress that might already be attached to skbuff->dst while following
the bridge path. If it is not released before setting a new
metadata_dst, it will be leaked. This is similar to what is done in
bpf_set_tunnel_key() or ip6_route_input().

It is important to note that the memory being leaked is not the dst
being set in the bridge code, but rather memory allocated from some
other code path that is not being freed correctly before the skb dst is
overwritten.

An example of the leakage fixed by this commit found using kmemleak:

unreferenced object 0xffff888010112b00 (size 256):
  comm "softirq", pid 0, jiffies 4294762496 (age 32.012s)
  hex dump (first 32 bytes):
    00 00 00 00 00 00 00 00 80 16 f1 83 ff ff ff ff  ................
    e1 4e f6 82 ff ff ff ff 00 00 00 00 00 00 00 00  .N..............
  backtrace:
    [<00000000d79567ea>] metadata_dst_alloc+0x1b/0xe0
    [<00000000be113e13>] udp_tun_rx_dst+0x174/0x1f0
    [<00000000a36848f4>] geneve_udp_encap_recv+0x350/0x7b0
    [<00000000d4afb476>] udp_queue_rcv_one_skb+0x380/0x560
    [<00000000ac064aea>] udp_unicast_rcv_skb+0x75/0x90
    [<000000009a8ee8c5>] ip_protocol_deliver_rcu+0xd8/0x230
    [<00000000ef4980bb>] ip_local_deliver_finish+0x7a/0xa0
    [<00000000d7533c8c>] __netif_receive_skb_one_core+0x89/0xa0
    [<00000000a879497d>] process_backlog+0x93/0x190
    [<00000000e41ade9f>] __napi_poll+0x28/0x170
    [<00000000b4c0906b>] net_rx_action+0x14f/0x2a0
    [<00000000b20dd5d4>] __do_softirq+0xf4/0x305
    [<000000003a7d7e15>] __irq_exit_rcu+0xc3/0x140
    [<00000000968d39a2>] sysvec_apic_timer_interrupt+0x9e/0xc0
    [<000000009e920794>] asm_sysvec_apic_timer_interrupt+0x16/0x20
    [<000000008942add0>] native_safe_halt+0x13/0x20

Florian Westphal says: "Original code was likely fine because nothing
ever did set a skb->dst entry earlier than bridge in those days."

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Harsh Modi <harshmodi@google.com>
Acked-by: Florian Westphal <fw@strlen.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
22 months agonetfilter: remove nf_conntrack_helper sysctl and modparam toggles
Pablo Neira Ayuso [Fri, 26 Aug 2022 06:49:16 +0000 (08:49 +0200)]
netfilter: remove nf_conntrack_helper sysctl and modparam toggles

__nf_ct_try_assign_helper() remains in place but it now requires a
template to configure the helper.

A toggle to disable automatic helper assignment was added by:

  a9006892643a ("netfilter: nf_ct_helper: allow to disable automatic helper assignment")

in 2012 to address the issues described in "Secure use of iptables and
connection tracking helpers". Automatic conntrack helper assignment was
disabled by:

  3bb398d925ec ("netfilter: nf_ct_helper: disable automatic helper assignment")

back in 2016.

This patch removes the sysctl and modparam toggles, users now have to
rely on explicit conntrack helper configuration via ruleset.

Update tools/testing/selftests/netfilter/nft_conntrack_helper.sh to
check that auto-assignment does not happen anymore.

Acked-by: Aaron Conole <aconole@redhat.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
22 months agonet: lan966x: improve error handle in lan966x_fdma_rx_get_frame()
Dan Carpenter [Fri, 26 Aug 2022 15:00:30 +0000 (18:00 +0300)]
net: lan966x: improve error handle in lan966x_fdma_rx_get_frame()

Don't just print a warning.  Clean up and return an error as well.

Fixes: c8349639324a ("net: lan966x: Add FDMA functionality")
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Reviewed-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Link: https://lore.kernel.org/r/YwjgDm/SVd5c1tQU@kili
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonfp: fix the access to management firmware hanging
Gao Xiao [Mon, 29 Aug 2022 10:16:51 +0000 (12:16 +0200)]
nfp: fix the access to management firmware hanging

When running `ethtool -p` with the old management firmware,
the management firmware resource is not correctly released,
which causes firmware related malfunction: all the access
to management firmware hangs.

It releases the management firmware resource when set id
mode operation is not supported.

Fixes: ccb9bc1dfa44 ("nfp: add 'ethtool --identify' support")
Signed-off-by: Gao Xiao <gao.xiao@corigine.com>
Reviewed-by: Louis Peens <louis.peens@corigine.com>
Signed-off-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Link: https://lore.kernel.org/r/20220829101651.633840-1-simon.horman@corigine.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agoMerge tag 'ieee802154-for-net-2022-08-29' of git://git.kernel.org/pub/scm/linux/kerne...
Jakub Kicinski [Wed, 31 Aug 2022 06:01:47 +0000 (23:01 -0700)]
Merge tag 'ieee802154-for-net-2022-08-29' of git://git./linux/kernel/git/sschmidt/wpan

Stefan Schmidt says:

====================
ieee802154 for net 2022-08-29

 - repeated word fix from Jilin Yuan.
 - missed return code setting in the cc2520 driver by Li Qiong.
 - fixing a potential race in by defering the workqueue destroy
   in the adf7242 driver by Lin Ma.
 - fixing a long standing problem in the mac802154 rx path to match
   corretcly by Miquel Raynal.

* tag 'ieee802154-for-net-2022-08-29' of git://git.kernel.org/pub/scm/linux/kernel/git/sschmidt/wpan:
  ieee802154: cc2520: add rc code in cc2520_tx()
  net: mac802154: Fix a condition in the receive path
  net/ieee802154: fix repeated words in comments
  ieee802154/adf7242: defer destroy_workqueue call
====================

Link: https://lore.kernel.org/r/20220829100308.2802578-1-stefan@datenfreihafen.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: phy: micrel: Make the GPIO to be non-exclusive
Horatiu Vultur [Tue, 30 Aug 2022 06:40:55 +0000 (08:40 +0200)]
net: phy: micrel: Make the GPIO to be non-exclusive

The same GPIO line can be shared by multiple phys for the coma mode pin.
If that is the case then, all the other phys that share the same line
will failed to be probed because the access to the gpio line is not
non-exclusive.
Fix this by making access to the gpio line to be nonexclusive using flag
GPIOD_FLAGS_BIT_NONEXCLUSIVE. This allows all the other PHYs to be
probed.

Fixes: 738871b09250ee ("net: phy: micrel: add coma mode GPIO")
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Horatiu Vultur <horatiu.vultur@microchip.com>
Link: https://lore.kernel.org/r/20220830064055.2340403-1-horatiu.vultur@microchip.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agonet: virtio_net: fix notification coalescing comments
Alvaro Karsz [Tue, 23 Aug 2022 07:39:47 +0000 (10:39 +0300)]
net: virtio_net: fix notification coalescing comments

Fix wording in comments for the notifications coalescing feature.

Signed-off-by: Alvaro Karsz <alvaro.karsz@solid-run.com>
Acked-by: Michael S. Tsirkin <mst@redhat.com>
Link: https://lore.kernel.org/r/20220823073947.14774-1-alvaro.karsz@solid-run.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
22 months agowifi: iwlegacy: 4965: corrected fix for potential off-by-one overflow in il4965_rs_fi...
Stanislaw Gruszka [Mon, 15 Aug 2022 07:37:37 +0000 (09:37 +0200)]
wifi: iwlegacy: 4965: corrected fix for potential off-by-one overflow in il4965_rs_fill_link_cmd()

This reverts commit a8eb8e6f7159c7c20c0ddac428bde3d110890aa7 as
it can cause invalid link quality command sent to the firmware
and address the off-by-one issue by fixing condition of while loop.

Cc: stable@vger.kernel.org
Fixes: a8eb8e6f7159 ("wifi: iwlegacy: 4965: fix potential off-by-one overflow in il4965_rs_fill_link_cmd()")
Signed-off-by: Stanislaw Gruszka <stf_xl@wp.pl>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220815073737.GA999388@wp.pl
22 months agowifi: wilc1000: fix DMA on stack objects
Ajay.Kathat@microchip.com [Tue, 9 Aug 2022 07:57:56 +0000 (07:57 +0000)]
wifi: wilc1000: fix DMA on stack objects

Sometimes 'wilc_sdio_cmd53' is called with addresses pointing to an
object on the stack. Use dynamically allocated memory for cmd53 instead
of stack address which is not DMA'able.

Fixes: 5625f965d764 ("wilc1000: move wilc driver out of staging")
Reported-by: Michael Walle <mwalle@kernel.org>
Suggested-by: Michael Walle <mwalle@kernel.org>
Signed-off-by: Ajay Singh <ajay.kathat@microchip.com>
Reviewed-by: Michael Walle <mwalle@kernel.org>
Tested-by: Michael Walle <mwalle@kernel.org>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/20220809075749.62752-1-ajay.kathat@microchip.com
22 months agonet/sched: fix netdevice reference leaks in attach_default_qdiscs()
Wang Hai [Fri, 26 Aug 2022 09:00:55 +0000 (17:00 +0800)]
net/sched: fix netdevice reference leaks in attach_default_qdiscs()

In attach_default_qdiscs(), if a dev has multiple queues and queue 0 fails
to attach qdisc because there is no memory in attach_one_default_qdisc().
Then dev->qdisc will be noop_qdisc by default. But the other queues may be
able to successfully attach to default qdisc.

In this case, the fallback to noqueue process will be triggered. If the
original attached qdisc is not released and a new one is directly
attached, this will cause netdevice reference leaks.

The following is the bug log:

veth0: default qdisc (fq_codel) fail, fallback to noqueue
unregister_netdevice: waiting for veth0 to become free. Usage count = 32
leaked reference.
 qdisc_alloc+0x12e/0x210
 qdisc_create_dflt+0x62/0x140
 attach_one_default_qdisc.constprop.41+0x44/0x70
 dev_activate+0x128/0x290
 __dev_open+0x12a/0x190
 __dev_change_flags+0x1a2/0x1f0
 dev_change_flags+0x23/0x60
 do_setlink+0x332/0x1150
 __rtnl_newlink+0x52f/0x8e0
 rtnl_newlink+0x43/0x70
 rtnetlink_rcv_msg+0x140/0x3b0
 netlink_rcv_skb+0x50/0x100
 netlink_unicast+0x1bb/0x290
 netlink_sendmsg+0x37c/0x4e0
 sock_sendmsg+0x5f/0x70
 ____sys_sendmsg+0x208/0x280

Fix this bug by clearing any non-noop qdiscs that may have been assigned
before trying to re-attach.

Fixes: bf6dba76d278 ("net: sched: fallback to qdisc noqueue if default qdisc setup fail")
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Link: https://lore.kernel.org/r/20220826090055.24424-1-wanghai38@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
22 months agonet: sched: tbf: don't call qdisc_put() while holding tree lock
Zhengchao Shao [Fri, 26 Aug 2022 01:39:30 +0000 (09:39 +0800)]
net: sched: tbf: don't call qdisc_put() while holding tree lock

The issue is the same to commit c2999f7fb05b ("net: sched: multiq: don't
call qdisc_put() while holding tree lock"). Qdiscs call qdisc_put() while
holding sch tree spinlock, which results sleeping-while-atomic BUG.

Fixes: c266f64dbfa2 ("net: sched: protect block state with mutex")
Signed-off-by: Zhengchao Shao <shaozhengchao@huawei.com>
Link: https://lore.kernel.org/r/20220826013930.340121-1-shaozhengchao@huawei.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
22 months agotracing: Define the is_signed_type() macro once
Bart Van Assche [Tue, 23 Aug 2022 19:59:25 +0000 (12:59 -0700)]
tracing: Define the is_signed_type() macro once

There are two definitions of the is_signed_type() macro: one in
<linux/overflow.h> and a second definition in <linux/trace_events.h>.

As suggested by Linus, move the definition of the is_signed_type() macro
into the <linux/compiler.h> header file.  Change the definition of the
is_signed_type() macro to make sure that it does not trigger any sparse
warnings with future versions of sparse for bitwise types.

Link: https://lore.kernel.org/all/CAHk-=whjH6p+qzwUdx5SOVVHjS3WvzJQr6mDUwhEyTf6pJWzaQ@mail.gmail.com/
Link: https://lore.kernel.org/all/CAHk-=wjQGnVfb4jehFR0XyZikdQvCZouE96xR_nnf5kqaM5qqQ@mail.gmail.com/
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Cc: Steven Rostedt <rostedt@goodmis.org>
Acked-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
22 months agoMerge tag 'docs-6.0-fixes' of git://git.lwn.net/linux
Linus Torvalds [Mon, 29 Aug 2022 16:49:48 +0000 (09:49 -0700)]
Merge tag 'docs-6.0-fixes' of git://git.lwn.net/linux

Pull documentation fixes from Jonathan Corbet:
 "A handful of fixes for documentation and the docs build system"

* tag 'docs-6.0-fixes' of git://git.lwn.net/linux:
  docs/conf.py: add function attribute '__fix_address' to conf.py
  Docs/admin-guide/mm/damon/usage: fix the example code snip
  docs: Update version number from 5.x to 6.x in README.rst
  docs/ja_JP/SubmittingPatches: Remove reference to submitting-drivers.rst
  docs: kerneldoc-preamble: Test xeCJK.sty before loading

22 months agoMerge branch 'u64_stats-fixups'
David S. Miller [Mon, 29 Aug 2022 12:02:27 +0000 (13:02 +0100)]
Merge branch 'u64_stats-fixups'

Sebastian Andrzej Siewior says:

====================
net: u64_stats fixups for 32bit.

while looking at the u64-stats patch
https://lore.kernel.org/all/20220817162703.728679-10-bigeasy@linutronix.de

I noticed that u64_stats_fetch_begin() is used. That suspicious thing
about it is that network processing, including stats update, is
performed in NAPI and so I would expect to see
u64_stats_fetch_begin_irq() in order to avoid updates from NAPI during
the read. This is only needed on 32bit-UP where the seqcount is not
used. This is address in 2/2. The remaining user take some kind of
precaution and may use u64_stats_fetch_begin().

I updated the previously mentioned patch to get rid of
u64_stats_fetch_begin_irq(). If this is not considered stable patch
worthy then it can be ignored and considred fixed by the other series
which removes the special 32bit cases.

The xrs700x driver reads and writes the counter from preemptible context
so the only missing piece here is at least disable preemption on the
writer side to avoid preemption while the writer is in progress. The
possible reader would spin then until the writer completes its write
critical section which is considered bad. This is addressed in 1/2 by
using u64_stats_update_begin_irqsave() and so disable interrupts during
the write critical section.
The other closet resemblance I found is mdio_bus.c::mdiobus_stats_acct()
where preemtion is disabled unconditionally. This is something I want to
avoid since it also affects 64bit.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: Use u64_stats_fetch_begin_irq() for stats fetch.
Sebastian Andrzej Siewior [Thu, 25 Aug 2022 11:36:45 +0000 (13:36 +0200)]
net: Use u64_stats_fetch_begin_irq() for stats fetch.

On 32bit-UP u64_stats_fetch_begin() disables only preemption. If the
reader is in preemptible context and the writer side
(u64_stats_update_begin*()) runs in an interrupt context (IRQ or
softirq) then the writer can update the stats during the read operation.
This update remains undetected.

Use u64_stats_fetch_begin_irq() to ensure the stats fetch on 32bit-UP
are not interrupted by a writer. 32bit-SMP remains unaffected by this
change.

Cc: "David S. Miller" <davem@davemloft.net>
Cc: Catherine Sullivan <csully@google.com>
Cc: David Awogbemila <awogbemila@google.com>
Cc: Dimitris Michailidis <dmichail@fungible.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: Hans Ulli Kroll <ulli.kroll@googlemail.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Jeroen de Borst <jeroendb@google.com>
Cc: Johannes Berg <johannes@sipsolutions.net>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Paolo Abeni <pabeni@redhat.com>
Cc: Simon Horman <simon.horman@corigine.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-wireless@vger.kernel.org
Cc: netdev@vger.kernel.org
Cc: oss-drivers@corigine.com
Cc: stable@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agonet: dsa: xrs700x: Use irqsave variant for u64 stats update
Sebastian Andrzej Siewior [Thu, 25 Aug 2022 11:36:44 +0000 (13:36 +0200)]
net: dsa: xrs700x: Use irqsave variant for u64 stats update

xrs700x_read_port_counters() updates the stats from a worker using the
u64_stats_update_begin() version. This is okay on 32-UP since on the
reader side preemption is disabled.
On 32bit-SMP the writer can be preempted by the reader at which point
the reader will spin on the seqcount until writer continues and
completes the update.

Assigning the mib_mutex mutex to the underlying seqcount would ensure
proper synchronisation. The API for that on the u64_stats_init() side
isn't available. Since it is the only user, just use disable interrupts
during the update.

Use u64_stats_update_begin_irqsave() on the writer side to ensure an
uninterrupted update.

Fixes: ee00b24f32eb8 ("net: dsa: add Arrow SpeedChips XRS700x driver")
Cc: Andrew Lunn <andrew@lunn.ch>
Cc: Florian Fainelli <f.fainelli@gmail.com>
Cc: George McCollister <george.mccollister@gmail.com>
Cc: Vivien Didelot <vivien.didelot@gmail.com>
Cc: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: George McCollister <george.mccollister@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
22 months agoieee802154: cc2520: add rc code in cc2520_tx()
Li Qiong [Mon, 29 Aug 2022 07:12:59 +0000 (15:12 +0800)]
ieee802154: cc2520: add rc code in cc2520_tx()

The rc code is 0 at the error path "status & CC2520_STATUS_TX_UNDERFLOW".
Assign rc code with '-EINVAL' at this error path to fix it.

Signed-off-by: Li Qiong <liqiong@nfschina.com>
Link: https://lore.kernel.org/r/20220829071259.18330-1-liqiong@nfschina.com
Signed-off-by: Stefan Schmidt <stefan@datenfreihafen.org>
22 months agonet: mac802154: Fix a condition in the receive path
Miquel Raynal [Fri, 26 Aug 2022 14:29:54 +0000 (16:29 +0200)]
net: mac802154: Fix a condition in the receive path

Upon reception, a packet must be categorized, either it's destination is
the host, or it is another host. A packet with no destination addressing
fields may be valid in two situations:
- the packet has no source field: only ACKs are built like that, we
  consider the host as the destination.
- the packet has a valid source field: it is directed to the PAN
  coordinator, as for know we don't have this information we consider we
  are not the PAN coordinator.

There was likely a copy/paste error made during a previous cleanup
because the if clause is now containing exactly the same condition as in
the switch case, which can never be true. In the past the destination
address was used in the switch and the source address was used in the
if, which matches what the spec says.

Cc: stable@vger.kernel.org
Fixes: ae531b9475f6 ("ieee802154: use ieee802154_addr instead of *_sa variants")
Signed-off-by: Miquel Raynal <miquel.raynal@bootlin.com>
Link: https://lore.kernel.org/r/20220826142954.254853-1-miquel.raynal@bootlin.com
Signed-off-by: Stefan Schmidt <stefan@datenfreihafen.org>
22 months agoLinux 6.0-rc3
Linus Torvalds [Sun, 28 Aug 2022 22:05:29 +0000 (15:05 -0700)]
Linux 6.0-rc3

22 months agoMerge tag 'mm-hotfixes-stable-2022-08-28' of git://git.kernel.org/pub/scm/linux/kerne...
Linus Torvalds [Sun, 28 Aug 2022 21:49:59 +0000 (14:49 -0700)]
Merge tag 'mm-hotfixes-stable-2022-08-28' of git://git./linux/kernel/git/akpm/mm

Pull more hotfixes from Andrew Morton:
 "Seventeen hotfixes.  Mostly memory management things.

  Ten patches are cc:stable, addressing pre-6.0 issues"

* tag 'mm-hotfixes-stable-2022-08-28' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
  .mailmap: update Luca Ceresoli's e-mail address
  mm/mprotect: only reference swap pfn page if type match
  squashfs: don't call kmalloc in decompressors
  mm/damon/dbgfs: avoid duplicate context directory creation
  mailmap: update email address for Colin King
  asm-generic: sections: refactor memory_intersects
  bootmem: remove the vmemmap pages from kmemleak in put_page_bootmem
  ocfs2: fix freeing uninitialized resource on ocfs2_dlm_shutdown
  Revert "memcg: cleanup racy sum avoidance code"
  mm/zsmalloc: do not attempt to free IS_ERR handle
  binder_alloc: add missing mmap_lock calls when using the VMA
  mm: re-allow pinning of zero pfns (again)
  vmcoreinfo: add kallsyms_num_syms symbol
  mailmap: update Guilherme G. Piccoli's email addresses
  writeback: avoid use-after-free after removing device
  shmem: update folio if shmem_replace_page() updates the page
  mm/hugetlb: avoid corrupting page->mapping in hugetlb_mcopy_atomic_pte

22 months agoMerge tag 'bitmap-6.0-rc3' of github.com:/norov/linux
Linus Torvalds [Sun, 28 Aug 2022 21:36:27 +0000 (14:36 -0700)]
Merge tag 'bitmap-6.0-rc3' of github.com:/norov/linux

Pull bitmap fixes from Yury Norov:
 "Fix the reported issues, and implements the suggested improvements,
  for the version of the cpumask tests [1] that was merged with commit
  c41e8866c28c ("lib/test: introduce cpumask KUnit test suite").

  These changes include fixes for the tests, and better alignment with
  the KUnit style guidelines"

* tag 'bitmap-6.0-rc3' of github.com:/norov/linux:
  lib/cpumask_kunit: add tests file to MAINTAINERS
  lib/cpumask_kunit: log mask contents
  lib/test_cpumask: follow KUnit style guidelines
  lib/test_cpumask: fix cpu_possible_mask last test
  lib/test_cpumask: drop cpu_possible_mask full test

22 months ago.mailmap: update Luca Ceresoli's e-mail address
Luca Ceresoli [Fri, 26 Aug 2022 13:05:15 +0000 (15:05 +0200)]
.mailmap: update Luca Ceresoli's e-mail address

My Bootlin address is preferred from now on.

Link: https://lkml.kernel.org/r/20220826130515.3011951-1-luca.ceresoli@bootlin.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Atish Patra <atishp@atishpatra.org>
Cc: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
22 months agomm/mprotect: only reference swap pfn page if type match
Peter Xu [Tue, 23 Aug 2022 22:11:38 +0000 (18:11 -0400)]
mm/mprotect: only reference swap pfn page if type match

Yu Zhao reported a bug after the commit "mm/swap: Add swp_offset_pfn() to
fetch PFN from swap entry" added a check in swp_offset_pfn() for swap type [1]:

  kernel BUG at include/linux/swapops.h:117!
  CPU: 46 PID: 5245 Comm: EventManager_De Tainted: G S         O L 6.0.0-dbg-DEV #2
  RIP: 0010:pfn_swap_entry_to_page+0x72/0xf0
  Code: c6 48 8b 36 48 83 fe ff 74 53 48 01 d1 48 83 c1 08 48 8b 09 f6
  c1 01 75 7b 66 90 48 89 c1 48 8b 09 f6 c1 01 74 74 5d c3 eb 9e <0f> 0b
  48 ba ff ff ff ff 03 00 00 00 eb ae a9 ff 0f 00 00 75 13 48
  RSP: 0018:ffffa59e73fabb80 EFLAGS: 00010282
  RAX: 00000000ffffffe8 RBX: 0c00000000000000 RCX: ffffcd5440000000
  RDX: 1ffffffffff7a80a RSI: 0000000000000000 RDI: 0c0000000000042b
  RBP: ffffa59e73fabb80 R08: ffff9965ca6e8bb8 R09: 0000000000000000
  R10: ffffffffa5a2f62d R11: 0000030b372e9fff R12: ffff997b79db5738
  R13: 000000000000042b R14: 0c0000000000042b R15: 1ffffffffff7a80a
  FS:  00007f549d1bb700(0000) GS:ffff99d3cf680000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 0000440d035b3180 CR3: 0000002243176004 CR4: 00000000003706e0
  DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
  DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
  Call Trace:
   <TASK>
   change_pte_range+0x36e/0x880
   change_p4d_range+0x2e8/0x670
   change_protection_range+0x14e/0x2c0
   mprotect_fixup+0x1ee/0x330
   do_mprotect_pkey+0x34c/0x440
   __x64_sys_mprotect+0x1d/0x30

It triggers because pfn_swap_entry_to_page() could be called upon e.g. a
genuine swap entry.

Fix it by only calling it when it's a write migration entry where the page*
is used.

[1] https://lore.kernel.org/lkml/CAOUHufaVC2Za-p8m0aiHw6YkheDcrO-C3wRGixwDS32VTS+k1w@mail.gmail.com/

Link: https://lkml.kernel.org/r/20220823221138.45602-1-peterx@redhat.com
Fixes: 6c287605fd56 ("mm: remember exclusively mapped anonymous pages with PG_anon_exclusive")
Signed-off-by: Peter Xu <peterx@redhat.com>
Reported-by: Yu Zhao <yuzhao@google.com>
Tested-by: Yu Zhao <yuzhao@google.com>
Reviewed-by: David Hildenbrand <david@redhat.com>
Cc: "Huang, Ying" <ying.huang@intel.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
22 months agosquashfs: don't call kmalloc in decompressors
Phillip Lougher [Mon, 22 Aug 2022 21:54:30 +0000 (22:54 +0100)]
squashfs: don't call kmalloc in decompressors

The decompressors may be called while in an atomic section.  So move the
kmalloc() out of this path, and into the "page actor" init function.

This fixes a regression introduced by commit
f268eedddf35 ("squashfs: extend "page actor" to handle missing pages")

Link: https://lkml.kernel.org/r/20220822215430.15933-1-phillip@squashfs.org.uk
Fixes: f268eedddf35 ("squashfs: extend "page actor" to handle missing pages")
Reported-by: Chris Murphy <lists@colorremedies.com>
Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>