platform/kernel/linux-rpi.git
3 years agonet: ftgmac100: Fix crash when removing driver
Joel Stanley [Tue, 17 Nov 2020 02:44:48 +0000 (13:14 +1030)]
net: ftgmac100: Fix crash when removing driver

When removing the driver we would hit BUG_ON(!list_empty(&dev->ptype_specific))
in net/core/dev.c due to still having the NC-SI packet handler
registered.

 # echo 1e660000.ethernet > /sys/bus/platform/drivers/ftgmac100/unbind
  ------------[ cut here ]------------
  kernel BUG at net/core/dev.c:10254!
  Internal error: Oops - BUG: 0 [#1] SMP ARM
  CPU: 0 PID: 115 Comm: sh Not tainted 5.10.0-rc3-next-20201111-00007-g02e0365710c4 #46
  Hardware name: Generic DT based system
  PC is at netdev_run_todo+0x314/0x394
  LR is at cpumask_next+0x20/0x24
  pc : [<806f5830>]    lr : [<80863cb0>]    psr: 80000153
  sp : 855bbd58  ip : 00000001  fp : 855bbdac
  r10: 80c03d00  r9 : 80c06228  r8 : 81158c54
  r7 : 00000000  r6 : 80c05dec  r5 : 80c05d18  r4 : 813b9280
  r3 : 813b9054  r2 : 8122c470  r1 : 00000002  r0 : 00000002
  Flags: Nzcv  IRQs on  FIQs off  Mode SVC_32  ISA ARM  Segment none
  Control: 00c5387d  Table: 85514008  DAC: 00000051
  Process sh (pid: 115, stack limit = 0x7cb5703d)
 ...
  Backtrace:
  [<806f551c>] (netdev_run_todo) from [<80707eec>] (rtnl_unlock+0x18/0x1c)
   r10:00000051 r9:854ed710 r8:81158c54 r7:80c76bb0 r6:81158c10 r5:8115b410
   r4:813b9000
  [<80707ed4>] (rtnl_unlock) from [<806f5db8>] (unregister_netdev+0x2c/0x30)
  [<806f5d8c>] (unregister_netdev) from [<805a8180>] (ftgmac100_remove+0x20/0xa8)
   r5:8115b410 r4:813b9000
  [<805a8160>] (ftgmac100_remove) from [<805355e4>] (platform_drv_remove+0x34/0x4c)

Fixes: bd466c3fb5a4 ("net/faraday: Support NCSI mode")
Signed-off-by: Joel Stanley <joel@jms.id.au>
Link: https://lore.kernel.org/r/20201117024448.1170761-1-joel@jms.id.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: b44: fix error return code in b44_init_one()
Zhang Changzhong [Tue, 17 Nov 2020 03:02:11 +0000 (11:02 +0800)]
net: b44: fix error return code in b44_init_one()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 39a6f4bce6b4 ("b44: replace the ssb_dma API with the generic DMA API")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Reviewed-by: Michael Chan <michael.chan@broadcom.com>
Link: https://lore.kernel.org/r/1605582131-36735-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoqed: fix error return code in qed_iwarp_ll2_start()
Zhang Changzhong [Mon, 16 Nov 2020 13:07:13 +0000 (21:07 +0800)]
qed: fix error return code in qed_iwarp_ll2_start()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 469981b17a4f ("qed: Add unaligned and packed packet processing")
Fixes: fcb39f6c10b2 ("qed: Add mpa buffer descriptors for storing and processing mpa fpdus")
Fixes: 1e28eaad07ea ("qed: Add iWARP support for fpdu spanned over more than two tcp packets")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Acked-by: Michal KalderonĀ <michal.kalderon@marvell.com>
Link: https://lore.kernel.org/r/1605532033-27373-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'bnxt_en-bug-fixes'
Jakub Kicinski [Tue, 17 Nov 2020 01:39:49 +0000 (17:39 -0800)]
Merge branch 'bnxt_en-bug-fixes'

Michael Chan says:

====================
bnxt_en: Bug fixes.

This first patch fixes a module eeprom A2h addressing issue.  The next
2 patches fix counter related issues.  The last one skips an
unsupported firmware call on the VF to avoid the error log.
====================

Link: https://lore.kernel.org/r/1605486472-28156-1-git-send-email-michael.chan@broadcom.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: Avoid unnecessary NVM_GET_DEV_INFO cmd error log on VFs.
Vasundhara Volam [Mon, 16 Nov 2020 00:27:52 +0000 (19:27 -0500)]
bnxt_en: Avoid unnecessary NVM_GET_DEV_INFO cmd error log on VFs.

VFs do not have access permissions to issue NVM_GET_DEV_INFO
firmware command.

Fixes: 4933f6753b50 ("bnxt_en: Add bnxt_hwrm_nvm_get_dev_info() to query NVM info.")
Signed-off-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: Fix counter overflow logic.
Michael Chan [Mon, 16 Nov 2020 00:27:51 +0000 (19:27 -0500)]
bnxt_en: Fix counter overflow logic.

bnxt_add_one_ctr() adds a hardware counter to a software counter and
adjusts for the hardware counter wraparound against the mask.  The logic
assumes that the hardware counter is always smaller than or equal to
the mask.

This assumption is mostly correct.  But in some cases if the firmware
is older and does not provide the accurate mask, the driver can use
a mask that is smaller than the actual hardware mask.  This can cause
some extra carry bits to be added to the software counter, resulting in
counters that far exceed the actual value.  Fix it by masking the
hardware counter with the mask passed into bnxt_add_one_ctr().

Fixes: fea6b3335527 ("bnxt_en: Accumulate all counters.")
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Reviewed-by: Pavan Chebbi <pavan.chebbi@broadcom.com>
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: Free port stats during firmware reset.
Michael Chan [Mon, 16 Nov 2020 00:27:50 +0000 (19:27 -0500)]
bnxt_en: Free port stats during firmware reset.

Firmware is unable to retain the port counters during any kind of
fatal or non-fatal resets, so we must clear the port counters to
avoid false detection of port counter overflow.

Fixes: fea6b3335527 ("bnxt_en: Accumulate all counters.")
Reviewed-by: Edwin Peer <edwin.peer@broadcom.com>
Reviewed-by: Vasundhara Volam <vasundhara-v.volam@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agobnxt_en: read EEPROM A2h address using page 0
Edwin Peer [Mon, 16 Nov 2020 00:27:49 +0000 (19:27 -0500)]
bnxt_en: read EEPROM A2h address using page 0

The module eeprom address range returned by bnxt_get_module_eeprom()
should be 256 bytes of A0h address space, the lower half of the A2h
address space, and page 0 for the upper half of the A2h address space.

Fix the firmware call by passing page_number 0 for the A2h slave address
space.

Fixes: 42ee18fe4ca2 ("bnxt_en: Add Support for ETHTOOL_GMODULEINFO and ETHTOOL_GMODULEEEPRO")
Signed-off-by: Edwin Peer <edwin.peer@broadcom.com>
Signed-off-by: Michael Chan <michael.chan@broadcom.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ipa: lock when freeing transaction
Alex Elder [Sat, 14 Nov 2020 18:20:17 +0000 (12:20 -0600)]
net: ipa: lock when freeing transaction

Transactions sit on one of several lists, depending on their state
(allocated, pending, complete, or polled).  A spinlock protects
against concurrent access when transactions are moved between these
lists.

Transactions are also reference counted.  A newly-allocated
transaction has an initial count of 1; a transaction is released in
gsi_trans_free() only if its decremented reference count reaches 0.
Releasing a transaction includes removing it from the polled (or if
unused, allocated) list, so the spinlock is acquired when we release
a transaction.

The reference count is used to allow a caller to synchronously wait
for a committed transaction to complete.  In this case, the waiter
takes an extra reference to the transaction *before* committing it
(so it won't be freed), and releases its reference (calls
gsi_trans_free()) when it is done with it.

Similarly, gsi_channel_update() takes an extra reference to ensure a
transaction isn't released before the function is done operating on
it.  Until the transaction is moved to the completed list (by this
function) it won't be freed, so this reference is taken "safely."

But in the quiesce path, we want to wait for the "last" transaction,
which we find in the completed or polled list.  Transactions on
these lists can be freed at any time, so we (try to) prevent that
by taking the reference while holding the spinlock.

Currently gsi_trans_free() decrements a transaction's reference
count unconditionally, acquiring the lock to remove the transaction
from its list *only* when the count reaches 0.  This does not
protect the quiesce path, which depends on the lock to ensure its
extra reference prevents release of the transaction.

Fix this by only dropping the last reference to a transaction
in gsi_trans_free() while holding the spinlock.

Fixes: 9dd441e4ed575 ("soc: qcom: ipa: GSI transactions")
Reported-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Alex Elder <elder@linaro.org>
Link: https://lore.kernel.org/r/20201114182017.28270-1-elder@linaro.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet/tls: fix corrupted data in recvmsg
Vadim Fedorenko [Sun, 15 Nov 2020 04:16:00 +0000 (07:16 +0300)]
net/tls: fix corrupted data in recvmsg

If tcp socket has more data than Encrypted Handshake Message then
tls_sw_recvmsg will try to decrypt next record instead of returning
full control message to userspace as mentioned in comment. The next
message - usually Application Data - gets corrupted because it uses
zero copy for decryption that's why the data is not stored in skb
for next iteration. Revert check to not decrypt next record if
current is not Application Data.

Fixes: 692d7b5d1f91 ("tls: Fix recvmsg() to be able to peek across multiple records")
Signed-off-by: Vadim Fedorenko <vfedorenko@novek.ru>
Link: https://lore.kernel.org/r/1605413760-21153-1-git-send-email-vfedorenko@novek.ru
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: qualcomm: rmnet: Fix incorrect receive packet handling during cleanup
Subash Abhinov Kasiviswanathan [Fri, 13 Nov 2020 20:12:05 +0000 (13:12 -0700)]
net: qualcomm: rmnet: Fix incorrect receive packet handling during cleanup

During rmnet unregistration, the real device rx_handler is first cleared
followed by the removal of rx_handler_data after the rcu synchronization.

Any packets in the receive path may observe that the rx_handler is NULL.
However, there is no check when dereferencing this value to use the
rmnet_port information.

This fixes following splat by adding the NULL check.

Unable to handle kernel NULL pointer dereference at virtual
address 000000000000000d
pc : rmnet_rx_handler+0x124/0x284
lr : rmnet_rx_handler+0x124/0x284
 rmnet_rx_handler+0x124/0x284
 __netif_receive_skb_core+0x758/0xd74
 __netif_receive_skb+0x50/0x17c
 process_backlog+0x15c/0x1b8
 napi_poll+0x88/0x284
 net_rx_action+0xbc/0x23c
 __do_softirq+0x20c/0x48c

Fixes: ceed73a2cf4a ("drivers: net: ethernet: qualcomm: rmnet: Initial implementation")
Signed-off-by: Sean Tranchetti <stranche@codeaurora.org>
Signed-off-by: Subash Abhinov Kasiviswanathan <subashab@codeaurora.org>
Link: https://lore.kernel.org/r/1605298325-3705-1-git-send-email-subashab@codeaurora.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: mvneta: fix possible memory leak in mvneta_swbm_add_rx_fragment
Lorenzo Bianconi [Fri, 13 Nov 2020 18:16:57 +0000 (19:16 +0100)]
net: mvneta: fix possible memory leak in mvneta_swbm_add_rx_fragment

Recycle the page running page_pool_put_full_page() in
mvneta_swbm_add_rx_fragment routine when the last descriptor
contains just the FCS or if the received packet contains more than
MAX_SKB_FRAGS fragments

Fixes: ca0e014609f0 ("net: mvneta: move skb build after descriptors processing")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Link: https://lore.kernel.org/r/df6a2bad70323ee58d3901491ada31c1ca2a40b9.1605291228.git.lorenzo@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: stmmac: Use rtnl_lock/unlock on netif_set_real_num_rx_queues() call
Wong Vee Khee [Sun, 15 Nov 2020 07:42:10 +0000 (15:42 +0800)]
net: stmmac: Use rtnl_lock/unlock on netif_set_real_num_rx_queues() call

Fix an issue where dump stack is printed on suspend resume flow due to
netif_set_real_num_rx_queues() is not called with rtnl_lock held().

Fixes: 686cff3d7022 ("net: stmmac: Fix incorrect location to set real_num_rx|tx_queues")
Reported-by: Christophe ROULLIER <christophe.roullier@st.com>
Tested-by: Christophe ROULLIER <christophe.roullier@st.com>
Cc: Alexandre TORGUE <alexandre.torgue@st.com>
Reviewed-by: Ong Boon Leong <boon.leong.ong@intel.com>
Signed-off-by: Wong Vee Khee <vee.khee.wong@intel.com>
Link: https://lore.kernel.org/r/20201115074210.23605-1-vee.khee.wong@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: bridge: add missing counters to ndo_get_stats64 callback
Heiner Kallweit [Fri, 13 Nov 2020 09:27:27 +0000 (10:27 +0100)]
net: bridge: add missing counters to ndo_get_stats64 callback

In br_forward.c and br_input.c fields dev->stats.tx_dropped and
dev->stats.multicast are populated, but they are ignored in
ndo_get_stats64.

Fixes: 28172739f0a2 ("net: fix 64 bit counters on 32 bit arches")
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Link: https://lore.kernel.org/r/58ea9963-77ad-a7cf-8dfd-fc95ab95f606@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: ethernet: ti: cpsw: fix error return code in cpsw_probe()
Zhang Changzhong [Fri, 13 Nov 2020 06:49:33 +0000 (14:49 +0800)]
net: ethernet: ti: cpsw: fix error return code in cpsw_probe()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 83a8471ba255 ("net: ethernet: ti: cpsw: refactor probe to group common hw initialization")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605250173-18438-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: stmmac: dwmac-intel-plat: fix error return code in intel_eth_plat_probe()
Zhang Changzhong [Fri, 13 Nov 2020 06:34:03 +0000 (14:34 +0800)]
net: stmmac: dwmac-intel-plat: fix error return code in intel_eth_plat_probe()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 9efc9b2b04c7 ("net: stmmac: Add dwmac-intel-plat for GBE driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605249243-17262-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoqlcnic: fix error return code in qlcnic_83xx_restart_hw()
Zhang Changzhong [Fri, 13 Nov 2020 06:16:26 +0000 (14:16 +0800)]
qlcnic: fix error return code in qlcnic_83xx_restart_hw()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 3ced0a88cd4c ("qlcnic: Add support to run firmware POST")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605248186-16013-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agocx82310_eth: fix error return code in cx82310_bind()
Zhang Changzhong [Fri, 13 Nov 2020 06:07:07 +0000 (14:07 +0800)]
cx82310_eth: fix error return code in cx82310_bind()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: ca139d76b0d9 ("cx82310_eth: re-enable ethernet mode after router reboot")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605247627-15385-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMAINTAINERS: update cxgb4 and cxgb3 maintainer
Raju Rangoju [Mon, 16 Nov 2020 10:43:22 +0000 (16:13 +0530)]
MAINTAINERS: update cxgb4 and cxgb3 maintainer

Update cxgb4 and cxgb3 driver maintainer

Signed-off-by: Raju Rangoju <rajur@chelsio.com>
Link: https://lore.kernel.org/r/20201116104322.3959-1-rajur@chelsio.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: lantiq: Wait for the GPHY firmware to be ready
Martin Blumenstingl [Sun, 15 Nov 2020 16:57:57 +0000 (17:57 +0100)]
net: lantiq: Wait for the GPHY firmware to be ready

A user reports (slightly shortened from the original message):
  libphy: lantiq,xrx200-mdio: probed
  mdio_bus 1e108000.switch-mii: MDIO device at address 17 is missing.
  gswip 1e108000.switch lan: no phy at 2
  gswip 1e108000.switch lan: failed to connect to port 2: -19
  lantiq,xrx200-net 1e10b308.eth eth0: error -19 setting up slave phy

This is a single-port board using the internal Fast Ethernet PHY. The
user reports that switching to PHY scanning instead of configuring the
PHY within device-tree works around this issue.

The documentation for the standalone variant of the PHY11G (which is
probably very similar to what is used inside the xRX200 SoCs but having
the firmware burnt onto that standalone chip in the factory) states that
the PHY needs 300ms to be ready for MDIO communication after releasing
the reset.

Add a 300ms delay after initializing all GPHYs to ensure that the GPHY
firmware had enough time to initialize and to appear on the MDIO bus.
Unfortunately there is no (known) documentation on what the minimum time
to wait after releasing the reset on an internal PHY so play safe and
take the one for the external variant. Only wait after the last GPHY
firmware is loaded to not slow down the initialization too much (
xRX200 has two GPHYs but newer SoCs have at least three GPHYs).

Fixes: 14fceff4771e51 ("net: dsa: Add Lantiq / Intel DSA driver for vrx200")
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Acked-by: Hauke Mehrtens <hauke@hauke-m.de>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20201115165757.552641-1-martin.blumenstingl@googlemail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMAINTAINERS: Add Martin Schiller as a maintainer for the X.25 stack
Xie He [Sat, 14 Nov 2020 11:10:29 +0000 (03:10 -0800)]
MAINTAINERS: Add Martin Schiller as a maintainer for the X.25 stack

Martin Schiller is an active developer and reviewer for the X.25 code.
His company is providing products based on the Linux X.25 stack.
So he is a good candidate for maintainers of the X.25 code.

The original maintainer of the X.25 network layer (Andrew Hendry) has
not sent any email to the netdev mail list since 2013. So he is probably
inactive now.

Cc: Andrew Hendry <andrew.hendry@gmail.com>
Signed-off-by: Xie He <xie.he.0141@gmail.com>
Acked-by: Martin Schiller <ms@dev.tdt.de>
Link: https://lore.kernel.org/r/20201114111029.326972-1-xie.he.0141@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoipv6/netfilter: Discard first fragment not including all headers
Georg Kohmann [Wed, 11 Nov 2020 11:50:25 +0000 (12:50 +0100)]
ipv6/netfilter: Discard first fragment not including all headers

Packets are processed even though the first fragment don't include all
headers through the upper layer header. This breaks TAHI IPv6 Core
Conformance Test v6LC.1.3.6.

Referring to RFC8200 SECTION 4.5: "If the first fragment does not include
all headers through an Upper-Layer header, then that fragment should be
discarded and an ICMP Parameter Problem, Code 3, message should be sent to
the source of the fragment, with the Pointer field set to zero."

The fragment needs to be validated the same way it is done in
commit 2efdaaaf883a ("IPv6: reply ICMP error if the first fragment don't
include all headers") for ipv6. Wrap the validation into a common function,
ipv6_frag_thdr_truncated() to check for truncation in the upper layer
header. This validation does not fullfill all aspects of RFC 8200,
section 4.5, but is at the moment sufficient to pass mentioned TAHI test.

In netfilter, utilize the fragment offset returned by find_prev_fhdr() to
let ipv6_frag_thdr_truncated() start it's traverse from the fragment
header.

Return 0 to drop the fragment in the netfilter. This is the same behaviour
as used on other protocol errors in this function, e.g. when
nf_ct_frag6_queue() returns -EPROTO. The Fragment will later be picked up
by ipv6_frag_rcv() in reassembly.c. ipv6_frag_rcv() will then send an
appropriate ICMP Parameter Problem message back to the source.

References commit 2efdaaaf883a ("IPv6: reply ICMP error if the first
fragment don't include all headers")

Signed-off-by: Georg Kohmann <geokohma@cisco.com>
Acked-by: Pablo Neira Ayuso <pablo@netfilter.org>
Link: https://lore.kernel.org/r/20201111115025.28879-1-geokohma@cisco.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge branch 'fix-usage-counter-leak-by-adding-a-general-sync-ops'
Jakub Kicinski [Mon, 16 Nov 2020 17:37:13 +0000 (09:37 -0800)]
Merge branch 'fix-usage-counter-leak-by-adding-a-general-sync-ops'

Zhang Qilong says:

====================
Fix usage counter leak by adding a general sync ops

In many case, we need to check return value of pm_runtime_get_sync,
but it brings a trouble to the usage counter processing. Many callers
forget to decrease the usage counter when it failed, which could
resulted in reference leak. It has been discussed a lot[0][1]. So we
add a function to deal with the usage counter for better coding and
view. Then, we replace pm_runtime_resume_and_get with it in fec_main.c
to avoid it.

[0] https://lkml.org/lkml/2020/6/14/88
[1] https://patchwork.ozlabs.org/project/linux-tegra/list/?series=178139
====================

Link: https://lore.kernel.org/r/20201110092933.3342784-1-zhangqilong3@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agonet: fec: Fix reference count leak in fec series ops
Zhang Qilong [Tue, 10 Nov 2020 09:29:33 +0000 (17:29 +0800)]
net: fec: Fix reference count leak in fec series ops

pm_runtime_get_sync() will increment pm usage at first and it will
resume the device later. If runtime of the device has error or
device is in inaccessible state(or other error state), resume
operation will fail. If we do not call put operation to decrease
the reference, it will result in reference count leak. Moreover,
this device cannot enter the idle state and always stay busy or other
non-idle state later. So we fixed it by replacing it with
pm_runtime_resume_and_get.

Fixes: 8fff755e9f8d0 ("net: fec: Ensure clocks are enabled while using mdio bus")
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoPM: runtime: Add pm_runtime_resume_and_get to deal with usage counter
Zhang Qilong [Tue, 10 Nov 2020 09:29:32 +0000 (17:29 +0800)]
PM: runtime: Add pm_runtime_resume_and_get to deal with usage counter

In many case, we need to check return value of pm_runtime_get_sync, but
it brings a trouble to the usage counter processing. Many callers forget
to decrease the usage counter when it failed, which could resulted in
reference leak. It has been discussed a lot[0][1]. So we add a function
to deal with the usage counter for better coding.

[0]https://lkml.org/lkml/2020/6/14/88
[1]https://patchwork.ozlabs.org/project/linux-tegra/list/?series=178139
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
3 years agoMerge tag 'linux-can-fixes-for-5.10-20201115' of git://git.kernel.org/pub/scm/linux...
Jakub Kicinski [Mon, 16 Nov 2020 15:34:30 +0000 (07:34 -0800)]
Merge tag 'linux-can-fixes-for-5.10-20201115' of git://git./linux/kernel/git/mkl/linux-can

Marc Kleine-Budde says:

====================
pull-request: can 2020-11-15

Anant Thazhemadam contributed two patches for the AF_CAN that prevent potential
access of uninitialized member in can_rcv() and canfd_rcv().

The next patch is by Alejandro Concepcion Rodriguez and changes can_restart()
to use the correct function to push a skb into the networking stack from
process context.

Zhang Qilong's patch fixes a memory leak in the error path of the ti_hecc's
probe function.

A patch by me fixes mcba_usb_start_xmit() function in the mcba_usb driver, to
first fill the skb and then pass it to can_put_echo_skb().

Colin Ian King's patch fixes a potential integer overflow on shift in the
peak_usb driver.

The next two patches target the flexcan driver, a patch by me adds the missing
"req_bit" to the stop mode property comment (which was broken during net-next
for v5.10). Zhang Qilong's patch fixes the failure handling of
pm_runtime_get_sync().

The next seven patches target the m_can driver including the tcan4x5x spi
driver glue code. Enric Balletbo i Serra's patch for the tcan4x5x Kconfig fix
the REGMAP_SPI dependency handling. A patch by me for the tcan4x5x driver's
probe() function adds missing error handling to for devm_regmap_init(), and in
tcan4x5x_can_remove() the order of deregistration is fixed. Wu Bo's patch for
the m_can driver fixes the state change handling in
m_can_handle_state_change(). Two patches by Dan Murphy first introduce
m_can_class_free_dev() and then make use of it to fix the freeing of the can
device. A patch by Faiz Abbas add a missing shutdown of the CAN controller in
the m_can_stop() function.

* tag 'linux-can-fixes-for-5.10-20201115' of git://git.kernel.org/pub/scm/linux/kernel/git/mkl/linux-can:
  can: m_can: m_can_stop(): set device to software init mode before closing
  can: m_can: Fix freeing of can device from peripherials
  can: m_can: m_can_class_free_dev(): introduce new function
  can: m_can: m_can_handle_state_change(): fix state change
  can: tcan4x5x: tcan4x5x_can_remove(): fix order of deregistration
  can: tcan4x5x: tcan4x5x_can_probe(): add missing error checking for devm_regmap_init()
  can: tcan4x5x: replace depends on REGMAP_SPI with depends on SPI
  can: flexcan: fix failure handling of pm_runtime_get_sync()
  can: flexcan: flexcan_setup_stop_mode(): add missing "req_bit" to stop mode property comment
  can: peak_usb: fix potential integer overflow on shift of a int
  can: mcba_usb: mcba_usb_start_xmit(): first fill skb, then pass to can_put_echo_skb()
  can: ti_hecc: Fix memleak in ti_hecc_probe
  can: dev: can_restart(): post buffer from the right context
  can: af_can: prevent potential access of uninitialized member in canfd_rcv()
  can: af_can: prevent potential access of uninitialized member in can_rcv()
====================

Link: https://lore.kernel.org/r/20201115174131.2089251-1-mkl@pengutronix.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agocan: m_can: m_can_stop(): set device to software init mode before closing
Faiz Abbas [Tue, 25 Aug 2020 05:54:42 +0000 (11:24 +0530)]
can: m_can: m_can_stop(): set device to software init mode before closing

There might be some requests pending in the buffer when the interface close
sequence occurs. In some devices, these pending requests might lead to the
module not shutting down properly when m_can_clk_stop() is called.

Therefore, move the device to init state before potentially powering it down.

Fixes: e0d1f4816f2a ("can: m_can: add Bosch M_CAN controller support")
Signed-off-by: Faiz Abbas <faiz_abbas@ti.com>
Acked-by: Dan Murphy <dmurphy@ti.com>
Link: https://lore.kernel.org/r/20200825055442.16994-1-faiz_abbas@ti.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: m_can: Fix freeing of can device from peripherials
Dan Murphy [Thu, 27 Feb 2020 18:38:29 +0000 (12:38 -0600)]
can: m_can: Fix freeing of can device from peripherials

Fix leaking netdev device from peripherial devices. The call to allocate the
netdev device is made from and managed by the peripherial.

Fixes: f524f829b75a ("can: m_can: Create a m_can platform framework")
Reported-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Link: http://lore.kernel.org/r/20200227183829.21854-2-dmurphy@ti.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: m_can: m_can_class_free_dev(): introduce new function
Dan Murphy [Thu, 27 Feb 2020 18:38:29 +0000 (12:38 -0600)]
can: m_can: m_can_class_free_dev(): introduce new function

This patch creates a common function that peripherials can call to free the
netdev device when failures occur.

Fixes: f524f829b75a ("can: m_can: Create a m_can platform framework")
Reported-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Dan Murphy <dmurphy@ti.com>
Link: http://lore.kernel.org/r/20200227183829.21854-2-dmurphy@ti.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: m_can: m_can_handle_state_change(): fix state change
Wu Bo [Wed, 29 Jan 2020 02:23:30 +0000 (10:23 +0800)]
can: m_can: m_can_handle_state_change(): fix state change

m_can_handle_state_change() is called with the new_state as an argument.

In the switch statements for CAN_STATE_ERROR_ACTIVE, the comment and the
following code indicate that a CAN_STATE_ERROR_WARNING is handled.

This patch fixes this problem by changing the case to CAN_STATE_ERROR_WARNING.

Signed-off-by: Wu Bo <wubo.oduw@gmail.com>
Link: http://lore.kernel.org/r/20200129022330.21248-2-wubo.oduw@gmail.com
Cc: Dan Murphy <dmurphy@ti.com>
Fixes: e0d1f4816f2a ("can: m_can: add Bosch M_CAN controller support")
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: tcan4x5x: tcan4x5x_can_remove(): fix order of deregistration
Marc Kleine-Budde [Mon, 10 Aug 2020 20:23:49 +0000 (22:23 +0200)]
can: tcan4x5x: tcan4x5x_can_remove(): fix order of deregistration

Change the order in tcan4x5x_can_remove() to be the exact inverse of
tcan4x5x_can_probe(). First m_can_class_unregister(), then power down the
device.

Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Cc: Dan Murphy <dmurphy@ti.com>
Link: http://lore.kernel.org/r/20201019154233.1262589-10-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: tcan4x5x: tcan4x5x_can_probe(): add missing error checking for devm_regmap_init()
Marc Kleine-Budde [Fri, 3 Jan 2020 10:30:34 +0000 (11:30 +0100)]
can: tcan4x5x: tcan4x5x_can_probe(): add missing error checking for devm_regmap_init()

This patch adds the missing error checking when initializing the regmap
interface fails.

Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Cc: Dan Murphy <dmurphy@ti.com>
Link: http://lore.kernel.org/r/20201019154233.1262589-7-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: tcan4x5x: replace depends on REGMAP_SPI with depends on SPI
Enric Balletbo i Serra [Mon, 13 Apr 2020 14:10:13 +0000 (16:10 +0200)]
can: tcan4x5x: replace depends on REGMAP_SPI with depends on SPI

regmap is a library function that gets selected by drivers that need it. No
driver modules should depend on it. Instead depends on SPI and select
REGMAP_SPI. Depending on REGMAP_SPI makes this driver only build if another
driver already selected REGMAP_SPI, as the symbol can't be selected through the
menu kernel configuration.

Signed-off-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Link: http://lore.kernel.org/r/20200413141013.506613-1-enric.balletbo@collabora.com
Reviewed-by: Dan Murphy <dmurphy@ti.com>
Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: flexcan: fix failure handling of pm_runtime_get_sync()
Zhang Qilong [Sun, 8 Nov 2020 08:30:00 +0000 (16:30 +0800)]
can: flexcan: fix failure handling of pm_runtime_get_sync()

pm_runtime_get_sync() will increment pm usage at first and it will resume the
device later. If runtime of the device has error or device is in inaccessible
state(or other error state), resume operation will fail. If we do not call put
operation to decrease the reference, it will result in reference leak in the
two functions flexcan_get_berr_counter() and flexcan_open().

Moreover, this device cannot enter the idle state and always stay busy or other
non-idle state later. So we should fix it through adding
pm_runtime_put_noidle().

Fixes: ca10989632d88 ("can: flexcan: implement can Runtime PM")
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Link: https://lore.kernel.org/r/20201108083000.2599705-1-zhangqilong3@huawei.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: flexcan: flexcan_setup_stop_mode(): add missing "req_bit" to stop mode property...
Marc Kleine-Budde [Wed, 14 Oct 2020 11:41:36 +0000 (13:41 +0200)]
can: flexcan: flexcan_setup_stop_mode(): add missing "req_bit" to stop mode property comment

In the patch

    d9b081e3fc4b ("can: flexcan: remove ack_grp and ack_bit handling from driver")

the unused ack_grp and ack_bit were removed from the driver. However in the
comment, the "req_bit" was accidentally removed, too.

This patch adds back the "req_bit" bit.

Fixes: d9b081e3fc4b ("can: flexcan: remove ack_grp and ack_bit handling from driver")
Reported-by: Joakim Zhang <qiangqing.zhang@nxp.com>
Link: http://lore.kernel.org/r/20201014114810.2911135-1-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: peak_usb: fix potential integer overflow on shift of a int
Colin Ian King [Thu, 5 Nov 2020 11:24:27 +0000 (11:24 +0000)]
can: peak_usb: fix potential integer overflow on shift of a int

The left shift of int 32 bit integer constant 1 is evaluated using 32 bit
arithmetic and then assigned to a signed 64 bit variable. In the case where
time_ref->adapter->ts_used_bits is 32 or more this can lead to an oveflow.
Avoid this by shifting using the BIT_ULL macro instead.

Fixes: bb4785551f64 ("can: usb: PEAK-System Technik USB adapters driver core")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Link: https://lore.kernel.org/r/20201105112427.40688-1-colin.king@canonical.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: mcba_usb: mcba_usb_start_xmit(): first fill skb, then pass to can_put_echo_skb()
Marc Kleine-Budde [Wed, 28 Aug 2019 19:16:55 +0000 (21:16 +0200)]
can: mcba_usb: mcba_usb_start_xmit(): first fill skb, then pass to can_put_echo_skb()

The driver has to first fill the skb with data and then handle it to
can_put_echo_skb(). This patch moves the can_put_echo_skb() down, right before
sending the skb out via USB.

Fixes: 51f3baad7de9 ("can: mcba_usb: Add support for Microchip CAN BUS Analyzer")
Cc: Remigiusz Kołłątaj <remigiusz.kollataj@mobica.com>
Link: https://lore.kernel.org/r/20201111221204.1639007-1-mkl@pengutronix.de
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: ti_hecc: Fix memleak in ti_hecc_probe
Zhang Qilong [Sat, 14 Nov 2020 11:17:08 +0000 (19:17 +0800)]
can: ti_hecc: Fix memleak in ti_hecc_probe

In the error handling, we should goto the probe_exit_candev
to free ndev to prevent memory leak.

Fixes: dabf54dd1c63 ("can: ti_hecc: Convert TI HECC driver to DT only driver")
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Link: https://lore.kernel.org/r/20201114111708.3465543-1-zhangqilong3@huawei.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: dev: can_restart(): post buffer from the right context
Alejandro Concepcion Rodriguez [Thu, 5 Nov 2020 21:51:47 +0000 (21:51 +0000)]
can: dev: can_restart(): post buffer from the right context

netif_rx() is meant to be called from interrupt contexts. can_restart() may be
called by can_restart_work(), which is called from a worqueue, so it may run in
process context. Use netif_rx_ni() instead.

Fixes: 39549eef3587 ("can: CAN Network device driver and Netlink interface")
Co-developed-by: Loris Fauster <loris.fauster@ttcontrol.com>
Signed-off-by: Loris Fauster <loris.fauster@ttcontrol.com>
Signed-off-by: Alejandro Concepcion Rodriguez <alejandro@acoro.eu>
Link: https://lore.kernel.org/r/4e84162b-fb31-3a73-fa9a-9438b4bd5234@acoro.eu
[mkl: use netif_rx_ni() instead of netif_rx_any_context()]
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: af_can: prevent potential access of uninitialized member in canfd_rcv()
Anant Thazhemadam [Tue, 3 Nov 2020 21:39:06 +0000 (03:09 +0530)]
can: af_can: prevent potential access of uninitialized member in canfd_rcv()

In canfd_rcv(), cfd->len is uninitialized when skb->len = 0, and this
uninitialized cfd->len is accessed nonetheless by pr_warn_once().

Fix this uninitialized variable access by checking cfd->len's validity
condition (cfd->len > CANFD_MAX_DLEN) separately after the skb->len's
condition is checked, and appropriately modify the log messages that
are generated as well.
In case either of the required conditions fail, the skb is freed and
NET_RX_DROP is returned, same as before.

Fixes: d4689846881d ("can: af_can: canfd_rcv(): replace WARN_ONCE by pr_warn_once")
Reported-by: syzbot+9bcb0c9409066696d3aa@syzkaller.appspotmail.com
Tested-by: Anant Thazhemadam <anant.thazhemadam@gmail.com>
Signed-off-by: Anant Thazhemadam <anant.thazhemadam@gmail.com>
Link: https://lore.kernel.org/r/20201103213906.24219-3-anant.thazhemadam@gmail.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agocan: af_can: prevent potential access of uninitialized member in can_rcv()
Anant Thazhemadam [Tue, 3 Nov 2020 21:39:05 +0000 (03:09 +0530)]
can: af_can: prevent potential access of uninitialized member in can_rcv()

In can_rcv(), cfd->len is uninitialized when skb->len = 0, and this
uninitialized cfd->len is accessed nonetheless by pr_warn_once().

Fix this uninitialized variable access by checking cfd->len's validity
condition (cfd->len > CAN_MAX_DLEN) separately after the skb->len's
condition is checked, and appropriately modify the log messages that
are generated as well.
In case either of the required conditions fail, the skb is freed and
NET_RX_DROP is returned, same as before.

Fixes: 8cb68751c115 ("can: af_can: can_rcv(): replace WARN_ONCE by pr_warn_once")
Reported-by: syzbot+9bcb0c9409066696d3aa@syzkaller.appspotmail.com
Tested-by: Anant Thazhemadam <anant.thazhemadam@gmail.com>
Signed-off-by: Anant Thazhemadam <anant.thazhemadam@gmail.com>
Link: https://lore.kernel.org/r/20201103213906.24219-2-anant.thazhemadam@gmail.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
4 years agodevlink: Add missing genlmsg_cancel() in devlink_nl_sb_port_pool_fill()
Wang Hai [Fri, 13 Nov 2020 11:16:22 +0000 (19:16 +0800)]
devlink: Add missing genlmsg_cancel() in devlink_nl_sb_port_pool_fill()

If sb_occ_port_pool_get() failed in devlink_nl_sb_port_pool_fill(),
msg should be canceled by genlmsg_cancel().

Fixes: df38dafd2559 ("devlink: implement shared buffer occupancy monitoring interface")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Link: https://lore.kernel.org/r/20201113111622.11040-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: stmmac: dwmac_lib: enlarge dma reset timeout
Jisheng Zhang [Fri, 13 Nov 2020 01:09:02 +0000 (09:09 +0800)]
net: stmmac: dwmac_lib: enlarge dma reset timeout

If the phy enables power saving technology, the dwmac's software reset
needs more time to complete, enlarge dma reset timeout to 200000us.

Signed-off-by: Jisheng Zhang <Jisheng.Zhang@synaptics.com>
Link: https://lore.kernel.org/r/20201113090902.5c7aab1a@xhacker.debian
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agolan743x: prevent entire kernel HANG on open, for some platforms
Sven Van Asbroeck [Thu, 12 Nov 2020 20:47:41 +0000 (15:47 -0500)]
lan743x: prevent entire kernel HANG on open, for some platforms

On arm imx6, when opening the chip's netdev, the whole Linux
kernel intermittently hangs/freezes.

This is caused by a bug in the driver code which tests if pcie
interrupts are working correctly, using the software interrupt:

1. open: enable the software interrupt
2. open: tell the chip to assert the software interrupt
3. open: wait for flag
4. ISR: acknowledge s/w interrupt, set flag
5. open: notice flag, disable the s/w interrupt, continue

Unfortunately the ISR only acknowledges the s/w interrupt, but
does not disable it. This will re-trigger the ISR in a tight
loop.

On some (lucky) platforms, open proceeds to disable the s/w
interrupt even while the ISR is 'spinning'. On arm imx6,
the spinning ISR does not allow open to proceed, resulting
in a hung Linux kernel.

Fix minimally by disabling the s/w interrupt in the ISR, which
will prevent it from spinning. This won't break anything because
the s/w interrupt is used as a one-shot interrupt.

Note that this is a minimal fix, overlooking many possible
cleanups, e.g.:
- lan743x_intr_software_isr() is completely redundant and reads
  INT_STS twice for no apparent reason
- disabling the s/w interrupt in lan743x_intr_test_isr() is now
  redundant, but harmless
- waiting on software_isr_flag can be converted from a sleeping
  poll loop to wait_event_timeout()

Fixes: 23f0703c125b ("lan743x: Add main source files for new lan743x driver")
Tested-by: Sven Van Asbroeck <thesven73@gmail.com> # arm imx6 lan7430
Signed-off-by: Sven Van Asbroeck <thesven73@gmail.com>
Link: https://lore.kernel.org/r/20201112204741.12375-1-TheSven73@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agolan743x: fix issue causing intermittent kernel log warnings
Sven Van Asbroeck [Thu, 12 Nov 2020 18:59:49 +0000 (13:59 -0500)]
lan743x: fix issue causing intermittent kernel log warnings

When running this chip on arm imx6, we intermittently observe
the following kernel warning in the log, especially when the
system is under high load:

[   50.119484] ------------[ cut here ]------------
[   50.124377] WARNING: CPU: 0 PID: 303 at kernel/softirq.c:169 __local_bh_enable_ip+0x100/0x184
[   50.132925] IRQs not enabled as expected
[   50.159250] CPU: 0 PID: 303 Comm: rngd Not tainted 5.7.8 #1
[   50.164837] Hardware name: Freescale i.MX6 Quad/DualLite (Device Tree)
[   50.171395] [<c0111a38>] (unwind_backtrace) from [<c010be28>] (show_stack+0x10/0x14)
[   50.179162] [<c010be28>] (show_stack) from [<c05b9dec>] (dump_stack+0xac/0xd8)
[   50.186408] [<c05b9dec>] (dump_stack) from [<c0122e40>] (__warn+0xd0/0x10c)
[   50.193391] [<c0122e40>] (__warn) from [<c0123238>] (warn_slowpath_fmt+0x98/0xc4)
[   50.200892] [<c0123238>] (warn_slowpath_fmt) from [<c012b010>] (__local_bh_enable_ip+0x100/0x184)
[   50.209860] [<c012b010>] (__local_bh_enable_ip) from [<bf09ecbc>] (destroy_conntrack+0x48/0xd8 [nf_conntrack])
[   50.220038] [<bf09ecbc>] (destroy_conntrack [nf_conntrack]) from [<c0ac9b58>] (nf_conntrack_destroy+0x94/0x168)
[   50.230160] [<c0ac9b58>] (nf_conntrack_destroy) from [<c0a4aaa0>] (skb_release_head_state+0xa0/0xd0)
[   50.239314] [<c0a4aaa0>] (skb_release_head_state) from [<c0a4aadc>] (skb_release_all+0xc/0x24)
[   50.247946] [<c0a4aadc>] (skb_release_all) from [<c0a4b4cc>] (consume_skb+0x74/0x17c)
[   50.255796] [<c0a4b4cc>] (consume_skb) from [<c081a2dc>] (lan743x_tx_release_desc+0x120/0x124)
[   50.264428] [<c081a2dc>] (lan743x_tx_release_desc) from [<c081a98c>] (lan743x_tx_napi_poll+0x5c/0x18c)
[   50.273755] [<c081a98c>] (lan743x_tx_napi_poll) from [<c0a6b050>] (net_rx_action+0x118/0x4a4)
[   50.282306] [<c0a6b050>] (net_rx_action) from [<c0101364>] (__do_softirq+0x13c/0x53c)
[   50.290157] [<c0101364>] (__do_softirq) from [<c012b29c>] (irq_exit+0x150/0x17c)
[   50.297575] [<c012b29c>] (irq_exit) from [<c0196a08>] (__handle_domain_irq+0x60/0xb0)
[   50.305423] [<c0196a08>] (__handle_domain_irq) from [<c05d44fc>] (gic_handle_irq+0x4c/0x90)
[   50.313790] [<c05d44fc>] (gic_handle_irq) from [<c0100ed4>] (__irq_usr+0x54/0x80)
[   50.321287] Exception stack(0xecd99fb0 to 0xecd99ff8)
[   50.326355] 9fa0:                                     1cf1aa74 00000001 00000001 00000000
[   50.334547] 9fc0: 00000001 00000000 00000000 00000000 00000000 00000000 00004097 b6d17d14
[   50.342738] 9fe0: 00000001 b6d17c60 00000000 b6e71f94 800b0010 ffffffff
[   50.349364] irq event stamp: 2525027
[   50.352955] hardirqs last  enabled at (2525026): [<c0a6afec>] net_rx_action+0xb4/0x4a4
[   50.360892] hardirqs last disabled at (2525027): [<c0d6d2fc>] _raw_spin_lock_irqsave+0x1c/0x50
[   50.369517] softirqs last  enabled at (2524660): [<c01015b4>] __do_softirq+0x38c/0x53c
[   50.377446] softirqs last disabled at (2524693): [<c012b29c>] irq_exit+0x150/0x17c
[   50.385027] ---[ end trace c0b571db4bc8087d ]---

The driver is calling dev_kfree_skb() from code inside a spinlock,
where h/w interrupts are disabled. This is forbidden, as documented
in include/linux/netdevice.h. The correct function to use
dev_kfree_skb_irq(), or dev_kfree_skb_any().

Fix by using the correct dev_kfree_skb_xxx() functions:

in lan743x_tx_release_desc():
  called by lan743x_tx_release_completed_descriptors()
    called by in lan743x_tx_napi_poll()
    which holds a spinlock
  called by lan743x_tx_release_all_descriptors()
    called by lan743x_tx_close()
    which can-sleep
conclusion: use dev_kfree_skb_any()

in lan743x_tx_xmit_frame():
  which holds a spinlock
conclusion: use dev_kfree_skb_irq()

in lan743x_tx_close():
  which can-sleep
conclusion: use dev_kfree_skb()

in lan743x_rx_release_ring_element():
  called by lan743x_rx_close()
    which can-sleep
  called by lan743x_rx_open()
    which can-sleep
conclusion: use dev_kfree_skb()

Fixes: 23f0703c125b ("lan743x: Add main source files for new lan743x driver")
Signed-off-by: Sven Van Asbroeck <thesven73@gmail.com>
Link: https://lore.kernel.org/r/20201112185949.11315-1-TheSven73@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonetlabel: fix an uninitialized warning in netlbl_unlabel_staticlist()
Paul Moore [Fri, 13 Nov 2020 21:30:40 +0000 (16:30 -0500)]
netlabel: fix an uninitialized warning in netlbl_unlabel_staticlist()

Static checking revealed that a previous fix to
netlbl_unlabel_staticlist() leaves a stack variable uninitialized,
this patches fixes that.

Fixes: 866358ec331f ("netlabel: fix our progress tracking in netlbl_unlabel_staticlist()")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Reviewed-by: James Morris <jamorris@linux.microsoft.com>
Link: https://lore.kernel.org/r/160530304068.15651.18355773009751195447.stgit@sifl
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agosctp: change to hold/put transport for proto_unreach_timer
Xin Long [Sat, 14 Nov 2020 05:22:53 +0000 (13:22 +0800)]
sctp: change to hold/put transport for proto_unreach_timer

A call trace was found in Hangbin's Codenomicon testing with debug kernel:

  [ 2615.981988] ODEBUG: free active (active state 0) object type: timer_list hint: sctp_generate_proto_unreach_event+0x0/0x3a0 [sctp]
  [ 2615.995050] WARNING: CPU: 17 PID: 0 at lib/debugobjects.c:328 debug_print_object+0x199/0x2b0
  [ 2616.095934] RIP: 0010:debug_print_object+0x199/0x2b0
  [ 2616.191533] Call Trace:
  [ 2616.194265]  <IRQ>
  [ 2616.202068]  debug_check_no_obj_freed+0x25e/0x3f0
  [ 2616.207336]  slab_free_freelist_hook+0xeb/0x140
  [ 2616.220971]  kfree+0xd6/0x2c0
  [ 2616.224293]  rcu_do_batch+0x3bd/0xc70
  [ 2616.243096]  rcu_core+0x8b9/0xd00
  [ 2616.256065]  __do_softirq+0x23d/0xacd
  [ 2616.260166]  irq_exit+0x236/0x2a0
  [ 2616.263879]  smp_apic_timer_interrupt+0x18d/0x620
  [ 2616.269138]  apic_timer_interrupt+0xf/0x20
  [ 2616.273711]  </IRQ>

This is because it holds asoc when transport->proto_unreach_timer starts
and puts asoc when the timer stops, and without holding transport the
transport could be freed when the timer is still running.

So fix it by holding/putting transport instead for proto_unreach_timer
in transport, just like other timers in transport.

v1->v2:
  - Also use sctp_transport_put() for the "out_unlock:" path in
    sctp_generate_proto_unreach_event(), as Marcelo noticed.

Fixes: 50b5d6ad6382 ("sctp: Fix a race between ICMP protocol unreachable and connect()")
Reported-by: Hangbin Liu <liuhangbin@gmail.com>
Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Link: https://lore.kernel.org/r/102788809b554958b13b95d33440f5448113b8d6.1605331373.git.lucien.xin@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agovsock: forward all packets to the host when no H2G is registered
Stefano Garzarella [Thu, 12 Nov 2020 13:38:37 +0000 (14:38 +0100)]
vsock: forward all packets to the host when no H2G is registered

Before commit c0cfa2d8a788 ("vsock: add multi-transports support"),
if a G2H transport was loaded (e.g. virtio transport), every packets
was forwarded to the host, regardless of the destination CID.
The H2G transports implemented until then (vhost-vsock, VMCI) always
responded with an error, if the destination CID was not
VMADDR_CID_HOST.

From that commit, we are using the remote CID to decide which
transport to use, so packets with remote CID > VMADDR_CID_HOST(2)
are sent only through H2G transport. If no H2G is available, packets
are discarded directly in the guest.

Some use cases (e.g. Nitro Enclaves [1]) rely on the old behaviour
to implement sibling VMs communication, so we restore the old
behavior when no H2G is registered.
It will be up to the host to discard packets if the destination is
not the right one. As it was already implemented before adding
multi-transport support.

Tested with nested QEMU/KVM by me and Nitro Enclaves by Andra.

[1] Documentation/virt/ne_overview.rst

Cc: Jorgen Hansen <jhansen@vmware.com>
Cc: Dexuan Cui <decui@microsoft.com>
Fixes: c0cfa2d8a788 ("vsock: add multi-transports support")
Reported-by: Andra Paraschiv <andraprs@amazon.com>
Tested-by: Andra Paraschiv <andraprs@amazon.com>
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
Link: https://lore.kernel.org/r/20201112133837.34183-1-sgarzare@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: dsa: mv88e6xxx: Avoid VTU corruption on 6097
Tobias Waldekranz [Thu, 12 Nov 2020 11:43:35 +0000 (12:43 +0100)]
net: dsa: mv88e6xxx: Avoid VTU corruption on 6097

As soon as you add the second port to a VLAN, all other port
membership configuration is overwritten with zeroes. The HW interprets
this as all ports being "unmodified members" of the VLAN.

In the simple case when all ports belong to the same VLAN, switching
will still work. But using multiple VLANs or trying to set multiple
ports as tagged members will not work.

On the 6352, doing a VTU GetNext op, followed by an STU GetNext op
will leave you with both the member- and state- data in the VTU/STU
data registers. But on the 6097 (which uses the same implementation),
the STU GetNext will override the information gathered from the VTU
GetNext.

Separate the two stages, parsing the result of the VTU GetNext before
doing the STU GetNext.

We opt to update the existing implementation for all applicable chips,
as opposed to creating a separate callback for 6097, because although
the previous implementation did work for (at least) 6352, the
datasheet does not mention the masking behavior.

Fixes: ef6fcea37f01 ("net: dsa: mv88e6xxx: get STU entry on VTU GetNext")
Signed-off-by: Tobias Waldekranz <tobias@waldekranz.com>
Link: https://lore.kernel.org/r/20201112114335.27371-1-tobias@waldekranz.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: phy: smsc: add missed clk_disable_unprepare in smsc_phy_probe()
Zhang Changzhong [Thu, 12 Nov 2020 11:23:59 +0000 (19:23 +0800)]
net: phy: smsc: add missed clk_disable_unprepare in smsc_phy_probe()

Add the missing clk_disable_unprepare() before return from
smsc_phy_probe() in the error handling case.

Fixes: bedd8d78aba3 ("net: phy: smsc: LAN8710/20: add phy refclk in support")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Link: https://lore.kernel.org/r/1605180239-1792-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoipv6: Fix error path to cancel the meseage
Zhang Qilong [Thu, 12 Nov 2020 08:09:50 +0000 (16:09 +0800)]
ipv6: Fix error path to cancel the meseage

genlmsg_cancel() needs to be called in the error path of
inet6_fill_ifmcaddr and inet6_fill_ifacaddr to cancel
the message.

Fixes: 6ecf4c37eb3e ("ipv6: enable IFA_TARGET_NETNSID for RTM_GETADDR")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Qilong <zhangqilong3@huawei.com>
Link: https://lore.kernel.org/r/20201112080950.1476302-1-zhangqilong3@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoip_tunnels: Set tunnel option flag when tunnel metadata is present
Yi-Hung Wei [Wed, 11 Nov 2020 00:16:40 +0000 (16:16 -0800)]
ip_tunnels: Set tunnel option flag when tunnel metadata is present

Currently, we may set the tunnel option flag when the size of metadata
is zero.  For example, we set TUNNEL_GENEVE_OPT in the receive function
no matter the geneve option is present or not.  As this may result in
issues on the tunnel flags consumers, this patch fixes the issue.

Related discussion:
* https://lore.kernel.org/netdev/1604448694-19351-1-git-send-email-yihung.wei@gmail.com/T/#u

Fixes: 256c87c17c53 ("net: check tunnel option type in tunnel flags")
Signed-off-by: Yi-Hung Wei <yihung.wei@gmail.com>
Link: https://lore.kernel.org/r/1605053800-74072-1-git-send-email-yihung.wei@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: ethernet: ti: am65-cpts: update ret when ptp_clock is ERROR
Wang Qing [Thu, 12 Nov 2020 16:45:41 +0000 (18:45 +0200)]
net: ethernet: ti: am65-cpts: update ret when ptp_clock is ERROR

We always have to update the value of ret, otherwise the
 error value may be the previous one.

Fixes: f6bd59526ca5 ("net: ethernet: ti: introduce am654 common platform time sync driver")
Signed-off-by: Wang Qing <wangqing@vivo.com>
[grygorii.strashko@ti.com: fix build warn, subj add fixes tag]
Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
Acked-by: Richard Cochran <richardcochran@gmail.com>
Link: https://lore.kernel.org/r/20201112164541.3223-1-grygorii.strashko@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: marvell: prestera: fix error return code in prestera_pci_probe()
Wang Hai [Fri, 13 Nov 2020 11:32:36 +0000 (19:32 +0800)]
net: marvell: prestera: fix error return code in prestera_pci_probe()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 4c2703dfd7fa ("net: marvell: prestera: Add PCI interface support")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Reviewed-by: Vadym Kochan <vadym.kochan@plvision.eu>
Acked-by: Vadym Kochan <vadym.kochan@plvision.eu>
Link: https://lore.kernel.org/r/20201113113236.71678-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: phy: mscc: remove non-MACSec compatible phy
Steen Hegelund [Fri, 13 Nov 2020 09:11:16 +0000 (10:11 +0100)]
net: phy: mscc: remove non-MACSec compatible phy

Selecting VSC8575 as a MACSec PHY was not correct

The relevant datasheet can be found here:
  - VSC8575: https://www.microchip.com/wwwproducts/en/VSC8575

History:
v1 -> v2:
   - Corrected the sha in the "Fixes:" tag

Fixes: 1bbe0ecc2a1a ("net: phy: mscc: macsec initialization")
Signed-off-by: Steen Hegelund <steen.hegelund@microchip.com>
Reviewed-by: Antoine Tenart <atenart@kernel.org>
Link: https://lore.kernel.org/r/20201113091116.1102450-1-steen.hegelund@microchip.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: Exempt multicast addresses from five-second neighbor lifetime
Jeff Dike [Fri, 13 Nov 2020 01:58:15 +0000 (20:58 -0500)]
net: Exempt multicast addresses from five-second neighbor lifetime

Commit 58956317c8de ("neighbor: Improve garbage collection")
guarantees neighbour table entries a five-second lifetime.  Processes
which make heavy use of multicast can fill the neighour table with
multicast addresses in five seconds.  At that point, neighbour entries
can't be GC-ed because they aren't five seconds old yet, the kernel
log starts to fill up with "neighbor table overflow!" messages, and
sends start to fail.

This patch allows multicast addresses to be thrown out before they've
lived out their five seconds.  This makes room for non-multicast
addresses and makes messages to all addresses more reliable in these
circumstances.

Fixes: 58956317c8de ("neighbor: Improve garbage collection")
Signed-off-by: Jeff Dike <jdike@akamai.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Link: https://lore.kernel.org/r/20201113015815.31397-1-jdike@akamai.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: ethernet: ti: cpsw: fix cpts irq after suspend
Grygorii Strashko [Thu, 12 Nov 2020 11:15:46 +0000 (13:15 +0200)]
net: ethernet: ti: cpsw: fix cpts irq after suspend

Depending on the SoC/platform the CPSW can completely lose context after a
suspend/resume cycle, including CPSW wrapper (WR) which will cause reset of
WR_C0_MISC_EN register, so CPTS IRQ will became disabled.

Fix it by moving CPTS IRQ enabling in cpsw_ndo_open() where CPTS is
actually started.

Fixes: 84ea9c0a95d7 ("net: ethernet: ti: cpsw: enable cpts irq")
Reported-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Grygorii Strashko <grygorii.strashko@ti.com>
Tested-by: Tony Lindgren <tony@atomide.com>
Link: https://lore.kernel.org/r/20201112111546.20343-1-grygorii.strashko@ti.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoMerge tag 'mac80211-for-net-2020-11-13' of git://git.kernel.org/pub/scm/linux/kernel...
Jakub Kicinski [Fri, 13 Nov 2020 19:58:55 +0000 (11:58 -0800)]
Merge tag 'mac80211-for-net-2020-11-13' of git://git./linux/kernel/git/jberg/mac80211

Johannes Berg says:

====================
A handful of fixes:
 * a use-after-free fix in rfkill
 * a memory leak fix in the mac80211 TX status path
 * some rate scaling fixes
 * a fix for the often-reported (by syzbot) sleeping
   in atomic issue with mac80211's station removal

* tag 'mac80211-for-net-2020-11-13' of git://git.kernel.org/pub/scm/linux/kernel/git/jberg/mac80211:
  mac80211: free sta in sta_info_insert_finish() on errors
  mac80211: minstrel: fix tx status processing corner case
  mac80211: minstrel: remove deferred sampling code
  mac80211: fix memory leak on filtered powersave frames
  rfkill: Fix use-after-free in rfkill_resume()
====================

Link: https://lore.kernel.org/r/20201113093421.24025-1-johannes@sipsolutions.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agomac80211: free sta in sta_info_insert_finish() on errors
Johannes Berg [Thu, 12 Nov 2020 10:22:04 +0000 (11:22 +0100)]
mac80211: free sta in sta_info_insert_finish() on errors

If sta_info_insert_finish() fails, we currently keep the station
around and free it only in the caller, but there's only one such
caller and it always frees it immediately.

As syzbot found, another consequence of this split is that we can
put things that sleep only into __cleanup_single_sta() and not in
sta_info_free(), but this is the only place that requires such of
sta_info_free() now.

Change this to free the station in sta_info_insert_finish(), in
which case we can still sleep. This will also let us unify the
cleanup code later.

Cc: stable@vger.kernel.org
Fixes: dcd479e10a05 ("mac80211: always wind down STA state")
Reported-by: syzbot+32c6c38c4812d22f2f0b@syzkaller.appspotmail.com
Reported-by: syzbot+4c81fe92e372d26c4246@syzkaller.appspotmail.com
Reported-by: syzbot+6a7fe9faf0d1d61bc24a@syzkaller.appspotmail.com
Reported-by: syzbot+abed06851c5ffe010921@syzkaller.appspotmail.com
Reported-by: syzbot+b7aeb9318541a1c709f1@syzkaller.appspotmail.com
Reported-by: syzbot+d5a9416c6cafe53b5dd0@syzkaller.appspotmail.com
Link: https://lore.kernel.org/r/20201112112201.ee6b397b9453.I9c31d667a0ea2151441cc64ed6613d36c18a48e0@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
4 years agonet: x25: Increase refcnt of "struct x25_neigh" in x25_rx_call_request
Xie He [Thu, 12 Nov 2020 10:35:06 +0000 (02:35 -0800)]
net: x25: Increase refcnt of "struct x25_neigh" in x25_rx_call_request

The x25_disconnect function in x25_subr.c would decrease the refcount of
"x25->neighbour" (struct x25_neigh) and reset this pointer to NULL.

However, the x25_rx_call_request function in af_x25.c, which is called
when we receive a connection request, does not increase the refcount when
it assigns the pointer.

Fix this issue by increasing the refcount of "struct x25_neigh" in
x25_rx_call_request.

This patch fixes frequent kernel crashes when using AF_X25 sockets.

Fixes: 4becb7ee5b3d ("net/x25: Fix x25_neigh refcnt leak when x25 disconnect")
Cc: Martin Schiller <ms@dev.tdt.de>
Signed-off-by: Xie He <xie.he.0141@gmail.com>
Link: https://lore.kernel.org/r/20201112103506.5875-1-xie.he.0141@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: ethernet: mtk-star-emac: fix error return code in mtk_star_enable()
Zhang Changzhong [Thu, 12 Nov 2020 11:34:39 +0000 (19:34 +0800)]
net: ethernet: mtk-star-emac: fix error return code in mtk_star_enable()

Fix to return a negative error code from the error handling
case instead of 0, as done elsewhere in this function.

Fixes: 8c7bd5a454ff ("net: ethernet: mtk-star-emac: new driver")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Acked-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Link: https://lore.kernel.org/r/1605180879-2573-1-git-send-email-zhangchangzhong@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: ethernet: mtk-star-emac: return ok when xmit drops
Vincent StehlƩ [Thu, 12 Nov 2020 08:48:33 +0000 (09:48 +0100)]
net: ethernet: mtk-star-emac: return ok when xmit drops

The ndo_start_xmit() method must return NETDEV_TX_OK if the DMA mapping
fails, after freeing the socket buffer.
Fix the mtk_star_netdev_start_xmit() function accordingly.

Fixes: 8c7bd5a454ff ("net: ethernet: mtk-star-emac: new driver")
Signed-off-by: Vincent StehlƩ <vincent.stehle@laposte.net>
Acked-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Link: https://lore.kernel.org/r/20201112084833.21842-1-vincent.stehle@laposte.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet/ncsi: Fix netlink registration
Joel Stanley [Thu, 12 Nov 2020 06:12:10 +0000 (16:42 +1030)]
net/ncsi: Fix netlink registration

If a user unbinds and re-binds a NC-SI aware driver the kernel will
attempt to register the netlink interface at runtime. The structure is
marked __ro_after_init so registration fails spectacularly at this point.

 # echo 1e660000.ethernet > /sys/bus/platform/drivers/ftgmac100/unbind
 # echo 1e660000.ethernet > /sys/bus/platform/drivers/ftgmac100/bind
  ftgmac100 1e660000.ethernet: Read MAC address 52:54:00:12:34:56 from chip
  ftgmac100 1e660000.ethernet: Using NCSI interface
  8<--- cut here ---
  Unable to handle kernel paging request at virtual address 80a8f858
  pgd = 8c768dd6
  [80a8f858] *pgd=80a0841e(bad)
  Internal error: Oops: 80d [#1] SMP ARM
  CPU: 0 PID: 116 Comm: sh Not tainted 5.10.0-rc3-next-20201111-00003-gdd25b227ec1e #51
  Hardware name: Generic DT based system
  PC is at genl_register_family+0x1f8/0x6d4
  LR is at 0xff26ffff
  pc : [<8073f930>]    lr : [<ff26ffff>]    psr: 20000153
  sp : 8553bc80  ip : 81406244  fp : 8553bd04
  r10: 8085d12c  r9 : 80a8f73c  r8 : 85739000
  r7 : 00000017  r6 : 80a8f860  r5 : 80c8ab98  r4 : 80a8f858
  r3 : 00000000  r2 : 00000000  r1 : 81406130  r0 : 00000017
  Flags: nzCv  IRQs on  FIQs off  Mode SVC_32  ISA ARM  Segment none
  Control: 00c5387d  Table: 85524008  DAC: 00000051
  Process sh (pid: 116, stack limit = 0x1f1988d6)
 ...
  Backtrace:
  [<8073f738>] (genl_register_family) from [<80860ac0>] (ncsi_init_netlink+0x20/0x48)
   r10:8085d12c r9:80c8fb0c r8:85739000 r7:00000000 r6:81218000 r5:85739000
   r4:8121c000
  [<80860aa0>] (ncsi_init_netlink) from [<8085d740>] (ncsi_register_dev+0x1b0/0x210)
   r5:8121c400 r4:8121c000
  [<8085d590>] (ncsi_register_dev) from [<805a8060>] (ftgmac100_probe+0x6e0/0x778)
   r10:00000004 r9:80950228 r8:8115bc10 r7:8115ab00 r6:9eae2c24 r5:813b6f88
   r4:85739000
  [<805a7980>] (ftgmac100_probe) from [<805355ec>] (platform_drv_probe+0x58/0xa8)
   r9:80c76bb0 r8:00000000 r7:80cd4974 r6:80c76bb0 r5:8115bc10 r4:00000000
  [<80535594>] (platform_drv_probe) from [<80532d58>] (really_probe+0x204/0x514)
   r7:80cd4974 r6:00000000 r5:80cd4868 r4:8115bc10

Jakub pointed out that ncsi_register_dev is obviously broken, because
there is only one family so it would never work if there was more than
one ncsi netdev.

Fix the crash by registering the netlink family once on boot, and drop
the code to unregister it.

Fixes: 955dc68cb9b2 ("net/ncsi: Add generic netlink family")
Signed-off-by: Joel Stanley <joel@jms.id.au>
Reviewed-by: Samuel Mendoza-Jonas <sam@mendozajonas.com>
Link: https://lore.kernel.org/r/20201112061210.914621-1-joel@jms.id.au
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoMerge tag 'net-5.10-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net
Linus Torvalds [Thu, 12 Nov 2020 22:02:04 +0000 (14:02 -0800)]
Merge tag 'net-5.10-rc4' of git://git./linux/kernel/git/netdev/net

Pull networking fixes from Jakub Kicinski:
 "Current release - regressions:

   - arm64: dts: fsl-ls1028a-kontron-sl28: specify in-band mode for
     ENETC

  Current release - bugs in new features:

   - mptcp: provide rmem[0] limit offset to fix oops

  Previous release - regressions:

   - IPv6: Set SIT tunnel hard_header_len to zero to fix path MTU
     calculations

   - lan743x: correctly handle chips with internal PHY

   - bpf: Don't rely on GCC __attribute__((optimize)) to disable GCSE

   - mlx5e: Fix VXLAN port table synchronization after function reload

  Previous release - always broken:

   - bpf: Zero-fill re-used per-cpu map element

   - fix out-of-order UDP packets when forwarding with UDP GSO fraglists
     turned on:
       - fix UDP header access on Fast/frag0 UDP GRO
       - fix IP header access and skb lookup on Fast/frag0 UDP GRO

   - ethtool: netlink: add missing netdev_features_change() call

   - net: Update window_clamp if SOCK_RCVBUF is set

   - igc: Fix returning wrong statistics

   - ch_ktls: fix multiple leaks and corner cases in Chelsio TLS offload

   - tunnels: Fix off-by-one in lower MTU bounds for ICMP/ICMPv6 replies

   - r8169: disable hw csum for short packets on all chip versions

   - vrf: Fix fast path output packet handling with async Netfilter
     rules"

* tag 'net-5.10-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (65 commits)
  lan743x: fix use of uninitialized variable
  net: udp: fix IP header access and skb lookup on Fast/frag0 UDP GRO
  net: udp: fix UDP header access on Fast/frag0 UDP GRO
  devlink: Avoid overwriting port attributes of registered port
  vrf: Fix fast path output packet handling with async Netfilter rules
  cosa: Add missing kfree in error path of cosa_write
  net: switch to the kernel.org patchwork instance
  ch_ktls: stop the txq if reaches threshold
  ch_ktls: tcb update fails sometimes
  ch_ktls/cxgb4: handle partial tag alone SKBs
  ch_ktls: don't free skb before sending FIN
  ch_ktls: packet handling prior to start marker
  ch_ktls: Correction in middle record handling
  ch_ktls: missing handling of header alone
  ch_ktls: Correction in trimmed_len calculation
  cxgb4/ch_ktls: creating skbs causes panic
  ch_ktls: Update cheksum information
  ch_ktls: Correction in finding correct length
  cxgb4/ch_ktls: decrypted bit is not enough
  net/x25: Fix null-ptr-deref in x25_connect
  ...

4 years agoMerge tag 'nfs-for-5.10-2' of git://git.linux-nfs.org/projects/anna/linux-nfs
Linus Torvalds [Thu, 12 Nov 2020 21:49:12 +0000 (13:49 -0800)]
Merge tag 'nfs-for-5.10-2' of git://git.linux-nfs.org/projects/anna/linux-nfs

Pull NFS client bugfixes from Anna Schumaker:
 "Stable fixes:
  - Fix failure to unregister shrinker

  Other fixes:
  - Fix unnecessary locking to clear up some contention
  - Fix listxattr receive buffer size
  - Fix default mount options for nfsroot"

* tag 'nfs-for-5.10-2' of git://git.linux-nfs.org/projects/anna/linux-nfs:
  NFS: Remove unnecessary inode lock in nfs_fsync_dir()
  NFS: Remove unnecessary inode locking in nfs_llseek_dir()
  NFS: Fix listxattr receive buffer size
  NFSv4.2: fix failure to unregister shrinker
  nfsroot: Default mount option should ask for built-in NFS version

4 years agoMerge tag 'acpi-5.10-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Thu, 12 Nov 2020 19:06:53 +0000 (11:06 -0800)]
Merge tag 'acpi-5.10-rc4' of git://git./linux/kernel/git/rafael/linux-pm

Pull ACPI fixes from Rafael Wysocki:
 "These are mostly docmentation fixes and janitorial changes plus some
  new device IDs and a new quirk.

  Specifics:

   - Fix documentation regarding GPIO properties (Andy Shevchenko)

   - Fix spelling mistakes in ACPI documentation (Flavio Suligoi)

   - Fix white space inconsistencies in ACPI code (Maximilian Luz)

   - Fix string formatting in the ACPI Generic Event Device (GED) driver
     (Nick Desaulniers)

   - Add Intel Alder Lake device IDs to the ACPI drivers used by the
     Dynamic Platform and Thermal Framework (Srinivas Pandruvada)

   - Add lid-related DMI quirk for Medion Akoya E2228T to the ACPI
     button driver (Hans de Goede)"

* tag 'acpi-5.10-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  ACPI: DPTF: Support Alder Lake
  Documentation: ACPI: fix spelling mistakes
  ACPI: button: Add DMI quirk for Medion Akoya E2228T
  ACPI: GED: fix -Wformat
  ACPI: Fix whitespace inconsistencies
  ACPI: scan: Fix acpi_dma_configure_id() kerneldoc name
  Documentation: firmware-guide: gpio-properties: Clarify initial output state
  Documentation: firmware-guide: gpio-properties: active_low only for GpioIo()
  Documentation: firmware-guide: gpio-properties: Fix factual mistakes

4 years agoMerge tag 'pm-5.10-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm
Linus Torvalds [Thu, 12 Nov 2020 19:03:38 +0000 (11:03 -0800)]
Merge tag 'pm-5.10-rc4' of git://git./linux/kernel/git/rafael/linux-pm

Pull power management fixes from Rafael Wysocki:
 "Make the intel_pstate driver behave as expected when it operates in
  the passive mode with HWP enabled and the 'powersave' governor on top
  of it"

* tag 'pm-5.10-rc4' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  cpufreq: intel_pstate: Take CPUFREQ_GOV_STRICT_TARGET into account
  cpufreq: Add strict_target to struct cpufreq_policy
  cpufreq: Introduce CPUFREQ_GOV_STRICT_TARGET
  cpufreq: Introduce governor flags

4 years agolan743x: fix use of uninitialized variable
Sven Van Asbroeck [Thu, 12 Nov 2020 15:25:13 +0000 (10:25 -0500)]
lan743x: fix use of uninitialized variable

When no devicetree is present, the driver will use an
uninitialized variable.

Fix by initializing this variable.

Fixes: 902a66e08cea ("lan743x: correctly handle chips with internal PHY")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Sven Van Asbroeck <thesven73@gmail.com>
Link: https://lore.kernel.org/r/20201112152513.1941-1-TheSven73@gmail.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoMerge branch 'net-udp-fix-fast-frag0-udp-gro'
Jakub Kicinski [Thu, 12 Nov 2020 17:55:58 +0000 (09:55 -0800)]
Merge branch 'net-udp-fix-fast-frag0-udp-gro'

Alexander Lobakin says:

====================
net: udp: fix Fast/frag0 UDP GRO

While testing UDP GSO fraglists forwarding through driver that uses
Fast GRO (via napi_gro_frags()), I was observing lots of out-of-order
iperf packets:

[ ID] Interval           Transfer     Bitrate         Jitter
[SUM]  0.0-40.0 sec  12106 datagrams received out-of-order

Simple switch to napi_gro_receive() or any other method without frag0
shortcut completely resolved them.

I've found two incorrect header accesses in GRO receive callback(s):
 - udp_hdr() (instead of udp_gro_udphdr()) that always points to junk
   in "fast" mode and could probably do this in "regular".
   This was the actual bug that caused all out-of-order delivers;
 - udp{4,6}_lib_lookup_skb() -> ip{,v6}_hdr() (instead of
   skb_gro_network_header()) that potentionally might return odd
   pointers in both modes.

Each patch addresses one of these two issues.

This doesn't cover a support for nested tunnels as it's out of the
subject and requires more invasive changes. It will be handled
separately in net-next series.

Credits:
Cc: Eric Dumazet <edumazet@google.com>
Cc: Jakub Kicinski <kuba@kernel.org>
Cc: Willem de Bruijn <willemb@google.com>
Since v4 [0]:
 - split the fix into two logical ones (Willem);
 - replace ternaries with plain ifs to beautify the code (Jakub);
 - drop p->data part to reintroduce it later in abovementioned set.

Since v3 [1]:
 - restore the original {,__}udp{4,6}_lib_lookup_skb() and use
   private versions of them inside GRO code (Willem).

Since v2 [2]:
 - dropped redundant check introduced in v2 as it's performed right
   before (thanks to Eric);
 - udp_hdr() switched to data + off for skbs from list (also Eric);
 - fixed possible malfunction of {,__}udp{4,6}_lib_lookup_skb() with
   Fast/frag0 due to ip{,v6}_hdr() usage (Willem).

Since v1 [3]:
 - added a NULL pointer check for "uh" as suggested by Willem.

[0] https://lore.kernel.org/netdev/Ha2hou5eJPcblo4abjAqxZRzIl1RaLs2Hy0oOAgFs@cp4-web-036.plabs.ch
[1] https://lore.kernel.org/netdev/MgZce9htmEtCtHg7pmWxXXfdhmQ6AHrnltXC41zOoo@cp7-web-042.plabs.ch
[2] https://lore.kernel.org/netdev/0eaG8xtbtKY1dEKCTKUBubGiC9QawGgB3tVZtNqVdY@cp4-web-030.plabs.ch
[3] https://lore.kernel.org/netdev/YazU6GEzBdpyZMDMwJirxDX7B4sualpDG68ADZYvJI@cp4-web-034.plabs.ch
====================

Link: https://lore.kernel.org/r/hjGOh0iCOYyo1FPiZh6TMXcx3YCgNs1T1eGKLrDz8@cp4-web-037.plabs.ch
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: udp: fix IP header access and skb lookup on Fast/frag0 UDP GRO
Alexander Lobakin [Wed, 11 Nov 2020 20:45:38 +0000 (20:45 +0000)]
net: udp: fix IP header access and skb lookup on Fast/frag0 UDP GRO

udp{4,6}_lib_lookup_skb() use ip{,v6}_hdr() to get IP header of the
packet. While it's probably OK for non-frag0 paths, this helpers
will also point to junk on Fast/frag0 GRO when all headers are
located in frags. As a result, sk/skb lookup may fail or give wrong
results. To support both GRO modes, skb_gro_network_header() might
be used. To not modify original functions, add private versions of
udp{4,6}_lib_lookup_skb() only to perform correct sk lookups on GRO.

Present since the introduction of "application-level" UDP GRO
in 4.7-rc1.

Misc: replace totally unneeded ternaries with plain ifs.

Fixes: a6024562ffd7 ("udp: Add GRO functions to UDP socket")
Suggested-by: Willem de Bruijn <willemb@google.com>
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Alexander Lobakin <alobakin@pm.me>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: udp: fix UDP header access on Fast/frag0 UDP GRO
Alexander Lobakin [Wed, 11 Nov 2020 20:45:25 +0000 (20:45 +0000)]
net: udp: fix UDP header access on Fast/frag0 UDP GRO

UDP GRO uses udp_hdr(skb) in its .gro_receive() callback. While it's
probably OK for non-frag0 paths (when all headers or even the entire
frame are already in skb head), this inline points to junk when
using Fast GRO (napi_gro_frags() or napi_gro_receive() with only
Ethernet header in skb head and all the rest in the frags) and breaks
GRO packet compilation and the packet flow itself.
To support both modes, skb_gro_header_fast() + skb_gro_header_slow()
are typically used. UDP even has an inline helper that makes use of
them, udp_gro_udphdr(). Use that instead of troublemaking udp_hdr()
to get rid of the out-of-order delivers.

Present since the introduction of plain UDP GRO in 5.0-rc1.

Fixes: e20cf8d3f1f7 ("udp: implement GRO for plain UDP sockets.")
Cc: Eric Dumazet <edumazet@google.com>
Signed-off-by: Alexander Lobakin <alobakin@pm.me>
Acked-by: Willem de Bruijn <willemb@google.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoMerge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Jakub Kicinski [Thu, 12 Nov 2020 16:47:22 +0000 (08:47 -0800)]
Merge branch '40GbE' of git://git./linux/kernel/git/tnguy/net-queue

Tony Nguyen says:

====================
Intel Wired LAN Driver Updates 2020-11-10

This series contains updates to i40e and igc drivers and the MAINTAINERS
file.

Slawomir fixes updating VF MAC addresses to fix various issues related
to reporting and setting of these addresses for i40e.

Dan Carpenter fixes a possible used before being initialized issue for
i40e.

Vinicius fixes reporting of netdev stats for igc.

Tony updates repositories for Intel Ethernet Drivers.

* '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
  MAINTAINERS: Update repositories for Intel Ethernet Drivers
  igc: Fix returning wrong statistics
  i40e, xsk: uninitialized variable in i40e_clean_rx_irq_zc()
  i40e: Fix MAC address setting for a VF via Host/VM
====================

Link: https://lore.kernel.org/r/20201111001955.533210-1-anthony.l.nguyen@intel.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agodevlink: Avoid overwriting port attributes of registered port
Parav Pandit [Wed, 11 Nov 2020 03:47:44 +0000 (05:47 +0200)]
devlink: Avoid overwriting port attributes of registered port

Cited commit in fixes tag overwrites the port attributes for the
registered port.

Avoid such error by checking registered flag before setting attributes.

Fixes: 71ad8d55f8e5 ("devlink: Replace devlink_port_attrs_set parameters with a struct")
Signed-off-by: Parav Pandit <parav@nvidia.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Link: https://lore.kernel.org/r/20201111034744.35554-1-parav@nvidia.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agovrf: Fix fast path output packet handling with async Netfilter rules
Martin Willi [Fri, 6 Nov 2020 07:30:30 +0000 (08:30 +0100)]
vrf: Fix fast path output packet handling with async Netfilter rules

VRF devices use an optimized direct path on output if a default qdisc
is involved, calling Netfilter hooks directly. This path, however, does
not consider Netfilter rules completing asynchronously, such as with
NFQUEUE. The Netfilter okfn() is called for asynchronously accepted
packets, but the VRF never passes that packet down the stack to send
it out over the slave device. Using the slower redirect path for this
seems not feasible, as we do not know beforehand if a Netfilter hook
has asynchronously completing rules.

Fix the use of asynchronously completing Netfilter rules in OUTPUT and
POSTROUTING by using a special completion function that additionally
calls dst_output() to pass the packet down the stack. Also, slightly
adjust the use of nf_reset_ct() so that is called in the asynchronous
case, too.

Fixes: dcdd43c41e60 ("net: vrf: performance improvements for IPv4")
Fixes: a9ec54d1b0cd ("net: vrf: performance improvements for IPv6")
Signed-off-by: Martin Willi <martin@strongswan.org>
Link: https://lore.kernel.org/r/20201106073030.3974927-1-martin@strongswan.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoNFS: Remove unnecessary inode lock in nfs_fsync_dir()
Trond Myklebust [Fri, 30 Oct 2020 21:57:30 +0000 (17:57 -0400)]
NFS: Remove unnecessary inode lock in nfs_fsync_dir()

nfs_inc_stats() is already thread-safe, and there are no other reasons
to hold the inode lock here.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
4 years agoNFS: Remove unnecessary inode locking in nfs_llseek_dir()
Trond Myklebust [Fri, 30 Oct 2020 21:57:29 +0000 (17:57 -0400)]
NFS: Remove unnecessary inode locking in nfs_llseek_dir()

Remove the contentious inode lock, and instead provide thread safety
using the file->f_lock spinlock.

Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
4 years agoNFS: Fix listxattr receive buffer size
Chuck Lever [Sat, 31 Oct 2020 16:44:25 +0000 (12:44 -0400)]
NFS: Fix listxattr receive buffer size

Certain NFSv4.2/RDMA tests fail with v5.9-rc1.

rpcrdma_convert_kvec() runs off the end of the rl_segments array
because rq_rcv_buf.tail[0].iov_len holds a very large positive
value. The resultant kernel memory corruption is enough to crash
the client system.

Callers of rpc_prepare_reply_pages() must reserve an extra XDR_UNIT
in the maximum decode size for a possible XDR pad of the contents
of the xdr_buf's pages. That guarantees the allocated receive buffer
will be large enough to accommodate the usual contents plus that XDR
pad word.

encode_op_hdr() cannot add that extra word. If it does,
xdr_inline_pages() underruns the length of the tail iovec.

Fixes: 3e1f02123fba ("NFSv4.2: add client side XDR handling for extended attributes")
Signed-off-by: Chuck Lever <chuck.lever@oracle.com>
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
4 years agoNFSv4.2: fix failure to unregister shrinker
J. Bruce Fields [Wed, 21 Oct 2020 14:34:15 +0000 (10:34 -0400)]
NFSv4.2: fix failure to unregister shrinker

We forgot to unregister the nfs4_xattr_large_entry_shrinker.

That leaves the global list of shrinkers corrupted after unload of the
nfs module, after which possibly unrelated code that calls
register_shrinker() or unregister_shrinker() gets a BUG() with
"supervisor write access in kernel mode".

And similarly for the nfs4_xattr_large_entry_lru.

Reported-by: Kris Karas <bugs-a17@moonlit-rail.com>
Tested-By: Kris Karas <bugs-a17@moonlit-rail.com>
Fixes: 95ad37f90c33 "NFSv4.2: add client side xattr caching."
Signed-off-by: J. Bruce Fields <bfields@redhat.com>
CC: stable@vger.kernel.org
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
4 years agoMerge branches 'acpi-scan', 'acpi-misc', 'acpi-button' and 'acpi-dptf'
Rafael J. Wysocki [Thu, 12 Nov 2020 15:11:48 +0000 (16:11 +0100)]
Merge branches 'acpi-scan', 'acpi-misc', 'acpi-button' and 'acpi-dptf'

* acpi-scan:
  ACPI: scan: Fix acpi_dma_configure_id() kerneldoc name

* acpi-misc:
  ACPI: GED: fix -Wformat
  ACPI: Fix whitespace inconsistencies

* acpi-button:
  ACPI: button: Add DMI quirk for Medion Akoya E2228T

* acpi-dptf:
  ACPI: DPTF: Support Alder Lake

4 years agomac80211: minstrel: fix tx status processing corner case
Felix Fietkau [Wed, 11 Nov 2020 18:33:59 +0000 (19:33 +0100)]
mac80211: minstrel: fix tx status processing corner case

Some drivers fill the status rate list without setting the rate index after
the final rate to -1. minstrel_ht already deals with this, but minstrel
doesn't, which causes it to get stuck at the lowest rate on these drivers.

Fix this by checking the count as well.

Cc: stable@vger.kernel.org
Fixes: cccf129f820e ("mac80211: add the 'minstrel' rate control algorithm")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20201111183359.43528-3-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
4 years agomac80211: minstrel: remove deferred sampling code
Felix Fietkau [Wed, 11 Nov 2020 18:33:58 +0000 (19:33 +0100)]
mac80211: minstrel: remove deferred sampling code

Deferring sampling attempts to the second stage has some bad interactions
with drivers that process the rate table in hardware and use the probe flag
to indicate probing packets (e.g. most mt76 drivers). On affected drivers
it can lead to probing not working at all.

If the link conditions turn worse, it might not be such a good idea to
do a lot of sampling for lower rates in this case.

Fix this by simply skipping the sample attempt instead of deferring it,
but keep the checks that would allow it to be sampled if it was skipped
too often, but only if it has less than 95% success probability.

Also ensure that IEEE80211_TX_CTL_RATE_CTRL_PROBE is set for all probing
packets.

Cc: stable@vger.kernel.org
Fixes: cccf129f820e ("mac80211: add the 'minstrel' rate control algorithm")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20201111183359.43528-2-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
4 years agomac80211: fix memory leak on filtered powersave frames
Felix Fietkau [Wed, 11 Nov 2020 18:33:57 +0000 (19:33 +0100)]
mac80211: fix memory leak on filtered powersave frames

After the status rework, ieee80211_tx_status_ext is leaking un-acknowledged
packets for stations in powersave mode.
To fix this, move the code handling those packets from __ieee80211_tx_status
into ieee80211_tx_status_ext

Reported-by: Tobias Waldvogel <tobias.waldvogel@gmail.com>
Fixes: 3318111cf63d ("mac80211: reduce duplication in tx status functions")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Link: https://lore.kernel.org/r/20201111183359.43528-1-nbd@nbd.name
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
4 years agorfkill: Fix use-after-free in rfkill_resume()
Claire Chang [Tue, 10 Nov 2020 08:49:08 +0000 (16:49 +0800)]
rfkill: Fix use-after-free in rfkill_resume()

If a device is getting removed or reprobed during resume, use-after-free
might happen. For example, h5_btrtl_resume() schedules a work queue for
device reprobing, which of course requires removal first.

If the removal happens in parallel with the device_resume() and wins the
race to acquire device_lock(), removal may remove the device from the PM
lists and all, but device_resume() is already running and will continue
when the lock can be acquired, thus calling rfkill_resume().

During this, if rfkill_set_block() is then called after the corresponding
*_unregister() and kfree() are called, there will be an use-after-free
in hci_rfkill_set_block():

BUG: KASAN: use-after-free in hci_rfkill_set_block+0x58/0xc0 [bluetooth]
...
Call trace:
  dump_backtrace+0x0/0x154
  show_stack+0x20/0x2c
  dump_stack+0xbc/0x12c
  print_address_description+0x88/0x4b0
  __kasan_report+0x144/0x168
  kasan_report+0x10/0x18
  check_memory_region+0x19c/0x1ac
  __kasan_check_write+0x18/0x24
  hci_rfkill_set_block+0x58/0xc0 [bluetooth]
  rfkill_set_block+0x9c/0x120
  rfkill_resume+0x34/0x70
  dpm_run_callback+0xf0/0x1f4
  device_resume+0x210/0x22c

Fix this by checking rfkill->registered in rfkill_resume(). device_del()
in rfkill_unregister() requires device_lock() and the whole rfkill_resume()
is also protected by the same lock via device_resume(), we can make sure
either the rfkill->registered is false before rfkill_resume() starts or the
rfkill device won't be unregistered before rfkill_resume() returns.

As async_resume() holds a reference to the device, at this level there can
be no use-after-free; only in the user that doesn't expect this scenario.

Fixes: 8589086f4efd ("Bluetooth: hci_h5: Turn off RTL8723BS on suspend, reprobe on resume")
Signed-off-by: Claire Chang <tientzu@chromium.org>
Link: https://lore.kernel.org/r/20201110084908.219088-1-tientzu@chromium.org
[edit commit message for clarity and add more info provided later]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
4 years agocosa: Add missing kfree in error path of cosa_write
Wang Hai [Tue, 10 Nov 2020 14:46:14 +0000 (22:46 +0800)]
cosa: Add missing kfree in error path of cosa_write

If memory allocation for 'kbuf' succeed, cosa_write() doesn't have a
corresponding kfree() in exception handling. Thus add kfree() for this
function implementation.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: Wang Hai <wanghai38@huawei.com>
Acked-by: Jan "Yenya" Kasprzak <kas@fi.muni.cz>
Link: https://lore.kernel.org/r/20201110144614.43194-1-wanghai38@huawei.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet: switch to the kernel.org patchwork instance
Jakub Kicinski [Tue, 10 Nov 2020 03:51:20 +0000 (19:51 -0800)]
net: switch to the kernel.org patchwork instance

Move to the kernel.org patchwork instance, it has significantly
lower latency for accessing from Europe and the US. Other quirks
include the reply bot.

Link: https://lore.kernel.org/r/20201110035120.642746-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoMerge branch 'cxgb4-ch_ktls-fixes-in-nic-tls-code'
Jakub Kicinski [Thu, 12 Nov 2020 00:30:41 +0000 (16:30 -0800)]
Merge branch 'cxgb4-ch_ktls-fixes-in-nic-tls-code'

Rohit Maheshwari says:

====================
cxgb4/ch_ktls: Fixes in nic tls code

This series helps in fixing multiple nic ktls issues. Series is broken
into 12 patches.

Patch 1 avoids deciding tls packet based on decrypted bit. If its a
retransmit packet which has tls handshake and finish (for encryption),
decrypted bit won't be set there, and so we can't rely on decrypted
bit.

Patch 2 helps supporting linear skb. SKBs were assumed non-linear.
Corrected the length extraction.

Patch 3 fixes the checksum offload update in WR.

Patch 4 fixes kernel panic happening due to creating new skb for each
record. As part of fix driver will use same skb to send out one tls
record (partial data) of the same SKB.

Patch 5 fixes the problem of skb data length smaller than remaining data
of the record.

Patch 6 fixes the handling of SKBs which has tls header alone pkt, but
not starting from beginning.

Patch 7 avoids sending extra data which is used to make a record 16 byte
aligned. We don't need to retransmit those extra few bytes.

Patch 8 handles the cases where retransmit packet has tls starting
exchanges which are prior to tls start marker.

Patch 9 fixes the problem os skb free before HW knows about tcp FIN.

Patch 10 handles the small packet case which has partial TAG bytes only.
HW can't handle those, hence using sw crypto for such pkts.

Patch 11 corrects the potential tcb update problem.

Patch 12 stops the queue if queue reaches threshold value.

v1->v2:
- Corrected fixes tag issue.
- Marked chcr_ktls_sw_fallback() static.

v2->v3:
- Replaced GFP_KERNEL with GFP_ATOMIC.
- Removed mixed fixes.

v3->v4:
- Corrected fixes tag issue.

v4->v5:
- Separated mixed fixes from patch 4.

v5-v6:
- Fixes tag should be at the end.
====================

Link: https://lore.kernel.org/r/20201109105142.15398-1-rohitm@chelsio.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: stop the txq if reaches threshold
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:42 +0000 (16:21 +0530)]
ch_ktls: stop the txq if reaches threshold

Stop the queue and ask for the credits if queue reaches to
threashold.

Fixes: 5a4b9fe7fece ("cxgb4/chcr: complete record tx handling")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: tcb update fails sometimes
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:41 +0000 (16:21 +0530)]
ch_ktls: tcb update fails sometimes

context id and port id should be filled while sending tcb update.

Fixes: 5a4b9fe7fece ("cxgb4/chcr: complete record tx handling")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls/cxgb4: handle partial tag alone SKBs
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:40 +0000 (16:21 +0530)]
ch_ktls/cxgb4: handle partial tag alone SKBs

If TCP congestion caused a very small packets which only has some
part fo the TAG, and that too is not till the end. HW can't handle
such case, so falling back to sw crypto in such cases.

v1->v2:
- Marked chcr_ktls_sw_fallback() static.

Fixes: dc05f3df8fac ("chcr: Handle first or middle part of record")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: don't free skb before sending FIN
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:39 +0000 (16:21 +0530)]
ch_ktls: don't free skb before sending FIN

If its a last packet and fin is set. Make sure FIN is informed
to HW before skb gets freed.

Fixes: 429765a149f1 ("chcr: handle partial end part of a record")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: packet handling prior to start marker
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:38 +0000 (16:21 +0530)]
ch_ktls: packet handling prior to start marker

There could be a case where ACK for tls exchanges prior to start
marker is missed out, and by the time tls is offloaded. This pkt
should not be discarded and handled carefully. It could be
plaintext alone or plaintext + finish as well.

Fixes: 5a4b9fe7fece ("cxgb4/chcr: complete record tx handling")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: Correction in middle record handling
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:37 +0000 (16:21 +0530)]
ch_ktls: Correction in middle record handling

If a record starts in middle, reset TCB UNA so that we could
avoid sending out extra packet which is needed to make it 16
byte aligned to start AES CTR.
Check also considers prev_seq, which should be what is
actually sent, not the skb data length.
Avoid updating partial TAG to HW at any point of time, that's
why we need to check if remaining part is smaller than TAG
size, then reset TX_MAX to be TAG starting sequence number.

Fixes: 5a4b9fe7fece ("cxgb4/chcr: complete record tx handling")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: missing handling of header alone
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:36 +0000 (16:21 +0530)]
ch_ktls: missing handling of header alone

If an skb has only header part which doesn't start from
beginning, is not being handled properly.

Fixes: dc05f3df8fac ("chcr: Handle first or middle part of record")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: Correction in trimmed_len calculation
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:35 +0000 (16:21 +0530)]
ch_ktls: Correction in trimmed_len calculation

trimmed length calculation goes wrong if skb has only tag part
to send. It should be zero if there is no data bytes apart from
TAG.

Fixes: dc05f3df8fac ("chcr: Handle first or middle part of record")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agocxgb4/ch_ktls: creating skbs causes panic
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:34 +0000 (16:21 +0530)]
cxgb4/ch_ktls: creating skbs causes panic

Creating SKB per tls record and freeing the original one causes
panic. There will be race if connection reset is requested. By
freeing original skb, refcnt will be decremented and that means,
there is no pending record to send, and so tls_dev_del will be
requested in control path while SKB of related connection is in
queue.
 Better approach is to use same SKB to send one record (partial
data) at a time. We still have to create a new SKB when partial
last part of a record is requested.
 This fix introduces new API cxgb4_write_partial_sgl() to send
partial part of skb. Present cxgb4_write_sgl can only provide
feasibility to start from an offset which limits to header only
and it can write sgls for the whole skb len. But this new API
will help in both. It can start from any offset and can end
writing in middle of the skb.

v4->v5:
- Removed extra changes.

Fixes: 429765a149f1 ("chcr: handle partial end part of a record")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: Update cheksum information
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:33 +0000 (16:21 +0530)]
ch_ktls: Update cheksum information

Checksum update was missing in the WR.

Fixes: 429765a149f1 ("chcr: handle partial end part of a record")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoch_ktls: Correction in finding correct length
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:32 +0000 (16:21 +0530)]
ch_ktls: Correction in finding correct length

There is a possibility of linear skbs coming in. Correcting
the length extraction logic.

v2->v3:
- Separated un-related changes from this patch.

Fixes: 5a4b9fe7fece ("cxgb4/chcr: complete record tx handling")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agocxgb4/ch_ktls: decrypted bit is not enough
Rohit Maheshwari [Mon, 9 Nov 2020 10:51:31 +0000 (16:21 +0530)]
cxgb4/ch_ktls: decrypted bit is not enough

If skb has retransmit data starting before start marker, e.g. ccs,
decrypted bit won't be set for that, and if it has some data to
encrypt, then it must be given to crypto ULD. So in place of
decrypted, check if socket is tls offloaded. Also, unless skb has
some data to encrypt, no need to give it for tls offload handling.

v2->v3:
- Removed ifdef.

Fixes: 5a4b9fe7fece ("cxgb4/chcr: complete record tx handling")
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agonet/x25: Fix null-ptr-deref in x25_connect
Martin Schiller [Mon, 9 Nov 2020 06:54:49 +0000 (07:54 +0100)]
net/x25: Fix null-ptr-deref in x25_connect

This fixes a regression for blocking connects introduced by commit
4becb7ee5b3d ("net/x25: Fix x25_neigh refcnt leak when x25 disconnect").

The x25->neighbour is already set to "NULL" by x25_disconnect() now,
while a blocking connect is waiting in
x25_wait_for_connection_establishment(). Therefore x25->neighbour must
not be accessed here again and x25->state is also already set to
X25_STATE_0 by x25_disconnect().

Fixes: 4becb7ee5b3d ("net/x25: Fix x25_neigh refcnt leak when x25 disconnect")
Signed-off-by: Martin Schiller <ms@dev.tdt.de>
Reviewed-by: Xie He <xie.he.0141@gmail.com>
Link: https://lore.kernel.org/r/20201109065449.9014-1-ms@dev.tdt.de
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
4 years agoarm64: dts: fsl-ls1028a-kontron-sl28: specify in-band mode for ENETC
Michael Walle [Mon, 9 Nov 2020 11:04:36 +0000 (12:04 +0100)]
arm64: dts: fsl-ls1028a-kontron-sl28: specify in-band mode for ENETC

Since commit 71b77a7a27a3 ("enetc: Migrate to PHYLINK and PCS_LYNX") the
network port of the Kontron sl28 board is broken. After the migration to
phylink the device tree has to specify the in-band-mode property. Add
it.

Fixes: 71b77a7a27a3 ("enetc: Migrate to PHYLINK and PCS_LYNX")
Suggested-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Signed-off-by: Michael Walle <michael@walle.cc>
Reviewed-by: Vladimir Oltean <vladimir.oltean@nxp.com>
Link: https://lore.kernel.org/r/20201109110436.5906-1-michael@walle.cc
Signed-off-by: Jakub Kicinski <kuba@kernel.org>