platform/kernel/linux-rpi.git
20 months agotest/vsock: new skbuff appending test
Arseniy Krasnov [Tue, 28 Mar 2023 11:33:07 +0000 (14:33 +0300)]
test/vsock: new skbuff appending test

This adds test which checks case when data of newly received skbuff is
appended to the last skbuff in the socket's queue. It looks like simple
test with 'send()' and 'recv()', but internally it triggers logic which
appends one received skbuff to another. Test checks that this feature
works correctly.

This test is actual only for virtio transport.

Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agovirtio/vsock: WARN_ONCE() for invalid state of socket
Arseniy Krasnov [Tue, 28 Mar 2023 11:32:12 +0000 (14:32 +0300)]
virtio/vsock: WARN_ONCE() for invalid state of socket

This adds WARN_ONCE() and return from stream dequeue callback when
socket's queue is empty, but 'rx_bytes' still non-zero. This allows
the detection of potential bugs due to packet merging (see previous
patch).

Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agovirtio/vsock: fix header length on skb merging
Arseniy Krasnov [Tue, 28 Mar 2023 11:31:28 +0000 (14:31 +0300)]
virtio/vsock: fix header length on skb merging

This fixes appending newly arrived skbuff to the last skbuff of the
socket's queue. Problem fires when we are trying to append data to skbuff
which was already processed in dequeue callback at least once. Dequeue
callback calls function 'skb_pull()' which changes 'skb->len'. In current
implementation 'skb->len' is used to update length in header of the last
skbuff after new data was copied to it. This is bug, because value in
header is used to calculate 'rx_bytes'/'fwd_cnt' and thus must be not
be changed during skbuff's lifetime.

Bug starts to fire since:

commit 077706165717
("virtio/vsock: don't use skbuff state to account credit")

It presents before, but didn't triggered due to a little bit buggy
implementation of credit calculation logic. So use Fixes tag for it.

Fixes: 077706165717 ("virtio/vsock: don't use skbuff state to account credit")
Signed-off-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agoMerge branch 'bnxt_en-3-bug-fixes'
Jakub Kicinski [Thu, 30 Mar 2023 04:48:18 +0000 (21:48 -0700)]
Merge branch 'bnxt_en-3-bug-fixes'

Michael Chan says:

====================
bnxt_en: 3 Bug fixes

This series contains 3 small bug fixes covering ethtool self test, PCI
ID string typos, and some missing 200G link speed ethtool reporting logic.
====================

Link: https://lore.kernel.org/r/20230329013021.5205-1-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agobnxt_en: Add missing 200G link speed reporting
Michael Chan [Wed, 29 Mar 2023 01:30:21 +0000 (18:30 -0700)]
bnxt_en: Add missing 200G link speed reporting

bnxt_fw_to_ethtool_speed() is missing the case statement for 200G
link speed reported by firmware.  As a result, ethtool will report
unknown speed when the firmware reports 200G link speed.

Fixes: 532262ba3b84 ("bnxt_en: ethtool: support PAM4 link speeds up to 200G")
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agobnxt_en: Fix typo in PCI id to device description string mapping
Kalesh AP [Wed, 29 Mar 2023 01:30:20 +0000 (18:30 -0700)]
bnxt_en: Fix typo in PCI id to device description string mapping

Fix 57502 and 57508 NPAR description string entries.  The typos
caused these devices to not match up with lspci output.

Fixes: 49c98421e6ab ("bnxt_en: Add PCI IDs for 57500 series NPAR devices.")
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agobnxt_en: Fix reporting of test result in ethtool selftest
Kalesh AP [Wed, 29 Mar 2023 01:30:19 +0000 (18:30 -0700)]
bnxt_en: Fix reporting of test result in ethtool selftest

When the selftest command fails, driver is not reporting the failure
by updating the "test->flags" when bnxt_close_nic() fails.

Fixes: eb51365846bc ("bnxt_en: Add basic ethtool -t selftest support.")
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Somnath Kotur <somnath.kotur@broadcom.com>
Signed-off-by: Kalesh AP <kalesh-anakkur.purayil@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agoi40e: fix registers dump after run ethtool adapter self test
Radoslaw Tyl [Tue, 28 Mar 2023 17:26:59 +0000 (10:26 -0700)]
i40e: fix registers dump after run ethtool adapter self test

Fix invalid registers dump from ethtool -d ethX after adapter self test
by ethtool -t ethY. It causes invalid data display.

The problem was caused by overwriting i40e_reg_list[].elements
which is common for ethtool self test and dump.

Fixes: 22dd9ae8afcc ("i40e: Rework register diagnostic")
Signed-off-by: Radoslaw Tyl <radoslawx.tyl@intel.com>
Reviewed-by: Michal Swiatkowski <michal.swiatkowski@linux.intel.com>
Tested-by: Arpana Arland <arpanax.arland@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20230328172659.3906413-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net...
Jakub Kicinski [Thu, 30 Mar 2023 04:46:18 +0000 (21:46 -0700)]
Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-03-28 (ice)

This series contains updates to ice driver only.

Jesse fixes mismatched header documentation reported when building with
W=1.

Brett restricts setting of VSI context to only applicable fields for the
given ICE_AQ_VSI_PROP_Q_OPT_VALID bit.

Junfeng adds check when adding Flow Director filters that conflict with
existing filter rules.

Jakob Koschel adds interim variable for iterating to prevent possible
misuse after looping.

* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  ice: fix invalid check for empty list in ice_sched_assoc_vsi_to_agg()
  ice: add profile conflict check for AVF FDIR
  ice: Fix ice_cfg_rdma_fltr() to only update relevant fields
  ice: fix W=1 headers mismatch
====================

Link: https://lore.kernel.org/r/20230328172035.3904953-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agoMerge tag 'ieee802154-for-net-2023-03-29' of git://git.kernel.org/pub/scm/linux/kerne...
Jakub Kicinski [Thu, 30 Mar 2023 04:41:12 +0000 (21:41 -0700)]
Merge tag 'ieee802154-for-net-2023-03-29' of git://git./linux/kernel/git/wpan/wpan

Stefan Schmidt says:

====================
ieee802154 for net 2023-03-29

Two small fixes this time.

Dongliang Mu removed an unnecessary null pointer check.

Harshit Mogalapalli fixed an int comparison unsigned against signed from a
recent other fix in the ca8210 driver.

* tag 'ieee802154-for-net-2023-03-29' of git://git.kernel.org/pub/scm/linux/kernel/git/wpan/wpan:
  net: ieee802154: remove an unnecessary null pointer check
  ca8210: Fix unsigned mac_len comparison with zero in ca8210_skb_tx()
====================

Link: https://lore.kernel.org/r/20230329064541.2147400-1-stefan@datenfreihafen.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agobnx2x: use the right build_skb() helper
Jakub Kicinski [Wed, 29 Mar 2023 00:00:13 +0000 (17:00 -0700)]
bnx2x: use the right build_skb() helper

build_skb() no longer accepts slab buffers. Since slab use is fairly
uncommon we prefer the drivers to call a separate slab_build_skb()
function appropriately.

bnx2x uses the old semantics where size of 0 meant buffer from slab.
It sets the fp->rx_frag_size to 0 for MTUs which don't fit in a page.
It needs to call slab_build_skb().

This fixes the WARN_ONCE() of incorrect API use seen with bnx2x.

Reported-by: Thomas Voegtle <tv@lio96.de>
Link: https://lore.kernel.org/all/b8f295e4-ba57-8bfb-7d9c-9d62a498a727@lio96.de/
Fixes: ce098da1497c ("skbuff: Introduce slab_build_skb()")
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20230329000013.2734957-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agonet: ipa: compute DMA pool size properly
Alex Elder [Tue, 28 Mar 2023 16:27:51 +0000 (11:27 -0500)]
net: ipa: compute DMA pool size properly

In gsi_trans_pool_init_dma(), the total size of a pool of memory
used for DMA transactions is calculated.  However the calculation is
done incorrectly.

For 4KB pages, this total size is currently always more than one
page, and as a result, the calculation produces a positive (though
incorrect) total size.  The code still works in this case; we just
end up with fewer DMA pool entries than we intended.

Bjorn Andersson tested booting a kernel with 16KB pages, and hit a
null pointer derereference in sg_alloc_append_table_from_pages(),
descending from gsi_trans_pool_init_dma().  The cause of this was
that a 16KB total size was going to be allocated, and with 16KB
pages the order of that allocation is 0.  The total_size calculation
yielded 0, which eventually led to the crash.

Correcting the total_size calculation fixes the problem.

Reported-by: Bjorn Andersson <quic_bjorande@quicinc.com>
Tested-by: Bjorn Andersson <quic_bjorande@quicinc.com>
Fixes: 9dd441e4ed57 ("soc: qcom: ipa: GSI transactions")
Reviewed-by: Mark Bloch <mbloch@nvidia.com>
Signed-off-by: Alex Elder <elder@linaro.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20230328162751.2861791-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agonet: wwan: iosm: fixes 7560 modem crash
M Chetan Kumar [Tue, 28 Mar 2023 06:28:44 +0000 (11:58 +0530)]
net: wwan: iosm: fixes 7560 modem crash

ModemManger/Apps probing the wwan0xmmrpc0 port for 7560 Modem results in
modem crash.

7560 Modem FW uses the MBIM interface for control command communication
whereas 7360 uses Intel RPC interface so disable wwan0xmmrpc0 port for
7560.

Fixes: d08b0f8f46e4 ("net: wwan: iosm: add rpc interface for xmm modems")
Reported-and-tested-by: Martin <mwolf@adiumentum.com>
Link: https://bugzilla.kernel.org/show_bug.cgi?id=217200
Signed-off-by: M Chetan Kumar <m.chetan.kumar@linux.intel.com>
Signed-off-by: Shane Parslow <shaneparslow808@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: ethernet: mtk_eth_soc: fix tx throughput regression with direct 1G links
Felix Fietkau [Fri, 24 Mar 2023 14:04:04 +0000 (15:04 +0100)]
net: ethernet: mtk_eth_soc: fix tx throughput regression with direct 1G links

Using the QDMA tx scheduler to throttle tx to line speed works fine for
switch ports, but apparently caused a regression on non-switch ports.

Based on a number of tests, it seems that this throttling can be safely
dropped without re-introducing the issues on switch ports that the
tx scheduling changes resolved.

Link: https://lore.kernel.org/netdev/trinity-92c3826f-c2c8-40af-8339-bc6d0d3ffea4-1678213958520@3c-app-gmx-bs16/
Fixes: f63959c7eec3 ("net: ethernet: mtk_eth_soc: implement multi-queue support for per-port queues")
Reported-by: Frank Wunderlich <frank-w@public-files.de>
Reported-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230324140404.95745-1-nbd@nbd.name
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agoice: fix invalid check for empty list in ice_sched_assoc_vsi_to_agg()
Jakob Koschel [Mon, 20 Mar 2023 12:48:15 +0000 (13:48 +0100)]
ice: fix invalid check for empty list in ice_sched_assoc_vsi_to_agg()

The code implicitly assumes that the list iterator finds a correct
handle. If 'vsi_handle' is not found the 'old_agg_vsi_info' was
pointing to an bogus memory location. For safety a separate list
iterator variable should be used to make the != NULL check on
'old_agg_vsi_info' correct under any circumstances.

Additionally Linus proposed to avoid any use of the list iterator
variable after the loop, in the attempt to move the list iterator
variable declaration into the macro to avoid any potential misuse after
the loop. Using it in a pointer comparison after the loop is undefined
behavior and should be omitted if possible [1].

Fixes: 37c592062b16 ("ice: remove the VSI info from previous agg")
Link: https://lore.kernel.org/all/CAHk-=wgRr_D8CB-D9Kg-c=EHreAsk5SqXPwr9Y7k9sA6cWXJ6w@mail.gmail.com/
Signed-off-by: Jakob Koschel <jkl820.git@gmail.com>
Tested-by: Arpana Arland <arpanax.arland@intel.com> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
20 months agoice: add profile conflict check for AVF FDIR
Junfeng Guo [Tue, 14 Mar 2023 02:03:15 +0000 (10:03 +0800)]
ice: add profile conflict check for AVF FDIR

Add profile conflict check while adding some FDIR rules to avoid
unexpected flow behavior, rules may have conflict including:
        IPv4 <---> {IPv4_UDP, IPv4_TCP, IPv4_SCTP}
        IPv6 <---> {IPv6_UDP, IPv6_TCP, IPv6_SCTP}

For example, when we create an FDIR rule for IPv4, this rule will work
on packets including IPv4, IPv4_UDP, IPv4_TCP and IPv4_SCTP. But if we
then create an FDIR rule for IPv4_UDP and then destroy it, the first
FDIR rule for IPv4 cannot work on pkt IPv4_UDP then.

To prevent this unexpected behavior, we add restriction in software
when creating FDIR rules by adding necessary profile conflict check.

Fixes: 1f7ea1cd6a37 ("ice: Enable FDIR Configure for AVF")
Signed-off-by: Junfeng Guo <junfeng.guo@intel.com>
Tested-by: Rafal Romanowski <rafal.romanowski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
20 months agoice: Fix ice_cfg_rdma_fltr() to only update relevant fields
Brett Creeley [Mon, 13 Mar 2023 20:36:08 +0000 (13:36 -0700)]
ice: Fix ice_cfg_rdma_fltr() to only update relevant fields

The current implementation causes ice_vsi_update() to update all VSI
fields based on the cached VSI context. This also assumes that the
ICE_AQ_VSI_PROP_Q_OPT_VALID bit is set. This can cause problems if the
VSI context is not correctly synced by the driver. Fix this by only
updating the fields that correspond to ICE_AQ_VSI_PROP_Q_OPT_VALID.
Also, make sure to save the updated result in the cached VSI context
on success.

Fixes: 348048e724a0 ("ice: Implement iidc operations")
Co-developed-by: Robert Malz <robertx.malz@intel.com>
Signed-off-by: Robert Malz <robertx.malz@intel.com>
Signed-off-by: Brett Creeley <brett.creeley@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reviewed-by: Piotr Raczynski <piotr.raczynski@intel.com>
Tested-by: Jakub Andrysiak <jakub.andrysiak@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
20 months agoice: fix W=1 headers mismatch
Jesse Brandeburg [Mon, 13 Mar 2023 20:36:07 +0000 (13:36 -0700)]
ice: fix W=1 headers mismatch

make modules W=1 returns:
.../ice/ice_txrx_lib.c:448: warning: Function parameter or member 'first_idx' not described in 'ice_finalize_xdp_rx'
.../ice/ice_txrx.c:948: warning: Function parameter or member 'ntc' not described in 'ice_get_rx_buf'
.../ice/ice_txrx.c:1038: warning: Excess function parameter 'rx_buf' description in 'ice_construct_skb'

Fix these warnings by adding and deleting the deviant arguments.

Fixes: 2fba7dc5157b ("ice: Add support for XDP multi-buffer on Rx side")
Fixes: d7956d81f150 ("ice: Pull out next_to_clean bump out of ice_put_rx_buf()")
CC: Maciej Fijalkowski <maciej.fijalkowski@intel.com>
Signed-off-by: Jesse Brandeburg <jesse.brandeburg@intel.com>
Reviewed-by: Piotr Raczynski <piotr.raczynski@intel.com>
Signed-off-by: Tony Nguyen <anthony.l.nguyen@intel.com>
20 months agoMerge branch 'xen-netback-fix-issue-introduced-recently'
Paolo Abeni [Tue, 28 Mar 2023 12:16:42 +0000 (14:16 +0200)]
Merge branch 'xen-netback-fix-issue-introduced-recently'

Juergen Gross says:

====================
xen/netback: fix issue introduced recently

The fix for XSA-423 introduced a bug which resulted in loss of network
connection in some configurations.

The first patch is fixing the issue, while the second one is removing
a test which isn't needed.
====================

Link: https://lore.kernel.org/r/20230327083646.18690-1-jgross@suse.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agoxen/netback: remove not needed test in xenvif_tx_build_gops()
Juergen Gross [Mon, 27 Mar 2023 08:36:46 +0000 (10:36 +0200)]
xen/netback: remove not needed test in xenvif_tx_build_gops()

The tests for the number of grant mapping or copy operations reaching
the array size of the operations buffer at the end of the main loop in
xenvif_tx_build_gops() isn't needed.

The loop can handle at maximum MAX_PENDING_REQS transfer requests, as
XEN_RING_NR_UNCONSUMED_REQUESTS() is taking unsent responses into
consideration, too.

Remove the tests.

Suggested-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agoxen/netback: don't do grant copy across page boundary
Juergen Gross [Mon, 27 Mar 2023 08:36:45 +0000 (10:36 +0200)]
xen/netback: don't do grant copy across page boundary

Fix xenvif_get_requests() not to do grant copy operations across local
page boundaries. This requires to double the maximum number of copy
operations per queue, as each copy could now be split into 2.

Make sure that struct xenvif_tx_cb doesn't grow too large.

Cc: stable@vger.kernel.org
Fixes: ad7f402ae4f4 ("xen/netback: Ensure protocol headers don't fall in the non-linear area")
Signed-off-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Paul Durrant <paul@xen.org>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agosmsc911x: avoid PHY being resumed when interface is not up
Wolfram Sang [Mon, 27 Mar 2023 08:31:38 +0000 (10:31 +0200)]
smsc911x: avoid PHY being resumed when interface is not up

SMSC911x doesn't need mdiobus suspend/resume, that's why it sets
'mac_managed_pm'. However, setting it needs to be moved from init to
probe, so mdiobus PM functions will really never be called (e.g. when
the interface is not up yet during suspend/resume).

Fixes: 3ce9f2bef755 ("net: smsc911x: Stop and start PHY during suspend and resume")
Suggested-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230327083138.6044-1-wsa+renesas@sang-engineering.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agoMerge branch 'net-mvpp2-rss-fixes'
Paolo Abeni [Tue, 28 Mar 2023 09:34:11 +0000 (11:34 +0200)]
Merge branch 'net-mvpp2-rss-fixes'

Sven Auhagen says:

====================
net: mvpp2: rss fixes

This patch series fixes up some rss problems
in the mvpp2 driver.

The classifier is missing some fragmentation flags,
the parser has the QinQ headers switched and
the PPPoE Layer 4 detecion is not working
correctly.

This is leading to no or bad rss for the default
settings.
====================

Link: https://lore.kernel.org/r/20230325163903.ofefgus43x66as7i@Svens-MacBookPro.local
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agonet: mvpp2: parser fix PPPoE
Sven Auhagen [Sat, 25 Mar 2023 16:41:05 +0000 (17:41 +0100)]
net: mvpp2: parser fix PPPoE

In PPPoE add all IPv4 header option length to the parser
and adjust the L3 and L4 offset accordingly.
Currently the L4 match does not work with PPPoE and
all packets are matched as L3 IP4 OPT.

Fixes: 3f518509dedc ("ethernet: Add new driver for Marvell Armada 375 network unit")
Signed-off-by: Sven Auhagen <sven.auhagen@voleatech.de>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agonet: mvpp2: parser fix QinQ
Sven Auhagen [Sat, 25 Mar 2023 16:40:53 +0000 (17:40 +0100)]
net: mvpp2: parser fix QinQ

The mvpp2 parser entry for QinQ has the inner and outer VLAN
in the wrong order.
Fix the problem by swapping them.

Fixes: 3f518509dedc ("ethernet: Add new driver for Marvell Armada 375 network unit")
Signed-off-by: Sven Auhagen <sven.auhagen@voleatech.de>
Reviewed-by: Marcin Wojtas <mw@semihalf.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agonet: mvpp2: classifier flow fix fragmentation flags
Sven Auhagen [Sat, 25 Mar 2023 16:40:29 +0000 (17:40 +0100)]
net: mvpp2: classifier flow fix fragmentation flags

Add missing IP Fragmentation Flag.

Fixes: f9358e12a0af ("net: mvpp2: split ingress traffic into multiple flows")
Signed-off-by: Sven Auhagen <sven.auhagen@voleatech.de>
Reviewed-by: Marcin Wojtas <mw@semihalf.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
20 months agoMerge tag 'linux-can-fixes-for-6.3-20230327' of git://git.kernel.org/pub/scm/linux...
Jakub Kicinski [Tue, 28 Mar 2023 02:47:42 +0000 (19:47 -0700)]
Merge tag 'linux-can-fixes-for-6.3-20230327' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2023-03-27

Oleksij Rempel and Hillf Danton contribute a patch for the CAN J1939
protocol that prevents a potential deadlock in j1939_sk_errqueue().

Ivan Orlov fixes an uninit-value in the CAN BCM protocol in the
bcm_tx_setup() function.

* tag 'linux-can-fixes-for-6.3-20230327' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can:
  can: bcm: bcm_tx_setup(): fix KMSAN uninit-value in vfs_write
  can: j1939: prevent deadlock by moving j1939_sk_errqueue()
====================

Link: https://lore.kernel.org/r/20230327124807.1157134-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agoMAINTAINERS: remove the linux-nfc@lists.01.org list
Lukas Bulwahn [Fri, 24 Mar 2023 08:16:13 +0000 (09:16 +0100)]
MAINTAINERS: remove the linux-nfc@lists.01.org list

Some MAINTAINERS sections mention to mail patches to the list
linux-nfc@lists.01.org. Probably due to changes on Intel's 01.org website
and servers, the list server lists.01.org/ml01.01.org is simply gone.

Considering emails recorded on lore.kernel.org, only a handful of emails
where sent to the linux-nfc@lists.01.org list, and they are usually also
sent to the netdev mailing list as well, where they are then picked up.
So, there is no big benefit in restoring the linux-nfc elsewhere.

Remove all occurrences of the linux-nfc@lists.01.org list in MAINTAINERS.

Suggested-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/all/CAKXUXMzggxQ43DUZZRkPMGdo5WkzgA=i14ySJUFw4kZfE5ZaZA@mail.gmail.com/
Signed-off-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Link: https://lore.kernel.org/r/20230324081613.32000-1-lukas.bulwahn@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agonet: fman: Add myself as a reviewer
Sean Anderson [Thu, 23 Mar 2023 14:59:57 +0000 (10:59 -0400)]
net: fman: Add myself as a reviewer

I've read through or reworked a good portion of this driver. Add myself
as a reviewer.

Signed-off-by: Sean Anderson <sean.anderson@seco.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Acked-by: Madalin Bucur <madalin.bucur@oss.nxp.com>
Link: https://lore.kernel.org/r/20230323145957.2999211-1-sean.anderson@seco.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agocan: bcm: bcm_tx_setup(): fix KMSAN uninit-value in vfs_write
Ivan Orlov [Tue, 14 Mar 2023 12:04:45 +0000 (16:04 +0400)]
can: bcm: bcm_tx_setup(): fix KMSAN uninit-value in vfs_write

Syzkaller reported the following issue:

=====================================================
BUG: KMSAN: uninit-value in aio_rw_done fs/aio.c:1520 [inline]
BUG: KMSAN: uninit-value in aio_write+0x899/0x950 fs/aio.c:1600
 aio_rw_done fs/aio.c:1520 [inline]
 aio_write+0x899/0x950 fs/aio.c:1600
 io_submit_one+0x1d1c/0x3bf0 fs/aio.c:2019
 __do_sys_io_submit fs/aio.c:2078 [inline]
 __se_sys_io_submit+0x293/0x770 fs/aio.c:2048
 __x64_sys_io_submit+0x92/0xd0 fs/aio.c:2048
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x63/0xcd

Uninit was created at:
 slab_post_alloc_hook mm/slab.h:766 [inline]
 slab_alloc_node mm/slub.c:3452 [inline]
 __kmem_cache_alloc_node+0x71f/0xce0 mm/slub.c:3491
 __do_kmalloc_node mm/slab_common.c:967 [inline]
 __kmalloc+0x11d/0x3b0 mm/slab_common.c:981
 kmalloc_array include/linux/slab.h:636 [inline]
 bcm_tx_setup+0x80e/0x29d0 net/can/bcm.c:930
 bcm_sendmsg+0x3a2/0xce0 net/can/bcm.c:1351
 sock_sendmsg_nosec net/socket.c:714 [inline]
 sock_sendmsg net/socket.c:734 [inline]
 sock_write_iter+0x495/0x5e0 net/socket.c:1108
 call_write_iter include/linux/fs.h:2189 [inline]
 aio_write+0x63a/0x950 fs/aio.c:1600
 io_submit_one+0x1d1c/0x3bf0 fs/aio.c:2019
 __do_sys_io_submit fs/aio.c:2078 [inline]
 __se_sys_io_submit+0x293/0x770 fs/aio.c:2048
 __x64_sys_io_submit+0x92/0xd0 fs/aio.c:2048
 do_syscall_x64 arch/x86/entry/common.c:50 [inline]
 do_syscall_64+0x3d/0xb0 arch/x86/entry/common.c:80
 entry_SYSCALL_64_after_hwframe+0x63/0xcd

CPU: 1 PID: 5034 Comm: syz-executor350 Not tainted 6.2.0-rc6-syzkaller-80422-geda666ff2276 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/12/2023
=====================================================

We can follow the call chain and find that 'bcm_tx_setup' function
calls 'memcpy_from_msg' to copy some content to the newly allocated
frame of 'op->frames'. After that the 'len' field of copied structure
being compared with some constant value (64 or 8). However, if
'memcpy_from_msg' returns an error, we will compare some uninitialized
memory. This triggers 'uninit-value' issue.

This patch will add 'memcpy_from_msg' possible errors processing to
avoid uninit-value issue.

Tested via syzkaller

Reported-by: syzbot+c9bfd85eca611ebf5db1@syzkaller.appspotmail.com
Link: https://syzkaller.appspot.com/bug?id=47f897f8ad958bbde5790ebf389b5e7e0a345089
Signed-off-by: Ivan Orlov <ivan.orlov0322@gmail.com>
Fixes: 6f3b911d5f29b ("can: bcm: add support for CAN FD frames")
Acked-by: Oliver Hartkopp <socketcan@hartkopp.net>
Link: https://lore.kernel.org/all/20230314120445.12407-1-ivan.orlov0322@gmail.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
20 months agonet: stmmac: don't reject VLANs when IFF_PROMISC is set
Vladimir Oltean [Sat, 25 Mar 2023 11:28:15 +0000 (13:28 +0200)]
net: stmmac: don't reject VLANs when IFF_PROMISC is set

The blamed commit has introduced the following tests to
dwmac4_add_hw_vlan_rx_fltr(), called from stmmac_vlan_rx_add_vid():

if (hw->promisc) {
netdev_err(dev,
   "Adding VLAN in promisc mode not supported\n");
return -EPERM;
}

"VLAN promiscuous" mode is keyed in this driver to IFF_PROMISC, and so,
vlan_vid_add() and vlan_vid_del() calls cannot take place in IFF_PROMISC
mode. I have the following 2 arguments that this restriction is.... hm,
how shall I put it nicely... unproductive :)

First, take the case of a Linux bridge. If the kernel is compiled with
CONFIG_BRIDGE_VLAN_FILTERING=y, then this bridge shall have a VLAN
database. The bridge shall try to call vlan_add_vid() on its bridge
ports for each VLAN in the VLAN table. It will do this irrespectively of
whether that port is *currently* VLAN-aware or not. So it will do this
even when the bridge was created with vlan_filtering 0.
But the Linux bridge, in VLAN-unaware mode, configures its ports in
promiscuous (IFF_PROMISC) mode, so that they accept packets with any
MAC DA (a switch must do this in order to forward those packets which
are not directly targeted to its MAC address).

As a result, the stmmac driver does not work as a bridge port, when the
kernel is compiled with CONFIG_BRIDGE_VLAN_FILTERING=y.

$ ip link add br0 type bridge && ip link set br0 up
$ ip link set eth0 master br0 && ip link set eth0 up
[ 2333.943296] br0: port 1(eth0) entered blocking state
[ 2333.943381] br0: port 1(eth0) entered disabled state
[ 2333.943782] device eth0 entered promiscuous mode
[ 2333.944080] 4033c000.ethernet eth0: Adding VLAN in promisc mode not supported
[ 2333.976509] 4033c000.ethernet eth0: failed to initialize vlan filtering on this port
RTNETLINK answers: Operation not permitted

Secondly, take the case of stmmac as DSA master. Some switch tagging
protocols are based on 802.1Q VLANs (tag_sja1105.c), and as such,
tag_8021q.c uses vlan_vid_add() to work with VLAN-filtering DSA masters.
But also, when a DSA port becomes promiscuous (for example when it joins
a bridge), the DSA framework also makes the DSA master promiscuous.

Moreover, for every VLAN that a DSA switch sends to the CPU, DSA also
programs a VLAN filter on the DSA master, because if the the DSA switch
uses a tail tag, then the hardware frame parser of the DSA master will
see VLAN as VLAN, and might filter them out, for being unknown.

Due to the above 2 reasons, my belief is that the stmmac driver does not
get to choose to not accept vlan_vid_add() calls while IFF_PROMISC is
enabled, because the 2 are completely independent and there are code
paths in the network stack which directly lead to this situation
occurring, without the user's direct input.

In fact, my belief is that "VLAN promiscuous" mode should have never
been keyed on IFF_PROMISC in the first place, but rather, on the
NETIF_F_HW_VLAN_CTAG_FILTER feature flag which can be toggled by the
user through ethtool -k, when present in netdev->hw_features.

In the stmmac driver, NETIF_F_HW_VLAN_CTAG_FILTER is only present in
"features", making this feature "on [fixed]".

I have this belief because I am unaware of any definition of promiscuity
which implies having an effect on anything other than MAC DA (therefore
not VLAN). However, I seem to be rather alone in having this opinion,
looking back at the disagreements from this discussion:
https://lore.kernel.org/netdev/20201110153958.ci5ekor3o2ekg3ky@ipetronik.com/

In any case, to remove the vlan_vid_add() dependency on !IFF_PROMISC,
one would need to remove the check and see what fails. I guess the test
was there because of the way in which dwmac4_vlan_promisc_enable() is
implemented.

For context, the dwmac4 supports Perfect Filtering for a limited number
of VLANs - dwmac4_get_num_vlan(), priv->hw->num_vlan, with a fallback on
Hash Filtering - priv->dma_cap.vlhash - see stmmac_vlan_update(), also
visible in cat /sys/kernel/debug/stmmaceth/eth0/dma_cap | grep 'VLAN
Hash Filtering'.

The perfect filtering is based on MAC_VLAN_Tag_Filter/MAC_VLAN_Tag_Data
registers, accessed in the driver through dwmac4_write_vlan_filter().

The hash filtering is based on the MAC_VLAN_Hash_Table register, named
GMAC_VLAN_HASH_TABLE in the driver and accessed by dwmac4_update_vlan_hash().
The control bit for enabling hash filtering is GMAC_VLAN_VTHM
(MAC_VLAN_Tag_Ctrl bit VTHM: VLAN Tag Hash Table Match Enable).

Now, the description of dwmac4_vlan_promisc_enable() is that it iterates
through the driver's cache of perfect filter entries (hw->vlan_filter[i],
added by dwmac4_add_hw_vlan_rx_fltr()), and evicts them from hardware by
unsetting their GMAC_VLAN_TAG_DATA_VEN (MAC_VLAN_Tag_Data bit VEN - VLAN
Tag Enable) bit. Then it unsets the GMAC_VLAN_VTHM bit, which disables
hash matching.

This leaves the MAC, according to table "VLAN Match Status" from the
documentation, to always enter these data paths:

VID    |VLAN Perfect Filter |VTHM Bit |VLAN Hash Filter |Final VLAN Match
       |Match Result        |         |Match Result     |Status
-------|--------------------|---------|-----------------|----------------
VID!=0 |Fail                |0        |don't care       |Pass

So, dwmac4_vlan_promisc_enable() does its job, but by unsetting
GMAC_VLAN_VTHM, it conflicts with the other code path which controls
this bit: dwmac4_update_vlan_hash(), called through stmmac_update_vlan_hash()
from stmmac_vlan_rx_add_vid() and from stmmac_vlan_rx_kill_vid().
This is, I guess, why dwmac4_add_hw_vlan_rx_fltr() is not allowed to run
after dwmac4_vlan_promisc_enable() has unset GMAC_VLAN_VTHM: because if
it did, then dwmac4_update_vlan_hash() would set GMAC_VLAN_VTHM again,
breaking the "VLAN promiscuity".

It turns out that dwmac4_vlan_promisc_enable() is way too complicated
for what needs to be done. The MAC_Packet_Filter register also has the
VTFE bit (VLAN Tag Filter Enable), which simply controls whether VLAN
tagged packets which don't match the filtering tables (either perfect or
hash) are dropped or not. At the moment, this driver unconditionally
sets GMAC_PACKET_FILTER_VTFE if NETIF_F_HW_VLAN_CTAG_FILTER was detected
through the priv->dma_cap.vlhash capability bits of the device, in
stmmac_dvr_probe().

I would suggest deleting the unnecessarily complex logic from
dwmac4_vlan_promisc_enable(), and simply unsetting GMAC_PACKET_FILTER_VTFE
when becoming IFF_PROMISC, which has the same effect of allowing packets
with any VLAN tags, but has the additional benefit of being able to run
concurrently with stmmac_vlan_rx_add_vid() and stmmac_vlan_rx_kill_vid().

As much as I believe that the VTFE bit should have been exclusively
controlled by NETIF_F_HW_VLAN_CTAG_FILTER through ethtool, and not by
IFF_PROMISC, changing that is not a punctual fix to the problem, and it
would probably break the VFFQ feature added by the later commit
e0f9956a3862 ("net: stmmac: Add option for VLAN filter fail queue
enable"). From the commit description, VFFQ needs IFF_PROMISC=on and
VTFE=off in order to work (and this change respects that). But if VTFE
was changed to be controlled through ethtool -k, then a user-visible
change would have been introduced in Intel's scripts (a need to run
"ethtool -k eth0 rx-vlan-filter off" which did not exist before).

The patch was tested with this set of commands:

  ip link set eth0 up
  ip link add link eth0 name eth0.100 type vlan id 100
  ip addr add 192.168.100.2/24 dev eth0.100 && ip link set eth0.100 up
  ip link set eth0 promisc on
  ip link add link eth0 name eth0.101 type vlan id 101
  ip addr add 192.168.101.2/24 dev eth0.101 && ip link set eth0.101 up
  ip link set eth0 promisc off
  ping -c 5 192.168.100.1
  ping -c 5 192.168.101.1
  ip link set eth0 promisc on
  ping -c 5 192.168.100.1
  ping -c 5 192.168.101.1
  ip link del eth0.100
  ip link del eth0.101
  # Wait for VLAN-tagged pings from the other end...
  # Check with "tcpdump -i eth0 -e -n -p" and we should see them
  ip link set eth0 promisc off
  # Wait for VLAN-tagged pings from the other end...
  # Check with "tcpdump -i eth0 -e -n -p" and we shouldn't see them
  # anymore, but remove the "-p" argument from tcpdump and they're there.

Fixes: c89f44ff10fd ("net: stmmac: Add support for VLAN promiscuous mode")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agocan: j1939: prevent deadlock by moving j1939_sk_errqueue()
Oleksij Rempel [Fri, 24 Mar 2023 13:01:41 +0000 (14:01 +0100)]
can: j1939: prevent deadlock by moving j1939_sk_errqueue()

This commit addresses a deadlock situation that can occur in certain
scenarios, such as when running data TP/ETP transfer and subscribing to
the error queue while receiving a net down event. The deadlock involves
locks in the following order:

3
  j1939_session_list_lock ->  active_session_list_lock
  j1939_session_activate
  ...
  j1939_sk_queue_activate_next -> sk_session_queue_lock
  ...
  j1939_xtp_rx_eoma_one

2
  j1939_sk_queue_drop_all  ->  sk_session_queue_lock
  ...
  j1939_sk_netdev_event_netdown -> j1939_socks_lock
  j1939_netdev_notify

1
  j1939_sk_errqueue -> j1939_socks_lock
  __j1939_session_cancel -> active_session_list_lock
  j1939_tp_rxtimer

       CPU0                    CPU1
       ----                    ----
  lock(&priv->active_session_list_lock);
                               lock(&jsk->sk_session_queue_lock);
                               lock(&priv->active_session_list_lock);
  lock(&priv->j1939_socks_lock);

The solution implemented in this commit is to move the
j1939_sk_errqueue() call out of the active_session_list_lock context,
thus preventing the deadlock situation.

Reported-by: syzbot+ee1cd780f69483a8616b@syzkaller.appspotmail.com
Fixes: 5b9272e93f2e ("can: j1939: extend UAPI to notify about RX status")
Co-developed-by: Hillf Danton <hdanton@sina.com>
Signed-off-by: Hillf Danton <hdanton@sina.com>
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Link: https://lore.kernel.org/all/20230324130141.2132787-1-o.rempel@pengutronix.de
Cc: stable@vger.kernel.org
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
20 months agonet: dsa: b53: mmap: add phy ops
Álvaro Fernández Rojas [Thu, 23 Mar 2023 19:48:41 +0000 (20:48 +0100)]
net: dsa: b53: mmap: add phy ops

Implement phy_read16() and phy_write16() ops for B53 MMAP to avoid accessing
B53_PORT_MII_PAGE registers which hangs the device.
This access should be done through the MDIO Mux bus controller.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: phy: micrel: correct KSZ9131RNX EEE capabilities and advertisement
Oleksij Rempel [Fri, 24 Mar 2023 13:39:08 +0000 (14:39 +0100)]
net: phy: micrel: correct KSZ9131RNX EEE capabilities and advertisement

The KSZ9131RNX incorrectly shows EEE capabilities in its registers.
Although the "EEE control and capability 1" (Register 3.20) is set to 0,
indicating no EEE support, the "EEE advertisement 1" (Register 7.60) is
set to 0x6, advertising EEE support for 1000BaseT/Full and
100BaseT/Full.
This inconsistency causes PHYlib to assume there is no EEE support,
preventing control over EEE advertisement, which is enabled by default.

This patch resolves the issue by utilizing the ksz9477_get_features()
function to correctly set the EEE capabilities for the KSZ9131RNX. This
adjustment allows proper control over EEE advertisement and ensures
accurate representation of the device's capabilities.

Fixes: 8b68710a3121 ("net: phy: start using genphy_c45_ethtool_get/set_eee()")
Reported-by: Marek Vasut <marex@denx.de>
Tested-by: Marek Vasut <marex@denx.de>
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agovsock/loopback: use only sk_buff_head.lock to protect the packet queue
Stefano Garzarella [Fri, 24 Mar 2023 11:54:50 +0000 (12:54 +0100)]
vsock/loopback: use only sk_buff_head.lock to protect the packet queue

pkt_list_lock was used before commit 71dc9ec9ac7d ("virtio/vsock:
replace virtio_vsock_pkt with sk_buff") to protect the packet queue.
After that commit we switched to sk_buff and we are using
sk_buff_head.lock in almost every place to protect the packet queue
except in vsock_loopback_work() when we call skb_queue_splice_init().

As reported by syzbot, this caused unlocked concurrent access to the
packet queue between vsock_loopback_work() and
vsock_loopback_cancel_pkt() since it is not holding pkt_list_lock.

With the introduction of sk_buff_head, pkt_list_lock is redundant and
can cause confusion, so let's remove it and use sk_buff_head.lock
everywhere to protect the packet queue access.

Fixes: 71dc9ec9ac7d ("virtio/vsock: replace virtio_vsock_pkt with sk_buff")
Cc: bobby.eshleman@bytedance.com
Reported-and-tested-by: syzbot+befff0a9536049e7902e@syzkaller.appspotmail.com
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Bobby Eshleman <bobby.eshleman@bytedance.com>
Reviewed-by: Arseniy Krasnov <AVKrasnov@sberdevices.ru>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agoMerge branch 'constify-sfp-phy-nodes'
David S. Miller [Mon, 27 Mar 2023 07:12:01 +0000 (08:12 +0100)]
Merge branch 'constify-sfp-phy-nodes'

Russell King says:

====================
Constify a few sfp/phy fwnodes

This series constifies a bunch of fwnode_handle pointers that are only
used to refer to but not modify the contents of the fwnode structures.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: phy: constify fwnode_get_phy_node() fwnode argument
Russell King (Oracle) [Fri, 24 Mar 2023 09:23:53 +0000 (09:23 +0000)]
net: phy: constify fwnode_get_phy_node() fwnode argument

fwnode_get_phy_node() does not motify the fwnode structure, so make
the argument const,

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: sfp: constify sfp-bus internal fwnode uses
Russell King (Oracle) [Fri, 24 Mar 2023 09:23:48 +0000 (09:23 +0000)]
net: sfp: constify sfp-bus internal fwnode uses

Constify sfp-bus internal fwnode uses, since we do not modify the
fwnode structures.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: sfp: make sfp_bus_find_fwnode() take a const fwnode
Russell King (Oracle) [Fri, 24 Mar 2023 09:23:42 +0000 (09:23 +0000)]
net: sfp: make sfp_bus_find_fwnode() take a const fwnode

sfp_bus_find_fwnode() does not write to the fwnode, so let's make it
const.

Signed-off-by: Russell King (Oracle) <rmk+kernel@armlinux.org.uk>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet/net_failover: fix txq exceeding warning
Faicker Mo [Fri, 24 Mar 2023 09:19:54 +0000 (17:19 +0800)]
net/net_failover: fix txq exceeding warning

The failover txq is inited as 16 queues.
when a packet is transmitted from the failover device firstly,
the failover device will select the queue which is returned from
the primary device if the primary device is UP and running.
If the primary device txq is bigger than the default 16,
it can lead to the following warning:
eth0 selects TX queue 18, but real number of TX queues is 16

The warning backtrace is:
[   32.146376] CPU: 18 PID: 9134 Comm: chronyd Tainted: G            E      6.2.8-1.el7.centos.x86_64 #1
[   32.147175] Hardware name: Red Hat KVM, BIOS 1.10.2-3.el7_4.1 04/01/2014
[   32.147730] Call Trace:
[   32.147971]  <TASK>
[   32.148183]  dump_stack_lvl+0x48/0x70
[   32.148514]  dump_stack+0x10/0x20
[   32.148820]  netdev_core_pick_tx+0xb1/0xe0
[   32.149180]  __dev_queue_xmit+0x529/0xcf0
[   32.149533]  ? __check_object_size.part.0+0x21c/0x2c0
[   32.149967]  ip_finish_output2+0x278/0x560
[   32.150327]  __ip_finish_output+0x1fe/0x2f0
[   32.150690]  ip_finish_output+0x2a/0xd0
[   32.151032]  ip_output+0x7a/0x110
[   32.151337]  ? __pfx_ip_finish_output+0x10/0x10
[   32.151733]  ip_local_out+0x5e/0x70
[   32.152054]  ip_send_skb+0x19/0x50
[   32.152366]  udp_send_skb.isra.0+0x163/0x3a0
[   32.152736]  udp_sendmsg+0xba8/0xec0
[   32.153060]  ? __folio_memcg_unlock+0x25/0x60
[   32.153445]  ? __pfx_ip_generic_getfrag+0x10/0x10
[   32.153854]  ? sock_has_perm+0x85/0xa0
[   32.154190]  inet_sendmsg+0x6d/0x80
[   32.154508]  ? inet_sendmsg+0x6d/0x80
[   32.154838]  sock_sendmsg+0x62/0x70
[   32.155152]  ____sys_sendmsg+0x134/0x290
[   32.155499]  ___sys_sendmsg+0x81/0xc0
[   32.155828]  ? _get_random_bytes.part.0+0x79/0x1a0
[   32.156240]  ? ip4_datagram_release_cb+0x5f/0x1e0
[   32.156649]  ? get_random_u16+0x69/0xf0
[   32.156989]  ? __fget_light+0xcf/0x110
[   32.157326]  __sys_sendmmsg+0xc4/0x210
[   32.157657]  ? __sys_connect+0xb7/0xe0
[   32.157995]  ? __audit_syscall_entry+0xce/0x140
[   32.158388]  ? syscall_trace_enter.isra.0+0x12c/0x1a0
[   32.158820]  __x64_sys_sendmmsg+0x24/0x30
[   32.159171]  do_syscall_64+0x38/0x90
[   32.159493]  entry_SYSCALL_64_after_hwframe+0x72/0xdc

Fix that by reducing txq number as the non-existent primary-dev does.

Fixes: cfc80d9a1163 ("net: Introduce net_failover driver")
Signed-off-by: Faicker Mo <faicker.mo@ucloud.cn>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agor8169: fix RTL8168H and RTL8107E rx crc error
ChunHao Lin [Thu, 23 Mar 2023 14:33:09 +0000 (22:33 +0800)]
r8169: fix RTL8168H and RTL8107E rx crc error

When link speed is 10 Mbps and temperature is under -20°C, RTL8168H and
RTL8107E may have rx crc error. Disable phy 10 Mbps pll off to fix this
issue.

Fixes: 6e1d0b898818 ("r8169:add support for RTL8168H and RTL8107E")
Signed-off-by: ChunHao Lin <hau@realtek.com>
Reviewed-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agoMerge branch 'ksz-fixes'
David S. Miller [Sat, 25 Mar 2023 11:17:32 +0000 (11:17 +0000)]
Merge branch 'ksz-fixes'

Oleksij Rempel says:

====================
net: dsa: microchip: ksz8: fixes for stable

changes v2:
- use proper Fixes tag
- add Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com> on all
  reviewed patches except the ksz8863_smi patch.

These fixes address issues such as incomplete FDB extraction, incorrect
FID extraction and configuration, incorrect timestamp extraction, and
ghost entry extraction from an empty dynamic MAC table. These updates
ensure proper functioning of the FDB/MDB functionality for the
ksz8863/ksz8873 series of chips.
====================

Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: dsa: microchip: ksz8: fix MDB configuration with non-zero VID
Oleksij Rempel [Fri, 24 Mar 2023 08:06:08 +0000 (09:06 +0100)]
net: dsa: microchip: ksz8: fix MDB configuration with non-zero VID

FID is directly mapped to VID. However, configuring a MAC address with a
VID != 0 resulted in incorrect configuration due to an incorrect bit
mask. This kernel commit fixed the issue by correcting the bit mask and
ensuring proper configuration of MAC addresses with non-zero VID.

Fixes: 4b20a07e103f ("net: dsa: microchip: ksz8795: add support for ksz88xx chips")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: dsa: microchip: ksz8863_smi: fix bulk access
Oleksij Rempel [Fri, 24 Mar 2023 08:06:07 +0000 (09:06 +0100)]
net: dsa: microchip: ksz8863_smi: fix bulk access

Current regmap bulk access is broken, resulting to wrong reads/writes
if ksz_read64/ksz_write64 functions are used.
Mostly this issue was visible by using ksz8_fdb_dump(), which returned
corrupt MAC address.

The reason is that regmap was configured to have max_raw_read/write,
even if ksz8863_mdio_read/write functions are able to handle unlimited
read/write accesses. On ksz_read64 function we are using multiple 32bit
accesses by incrementing each access by 1 instead of 4. Resulting buffer
had 01234567.12345678 instead of 01234567.89abcdef.

We have multiple ways to fix it:
- enable 4 byte alignment for 32bit accesses. Since the HW do not have
  this requirement. It will break driver.
- disable max_raw_* limit.

This patch is removing max_raw_* limit for regmap accesses in ksz8863_smi.

Fixes: 60a364760002 ("net: dsa: microchip: Add Microchip KSZ8863 SMI based driver support")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: dsa: microchip: ksz8: ksz8_fdb_dump: avoid extracting ghost entry from empty...
Oleksij Rempel [Fri, 24 Mar 2023 08:06:06 +0000 (09:06 +0100)]
net: dsa: microchip: ksz8: ksz8_fdb_dump: avoid extracting ghost entry from empty dynamic MAC table.

If the dynamic MAC table is empty, we will still extract one outdated
entry. Fix it by using correct bit offset.

Fixes: 4b20a07e103f ("net: dsa: microchip: ksz8795: add support for ksz88xx chips")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: dsa: microchip: ksz8: fix offset for the timestamp filed
Oleksij Rempel [Fri, 24 Mar 2023 08:06:05 +0000 (09:06 +0100)]
net: dsa: microchip: ksz8: fix offset for the timestamp filed

We are using wrong offset, so we will get not a timestamp.

Fixes: 4b20a07e103f ("net: dsa: microchip: ksz8795: add support for ksz88xx chips")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: dsa: microchip: ksz8: fix ksz8_fdb_dump() to extract all 1024 entries
Oleksij Rempel [Fri, 24 Mar 2023 08:06:04 +0000 (09:06 +0100)]
net: dsa: microchip: ksz8: fix ksz8_fdb_dump() to extract all 1024 entries

Current ksz8_fdb_dump() is able to extract only max 249 entries on
the ksz8863/ksz8873 series of switches. This happened due to wrong
bit mask and offset calculation.

This commit corrects the issue and allows for the complete extraction of
all 1024 entries.

Fixes: 4b20a07e103f ("net: dsa: microchip: ksz8795: add support for ksz88xx chips")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agonet: dsa: microchip: ksz8: fix ksz8_fdb_dump()
Oleksij Rempel [Fri, 24 Mar 2023 08:06:03 +0000 (09:06 +0100)]
net: dsa: microchip: ksz8: fix ksz8_fdb_dump()

Before this patch, the ksz8_fdb_dump() function had several issues, such
as uninitialized variables and incorrect usage of source port as a bit
mask. These problems caused inaccurate reporting of vid information and
port assignment in the bridge fdb.

Fixes: e587be759e6e ("net: dsa: microchip: update fdb add/del/dump in ksz_common")
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Acked-by: Arun Ramadoss <arun.ramadoss@microchip.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
20 months agoptp_qoriq: fix memory leak in probe()
SongJingyi [Fri, 24 Mar 2023 03:14:06 +0000 (11:14 +0800)]
ptp_qoriq: fix memory leak in probe()

Smatch complains that:
drivers/ptp/ptp_qoriq.c ptp_qoriq_probe()
warn: 'base' from ioremap() not released.

Fix this by revising the parameter from 'ptp_qoriq->base' to 'base'.
This is only a bug if ptp_qoriq_init() returns on the
first -ENODEV error path.
For other error paths ptp_qoriq->base and base are the same.
And this change makes the code more readable.

Fixes: 7f4399ba405b ("ptp_qoriq: fix NULL access if ptp dt node missing")
Signed-off-by: SongJingyi <u201912584@hust.edu.cn>
Reviewed-by: Dan Carpenter <error27@gmail.com>
Reviewed-by: Dongliang Mu <dzm91@hust.edu.cn>
Link: https://lore.kernel.org/r/20230324031406.1895159-1-u201912584@hust.edu.cn
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agovmxnet3: use gro callback when UPT is enabled
Ronak Doshi [Thu, 23 Mar 2023 20:07:21 +0000 (13:07 -0700)]
vmxnet3: use gro callback when UPT is enabled

Currently, vmxnet3 uses GRO callback only if LRO is disabled. However,
on smartNic based setups where UPT is supported, LRO can be enabled
from guest VM but UPT devicve does not support LRO as of now. In such
cases, there can be performance degradation as GRO is not being done.

This patch fixes this issue by calling GRO API when UPT is enabled. We
use updateRxProd to determine if UPT mode is active or not.

To clarify few things discussed over the thread:
The patch is not neglecting any feature bits nor disabling GRO. It uses
GRO callback when UPT is active as LRO is not available in UPT.
GRO callback cannot be used as default for all cases as it degrades
performance for non-UPT cases or for cases when LRO is already done in
ESXi.

Cc: stable@vger.kernel.org
Fixes: 6f91f4ba046e ("vmxnet3: add support for capability registers")
Signed-off-by: Ronak Doshi <doshir@vmware.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230323200721.27622-1-doshir@vmware.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agonet: dsa: realtek: fix out-of-bounds access
Ahmad Fatoum [Thu, 23 Mar 2023 10:37:35 +0000 (11:37 +0100)]
net: dsa: realtek: fix out-of-bounds access

The probe function sets priv->chip_data to (void *)priv + sizeof(*priv)
with the expectation that priv has enough trailing space.

However, only realtek-smi actually allocated this chip_data space.
Do likewise in realtek-mdio to fix out-of-bounds accesses.

These accesses likely went unnoticed so far, because of an (unused)
buf[4096] member in struct realtek_priv, which caused kmalloc to
round up the allocated buffer to a big enough size, so nothing of
value was overwritten. With a different allocator (like in the barebox
bootloader port of the driver) or with KASAN, the memory corruption
becomes quickly apparent.

Fixes: aac94001067d ("net: dsa: realtek: add new mdio interface for drivers")
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Reviewed-by: Alvin Šipraga <alsi@bang-olufsen.dk>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Ahmad Fatoum <a.fatoum@pengutronix.de>
Link: https://lore.kernel.org/r/20230323103735.2331786-1-a.fatoum@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agonet: phy: dp83869: fix default value for tx-/rx-internal-delay
Josua Mayer [Thu, 23 Mar 2023 10:25:36 +0000 (12:25 +0200)]
net: phy: dp83869: fix default value for tx-/rx-internal-delay

dp83869 internally uses a look-up table for mapping supported delays in
nanoseconds to register values.
When specific delays are defined in device-tree, phy_get_internal_delay
does the lookup automatically returning an index.

The default case wrongly assigns the nanoseconds value from the lookup
table, resulting in numeric value 2000 applied to delay configuration
register, rather than the expected index values 0-7 (7 for 2000).
Ultimately this issue broke RX for 1Gbps links.

Fix default delay configuration by assigning the intended index value
directly.

Cc: stable@vger.kernel.org
Fixes: 736b25afe284 ("net: dp83869: Add RGMII internal delay configuration")
Co-developed-by: Yazan Shhady <yazan.shhady@solid-run.com>
Signed-off-by: Yazan Shhady <yazan.shhady@solid-run.com>
Signed-off-by: Josua Mayer <josua@solid-run.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230323102536.31988-1-josua@solid-run.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agosfc: ef10: don't overwrite offload features at NIC reset
Íñigo Huguet [Thu, 23 Mar 2023 08:34:17 +0000 (09:34 +0100)]
sfc: ef10: don't overwrite offload features at NIC reset

At NIC reset, some offload features related to encapsulated traffic
might have changed (this mainly happens if the firmware-variant is
changed with the sfboot userspace tool). Because of this, features are
checked and set again at reset time.

However, this was not done right, and some features were improperly
overwritten at NIC reset:
- Tunneled IPv6 segmentation was always disabled
- Features disabled with ethtool were reenabled
- Features that becomes unsupported after the reset were not disabled

Also, checking if the device supports IPV6_CSUM to enable TSO6 is no
longer necessary because all currently supported devices support it.
Additionally, move the assignment of some other features to the
EF10_OFFLOAD_FEATURES macro, like it is done in ef100, leaving the
selection of features in efx_pci_probe_post_io a bit cleaner.

Fixes: ffffd2454a7a ("sfc: correctly advertise tunneled IPv6 segmentation")
Fixes: 24b2c3751aa3 ("sfc: advertise encapsulated offloads on EF10")
Reported-by: Tianhao Zhao <tizhao@redhat.com>
Suggested-by: Jonathan Cooper <jonathan.s.cooper@amd.com>
Tested-by: Jonathan Cooper <jonathan.s.cooper@amd.com>
Signed-off-by: Íñigo Huguet <ihuguet@redhat.com>
Acked-by: Edward Cree <ecree.xilinx@gmail.com>
Link: https://lore.kernel.org/r/20230323083417.7345-1-ihuguet@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agonet: wangxun: Fix vector length of interrupt cause
Jiawen Wu [Wed, 22 Mar 2023 10:36:32 +0000 (18:36 +0800)]
net: wangxun: Fix vector length of interrupt cause

There is 64-bit interrupt cause register for txgbe. Fix to clear upper
32 bits.

Fixes: 3f703186113f ("net: libwx: Add irq flow functions")
Signed-off-by: Jiawen Wu <jiawenwu@trustnetic.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/20230322103632.132011-1-jiawenwu@trustnetic.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
20 months agoMerge tag 'net-6.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Fri, 24 Mar 2023 15:48:12 +0000 (08:48 -0700)]
Merge tag 'net-6.3-rc4' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Including fixes from bpf, wifi and bluetooth.

  Current release - regressions:

   - wifi: mt76: mt7915: add back 160MHz channel width support for
     MT7915

   - libbpf: revert poisoning of strlcpy, it broke uClibc-ng

  Current release - new code bugs:

   - bpf: improve the coverage of the "allow reads from uninit stack"
     feature to fix verification complexity problems

   - eth: am65-cpts: reset PPS genf adj settings on enable

  Previous releases - regressions:

   - wifi: mac80211: serialize ieee80211_handle_wake_tx_queue()

   - wifi: mt76: do not run mt76_unregister_device() on unregistered hw,
     fix null-deref

   - Bluetooth: btqcomsmd: fix command timeout after setting BD address

   - eth: igb: revert rtnl_lock() that causes a deadlock

   - dsa: mscc: ocelot: fix device specific statistics

  Previous releases - always broken:

   - xsk: add missing overflow check in xdp_umem_reg()

   - wifi: mac80211:
      - fix QoS on mesh interfaces
      - fix mesh path discovery based on unicast packets

   - Bluetooth:
      - ISO: fix timestamped HCI ISO data packet parsing
      - remove "Power-on" check from Mesh feature

   - usbnet: more fixes to drivers trusting packet length

   - wifi: iwlwifi: mvm: fix mvmtxq->stopped handling

   - Bluetooth: btintel: iterate only bluetooth device ACPI entries

   - eth: iavf: fix inverted Rx hash condition leading to disabled hash

   - eth: igc: fix the validation logic for taprio's gate list

   - dsa: tag_brcm: legacy: fix daisy-chained switches

  Misc:

   - bpf: adjust insufficient default bpf_jit_limit to account for
     growth of BPF use over the last 5 years

   - xdp: bpf_xdp_metadata() use EOPNOTSUPP as unique errno indicating
     no driver support"

* tag 'net-6.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (84 commits)
  Bluetooth: HCI: Fix global-out-of-bounds
  Bluetooth: mgmt: Fix MGMT add advmon with RSSI command
  Bluetooth: btsdio: fix use after free bug in btsdio_remove due to unfinished work
  Bluetooth: L2CAP: Fix responding with wrong PDU type
  Bluetooth: btqcomsmd: Fix command timeout after setting BD address
  Bluetooth: btinel: Check ACPI handle for NULL before accessing
  net: mdio: thunder: Add missing fwnode_handle_put()
  net: dsa: mt7530: move setting ssc_delta to PHY_INTERFACE_MODE_TRGMII case
  net: dsa: mt7530: move lowering TRGMII driving to mt7530_setup()
  net: dsa: mt7530: move enabling disabling core clock to mt7530_pll_setup()
  net: asix: fix modprobe "sysfs: cannot create duplicate filename"
  gve: Cache link_speed value from device
  tools: ynl: Fix genlmsg header encoding formats
  net: enetc: fix aggregate RMON counters not showing the ranges
  Bluetooth: Remove "Power-on" check from Mesh feature
  Bluetooth: Fix race condition in hci_cmd_sync_clear
  Bluetooth: btintel: Iterate only bluetooth device ACPI entries
  Bluetooth: ISO: fix timestamped HCI ISO data packet parsing
  Bluetooth: btusb: Remove detection of ISO packets over bulk
  Bluetooth: hci_core: Detect if an ACL packet is in fact an ISO packet
  ...

20 months agoMerge tag 'for-6.3-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave...
Linus Torvalds [Fri, 24 Mar 2023 15:32:10 +0000 (08:32 -0700)]
Merge tag 'for-6.3-rc3-tag' of git://git./linux/kernel/git/kdave/linux

Pull btrfs fixes from David Sterba:
 "A few more fixes, the zoned accounting fix is spread across a few
  patches, preparatory and the actual fixes:

   - zoned mode:
      - fix accounting of unusable zone space
      - fix zone activation condition for DUP profile
      - preparatory patches

   - improved error handling of missing chunks

   - fix compiler warning"

* tag 'for-6.3-rc3-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/kdave/linux:
  btrfs: zoned: drop space_info->active_total_bytes
  btrfs: zoned: count fresh BG region as zone unusable
  btrfs: use temporary variable for space_info in btrfs_update_block_group
  btrfs: rename BTRFS_FS_NO_OVERCOMMIT to BTRFS_FS_ACTIVE_ZONE_TRACKING
  btrfs: zoned: fix btrfs_can_activate_zone() to support DUP profile
  btrfs: fix compiler warning on SPARC/PA-RISC handling fscrypt_setup_filename
  btrfs: handle missing chunk mapping more gracefully

20 months agoMerge tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Fri, 24 Mar 2023 15:27:13 +0000 (08:27 -0700)]
Merge tag 'scsi-fixes' of git://git./linux/kernel/git/jejb/scsi

Pull SCSI fixes from James Bottomley:
 "Four small fixes, three in drivers.

  The core fix adds a UFS device to an existing quirk to avoid a huge
  delay on boot"

* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: scsi_dh_alua: Fix memleak for 'qdata' in alua_activate()
  scsi: qla2xxx: Synchronize the IOCB count to be in order
  scsi: qla2xxx: Perform lockless command completion in abort path
  scsi: core: Add BLIST_SKIP_VPD_PAGES for SKhynix H28U74301AMR

21 months agoMerge tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf
Jakub Kicinski [Thu, 23 Mar 2023 23:03:33 +0000 (16:03 -0700)]
Merge tag 'for-netdev' of https://git./linux/kernel/git/bpf/bpf

Daniel Borkmann says:

====================
pull-request: bpf 2023-03-23

We've added 8 non-merge commits during the last 13 day(s) which contain
a total of 21 files changed, 238 insertions(+), 161 deletions(-).

The main changes are:

1) Fix verification issues in some BPF programs due to their stack usage
   patterns, from Eduard Zingerman.

2) Fix to add missing overflow checks in xdp_umem_reg and return an error
   in such case, from Kal Conley.

3) Fix and undo poisoning of strlcpy in libbpf given it broke builds for
   libcs which provided the former like uClibc-ng, from Jesus Sanchez-Palencia.

4) Fix insufficient bpf_jit_limit default to avoid users running into hard
   to debug seccomp BPF errors, from Daniel Borkmann.

5) Fix driver return code when they don't support a bpf_xdp_metadata kfunc
   to make it unambiguous from other errors, from Jesper Dangaard Brouer.

6) Two BPF selftest fixes to address compilation errors from recent changes
   in kernel structures, from Alexei Starovoitov.

* tag 'for-netdev' of https://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf:
  xdp: bpf_xdp_metadata use EOPNOTSUPP for no driver support
  bpf: Adjust insufficient default bpf_jit_limit
  xsk: Add missing overflow check in xdp_umem_reg
  selftests/bpf: Fix progs/test_deny_namespace.c issues.
  selftests/bpf: Fix progs/find_vma_fail1.c build error.
  libbpf: Revert poisoning of strlcpy
  selftests/bpf: Tests for uninitialized stack reads
  bpf: Allow reads from uninit stack
====================

Link: https://lore.kernel.org/r/20230323225221.6082-1-daniel@iogearbox.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agoMerge tag 'for-net-2023-03-23' of git://git.kernel.org/pub/scm/linux/kernel/git/bluet...
Jakub Kicinski [Thu, 23 Mar 2023 22:56:20 +0000 (15:56 -0700)]
Merge tag 'for-net-2023-03-23' of git://git./linux/kernel/git/bluetooth/bluetooth

Luiz Augusto von Dentz says:

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

 - Fix MGMT add advmon with RSSI command
 - L2CAP: Fix responding with wrong PDU type
 - Fix race condition in hci_cmd_sync_clear
 - ISO: Fix timestamped HCI ISO data packet parsing
 - HCI: Fix global-out-of-bounds
 - hci_sync: Resume adv with no RPA when active scan

* tag 'for-net-2023-03-23' of git://git.kernel.org/pub/scm/linux/kernel/git/bluetooth/bluetooth:
  Bluetooth: HCI: Fix global-out-of-bounds
  Bluetooth: mgmt: Fix MGMT add advmon with RSSI command
  Bluetooth: btsdio: fix use after free bug in btsdio_remove due to unfinished work
  Bluetooth: L2CAP: Fix responding with wrong PDU type
  Bluetooth: btqcomsmd: Fix command timeout after setting BD address
  Bluetooth: btinel: Check ACPI handle for NULL before accessing
  Bluetooth: Remove "Power-on" check from Mesh feature
  Bluetooth: Fix race condition in hci_cmd_sync_clear
  Bluetooth: btintel: Iterate only bluetooth device ACPI entries
  Bluetooth: ISO: fix timestamped HCI ISO data packet parsing
  Bluetooth: btusb: Remove detection of ISO packets over bulk
  Bluetooth: hci_core: Detect if an ACL packet is in fact an ISO packet
  Bluetooth: hci_sync: Resume adv with no RPA when active scan
====================

Link: https://lore.kernel.org/r/20230323202335.3380841-1-luiz.dentz@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agoMerge tag 'wireless-2023-03-23' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Thu, 23 Mar 2023 22:47:53 +0000 (15:47 -0700)]
Merge tag 'wireless-2023-03-23' of git://git./linux/kernel/git/wireless/wireless

Kalle Valo says:

====================
wireless fixes for v6.3

Third set of fixes for v6.3. mt76 has two kernel crash fixes and
adding back 160 MHz channel support for mt7915. mac80211 has fixes for
a race in transmit path and two mesh related fixes. iwlwifi also has
fixes for races.

* tag 'wireless-2023-03-23' of git://git.kernel.org/pub/scm/linux/kernel/git/wireless/wireless:
  wifi: mac80211: fix mesh path discovery based on unicast packets
  wifi: mac80211: fix qos on mesh interfaces
  wifi: iwlwifi: mvm: protect TXQ list manipulation
  wifi: iwlwifi: mvm: fix mvmtxq->stopped handling
  wifi: mac80211: Serialize ieee80211_handle_wake_tx_queue()
  wifi: mwifiex: mark OF related data as maybe unused
  wifi: mt76: connac: do not check WED status for non-mmio devices
  wifi: mt76: mt7915: add back 160MHz channel width support for MT7915
  wifi: mt76: do not run mt76_unregister_device() on unregistered hw
====================

Link: https://lore.kernel.org/r/20230323110332.C4FE4C433D2@smtp.kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agoMerge tag 'gfs2-v6.3-rc3-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2...
Linus Torvalds [Thu, 23 Mar 2023 22:25:49 +0000 (15:25 -0700)]
Merge tag 'gfs2-v6.3-rc3-fix' of git://git./linux/kernel/git/gfs2/linux-gfs2

Pull gfs2 fix from Andreas Gruenbacher:

 - Reinstate commit 970343cd4904 ("GFS2: free disk inode which is
   deleted by remote node -V2") as reverting that commit could cause
   gfs2_put_super() to hang.

* tag 'gfs2-v6.3-rc3-fix' of git://git.kernel.org/pub/scm/linux/kernel/git/gfs2/linux-gfs2:
  Reinstate "GFS2: free disk inode which is deleted by remote node -V2"

21 months agoBluetooth: HCI: Fix global-out-of-bounds
Sungwoo Kim [Tue, 21 Mar 2023 01:50:18 +0000 (21:50 -0400)]
Bluetooth: HCI: Fix global-out-of-bounds

To loop a variable-length array, hci_init_stage_sync(stage) considers
that stage[i] is valid as long as stage[i-1].func is valid.
Thus, the last element of stage[].func should be intentionally invalid
as hci_init0[], le_init2[], and others did.
However, amp_init1[] and amp_init2[] have no invalid element, letting
hci_init_stage_sync() keep accessing amp_init1[] over its valid range.
This patch fixes this by adding {} in the last of amp_init1[] and
amp_init2[].

==================================================================
BUG: KASAN: global-out-of-bounds in hci_dev_open_sync (
/v6.2-bzimage/net/bluetooth/hci_sync.c:3154
/v6.2-bzimage/net/bluetooth/hci_sync.c:3343
/v6.2-bzimage/net/bluetooth/hci_sync.c:4418
/v6.2-bzimage/net/bluetooth/hci_sync.c:4609
/v6.2-bzimage/net/bluetooth/hci_sync.c:4689)
Read of size 8 at addr ffffffffaed1ab70 by task kworker/u5:0/1032
CPU: 0 PID: 1032 Comm: kworker/u5:0 Not tainted 6.2.0 #3
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.15.0-1 04
Workqueue: hci1 hci_power_on
Call Trace:
 <TASK>
dump_stack_lvl (/v6.2-bzimage/lib/dump_stack.c:107 (discriminator 1))
print_report (/v6.2-bzimage/mm/kasan/report.c:307
  /v6.2-bzimage/mm/kasan/report.c:417)
? hci_dev_open_sync (/v6.2-bzimage/net/bluetooth/hci_sync.c:3154
  /v6.2-bzimage/net/bluetooth/hci_sync.c:3343
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4418
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4609
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4689)
kasan_report (/v6.2-bzimage/mm/kasan/report.c:184
  /v6.2-bzimage/mm/kasan/report.c:519)
? hci_dev_open_sync (/v6.2-bzimage/net/bluetooth/hci_sync.c:3154
  /v6.2-bzimage/net/bluetooth/hci_sync.c:3343
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4418
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4609
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4689)
hci_dev_open_sync (/v6.2-bzimage/net/bluetooth/hci_sync.c:3154
  /v6.2-bzimage/net/bluetooth/hci_sync.c:3343
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4418
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4609
  /v6.2-bzimage/net/bluetooth/hci_sync.c:4689)
? __pfx_hci_dev_open_sync (/v6.2-bzimage/net/bluetooth/hci_sync.c:4635)
? mutex_lock (/v6.2-bzimage/./arch/x86/include/asm/atomic64_64.h:190
  /v6.2-bzimage/./include/linux/atomic/atomic-long.h:443
  /v6.2-bzimage/./include/linux/atomic/atomic-instrumented.h:1781
  /v6.2-bzimage/kernel/locking/mutex.c:171
  /v6.2-bzimage/kernel/locking/mutex.c:285)
? __pfx_mutex_lock (/v6.2-bzimage/kernel/locking/mutex.c:282)
hci_power_on (/v6.2-bzimage/net/bluetooth/hci_core.c:485
  /v6.2-bzimage/net/bluetooth/hci_core.c:984)
? __pfx_hci_power_on (/v6.2-bzimage/net/bluetooth/hci_core.c:969)
? read_word_at_a_time (/v6.2-bzimage/./include/asm-generic/rwonce.h:85)
? strscpy (/v6.2-bzimage/./arch/x86/include/asm/word-at-a-time.h:62
  /v6.2-bzimage/lib/string.c:161)
process_one_work (/v6.2-bzimage/kernel/workqueue.c:2294)
worker_thread (/v6.2-bzimage/./include/linux/list.h:292
  /v6.2-bzimage/kernel/workqueue.c:2437)
? __pfx_worker_thread (/v6.2-bzimage/kernel/workqueue.c:2379)
kthread (/v6.2-bzimage/kernel/kthread.c:376)
? __pfx_kthread (/v6.2-bzimage/kernel/kthread.c:331)
ret_from_fork (/v6.2-bzimage/arch/x86/entry/entry_64.S:314)
 </TASK>
The buggy address belongs to the variable:
amp_init1+0x30/0x60
The buggy address belongs to the physical page:
page:000000003a157ec6 refcount:1 mapcount:0 mapping:0000000000000000 ia
flags: 0x200000000001000(reserved|node=0|zone=2)
raw: 0200000000001000 ffffea0005054688 ffffea0005054688 000000000000000
raw: 0000000000000000 0000000000000000 00000001ffffffff 000000000000000
page dumped because: kasan: bad access detected
Memory state around the buggy address:
 ffffffffaed1aa00: f9 f9 f9 f9 00 00 00 00 f9 f9 f9 f9 00 00 00 00
 ffffffffaed1aa80: 00 00 00 00 f9 f9 f9 f9 00 00 00 00 00 00 00 00
>ffffffffaed1ab00: 00 f9 f9 f9 f9 f9 f9 f9 00 00 00 00 00 00 f9 f9
                                                             ^
 ffffffffaed1ab80: f9 f9 f9 f9 00 00 00 00 f9 f9 f9 f9 00 00 00 f9
 ffffffffaed1ac00: f9 f9 f9 f9 00 06 f9 f9 f9 f9 f9 f9 00 00 02 f9

This bug is found by FuzzBT, a modified version of Syzkaller.
Other contributors for this bug are Ruoyu Wu and Peng Hui.

Fixes: d0b137062b2d ("Bluetooth: hci_sync: Rework init stages")
Signed-off-by: Sungwoo Kim <iam@sung-woo.kim>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: mgmt: Fix MGMT add advmon with RSSI command
Howard Chung [Thu, 16 Mar 2023 10:11:38 +0000 (18:11 +0800)]
Bluetooth: mgmt: Fix MGMT add advmon with RSSI command

The MGMT command: MGMT_OP_ADD_ADV_PATTERNS_MONITOR_RSSI uses variable
length argument. This causes host not able to register advmon with rssi.

This patch has been locally tested by adding monitor with rssi via
btmgmt on a kernel 6.1 machine.

Reviewed-by: Archie Pusaka <apusaka@chromium.org>
Fixes: b338d91703fa ("Bluetooth: Implement support for Mesh")
Signed-off-by: Howard Chung <howardchung@google.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: btsdio: fix use after free bug in btsdio_remove due to unfinished work
Zheng Wang [Thu, 9 Mar 2023 08:07:39 +0000 (16:07 +0800)]
Bluetooth: btsdio: fix use after free bug in btsdio_remove due to unfinished work

In btsdio_probe, &data->work was bound with btsdio_work.In
btsdio_send_frame, it was started by schedule_work.

If we call btsdio_remove with an unfinished job, there may
be a race condition and cause UAF bug on hdev.

Fixes: ddbaf13e3609 ("[Bluetooth] Add generic driver for Bluetooth SDIO devices")
Signed-off-by: Zheng Wang <zyytlz.wz@163.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: L2CAP: Fix responding with wrong PDU type
Luiz Augusto von Dentz [Wed, 8 Mar 2023 22:20:34 +0000 (14:20 -0800)]
Bluetooth: L2CAP: Fix responding with wrong PDU type

L2CAP_ECRED_CONN_REQ shall be responded with L2CAP_ECRED_CONN_RSP not
L2CAP_LE_CONN_RSP:

L2CAP LE EATT Server - Reject - run
  Listening for connections
  New client connection with handle 0x002a
  Sending L2CAP Request from client
  Client received response code 0x15
  Unexpected L2CAP response code (expected 0x18)
L2CAP LE EATT Server - Reject - test failed

> ACL Data RX: Handle 42 flags 0x02 dlen 26
      LE L2CAP: Enhanced Credit Connection Request (0x17) ident 1 len 18
        PSM: 39 (0x0027)
        MTU: 64
        MPS: 64
        Credits: 5
        Source CID: 65
        Source CID: 66
        Source CID: 67
        Source CID: 68
        Source CID: 69
< ACL Data TX: Handle 42 flags 0x00 dlen 16
      LE L2CAP: LE Connection Response (0x15) ident 1 len 8
        invalid size
        00 00 00 00 00 00 06 00

L2CAP LE EATT Server - Reject - run
  Listening for connections
  New client connection with handle 0x002a
  Sending L2CAP Request from client
  Client received response code 0x18
L2CAP LE EATT Server - Reject - test passed

Fixes: 15f02b910562 ("Bluetooth: L2CAP: Add initial code for Enhanced Credit Based Mode")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: btqcomsmd: Fix command timeout after setting BD address
Stephan Gerhold [Wed, 8 Mar 2023 13:31:55 +0000 (14:31 +0100)]
Bluetooth: btqcomsmd: Fix command timeout after setting BD address

On most devices using the btqcomsmd driver (e.g. the DragonBoard 410c
and other devices based on the Qualcomm MSM8916/MSM8909/... SoCs)
the Bluetooth firmware seems to become unresponsive for a while after
setting the BD address. On recent kernel versions (at least 5.17+)
this often causes timeouts for subsequent commands, e.g. the HCI reset
sent by the Bluetooth core during initialization:

    Bluetooth: hci0: Opcode 0x c03 failed: -110

Unfortunately this behavior does not seem to be documented anywhere.
Experimentation suggests that the minimum necessary delay to avoid
the problem is ~150us. However, to be sure add a sleep for > 1ms
in case it is a bit longer on other firmware versions.

Older kernel versions are likely also affected, although perhaps with
slightly different errors or less probability. Side effects can easily
hide the issue in most cases, e.g. unrelated incoming interrupts that
cause the necessary delay.

Fixes: 1511cc750c3d ("Bluetooth: Introduce Qualcomm WCNSS SMD based HCI driver")
Signed-off-by: Stephan Gerhold <stephan.gerhold@kernkonzept.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: btinel: Check ACPI handle for NULL before accessing
Kiran K [Wed, 8 Mar 2023 07:58:37 +0000 (13:28 +0530)]
Bluetooth: btinel: Check ACPI handle for NULL before accessing

Older platforms and Virtual platforms which doesn't have support for
bluetooth device in ACPI firmware will not have valid ACPI handle.
Check for validity of handle before accessing.

dmesg log from simics environment (virtual platform):

BUG: unable to handle kernel NULL pointer dereference at
0000000000000018
IP: acpi_ns_walk_namespace+0x5c/0x278
PGD 0 P4D 0
Oops: 0000 [#1] SMP PTI
Modules linked in: bnep intel_powerclamp coretemp kvm_intel
kvm irqbypass intel_cstate input_leds joydev serio_raw mac_hid
btusb(OE) btintel(OE) bluetooth(OE) lpc_ich compat(OE) ecdh_generic
i7core_edac i5500_temp shpchp binfmt_misc sch_fq_codel parport_pc ppdev
lp parport ip_tables x_tables autofs4 hid_generic usbhid hid e1000e
psmouse ahci pata_acpi libahci ptp pps_core floppy
CPU: 0 PID: 35 Comm: kworker/u3:0 Tainted: G           OE
4.15.0-140-generic #144-Ubuntu
Hardware name: Simics Simics, BIOS Simics 01/01/2011
Workqueue: hci0 hci_power_on [bluetooth]
RIP: 0010:acpi_ns_walk_namespace+0x5c/0x278
RSP: 0000:ffffaa9c0049bba8 EFLAGS: 00010246
RAX: 0000000000000001 RBX: 0000000000001001 RCX: 0000000000000010
RDX: ffffffff92ea7e27 RSI: ffffffff92ea7e10 RDI: 00000000000000c8
RBP: ffffaa9c0049bbf8 R08: 0000000000000000 R09: ffffffffc05b39d0
R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000001
R13: 0000000000000000 R14: ffffffffc05b39d0 R15: ffffaa9c0049bc70
FS:  0000000000000000(0000) GS:ffff8be73fc00000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000018 CR3: 0000000075f0e000 CR4: 00000000000006f0

Fixes: 294d749b5df5 ("Bluetooth: btintel: Iterate only bluetooth device ACPI entries")
Signed-off-by: Kiran K <kiran.k@intel.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoReinstate "GFS2: free disk inode which is deleted by remote node -V2"
Bob Peterson [Thu, 23 Mar 2023 16:26:02 +0000 (12:26 -0400)]
Reinstate "GFS2: free disk inode which is deleted by remote node -V2"

It turns out that reverting commit 970343cd4904 ("GFS2: free disk inode
which is deleted by remote node -V2") causes a regression related to
evicting inodes that were unlinked on a different cluster node.

We could also have simply added a call to d_mark_dontcache() to function
gfs2_try_evict(), but the original pre-revert code is better tested and
proven.

This reverts commit 445cb1277e10d7e19b631ef8a64aa3f055df377d.

Signed-off-by: Bob Peterson <rpeterso@redhat.com>
Signed-off-by: Andreas Gruenbacher <agruenba@redhat.com>
21 months agoMerge tag 'zonefs-6.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal...
Linus Torvalds [Thu, 23 Mar 2023 18:05:08 +0000 (11:05 -0700)]
Merge tag 'zonefs-6.3-rc4' of git://git./linux/kernel/git/dlemoal/zonefs

Pull zonefs fixes from Damien Le Moal:

 - Silence a false positive smatch warning about an uninitialized
   variable

 - Fix an error message to provide more useful information about invalid
   zone append write results

* tag 'zonefs-6.3-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/dlemoal/zonefs:
  zonefs: Fix error message in zonefs_file_dio_append()
  zonefs: Prevent uninitialized symbol 'size' warning

21 months agonet: mdio: thunder: Add missing fwnode_handle_put()
Liang He [Wed, 22 Mar 2023 06:20:57 +0000 (14:20 +0800)]
net: mdio: thunder: Add missing fwnode_handle_put()

In device_for_each_child_node(), we should add fwnode_handle_put()
when break out of the iteration device_for_each_child_node()
as it will automatically increase and decrease the refcounter.

Fixes: 379d7ac7ca31 ("phy: mdio-thunder: Add driver for Cavium Thunder SoC MDIO buses.")
Signed-off-by: Liang He <windhl@126.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
21 months agoMerge tag 'mlx5-fixes-2023-03-21' of git://git.kernel.org/pub/scm/linux/kernel/git...
Jakub Kicinski [Thu, 23 Mar 2023 05:50:01 +0000 (22:50 -0700)]
Merge tag 'mlx5-fixes-2023-03-21' of git://git./linux/kernel/git/saeed/linux

Saeed Mahameed says:

====================
mlx5 fixes 2023-03-21

This series provides bug fixes to mlx5 driver.

* tag 'mlx5-fixes-2023-03-21' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
  net/mlx5: E-Switch, Fix an Oops in error handling code
  net/mlx5: Read the TC mapping of all priorities on ETS query
  net/mlx5e: Overcome slow response for first macsec ASO WQE
  net/mlx5e: Initialize link speed to zero
  net/mlx5: Fix steering rules cleanup
  net/mlx5e: Block entering switchdev mode with ns inconsistency
  net/mlx5e: Set uplink rep as NETNS_LOCAL
====================

Link: https://lore.kernel.org/r/20230321211135.47711-1-saeed@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agoMerge branch '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net...
Jakub Kicinski [Thu, 23 Mar 2023 05:45:22 +0000 (22:45 -0700)]
Merge branch '100GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-03-21 (ice)

This series contains updates to ice driver only.

Piotr sets first_desc field for proper handling of Flow Director
packets.

Michal moves error checking for VF earlier in function to properly return
error before other checks/reporting; he also corrects VSI filter removal to
be done during VSI removal and not rebuild.

* '100GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  ice: remove filters only if VSI is deleted
  ice: check if VF exists before mode check
  ice: fix rx buffers handling for flow director packets
====================

Link: https://lore.kernel.org/r/20230321183641.2849726-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agoMerge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Jakub Kicinski [Thu, 23 Mar 2023 05:42:15 +0000 (22:42 -0700)]
Merge branch '40GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2023-03-21 (iavf, i40e)

This series contains updates to iavf and i40e drivers.

Stefan Assmann adds check, and return, if driver has already gone
through remove to prevent hang for iavf.

Radoslaw adds zero initialization to ensure Flow Director packets are
populated with correct values for i40e.

* '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  i40e: fix flow director packet filter programming
  iavf: fix hang on reboot with ice
====================

Link: https://lore.kernel.org/r/20230321183548.2849671-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agonet: dsa: mt7530: move setting ssc_delta to PHY_INTERFACE_MODE_TRGMII case
Arınç ÜNAL [Mon, 20 Mar 2023 19:05:20 +0000 (22:05 +0300)]
net: dsa: mt7530: move setting ssc_delta to PHY_INTERFACE_MODE_TRGMII case

Move setting the ssc_delta variable to under the PHY_INTERFACE_MODE_TRGMII
case as it's only needed when trgmii is used.

Fixes: b8f126a8d543 ("net-next: dsa: add dsa support for Mediatek MT7530 switch")
Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Link: https://lore.kernel.org/r/20230320190520.124513-3-arinc.unal@arinc9.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agonet: dsa: mt7530: move lowering TRGMII driving to mt7530_setup()
Arınç ÜNAL [Mon, 20 Mar 2023 19:05:19 +0000 (22:05 +0300)]
net: dsa: mt7530: move lowering TRGMII driving to mt7530_setup()

Move lowering the TRGMII Tx clock driving to mt7530_setup(), after setting
the core clock, as seen on the U-Boot MediaTek ethernet driver.

Move the code which looks like it lowers the TRGMII Rx clock driving to
after the TRGMII Tx clock driving is lowered. This is run after lowering
the Tx clock driving on the U-Boot MediaTek ethernet driver as well.

This way, the switch should consume less power regardless of port 6 being
used.

Update the comment explaining mt7530_pad_clk_setup().

Tested rgmii and trgmii modes of port 6 and rgmii mode of port 5 on MCM
MT7530 on MT7621AT Unielec U7621-06 and standalone MT7530 on MT7623NI
Bananapi BPI-R2.

Fixes: b8f126a8d543 ("net-next: dsa: add dsa support for Mediatek MT7530 switch")
Link: https://source.denx.de/u-boot/u-boot/-/blob/29a48bf9ccba45a5e560bb564bbe76e42629325f/drivers/net/mtk_eth.c#L682
Tested-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Link: https://lore.kernel.org/r/20230320190520.124513-2-arinc.unal@arinc9.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agonet: dsa: mt7530: move enabling disabling core clock to mt7530_pll_setup()
Arınç ÜNAL [Mon, 20 Mar 2023 19:05:18 +0000 (22:05 +0300)]
net: dsa: mt7530: move enabling disabling core clock to mt7530_pll_setup()

Split the code that enables and disables TRGMII clocks and core clock.
Move enabling and disabling core clock to mt7530_pll_setup() as it's
supposed to be run there.

Add 20 ms delay before enabling the core clock as seen on the U-Boot
MediaTek ethernet driver.

Change the comment for enabling and disabling TRGMII clocks as the code
seems to affect both TXC and RXC.

Tested rgmii and trgmii modes of port 6 and rgmii mode of port 5 on MCM
MT7530 on MT7621AT Unielec U7621-06 and standalone MT7530 on MT7623NI
Bananapi BPI-R2.

Fixes: b8f126a8d543 ("net-next: dsa: add dsa support for Mediatek MT7530 switch")
Link: https://source.denx.de/u-boot/u-boot/-/blob/29a48bf9ccba45a5e560bb564bbe76e42629325f/drivers/net/mtk_eth.c#L589
Tested-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Link: https://lore.kernel.org/r/20230320190520.124513-1-arinc.unal@arinc9.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agonet: asix: fix modprobe "sysfs: cannot create duplicate filename"
Grant Grundler [Tue, 21 Mar 2023 17:05:39 +0000 (10:05 -0700)]
net: asix: fix modprobe "sysfs: cannot create duplicate filename"

"modprobe asix ; rmmod asix ; modprobe asix" fails with:
   sysfs: cannot create duplicate filename \
    '/devices/virtual/mdio_bus/usb-003:004'

Issue was originally reported by Anton Lundin on 2022-06-22 (link below).

Chrome OS team hit the same issue in Feb, 2023 when trying to find
work arounds for other issues with AX88172 devices.

The use of devm_mdiobus_register() with usbnet devices results in the
MDIO data being associated with the USB device. When the asix driver
is unloaded, the USB device continues to exist and the corresponding
"mdiobus_unregister()" is NOT called until the USB device is unplugged
or unauthorized. So the next "modprobe asix" will fail because the MDIO
phy sysfs attributes still exist.

The 'easy' (from a design PoV) fix is to use the non-devm variants of
mdiobus_* functions and explicitly manage this use in the asix_bind
and asix_unbind function calls. I've not explored trying to fix usbnet
initialization so devm_* stuff will work.

Fixes: e532a096be0e5 ("net: usb: asix: ax88772: add phylib support")
Reported-by: Anton Lundin <glance@acc.umu.se>
Link: https://lore.kernel.org/netdev/20220623063649.GD23685@pengutronix.de/T/
Tested-by: Eizan Miyamoto <eizan@chromium.org>
Signed-off-by: Grant Grundler <grundler@chromium.org>
Link: https://lore.kernel.org/r/20230321170539.732147-1-grundler@chromium.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agogve: Cache link_speed value from device
Joshua Washington [Tue, 21 Mar 2023 17:23:32 +0000 (10:23 -0700)]
gve: Cache link_speed value from device

The link speed is never changed for the uptime of a VM, and the current
implementation sends an admin queue command for each call. Admin queue
command invocations have nontrivial overhead (e.g., VM exits), which can
be disruptive to users if triggered frequently. Our telemetry data shows
that there are VMs that make frequent calls to this admin queue command.
Caching the result of the original admin queue command would eliminate
the need to send multiple admin queue commands on subsequent calls to
retrieve link speed.

Fixes: 7e074d5a76ca ("gve: Enable Link Speed Reporting in the driver.")
Signed-off-by: Joshua Washington <joshwash@google.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230321172332.91678-1-joshwash@google.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agotools: ynl: Fix genlmsg header encoding formats
Donald Hunter [Sun, 19 Mar 2023 19:37:58 +0000 (19:37 +0000)]
tools: ynl: Fix genlmsg header encoding formats

The pack strings use 'b' signed char for cmd and version but struct
genlmsghdr defines them as unsigned char. Use 'B' instead.

Fixes: 4e4480e89c47 ("tools: ynl: move the cli and netlink code around")
Signed-off-by: Donald Hunter <donald.hunter@gmail.com>
Link: https://lore.kernel.org/r/20230319193803.97453-1-donald.hunter@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agonet: enetc: fix aggregate RMON counters not showing the ranges
Vladimir Oltean [Tue, 21 Mar 2023 23:28:31 +0000 (01:28 +0200)]
net: enetc: fix aggregate RMON counters not showing the ranges

When running "ethtool -S eno0 --groups rmon" without an explicit "--src
emac|pmac" argument, the kernel will not report
rx-rmon-etherStatsPkts64to64Octets, rx-rmon-etherStatsPkts65to127Octets,
etc. This is because on ETHTOOL_MAC_STATS_SRC_AGGREGATE, we do not
populate the "ranges" argument.

ocelot_port_get_rmon_stats() does things differently and things work
there. I had forgotten to make sure that the code is structured the same
way in both drivers, so do that now.

Fixes: cf52bd238b75 ("net: enetc: add support for MAC Merge statistics counters")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/20230321232831.1200905-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agoBluetooth: Remove "Power-on" check from Mesh feature
Brian Gix [Mon, 6 Mar 2023 22:32:21 +0000 (14:32 -0800)]
Bluetooth: Remove "Power-on" check from Mesh feature

The Bluetooth mesh experimental feature enable was requiring the
controller to be powered off in order for the Enable to work. Mesh is
supposed to be enablable regardless of the controller state, and created
an unintended requirement that the mesh daemon be started before the
classic bluetoothd daemon.

Fixes: af6bcc1921ff ("Bluetooth: Add experimental wrapper for MGMT based mesh")
Signed-off-by: Brian Gix <brian.gix@gmail.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: Fix race condition in hci_cmd_sync_clear
Min Li [Sat, 4 Mar 2023 13:50:35 +0000 (21:50 +0800)]
Bluetooth: Fix race condition in hci_cmd_sync_clear

There is a potential race condition in hci_cmd_sync_work and
hci_cmd_sync_clear, and could lead to use-after-free. For instance,
hci_cmd_sync_work is added to the 'req_workqueue' after cancel_work_sync
The entry of 'cmd_sync_work_list' may be freed in hci_cmd_sync_clear, and
causing kernel panic when it is used in 'hci_cmd_sync_work'.

Here's the call trace:

dump_stack_lvl+0x49/0x63
print_report.cold+0x5e/0x5d3
? hci_cmd_sync_work+0x282/0x320
kasan_report+0xaa/0x120
? hci_cmd_sync_work+0x282/0x320
__asan_report_load8_noabort+0x14/0x20
hci_cmd_sync_work+0x282/0x320
process_one_work+0x77b/0x11c0
? _raw_spin_lock_irq+0x8e/0xf0
worker_thread+0x544/0x1180
? poll_idle+0x1e0/0x1e0
kthread+0x285/0x320
? process_one_work+0x11c0/0x11c0
? kthread_complete_and_exit+0x30/0x30
ret_from_fork+0x22/0x30
</TASK>

Allocated by task 266:
kasan_save_stack+0x26/0x50
__kasan_kmalloc+0xae/0xe0
kmem_cache_alloc_trace+0x191/0x350
hci_cmd_sync_queue+0x97/0x2b0
hci_update_passive_scan+0x176/0x1d0
le_conn_complete_evt+0x1b5/0x1a00
hci_le_conn_complete_evt+0x234/0x340
hci_le_meta_evt+0x231/0x4e0
hci_event_packet+0x4c5/0xf00
hci_rx_work+0x37d/0x880
process_one_work+0x77b/0x11c0
worker_thread+0x544/0x1180
kthread+0x285/0x320
ret_from_fork+0x22/0x30

Freed by task 269:
kasan_save_stack+0x26/0x50
kasan_set_track+0x25/0x40
kasan_set_free_info+0x24/0x40
____kasan_slab_free+0x176/0x1c0
__kasan_slab_free+0x12/0x20
slab_free_freelist_hook+0x95/0x1a0
kfree+0xba/0x2f0
hci_cmd_sync_clear+0x14c/0x210
hci_unregister_dev+0xff/0x440
vhci_release+0x7b/0xf0
__fput+0x1f3/0x970
____fput+0xe/0x20
task_work_run+0xd4/0x160
do_exit+0x8b0/0x22a0
do_group_exit+0xba/0x2a0
get_signal+0x1e4a/0x25b0
arch_do_signal_or_restart+0x93/0x1f80
exit_to_user_mode_prepare+0xf5/0x1a0
syscall_exit_to_user_mode+0x26/0x50
ret_from_fork+0x15/0x30

Fixes: 6a98e3836fa2 ("Bluetooth: Add helper for serialized HCI command execution")
Cc: stable@vger.kernel.org
Signed-off-by: Min Li <lm0963hack@gmail.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: btintel: Iterate only bluetooth device ACPI entries
Kiran K [Tue, 28 Feb 2023 11:01:54 +0000 (16:31 +0530)]
Bluetooth: btintel: Iterate only bluetooth device ACPI entries

Current flow interates over entire ACPI table entries looking for
Bluetooth Per Platform Antenna Gain(PPAG) entry. This patch iterates
over ACPI entries relvant to Bluetooth device only.

Fixes: c585a92b2f9c ("Bluetooth: btintel: Set Per Platform Antenna Gain(PPAG)")
Signed-off-by: Kiran K <kiran.k@intel.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: ISO: fix timestamped HCI ISO data packet parsing
Pauli Virtanen [Mon, 20 Feb 2023 19:38:24 +0000 (19:38 +0000)]
Bluetooth: ISO: fix timestamped HCI ISO data packet parsing

Use correct HCI ISO data packet header struct when the packet has
timestamp. The timestamp, when present, goes before the other fields
(Core v5.3 4E 5.4.5), so the structs are not compatible.

Fixes: ccf74f2390d6 ("Bluetooth: Add BTPROTO_ISO socket type")
Signed-off-by: Pauli Virtanen <pav@iki.fi>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: btusb: Remove detection of ISO packets over bulk
Luiz Augusto von Dentz [Fri, 24 Feb 2023 23:54:31 +0000 (15:54 -0800)]
Bluetooth: btusb: Remove detection of ISO packets over bulk

This removes the code introduced by
14202eff214e1e941fefa0366d4c3bc4b1a0d500 as hci_recv_frame is now able
to detect ACL packets that are in fact ISO packets.

Fixes: 14202eff214e ("Bluetooth: btusb: Detect if an ACL packet is in fact an ISO packet")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: hci_core: Detect if an ACL packet is in fact an ISO packet
Luiz Augusto von Dentz [Fri, 24 Feb 2023 23:43:31 +0000 (15:43 -0800)]
Bluetooth: hci_core: Detect if an ACL packet is in fact an ISO packet

Because some transports don't have a dedicated type for ISO packets
(see 14202eff214e1e941fefa0366d4c3bc4b1a0d500) they may use ACL type
when in fact they are ISO packets.

In the past this was left for the driver to detect such thing but it
creates a problem when using the likes of btproxy when used by a VM as
the host would not be aware of the connection the guest is doing it
won't be able to detect such behavior, so this make bt_recv_frame
detect when it happens as it is the common interface to all drivers
including guest VMs.

Fixes: 14202eff214e ("Bluetooth: btusb: Detect if an ACL packet is in fact an ISO packet")
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoBluetooth: hci_sync: Resume adv with no RPA when active scan
Zhengping Jiang [Wed, 22 Feb 2023 00:17:56 +0000 (16:17 -0800)]
Bluetooth: hci_sync: Resume adv with no RPA when active scan

The address resolution should be disabled during the active scan,
so all the advertisements can reach the host. The advertising
has to be paused before disabling the address resolution,
because the advertising will prevent any changes to the resolving
list and the address resolution status. Skipping this will cause
the hci error and the discovery failure.

According to the bluetooth specification:
"7.8.44 LE Set Address Resolution Enable command

This command shall not be used when:
- Advertising (other than periodic advertising) is enabled,
- Scanning is enabled, or
- an HCI_LE_Create_Connection, HCI_LE_Extended_Create_Connection, or
  HCI_LE_Periodic_Advertising_Create_Sync command is outstanding."

If the host is using RPA, the controller needs to generate RPA for
the advertising, so the advertising must remain paused during the
active scan.

If the host is not using RPA, the advertising can be resumed after
disabling the address resolution.

Fixes: 9afc675edeeb ("Bluetooth: hci_sync: allow advertise when scan without RPA")
Signed-off-by: Zhengping Jiang <jiangzp@google.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
21 months agoMerge tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm
Linus Torvalds [Wed, 22 Mar 2023 21:15:05 +0000 (14:15 -0700)]
Merge tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm

Pull ARM fix from Russell King:
 "Just one fix for now to eliminate a KASAN false positive"

* tag 'for-linus' of git://git.armlinux.org.uk/~rmk/linux-arm:
  ARM: 9290/1: uaccess: Fix KASAN false-positives

21 months agoMerge tag 'bootconfig-fixes-v6.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Wed, 22 Mar 2023 21:09:20 +0000 (14:09 -0700)]
Merge tag 'bootconfig-fixes-v6.3-rc3' of git://git./linux/kernel/git/trace/linux-trace

Pull bootconfig fixes from Masami Hiramatsu:

 - Fix bootconfig test script to test increased maximum number (8192)
   node correctly

 - Change the console message if there is no bootconfig data and the
   kernel is compiled with CONFIG_BOOT_CONFIG_FORCE=y

* tag 'bootconfig-fixes-v6.3-rc3' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  bootconfig: Change message if no bootconfig with CONFIG_BOOT_CONFIG_FORCE=y
  bootconfig: Fix testcase to increase max node

21 months agoxdp: bpf_xdp_metadata use EOPNOTSUPP for no driver support
Jesper Dangaard Brouer [Tue, 21 Mar 2023 13:52:31 +0000 (14:52 +0100)]
xdp: bpf_xdp_metadata use EOPNOTSUPP for no driver support

When driver doesn't implement a bpf_xdp_metadata kfunc the fallback
implementation returns EOPNOTSUPP, which indicate device driver doesn't
implement this kfunc.

Currently many drivers also return EOPNOTSUPP when the hint isn't
available, which is ambiguous from an API point of view. Instead
change drivers to return ENODATA in these cases.

There can be natural cases why a driver doesn't provide any hardware
info for a specific hint, even on a frame to frame basis (e.g. PTP).
Lets keep these cases as separate return codes.

When describing the return values, adjust the function kernel-doc layout
to get proper rendering for the return values.

Fixes: ab46182d0dcb ("net/mlx4_en: Support RX XDP metadata")
Fixes: bc8d405b1ba9 ("net/mlx5e: Support RX XDP metadata")
Fixes: 306531f0249f ("veth: Support RX XDP metadata")
Fixes: 3d76a4d3d4e5 ("bpf: XDP metadata RX kfuncs")
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Acked-by: Stanislav Fomichev <sdf@google.com>
Acked-by: Toke Høiland-Jørgensen <toke@redhat.com>
Acked-by: Tariq Toukan <tariqt@nvidia.com>
Link: https://lore.kernel.org/r/167940675120.2718408.8176058626864184420.stgit@firesoul
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
21 months agomlxsw: spectrum_fid: Fix incorrect local port type
Ido Schimmel [Tue, 21 Mar 2023 11:42:00 +0000 (12:42 +0100)]
mlxsw: spectrum_fid: Fix incorrect local port type

Local port is a 10-bit number, but it was mistakenly stored in a u8,
resulting in firmware errors when using a netdev corresponding to a
local port higher than 255.

Fix by storing the local port in u16, as is done in the rest of the
code.

Fixes: bf73904f5fba ("mlxsw: Add support for 802.1Q FID family")
Signed-off-by: Ido Schimmel <idosch@nvidia.com>
Reviewed-by: Danielle Ratson <danieller@nvidia.com>
Signed-off-by: Petr Machata <petrm@nvidia.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Link: https://lore.kernel.org/r/eace1f9d96545ab8a2775db857cb7e291a9b166b.1679398549.git.petrm@nvidia.com
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
21 months agobootconfig: Change message if no bootconfig with CONFIG_BOOT_CONFIG_FORCE=y
Masami Hiramatsu (Google) [Tue, 28 Feb 2023 01:01:42 +0000 (10:01 +0900)]
bootconfig: Change message if no bootconfig with CONFIG_BOOT_CONFIG_FORCE=y

Change no bootconfig data error message if user do not specify 'bootconfig'
option but CONFIG_BOOT_CONFIG_FORCE=y.
With CONFIG_BOOT_CONFIG_FORCE=y, the kernel proceeds bootconfig check even
if user does not specify 'bootconfig' option. So the current error message
is confusing. Let's show just an information message to notice skipping
the bootconfig in that case.

Link: https://lore.kernel.org/all/167754610254.318944.16848412476667893329.stgit@devnote2/
Fixes: b743852ccc1d ("Allow forcing unconditional bootconfig processing")
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Link: https://lore.kernel.org/all/CAMuHMdV9jJvE2y8gY5V_CxidUikCf5515QMZHzTA3rRGEOj6=w@mail.gmail.com/
Suggested-by: Paul E. McKenney <paulmck@kernel.org>
Signed-off-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Tested-by: Paul E. McKenney <paulmck@kernel.org>
Acked-by: Mukesh Ojha <quic_mojha@quicinc.com>
21 months agowifi: mac80211: fix mesh path discovery based on unicast packets
Felix Fietkau [Tue, 14 Mar 2023 09:59:52 +0000 (10:59 +0100)]
wifi: mac80211: fix mesh path discovery based on unicast packets

If a packet has reached its intended destination, it was bumped to the code
that accepts it, without first checking if a mesh_path needs to be created
based on the discovered source.
Fix this by moving the destination address check further down.

Cc: stable@vger.kernel.org
Fixes: 986e43b19ae9 ("wifi: mac80211: fix receiving A-MSDU frames on mesh interfaces")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230314095956.62085-3-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
21 months agowifi: mac80211: fix qos on mesh interfaces
Felix Fietkau [Tue, 14 Mar 2023 09:59:50 +0000 (10:59 +0100)]
wifi: mac80211: fix qos on mesh interfaces

When ieee80211_select_queue is called for mesh, the sta pointer is usually
NULL, since the nexthop is looked up much later in the tx path.
Explicitly check for unicast address in that case in order to make qos work
again.

Cc: stable@vger.kernel.org
Fixes: 50e2ab392919 ("wifi: mac80211: fix queue selection for mesh/OCB interfaces")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20230314095956.62085-1-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
21 months agowifi: iwlwifi: mvm: protect TXQ list manipulation
Johannes Berg [Fri, 17 Mar 2023 09:53:25 +0000 (10:53 +0100)]
wifi: iwlwifi: mvm: protect TXQ list manipulation

Some recent upstream debugging uncovered the fact that in
iwlwifi, the TXQ list manipulation is racy.

Introduce a new state bit for when the TXQ is completely
ready and can be used without locking, and if that's not
set yet acquire the lock to check everything correctly.

Reviewed-by: Benjamin Berg <benjamin.berg@intel.com>
Tested-by: Jose Ignacio Tornos Martinez <jtornosm@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
21 months agowifi: iwlwifi: mvm: fix mvmtxq->stopped handling
Johannes Berg [Fri, 17 Mar 2023 09:53:24 +0000 (10:53 +0100)]
wifi: iwlwifi: mvm: fix mvmtxq->stopped handling

This could race if the queue is redirected while full, then
the flushing internally would start it while it's not yet
usable again. Fix it by using two state bits instead of just
one.

Reviewed-by: Benjamin Berg <benjamin.berg@intel.com>
Tested-by: Jose Ignacio Tornos Martinez <jtornosm@redhat.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
21 months agonet/sonic: use dma_mapping_error() for error check
Zhang Changzhong [Tue, 21 Mar 2023 03:45:43 +0000 (14:45 +1100)]
net/sonic: use dma_mapping_error() for error check

The DMA address returned by dma_map_single() should be checked with
dma_mapping_error(). Fix it accordingly.

Fixes: efcce839360f ("[PATCH] macsonic/jazzsonic network drivers update")
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Tested-by: Stan Johnson <userm57@yahoo.com>
Signed-off-by: Finn Thain <fthain@linux-m68k.org>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Link: https://lore.kernel.org/r/6645a4b5c1e364312103f48b7b36783b94e197a2.1679370343.git.fthain@linux-m68k.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agoMerge branch 'fix-trainwreck-with-ocelot-switch-statistics-counters'
Jakub Kicinski [Wed, 22 Mar 2023 04:27:12 +0000 (21:27 -0700)]
Merge branch 'fix-trainwreck-with-ocelot-switch-statistics-counters'

Vladimir Oltean says:

====================
Fix trainwreck with Ocelot switch statistics counters

While testing the patch set for preemptible traffic classes with some
controlled traffic and measuring counter deltas:
https://lore.kernel.org/netdev/20230220122343.1156614-1-vladimir.oltean@nxp.com/

I noticed that in the output of "ethtool -S swp0 --groups eth-mac
eth-phy eth-ctrl rmon -- --src emac | grep -v ': 0'", the TX counters
were off. Quickly I realized that their values were permutated by 1
compared to their names, and that for example
tx-rmon-etherStatsPkts64to64Octets was incrementing when
tx-rmon-etherStatsPkts65to127Octets should have.

Initially I suspected something having to do with the bulk reading
logic, and indeed I found a bug there (fixed as 1/3), but that was not
the source of the problems. Instead it revealed other problems.

While dumping the regions created by the driver on my switch, I figured
out that it sees a discontinuity which shouldn't have existed between
reg 0x278 and reg 0x280.

Discontinuity between last reg 0x0 and new reg 0x0, creating new region
Discontinuity between last reg 0x108 and new reg 0x200, creating new region
Discontinuity between last reg 0x278 and new reg 0x280, creating new region
Discontinuity between last reg 0x2b0 and new reg 0x400, creating new region
region of 67 contiguous counters starting with SYS:STAT:CNT[0x000]
region of 31 contiguous counters starting with SYS:STAT:CNT[0x080]
region of 13 contiguous counters starting with SYS:STAT:CNT[0x0a0]
region of 18 contiguous counters starting with SYS:STAT:CNT[0x100]

That is where TX_MM_HOLD should have been, and that was the bug, since
it was missing. After adding it, the regions look like this and the
off-by-one issue is resolved:

Discontinuity between last reg 0x000000 and new reg 0x000000, creating new region
Discontinuity between last reg 0x000108 and new reg 0x000200, creating new region
Discontinuity between last reg 0x0002b0 and new reg 0x000400, creating new region
region of 67 contiguous counters starting with SYS:STAT:CNT[0x000]
region of 45 contiguous counters starting with SYS:STAT:CNT[0x080]
region of 18 contiguous counters starting with SYS:STAT:CNT[0x100]

However, as I am thinking out loud, it should have not reported the
other counters as off by one even when skipping TX_MM_HOLD... after all,
on Ocelot/Seville, there are more counters which need to be skipped.

Which is when I investigated and noticed the bug solved in 2/3.
I've validated that both on native VSC9959 (which uses
ocelot_mm_stats_layout) as well as by faking the other switches by
making VSC9959 use the plain ocelot_stats_layout.

To summarize: on all Ocelot switches, the TX counters and drop counters
are completely broken. The RX counters are mostly fine.

With this occasion, I have collected more cleanup patches in this area,
which I'm going to submit after the net -> net-next merge.
====================

Link: https://lore.kernel.org/r/20230321010325.897817-1-vladimir.oltean@nxp.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agonet: mscc: ocelot: add TX_MM_HOLD to ocelot_mm_stats_layout
Vladimir Oltean [Tue, 21 Mar 2023 01:03:25 +0000 (03:03 +0200)]
net: mscc: ocelot: add TX_MM_HOLD to ocelot_mm_stats_layout

The lack of a definition for this counter is what initially prompted me
to investigate a problem which really manifested itself as the previous
change, "net: mscc: ocelot: fix transfer from region->buf to ocelot->stats".

When TX_MM_HOLD is defined in enum ocelot_stat but not in struct
ocelot_stat_layout ocelot_mm_stats_layout, this creates a hole, which
due to the aforementioned bug, makes all counters following TX_MM_HOLD
be recorded off by one compared to their correct position. So for
example, a non-zero TX_PMAC_OCTETS would be reported as TX_MERGE_FRAGMENTS,
TX_PMAC_UNICAST would be reported as TX_PMAC_OCTETS, TX_PMAC_64 would be
reported as TX_PMAC_PAUSE, etc etc. This is because the size of the hole
(1) is much smaller than the size of the region, so the phenomenon where
the stats are off-by-one, rather than lost, prevails.

However, the phenomenon where stats are lost can be seen too, for
example with DROP_LOCAL, which is at the beginning of its own region
(offset 0x000400 vs the previous 0x0002b0 constitutes a discontinuity).
This is also reported as off by one and saved to TX_PMAC_1527_MAX, but
that counter is not reported to the unstructured "ethtool -S", as
opposed to DROP_LOCAL which is (as "drop_local").

Fixes: ab3f97a9610a ("net: mscc: ocelot: export ethtool MAC Merge stats for Felix VSC9959")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
21 months agonet: mscc: ocelot: fix transfer from region->buf to ocelot->stats
Vladimir Oltean [Tue, 21 Mar 2023 01:03:24 +0000 (03:03 +0200)]
net: mscc: ocelot: fix transfer from region->buf to ocelot->stats

To understand the problem, we need some definitions.

The driver is aware of multiple counters (enum ocelot_stat), yet not all
switches supported by the driver implement all counters. There are 2
statistics layouts: ocelot_stats_layout and ocelot_mm_stats_layout, the
latter having 36 counters more than the former.

ocelot->stats[] is not a compact array, i.e. there are elements within
it which are not going to be populated for ocelot_stats_layout. On the
other hand, ocelot->stats[] is easily indexable, for example "tx_octets"
for port 3 can be found at ocelot->stats[3 * OCELOT_NUM_STATS +
OCELOT_STAT_TX_OCTETS], and that is why we keep it sparse.

Regions, as created by ocelot_prepare_stats_regions(), are compact
(every element from region->buf will correspond to a counter that is
present in this switch's layout) but are not easily indexable.

Let's define holes as the ranges of values of enum ocelot_stat for which
ocelot_stats_layout doesn't have a "reg" defined. For example, there is
a hole between OCELOT_STAT_RX_GREEN_PRIO_7 and OCELOT_STAT_TX_OCTETS
which is of 23 elements that are only present on ocelot_mm_stats_layout,
and as such, they are also present in enum ocelot_stat. Let's define the
left extremity of the hole - the last enum ocelot_stat still defined -
as A (in this case OCELOT_STAT_RX_GREEN_PRIO_7) and the right extremity -
the first enum ocelot_stat that is defined after a series of undefined
ones - as B (in this case OCELOT_STAT_TX_OCTETS).

There is a bug in the procedure which transfers stats from region->buf[]
to ocelot->stats[].

For each hole in the ocelot_stats_layout, the logic transfers the stats
starting with enum ocelot_stat B to ocelot->stats[] index A + 1. So all
stats after a hole are saved to a position which is off by B - A + 1
elements.

This causes 2 kinds of issues:
(a) counters which shouldn't increment increment
(b) counters which should increment don't

Holes in the ocelot_stat_layout automatically imply the end of a region
and the beginning of a new one; however the reverse is not necessarily
true. For example, for ocelot_mm_stat_layout, there could be multiple
regions (which indicate discontinuities in register addresses) while
there is no hole (which indicates discontinuities in enum ocelot_stat
values).

In the example above, the stats from the second region->buf[] are not
transferred to ocelot->stats starting with index
"port * OCELOT_NUM_STATS + OCELOT_STAT_TX_OCTETS" as they should, but
rather, starting with element
"port * OCELOT_NUM_STATS + OCELOT_STAT_RX_GREEN_PRIO_7 + 1".

That stats[] array element is not reported to user space for switches
that use ocelot_stat_layout, and that is how issue (b) occurs.

However, if the length of the second region is larger than the hole,
then some stats will start to be transferred to the ocelot->stats[]
indices which *are* reported to user space, but those indices contain
wrong values (corresponding to unexpected counters). This is how issue
(a) occurs.

The procedure, as it was introduced in commit d87b1c08f38a ("net: mscc:
ocelot: use bulk reads for stats"), was not buggy, because there were no
holes in the struct ocelot_stat_layout instances at that time. The
problem is that when those holes were introduced, the function was not
updated to take them into consideration.

To update the procedure, we need to know, for each region, which enum
ocelot_stat corresponds to its region->base. We have no way of deducing
that based on the contents of struct ocelot_stats_region, so we need to
add this information.

Fixes: ab3f97a9610a ("net: mscc: ocelot: export ethtool MAC Merge stats for Felix VSC9959")
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>