platform/kernel/linux-starfive.git
18 months agoBluetooth: MGMT: Fix error report for ADD_EXT_ADV_PARAMS
Inga Stotland [Sat, 22 Oct 2022 00:48:56 +0000 (17:48 -0700)]
Bluetooth: MGMT: Fix error report for ADD_EXT_ADV_PARAMS

[ Upstream commit 3b1c7c00b8c22b3cb79532252c59eb0b287bb86d ]

When validating the parameter length for MGMT_OP_ADD_EXT_ADV_PARAMS
command, use the correct op code in error status report:
was MGMT_OP_ADD_ADVERTISING, changed to MGMT_OP_ADD_EXT_ADV_PARAMS.

Fixes: 12410572833a2 ("Bluetooth: Break add adv into two mgmt commands")
Signed-off-by: Inga Stotland <inga.stotland@intel.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoBluetooth: hci_core: fix error handling in hci_register_dev()
Yang Yingliang [Thu, 20 Oct 2022 02:16:56 +0000 (10:16 +0800)]
Bluetooth: hci_core: fix error handling in hci_register_dev()

[ Upstream commit 0d75da38e060d21f948b3df5f5e349c962cf1ed2 ]

If hci_register_suspend_notifier() returns error, the hdev and rfkill
are leaked. We could disregard the error and print a warning message
instead to avoid leaks, as it just means we won't be handing suspend
requests.

Fixes: 9952d90ea288 ("Bluetooth: Handle PM_SUSPEND_PREPARE and PM_POST_SUSPEND")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agosctp: sysctl: make extra pointers netns aware
Firo Yang [Fri, 9 Dec 2022 05:48:54 +0000 (13:48 +0800)]
sctp: sysctl: make extra pointers netns aware

[ Upstream commit da05cecc4939c0410d56c29e252998b192756318 ]

Recently, a customer reported that from their container whose
net namespace is different to the host's init_net, they can't set
the container's net.sctp.rto_max to any value smaller than
init_net.sctp.rto_min.

For instance,
Host:
sudo sysctl net.sctp.rto_min
net.sctp.rto_min = 1000

Container:
echo 100 > /mnt/proc-net/sctp/rto_min
echo 400 > /mnt/proc-net/sctp/rto_max
echo: write error: Invalid argument

This is caused by the check made from this'commit 4f3fdf3bc59c
("sctp: add check rto_min and rto_max in sysctl")'
When validating the input value, it's always referring the boundary
value set for the init_net namespace.

Having container's rto_max smaller than host's init_net.sctp.rto_min
does make sense. Consider that the rto between two containers on the
same host is very likely smaller than it for two hosts.

So to fix this problem, as suggested by Marcelo, this patch makes the
extra pointers of rto_min, rto_max, pf_retrans, and ps_retrans point
to the corresponding variables from the newly created net namespace while
the new net namespace is being registered in sctp_sysctl_net_register.

Fixes: 4f3fdf3bc59c ("sctp: add check rto_min and rto_max in sysctl")
Reviewed-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Reviewed-by: Jakub Kicinski <kuba@kernel.org>
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
Signed-off-by: Firo Yang <firo.yang@suse.com>
Link: https://lore.kernel.org/r/20221209054854.23889-1-firo.yang@suse.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agontb_netdev: Use dev_kfree_skb_any() in interrupt context
Eric Pilmore [Fri, 9 Dec 2022 00:06:59 +0000 (16:06 -0800)]
ntb_netdev: Use dev_kfree_skb_any() in interrupt context

[ Upstream commit 5f7d78b2b12a9d561f48fa00bab29b40f4616dad ]

TX/RX callback handlers (ntb_netdev_tx_handler(),
ntb_netdev_rx_handler()) can be called in interrupt
context via the DMA framework when the respective
DMA operations have completed. As such, any calls
by these routines to free skb's, should use the
interrupt context safe dev_kfree_skb_any() function.

Previously, these callback handlers would call the
interrupt unsafe version of dev_kfree_skb(). This has
not presented an issue on Intel IOAT DMA engines as
that driver utilizes tasklets rather than a hard
interrupt handler, like the AMD PTDMA DMA driver.
On AMD systems, a kernel WARNING message is
encountered, which is being issued from
skb_release_head_state() due to in_hardirq()
being true.

Besides the user visible WARNING from the kernel,
the other symptom of this bug was that TCP/IP performance
across the ntb_netdev interface was very poor, i.e.
approximately an order of magnitude below what was
expected. With the repair to use dev_kfree_skb_any(),
kernel WARNINGs from skb_release_head_state() ceased
and TCP/IP performance, as measured by iperf, was on
par with expected results, approximately 20 Gb/s on
AMD Milan based server. Note that this performance
is comparable with Intel based servers.

Fixes: 765ccc7bc3d91 ("ntb_netdev: correct skb leak")
Fixes: 548c237c0a997 ("net: Add support for NTB virtual ethernet device")
Signed-off-by: Eric Pilmore <epilmore@gigaio.com>
Reviewed-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/20221209000659.8318-1-epilmore@gigaio.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: lan9303: Fix read error execution path
Jerry Ray [Fri, 9 Dec 2022 15:35:02 +0000 (09:35 -0600)]
net: lan9303: Fix read error execution path

[ Upstream commit 8964916d206071b058c6351f88b1966bd58cbde0 ]

This patch fixes an issue where a read failure of a port statistic counter
will return unknown results.  While it is highly unlikely the read will
ever fail, it is much cleaner to return a zero for the stat count.

Fixes: a1292595e006 ("net: dsa: add new DSA switch driver for the SMSC-LAN9303")
Signed-off-by: Jerry Ray <jerry.ray@microchip.com>
Reviewed-by: Vladimir Oltean <olteanv@gmail.com>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Link: https://lore.kernel.org/r/20221209153502.7429-1-jerry.ray@microchip.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: ethernet: ti: am65-cpsw: Fix PM runtime leakage in am65_cpsw_nuss_ndo_slave_open()
Roger Quadros [Thu, 8 Dec 2022 10:55:34 +0000 (12:55 +0200)]
net: ethernet: ti: am65-cpsw: Fix PM runtime leakage in am65_cpsw_nuss_ndo_slave_open()

[ Upstream commit 5821504f5073983733465b8bc430049c4343bbd7 ]

Ensure pm_runtime_put() is issued in error path.

Reported-by: Jakub Kicinski <kuba@kernel.org>
Fixes: 93a76530316a ("net: ethernet: ti: introduce am65x/j721e gigabit eth subsystem driver")
Signed-off-by: Roger Quadros <rogerq@kernel.org>
Reviewed-by: Saeed Mahameed <saeed@kernel.org>
Link: https://lore.kernel.org/r/20221208105534.63709-1-rogerq@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agocan: tcan4x5x: Fix use of register error status mask
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:26 +0000 (12:57 +0100)]
can: tcan4x5x: Fix use of register error status mask

[ Upstream commit 67727a17a6b375d68fe569b77e6516b034b834c0 ]

TCAN4X5X_ERROR_STATUS is not a status register that needs clearing
during interrupt handling. Instead this is a masking register that masks
error interrupts. Writing TCAN4X5X_CLEAR_ALL_INT to this register
effectively masks everything.

Rename the register and mask all error interrupts only once by writing
to the register in tcan4x5x_init.

Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-10-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agocan: m_can: Call the RAM init directly from m_can_chip_config
Vivek Yadav [Wed, 7 Dec 2022 10:06:31 +0000 (15:36 +0530)]
can: m_can: Call the RAM init directly from m_can_chip_config

[ Upstream commit eaacfeaca7ad0804b9a6eff7afeba93a87db7638 ]

When we try to access the mcan message ram addresses during the probe,
hclk is gated by any other drivers or disabled, because of that probe
gets failed.

Move the mram init functionality to mcan chip config called by
m_can_start from mcan open function, by that time clocks are
enabled.

Suggested-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Vivek Yadav <vivek.2311@samsung.com>
Link: https://lore.kernel.org/all/20221207100632.96200-2-vivek.2311@samsung.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Stable-dep-of: 67727a17a6b3 ("can: tcan4x5x: Fix use of register error status mask")
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agocan: tcan4x5x: Remove invalid write in clear_interrupts
Markus Schneider-Pargmann [Tue, 6 Dec 2022 11:57:25 +0000 (12:57 +0100)]
can: tcan4x5x: Remove invalid write in clear_interrupts

[ Upstream commit 40c9e4f676abbe194541d88e796341c92d5a13c0 ]

Register 0x824 TCAN4X5X_MCAN_INT_REG is a read-only register. Any writes
to this register do not have any effect.

Remove this write. The m_can driver aldready clears the interrupts in
m_can_isr() by writing to M_CAN_IR which is translated to register
0x1050 which is a writable version of this register.

Fixes: 5443c226ba91 ("can: tcan4x5x: Add tcan4x5x driver to the kernel")
Signed-off-by: Markus Schneider-Pargmann <msp@baylibre.com>
Link: https://lore.kernel.org/all/20221206115728.1056014-9-msp@baylibre.com
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: amd-xgbe: Check only the minimum speed for active/passive cables
Tom Lendacky [Thu, 8 Dec 2022 16:22:25 +0000 (10:22 -0600)]
net: amd-xgbe: Check only the minimum speed for active/passive cables

[ Upstream commit f8ab263d4d48e6dab752029bf562f20a2ee630ed ]

There are cables that exist that can support speeds in excess of 10GbE.
The driver, however, restricts the EEPROM advertised nominal bitrate to
a specific range, which can prevent usage of cables that can support,
for example, up to 25GbE.

Rather than checking that an active or passive cable supports a specific
range, only check for a minimum supported speed.

Fixes: abf0a1c2b26a ("amd-xgbe: Add support for SFP+ modules")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: amd-xgbe: Fix logic around active and passive cables
Tom Lendacky [Thu, 8 Dec 2022 16:22:24 +0000 (10:22 -0600)]
net: amd-xgbe: Fix logic around active and passive cables

[ Upstream commit 4998006c73afe44e2f639d55bd331c6c26eb039f ]

SFP+ active and passive cables are copper cables with fixed SFP+ end
connectors. Due to a misinterpretation of this, SFP+ active cables could
end up not being recognized, causing the driver to fail to establish a
connection.

Introduce a new enum in SFP+ cable types, XGBE_SFP_CABLE_FIBER, that is
the default cable type, and handle active and passive cables when they are
specifically detected.

Fixes: abf0a1c2b26a ("amd-xgbe: Add support for SFP+ modules")
Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoaf_unix: call proto_unregister() in the error path in af_unix_init()
Yang Yingliang [Thu, 8 Dec 2022 15:01:58 +0000 (23:01 +0800)]
af_unix: call proto_unregister() in the error path in af_unix_init()

[ Upstream commit 73e341e0281a35274629e9be27eae2f9b1b492bf ]

If register unix_stream_proto returns error, unix_dgram_proto needs
be unregistered.

Fixes: 94531cfcbe79 ("af_unix: Add unix_stream_proto for sockmap")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Simon Horman <simon.horman@corigine.com>
Reviewed-by: Kuniyuki Iwashima <kuniyu@amazon.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: setsockopt: fix IPV6_UNICAST_IF option for connected sockets
Richard Gobert [Thu, 8 Dec 2022 14:54:46 +0000 (15:54 +0100)]
net: setsockopt: fix IPV6_UNICAST_IF option for connected sockets

[ Upstream commit 526682b458b1b56d2e0db027df535cb5cdcfde59 ]

Change the behaviour of ip6_datagram_connect to consider the interface
set by the IPV6_UNICAST_IF socket option, similarly to udpv6_sendmsg.

This change is the IPv6 counterpart of the fix for IP_UNICAST_IF.
The tests introduced by that patch showed that the incorrect
behavior is present in IPv6 as well.
This patch fixes the broken test.

Reported-by: kernel test robot <oliver.sang@intel.com>
Link: https://lore.kernel.org/r/202210062117.c7eef1a3-oliver.sang@intel.com
Fixes: 0e4d354762ce ("net-next: Fix IP_UNICAST_IF option behavior for connected sockets")

Signed-off-by: Richard Gobert <richardbgobert@gmail.com>
Reviewed-by: David Ahern <dsahern@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: amd: lance: don't call dev_kfree_skb() under spin_lock_irqsave()
Yang Yingliang [Thu, 8 Dec 2022 14:21:47 +0000 (22:21 +0800)]
net: amd: lance: don't call dev_kfree_skb() under spin_lock_irqsave()

[ Upstream commit 6151d105dfce8c23edf30eed35e97f3d9b96a35c ]

It is not allowed to call kfree_skb() or consume_skb() from hardware
interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead.
The difference between them is free reason, dev_kfree_skb_irq() means
the SKB is dropped in error and dev_consume_skb_irq() means the SKB
is consumed in normal.

In these two cases, dev_kfree_skb() is called consume the xmited SKB,
so replace it with dev_consume_skb_irq().

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohamradio: don't call dev_kfree_skb() under spin_lock_irqsave()
Yang Yingliang [Thu, 8 Dec 2022 14:21:46 +0000 (22:21 +0800)]
hamradio: don't call dev_kfree_skb() under spin_lock_irqsave()

[ Upstream commit 3727f742915f04f6fc550b80cf406999bd4e90d0 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware
interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead.
The difference between them is free reason, dev_kfree_skb_irq() means
the SKB is dropped in error and dev_consume_skb_irq() means the SKB
is consumed in normal.

In scc_discard_buffers(), dev_kfree_skb() is called to discard the SKBs,
so replace it with dev_kfree_skb_irq().

In scc_net_tx(), dev_kfree_skb() is called to drop the SKB that exceed
queue length, so replace it with dev_kfree_skb_irq().

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: ethernet: dnet: don't call dev_kfree_skb() under spin_lock_irqsave()
Yang Yingliang [Thu, 8 Dec 2022 14:21:45 +0000 (22:21 +0800)]
net: ethernet: dnet: don't call dev_kfree_skb() under spin_lock_irqsave()

[ Upstream commit f07fadcbee2a5e84caa67c7c445424200bffb60b ]

It is not allowed to call kfree_skb() or consume_skb() from hardware
interrupt context or with hardware interrupts being disabled.

In this case, the lock is used to protected 'bp', so we can move
dev_kfree_skb() after the spin_unlock_irqrestore().

Fixes: 4796417417a6 ("dnet: Dave DNET ethernet controller driver (updated)")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: emaclite: don't call dev_kfree_skb() under spin_lock_irqsave()
Yang Yingliang [Thu, 8 Dec 2022 14:21:44 +0000 (22:21 +0800)]
net: emaclite: don't call dev_kfree_skb() under spin_lock_irqsave()

[ Upstream commit d1678bf45f21fa5ae4a456f821858679556ea5f8 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware
interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead.
The difference between them is free reason, dev_kfree_skb_irq() means
the SKB is dropped in error and dev_consume_skb_irq() means the SKB
is consumed in normal.

In this case, dev_kfree_skb() is called in xemaclite_tx_timeout() to
drop the SKB, when tx timeout, so replace it with dev_kfree_skb_irq().

Fixes: bb81b2ddfa19 ("net: add Xilinx emac lite device driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: apple: bmac: don't call dev_kfree_skb() under spin_lock_irqsave()
Yang Yingliang [Thu, 8 Dec 2022 13:37:35 +0000 (21:37 +0800)]
net: apple: bmac: don't call dev_kfree_skb() under spin_lock_irqsave()

[ Upstream commit 5fe02e046e6422c4adfdbc50206ec7186077da24 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware
interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead.
The difference between them is free reason, dev_kfree_skb_irq() means
the SKB is dropped in error and dev_consume_skb_irq() means the SKB
is consumed in normal.

In this case, dev_kfree_skb() is called in bmac_tx_timeout() to drop
the SKB, when tx timeout, so replace it with dev_kfree_skb_irq().

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: apple: mace: don't call dev_kfree_skb() under spin_lock_irqsave()
Yang Yingliang [Thu, 8 Dec 2022 13:37:34 +0000 (21:37 +0800)]
net: apple: mace: don't call dev_kfree_skb() under spin_lock_irqsave()

[ Upstream commit 3dfe3486c1cd4f82b466b7d307f23777137b8acc ]

It is not allowed to call kfree_skb() or consume_skb() from hardware
interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead.
The difference between them is free reason, dev_kfree_skb_irq() means
the SKB is dropped in error and dev_consume_skb_irq() means the SKB
is consumed in normal.

In this case, dev_kfree_skb() is called in mace_tx_timeout() to drop
the SKB, when tx timeout, so replace it with dev_kfree_skb_irq().

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet/tunnel: wait until all sk_user_data reader finish before releasing the sock
Hangbin Liu [Thu, 8 Dec 2022 12:04:52 +0000 (20:04 +0800)]
net/tunnel: wait until all sk_user_data reader finish before releasing the sock

[ Upstream commit 3cf7203ca620682165706f70a1b12b5194607dce ]

There is a race condition in vxlan that when deleting a vxlan device
during receiving packets, there is a possibility that the sock is
released after getting vxlan_sock vs from sk_user_data. Then in
later vxlan_ecn_decapsulate(), vxlan_get_sk_family() we will got
NULL pointer dereference. e.g.

   #0 [ffffa25ec6978a38] machine_kexec at ffffffff8c669757
   #1 [ffffa25ec6978a90] __crash_kexec at ffffffff8c7c0a4d
   #2 [ffffa25ec6978b58] crash_kexec at ffffffff8c7c1c48
   #3 [ffffa25ec6978b60] oops_end at ffffffff8c627f2b
   #4 [ffffa25ec6978b80] page_fault_oops at ffffffff8c678fcb
   #5 [ffffa25ec6978bd8] exc_page_fault at ffffffff8d109542
   #6 [ffffa25ec6978c00] asm_exc_page_fault at ffffffff8d200b62
      [exception RIP: vxlan_ecn_decapsulate+0x3b]
      RIP: ffffffffc1014e7b  RSP: ffffa25ec6978cb0  RFLAGS: 00010246
      RAX: 0000000000000008  RBX: ffff8aa000888000  RCX: 0000000000000000
      RDX: 000000000000000e  RSI: ffff8a9fc7ab803e  RDI: ffff8a9fd1168700
      RBP: ffff8a9fc7ab803e   R8: 0000000000700000   R9: 00000000000010ae
      R10: ffff8a9fcb748980  R11: 0000000000000000  R12: ffff8a9fd1168700
      R13: ffff8aa000888000  R14: 00000000002a0000  R15: 00000000000010ae
      ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018
   #7 [ffffa25ec6978ce8] vxlan_rcv at ffffffffc10189cd [vxlan]
   #8 [ffffa25ec6978d90] udp_queue_rcv_one_skb at ffffffff8cfb6507
   #9 [ffffa25ec6978dc0] udp_unicast_rcv_skb at ffffffff8cfb6e45
  #10 [ffffa25ec6978dc8] __udp4_lib_rcv at ffffffff8cfb8807
  #11 [ffffa25ec6978e20] ip_protocol_deliver_rcu at ffffffff8cf76951
  #12 [ffffa25ec6978e48] ip_local_deliver at ffffffff8cf76bde
  #13 [ffffa25ec6978ea0] __netif_receive_skb_one_core at ffffffff8cecde9b
  #14 [ffffa25ec6978ec8] process_backlog at ffffffff8cece139
  #15 [ffffa25ec6978f00] __napi_poll at ffffffff8ceced1a
  #16 [ffffa25ec6978f28] net_rx_action at ffffffff8cecf1f3
  #17 [ffffa25ec6978fa0] __softirqentry_text_start at ffffffff8d4000ca
  #18 [ffffa25ec6978ff0] do_softirq at ffffffff8c6fbdc3

Reproducer: https://github.com/Mellanox/ovs-tests/blob/master/test-ovs-vxlan-remove-tunnel-during-traffic.sh

Fix this by waiting for all sk_user_data reader to finish before
releasing the sock.

Reported-by: Jianlin Shi <jishi@redhat.com>
Suggested-by: Jakub Sitnicki <jakub@cloudflare.com>
Fixes: 6a93cc905274 ("udp-tunnel: Add a few more UDP tunnel APIs")
Signed-off-by: Hangbin Liu <liuhangbin@gmail.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: farsync: Fix kmemleak when rmmods farsync
Li Zetao [Thu, 8 Dec 2022 12:05:40 +0000 (20:05 +0800)]
net: farsync: Fix kmemleak when rmmods farsync

[ Upstream commit 2f623aaf9f31de968dea6169849706a2f9be444c ]

There are two memory leaks reported by kmemleak:

  unreferenced object 0xffff888114b20200 (size 128):
    comm "modprobe", pid 4846, jiffies 4295146524 (age 401.345s)
    hex dump (first 32 bytes):
      e0 62 57 09 81 88 ff ff e0 62 57 09 81 88 ff ff  .bW......bW.....
      01 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
    backtrace:
      [<ffffffff815bcd82>] kmalloc_trace+0x22/0x60
      [<ffffffff83d35c78>] __hw_addr_add_ex+0x198/0x6c0
      [<ffffffff83d3989d>] dev_addr_init+0x13d/0x230
      [<ffffffff83d1063d>] alloc_netdev_mqs+0x10d/0xe50
      [<ffffffff82b4a06e>] alloc_hdlcdev+0x2e/0x80
      [<ffffffffa016a741>] fst_add_one+0x601/0x10e0 [farsync]
      ...

  unreferenced object 0xffff88810b85b000 (size 1024):
    comm "modprobe", pid 4846, jiffies 4295146523 (age 401.346s)
    hex dump (first 32 bytes):
      00 00 b0 02 00 c9 ff ff 00 70 0a 00 00 c9 ff ff  .........p......
      00 00 00 f2 00 00 00 f3 0a 00 00 00 02 00 00 00  ................
    backtrace:
      [<ffffffff815bcd82>] kmalloc_trace+0x22/0x60
      [<ffffffffa016a294>] fst_add_one+0x154/0x10e0 [farsync]
      [<ffffffff82060e83>] local_pci_probe+0xd3/0x170
      ...

The root cause is traced to the netdev and fst_card_info are not freed
when removes one fst in fst_remove_one(), which may trigger oom if
repeated insmod and rmmod module.

Fix it by adding free_netdev() and kfree() in fst_remove_one(), just as
the operations on the error handling path in fst_add_one().

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Li Zetao <lizetao1@huawei.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoethernet: s2io: don't call dev_kfree_skb() under spin_lock_irqsave()
Yang Yingliang [Thu, 8 Dec 2022 12:01:21 +0000 (20:01 +0800)]
ethernet: s2io: don't call dev_kfree_skb() under spin_lock_irqsave()

[ Upstream commit 6cee96e09df54ae17784c0f38a49e0ed8229b825 ]

It is not allowed to call kfree_skb() or consume_skb() from hardware
interrupt context or with hardware interrupts being disabled.

It should use dev_kfree_skb_irq() or dev_consume_skb_irq() instead.
The difference between them is free reason, dev_kfree_skb_irq() means
the SKB is dropped in error and dev_consume_skb_irq() means the SKB
is consumed in normal.

In this case, dev_kfree_skb() is called in free_tx_buffers() to drop
the SKBs in tx buffers, when the card is down, so replace it with
dev_kfree_skb_irq() here.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoof: overlay: fix null pointer dereferencing in find_dup_cset_node_entry() and find_du...
ruanjinjie [Sun, 11 Dec 2022 02:33:37 +0000 (10:33 +0800)]
of: overlay: fix null pointer dereferencing in find_dup_cset_node_entry() and find_dup_cset_prop()

[ Upstream commit ee9d7a0e754568180a2f8ebc4aad226278a9116f ]

When kmalloc() fail to allocate memory in kasprintf(), fn_1 or fn_2 will
be NULL, and strcmp() will cause null pointer dereference.

Fixes: 2fe0e8769df9 ("of: overlay: check prevents multiple fragments touching same property")
Signed-off-by: ruanjinjie <ruanjinjie@huawei.com>
Link: https://lore.kernel.org/r/20221211023337.592266-1-ruanjinjie@huawei.com
Signed-off-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoipvs: use u64_stats_t for the per-cpu counters
Julian Anastasov [Tue, 22 Nov 2022 16:46:01 +0000 (18:46 +0200)]
ipvs: use u64_stats_t for the per-cpu counters

[ Upstream commit 1dbd8d9a82e3f26b9d063292d47ece673f48fce2 ]

Use the provided u64_stats_t type to avoid
load/store tearing.

Fixes: 316580b69d0a ("u64_stats: provide u64_stats_t type")
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Cc: yunhong-cgl jiang <xintian1976@gmail.com>
Cc: "dust.li" <dust.li@linux.alibaba.com>
Reviewed-by: Jiri Wiesner <jwiesner@suse.de>
Tested-by: Jiri Wiesner <jwiesner@suse.de>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agodrivers: net: qlcnic: Fix potential memory leak in qlcnic_sriov_init()
Yuan Can [Wed, 7 Dec 2022 08:54:10 +0000 (08:54 +0000)]
drivers: net: qlcnic: Fix potential memory leak in qlcnic_sriov_init()

[ Upstream commit 01de1123322e4fe1bbd0fcdf0982511b55519c03 ]

If vp alloc failed in qlcnic_sriov_init(), all previously allocated vp
needs to be freed.

Fixes: f197a7aa6288 ("qlcnic: VF-PF communication channel implementation")
Signed-off-by: Yuan Can <yuancan@huawei.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: stmmac: fix possible memory leak in stmmac_dvr_probe()
Gaosheng Cui [Wed, 7 Dec 2022 08:34:13 +0000 (16:34 +0800)]
net: stmmac: fix possible memory leak in stmmac_dvr_probe()

[ Upstream commit a137f3f27f9290933fe7e40e6dc8a445781c31a2 ]

The bitmap_free() should be called to free priv->af_xdp_zc_qps
when create_singlethread_workqueue() fails, otherwise there will
be a memory leak, so we add the err path error_wq_init to fix it.

Fixes: bba2556efad6 ("net: stmmac: Enable RX via AF_XDP zero-copy")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Reviewed-by: Leon Romanovsky <leonro@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: stmmac: selftests: fix potential memleak in stmmac_test_arpoffload()
Zhang Changzhong [Wed, 7 Dec 2022 08:31:59 +0000 (16:31 +0800)]
net: stmmac: selftests: fix potential memleak in stmmac_test_arpoffload()

[ Upstream commit f150b63f3fa5fdd81e0dd6151e8850268e29438c ]

The skb allocated by stmmac_test_get_arp_skb() hasn't been released in
some error handling case, which will lead to a memory leak. Fix this up
by adding kfree_skb() to release skb.

Compile tested only.

Fixes: 5e3fb0a6e2b3 ("net: stmmac: selftests: Implement the ARP Offload test")
Signed-off-by: Zhang Changzhong <zhangchangzhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: defxx: Fix missing err handling in dfx_init()
Yongqiang Liu [Wed, 7 Dec 2022 07:20:45 +0000 (07:20 +0000)]
net: defxx: Fix missing err handling in dfx_init()

[ Upstream commit ae18dcdff0f8d7e84cd3fd9f496518b5e72d185d ]

When eisa_driver_register() or tc_register_driver() failed,
the modprobe defxx would fail with some err log as follows:

 Error: Driver 'defxx' is already registered, aborting...

Fix this issue by adding err hanling in dfx_init().

Fixes: e89a2cfb7d7b5 ("[TC] defxx: TURBOchannel support")
Signed-off-by: Yongqiang Liu <liuyongqiang13@huawei.com>
Reviewed-by: Jiri Pirko <jiri@nvidia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonet: vmw_vsock: vmci: Check memcpy_from_msg()
Artem Chernyshev [Tue, 6 Dec 2022 06:58:34 +0000 (09:58 +0300)]
net: vmw_vsock: vmci: Check memcpy_from_msg()

[ Upstream commit 44aa5a6dba8283bfda28b1517af4de711c5652a4 ]

vmci_transport_dgram_enqueue() does not check the return value
of memcpy_from_msg().  If memcpy_from_msg() fails, it is possible that
uninitialized memory contents are sent unintentionally instead of user's
message in the datagram to the destination.  Return with an error if
memcpy_from_msg() fails.

Found by Linux Verification Center (linuxtesting.org) with SVACE.

Fixes: 0f7db23a07af ("vmci_transport: switch ->enqeue_dgram, ->enqueue_stream and ->dequeue_stream to msghdr")
Signed-off-by: Artem Chernyshev <artem.chernyshev@red-soft.ru>
Reviewed-by: Stefano Garzarella <sgarzare@redhat.com>
Reviewed-by: Vishnu Dasa <vdasa@vmware.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoclk: socfpga: Fix memory leak in socfpga_gate_init()
Xiu Jianfeng [Wed, 23 Nov 2022 03:16:22 +0000 (11:16 +0800)]
clk: socfpga: Fix memory leak in socfpga_gate_init()

[ Upstream commit 0b8ba891ad4d1ef6bfa4c72efc83f9f9f855f68b ]

Free @socfpga_clk and @ops on the error path to avoid memory leak issue.

Fixes: a30a67be7b6e ("clk: socfpga: Don't have get_parent for single parent ops")
Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
Link: https://lore.kernel.org/r/20221123031622.63171-1-xiujianfeng@huawei.com
Acked-by: Dinh Nguyen <dinguyen@kernel.org>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agobpf: Do not zero-extend kfunc return values
Björn Töpel [Wed, 7 Dec 2022 10:35:40 +0000 (11:35 +0100)]
bpf: Do not zero-extend kfunc return values

[ Upstream commit d35af0a7feb077c43ff0233bba5a8c6e75b73e35 ]

In BPF all global functions, and BPF helpers return a 64-bit
value. For kfunc calls, this is not the case, and they can return
e.g. 32-bit values.

The return register R0 for kfuncs calls can therefore be marked as
subreg_def != DEF_NOT_SUBREG. In general, if a register is marked with
subreg_def != DEF_NOT_SUBREG, some archs (where bpf_jit_needs_zext()
returns true) require the verifier to insert explicit zero-extension
instructions.

For kfuncs calls, however, the caller should do sign/zero extension
for return values. In other words, the compiler is responsible to
insert proper instructions, not the verifier.

An example, provided by Yonghong Song:

$ cat t.c
extern unsigned foo(void);
unsigned bar1(void) {
     return foo();
}
unsigned bar2(void) {
     if (foo()) return 10; else return 20;
}

$ clang -target bpf -mcpu=v3 -O2 -c t.c && llvm-objdump -d t.o
t.o:    file format elf64-bpf

Disassembly of section .text:

0000000000000000 <bar1>:
0:       85 10 00 00 ff ff ff ff call -0x1
1:       95 00 00 00 00 00 00 00 exit

0000000000000010 <bar2>:
2:       85 10 00 00 ff ff ff ff call -0x1
3:       bc 01 00 00 00 00 00 00 w1 = w0
4:       b4 00 00 00 14 00 00 00 w0 = 0x14
5:       16 01 01 00 00 00 00 00 if w1 == 0x0 goto +0x1 <LBB1_2>
6:       b4 00 00 00 0a 00 00 00 w0 = 0xa

0000000000000038 <LBB1_2>:
7:       95 00 00 00 00 00 00 00 exit

If the return value of 'foo()' is used in the BPF program, the proper
zero-extension will be done.

Currently, the verifier correctly marks, say, a 32-bit return value as
subreg_def != DEF_NOT_SUBREG, but will fail performing the actual
zero-extension, due to a verifier bug in
opt_subreg_zext_lo32_rnd_hi32(). load_reg is not properly set to R0,
and the following path will be taken:

if (WARN_ON(load_reg == -1)) {
verbose(env, "verifier bug. zext_dst is set, but no reg is defined\n");
return -EFAULT;
}

A longer discussion from v1 can be found in the link below.

Correct the verifier by avoiding doing explicit zero-extension of R0
for kfunc calls. Note that R0 will still be marked as a sub-register
for return values smaller than 64-bit.

Fixes: 83a2881903f3 ("bpf: Account for BPF_FETCH in insn_has_def32()")
Link: https://lore.kernel.org/bpf/20221202103620.1915679-1-bjorn@kernel.org/
Suggested-by: Yonghong Song <yhs@meta.com>
Signed-off-by: Björn Töpel <bjorn@rivosinc.com>
Acked-by: Yonghong Song <yhs@fb.com>
Link: https://lore.kernel.org/r/20221207103540.396496-1-bjorn@kernel.org
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoblktrace: Fix output non-blktrace event when blk_classic option enabled
Yang Jihong [Tue, 22 Nov 2022 04:04:10 +0000 (12:04 +0800)]
blktrace: Fix output non-blktrace event when blk_classic option enabled

[ Upstream commit f596da3efaf4130ff61cd029558845808df9bf99 ]

When the blk_classic option is enabled, non-blktrace events must be
filtered out. Otherwise, events of other types are output in the blktrace
classic format, which is unexpected.

The problem can be triggered in the following ways:

  # echo 1 > /sys/kernel/debug/tracing/options/blk_classic
  # echo 1 > /sys/kernel/debug/tracing/events/enable
  # echo blk > /sys/kernel/debug/tracing/current_tracer
  # cat /sys/kernel/debug/tracing/trace_pipe

Fixes: c71a89615411 ("blktrace: add ftrace plugin")
Signed-off-by: Yang Jihong <yangjihong1@huawei.com>
Link: https://lore.kernel.org/r/20221122040410.85113-1-yangjihong1@huawei.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: brcmfmac: Fix error return code in brcmf_sdio_download_firmware()
Wang Yufen [Fri, 2 Dec 2022 05:35:42 +0000 (13:35 +0800)]
wifi: brcmfmac: Fix error return code in brcmf_sdio_download_firmware()

[ Upstream commit c2f2924bc7f9ea75ef8d95863e710168f8196256 ]

Fix to return a negative error code instead of 0 when
brcmf_chip_set_active() fails. In addition, change the return
value for brcmf_pcie_exit_download_state() to keep consistent.

Fixes: d380ebc9b6fb ("brcmfmac: rename chip download functions")
Signed-off-by: Wang Yufen <wangyufen@huawei.com>
Reviewed-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/1669959342-27144-1-git-send-email-wangyufen@huawei.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: rtl8xxxu: Fix the channel width reporting
Bitterblue Smith [Thu, 1 Dec 2022 14:15:08 +0000 (16:15 +0200)]
wifi: rtl8xxxu: Fix the channel width reporting

[ Upstream commit 76c16af2cb10282274596e21add2c9f0b95c941b ]

The gen 2 chips RTL8192EU and RTL8188FU periodically send the driver
reports about the TX rate, and the driver passes these reports to
sta_statistics. The reports from RTL8192EU may or may not include the
channel width. The reports from RTL8188FU do not include it.

Only access the c2h->ra_report.bw field if the report (skb) is big
enough.

The other problem fixed here is that the code was actually never
changing the channel width initially reported by
rtl8xxxu_bss_info_changed because the value of RATE_INFO_BW_20 is 0.

Fixes: 0985d3a410ac ("rtl8xxxu: Feed current txrate information for mac80211")
Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Reviewed-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/5b41f1ae-72e7-6b7a-2459-b736399a1c40@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: rtl8xxxu: Add __packed to struct rtl8723bu_c2h
Bitterblue Smith [Thu, 1 Dec 2022 14:13:57 +0000 (16:13 +0200)]
wifi: rtl8xxxu: Add __packed to struct rtl8723bu_c2h

[ Upstream commit dd469a754afdb782ba3033cee102147493dc39f4 ]

This struct is used to access a sequence of bytes received from the
wifi chip. It must not have any padding bytes between the members.

This doesn't change anything on my system, possibly because currently
none of the members need more than byte alignment.

Fixes: b2b43b7837ba ("rtl8xxxu: Initial functionality to handle C2H events for 8723bu")
Signed-off-by: Bitterblue Smith <rtl8821cerfe2@gmail.com>
Reviewed-by: Ping-Ke Shih <pkshih@realtek.com>
Signed-off-by: Kalle Valo <kvalo@kernel.org>
Link: https://lore.kernel.org/r/1a270918-da22-ff5f-29fc-7855f740c5ba@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agospi: spi-gpio: Don't set MOSI as an input if not 3WIRE mode
Kris Bahnsen [Wed, 7 Dec 2022 23:08:53 +0000 (15:08 -0800)]
spi: spi-gpio: Don't set MOSI as an input if not 3WIRE mode

[ Upstream commit 3a6f994f848a69deb2bf3cd9d130dd0c09730e55 ]

The addition of 3WIRE support would affect MOSI direction even
when still in standard (4 wire) mode. This can lead to MOSI being
at an invalid logic level when a device driver sets an SPI
message with a NULL tx_buf.

spi.h states that if tx_buf is NULL then "zeros will be shifted
out ... " If MOSI is tristated then the data shifted out is subject
to pull resistors, keepers, or in the absence of those, noise.

This issue came to light when using spi-gpio connected to an
ADS7843 touchscreen controller. MOSI pulled high when clocking
MISO data in caused the SPI device to interpret this as a command
which would put the device in an unexpected and non-functional
state.

Fixes: 4b859db2c606 ("spi: spi-gpio: add SPI_3WIRE support")
Fixes: 5132b3d28371 ("spi: gpio: Support 3WIRE high-impedance turn-around")
Signed-off-by: Kris Bahnsen <kris@embeddedTS.com>
Link: https://lore.kernel.org/r/20221207230853.6174-1-kris@embeddedTS.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoclk: samsung: Fix memory leak in _samsung_clk_register_pll()
Xiu Jianfeng [Wed, 23 Nov 2022 03:20:15 +0000 (11:20 +0800)]
clk: samsung: Fix memory leak in _samsung_clk_register_pll()

[ Upstream commit 5174e5b0d1b669a489524192b6adcbb3c54ebc72 ]

If clk_register() fails, @pll->rate_table may have allocated memory by
kmemdup(), so it needs to be freed, otherwise will cause memory leak
issue, this patch fixes it.

Fixes: 3ff6e0d8d64d ("clk: samsung: Add support to register rate_table for samsung plls")
Signed-off-by: Xiu Jianfeng <xiujianfeng@huawei.com>
Link: https://lore.kernel.org/r/20221123032015.63980-1-xiujianfeng@huawei.com
Reviewed-by: Alim Akhtar <alim.akhtar@samsung.com>
Signed-off-by: Stephen Boyd <sboyd@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomedia: staging: stkwebcam: Restore MEDIA_{USB,CAMERA}_SUPPORT dependencies
Geert Uytterhoeven [Mon, 21 Nov 2022 15:58:33 +0000 (16:58 +0100)]
media: staging: stkwebcam: Restore MEDIA_{USB,CAMERA}_SUPPORT dependencies

[ Upstream commit faaf901727eddcfbe889fe172ec9cdb5e63c8236 ]

By moving support for the USB Syntek DC1125 Camera to staging, the
dependencies on MEDIA_USB_SUPPORT and MEDIA_CAMERA_SUPPORT were lost.

Fixes: 56280c64ecac ("media: stkwebcam: deprecate driver, move to staging")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Ricardo Ribalda <ribalda@chromium.org>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomedia: coda: Add check for kmalloc
Jiasheng Jiang [Thu, 17 Nov 2022 07:02:36 +0000 (15:02 +0800)]
media: coda: Add check for kmalloc

[ Upstream commit 6e5e5defdb8b0186312c2f855ace175aee6daf9b ]

As the kmalloc may return NULL pointer,
it should be better to check the return value
in order to avoid NULL poineter dereference,
same as the others.

Fixes: cb1d3a336371 ("[media] coda: add CODA7541 JPEG support")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomedia: coda: Add check for dcoda_iram_alloc
Jiasheng Jiang [Thu, 17 Nov 2022 06:56:52 +0000 (14:56 +0800)]
media: coda: Add check for dcoda_iram_alloc

[ Upstream commit 6b8082238fb8bb20f67e46388123e67a5bbc558d ]

As the coda_iram_alloc may return NULL pointer,
it should be better to check the return value
in order to avoid NULL poineter dereference,
same as the others.

Fixes: b313bcc9a467 ("[media] coda: simplify IRAM setup")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomedia: c8sectpfe: Add of_node_put() when breaking out of loop
Liang He [Tue, 19 Jul 2022 14:10:23 +0000 (22:10 +0800)]
media: c8sectpfe: Add of_node_put() when breaking out of loop

[ Upstream commit 63ff05a1ad242a5a0f897921c87b70d601bda59c ]

In configure_channels(), we should call of_node_put() when breaking
out of for_each_child_of_node() which will automatically increase
and decrease the refcount.

Fixes: c5f5d0f99794 ("[media] c8sectpfe: STiH407/10 Linux DVB demux support")
Signed-off-by: Liang He <windhl@126.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoregulator: qcom-labibb: Fix missing of_node_put() in qcom_labibb_regulator_probe()
Yuan Can [Sat, 3 Dec 2022 06:21:09 +0000 (06:21 +0000)]
regulator: qcom-labibb: Fix missing of_node_put() in qcom_labibb_regulator_probe()

[ Upstream commit cf34ac6aa2b12fb0c3aacfdcae8acd7904b949ec ]

The reg_node needs to be released through of_node_put() in the error
handling path when of_irq_get_byname() failed.

Fixes: 390af53e0411 ("regulator: qcom-labibb: Implement short-circuit and over-current IRQs")
Signed-off-by: Yuan Can <yuancan@huawei.com>
Link: https://lore.kernel.org/r/20221203062109.115043-1-yuancan@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agonvme: pass nr_maps explicitly to nvme_alloc_io_tag_set
Christoph Hellwig [Wed, 30 Nov 2022 16:16:52 +0000 (17:16 +0100)]
nvme: pass nr_maps explicitly to nvme_alloc_io_tag_set

[ Upstream commit dcef77274ae52136925287b6b59d5c6e6a4adfb9 ]

Don't look at ctrl->ops as only RDMA and TCP actually support multiple
maps.

Fixes: 6dfba1c09c10 ("nvme-fc: use the tagset alloc/free helpers")
Fixes: ceee1953f923 ("nvme-loop: use the tagset alloc/free helpers")
Signed-off-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Sagi Grimberg <sagi@grimberg.me>
Reviewed-by: Chaitanya Kulkarni <kch@nvidia.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: core: Normalize the error handling branch in sd_read_ext_regs()
Zhen Lei [Wed, 30 Nov 2022 13:49:20 +0000 (21:49 +0800)]
mmc: core: Normalize the error handling branch in sd_read_ext_regs()

[ Upstream commit fc02e2b52389c8fde02852b2f959c0b45f042bbd ]

Let's use pr_err() to output the error messages and let's extend a comment
to clarify why returning 0 (success) in one case make sense.

Fixes: c784f92769ae ("mmc: core: Read the SD function extension registers for power management")
Signed-off-by: Zhen Lei <thunder.leizhen@huawei.com>
[Ulf: Clarified the comment and the commit-msg]
Link: https://lore.kernel.org/r/20221130134920.2109-1-thunder.leizhen@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomemstick/ms_block: Add check for alloc_ordered_workqueue
Jiasheng Jiang [Sat, 26 Nov 2022 01:25:58 +0000 (09:25 +0800)]
memstick/ms_block: Add check for alloc_ordered_workqueue

[ Upstream commit 4f431a047a5c8698ed4b67e2760cfbeb5fffb69d ]

As the alloc_ordered_workqueue may return NULL pointer, it should be better
to add check for the return value. Moreover, the msb->io_queue should be
freed if error occurs later.

Fixes: 0ab30494bc4f ("memstick: add support for legacy memorysticks")
Signed-off-by: Jiasheng Jiang <jiasheng@iscas.ac.cn>
Link: https://lore.kernel.org/r/20221126012558.34374-1-jiasheng@iscas.ac.cn
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: renesas_sdhi: alway populate SCC pointer
Wolfram Sang [Sun, 20 Nov 2022 11:34:54 +0000 (12:34 +0100)]
mmc: renesas_sdhi: alway populate SCC pointer

[ Upstream commit 3d4f9898c1c74323dd61d6a8a0efca9401232ad4 ]

We need the SCC pointer to reset the device, so populate it even when we
don't need it for tuning.

Fixes: 45bffc371fef ("mmc: renesas_sdhi: only reset SCC when its pointer is populated")
Signed-off-by: Takeshi Saito <takeshi.saito.xv@renesas.com>
Signed-off-by: Takeshi Kihara <takeshi.kihara.df@renesas.com>
Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Tested-by: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Link: https://lore.kernel.org/r/20221120113457.42010-2-wsa+renesas@sang-engineering.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: mmci: fix return value check of mmc_add_host()
Yang Yingliang [Wed, 9 Nov 2022 13:35:39 +0000 (21:35 +0800)]
mmc: mmci: fix return value check of mmc_add_host()

[ Upstream commit b38a20f29a49ae04d23750d104b25400b792b98c ]

mmc_add_host() may return error, if we ignore its return value,
it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to
   delete device, but it's not added yet, it will lead a kernel
   crash because of null-ptr-deref in device_del().

So fix this by checking the return value and goto error path which
will call mmc_free_host().

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221109133539.3275664-1-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: wbsd: fix return value check of mmc_add_host()
Yang Yingliang [Wed, 9 Nov 2022 13:32:37 +0000 (21:32 +0800)]
mmc: wbsd: fix return value check of mmc_add_host()

[ Upstream commit dc5b9b50fc9d1334407e316e6e29a5097ef833bd ]

mmc_add_host() may return error, if we ignore its return value,
it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to
   delete device, but it's not added yet, it will lead a kernel
   crash because of null-ptr-deref in device_del().

So fix this by checking the return value and goto error path which
will call mmc_free_host(), besides, other resources also need be
released.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221109133237.3273558-1-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: via-sdmmc: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 8 Nov 2022 13:09:49 +0000 (21:09 +0800)]
mmc: via-sdmmc: fix return value check of mmc_add_host()

[ Upstream commit e4e46fb61e3bb4628170810d3f2b996b709b90d9 ]

mmc_add_host() may return error, if we ignore its return value,
it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to
   delete device, but it's not added yet, it will lead a kernel
   crash because of null-ptr-deref in device_del().

Fix this by checking the return value and goto error path which
will call mmc_free_host().

Fixes: f0bf7f61b840 ("mmc: Add new via-sdmmc host controller driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221108130949.1067699-1-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: meson-gx: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 8 Nov 2022 12:34:17 +0000 (20:34 +0800)]
mmc: meson-gx: fix return value check of mmc_add_host()

[ Upstream commit 90935f16f2650ab7416fa2ffbe5c28cb39cf3f1e ]

mmc_add_host() may return error, if we ignore its return value,
it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to
   delete device, but it's not added yet, it will lead a kernel
   crash because of null-ptr-deref in device_del().

Fix this by checking the return value and goto error path which
will call mmc_free_host().

Fixes: 51c5d8447bd7 ("MMC: meson: initial support for GX platforms")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Reviewed-by: Neil Armstrong <neil.armstrong@linaro.org>
Link: https://lore.kernel.org/r/20221108123417.479045-1-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: omap_hsmmc: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 8 Nov 2022 12:13:16 +0000 (20:13 +0800)]
mmc: omap_hsmmc: fix return value check of mmc_add_host()

[ Upstream commit a525cad241c339ca00bf7ebf03c5180f2a9b767c ]

mmc_add_host() may return error, if we ignore its return value,
it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to
   delete device, but it's not added yet, it will lead a kernel
   crash because of null-ptr-deref in device_del().

Fix this by checking the return value and goto error path wihch
will call mmc_free_host().

Fixes: a45c6cb81647 ("[ARM] 5369/1: omap mmc: Add new omap hsmmc controller for 2430 and 34xx, v3")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221108121316.340354-1-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: atmel-mci: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 8 Nov 2022 12:28:19 +0000 (20:28 +0800)]
mmc: atmel-mci: fix return value check of mmc_add_host()

[ Upstream commit 9e6e8c43726673ca2abcaac87640b9215fd72f4c ]

mmc_add_host() may return error, if we ignore its return value,
it will lead two issues:
1. The memory that allocated in mmc_alloc_host() is leaked.
2. In the remove() path, mmc_remove_host() will be called to
   delete device, but it's not added yet, it will lead a kernel
   crash because of null-ptr-deref in device_del().

So fix this by checking the return value and calling mmc_free_host()
in the error path.

Fixes: 7d2be0749a59 ("atmel-mci: Driver for Atmel on-chip MMC controllers")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221108122819.429975-1-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: litex_mmc: ensure `host->irq == 0` if polling
Gabriel Somlo [Mon, 7 Nov 2022 15:55:16 +0000 (10:55 -0500)]
mmc: litex_mmc: ensure `host->irq == 0` if polling

[ Upstream commit 5c1a2b77cd1b59112cf22b3e338f7e416797ad32 ]

Ensure the flag is explicitly set to 0 if we determine that polling is
needed during driver probe, to cover all possible cases.

Fixes: 92e099104729 ("mmc: Add driver for LiteX's LiteSDCard interface")
Signed-off-by: Gabriel Somlo <gsomlo@gmail.com>
Link: https://lore.kernel.org/r/20221107155516.2535912-1-gsomlo@gmail.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: wmt-sdmmc: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:23 +0000 (14:30 +0800)]
mmc: wmt-sdmmc: fix return value check of mmc_add_host()

[ Upstream commit 29276d56f6ed138db0f38cd31aedc0b725c8c76c ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call
mmc_free_host(), besides, clk_disable_unprepare() also needs be called.

Fixes: 3a96dff0f828 ("mmc: SD/MMC Host Controller for Wondermedia WM8505/WM8650")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-10-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: vub300: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:22 +0000 (14:30 +0800)]
mmc: vub300: fix return value check of mmc_add_host()

[ Upstream commit 0613ad2401f88bdeae5594c30afe318e93b14676 ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call
mmc_free_host(), besides, the timer added before mmc_add_host() needs be del.

And this patch fixes another missing call mmc_free_host() if usb_control_msg()
fails.

Fixes: 88095e7b473a ("mmc: Add new VUB300 USB-to-SD/SDIO/MMC driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-9-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: toshsd: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:21 +0000 (14:30 +0800)]
mmc: toshsd: fix return value check of mmc_add_host()

[ Upstream commit f670744a316ea983113a65313dcd387b5a992444 ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call
mmc_free_host(), besides, free_irq() also needs be called.

Fixes: a5eb8bbd66cc ("mmc: add Toshiba PCI SD controller driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-8-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: rtsx_usb_sdmmc: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:20 +0000 (14:30 +0800)]
mmc: rtsx_usb_sdmmc: fix return value check of mmc_add_host()

[ Upstream commit fc38a5a10e9e5a75eb9189854abeb8405b214cc9 ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and calling mmc_free_host() in the
error path, besides, led_classdev_unregister() and pm_runtime_disable() also
need be called.

Fixes: c7f6558d84af ("mmc: Add realtek USB sdmmc host driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-7-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: rtsx_pci: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:19 +0000 (14:30 +0800)]
mmc: rtsx_pci: fix return value check of mmc_add_host()

[ Upstream commit 0c87db77423a282b3b38b8a6daf057b822680516 ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and calling mmc_free_host() in the
error path, beside, runtime PM also needs be disabled.

Fixes: ff984e57d36e ("mmc: Add realtek pcie sdmmc host driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-6-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: pxamci: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:18 +0000 (14:30 +0800)]
mmc: pxamci: fix return value check of mmc_add_host()

[ Upstream commit 80e1ef3afb8bfbe768380b70ffe1b6cab87d1a3b ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call
mmc_free_host(), besides, ->exit() need be called to uninit the pdata.

Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-5-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: mxcmmc: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:17 +0000 (14:30 +0800)]
mmc: mxcmmc: fix return value check of mmc_add_host()

[ Upstream commit cde600af7b413c9fe03e85c58c4279df90e91d13 ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call
mmc_free_host().

Fixes: d96be879ff46 ("mmc: Add a MX2/MX3 specific SDHC driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-4-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: moxart: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:16 +0000 (14:30 +0800)]
mmc: moxart: fix return value check of mmc_add_host()

[ Upstream commit 0ca18d09c744fb030ae9bc5836c3e357e0237dea ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and goto error path which will call
mmc_free_host().

Fixes: 1b66e94e6b99 ("mmc: moxart: Add MOXA ART SD/MMC driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-3-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agommc: alcor: fix return value check of mmc_add_host()
Yang Yingliang [Tue, 1 Nov 2022 06:30:15 +0000 (14:30 +0800)]
mmc: alcor: fix return value check of mmc_add_host()

[ Upstream commit e93d1468f429475a753d6baa79b853b7ee5ef8c0 ]

mmc_add_host() may return error, if we ignore its return value, the memory
that allocated in mmc_alloc_host() will be leaked and it will lead a kernel
crash because of deleting not added device in the remove path.

So fix this by checking the return value and calling mmc_free_host() in the
error path.

Fixes: c5413ad815a6 ("mmc: add new Alcor Micro Cardreader SD/MMC driver")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221101063023.1664968-2-yangyingliang@huawei.com
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohwmon: (emc2305) fix pwm never being able to set lower
Xingjiang Qiao [Tue, 6 Dec 2022 05:53:31 +0000 (13:53 +0800)]
hwmon: (emc2305) fix pwm never being able to set lower

[ Upstream commit 364ffd2537c44cb6914ff5669153f4a86fffad29 ]

There are fields 'last_hwmon_state' and 'last_thermal_state' in the
structure 'emc2305_cdev_data', which respectively store the cooling state
set by the 'hwmon' and 'thermal' subsystem, and the driver author hopes
that if the state set by 'hwmon' is lower than the value set by 'thermal',
the driver will just save it without actually setting the pwm. Currently,
the 'last_thermal_state' also be updated by 'hwmon', which will cause the
cooling state to never be set to a lower value. This patch fixes that.

Signed-off-by: Xingjiang Qiao <nanpuyue@gmail.com>
Link: https://lore.kernel.org/r/20221206055331.170459-2-nanpuyue@gmail.com
Fixes: 0d8400c5a2ce1 ("hwmon: (emc2305) add support for EMC2301/2/3/5 RPM-based PWM Fan Speed Controller.")
[groeck: renamed emc2305_set_cur_state_shim -> __emc2305_set_cur_state]
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohwmon: (emc2305) fix unable to probe emc2301/2/3
Xingjiang Qiao [Tue, 6 Dec 2022 05:53:30 +0000 (13:53 +0800)]
hwmon: (emc2305) fix unable to probe emc2301/2/3

[ Upstream commit 4d50591ebf60ccf79380fff3a4c23659c61c482f ]

The definitions of 'EMC2305_REG_PRODUCT_ID' and 'EMC2305_REG_DEVICE' are
both '0xfd', they actually return the same value, but the values returned
by emc2301/2/3/5 are different, so probe emc2301/2/3 will fail, This patch
fixes that.

Signed-off-by: Xingjiang Qiao <nanpuyue@gmail.com>
Link: https://lore.kernel.org/r/20221206055331.170459-1-nanpuyue@gmail.com
Fixes: 0d8400c5a2ce1 ("hwmon: (emc2305) add support for EMC2301/2/3/5 RPM-based PWM Fan Speed Controller.")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agobpftool: Fix memory leak in do_build_table_cb
Miaoqian Lin [Tue, 6 Dec 2022 07:19:06 +0000 (11:19 +0400)]
bpftool: Fix memory leak in do_build_table_cb

[ Upstream commit fa55ef14ef4fe06198c0ce811b603aec24134bc2 ]

strdup() allocates memory for path. We need to release the memory in the
following error path. Add free() to avoid memory leak.

Fixes: 8f184732b60b ("bpftool: Switch to libbpf's hashmap for pinned paths of BPF objects")
Signed-off-by: Miaoqian Lin <linmq006@gmail.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Link: https://lore.kernel.org/bpf/20221206071906.806384-1-linmq006@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoriscv, bpf: Emit fixed-length instructions for BPF_PSEUDO_FUNC
Pu Lehui [Tue, 6 Dec 2022 09:14:10 +0000 (17:14 +0800)]
riscv, bpf: Emit fixed-length instructions for BPF_PSEUDO_FUNC

[ Upstream commit b54b6003612a376e7be32cbc5c1af3754bbbbb3d ]

For BPF_PSEUDO_FUNC instruction, verifier will refill imm with
correct addresses of bpf_calls and then run last pass of JIT.
Since the emit_imm of RV64 is variable-length, which will emit
appropriate length instructions accorroding to the imm, it may
broke ctx->offset, and lead to unpredictable problem, such as
inaccurate jump. So let's fix it with fixed-length instructions.

Fixes: 69c087ba6225 ("bpf: Add bpf_for_each_map_elem() helper")
Suggested-by: Björn Töpel <bjorn@rivosinc.com>
Signed-off-by: Pu Lehui <pulehui@huawei.com>
Signed-off-by: Daniel Borkmann <daniel@iogearbox.net>
Reviewed-by: Björn Töpel <bjorn@kernel.org>
Acked-by: Björn Töpel <bjorn@kernel.org>
Link: https://lore.kernel.org/bpf/20221206091410.1584784-1-pulehui@huaweicloud.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoNFSv4.x: Fail client initialisation if state manager thread can't run
Trond Myklebust [Tue, 6 Dec 2022 17:42:59 +0000 (12:42 -0500)]
NFSv4.x: Fail client initialisation if state manager thread can't run

[ Upstream commit b4e4f66901658fae0614dea5bf91062a5387eda7 ]

If the state manager thread fails to start, then we should just mark the
client initialisation as failed so that other processes or threads don't
get stuck in nfs_wait_client_init_complete().

Reported-by: ChenXiaoSong <chenxiaosong2@huawei.com>
Fixes: 4697bd5e9419 ("NFSv4: Fix a race in the net namespace mount notification")
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoNFS: Allow very small rsize & wsize again
Anna Schumaker [Wed, 30 Nov 2022 20:30:47 +0000 (15:30 -0500)]
NFS: Allow very small rsize & wsize again

[ Upstream commit a60214c2465493aac0b014d87ee19327b6204c42 ]

940261a19508 introduced nfs_io_size() to clamp the iosize to a multiple
of PAGE_SIZE. This had the unintended side effect of no longer allowing
iosizes less than a page, which could be useful in some situations.

UDP already has an exception that causes it to fall back on the
power-of-two style sizes instead. This patch adds an additional
exception for very small iosizes.

Reported-by: Jeff Layton <jlayton@kernel.org>
Fixes: 940261a19508 ("NFS: Allow setting rsize / wsize to a multiple of PAGE_SIZE")
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Reviewed-by: Jeff Layton <jlayton@kernel.org>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoNFSv4.2: Set the correct size scratch buffer for decoding READ_PLUS
Anna Schumaker [Wed, 30 Nov 2022 18:15:25 +0000 (13:15 -0500)]
NFSv4.2: Set the correct size scratch buffer for decoding READ_PLUS

[ Upstream commit 36357fe74ef736524a29fbd3952948768510a8b9 ]

The scratch_buf array is 16 bytes, but I was passing 32 to the
xdr_set_scratch_buffer() function. Fix this by using sizeof(), which is
what I probably should have been doing this whole time.

Fixes: d3b00a802c84 ("NFS: Replace the READ_PLUS decoding code")
Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoSUNRPC: Fix missing release socket in rpc_sockname()
Wang ShaoBo [Thu, 24 Nov 2022 09:23:42 +0000 (17:23 +0800)]
SUNRPC: Fix missing release socket in rpc_sockname()

[ Upstream commit 50fa355bc0d75911fe9d5072a5ba52cdb803aff7 ]

socket dynamically created is not released when getting an unintended
address family type in rpc_sockname(), direct to out_release for calling
sock_release().

Fixes: 2e738fdce22f ("SUNRPC: Add API to acquire source address")
Signed-off-by: Wang ShaoBo <bobo.shaobowang@huawei.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoxprtrdma: Fix regbuf data not freed in rpcrdma_req_create()
Zhang Xiaoxu [Sun, 20 Nov 2022 07:34:29 +0000 (15:34 +0800)]
xprtrdma: Fix regbuf data not freed in rpcrdma_req_create()

[ Upstream commit 9181f40fb2952fd59ecb75e7158620c9c669eee3 ]

If rdma receive buffer allocate failed, should call rpcrdma_regbuf_free()
to free the send buffer, otherwise, the buffer data will be leaked.

Fixes: bb93a1ae2bf4 ("xprtrdma: Allocate req's regbufs at xprt create time")
Signed-off-by: Zhang Xiaoxu <zhangxiaoxu5@huawei.com>
Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agopinctrl: thunderbay: fix possible memory leak in thunderbay_build_functions()
Gaosheng Cui [Tue, 29 Nov 2022 12:01:26 +0000 (20:01 +0800)]
pinctrl: thunderbay: fix possible memory leak in thunderbay_build_functions()

[ Upstream commit 83e1bcaf8cef26edaaf2a6098ef760f563683483 ]

The thunderbay_add_functions() will free memory of thunderbay_funcs
when everything is ok, but thunderbay_funcs will not be freed when
thunderbay_add_functions() fails, then there will be a memory leak,
so we need to add kfree() when thunderbay_add_functions() fails to
fix it.

In addition, doing some cleaner works, moving kfree(funcs) from
thunderbay_add_functions() to thunderbay_build_functions().

Fixes: 12422af8194d ("pinctrl: Add Intel Thunder Bay pinctrl driver")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Reviewed-by: Rafał Miłecki <rafal@milecki.pl>
Link: https://lore.kernel.org/r/20221129120126.1567338-1-cuigaosheng1@huawei.com
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoALSA: mts64: fix possible null-ptr-defer in snd_mts64_interrupt
Gaosheng Cui [Tue, 6 Dec 2022 06:10:04 +0000 (14:10 +0800)]
ALSA: mts64: fix possible null-ptr-defer in snd_mts64_interrupt

[ Upstream commit cf2ea3c86ad90d63d1c572b43e1ca9276b0357ad ]

I got a null-ptr-defer error report when I do the following tests
on the qemu platform:

make defconfig and CONFIG_PARPORT=m, CONFIG_PARPORT_PC=m,
CONFIG_SND_MTS64=m

Then making test scripts:
cat>test_mod1.sh<<EOF
modprobe snd-mts64
modprobe snd-mts64
EOF

Executing the script, perhaps several times, we will get a null-ptr-defer
report, as follow:

syzkaller:~# ./test_mod.sh
snd_mts64: probe of snd_mts64.0 failed with error -5
modprobe: ERROR: could not insert 'snd_mts64': No such device
 BUG: kernel NULL pointer dereference, address: 0000000000000000
 #PF: supervisor write access in kernel mode
 #PF: error_code(0x0002) - not-present page
 PGD 0 P4D 0
 Oops: 0002 [#1] PREEMPT SMP PTI
 CPU: 0 PID: 205 Comm: modprobe Not tainted 6.1.0-rc8-00588-g76dcd734eca2 #6
 Call Trace:
  <IRQ>
  snd_mts64_interrupt+0x24/0xa0 [snd_mts64]
  parport_irq_handler+0x37/0x50 [parport]
  __handle_irq_event_percpu+0x39/0x190
  handle_irq_event_percpu+0xa/0x30
  handle_irq_event+0x2f/0x50
  handle_edge_irq+0x99/0x1b0
  __common_interrupt+0x5d/0x100
  common_interrupt+0xa0/0xc0
  </IRQ>
  <TASK>
  asm_common_interrupt+0x22/0x40
 RIP: 0010:_raw_write_unlock_irqrestore+0x11/0x30
  parport_claim+0xbd/0x230 [parport]
  snd_mts64_probe+0x14a/0x465 [snd_mts64]
  platform_probe+0x3f/0xa0
  really_probe+0x129/0x2c0
  __driver_probe_device+0x6d/0xc0
  driver_probe_device+0x1a/0xa0
  __device_attach_driver+0x7a/0xb0
  bus_for_each_drv+0x62/0xb0
  __device_attach+0xe4/0x180
  bus_probe_device+0x82/0xa0
  device_add+0x550/0x920
  platform_device_add+0x106/0x220
  snd_mts64_attach+0x2e/0x80 [snd_mts64]
  port_check+0x14/0x20 [parport]
  bus_for_each_dev+0x6e/0xc0
  __parport_register_driver+0x7c/0xb0 [parport]
  snd_mts64_module_init+0x31/0x1000 [snd_mts64]
  do_one_initcall+0x3c/0x1f0
  do_init_module+0x46/0x1c6
  load_module+0x1d8d/0x1e10
  __do_sys_finit_module+0xa2/0xf0
  do_syscall_64+0x37/0x90
  entry_SYSCALL_64_after_hwframe+0x63/0xcd
  </TASK>
 Kernel panic - not syncing: Fatal exception in interrupt
 Rebooting in 1 seconds..

The mts wa not initialized during interrupt,  we add check for
mts to fix this bug.

Fixes: 68ab801e32bb ("[ALSA] Add snd-mts64 driver for ESI Miditerminal 4140")
Signed-off-by: Gaosheng Cui <cuigaosheng1@huawei.com>
Link: https://lore.kernel.org/r/20221206061004.1222966-1-cuigaosheng1@huawei.com
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomedia: ov5640: set correct default link frequency
Guoniu.zhou [Fri, 25 Nov 2022 09:20:24 +0000 (09:20 +0000)]
media: ov5640: set correct default link frequency

[ Upstream commit d7b41196927ba2a2b5badad1a85f9087eb90b076 ]

current_link_freq field in ov5640_dev structure is link frequency,
not link frequency array index, so correct it.

Fixes: 3c28588f35d3 ("media: ov5640: Update pixel_rate and link_freq")
Signed-off-by: Guoniu.zhou <guoniu.zhou@nxp.com>
Acked-by: Jacopo Mondi <jacopo@jmondi.org>
Signed-off-by: Sakari Ailus <sakari.ailus@linux.intel.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomedia: saa7164: fix missing pci_disable_device()
Liu Shixin [Sat, 26 Nov 2022 11:31:26 +0000 (11:31 +0000)]
media: saa7164: fix missing pci_disable_device()

[ Upstream commit 57fb35d7542384cac8f198cd1c927540ad38b61a ]

Add missing pci_disable_device() in the error path in saa7164_initdev().

Fixes: 443c1228d505 ("V4L/DVB (12923): SAA7164: Add support for the NXP SAA7164 silicon")
Signed-off-by: Liu Shixin <liushixin2@huawei.com>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoALSA: pcm: Set missing stop_operating flag at undoing trigger start
Takashi Iwai [Mon, 5 Dec 2022 13:21:22 +0000 (14:21 +0100)]
ALSA: pcm: Set missing stop_operating flag at undoing trigger start

[ Upstream commit 5c8cc93b06d1ff860327a273abf3ac006290d242 ]

When a PCM trigger-start fails at snd_pcm_do_start(), PCM core tries
to undo the action at snd_pcm_undo_start() by issuing the trigger STOP
manually.  At that point, we forgot to set the stop_operating flag,
hence the sync-stop won't be issued at the next prepare or other
calls.

This patch adds the missing stop_operating flag at
snd_pcm_undo_start().

Fixes: 1e850beea278 ("ALSA: pcm: Add the support for sync-stop operation")
Link: https://lore.kernel.org/r/b4e71631-4a94-613-27b2-fb595792630@carlh.net
Link: https://lore.kernel.org/r/20221205132124.11585-2-tiwai@suse.de
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agobpf, sockmap: fix race in sock_map_free()
Eric Dumazet [Fri, 2 Dec 2022 11:16:40 +0000 (11:16 +0000)]
bpf, sockmap: fix race in sock_map_free()

[ Upstream commit 0a182f8d607464911756b4dbef5d6cad8de22469 ]

sock_map_free() calls release_sock(sk) without owning a reference
on the socket. This can cause use-after-free as syzbot found [1]

Jakub Sitnicki already took care of a similar issue
in sock_hash_free() in commit 75e68e5bf2c7 ("bpf, sockhash:
Synchronize delete from bucket list on map free")

[1]
refcount_t: decrement hit 0; leaking memory.
WARNING: CPU: 0 PID: 3785 at lib/refcount.c:31 refcount_warn_saturate+0x17c/0x1a0 lib/refcount.c:31
Modules linked in:
CPU: 0 PID: 3785 Comm: kworker/u4:6 Not tainted 6.1.0-rc7-syzkaller-00103-gef4d3ea40565 #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/26/2022
Workqueue: events_unbound bpf_map_free_deferred
RIP: 0010:refcount_warn_saturate+0x17c/0x1a0 lib/refcount.c:31
Code: 68 8b 31 c0 e8 75 71 15 fd 0f 0b e9 64 ff ff ff e8 d9 6e 4e fd c6 05 62 9c 3d 0a 01 48 c7 c7 80 bb 68 8b 31 c0 e8 54 71 15 fd <0f> 0b e9 43 ff ff ff 89 d9 80 e1 07 80 c1 03 38 c1 0f 8c a2 fe ff
RSP: 0018:ffffc9000456fb60 EFLAGS: 00010246
RAX: eae59bab72dcd700 RBX: 0000000000000004 RCX: ffff8880207057c0
RDX: 0000000000000000 RSI: 0000000000000201 RDI: 0000000000000000
RBP: 0000000000000004 R08: ffffffff816fdabd R09: fffff520008adee5
R10: fffff520008adee5 R11: 1ffff920008adee4 R12: 0000000000000004
R13: dffffc0000000000 R14: ffff88807b1c6c00 R15: 1ffff1100f638dcf
FS: 0000000000000000(0000) GS:ffff8880b9800000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000001b30c30000 CR3: 000000000d08e000 CR4: 00000000003506f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
Call Trace:
<TASK>
__refcount_dec include/linux/refcount.h:344 [inline]
refcount_dec include/linux/refcount.h:359 [inline]
__sock_put include/net/sock.h:779 [inline]
tcp_release_cb+0x2d0/0x360 net/ipv4/tcp_output.c:1092
release_sock+0xaf/0x1c0 net/core/sock.c:3468
sock_map_free+0x219/0x2c0 net/core/sock_map.c:356
process_one_work+0x81c/0xd10 kernel/workqueue.c:2289
worker_thread+0xb14/0x1330 kernel/workqueue.c:2436
kthread+0x266/0x300 kernel/kthread.c:376
ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:306
</TASK>

Fixes: 7e81a3530206 ("bpf: Sockmap, ensure sock lock held during tear down")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Cc: Jakub Sitnicki <jakub@cloudflare.com>
Cc: John Fastabend <john.fastabend@gmail.com>
Cc: Alexei Starovoitov <ast@kernel.org>
Cc: Daniel Borkmann <daniel@iogearbox.net>
Cc: Song Liu <songliubraving@fb.com>
Acked-by: John Fastabend <john.fastabend@gmail.com>
Link: https://lore.kernel.org/r/20221202111640.2745533-1-edumazet@google.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agobpf: Add dummy type reference to nf_conn___init to fix type deduplication
Toke Høiland-Jørgensen [Thu, 1 Dec 2022 12:39:39 +0000 (13:39 +0100)]
bpf: Add dummy type reference to nf_conn___init to fix type deduplication

[ Upstream commit 578ce69ffda49d6c1a252490553290d1f27199f0 ]

The bpf_ct_set_nat_info() kfunc is defined in the nf_nat.ko module, and
takes as a parameter the nf_conn___init struct, which is allocated through
the bpf_xdp_ct_alloc() helper defined in the nf_conntrack.ko module.
However, because kernel modules can't deduplicate BTF types between each
other, and the nf_conn___init struct is not referenced anywhere in vmlinux
BTF, this leads to two distinct BTF IDs for the same type (one in each
module). This confuses the verifier, as described here:

https://lore.kernel.org/all/87leoh372s.fsf@toke.dk/

As a workaround, add an explicit BTF_TYPE_EMIT for the type in
net/filter.c, so the type definition gets included in vmlinux BTF. This
way, both modules can refer to the same type ID (as they both build on top
of vmlinux BTF), and the verifier is no longer confused.

v2:

- Use BTF_TYPE_EMIT (which is a statement so it has to be inside a function
  definition; use xdp_func_proto() for this, since this is mostly
  xdp-related).

Fixes: 820dc0523e05 ("net: netfilter: move bpf_ct_set_nat_info kfunc in nf_nat_bpf.c")
Signed-off-by: Toke Høiland-Jørgensen <toke@redhat.com>
Link: https://lore.kernel.org/r/20221201123939.696558-1-toke@redhat.com
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohwmon: (jc42) Restore the min/max/critical temperatures on resume
Martin Blumenstingl [Sun, 23 Oct 2022 21:31:57 +0000 (23:31 +0200)]
hwmon: (jc42) Restore the min/max/critical temperatures on resume

[ Upstream commit 084ed144c448fd5bc8ed5a58247153fbbfd115c3 ]

The JC42 compatible thermal sensor on Kingston KSM32ES8/16ME DIMMs
(using Micron E-Die) is an ST Microelectronics STTS2004 (manufacturer
0x104a, device 0x2201). It does not keep the previously programmed
minimum, maximum and critical temperatures after system suspend and
resume (which is a shutdown / startup cycle for the JC42 temperature
sensor). This results in an alarm on system resume because the hardware
default for these values is 0°C (so any environment temperature greater
than 0°C will trigger the alarm).

Example before system suspend:
  jc42-i2c-0-1a
  Adapter: SMBus PIIX4 adapter port 0 at 0b00
  temp1:        +34.8°C  (low  =  +0.0°C)
                         (high = +85.0°C, hyst = +85.0°C)
                         (crit = +95.0°C, hyst = +95.0°C)

Example after system resume (without this change):
  jc42-i2c-0-1a
  Adapter: SMBus PIIX4 adapter port 0 at 0b00
  temp1:        +34.8°C  (low  =  +0.0°C)             ALARM (HIGH, CRIT)
                         (high =  +0.0°C, hyst =  +0.0°C)
                         (crit =  +0.0°C, hyst =  +0.0°C)

Apply the cached values from the JC42_REG_TEMP_UPPER,
JC42_REG_TEMP_LOWER, JC42_REG_TEMP_CRITICAL and JC42_REG_SMBUS (where
the SMBUS register is not related to this issue but a side-effect of
using regcache_sync() during system resume with the previously
cached/programmed values. This fixes the alarm due to the hardware
defaults of 0°C because the previously applied limits (set by userspace)
are re-applied on system resume.

Fixes: 175c490c9e7f ("hwmon: (jc42) Add support for STTS2004 and AT30TSE004")
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20221023213157.11078-3-martin.blumenstingl@googlemail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohwmon: (jc42) Convert register access and caching to regmap/regcache
Martin Blumenstingl [Sun, 23 Oct 2022 21:31:56 +0000 (23:31 +0200)]
hwmon: (jc42) Convert register access and caching to regmap/regcache

[ Upstream commit 8f2fa4726faf01094d7a5be7bd0c120c565f54d9 ]

Switch the jc42 driver to use an I2C regmap to access the registers.
Also move over to regmap's built-in caching instead of adding a
custom caching implementation. This works for JC42_REG_TEMP_UPPER,
JC42_REG_TEMP_LOWER and JC42_REG_TEMP_CRITICAL as these values never
change except when explicitly written. The cache For JC42_REG_TEMP is
dropped (regmap can't cache it because it's volatile, meaning it can
change at any time) as well for simplicity and consistency with other
drivers.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://lore.kernel.org/r/20221023213157.11078-2-martin.blumenstingl@googlemail.com
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Stable-dep-of: 084ed144c448 ("hwmon: (jc42) Restore the min/max/critical temperatures on resume")
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoregulator: core: fix resource leak in regulator_register()
Yang Yingliang [Fri, 2 Dec 2022 02:51:11 +0000 (10:51 +0800)]
regulator: core: fix resource leak in regulator_register()

[ Upstream commit ba62319a42c50e6254e98b3f316464fac8e77968 ]

I got some resource leak reports while doing fault injection test:

  OF: ERROR: memory leak, expected refcount 1 instead of 100,
  of_node_get()/of_node_put() unbalanced - destroy cset entry:
  attach overlay node /i2c/pmic@64/regulators/buck1

unreferenced object 0xffff88810deea000 (size 512):
  comm "490-i2c-rt5190a", pid 253, jiffies 4294859840 (age 5061.046s)
  hex dump (first 32 bytes):
    00 00 00 00 ad 4e ad de ff ff ff ff 00 00 00 00  .....N..........
    ff ff ff ff ff ff ff ff a0 1e 00 a1 ff ff ff ff  ................
  backtrace:
    [<00000000d78541e2>] kmalloc_trace+0x21/0x110
    [<00000000b343d153>] device_private_init+0x32/0xd0
    [<00000000be1f0c70>] device_add+0xb2d/0x1030
    [<00000000e3e6344d>] regulator_register+0xaf2/0x12a0
    [<00000000e2f5e754>] devm_regulator_register+0x57/0xb0
    [<000000008b898197>] rt5190a_probe+0x52a/0x861 [rt5190a_regulator]

unreferenced object 0xffff88810b617b80 (size 32):
  comm "490-i2c-rt5190a", pid 253, jiffies 4294859904 (age 5060.983s)
  hex dump (first 32 bytes):
    72 65 67 75 6c 61 74 6f 72 2e 32 38 36 38 2d 53  regulator.2868-S
    55 50 50 4c 59 00 ff ff 29 00 00 00 2b 00 00 00  UPPLY...)...+...
  backtrace:
    [<000000009da9280d>] __kmalloc_node_track_caller+0x44/0x1b0
    [<0000000025c6a4e5>] kstrdup+0x3a/0x70
    [<00000000790efb69>] create_regulator+0xc0/0x4e0
    [<0000000005ed203a>] regulator_resolve_supply+0x2d4/0x440
    [<0000000045796214>] regulator_register+0x10b3/0x12a0
    [<00000000e2f5e754>] devm_regulator_register+0x57/0xb0
    [<000000008b898197>] rt5190a_probe+0x52a/0x861 [rt5190a_regulator]

After calling regulator_resolve_supply(), the 'rdev->supply' is set
by set_supply(), after this set, in the error path, the resources
need be released, so call regulator_put() to avoid the leaks.

Fixes: aea6cb99703e ("regulator: resolve supply after creating regulator")
Fixes: 8a866d527ac0 ("regulator: core: Resolve supply name earlier to prevent double-init")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221202025111.496402-1-yangyingliang@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoconfigfs: fix possible memory leak in configfs_create_dir()
Chen Zhongjin [Mon, 17 Oct 2022 01:42:30 +0000 (09:42 +0800)]
configfs: fix possible memory leak in configfs_create_dir()

[ Upstream commit c65234b283a65cfbfc94619655e820a5e55199eb ]

kmemleak reported memory leaks in configfs_create_dir():

unreferenced object 0xffff888009f6af00 (size 192):
  comm "modprobe", pid 3777, jiffies 4295537735 (age 233.784s)
  backtrace:
    kmem_cache_alloc (mm/slub.c:3250 mm/slub.c:3256 mm/slub.c:3263 mm/slub.c:3273)
    new_fragment (./include/linux/slab.h:600 fs/configfs/dir.c:163)
    configfs_register_subsystem (fs/configfs/dir.c:1857)
    basic_write (drivers/hwtracing/stm/p_basic.c:14) stm_p_basic
    do_one_initcall (init/main.c:1296)
    do_init_module (kernel/module/main.c:2455)
    ...

unreferenced object 0xffff888003ba7180 (size 96):
  comm "modprobe", pid 3777, jiffies 4295537735 (age 233.784s)
  backtrace:
    kmem_cache_alloc (mm/slub.c:3250 mm/slub.c:3256 mm/slub.c:3263 mm/slub.c:3273)
    configfs_new_dirent (./include/linux/slab.h:723 fs/configfs/dir.c:194)
    configfs_make_dirent (fs/configfs/dir.c:248)
    configfs_create_dir (fs/configfs/dir.c:296)
    configfs_attach_group.isra.28 (fs/configfs/dir.c:816 fs/configfs/dir.c:852)
    configfs_register_subsystem (fs/configfs/dir.c:1881)
    basic_write (drivers/hwtracing/stm/p_basic.c:14) stm_p_basic
    do_one_initcall (init/main.c:1296)
    do_init_module (kernel/module/main.c:2455)
    ...

This is because the refcount is not correct in configfs_make_dirent().
For normal stage, the refcount is changing as:

configfs_register_subsystem()
  configfs_create_dir()
    configfs_make_dirent()
      configfs_new_dirent() # set s_count = 1
      dentry->d_fsdata = configfs_get(sd); # s_count = 2
...
configfs_unregister_subsystem()
  configfs_remove_dir()
    remove_dir()
      configfs_remove_dirent() # s_count = 1
    dput() ...
      *dentry_unlink_inode()*
        configfs_d_iput() # s_count = 0, release

However, if we failed in configfs_create():

configfs_register_subsystem()
  configfs_create_dir()
    configfs_make_dirent() # s_count = 2
    ...
    configfs_create() # fail
    ->out_remove:
    configfs_remove_dirent(dentry)
      configfs_put(sd) # s_count = 1
      return PTR_ERR(inode);

There is no inode in the error path, so the configfs_d_iput() is lost
and makes sd and fragment memory leaked.

To fix this, when we failed in configfs_create(), manually call
configfs_put(sd) to keep the refcount correct.

Fixes: 7063fbf22611 ("[PATCH] configfs: User-driven configuration filesystem")
Signed-off-by: Chen Zhongjin <chenzhongjin@huawei.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohsr: Synchronize sequence number updates.
Sebastian Andrzej Siewior [Tue, 29 Nov 2022 16:48:13 +0000 (17:48 +0100)]
hsr: Synchronize sequence number updates.

[ Upstream commit 5c7aa13210c3abdd34fd421f62347665ec6eb551 ]

hsr_register_frame_out() compares new sequence_nr vs the old one
recorded in hsr_node::seq_out and if the new sequence_nr is higher then
it will be written to hsr_node::seq_out as the new value.

This operation isn't locked so it is possible that two frames with the
same sequence number arrive (via the two slave devices) and are fed to
hsr_register_frame_out() at the same time. Both will pass the check and
update the sequence counter later to the same value. As a result the
content of the same packet is fed into the stack twice.

This was noticed by running ping and observing DUP being reported from
time to time.

Instead of using the hsr_priv::seqnr_lock for the whole receive path (as
it is for sending in the master node) add an additional lock that is only
used for sequence number checks and updates.

Add a per-node lock that is used during sequence number reads and
updates.

Fixes: f421436a591d3 ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohsr: Synchronize sending frames to have always incremented outgoing seq nr.
Sebastian Andrzej Siewior [Tue, 29 Nov 2022 16:48:12 +0000 (17:48 +0100)]
hsr: Synchronize sending frames to have always incremented outgoing seq nr.

[ Upstream commit 06afd2c31d338fa762548580c1bf088703dd1e03 ]

Sending frames via the hsr (master) device requires a sequence number
which is tracked in hsr_priv::sequence_nr and protected by
hsr_priv::seqnr_lock. Each time a new frame is sent, it will obtain a
new id and then send it via the slave devices.
Each time a packet is sent (via hsr_forward_do()) the sequence number is
checked via hsr_register_frame_out() to ensure that a frame is not
handled twice. This make sense for the receiving side to ensure that the
frame is not injected into the stack twice after it has been received
from both slave ports.

There is no locking to cover the sending path which means the following
scenario is possible:

  CPU0 CPU1
  hsr_dev_xmit(skb1) hsr_dev_xmit(skb2)
   fill_frame_info()             fill_frame_info()
    hsr_fill_frame_info()         hsr_fill_frame_info()
     handle_std_frame()            handle_std_frame()
      skb1's sequence_nr = 1
                                    skb2's sequence_nr = 2
   hsr_forward_do()              hsr_forward_do()

                                   hsr_register_frame_out(, 2)  // okay, send)

    hsr_register_frame_out(, 1) // stop, lower seq duplicate

Both skbs (or their struct hsr_frame_info) received an unique id.
However since skb2 was sent before skb1, the higher sequence number was
recorded in hsr_register_frame_out() and the late arriving skb1 was
dropped and never sent.

This scenario has been observed in a three node HSR setup, with node1 +
node2 having ping and iperf running in parallel. From time to time ping
reported a missing packet. Based on tracing that missing ping packet did
not leave the system.

It might be possible (didn't check) to drop the sequence number check on
the sending side. But if the higher sequence number leaves on wire
before the lower does and the destination receives them in that order
and it will drop the packet with the lower sequence number and never
inject into the stack.
Therefore it seems the only way is to lock the whole path from obtaining
the sequence number and sending via dev_queue_xmit() and assuming the
packets leave on wire in the same order (and don't get reordered by the
NIC).

Cover the whole path for the master interface from obtaining the ID
until after it has been forwarded via hsr_forward_skb() to ensure the
skbs are sent to the NIC in the order of the assigned sequence numbers.

Fixes: f421436a591d3 ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohsr: Disable netpoll.
Sebastian Andrzej Siewior [Tue, 29 Nov 2022 16:48:11 +0000 (17:48 +0100)]
hsr: Disable netpoll.

[ Upstream commit d5c7652eb16fa203d82546e0285136d7b321ffa9 ]

The hsr device is a software device. Its
net_device_ops::ndo_start_xmit() routine will process the packet and
then pass the resulting skb to dev_queue_xmit().
During processing, hsr acquires a lock with spin_lock_bh()
(hsr_add_node()) which needs to be promoted to the _irq() suffix in
order to avoid a potential deadlock.
Then there are the warnings in dev_queue_xmit() (due to
local_bh_disable() with disabled interrupts) left.

Instead trying to address those (there is qdisc and…) for netpoll sake,
just disable netpoll on hsr.

Disable netpoll on hsr and replace the _irqsave() locking with _bh().

Fixes: f421436a591d3 ("net/hsr: Add support for the High-availability Seamless Redundancy protocol (HSRv0)")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohsr: Avoid double remove of a node.
Sebastian Andrzej Siewior [Tue, 29 Nov 2022 16:48:10 +0000 (17:48 +0100)]
hsr: Avoid double remove of a node.

[ Upstream commit 0c74d9f79ec4299365bbe803baa736ae0068179e ]

Due to the hashed-MAC optimisation one problem become visible:
hsr_handle_sup_frame() walks over the list of available nodes and merges
two node entries into one if based on the information in the supervision
both MAC addresses belong to one node. The list-walk happens on a RCU
protected list and delete operation happens under a lock.

If the supervision arrives on both slave interfaces at the same time
then this delete operation can occur simultaneously on two CPUs. The
result is the first-CPU deletes the from the list and the second CPUs
BUGs while attempting to dereference a poisoned list-entry. This happens
more likely with the optimisation because a new node for the mac_B entry
is created once a packet has been received and removed (merged) once the
supervision frame has been received.

Avoid removing/ cleaning up a hsr_node twice by adding a `removed' field
which is set to true after the removal and checked before the removal.

Fixes: f266a683a4804 ("net/hsr: Better frame dispatch")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agohsr: Add a rcu-read lock to hsr_forward_skb().
Sebastian Andrzej Siewior [Tue, 29 Nov 2022 16:48:09 +0000 (17:48 +0100)]
hsr: Add a rcu-read lock to hsr_forward_skb().

[ Upstream commit 5aa2820177af650293b2f9f1873c1f6f8e4ad7a4 ]

hsr_forward_skb() a skb and keeps information in an on-stack
hsr_frame_info. hsr_get_node() assigns hsr_frame_info::node_src which is
from a RCU list. This pointer is used later in hsr_forward_do().
I don't see a reason why this pointer can't vanish midway since there is
no guarantee that hsr_forward_skb() is invoked from an RCU read section.

Use rcu_read_lock() to protect hsr_frame_info::node_src from its
assignment until it is no longer used.

Fixes: f266a683a4804 ("net/hsr: Better frame dispatch")
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoRevert "net: hsr: use hlist_head instead of list_head for mac addresses"
Sebastian Andrzej Siewior [Tue, 29 Nov 2022 16:48:08 +0000 (17:48 +0100)]
Revert "net: hsr: use hlist_head instead of list_head for mac addresses"

[ Upstream commit e012764cebf6e33097f6833ff15a936fbe7b846c ]

The hlist optimisation (which not only uses hlist_head instead of
list_head but also splits hsr_priv::node_db into an array of 256 slots)
does not consider the "node merge":
Upon starting the hsr network (with three nodes) a packet that is
sent from node1 to node3 will also be sent from node1 to node2 and then
forwarded to node3.
As a result node3 will receive 2 packets because it is not able
to filter out the duplicate. Each packet received will create a new
struct hsr_node with macaddress_A only set the MAC address it received
from (the two MAC addesses from node1).
At some point (early in the process) two supervision frames will be
received from node1. They will be processed by hsr_handle_sup_frame()
and one frame will leave early ("Node has already been merged") and does
nothing. The other frame will be merged as portB and have its MAC
address written to macaddress_B and the hsr_node (that was created for
it as macaddress_A) will be removed.
From now on HSR is able to identify a duplicate because both packets
sent from one node will result in the same struct hsr_node because
hsr_get_node() will find the MAC address either on macaddress_A or
macaddress_B.

Things get tricky with the optimisation: If sender's MAC address is
saved as macaddress_A then the lookup will work as usual. If the MAC
address has been merged into macaddress_B of another hsr_node then the
lookup won't work because it is likely that the data structure is in
another bucket. This results in creating a new struct hsr_node and not
recognising a possible duplicate.

A way around it would be to add another hsr_node::mac_list_B and attach
it to the other bucket to ensure that this hsr_node will be looked up
either via macaddress_A _or_ macaddress_B.

I however prefer to revert it because it sounds like an academic problem
rather than real life workload plus it adds complexity. I'm not an HSR
expert with what is usual size of a network but I would guess 40 to 60
nodes. With 10.000 nodes and assuming 60us for pass-through (from node
to node) then it would take almost 600ms for a packet to almost wrap
around which sounds a lot.

Revert the hash MAC addresses optimisation.

Fixes: 4acc45db71158 ("net: hsr: use hlist_head instead of list_head for mac addresses")
Cc: Juhee Kang <claudiajkang@gmail.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoclk: qcom: clk-krait: fix wrong div2 functions
Christian Marangi [Tue, 8 Nov 2022 21:56:25 +0000 (22:56 +0100)]
clk: qcom: clk-krait: fix wrong div2 functions

[ Upstream commit d676d3a3717cf726d3affedbe5ba98fc4ccad7b3 ]

Currently div2 value is applied to the wrong bits. This is caused by a
bug in the code where the shift is done only for lpl, for anything
else the mask is not shifted to the correct bits.

Fix this by correctly shift if lpl is not supported.

Fixes: 4d7dc77babfe ("clk: qcom: Add support for Krait clocks")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221108215625.30186-1-ansuelsmth@gmail.com
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoclk: qcom: lpass-sc7180: Fix pm_runtime usage
Douglas Anderson [Fri, 4 Nov 2022 13:56:29 +0000 (06:56 -0700)]
clk: qcom: lpass-sc7180: Fix pm_runtime usage

[ Upstream commit ff1ccf59eaffd192efe21f7de9fb0c130faf1b1b ]

The sc7180 lpass clock controller's pm_runtime usage wasn't broken
quite as spectacularly as the sc7280's pm_runtime usage, but it was
still broken. Putting some printouts in at boot showed me this (with
serial console enabled, which makes the prints slow and thus changes
timing):
  [    3.109951] DOUG: my_pm_clk_resume, usage=1
  [    3.114767] DOUG: my_pm_clk_resume, usage=1
  [    3.664443] DOUG: my_pm_clk_suspend, usage=0
  [    3.897566] DOUG: my_pm_clk_suspend, usage=0
  [    3.910137] DOUG: my_pm_clk_resume, usage=1
  [    3.923217] DOUG: my_pm_clk_resume, usage=0
  [    4.440116] DOUG: my_pm_clk_suspend, usage=-1
  [    4.444982] DOUG: my_pm_clk_suspend, usage=0
  [   14.170501] DOUG: my_pm_clk_resume, usage=1
  [   14.176245] DOUG: my_pm_clk_resume, usage=0

...or this w/out serial console:
  [    0.556139] DOUG: my_pm_clk_resume, usage=1
  [    0.556279] DOUG: my_pm_clk_resume, usage=1
  [    1.058422] DOUG: my_pm_clk_suspend, usage=-1
  [    1.058464] DOUG: my_pm_clk_suspend, usage=0
  [    1.186250] DOUG: my_pm_clk_resume, usage=1
  [    1.186292] DOUG: my_pm_clk_resume, usage=0
  [    1.731536] DOUG: my_pm_clk_suspend, usage=-1
  [    1.731557] DOUG: my_pm_clk_suspend, usage=0
  [   10.288910] DOUG: my_pm_clk_resume, usage=1
  [   10.289496] DOUG: my_pm_clk_resume, usage=0

It seems to be doing roughly the right sequence of calls, but just
like with sc7280 this is more by luck than anything. Having a usage of
-1 is just not OK.

Let's fix this like we did with sc7280.

Signed-off-by: Douglas Anderson <dianders@chromium.org>
Fixes: ce8c195e652f ("clk: qcom: lpasscc: Introduce pm autosuspend for SC7180")
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221104064055.2.I49b25b9bda9430fc7ea21e5a708ca5a0aced2798@changeid
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoclk: qcom: lpass-sc7280: Fix pm_runtime usage
Douglas Anderson [Fri, 4 Nov 2022 13:56:28 +0000 (06:56 -0700)]
clk: qcom: lpass-sc7280: Fix pm_runtime usage

[ Upstream commit d470be3c4f30b4666e43eef6bab80f543563cdb0 ]

The pm_runtime usage in lpass-sc7280 was broken in quite a few
ways. Specifically:

1. At the end of probe it called "put" twice. This is a no-no and will
   end us up with a negative usage count. Even worse than calling
   "put" twice, it never called "get" once. Thus after bootup it could
   be seen that the runtime usage of the devices managed by this
   driver was -2.
2. In some error cases it manually called pm_runtime_disable() even
   though it had previously used devm_add_action_or_reset() to set
   this up to be called automatically. This meant that in these error
   cases we'd double-call pm_runtime_disable().
3. It forgot to call undo pm_runtime_use_autosuspend(), which can
   sometimes have subtle problems (and the docs specifically mention
   that you need to undo this function).

Overall the above seriously calls into question how this driver is
working. It seems like a combination of "it doesn't", "by luck", and
"because of the weirdness of runtime_pm". Specifically I put a
printout to the serial console every time the runtime suspend/resume
was called for the two devices created by this driver (I wrapped the
pm_clk calls). When I had serial console enabled, I found that the
calls got resumed at bootup (when the clk core probed and before our
double-put) and then never touched again. That's no good.
  [    0.829997] DOUG: my_pm_clk_resume, usage=1
  [    0.835487] DOUG: my_pm_clk_resume, usage=1

When I disabled serial console (speeding up boot), I got a different
pattern, which I guess (?) is better:
  [    0.089767] DOUG: my_pm_clk_resume, usage=1
  [    0.090507] DOUG: my_pm_clk_resume, usage=1
  [    0.151885] DOUG: my_pm_clk_suspend, usage=-2
  [    0.151914] DOUG: my_pm_clk_suspend, usage=-2
  [    1.825747] DOUG: my_pm_clk_resume, usage=-1
  [    1.825774] DOUG: my_pm_clk_resume, usage=-1
  [    1.888269] DOUG: my_pm_clk_suspend, usage=-2
  [    1.888282] DOUG: my_pm_clk_suspend, usage=-2

These different patterns have to do with the fact that the core PM
Runtime code really isn't designed to be robust to negative usage
counts and sometimes may happen to stumble upon a behavior that
happens to "work". For instance, you can see that
__pm_runtime_suspend() will treat any non-zero value (including
negative numbers) as if the device is in use.

In any case, let's fix the driver to be correct. We'll hold a
pm_runtime reference for the whole probe and then drop it (once!) at
the end. We'll get rid of manual pm_runtime_disable() calls in the
error handling. We'll also switch to devm_pm_runtime_enable(), which
magically handles undoing pm_runtime_use_autosuspend() as of commit
b4060db9251f ("PM: runtime: Have devm_pm_runtime_enable() handle
pm_runtime_dont_use_autosuspend()").

While we're at this, let's also use devm_pm_clk_create() instead of
rolling it ourselves.

Note that the above changes make it obvious that
lpassaudio_create_pm_clks() was doing more than just creating
clocks. It was also setting up pm_runtime parameters. Let's rename it.

All of these problems were found by code inspection. I started looking
at this driver because it was involved in a deadlock that I reported a
while ago [1]. Though I bisected the deadlock to commit 1b771839de05
("clk: qcom: gdsc: enable optional power domain support"), it was
never really clear why that patch affected it other than a luck of
timing changes. I'll also note that by fixing the timing (as done in
this change) we also seem to aboid the deadlock, which is a nice
benefit.

Also note that some of the fixes here are much the same type of stuff
that Dmitry did in commit 72cfc73f4663 ("clk: qcom: use
devm_pm_runtime_enable and devm_pm_clk_create"), but I guess
lpassaudiocc-sc7280.c didn't exist then.

[1] https://lore.kernel.org/r/20220922154354.2486595-1-dianders@chromium.org

Fixes: a9dd26639d05 ("clk: qcom: lpass: Add support for LPASS clock controller for SC7280")
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Bjorn Andersson <andersson@kernel.org>
Link: https://lore.kernel.org/r/20221104064055.1.I00a0e4564a25489e85328ec41636497775627564@changeid
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agoregulator: core: fix module refcount leak in set_supply()
Yang Yingliang [Thu, 1 Dec 2022 12:27:05 +0000 (20:27 +0800)]
regulator: core: fix module refcount leak in set_supply()

[ Upstream commit da46ee19cbd8344d6860816b4827a7ce95764867 ]

If create_regulator() fails in set_supply(), the module refcount
needs be put to keep refcount balanced.

Fixes: e2c09ae7a74d ("regulator: core: Increase refcount for regulator supply's module")
Signed-off-by: Yang Yingliang <yangyingliang@huawei.com>
Link: https://lore.kernel.org/r/20221201122706.4055992-2-yangyingliang@huawei.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agomt76: mt7915: Fix PCI device refcount leak in mt7915_pci_init_hif2()
Xiongfeng Wang [Fri, 25 Nov 2022 02:58:31 +0000 (10:58 +0800)]
mt76: mt7915: Fix PCI device refcount leak in mt7915_pci_init_hif2()

[ Upstream commit 5938196cc188ba4323bc6357f5ac55127d715888 ]

As comment of pci_get_device() says, it returns a pci_device with its
refcount increased. We need to call pci_dev_put() to decrease the
refcount. Save the return value of pci_get_device() and call
pci_dev_put() to decrease the refcount.

Fixes: 9093cfff72e3 ("mt76: mt7915: add support for using a secondary PCIe link for gen1")
Fixes: 2e30db0dde61 ("mt76: mt7915: add device id for mt7916")
Signed-off-by: Xiongfeng Wang <wangxiongfeng2@huawei.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: mt76: do not send firmware FW_FEATURE_NON_DL region
Deren Wu [Thu, 24 Nov 2022 14:20:38 +0000 (22:20 +0800)]
wifi: mt76: do not send firmware FW_FEATURE_NON_DL region

[ Upstream commit f37f76d43865c58cb96aa13c87164abb41f22d0b ]

skip invalid section to avoid potential risks

Fixes: 23bdc5d8cadf ("wifi: mt76: mt7921: introduce Country Location Control support")
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: mt76: mt7921: Add missing __packed annotation of struct mt7921_clc
Deren Wu [Mon, 28 Nov 2022 07:04:21 +0000 (15:04 +0800)]
wifi: mt76: mt7921: Add missing __packed annotation of struct mt7921_clc

[ Upstream commit e5c6bc6f19d8c293f86b347cddab54d5a3830b38 ]

Add __packed annotation to avoid potential CLC parsing error

Fixes: 23bdc5d8cadf ("wifi: mt76: mt7921: introduce Country Location Control support")
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: mt76: fix coverity overrun-call in mt76_get_txpower()
Deren Wu [Sun, 27 Nov 2022 02:35:37 +0000 (10:35 +0800)]
wifi: mt76: fix coverity overrun-call in mt76_get_txpower()

[ Upstream commit 03dd0d49de7db680a856fa566963bb8421f46368 ]

Make sure the nss is valid for nss_delta array. Return zero
if the index is invalid.

Coverity message:
Event overrun-call: Overrunning callee's array of size 4 by passing
argument "n_chains" (which evaluates to 15) in call to
"mt76_tx_power_nss_delta".
int delta = mt76_tx_power_nss_delta(n_chains);

Fixes: 07cda406308b ("mt76: fix rounding issues on converting per-chain and combined txpower")
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: mt76: mt7921: fix wrong power after multiple SAR set
YN Chen [Wed, 16 Nov 2022 14:43:02 +0000 (22:43 +0800)]
wifi: mt76: mt7921: fix wrong power after multiple SAR set

[ Upstream commit 7eefb93d4a6fbccd859e538d208c50fd10b44cb7 ]

We should update CLC config before SAR set to synchronize all
related settings.

Fixes: 23bdc5d8cadf ("wifi: mt76: mt7921: introduce Country Location Control support")
Signed-off-by: YN Chen <YN.Chen@mediatek.com>
Signed-off-by: Deren Wu <deren.wu@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: mt76: mt7915: Fix chainmask calculation on mt7915 DBDC
Nicolas Cavallari [Thu, 10 Nov 2022 15:39:51 +0000 (16:39 +0100)]
wifi: mt76: mt7915: Fix chainmask calculation on mt7915 DBDC

[ Upstream commit de147cc28985a2a09e5d6d179fc5ef59b22fc058 ]

mt7915 does not have a per-band number of chains unlike the other chips,
it only has a total number of chains.  Yet the current code would
consider the total number as a per-band number.

For example, it would report that a 2x2 + 2x2 DBDC card have 4 chains on
each band and set chainmask to 0b1111 for the first interface and
0b11110000 for the second.

Fixes: 99ad32a4ca3a ("mt76: mt7915: add support for MT7986")
Co-developed-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Nicolas Cavallari <nicolas.cavallari@green-communications.fr>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: mt76: mt7915: rework eeprom tx paths and streams init
Shayne Chen [Fri, 30 Sep 2022 15:13:10 +0000 (23:13 +0800)]
wifi: mt76: mt7915: rework eeprom tx paths and streams init

[ Upstream commit a7ec8bcf00034ce84d4c9a15dffd7577fbed4db2 ]

Rework tx paths and streams init part to improve readability, and make
sure that the available tx streams should be smaller than or equal to
the available tx paths.

Signed-off-by: Shayne Chen <shayne.chen@mediatek.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Stable-dep-of: de147cc28985 ("wifi: mt76: mt7915: Fix chainmask calculation on mt7915 DBDC")
Signed-off-by: Sasha Levin <sashal@kernel.org>
18 months agowifi: mt76: mt7921: fix reporting of TX AGGR histogram
Lorenzo Bianconi [Wed, 2 Nov 2022 12:46:50 +0000 (13:46 +0100)]
wifi: mt76: mt7921: fix reporting of TX AGGR histogram

[ Upstream commit 028b4f22b37b88821fd87b56ce47b180583c774e ]

Similar to mt7915, fix stats clash between bins [4-7] in 802.11 tx
aggregation histogram.

Fixes: 163f4d22c118d ("mt76: mt7921: add MAC support")
Signed-off-by: Lorenzo Bianconi <lorenzo@kernel.org>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Sasha Levin <sashal@kernel.org>