platform/kernel/linux-starfive.git
3 years agobnxt_en: Clear DEFRAG flag in firmware message when retry flashing.
Pavan Chebbi [Mon, 11 Jan 2021 09:26:40 +0000 (04:26 -0500)]
bnxt_en: Clear DEFRAG flag in firmware message when retry flashing.

When the FW tells the driver to retry the INSTALL_UPDATE command after
it has cleared the NVM area, the driver is not clearing the previously
used ALLOWED_TO_DEFRAG flag. As a result the FW tries to defrag the NVM
area a second time in a loop and can fail the request.

Fixes: 1432c3f6a6ca ("bnxt_en: Retry installing FW package under NO_SPACE error condition.")
Signed-off-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: Improve stats context resource accounting with RDMA driver loaded.
Michael Chan [Mon, 11 Jan 2021 09:26:39 +0000 (04:26 -0500)]
bnxt_en: Improve stats context resource accounting with RDMA driver loaded.

The function bnxt_get_ulp_stat_ctxs() does not count the stats contexts
used by the RDMA driver correctly when the RDMA driver is freeing the
MSIX vectors.  It assumes that if the RDMA driver is registered, the
additional stats contexts will be needed.  This is not true when the
RDMA driver is about to unregister and frees the MSIX vectors.

This slight error leads to over accouting of the stats contexts needed
after the RDMA driver has unloaded.  This will cause some firmware
warning and error messages in dmesg during subsequent config. changes
or ifdown/ifup.

Fix it by properly accouting for extra stats contexts only if the
RDMA driver is registered and MSIX vectors have been successfully
requested.

Fixes: c027c6b4e91f ("bnxt_en: get rid of num_stat_ctxs variable")
Reviewed-by: Yongping Zhang <yongping.zhang@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agor8153_ecm: Add Lenovo Powered USB-C Hub as a fallback of r8152
Leon Schuermann [Mon, 11 Jan 2021 19:03:15 +0000 (20:03 +0100)]
r8153_ecm: Add Lenovo Powered USB-C Hub as a fallback of r8152

This commit enables the use of the r8153_ecm driver, introduced with
commit c1aedf015ebdd0 ("net/usb/r8153_ecm: support ECM mode for
RTL8153") for the Lenovo Powered USB-C Hub (17ef:721e) based on the
Realtek RTL8153B chip.

This results in the following driver preference:

- if r8152 is available, use the r8152 driver
- if r8152 is not available, use the r8153_ecm driver

This is done to prevent the NIC from constantly sending pause frames
when the host system enters standby (fixed by using the r8152 driver
in "r8152: Add Lenovo Powered USB-C Travel Hub"), while still allowing
the device to work with the r8153_ecm driver as a fallback.

Signed-off-by: Leon Schuermann <leon@is.currently.online>
Tested-by: Leon Schuermann <leon@is.currently.online>
Link: https://lore.kernel.org/r/20210111190312.12589-3-leon@is.currently.online
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agor8152: Add Lenovo Powered USB-C Travel Hub
Leon Schuermann [Mon, 11 Jan 2021 19:03:13 +0000 (20:03 +0100)]
r8152: Add Lenovo Powered USB-C Travel Hub

This USB-C Hub (17ef:721e) based on the Realtek RTL8153B chip used to
use the cdc_ether driver. However, using this driver, with the system
suspended the device constantly sends pause-frames as soon as the
receive buffer fills up. This causes issues with other devices, where
some Ethernet switches stop forwarding packets altogether.

Using the Realtek driver (r8152) fixes this issue. Pause frames are no
longer sent while the host system is suspended.

Signed-off-by: Leon Schuermann <leon@is.currently.online>
Tested-by: Leon Schuermann <leon@is.currently.online>
Link: https://lore.kernel.org/r/20210111190312.12589-2-leon@is.currently.online
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: dsa: clear devlink port type before unregistering slave netdevs
Vladimir Oltean [Tue, 12 Jan 2021 00:48:31 +0000 (02:48 +0200)]
net: dsa: clear devlink port type before unregistering slave netdevs

Florian reported a use-after-free bug in devlink_nl_port_fill found with
KASAN:

(devlink_nl_port_fill)
(devlink_port_notify)
(devlink_port_unregister)
(dsa_switch_teardown.part.3)
(dsa_tree_teardown_switches)
(dsa_unregister_switch)
(bcm_sf2_sw_remove)
(platform_remove)
(device_release_driver_internal)
(device_links_unbind_consumers)
(device_release_driver_internal)
(device_driver_detach)
(unbind_store)

Allocated by task 31:
 alloc_netdev_mqs+0x5c/0x50c
 dsa_slave_create+0x110/0x9c8
 dsa_register_switch+0xdb0/0x13a4
 b53_switch_register+0x47c/0x6dc
 bcm_sf2_sw_probe+0xaa4/0xc98
 platform_probe+0x90/0xf4
 really_probe+0x184/0x728
 driver_probe_device+0xa4/0x278
 __device_attach_driver+0xe8/0x148
 bus_for_each_drv+0x108/0x158

Freed by task 249:
 free_netdev+0x170/0x194
 dsa_slave_destroy+0xac/0xb0
 dsa_port_teardown.part.2+0xa0/0xb4
 dsa_tree_teardown_switches+0x50/0xc4
 dsa_unregister_switch+0x124/0x250
 bcm_sf2_sw_remove+0x98/0x13c
 platform_remove+0x44/0x5c
 device_release_driver_internal+0x150/0x254
 device_links_unbind_consumers+0xf8/0x12c
 device_release_driver_internal+0x84/0x254
 device_driver_detach+0x30/0x34
 unbind_store+0x90/0x134

What happens is that devlink_port_unregister emits a netlink
DEVLINK_CMD_PORT_DEL message which associates the devlink port that is
getting unregistered with the ifindex of its corresponding net_device.
Only trouble is, the net_device has already been unregistered.

It looks like we can stub out the search for a corresponding net_device
if we clear the devlink_port's type. This looks like a bit of a hack,
but also seems to be the reason why the devlink_port_type_clear function
exists in the first place.

Fixes: 3122433eb533 ("net: dsa: Register devlink ports before calling DSA driver setup()")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Florian fainelli <f.fainelli@gmail.com>
Reported-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20210112004831.3778323-1-olteanv@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: dsa: unbind all switches from tree when DSA master unbinds
Vladimir Oltean [Mon, 11 Jan 2021 23:09:43 +0000 (01:09 +0200)]
net: dsa: unbind all switches from tree when DSA master unbinds

Currently the following happens when a DSA master driver unbinds while
there are DSA switches attached to it:

$ echo 0000:00:00.5 > /sys/bus/pci/drivers/mscc_felix/unbind
------------[ cut here ]------------
WARNING: CPU: 0 PID: 392 at net/core/dev.c:9507
Call trace:
 rollback_registered_many+0x5fc/0x688
 unregister_netdevice_queue+0x98/0x120
 dsa_slave_destroy+0x4c/0x88
 dsa_port_teardown.part.16+0x78/0xb0
 dsa_tree_teardown_switches+0x58/0xc0
 dsa_unregister_switch+0x104/0x1b8
 felix_pci_remove+0x24/0x48
 pci_device_remove+0x48/0xf0
 device_release_driver_internal+0x118/0x1e8
 device_driver_detach+0x28/0x38
 unbind_store+0xd0/0x100

Located at the above location is this WARN_ON:

/* Notifier chain MUST detach us all upper devices. */
WARN_ON(netdev_has_any_upper_dev(dev));

Other stacked interfaces, like VLAN, do indeed listen for
NETDEV_UNREGISTER on the real_dev and also unregister themselves at that
time, which is clearly the behavior that rollback_registered_many
expects. But DSA interfaces are not VLAN. They have backing hardware
(platform devices, PCI devices, MDIO, SPI etc) which have a life cycle
of their own and we can't just trigger an unregister from the DSA
framework when we receive a netdev notifier that the master unregisters.

Luckily, there is something we can do, and that is to inform the driver
core that we have a runtime dependency to the DSA master interface's
device, and create a device link where that is the supplier and we are
the consumer. Having this device link will make the DSA switch unbind
before the DSA master unbinds, which is enough to avoid the WARN_ON from
rollback_registered_many.

Note that even before the blamed commit, DSA did nothing intelligent
when the master interface got unregistered either. See the discussion
here:
https://lore.kernel.org/netdev/20200505210253.20311-1-f.fainelli@gmail.com/
But this time, at least the WARN_ON is loud enough that the
upper_dev_link commit can be blamed.

The advantage with this approach vs dev_hold(master) in the attached
link is that the latter is not meant for long term reference counting.
With dev_hold, the only thing that will happen is that when the user
attempts an unbind of the DSA master, netdev_wait_allrefs will keep
waiting and waiting, due to DSA keeping the refcount forever. DSA would
not access freed memory corresponding to the master interface, but the
unbind would still result in a freeze. Whereas with device links,
graceful teardown is ensured. It even works with cascaded DSA trees.

$ echo 0000:00:00.2 > /sys/bus/pci/drivers/fsl_enetc/unbind
[ 1818.797546] device swp0 left promiscuous mode
[ 1819.301112] sja1105 spi2.0: Link is Down
[ 1819.307981] DSA: tree 1 torn down
[ 1819.312408] device eno2 left promiscuous mode
[ 1819.656803] mscc_felix 0000:00:00.5: Link is Down
[ 1819.667194] DSA: tree 0 torn down
[ 1819.711557] fsl_enetc 0000:00:00.2 eno2: Link is Down

This approach allows us to keep the DSA framework absolutely unchanged,
and the driver core will just know to unbind us first when the master
goes away - as opposed to the large (and probably impossible) rework
required if attempting to listen for NETDEV_UNREGISTER.

As per the documentation at Documentation/driver-api/device_link.rst,
specifying the DL_FLAG_AUTOREMOVE_CONSUMER flag causes the device link
to be automatically purged when the consumer fails to probe or later
unbinds. So we don't need to keep the consumer_link variable in struct
dsa_switch.

Fixes: 2f1e8ea726e9 ("net: dsa: link interfaces with the DSA master to get rid of lockdep warnings")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Tested-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20210111230943.3701806-1-olteanv@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: phy: smsc: fix clk error handling
Marco Felsch [Mon, 11 Jan 2021 08:59:32 +0000 (09:59 +0100)]
net: phy: smsc: fix clk error handling

Commit bedd8d78aba3 ("net: phy: smsc: LAN8710/20: add phy refclk in
support") added the phy clk support. The commit already checks if
clk_get_optional() throw an error but instead of returning the error it
ignores it.

Fixes: bedd8d78aba3 ("net: phy: smsc: LAN8710/20: add phy refclk in support")
Suggested-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Marco Felsch <m.felsch@pengutronix.de>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Link: https://lore.kernel.org/r/20210111085932.28680-1-m.felsch@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: dcb: Accept RTM_GETDCB messages carrying set-like DCB commands
Petr Machata [Mon, 11 Jan 2021 17:07:07 +0000 (18:07 +0100)]
net: dcb: Accept RTM_GETDCB messages carrying set-like DCB commands

In commit 826f328e2b7e ("net: dcb: Validate netlink message in DCB
handler"), Linux started rejecting RTM_GETDCB netlink messages if they
contained a set-like DCB_CMD_ command.

The reason was that privileges were only verified for RTM_SETDCB messages,
but the value that determined the action to be taken is the command, not
the message type. And validation of message type against the DCB command
was the obvious missing piece.

Unfortunately it turns out that mlnx_qos, a somewhat widely deployed tool
for configuration of DCB, accesses the DCB set-like APIs through
RTM_GETDCB.

Therefore do not bounce the discrepancy between message type and command.
Instead, in addition to validating privileges based on the actual message
type, validate them also based on the expected message type. This closes
the loophole of allowing DCB configuration on non-admin accounts, while
maintaining backward compatibility.

Fixes: 2f90b8657ec9 ("ixgbe: this patch adds support for DCB to the kernel and ixgbe driver")
Fixes: 826f328e2b7e ("net: dcb: Validate netlink message in DCB handler")
Signed-off-by: Petr Machata <petrm@nvidia.com>
Link: https://lore.kernel.org/r/a3edcfda0825f2aa2591801c5232f2bbf2d8a554.1610384801.git.me@pmachata.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'skb-frag-kmap_atomic-fixes'
Jakub Kicinski [Tue, 12 Jan 2021 02:20:12 +0000 (18:20 -0800)]
Merge branch 'skb-frag-kmap_atomic-fixes'

Willem de Bruijn says:

====================
skb frag: kmap_atomic fixes

skb frags may be backed by highmem and/or compound pages. Various
code calls kmap_atomic to safely access highmem pages. But this
needs additional care for compound pages. Fix a few issues:

patch 1 expect kmap mappings with CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP
patch 2 fixes kmap_atomic + compound page support in skb_seq_read
patch 3 fixes kmap_atomic + compound page support in esp
====================

Link: https://lore.kernel.org/r/20210109221834.3459768-1-willemdebruijn.kernel@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoesp: avoid unneeded kmap_atomic call
Willem de Bruijn [Sat, 9 Jan 2021 22:18:34 +0000 (17:18 -0500)]
esp: avoid unneeded kmap_atomic call

esp(6)_output_head uses skb_page_frag_refill to allocate a buffer for
the esp trailer.

It accesses the page with kmap_atomic to handle highmem. But
skb_page_frag_refill can return compound pages, of which
kmap_atomic only maps the first underlying page.

skb_page_frag_refill does not return highmem, because flag
__GFP_HIGHMEM is not set. ESP uses it in the same manner as TCP.
That also does not call kmap_atomic, but directly uses page_address,
in skb_copy_to_page_nocache. Do the same for ESP.

This issue has become easier to trigger with recent kmap local
debugging feature CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP.

Fixes: cac2661c53f3 ("esp4: Avoid skb_cow_data whenever possible")
Fixes: 03e2a30f6a27 ("esp6: Avoid skb_cow_data whenever possible")
Signed-off-by: Willem de Bruijn <willemb@google.com>
Acked-by: Steffen Klassert <steffen.klassert@secunet.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: compound page support in skb_seq_read
Willem de Bruijn [Sat, 9 Jan 2021 22:18:33 +0000 (17:18 -0500)]
net: compound page support in skb_seq_read

skb_seq_read iterates over an skb, returning pointer and length of
the next data range with each call.

It relies on kmap_atomic to access highmem pages when needed.

An skb frag may be backed by a compound page, but kmap_atomic maps
only a single page. There are not enough kmap slots to always map all
pages concurrently.

Instead, if kmap_atomic is needed, iterate over each page.

As this increases the number of calls, avoid this unless needed.
The necessary condition is captured in skb_frag_must_loop.

I tried to make the change as obvious as possible. It should be easy
to verify that nothing changes if skb_frag_must_loop returns false.

Tested:
  On an x86 platform with
    CONFIG_HIGHMEM=y
    CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP=y
    CONFIG_NETFILTER_XT_MATCH_STRING=y

  Run
    ip link set dev lo mtu 1500
    iptables -A OUTPUT -m string --string 'badstring' -algo bm -j ACCEPT
    dd if=/dev/urandom of=in bs=1M count=20
    nc -l -p 8000 > /dev/null &
    nc -w 1 -q 0 localhost 8000 < in

Signed-off-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: support kmap_local forced debugging in skb_frag_foreach
Willem de Bruijn [Sat, 9 Jan 2021 22:18:32 +0000 (17:18 -0500)]
net: support kmap_local forced debugging in skb_frag_foreach

Skb frags may be backed by highmem and/or compound pages. Highmem
pages need kmap_atomic mappings to access. But kmap_atomic maps a
single page, not the entire compound page.

skb_foreach_page iterates over an skb frag, in one step in the common
case, page by page only if kmap_atomic must be called for each page.
The decision logic is captured in skb_frag_must_loop.

CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP extends kmap from highmem to all
pages, to increase code coverage.

Extend skb_frag_must_loop to this new condition.

Link: https://lore.kernel.org/linux-mm/20210106180132.41dc249d@gandalf.local.home/
Fixes: 0e91a0c6984c ("mm/highmem: Provide CONFIG_DEBUG_KMAP_LOCAL_FORCE_MAP")
Reported-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Willem de Bruijn <willemb@google.com>
Tested-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agorndis_host: set proper input size for OID_GEN_PHYSICAL_MEDIUM request
Andrey Zhizhikin [Fri, 8 Jan 2021 09:58:39 +0000 (09:58 +0000)]
rndis_host: set proper input size for OID_GEN_PHYSICAL_MEDIUM request

MSFT ActiveSync implementation requires that the size of the response for
incoming query is to be provided in the request input length. Failure to
set the input size proper results in failed request transfer, where the
ActiveSync counterpart reports the NDIS_STATUS_INVALID_LENGTH (0xC0010014L)
error.

Set the input size for OID_GEN_PHYSICAL_MEDIUM query to the expected size
of the response in order for the ActiveSync to properly respond to the
request.

Fixes: 039ee17d1baa ("rndis_host: Add RNDIS physical medium checking into generic_rndis_bind()")
Signed-off-by: Andrey Zhizhikin <andrey.zhizhikin@leica-geosystems.com>
Link: https://lore.kernel.org/r/20210108095839.3335-1-andrey.zhizhikin@leica-geosystems.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: mvpp2: Remove Pause and Asym_Pause support
Stefan Chulski [Sun, 10 Jan 2021 19:23:02 +0000 (21:23 +0200)]
net: mvpp2: Remove Pause and Asym_Pause support

Packet Processor hardware not connected to MAC flow control unit and
cannot support TX flow control.
This patch disable flow control support.

Fixes: 3f518509dedc ("ethernet: Add new driver for Marvell Armada 375 network unit")
Signed-off-by: Stefan Chulski <stefanc@marvell.com>
Acked-by: Marcin Wojtas <mw@semihalf.com>
Link: https://lore.kernel.org/r/1610306582-16641-1-git-send-email-stefanc@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agodt-bindings: net: dwmac: fix queue priority documentation
Seb Laveze [Mon, 11 Jan 2021 08:14:07 +0000 (09:14 +0100)]
dt-bindings: net: dwmac: fix queue priority documentation

The priority field is not the queue priority (queue priority is fixed)
but a bitmask of priorities assigned to this queue.

In receive, priorities relate to tagged frames priorities.

In transmit, priorities relate to PFC frames.

Signed-off-by: Seb Laveze <sebastien.laveze@nxp.com>
Link: https://lore.kernel.org/r/20210111081406.1348622-1-sebastien.laveze@oss.nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agodt-bindings: net: renesas,etheravb: RZ/G2H needs tx-internal-delay-ps
Geert Uytterhoeven [Tue, 5 Jan 2021 15:15:16 +0000 (16:15 +0100)]
dt-bindings: net: renesas,etheravb: RZ/G2H needs tx-internal-delay-ps

The merge resolution of the interaction of commits 307eea32b202864c
("dt-bindings: net: renesas,ravb: Add support for r8a774e1 SoC") and
d7adf6331189cbe9 ("dt-bindings: net: renesas,etheravb: Convert to
json-schema") missed that "tx-internal-delay-ps" should be a required
property on RZ/G2H.

Fixes: 8b0308fe319b8002 ("Merge git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20210105151516.1540653-1-geert+renesas@glider.be
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'mlxsw-core-thermal-control-fixes'
Jakub Kicinski [Sun, 10 Jan 2021 00:25:14 +0000 (16:25 -0800)]
Merge branch 'mlxsw-core-thermal-control-fixes'

Ido Schimmel says:

====================
mlxsw: core: Thermal control fixes

This series includes two fixes for thermal control in mlxsw.

Patch #1 validates that the alarm temperature threshold read from a
transceiver is above the warning temperature threshold. If not, the
current thresholds are maintained. It was observed that some transceiver
might be unreliable and sometimes report a too low alarm temperature
threshold which would result in thermal shutdown of the system.

Patch #2 increases the temperature threshold above which thermal
shutdown is triggered for the ASIC thermal zone. It is currently too low
and might result in thermal shutdown under perfectly fine operational
conditions.
====================

Link: https://lore.kernel.org/r/20210108145210.1229820-1-idosch@idosch.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agomlxsw: core: Increase critical threshold for ASIC thermal zone
Vadim Pasternak [Fri, 8 Jan 2021 14:52:10 +0000 (16:52 +0200)]
mlxsw: core: Increase critical threshold for ASIC thermal zone

Increase critical threshold for ASIC thermal zone from 110C to 140C
according to the system hardware requirements. All the supported ASICs
(Spectrum-1, Spectrum-2, Spectrum-3) could be still operational with ASIC
temperature below 140C. With the old critical threshold value system
can perform unjustified shutdown.

All the systems equipped with the above ASICs implement thermal
protection mechanism at firmware level and firmware could decide to
perform system thermal shutdown in case the temperature is below 140C.
So with the new threshold system will not meltdown, while thermal
operating range will be aligned with hardware abilities.

Fixes: 41e760841d26 ("mlxsw: core: Replace thermal temperature trips with defines")
Fixes: a50c1e35650b ("mlxsw: core: Implement thermal zone")
Signed-off-by: Vadim Pasternak <vadimp@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agomlxsw: core: Add validation of transceiver temperature thresholds
Vadim Pasternak [Fri, 8 Jan 2021 14:52:09 +0000 (16:52 +0200)]
mlxsw: core: Add validation of transceiver temperature thresholds

Validate thresholds to avoid a single failure due to some transceiver
unreliability. Ignore the last readouts in case warning temperature is
above alarm temperature, since it can cause unexpected thermal
shutdown. Stay with the previous values and refresh threshold within
the next iteration.

This is a rare scenario, but it was observed at a customer site.

Fixes: 6a79507cfe94 ("mlxsw: core: Extend thermal module with per QSFP module thermal zones")
Signed-off-by: Vadim Pasternak <vadimp@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agotipc: fix NULL deref in tipc_link_xmit()
Hoang Le [Fri, 8 Jan 2021 07:13:37 +0000 (14:13 +0700)]
tipc: fix NULL deref in tipc_link_xmit()

The buffer list can have zero skb as following path:
tipc_named_node_up()->tipc_node_xmit()->tipc_link_xmit(), so
we need to check the list before casting an &sk_buff.

Fault report:
 [] tipc: Bulk publication failure
 [] general protection fault, probably for non-canonical [#1] PREEMPT [...]
 [] KASAN: null-ptr-deref in range [0x00000000000000c8-0x00000000000000cf]
 [] CPU: 0 PID: 0 Comm: swapper/0 Kdump: loaded Not tainted 5.10.0-rc4+ #2
 [] Hardware name: Bochs ..., BIOS Bochs 01/01/2011
 [] RIP: 0010:tipc_link_xmit+0xc1/0x2180
 [] Code: 24 b8 00 00 00 00 4d 39 ec 4c 0f 44 e8 e8 d7 0a 10 f9 48 [...]
 [] RSP: 0018:ffffc90000006ea0 EFLAGS: 00010202
 [] RAX: dffffc0000000000 RBX: ffff8880224da000 RCX: 1ffff11003d3cc0d
 [] RDX: 0000000000000019 RSI: ffffffff886007b9 RDI: 00000000000000c8
 [] RBP: ffffc90000007018 R08: 0000000000000001 R09: fffff52000000ded
 [] R10: 0000000000000003 R11: fffff52000000dec R12: ffffc90000007148
 [] R13: 0000000000000000 R14: 0000000000000000 R15: ffffc90000007018
 [] FS:  0000000000000000(0000) GS:ffff888037400000(0000) knlGS:000[...]
 [] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 [] CR2: 00007fffd2db5000 CR3: 000000002b08f000 CR4: 00000000000006f0

Fixes: af9b028e270fd ("tipc: make media xmit call outside node spinlock context")
Acked-by: Jon Maloy <jmaloy@redhat.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Link: https://lore.kernel.org/r/20210108071337.3598-1-hoang.h.le@dektech.com.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoselftests/tls: fix selftests after adding ChaCha20-Poly1305
Vadim Fedorenko [Fri, 8 Jan 2021 21:37:45 +0000 (00:37 +0300)]
selftests/tls: fix selftests after adding ChaCha20-Poly1305

TLS selftests where broken because of wrong variable types used.
Fix it by changing u16 -> uint16_t

Fixes: 4f336e88a870 ("selftests/tls: add CHACHA20-POLY1305 to tls selftests")
Reported-by: kernel test robot <oliver.sang@intel.com>
Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Link: https://lore.kernel.org/r/1610141865-7142-1-git-send-email-vfedorenko@novek.ru
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ipv6: Validate GSO SKB before finish IPv6 processing
Aya Levin [Thu, 7 Jan 2021 13:50:18 +0000 (15:50 +0200)]
net: ipv6: Validate GSO SKB before finish IPv6 processing

There are cases where GSO segment's length exceeds the egress MTU:
 - Forwarding of a TCP GRO skb, when DF flag is not set.
 - Forwarding of an skb that arrived on a virtualisation interface
   (virtio-net/vhost/tap) with TSO/GSO size set by other network
   stack.
 - Local GSO skb transmitted on an NETIF_F_TSO tunnel stacked over an
   interface with a smaller MTU.
 - Arriving GRO skb (or GSO skb in a virtualised environment) that is
   bridged to a NETIF_F_TSO tunnel stacked over an interface with an
   insufficient MTU.

If so:
 - Consume the SKB and its segments.
 - Issue an ICMP packet with 'Packet Too Big' message containing the
   MTU, allowing the source host to reduce its Path MTU appropriately.

Note: These cases are handled in the same manner in IPv4 output finish.
This patch aligns the behavior of IPv6 and the one of IPv4.

Fixes: 9e50849054a4 ("netfilter: ipv6: move POSTROUTING invocation before fragmentation")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://lore.kernel.org/r/1610027418-30438-1-git-send-email-ayal@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonetxen_nic: fix MSI/MSI-x interrupts
Manish Chopra [Thu, 7 Jan 2021 10:15:20 +0000 (02:15 -0800)]
netxen_nic: fix MSI/MSI-x interrupts

For all PCI functions on the netxen_nic adapter, interrupt
mode (INTx or MSI) configuration is dependent on what has
been configured by the PCI function zero in the shared
interrupt register, as these adapters do not support mixed
mode interrupts among the functions of a given adapter.

Logic for setting MSI/MSI-x interrupt mode in the shared interrupt
register based on PCI function id zero check is not appropriate for
all family of netxen adapters, as for some of the netxen family
adapters PCI function zero is not really meant to be probed/loaded
in the host but rather just act as a management function on the device,
which caused all the other PCI functions on the adapter to always use
legacy interrupt (INTx) mode instead of choosing MSI/MSI-x interrupt mode.

This patch replaces that check with port number so that for all
type of adapters driver attempts for MSI/MSI-x interrupt modes.

Fixes: b37eb210c076 ("netxen_nic: Avoid mixed mode interrupts")
Signed-off-by: Manish Chopra <manishc@marvell.com>
Signed-off-by: Igor Russkikh <irusskikh@marvell.com>
Link: https://lore.kernel.org/r/20210107101520.6735-1-manishc@marvell.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'net-fix-issues-around-register_netdevice-failures'
Jakub Kicinski [Sat, 9 Jan 2021 03:27:44 +0000 (19:27 -0800)]
Merge branch 'net-fix-issues-around-register_netdevice-failures'

Jakub Kicinski says:

====================
net: fix issues around register_netdevice() failures

This series attempts to clean up the life cycle of struct
net_device. Dave has added dev->needs_free_netdev in the
past to fix double frees, we can lean on that mechanism
a little more to fix remaining issues with register_netdevice().

This is the next chapter of the saga which already includes:
commit 0e0eee2465df ("net: correct error path in rtnl_newlink()")
commit e51fb152318e ("rtnetlink: fix a memory leak when ->newlink fails")
commit cf124db566e6 ("net: Fix inconsistent teardown and release of private netdev state.")
commit 93ee31f14f6f ("[NET]: Fix free_netdev on register_netdev failure.")
commit 814152a89ed5 ("net: fix memleak in register_netdevice()")
commit 10cc514f451a ("net: Fix null de-reference of device refcount")

The immediate problem which gets fixed here is that calling
free_netdev() right after unregister_netdevice() is illegal
because we need to release rtnl_lock first, to let the
unregistration finish. Note that unregister_netdevice() is
just a wrapper of unregister_netdevice_queue(), it only
does half of the job.

Where this limitation becomes most problematic is in failure
modes of register_netdevice(). There is a notifier call right
at the end of it, which lets other subsystems veto the entire
thing. At which point we should really go through a full
unregister_netdevice(), but we can't because callers may
go straight to free_netdev() after the failure, and that's
no bueno (see the previous paragraph).

This set makes free_netdev() more lenient, when device
is still being unregistered free_netdev() will simply set
dev->needs_free_netdev and let the unregister process do
the freeing.

With the free_netdev() problem out of the way failures in
register_netdevice() can make use of net_todo, again.
Users are still expected to call free_netdev() right after
failure but that will only set dev->needs_free_netdev.

To prevent the pathological case of:

 dev->needs_free_netdev = true;
 if (register_netdevice(dev)) {
   rtnl_unlock();
   free_netdev(dev);
 }

make register_netdevice()'s failure clear dev->needs_free_netdev.

Problems described above are only present with register_netdevice() /
unregister_netdevice(). We have two parallel APIs for registration
of devices:
 - those called outside rtnl_lock (register_netdev(), and
   unregister_netdev());
 - and those to be used under rtnl_lock - register_netdevice()
   and unregister_netdevice().
The former is trivial and has no problems. The alternative
approach to fix the latter would be to also separate the
freeing functions - i.e. add free_netdevice(). This has been
implemented (incl. converting all relevant calls in the tree)
but it feels a little unnecessary to put the burden of choosing
the right free_netdev{,ice}() call on the programmer when we
can "just do the right thing" by default.
====================

Link: https://lore.kernel.org/r/20210106184007.1821480-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: make sure devices go through netdev_wait_all_refs
Jakub Kicinski [Wed, 6 Jan 2021 18:40:07 +0000 (10:40 -0800)]
net: make sure devices go through netdev_wait_all_refs

If register_netdevice() fails at the very last stage - the
notifier call - some subsystems may have already seen it and
grabbed a reference. struct net_device can't be freed right
away without calling netdev_wait_all_refs().

Now that we have a clean interface in form of dev->needs_free_netdev
and lenient free_netdev() we can undo what commit 93ee31f14f6f ("[NET]:
Fix free_netdev on register_netdev failure.") has done and complete
the unregistration path by bringing the net_set_todo() call back.

After registration fails user is still expected to explicitly
free the net_device, so make sure ->needs_free_netdev is cleared,
otherwise rolling back the registration will cause the old double
free for callers who release rtnl_lock before the free.

This also solves the problem of priv_destructor not being called
on notifier error.

net_set_todo() will be moved back into unregister_netdevice_queue()
in a follow up.

Reported-by: Hulk Robot <hulkci@huawei.com>
Reported-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: make free_netdev() more lenient with unregistering devices
Jakub Kicinski [Wed, 6 Jan 2021 18:40:06 +0000 (10:40 -0800)]
net: make free_netdev() more lenient with unregistering devices

There are two flavors of handling netdev registration:
 - ones called without holding rtnl_lock: register_netdev() and
   unregister_netdev(); and
 - those called with rtnl_lock held: register_netdevice() and
   unregister_netdevice().

While the semantics of the former are pretty clear, the same can't
be said about the latter. The netdev_todo mechanism is utilized to
perform some of the device unregistering tasks and it hooks into
rtnl_unlock() so the locked variants can't actually finish the work.
In general free_netdev() does not mix well with locked calls. Most
drivers operating under rtnl_lock set dev->needs_free_netdev to true
and expect core to make the free_netdev() call some time later.

The part where this becomes most problematic is error paths. There is
no way to unwind the state cleanly after a call to register_netdevice(),
since unreg can't be performed fully without dropping locks.

Make free_netdev() more lenient, and defer the freeing if device
is being unregistered. This allows error paths to simply call
free_netdev() both after register_netdevice() failed, and after
a call to unregister_netdevice() but before dropping rtnl_lock.

Simplify the error paths which are currently doing gymnastics
around free_netdev() handling.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agodocs: net: explain struct net_device lifetime
Jakub Kicinski [Wed, 6 Jan 2021 18:40:05 +0000 (10:40 -0800)]
docs: net: explain struct net_device lifetime

Explain the two basic flows of struct net_device's operation.

Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoppp: fix refcount underflow on channel unbridge
Tom Parkin [Thu, 7 Jan 2021 18:13:15 +0000 (18:13 +0000)]
ppp: fix refcount underflow on channel unbridge

When setting up a channel bridge, ppp_bridge_channels sets the
pch->bridge field before taking the associated reference on the bridge
file instance.

This opens up a refcount underflow bug if ppp_bridge_channels called
via. iotcl runs concurrently with ppp_unbridge_channels executing via.
file release.

The bug is triggered by ppp_bridge_channels taking the error path
through the 'err_unset' label.  In this scenario, pch->bridge is set,
but the reference on the bridged channel will not be taken because
the function errors out.  If ppp_unbridge_channels observes pch->bridge
before it is unset by the error path, it will erroneously drop the
reference on the bridged channel and cause a refcount underflow.

To avoid this, ensure that ppp_bridge_channels holds a reference on
each channel in advance of setting the bridge pointers.

Signed-off-by: Tom Parkin <tparkin@katalix.com>
Fixes: 4cf476ced45d ("ppp: add PPPIOCBRIDGECHAN and PPPIOCUNBRIDGECHAN ioctls")
Acked-by: Guillaume Nault <gnault@redhat.com>
Link: https://lore.kernel.org/r/20210107181315.3128-1-tparkin@katalix.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoudp: Prevent reuseport_select_sock from reading uninitialized socks
Baptiste Lepers [Thu, 7 Jan 2021 05:11:10 +0000 (16:11 +1100)]
udp: Prevent reuseport_select_sock from reading uninitialized socks

reuse->socks[] is modified concurrently by reuseport_add_sock. To
prevent reading values that have not been fully initialized, only read
the array up until the last known safe index instead of incorrectly
re-reading the last index of the array.

Fixes: acdcecc61285f ("udp: correct reuseport selection with connected sockets")
Signed-off-by: Baptiste Lepers <baptiste.lepers@gmail.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Link: https://lore.kernel.org/r/20210107051110.12247-1-baptiste.lepers@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: fix use-after-free when UDP GRO with shared fraglist
Dongseok Yi [Fri, 8 Jan 2021 02:28:38 +0000 (11:28 +0900)]
net: fix use-after-free when UDP GRO with shared fraglist

skbs in fraglist could be shared by a BPF filter loaded at TC. If TC
writes, it will call skb_ensure_writable -> pskb_expand_head to create
a private linear section for the head_skb. And then call
skb_clone_fraglist -> skb_get on each skb in the fraglist.

skb_segment_list overwrites part of the skb linear section of each
fragment itself. Even after skb_clone, the frag_skbs share their
linear section with their clone in PF_PACKET.

Both sk_receive_queue of PF_PACKET and PF_INET (or PF_INET6) can have
a link for the same frag_skbs chain. If a new skb (not frags) is
queued to one of the sk_receive_queue, multiple ptypes can see and
release this. It causes use-after-free.

[ 4443.426215] ------------[ cut here ]------------
[ 4443.426222] refcount_t: underflow; use-after-free.
[ 4443.426291] WARNING: CPU: 7 PID: 28161 at lib/refcount.c:190
refcount_dec_and_test_checked+0xa4/0xc8
[ 4443.426726] pstate: 60400005 (nZCv daif +PAN -UAO)
[ 4443.426732] pc : refcount_dec_and_test_checked+0xa4/0xc8
[ 4443.426737] lr : refcount_dec_and_test_checked+0xa0/0xc8
[ 4443.426808] Call trace:
[ 4443.426813]  refcount_dec_and_test_checked+0xa4/0xc8
[ 4443.426823]  skb_release_data+0x144/0x264
[ 4443.426828]  kfree_skb+0x58/0xc4
[ 4443.426832]  skb_queue_purge+0x64/0x9c
[ 4443.426844]  packet_set_ring+0x5f0/0x820
[ 4443.426849]  packet_setsockopt+0x5a4/0xcd0
[ 4443.426853]  __sys_setsockopt+0x188/0x278
[ 4443.426858]  __arm64_sys_setsockopt+0x28/0x38
[ 4443.426869]  el0_svc_common+0xf0/0x1d0
[ 4443.426873]  el0_svc_handler+0x74/0x98
[ 4443.426880]  el0_svc+0x8/0xc

Fixes: 3a1296a38d0c (net: Support GRO/GSO fraglist chaining.)
Signed-off-by: Dongseok Yi <dseok.yi@samsung.com>
Acked-by: Willem de Bruijn <willemb@google.com>
Acked-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/r/1610072918-174177-1-git-send-email-dseok.yi@samsung.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ipa: modem: add missing SET_NETDEV_DEV() for proper sysfs links
Stephan Gerhold [Wed, 6 Jan 2021 10:07:55 +0000 (11:07 +0100)]
net: ipa: modem: add missing SET_NETDEV_DEV() for proper sysfs links

At the moment it is quite hard to identify the network interface
provided by IPA in userspace components: The network interface is
created as virtual device, without any link to the IPA device.
The interface name ("rmnet_ipa%d") is the only indication that the
network interface belongs to IPA, but this is not very reliable.

Add SET_NETDEV_DEV() to associate the network interface with the
IPA parent device. This allows userspace services like ModemManager
to properly identify that this network interface is provided by IPA
and belongs to the modem.

Cc: Alex Elder <elder@kernel.org>
Fixes: a646d6ec9098 ("soc: qcom: ipa: modem and microcontroller")
Signed-off-by: Stephan Gerhold <stephan@gerhold.net>
Link: https://lore.kernel.org/r/20210106100755.56800-1-stephan@gerhold.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge tag 'net-5.11-rc3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Fri, 8 Jan 2021 20:12:30 +0000 (12:12 -0800)]
Merge tag 'net-5.11-rc3-2' of git://git./linux/kernel/git/netdev/net

Pull more networking fixes from Jakub Kicinski:
 "Slightly lighter pull request to get back into the Thursday cadence.

  Current release - always broken:

   - can: mcp251xfd: fix Tx/Rx ring buffer driver race conditions

   - dsa: hellcreek: fix led_classdev build errors

  Previous releases - regressions:

   - ipv6: fib: flush exceptions when purging route to avoid netdev
     reference leak

   - ip_tunnels: fix pmtu check in nopmtudisc mode

   - ip: always refragment ip defragmented packets to avoid MTU issues
     when forwarding through tunnels, correct "packet too big" message
     is prohibitively tricky to generate

   - s390/qeth: fix locking for discipline setup / removal and during
     recovery to prevent both deadlocks and races

   - mlx5: Use port_num 1 instead of 0 when delete a RoCE address

  Previous releases - always broken:

   - cdc_ncm: correct overhead calculation in delayed_ndp_size to
     prevent out of bound accesses with Huawei 909s-120 LTE module

   - fix stmmac dwmac-sun8i suspend/resume:
           - PHY being left powered off
           - MAC syscon configuration being reset
           - reference to the reset controller being improperly dropped

   - qrtr: fix null-ptr-deref in qrtr_ns_remove

   - can: tcan4x5x: fix bittiming const, use common bittiming from m_can
     driver

   - mlx5e: CT: Use per flow counter when CT flow accounting is enabled

   - mlx5e: Fix SWP offsets when vlan inserted by driver

  Misc:

   - bpf: Fix a task_iter bug caused by a bpf -> net merge conflict
     resolution

  And the usual many fixes to various error paths"

* tag 'net-5.11-rc3-2' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (69 commits)
  net: dsa: lantiq_gswip: Exclude RMII from modes that report 1 GbE
  s390/qeth: fix L2 header access in qeth_l3_osa_features_check()
  s390/qeth: fix locking for discipline setup / removal
  s390/qeth: fix deadlock during recovery
  selftests: fib_nexthops: Fix wrong mausezahn invocation
  nexthop: Bounce NHA_GATEWAY in FDB nexthop groups
  nexthop: Unlink nexthop group entry in error path
  nexthop: Fix off-by-one error in error path
  octeontx2-af: fix memory leak of lmac and lmac->name
  chtls: Fix chtls resources release sequence
  chtls: Added a check to avoid NULL pointer dereference
  chtls: Replace skb_dequeue with skb_peek
  chtls: Avoid unnecessary freeing of oreq pointer
  chtls: Fix panic when route to peer not configured
  chtls: Remove invalid set_tcb call
  chtls: Fix hardware tid leak
  net: ip: always refragment ip defragmented packets
  net: fix pmtu check in nopmtudisc mode
  selftests: netfilter: add selftest for ipip pmtu discovery with enabled connection tracking
  docs: octeontx2: tune rst markup
  ...

3 years agoMerge branch 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6
Linus Torvalds [Fri, 8 Jan 2021 20:05:11 +0000 (12:05 -0800)]
Merge branch 'linus' of git://git./linux/kernel/git/herbert/crypto-2.6

Pull crypto fixes from Herbert Xu:
 "This fixes a functional bug in arm/chacha-neon as well as a potential
  buffer overflow in ecdh"

* 'linus' of git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6:
  crypto: ecdh - avoid buffer overflow in ecdh_set_secret()
  crypto: arm/chacha-neon - add missing counter increment

3 years agopoll: fix performance regression due to out-of-line __put_user()
Linus Torvalds [Thu, 7 Jan 2021 17:43:54 +0000 (09:43 -0800)]
poll: fix performance regression due to out-of-line __put_user()

The kernel test robot reported a -5.8% performance regression on the
"poll2" test of will-it-scale, and bisected it to commit d55564cfc222
("x86: Make __put_user() generate an out-of-line call").

I didn't expect an out-of-line __put_user() to matter, because no normal
core code should use that non-checking legacy version of user access any
more.  But I had overlooked the very odd poll() usage, which does a
__put_user() to update the 'revents' values of the poll array.

Now, Al Viro correctly points out that instead of updating just the
'revents' field, it would be much simpler to just copy the _whole_
pollfd entry, and then we could just use "copy_to_user()" on the whole
array of entries, the same way we use "copy_from_user()" a few lines
earlier to get the original values.

But that is not what we've traditionally done, and I worry that threaded
applications might be concurrently modifying the other fields of the
pollfd array.  So while Al's suggestion is simpler - and perhaps worth
trying in the future - this instead keeps the "just update revents"
model.

To fix the performance regression, use the modern "unsafe_put_user()"
instead of __put_user(), with the proper "user_write_access_begin()"
guarding in place. This improves code generation enormously.

Link: https://lore.kernel.org/lkml/20210107134723.GA28532@xsang-OptiPlex-9020/
Reported-by: kernel test robot <oliver.sang@intel.com>
Tested-by: Oliver Sang <oliver.sang@intel.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: David Laight <David.Laight@aculab.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agoRevert "init/console: Use ttynull as a fallback when there is no console"
Petr Mladek [Fri, 8 Jan 2021 11:48:47 +0000 (12:48 +0100)]
Revert "init/console: Use ttynull as a fallback when there is no console"

This reverts commit 757055ae8dedf5333af17b3b5b4b70ba9bc9da4e.

The commit caused that ttynull was used as the default console
on several systems[1][2][3]. As a result, the console was
blank even when a better alternative existed.

It happened when there was no console configured
on the command line and ttynull_init() was the first initcall
calling register_console().

Or it happened when /dev/ did not exist when console_on_rootfs()
was called. It was not able to open /dev/console even though
a console driver was registered. It tried to add ttynull console
but it obviously did not help. But ttynull became the preferred
console and was used by /dev/console when it was available later.

The commit tried to fix a historical problem that have been there
for ages. The primary motivation was the commit 3cffa06aeef7ece30f6
("printk/console: Allow to disable console output by using console=""
 or console=null"). It provided a clean solution for a workaround
 that was widely used and worked only by chance.

This revert causes that the console="" or console=null command line
options will again work only by chance. These options will cause that
a particular console will be preferred and the default (tty) ones
will not get enabled. There will be no console registered at
all. As a result there won't be stdin, stdout, and stderr for
the init process. But it worked exactly this way even before.

The proper solution has to fulfill many conditions:

  + Register ttynull only when explicitly required or as
    the ultimate fallback.

  + ttynull should get associated with /dev/console but it must
    not become preferred console when used as a fallback.
    Especially, it must still be possible to replace it
    by a better console later.

Such a change requires clean up of the register_console() code.
Otherwise, it would be even harder to follow. Especially, the use
of has_preferred_console and CON_CONSDEV flag is tricky. The clean
up is risky. The ordering of consoles is not well defined. And
any changes tend to break existing user settings.

Do the revert at the least risky solution for now.

[1] https://lore.kernel.org/linux-kselftest/20201221144302.GR4077@smile.fi.intel.com/
[2] https://lore.kernel.org/lkml/d2a3b3c0-e548-7dd1-730f-59bc5c04e191@synopsys.com/
[3] https://patchwork.ozlabs.org/project/linux-um/patch/20210105120128.10854-1-thomas@m3y3r.de/

Reported-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reported-by: Vineet Gupta <vgupta@synopsys.com>
Reported-by: Thomas Meyer <thomas@m3y3r.de>
Signed-off-by: Petr Mladek <pmladek@suse.com>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Acked-by: Sergey Senozhatsky <sergey.senozhatsky@gmail.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
3 years agoMerge tag 'mlx5-fixes-2021-01-07' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Fri, 8 Jan 2021 03:13:29 +0000 (19:13 -0800)]
Merge tag 'mlx5-fixes-2021-01-07' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2021-01-07

* tag 'mlx5-fixes-2021-01-07' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5e: Fix memleak in mlx5e_create_l2_table_groups
  net/mlx5e: Fix two double free cases
  net/mlx5: Release devlink object if adev fails
  net/mlx5e: ethtool, Fix restriction of autoneg with 56G
  net/mlx5e: In skb build skip setting mark in switchdev mode
  net/mlx5: E-Switch, fix changing vf VLANID
  net/mlx5e: Fix SWP offsets when vlan inserted by driver
  net/mlx5e: CT: Use per flow counter when CT flow accounting is enabled
  net/mlx5: Use port_num 1 instead of 0 when delete a RoCE address
  net/mlx5e: Add missing capability check for uplink follow
  net/mlx5: Check if lag is supported before creating one
====================

Link: https://lore.kernel.org/r/20210107202845.470205-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: dsa: lantiq_gswip: Exclude RMII from modes that report 1 GbE
Aleksander Jan Bajkowski [Thu, 7 Jan 2021 19:58:18 +0000 (20:58 +0100)]
net: dsa: lantiq_gswip: Exclude RMII from modes that report 1 GbE

Exclude RMII from modes that report 1 GbE support. Reduced MII supports
up to 100 MbE.

Fixes: 14fceff4771e ("net: dsa: Add Lantiq / Intel DSA driver for vrx200")
Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20210107195818.3878-1-olek2@wp.pl
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 's390-qeth-fixes-2021-01-07'
Jakub Kicinski [Fri, 8 Jan 2021 02:54:08 +0000 (18:54 -0800)]
Merge branch 's390-qeth-fixes-2021-01-07'

Julian Wiedmann says:

====================
s390/qeth: fixes 2021-01-07

This brings two locking fixes for the device control path.
Also one fix for a path where our .ndo_features_check() attempts to
access a non-existent L2 header.
====================

Link: https://lore.kernel.org/r/20210107172442.1737-1-jwi@linux.ibm.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agos390/qeth: fix L2 header access in qeth_l3_osa_features_check()
Julian Wiedmann [Thu, 7 Jan 2021 17:24:42 +0000 (18:24 +0100)]
s390/qeth: fix L2 header access in qeth_l3_osa_features_check()

ip_finish_output_gso() may call .ndo_features_check() even before the
skb has a L2 header. This conflicts with qeth_get_ip_version()'s attempt
to inspect the L2 header via vlan_eth_hdr().

Switch to vlan_get_protocol(), as already used further down in the
common qeth_features_check() path.

Fixes: f13ade199391 ("s390/qeth: run non-offload L3 traffic over common xmit path")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agos390/qeth: fix locking for discipline setup / removal
Julian Wiedmann [Thu, 7 Jan 2021 17:24:41 +0000 (18:24 +0100)]
s390/qeth: fix locking for discipline setup / removal

Due to insufficient locking, qeth_core_set_online() and
qeth_dev_layer2_store() can run in parallel, both attempting to load &
setup the discipline (and stepping on each other toes along the way).
A similar race can also occur between qeth_core_remove_device() and
qeth_dev_layer2_store().

Access to .discipline is meant to be protected by the discipline_mutex,
so add/expand the locking in qeth_core_remove_device() and
qeth_core_set_online().
Adjust the locking in qeth_l*_remove_device() accordingly, as it's now
handled by the callers in a consistent manner.

Based on an initial patch by Ursula Braun.

Fixes: 9dc48ccc68b9 ("qeth: serialize sysfs-triggered device configurations")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agos390/qeth: fix deadlock during recovery
Julian Wiedmann [Thu, 7 Jan 2021 17:24:40 +0000 (18:24 +0100)]
s390/qeth: fix deadlock during recovery

When qeth_dev_layer2_store() - holding the discipline_mutex - waits
inside qeth_l*_remove_device() for a qeth_do_reset() thread to complete,
we can hit a deadlock if qeth_do_reset() concurrently calls
qeth_set_online() and thus tries to aquire the discipline_mutex.

Move the discipline_mutex locking outside of qeth_set_online() and
qeth_set_offline(), and turn the discipline into a parameter so that
callers understand the dependency.

To fix the deadlock, we can now relax the locking:
As already established, qeth_l*_remove_device() waits for
qeth_do_reset() to complete. So qeth_do_reset() itself is under no risk
of having card->discipline ripped out while it's running, and thus
doesn't need to take the discipline_mutex.

Fixes: 9dc48ccc68b9 ("qeth: serialize sysfs-triggered device configurations")
Signed-off-by: Julian Wiedmann <jwi@linux.ibm.com>
Reviewed-by: Alexandra Winter <wintera@linux.ibm.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'nexthop-various-fixes'
Jakub Kicinski [Fri, 8 Jan 2021 02:47:21 +0000 (18:47 -0800)]
Merge branch 'nexthop-various-fixes'

Ido Schimmel says:

====================
nexthop: Various fixes

This series contains various fixes for the nexthop code. The bugs were
uncovered during the development of resilient nexthop groups.

Patches #1-#2 fix the error path of nexthop_create_group(). I was not
able to trigger these bugs with current code, but it is possible with
the upcoming resilient nexthop groups code which adds a user
controllable memory allocation further in the function.

Patch #3 fixes wrong validation of netlink attributes.

Patch #4 fixes wrong invocation of mausezahn in a selftest.
====================

Link: https://lore.kernel.org/r/20210107144824.1135691-1-idosch@idosch.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoselftests: fib_nexthops: Fix wrong mausezahn invocation
Ido Schimmel [Thu, 7 Jan 2021 14:48:24 +0000 (16:48 +0200)]
selftests: fib_nexthops: Fix wrong mausezahn invocation

For IPv6 traffic, mausezahn needs to be invoked with '-6'. Otherwise an
error is returned:

 # ip netns exec me mausezahn veth1 -B 2001:db8:101::2 -A 2001:db8:91::1 -c 0 -t tcp "dp=1-1023, flags=syn"
 Failed to set source IPv4 address. Please check if source is set to a valid IPv4 address.
  Invalid command line parameters!

Fixes: 7c741868ceab ("selftests: Add torture tests to nexthop tests")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonexthop: Bounce NHA_GATEWAY in FDB nexthop groups
Petr Machata [Thu, 7 Jan 2021 14:48:23 +0000 (16:48 +0200)]
nexthop: Bounce NHA_GATEWAY in FDB nexthop groups

The function nh_check_attr_group() is called to validate nexthop groups.
The intention of that code seems to have been to bounce all attributes
above NHA_GROUP_TYPE except for NHA_FDB. However instead it bounces all
these attributes except when NHA_FDB attribute is present--then it accepts
them.

NHA_FDB validation that takes place before, in rtm_to_nh_config(), already
bounces NHA_OIF, NHA_BLACKHOLE, NHA_ENCAP and NHA_ENCAP_TYPE. Yet further
back, NHA_GROUPS and NHA_MASTER are bounced unconditionally.

But that still leaves NHA_GATEWAY as an attribute that would be accepted in
FDB nexthop groups (with no meaning), so long as it keeps the address
family as unspecified:

 # ip nexthop add id 1 fdb via 127.0.0.1
 # ip nexthop add id 10 fdb via default group 1

The nexthop code is still relatively new and likely not used very broadly,
and the FDB bits are newer still. Even though there is a reproducer out
there, it relies on an improbable gateway arguments "via default", "via
all" or "via any". Given all this, I believe it is OK to reformulate the
condition to do the right thing and bounce NHA_GATEWAY.

Fixes: 38428d68719c ("nexthop: support for fdb ecmp nexthops")
Signed-off-by: Petr Machata <petrm@nvidia.com>
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonexthop: Unlink nexthop group entry in error path
Ido Schimmel [Thu, 7 Jan 2021 14:48:22 +0000 (16:48 +0200)]
nexthop: Unlink nexthop group entry in error path

In case of error, remove the nexthop group entry from the list to which
it was previously added.

Fixes: 430a049190de ("nexthop: Add support for nexthop groups")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonexthop: Fix off-by-one error in error path
Ido Schimmel [Thu, 7 Jan 2021 14:48:21 +0000 (16:48 +0200)]
nexthop: Fix off-by-one error in error path

A reference was not taken for the current nexthop entry, so do not try
to put it in the error path.

Fixes: 430a049190de ("nexthop: Add support for nexthop groups")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoocteontx2-af: fix memory leak of lmac and lmac->name
Colin Ian King [Thu, 7 Jan 2021 12:39:16 +0000 (12:39 +0000)]
octeontx2-af: fix memory leak of lmac and lmac->name

Currently the error return paths don't kfree lmac and lmac->name
leading to some memory leaks.  Fix this by adding two error return
paths that kfree these objects

Addresses-Coverity: ("Resource leak")
Fixes: 1463f382f58d ("octeontx2-af: Add support for CGX link management")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20210107123916.189748-1-colin.king@canonical.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'bug-fixes-for-chtls-driver'
Jakub Kicinski [Fri, 8 Jan 2021 01:06:05 +0000 (17:06 -0800)]
Merge branch 'bug-fixes-for-chtls-driver'

Ayush Sawal says:

====================
Bug fixes for chtls driver

patch 1: Fix hardware tid leak.
patch 2: Remove invalid set_tcb call.
patch 3: Fix panic when route to peer not configured.
patch 4: Avoid unnecessary freeing of oreq pointer.
patch 5: Replace skb_dequeue with skb_peek.
patch 6: Added a check to avoid NULL pointer dereference patch.
patch 7: Fix chtls resources release sequence.
====================

Link: https://lore.kernel.org/r/20210106042912.23512-1-ayush.sawal@chelsio.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agochtls: Fix chtls resources release sequence
Ayush Sawal [Wed, 6 Jan 2021 04:29:12 +0000 (09:59 +0530)]
chtls: Fix chtls resources release sequence

CPL_ABORT_RPL is sent after releasing the resources by calling
chtls_release_resources(sk); and chtls_conn_done(sk);
eventually causing kernel panic. Fixing it by calling release
in appropriate order.

Fixes: cc35c88ae4db ("crypto : chtls - CPL handler definition")
Signed-off-by: Vinay Kumar Yadav <vinay.yadav@chelsio.com>
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agochtls: Added a check to avoid NULL pointer dereference
Ayush Sawal [Wed, 6 Jan 2021 04:29:11 +0000 (09:59 +0530)]
chtls: Added a check to avoid NULL pointer dereference

In case of server removal lookup_stid() may return NULL pointer, which
is used as listen_ctx. So added a check before accessing this pointer.

Fixes: cc35c88ae4db ("crypto : chtls - CPL handler definition")
Signed-off-by: Vinay Kumar Yadav <vinay.yadav@chelsio.com>
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agochtls: Replace skb_dequeue with skb_peek
Ayush Sawal [Wed, 6 Jan 2021 04:29:10 +0000 (09:59 +0530)]
chtls: Replace skb_dequeue with skb_peek

The skb is unlinked twice, one in __skb_dequeue in function
chtls_reset_synq() and another in cleanup_syn_rcv_conn().
So in this patch using skb_peek() instead of __skb_dequeue(),
so that unlink will be handled only in cleanup_syn_rcv_conn().

Fixes: cc35c88ae4db ("crypto : chtls - CPL handler definition")
Signed-off-by: Vinay Kumar Yadav <vinay.yadav@chelsio.com>
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agochtls: Avoid unnecessary freeing of oreq pointer
Ayush Sawal [Wed, 6 Jan 2021 04:29:09 +0000 (09:59 +0530)]
chtls: Avoid unnecessary freeing of oreq pointer

In chtls_pass_accept_request(), removing the chtls_reqsk_free()
call to avoid oreq freeing twice. Here oreq is the pointer to
struct request_sock.

Fixes: cc35c88ae4db ("crypto : chtls - CPL handler definition")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agochtls: Fix panic when route to peer not configured
Ayush Sawal [Wed, 6 Jan 2021 04:29:08 +0000 (09:59 +0530)]
chtls: Fix panic when route to peer not configured

If route to peer is not configured, we might get non tls
devices from dst_neigh_lookup() which is invalid, adding a
check to avoid it.

Fixes: cc35c88ae4db ("crypto : chtls - CPL handler definition")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agochtls: Remove invalid set_tcb call
Ayush Sawal [Wed, 6 Jan 2021 04:29:07 +0000 (09:59 +0530)]
chtls: Remove invalid set_tcb call

At the time of SYN_RECV, connection information is not
initialized at FW, updating tcb flag over uninitialized
connection causes adapter crash. We don't need to
update the flag during SYN_RECV state, so avoid this.

Fixes: cc35c88ae4db ("crypto : chtls - CPL handler definition")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agochtls: Fix hardware tid leak
Ayush Sawal [Wed, 6 Jan 2021 04:29:06 +0000 (09:59 +0530)]
chtls: Fix hardware tid leak

send_abort_rpl() is not calculating cpl_abort_req_rss offset and
ends up sending wrong TID with abort_rpl WR causng tid leaks.
Replaced send_abort_rpl() with chtls_send_abort_rpl() as it is
redundant.

Fixes: cc35c88ae4db ("crypto : chtls - CPL handler definition")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge tag 'gcc-plugins-v5.11-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 8 Jan 2021 00:03:19 +0000 (16:03 -0800)]
Merge tag 'gcc-plugins-v5.11-rc3' of git://git./linux/kernel/git/kees/linux

Pull gcc-plugins fix from Kees Cook:
 "Bump c++ standard version for latest GCC versions (Valdis Kletnieks)"

* tag 'gcc-plugins-v5.11-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  gcc-plugins: fix gcc 11 indigestion with plugins...

3 years agoMerge https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 7 Jan 2021 23:10:26 +0000 (15:10 -0800)]
Merge https://git./linux/kernel/git/bpf/bpf

Alexei Starovoitov says:

====================
pull-request: bpf 2021-01-07

We've added 4 non-merge commits during the last 10 day(s) which contain
a total of 4 files changed, 14 insertions(+), 7 deletions(-).

The main changes are:

1) Fix task_iter bug caused by the merge conflict resolution, from Yonghong.

2) Fix resolve_btfids for multiple type hierarchies, from Jiri.

* https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  bpftool: Fix compilation failure for net.o with older glibc
  tools/resolve_btfids: Warn when having multiple IDs for single type
  bpf: Fix a task_iter bug caused by a merge conflict resolution
  selftests/bpf: Fix a compile error for BPF_F_BPRM_SECUREEXEC
====================

Link: https://lore.kernel.org/r/20210107221555.64959-1-alexei.starovoitov@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'net-fix-netfilter-defrag-ip-tunnel-pmtu-blackhole'
Jakub Kicinski [Thu, 7 Jan 2021 22:42:36 +0000 (14:42 -0800)]
Merge branch 'net-fix-netfilter-defrag-ip-tunnel-pmtu-blackhole'

Florian Westphal says:

====================
net: fix netfilter defrag/ip tunnel pmtu blackhole

Christian Perle reported a PMTU blackhole due to unexpected interaction
between the ip defragmentation that comes with connection tracking and
ip tunnels.

Unfortunately setting 'nopmtudisc' on the tunnel breaks the test
scenario even without netfilter.

Christinas setup looks like this:
     +--------+       +---------+       +--------+
     |Router A|-------|Wanrouter|-------|Router B|
     |        |.IPIP..|         |..IPIP.|        |
     +--------+       +---------+       +--------+
          /             mtu 1400           \
         /                                  \
 +--------+                                  +--------+
 |Client A|                                  |Client B|
 +--------+                                  +--------+

MTU is 1500 everywhere, except on Router A to Wanrouter and
Wanrouter to Router B.

Router A and Router B use IPIP tunnel interfaces to tunnel traffic
between Client A and Client B over WAN.

Client A sends a 1400 byte UDP datagram to Client B.
This packet gets encapsulated in the IPIP tunnel.

This works, packet is received on client B.

When conntrack (or anything else that forces ip defragmentation) is
enabled on Router A, the packet gets dropped on Router A after
encapsulation because they exceed the link MTU.

Setting the 'nopmtudisc' flag on the IPIP tunnel makes things worse,
no packets pass even in the no-netfilter scenario.

Patch one is a reproducer script for selftest infra.

Patch two is a fix for 'nopmtudisc' behaviour so ip_tunnel will send
an icmp error to Client A.  This allows 'nopmtudisc' tunnel to forward
the UDP datagrams.

Patch three enables ip refragmentation for all reassembled packets, just
like ipv6.
====================

Link: https://lore.kernel.org/r/20210105231523.622-1-fw@strlen.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ip: always refragment ip defragmented packets
Florian Westphal [Tue, 5 Jan 2021 23:15:23 +0000 (00:15 +0100)]
net: ip: always refragment ip defragmented packets

Conntrack reassembly records the largest fragment size seen in IPCB.
However, when this gets forwarded/transmitted, fragmentation will only
be forced if one of the fragmented packets had the DF bit set.

In that case, a flag in IPCB will force fragmentation even if the
MTU is large enough.

This should work fine, but this breaks with ip tunnels.
Consider client that sends a UDP datagram of size X to another host.

The client fragments the datagram, so two packets, of size y and z, are
sent. DF bit is not set on any of these packets.

Middlebox netfilter reassembles those packets back to single size-X
packet, before routing decision.

packet-size-vs-mtu checks in ip_forward are irrelevant, because DF bit
isn't set.  At output time, ip refragmentation is skipped as well
because x is still smaller than the mtu of the output device.

If ttransmit device is an ip tunnel, the packet size increases to
x+overhead.

Also, tunnel might be configured to force DF bit on outer header.

In this case, packet will be dropped (exceeds MTU) and an ICMP error is
generated back to sender.

But sender already respects the announced MTU, all the packets that
it sent did fit the announced mtu.

Force refragmentation as per original sizes unconditionally so ip tunnel
will encapsulate the fragments instead.

The only other solution I see is to place ip refragmentation in
the ip_tunnel code to handle this case.

Fixes: d6b915e29f4ad ("ip_fragment: don't forward defragmented DF packet")
Reported-by: Christian Perle <christian.perle@secunet.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: fix pmtu check in nopmtudisc mode
Florian Westphal [Tue, 5 Jan 2021 23:15:22 +0000 (00:15 +0100)]
net: fix pmtu check in nopmtudisc mode

For some reason ip_tunnel insist on setting the DF bit anyway when the
inner header has the DF bit set, EVEN if the tunnel was configured with
'nopmtudisc'.

This means that the script added in the previous commit
cannot be made to work by adding the 'nopmtudisc' flag to the
ip tunnel configuration. Doing so breaks connectivity even for the
without-conntrack/netfilter scenario.

When nopmtudisc is set, the tunnel will skip the mtu check, so no
icmp error is sent to client. Then, because inner header has DF set,
the outer header gets added with DF bit set as well.

IP stack then sends an error to itself because the packet exceeds
the device MTU.

Fixes: 23a3647bc4f93 ("ip_tunnels: Use skb-len to PMTU check.")
Cc: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoselftests: netfilter: add selftest for ipip pmtu discovery with enabled connection...
Florian Westphal [Tue, 5 Jan 2021 23:15:21 +0000 (00:15 +0100)]
selftests: netfilter: add selftest for ipip pmtu discovery with enabled connection tracking

Convert Christians bug description into a reproducer.

Cc: Shuah Khan <shuah@kernel.org>
Reported-by: Christian Perle <christian.perle@secunet.com>
Signed-off-by: Florian Westphal <fw@strlen.de>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agodocs: octeontx2: tune rst markup
Lukas Bulwahn [Wed, 6 Jan 2021 16:17:35 +0000 (17:17 +0100)]
docs: octeontx2: tune rst markup

Commit 80b9414832a1 ("docs: octeontx2: Add Documentation for NPA health
reporters") added new documentation with improper formatting for rst, and
caused a few new warnings for make htmldocs in octeontx2.rst:169--202.

Tune markup and formatting for better presentation in the HTML view.

Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Acked-by: George Cherian <george.cherian@marvell.com>
Link: https://lore.kernel.org/r/20210106161735.21751-1-lukas.bulwahn@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet/mlx5e: Fix memleak in mlx5e_create_l2_table_groups
Dinghao Liu [Mon, 21 Dec 2020 11:27:31 +0000 (19:27 +0800)]
net/mlx5e: Fix memleak in mlx5e_create_l2_table_groups

When mlx5_create_flow_group() fails, ft->g should be
freed just like when kvzalloc() fails. The caller of
mlx5e_create_l2_table_groups() does not catch this
issue on failure, which leads to memleak.

Fixes: 33cfaaa8f36f ("net/mlx5e: Split the main flow steering table")
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix two double free cases
Dinghao Liu [Mon, 28 Dec 2020 08:48:40 +0000 (16:48 +0800)]
net/mlx5e: Fix two double free cases

mlx5e_create_ttc_table_groups() frees ft->g on failure of
kvzalloc(), but such failure will be caught by its caller
in mlx5e_create_ttc_table() and ft->g will be freed again
in mlx5e_destroy_flow_table(). The same issue also occurs
in mlx5e_create_ttc_table_groups(). Set ft->g to NULL after
kfree() to avoid double free.

Fixes: 7b3722fa9ef6 ("net/mlx5e: Support RSS for GRE tunneled packets")
Fixes: 33cfaaa8f36f ("net/mlx5e: Split the main flow steering table")
Signed-off-by: Dinghao Liu <dinghao.liu@zju.edu.cn>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Release devlink object if adev fails
Leon Romanovsky [Mon, 4 Jan 2021 08:08:36 +0000 (10:08 +0200)]
net/mlx5: Release devlink object if adev fails

Add missed freeing previously allocated devlink object.

Fixes: a925b5e309c9 ("net/mlx5: Register mlx5 devices to auxiliary virtual bus")
Signed-off-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: ethtool, Fix restriction of autoneg with 56G
Aya Levin [Sun, 27 Dec 2020 14:33:19 +0000 (16:33 +0200)]
net/mlx5e: ethtool, Fix restriction of autoneg with 56G

Prior to this patch, configuring speed to 50G with autoneg off over
devices supporting 50G per lane failed.
Support for 50G per lane introduced a new set of link-modes, on which
driver always performed a speed validation as if only legacy link-modes
were configured. Fix driver speed validation to force setting autoneg
over 56G only if in legacy link-mode.

Fixes: 3d7cadae51f1 ("net/mlx5e: ethtool, Fix analysis of speed setting")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Eran Ben Elisha <eranbe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: In skb build skip setting mark in switchdev mode
Maor Dickman [Mon, 14 Dec 2020 11:53:03 +0000 (13:53 +0200)]
net/mlx5e: In skb build skip setting mark in switchdev mode

sop_drop_qpn field in the cqe is used by two features, in SWITCHDEV mode
to restore the chain id in case of a miss and in LEGACY mode to support
skbedit mark action. In build RX skb, the skb mark field is set regardless
of the configured mode which cause a corruption of the mark field in case
of switchdev mode.

Fix by overriding the mark value back to 0 in the representor tc update
skb flow.

Fixes: 8f1e0b97cc70 ("net/mlx5: E-Switch, Mark miss packets with new chain id mapping")
Signed-off-by: Maor Dickman <maord@nvidia.com>
Reviewed-by: Raed Salem <raeds@nvidia.com>
Reviewed-by: Oz Shlomo <ozsh@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: E-Switch, fix changing vf VLANID
Alaa Hleihel [Mon, 4 Jan 2021 10:54:40 +0000 (12:54 +0200)]
net/mlx5: E-Switch, fix changing vf VLANID

Adding vf VLANID for the first time, or after having cleared previously
defined VLANID works fine, however, attempting to change an existing vf
VLANID clears the rules on the firmware, but does not add new rules for
the new vf VLANID.

Fix this by changing the logic in function esw_acl_egress_lgcy_setup()
so that it will always configure egress rules.

Fixes: ea651a86d468 ("net/mlx5: E-Switch, Refactor eswitch egress acl codes")
Signed-off-by: Alaa Hleihel <alaa@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Fix SWP offsets when vlan inserted by driver
Moshe Shemesh [Fri, 13 Nov 2020 04:06:28 +0000 (06:06 +0200)]
net/mlx5e: Fix SWP offsets when vlan inserted by driver

In case WQE includes inline header the vlan is inserted by driver even
if vlan offload is set. On geneve over vlan interface where software
parser is used the SWP offsets should be updated according to the added
vlan.

Fixes: e3cfc7e6b7bd ("net/mlx5e: TX, Add geneve tunnel stateless offload support")
Signed-off-by: Moshe Shemesh <moshe@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: CT: Use per flow counter when CT flow accounting is enabled
Oz Shlomo [Mon, 7 Dec 2020 08:15:18 +0000 (08:15 +0000)]
net/mlx5e: CT: Use per flow counter when CT flow accounting is enabled

Connection counters may be shared for both directions when the counter
is used for connection aging purposes. However, if TC flow
accounting is enabled then a unique counter is required per direction.

Instantiate a unique counter per direction if the conntrack accounting
extension is enabled. Use a shared counter when the connection accounting
extension is disabled.

Fixes: 1edae2335adf ("net/mlx5e: CT: Use the same counter for both directions")
Signed-off-by: Oz Shlomo <ozsh@nvidia.com>
Reported-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Roi Dayan <roid@nvidia.com>
Reviewed-by: Paul Blakey <paulb@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Use port_num 1 instead of 0 when delete a RoCE address
Mark Zhang [Mon, 14 Dec 2020 01:38:40 +0000 (03:38 +0200)]
net/mlx5: Use port_num 1 instead of 0 when delete a RoCE address

In multi-port mode, FW reports syndrome 0x2ea48 (invalid vhca_port_number)
if the port_num is not 1 or 2.

Fixes: 80f09dfc237f ("net/mlx5: Eswitch, enable RoCE loopback traffic")
Signed-off-by: Mark Zhang <markzhang@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5e: Add missing capability check for uplink follow
Aya Levin [Tue, 24 Nov 2020 20:16:23 +0000 (22:16 +0200)]
net/mlx5e: Add missing capability check for uplink follow

Expose firmware indication that it supports setting eswitch uplink state
to follow (follow the physical link). Condition setting the eswitch
uplink admin-state with this capability bit. Older FW may not support
the uplink state setting.

Fixes: 7d0314b11cdd ("net/mlx5e: Modify uplink state on interface up/down")
Signed-off-by: Aya Levin <ayal@nvidia.com>
Reviewed-by: Moshe Shemesh <moshe@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agonet/mlx5: Check if lag is supported before creating one
Mark Zhang [Mon, 30 Nov 2020 02:38:11 +0000 (04:38 +0200)]
net/mlx5: Check if lag is supported before creating one

This patch fixes a memleak issue by preventing to create a lag and
add PFs if lag is not supported.

comm “python3”, pid 349349, jiffies 4296985507 (age 1446.976s)
hex dump (first 32 bytes):
  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  …………….
  00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  …………….
 backtrace:
  [<000000005b216ae7>] mlx5_lag_add+0x1d5/0×3f0 [mlx5_core]
  [<000000000445aa55>] mlx5e_nic_enable+0x66/0×1b0 [mlx5_core]
  [<00000000c56734c3>] mlx5e_attach_netdev+0x16e/0×200 [mlx5_core]
  [<0000000030439d1f>] mlx5e_attach+0x5c/0×90 [mlx5_core]
  [<0000000018fd8615>] mlx5e_add+0x1a4/0×410 [mlx5_core]
  [<0000000068bc504b>] mlx5_add_device+0x72/0×120 [mlx5_core]
  [<000000009fce51f9>] mlx5_register_device+0x77/0xb0 [mlx5_core]
  [<00000000d0d81ff3>] mlx5_load_one+0xc58/0×1eb0 [mlx5_core]
  [<0000000045077adc>] init_one+0x3ea/0×920 [mlx5_core]
  [<0000000043287674>] pci_device_probe+0xcd/0×150
  [<00000000dafd3279>] really_probe+0x1c9/0×4b0
  [<00000000f06bdd84>] driver_probe_device+0x5d/0×140
  [<00000000e3d508b6>] device_driver_attach+0x4f/0×60
  [<0000000084fba0f0>] bind_store+0xbf/0×120
  [<00000000bf6622b3>] kernfs_fop_write+0x114/0×1b0

Fixes: 9b412cc35f00 ("net/mlx5e: Add LAG warning if bond slave is not lag master")
Signed-off-by: Mark Zhang <markzhang@nvidia.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Reviewed-by: Maor Gottlieb <maorg@nvidia.com>
Signed-off-by: Saeed Mahameed <saeedm@nvidia.com>
3 years agoMerge tag 'spi-fix-v5.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/brooni...
Linus Torvalds [Thu, 7 Jan 2021 20:21:32 +0000 (12:21 -0800)]
Merge tag 'spi-fix-v5.11-rc2' of git://git./linux/kernel/git/broonie/spi

Pull spi fixes from Mark Brown:
 "A couple of core fixes here, both to do with handling of drivers which
  don't report their maximum speed since we factored some of the
  handling for transfer speeds out into the core in the previous
  release.

  There's also some driver specific fixes, including a relatively large
  set for some races around timeouts in spi-geni-qcom"

* tag 'spi-fix-v5.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi:
  spi: fix the divide by 0 error when calculating xfer waiting time
  spi: Fix the clamping of spi->max_speed_hz
  spi: altera: fix return value for altera_spi_txrx()
  spi: stm32: FIFO threshold level - fix align packet size
  spi: spi-geni-qcom: Print an error when we timeout setting the CS
  spi: spi-geni-qcom: Don't try to set CS if an xfer is pending
  spi: spi-geni-qcom: Fail new xfers if xfer/cancel/abort pending
  spi: spi-geni-qcom: Fix geni_spi_isr() NULL dereference in timeout case

3 years agoMerge tag 'regulator-fix-v5.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Thu, 7 Jan 2021 20:18:22 +0000 (12:18 -0800)]
Merge tag 'regulator-fix-v5.11-rc2' of git://git./linux/kernel/git/broonie/regulator

Pull regulator fixes from Mark Brown:
 "A few minor driver specific fixes, mostly DT bindings document bits,
  plus a new device ID"

* tag 'regulator-fix-v5.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regulator:
  regulator: qcom-rpmh: add QCOM_COMMAND_DB dependency
  regulator: qcom-rpmh-regulator: correct hfsmps515 definition
  dt-bindings: regulator: qcom,rpmh-regulator: add pm8009 revision
  regulator: bd718x7: Add enable times
  regulator: pf8x00: Use specific compatible strings for devices

3 years agotools: selftests: add test for changing routes with PTMU exceptions
Sean Tranchetti [Wed, 6 Jan 2021 00:22:26 +0000 (16:22 -0800)]
tools: selftests: add test for changing routes with PTMU exceptions

Adds new 2 new tests to the PTMU script: pmtu_ipv4/6_route_change.

These tests explicitly test for a recently discovered problem in the
IPv6 routing framework where PMTU exceptions were not properly released
when replacing a route via "ip route change ...".

After creating PMTU exceptions, the route from the device A to R1 will be
replaced with a new route, then device A will be deleted. If the PMTU
exceptions were properly cleaned up by the kernel, this device deletion
will succeed. Otherwise, the unregistration of the device will stall, and
messages such as the following will be logged in dmesg:

unregister_netdevice: waiting for veth_A-R1 to become free. Usage count = 4

Signed-off-by: Sean Tranchetti <stranche@codeaurora.org>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/1609892546-11389-2-git-send-email-stranche@quicinc.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ipv6: fib: flush exceptions when purging route
Sean Tranchetti [Wed, 6 Jan 2021 00:22:25 +0000 (16:22 -0800)]
net: ipv6: fib: flush exceptions when purging route

Route removal is handled by two code paths. The main removal path is via
fib6_del_route() which will handle purging any PMTU exceptions from the
cache, removing all per-cpu copies of the DST entry used by the route, and
releasing the fib6_info struct.

The second removal location is during fib6_add_rt2node() during a route
replacement operation. This path also calls fib6_purge_rt() to handle
cleaning up the per-cpu copies of the DST entries and releasing the
fib6_info associated with the older route, but it does not flush any PMTU
exceptions that the older route had. Since the older route is removed from
the tree during the replacement, we lose any way of accessing it again.

As these lingering DSTs and the fib6_info struct are holding references to
the underlying netdevice struct as well, unregistering that device from the
kernel can never complete.

Fixes: 2b760fcf5cfb3 ("ipv6: hook up exception table to store dst cache")
Signed-off-by: Sean Tranchetti <stranche@codeaurora.org>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/1609892546-11389-1-git-send-email-stranche@quicinc.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge tag 'regmap-fix-v5.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Thu, 7 Jan 2021 19:57:56 +0000 (11:57 -0800)]
Merge tag 'regmap-fix-v5.11-rc2' of git://git./linux/kernel/git/broonie/regmap

Pull regmap fixes from Mark Brown:
 "A couple of small fixes for leaks when attaching a device to a
  preexisting regmap"

* tag 'regmap-fix-v5.11-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/broonie/regmap:
  regmap: debugfs: Fix a reversed if statement in regmap_debugfs_init()
  regmap: debugfs: Fix a memory leak when calling regmap_attach_dev

3 years agoMerge tag 'linux-can-fixes-for-5.11-20210107' of git://git.kernel.org/pub/scm/linux...
Jakub Kicinski [Thu, 7 Jan 2021 19:08:08 +0000 (11:08 -0800)]
Merge tag 'linux-can-fixes-for-5.11-20210107' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2021-01-07

The first patch is by me for the m_can driver and removes an erroneous
m_can_clk_stop() from the driver's unregister function.

The second patch targets the tcan4x5x driver, is by me, and fixes the bit
timing constant parameters.

The next two patches are by me, target the mcp251xfd driver, and fix a race
condition in the optimized TEF path (which was added in net-next for v5.11).
The similar code in the RX path is changed to look the same, although it
doesn't suffer from the race condition.

A patch by Lad Prabhakar updates the description and help text for the rcar CAN
driver to reflect all supported SoCs.

In the last patch Sriram Dash transfers the maintainership of the m_can driver
to Pankaj Sharma.

* tag 'linux-can-fixes-for-5.11-20210107' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can:
  MAINTAINERS: Update MCAN MMIO device driver maintainer
  can: rcar: Kconfig: update help description for CAN_RCAR config
  can: mcp251xfd: mcp251xfd_handle_rxif_ring(): first increment RX tail pointer in HW, then in driver
  can: mcp251xfd: mcp251xfd_handle_tefif(): fix TEF vs. TX race condition
  can: tcan4x5x: fix bittiming const, use common bittiming from m_can driver
  can: m_can: m_can_class_unregister(): remove erroneous m_can_clk_stop()
====================

Link: https://lore.kernel.org/r/20210107103451.183477-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMAINTAINERS: Update MCAN MMIO device driver maintainer
Sriram Dash [Mon, 4 Jan 2021 12:31:34 +0000 (18:01 +0530)]
MAINTAINERS: Update MCAN MMIO device driver maintainer

Update Pankaj Sharma as maintainer for mcan mmio device driver as I
will be moving to a different role.

Signed-off-by: Sriram Dash <sriram.dash@samsung.com>
Acked-by: Pankaj Sharma <pankj.sharma@samsung.com>
Link: https://lore.kernel.org/r/20210104123134.16930-1-sriram.dash@samsung.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agocan: rcar: Kconfig: update help description for CAN_RCAR config
Lad Prabhakar [Mon, 4 Jan 2021 09:03:27 +0000 (09:03 +0000)]
can: rcar: Kconfig: update help description for CAN_RCAR config

The rcar_can driver also supports RZ/G SoC's, update the description to reflect
this.

Signed-off-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@bp.renesas.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Link: https://lore.kernel.org/r/20210104090327.6547-1-prabhakar.mahadev-lad.rj@bp.renesas.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agocan: mcp251xfd: mcp251xfd_handle_rxif_ring(): first increment RX tail pointer in...
Marc Kleine-Budde [Tue, 5 Jan 2021 21:41:38 +0000 (22:41 +0100)]
can: mcp251xfd: mcp251xfd_handle_rxif_ring(): first increment RX tail pointer in HW, then in driver

The previous patch fixes a TEF vs. TX race condition, by first updating the TEF
tail pointer in hardware, and then updating the driver internal pointer.

The same pattern exists in the RX-path, too. This should be no problem, as the
driver accesses the RX-FIFO from the interrupt handler only, thus the access is
properly serialized. Fix the order here, too, so that the TEF- and RX-path look
similar.

Fixes: 1f652bb6bae7 ("can: mcp25xxfd: rx-path: reduce number of SPI core requests to set UINC bit")
Link: https://lore.kernel.org/r/20210105214138.3150886-3-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agocan: mcp251xfd: mcp251xfd_handle_tefif(): fix TEF vs. TX race condition
Marc Kleine-Budde [Tue, 5 Jan 2021 21:41:37 +0000 (22:41 +0100)]
can: mcp251xfd: mcp251xfd_handle_tefif(): fix TEF vs. TX race condition

The mcp251xfd driver uses a TX FIFO for sending CAN frames and a TX Event FIFO
(TEF) for completed TX-requests.

The TEF event handling in the mcp251xfd_handle_tefif() function has a race
condition. It first increments the tx-ring's tail counter to signal that
there's room in the TX and TEF FIFO, then it increments the TEF FIFO in
hardware.

A running mcp251xfd_start_xmit() on a different CPU might not stop the txqueue
(as the tx-ring still shows free space). The next mcp251xfd_start_xmit() will
push a message into the chip and the TX complete event might overflow the TEF
FIFO.

This patch changes the order to fix the problem.

Fixes: 68c0c1c7f966 ("can: mcp251xfd: tef-path: reduce number of SPI core requests to set UINC bit")
Link: https://lore.kernel.org/r/20210105214138.3150886-2-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agocan: tcan4x5x: fix bittiming const, use common bittiming from m_can driver
Marc Kleine-Budde [Tue, 15 Dec 2020 10:32:38 +0000 (11:32 +0100)]
can: tcan4x5x: fix bittiming const, use common bittiming from m_can driver

According to the TCAN4550 datasheet "SLLSF91 - DECEMBER 2018" the tcan4x5x has
the same bittiming constants as a m_can revision 3.2.x/3.3.0.

The tcan4x5x chip I'm using identifies itself as m_can revision 3.2.1, so
remove the tcan4x5x specific bittiming values and rely on the values in the
m_can driver, which are selected according to core revision.

Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Cc: Dan Murphy <dmurphy@ti.com>
Reviewed-by: Sean Nyekjaer <sean@geanix.com>
Link: https://lore.kernel.org/r/20201215103238.524029-3-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agocan: m_can: m_can_class_unregister(): remove erroneous m_can_clk_stop()
Marc Kleine-Budde [Tue, 15 Dec 2020 10:32:37 +0000 (11:32 +0100)]
can: m_can: m_can_class_unregister(): remove erroneous m_can_clk_stop()

In m_can_class_register() the clock is started, but stopped on exit. When
calling m_can_class_unregister(), the clock is stopped a second time.

This patch removes the erroneous m_can_clk_stop() in  m_can_class_unregister().

Fixes: f524f829b75a ("can: m_can: Create a m_can platform framework")
Cc: Dan Murphy <dmurphy@ti.com>
Cc: Sriram Dash <sriram.dash@samsung.com>
Reviewed-by: Sean Nyekjaer <sean@geanix.com>
Link: https://lore.kernel.org/r/20201215103238.524029-2-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
3 years agoptp: ptp_ines: prevent build when HAS_IOMEM is not set
Randy Dunlap [Wed, 6 Jan 2021 04:25:31 +0000 (20:25 -0800)]
ptp: ptp_ines: prevent build when HAS_IOMEM is not set

ptp_ines.c uses devm_platform_ioremap_resource(), which is only
built/available when CONFIG_HAS_IOMEM is enabled.
CONFIG_HAS_IOMEM is not enabled for arch/s390/, so builds on S390
have a build error:

s390-linux-ld: drivers/ptp/ptp_ines.o: in function `ines_ptp_ctrl_probe':
ptp_ines.c:(.text+0x17e6): undefined reference to `devm_platform_ioremap_resource'

Prevent builds of ptp_ines.c when HAS_IOMEM is not set.

Fixes: bad1eaa6ac31 ("ptp: Add a driver for InES time stamping IP core.")
Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com>
Link: lore.kernel.org/r/202101031125.ZEFCUiKi-lkp@intel.com
Acked-by: Richard Cochran <richardcochran@gmail.com>
Link: https://lore.kernel.org/r/20210106042531.1351-1-rdunlap@infradead.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: dsa: fix led_classdev build errors
Randy Dunlap [Wed, 6 Jan 2021 02:18:15 +0000 (18:18 -0800)]
net: dsa: fix led_classdev build errors

Fix build errors when LEDS_CLASS=m and NET_DSA_HIRSCHMANN_HELLCREEK=y.
This limits the latter to =m when LEDS_CLASS=m.

microblaze-linux-ld: drivers/net/dsa/hirschmann/hellcreek_ptp.o: in function `hellcreek_ptp_setup':
(.text+0xf80): undefined reference to `led_classdev_register_ext'
microblaze-linux-ld: (.text+0xf94): undefined reference to `led_classdev_register_ext'
microblaze-linux-ld: drivers/net/dsa/hirschmann/hellcreek_ptp.o: in function `hellcreek_ptp_free':
(.text+0x1018): undefined reference to `led_classdev_unregister'
microblaze-linux-ld: (.text+0x1024): undefined reference to `led_classdev_unregister'

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reported-by: kernel test robot <lkp@intel.com>
Link: lore.kernel.org/r/202101060655.iUvMJqS2-lkp@intel.com
Cc: Kurt Kanzenbach <kurt@linutronix.de>
Link: https://lore.kernel.org/r/20210106021815.31796-1-rdunlap@infradead.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: bareudp: add missing error handling for bareudp_link_config()
Jakub Kicinski [Tue, 5 Jan 2021 19:07:25 +0000 (11:07 -0800)]
net: bareudp: add missing error handling for bareudp_link_config()

.dellink does not get called after .newlink fails,
bareudp_newlink() must undo what bareudp_configure()
has done if bareudp_link_config() fails.

v2: call bareudp_dellink(), like bareudp_dev_create() does

Fixes: 571912c69f0e ("net: UDP tunnel encapsulation module for tunnelling different protocols like MPLS, IP, NSH etc.")
Link: https://lore.kernel.org/r/20210105190725.1736246-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: mvneta: fix error message when MTU too large for XDP
Marek Behún [Tue, 5 Jan 2021 17:23:33 +0000 (18:23 +0100)]
net: mvneta: fix error message when MTU too large for XDP

The error message says that "Jumbo frames are not supported on XDP", but
the code checks for mtu > MVNETA_MAX_RX_BUF_SIZE, not mtu > 1500.

Fix this error message.

Signed-off-by: Marek Behún <kabel@kernel.org>
Fixes: 0db51da7a8e9 ("net: mvneta: add basic XDP support")
Cc: Lorenzo Bianconi <lorenzo@kernel.org>
Cc: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Link: https://lore.kernel.org/r/20210105172333.21613-1-kabel@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobpftool: Fix compilation failure for net.o with older glibc
Alan Maguire [Wed, 6 Jan 2021 15:59:06 +0000 (15:59 +0000)]
bpftool: Fix compilation failure for net.o with older glibc

For older glibc ~2.17, #include'ing both linux/if.h and net/if.h
fails due to complaints about redefinition of interface flags:

  CC       net.o
In file included from net.c:13:0:
/usr/include/linux/if.h:71:2: error: redeclaration of enumerator ‘IFF_UP’
  IFF_UP    = 1<<0,  /* sysfs */
  ^
/usr/include/net/if.h:44:5: note: previous definition of ‘IFF_UP’ was here
     IFF_UP = 0x1,  /* Interface is up.  */

The issue was fixed in kernel headers in [1], but since compilation
of net.c picks up system headers the problem can recur.

Dropping #include <linux/if.h> resolves the issue and it is
not needed for compilation anyhow.

[1] https://lore.kernel.org/netdev/1461512707-23058-1-git-send-email-mikko.rapeli__34748.27880641$1462831734$gmane$org@iki.fi/

Fixes: f6f3bac08ff9 ("tools/bpf: bpftool: add net support")
Signed-off-by: Alan Maguire <alan.maguire@oracle.com>
Signed-off-by: Andrii Nakryiko <andrii@kernel.org>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Link: https://lore.kernel.org/bpf/1609948746-15369-1-git-send-email-alan.maguire@oracle.com
3 years agogcc-plugins: fix gcc 11 indigestion with plugins...
Valdis Klētnieks [Sat, 26 Dec 2020 18:21:58 +0000 (13:21 -0500)]
gcc-plugins: fix gcc 11 indigestion with plugins...

Fedora Rawhide has started including gcc 11,and the g++ compiler
throws a wobbly when it hits scripts/gcc-plugins:

  HOSTCXX scripts/gcc-plugins/latent_entropy_plugin.so
In file included from /usr/include/c++/11/type_traits:35,
                 from /usr/lib/gcc/x86_64-redhat-linux/11/plugin/include/system.h:244,
                 from /usr/lib/gcc/x86_64-redhat-linux/11/plugin/include/gcc-plugin.h:28,
                 from scripts/gcc-plugins/gcc-common.h:7,
                 from scripts/gcc-plugins/latent_entropy_plugin.c:78:
/usr/include/c++/11/bits/c++0x_warning.h:32:2: error: #error This file requires compiler and library support for the ISO
 C++ 2011 standard. This support must be enabled with the -std=c++11 or -std=gnu++11 compiler options.
   32 | #error This file requires compiler and library support \

In fact, it works just fine with c++11, which has been in gcc since 4.8,
and we now require 4.9 as a minimum.

Signed-off-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Acked-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/82487.1609006918@turing-police
3 years agoMerge tag 'for-5.11-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Wed, 6 Jan 2021 19:19:08 +0000 (11:19 -0800)]
Merge tag 'for-5.11-rc2-tag' of git://git./linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "A few more fixes that arrived before the end of the year:

   - a bunch of fixes related to transaction handle lifetime wrt various
     operations (umount, remount, qgroup scan, orphan cleanup)

   - async discard scheduling fixes

   - fix item size calculation when item keys collide for extend refs
     (hardlinks)

   - fix qgroup flushing from running transaction

   - fix send, wrong file path when there is an inode with a pending
     rmdir

   - fix deadlock when cloning inline extent and low on free metadata
     space"

* tag 'for-5.11-rc2-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: run delayed iputs when remounting RO to avoid leaking them
  btrfs: add assertion for empty list of transactions at late stage of umount
  btrfs: fix race between RO remount and the cleaner task
  btrfs: fix transaction leak and crash after cleaning up orphans on RO mount
  btrfs: fix transaction leak and crash after RO remount caused by qgroup rescan
  btrfs: merge critical sections of discard lock in workfn
  btrfs: fix racy access to discard_ctl data
  btrfs: fix async discard stall
  btrfs: tests: initialize test inodes location
  btrfs: send: fix wrong file path when there is an inode with a pending rmdir
  btrfs: qgroup: don't try to wait flushing if we're already holding a transaction
  btrfs: correctly calculate item size used when item key collision happens
  btrfs: fix deadlock when cloning inline extent and low on free metadata space

3 years agotools/resolve_btfids: Warn when having multiple IDs for single type
Jiri Olsa [Tue, 5 Jan 2021 23:42:19 +0000 (00:42 +0100)]
tools/resolve_btfids: Warn when having multiple IDs for single type

The kernel image can contain multiple types (structs/unions)
with the same name. This causes distinct type hierarchies in
BTF data and makes resolve_btfids fail with error like:

  BTFIDS  vmlinux
FAILED unresolved symbol udp6_sock

as reported by Qais Yousef [1].

This change adds warning when multiple types of the same name
are detected:

  BTFIDS  vmlinux
WARN: multiple IDs found for 'file': 526, 113351 - using 526
WARN: multiple IDs found for 'sk_buff': 2744, 113958 - using 2744

We keep the lower ID for the given type instance and let the
build continue.

Also changing the 'nr' variable name to 'nr_types' to avoid confusion.

[1] https://lore.kernel.org/lkml/20201229151352.6hzmjvu3qh6p2qgg@e107158-lin/

Signed-off-by: Jiri Olsa <jolsa@kernel.org>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Acked-by: Andrii Nakryiko <andrii@kernel.org>
Link: https://lore.kernel.org/bpf/20210105234219.970039-1-jolsa@kernel.org
3 years agoselftests: fix the return value for UDP GRO test
Po-Hsu Lin [Tue, 5 Jan 2021 10:17:40 +0000 (18:17 +0800)]
selftests: fix the return value for UDP GRO test

The udpgro.sh will always return 0 (unless the bpf selftest was not
build first) even if there are some failed sub test-cases.

Therefore the kselftest framework will report this case is OK.

Check and return the exit status of each test to make it easier to
spot real failures.

Signed-off-by: Po-Hsu Lin <po-hsu.lin@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: ethernet: fs_enet: Add missing MODULE_LICENSE
Michael Ellerman [Tue, 5 Jan 2021 09:15:15 +0000 (20:15 +1100)]
net: ethernet: fs_enet: Add missing MODULE_LICENSE

Since commit 1d6cd3929360 ("modpost: turn missing MODULE_LICENSE()
into error") the ppc32_allmodconfig build fails with:

  ERROR: modpost: missing MODULE_LICENSE() in drivers/net/ethernet/freescale/fs_enet/mii-fec.o
  ERROR: modpost: missing MODULE_LICENSE() in drivers/net/ethernet/freescale/fs_enet/mii-bitbang.o

Add the missing MODULE_LICENSEs to fix the build. Both files include a
copyright header indicating they are GPL v2.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: qrtr: fix null-ptr-deref in qrtr_ns_remove
Qinglang Miao [Tue, 5 Jan 2021 05:57:54 +0000 (13:57 +0800)]
net: qrtr: fix null-ptr-deref in qrtr_ns_remove

A null-ptr-deref bug is reported by Hulk Robot like this:
--------------
KASAN: null-ptr-deref in range [0x0000000000000128-0x000000000000012f]
Call Trace:
qrtr_ns_remove+0x22/0x40 [ns]
qrtr_proto_fini+0xa/0x31 [qrtr]
__x64_sys_delete_module+0x337/0x4e0
do_syscall_64+0x34/0x80
entry_SYSCALL_64_after_hwframe+0x44/0xa9
RIP: 0033:0x468ded
--------------

When qrtr_ns_init fails in qrtr_proto_init, qrtr_ns_remove which would
be called later on would raise a null-ptr-deref because qrtr_ns.workqueue
has been destroyed.

Fix it by making qrtr_ns_init have a return value and adding a check in
qrtr_proto_init.

Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Qinglang Miao <miaoqinglang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: cdc_ncm: correct overhead in delayed_ndp_size
Jouni K. Seppänen [Tue, 5 Jan 2021 04:52:49 +0000 (06:52 +0200)]
net: cdc_ncm: correct overhead in delayed_ndp_size

Aligning to tx_ndp_modulus is not sufficient because the next align
call can be cdc_ncm_align_tail, which can add up to ctx->tx_modulus +
ctx->tx_remainder - 1 bytes. This used to lead to occasional crashes
on a Huawei 909s-120 LTE module as follows:

- the condition marked /* if there is a remaining skb [...] */ is true
  so the swaps happen
- skb_out is set from ctx->tx_curr_skb
- skb_out->len is exactly 0x3f52
- ctx->tx_curr_size is 0x4000 and delayed_ndp_size is 0xac
  (note that the sum of skb_out->len and delayed_ndp_size is 0x3ffe)
- the for loop over n is executed once
- the cdc_ncm_align_tail call marked /* align beginning of next frame */
  increases skb_out->len to 0x3f56 (the sum is now 0x4002)
- the condition marked /* check if we had enough room left [...] */ is
  false so we break out of the loop
- the condition marked /* If requested, put NDP at end of frame. */ is
  true so the NDP is written into skb_out
- now skb_out->len is 0x4002, so padding_count is minus two interpreted
  as an unsigned number, which is used as the length argument to memset,
  leading to a crash with various symptoms but usually including

> Call Trace:
>  <IRQ>
>  cdc_ncm_fill_tx_frame+0x83a/0x970 [cdc_ncm]
>  cdc_mbim_tx_fixup+0x1d9/0x240 [cdc_mbim]
>  usbnet_start_xmit+0x5d/0x720 [usbnet]

The cdc_ncm_align_tail call first aligns on a ctx->tx_modulus
boundary (adding at most ctx->tx_modulus-1 bytes), then adds
ctx->tx_remainder bytes. Alternatively, the next alignment call can
occur in cdc_ncm_ndp16 or cdc_ncm_ndp32, in which case at most
ctx->tx_ndp_modulus-1 bytes are added.

A similar problem has occurred before, and the code is nontrivial to
reason about, so add a guard before the crashing call. By that time it
is too late to prevent any memory corruption (we'll have written past
the end of the buffer already) but we can at least try to get a warning
written into an on-disk log by avoiding the hard crash caused by padding
past the buffer with a huge number of zeros.

Signed-off-by: Jouni K. Seppänen <jks@iki.fi>
Fixes: 4a0e3e989d66 ("cdc_ncm: Add support for moving NDP to end of NCM frame")
Link: https://bugzilla.kernel.org/show_bug.cgi?id=209407
Reported-by: kernel test robot <lkp@intel.com>
Reviewed-by: Bjørn Mork <bjorn@mork.no>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agoMerge branch 'hns3-fixes'
David S. Miller [Wed, 6 Jan 2021 00:46:41 +0000 (16:46 -0800)]
Merge branch 'hns3-fixes'

Huazhong Tan says:

====================
net: hns3: fixes for -net

There are some bugfixes for the HNS3 ethernet driver.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hns3: fix incorrect handling of sctp6 rss tuple
Jian Shen [Tue, 5 Jan 2021 03:37:28 +0000 (11:37 +0800)]
net: hns3: fix incorrect handling of sctp6 rss tuple

For DEVICE_VERSION_V2, the hardware only supports src-ip,
dst-ip and verification-tag for rss tuple set of sctp6
packet. For DEVICE_VERSION_V3, the hardware supports
src-port and dst-port as well.

Currently, when user queries the sctp6 rss tuples info,
some unsupported information will be showed on V2. So add
a check for hardware version when initializing and queries
sctp6 rss tuple to fix this issue.

Fixes: 46a3df9f9718 ("net: hns3: Add HNS3 Acceleration Engine & Compatibility Layer Support")
Signed-off-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
3 years agonet: hns3: fix the number of queues actually used by ARQ
Yufeng Mo [Tue, 5 Jan 2021 03:37:27 +0000 (11:37 +0800)]
net: hns3: fix the number of queues actually used by ARQ

HCLGE_MBX_MAX_ARQ_MSG_NUM is used to apply memory for the number
of queues used by ARQ(Asynchronous Receive Queue), so the head
and tail pointers should also use this macro.

Fixes: 07a0556a3a73 ("net: hns3: Changes to support ARQ(Asynchronous Receive Queue)")
Signed-off-by: Yufeng Mo <moyufeng@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>