Jeff Kirsher says:
====================
Intel Wired LAN Driver Fixes 2019-11-08
This series contains fixes to igb, igc, ixgbe, i40e, iavf and ice
drivers.
Colin Ian King fixes a potentially wrap-around counter in a for-loop.
Nick fixes the default ITR values for the iavf driver to 50 usecs
interval.
Arkadiusz fixes 'ethtool -m' for X722 devices where the correct value
cannot be obtained from the firmware, so add X722 to the check to ensure
the wrong value is not returned.
Jake fixes igb and igc drivers in their implementation of launch time
support by declaring skb->tstamp value as ktime_t instead of s64.
Magnus fixes ixgbe and i40e where the need_wakeup flag for transmit may
not be set for AF_XDP sockets that are only used to send packets.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
The need_wakeup flag for Tx might not be set for AF_XDP sockets that
are only used to send packets. This happens if there is at least one
outstanding packet that has not been completed by the hardware and we
get that corresponding completion (which will not generate an
interrupt since interrupts are disabled in the napi poll loop) between
the time we stopped processing the Tx completions and interrupts are
enabled again. In this case, the need_wakeup flag will have been
cleared at the end of the Tx completion processing as we believe we
will get an interrupt from the outstanding completion at a later point
in time. But if this completion interrupt occurs before interrupts
are enable, we lose it and should at that point really have set the
need_wakeup flag since there are no more outstanding completions that
can generate an interrupt to continue the processing. When this
happens, user space will see a Tx queue need_wakeup of 0 and skip
issuing a syscall, which means will never get into the Tx processing
again and we have a deadlock.
This patch introduces a quick fix for this issue by just setting the
need_wakeup flag for Tx to 1 all the time. I am working on a proper
fix for this that will toggle the flag appropriately, but it is more
challenging than I anticipated and I am afraid that this patch will
not be completed before the merge window closes, therefore this easier
fix for now. This fix has a negative performance impact in the range
of 0% to 4%. Towards the higher end of the scale if you have driver
and application on the same core and issue a lot of packets, and
towards no negative impact if you use two cores, lower transmission
speeds and/or a workload that also receives packets.
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
The need_wakeup flag for Tx might not be set for AF_XDP sockets that
are only used to send packets. This happens if there is at least one
outstanding packet that has not been completed by the hardware and we
get that corresponding completion (which will not generate an
interrupt since interrupts are disabled in the napi poll loop) between
the time we stopped processing the Tx completions and interrupts are
enabled again. In this case, the need_wakeup flag will have been
cleared at the end of the Tx completion processing as we believe we
will get an interrupt from the outstanding completion at a later point
in time. But if this completion interrupt occurs before interrupts
are enable, we lose it and should at that point really have set the
need_wakeup flag since there are no more outstanding completions that
can generate an interrupt to continue the processing. When this
happens, user space will see a Tx queue need_wakeup of 0 and skip
issuing a syscall, which means will never get into the Tx processing
again and we have a deadlock.
This patch introduces a quick fix for this issue by just setting the
need_wakeup flag for Tx to 1 all the time. I am working on a proper
fix for this that will toggle the flag appropriately, but it is more
challenging than I anticipated and I am afraid that this patch will
not be completed before the merge window closes, therefore this easier
fix for now. This fix has a negative performance impact in the range
of 0% to 4%. Towards the higher end of the scale if you have driver
and application on the same core and issue a lot of packets, and
towards no negative impact if you use two cores, lower transmission
speeds and/or a workload that also receives packets.
Signed-off-by: Magnus Karlsson <magnus.karlsson@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
When implementing launch time support in the igb and igc drivers, the
skb->tstamp value is assumed to be a s64, but it's declared as a ktime_t
value.
Although ktime_t is typedef'd to s64 it wasn't always, and the kernel
provides accessors for ktime_t values.
Use the ktime_to_timespec64 and ktime_set accessors instead of directly
assuming that the variable is always an s64.
This improves portability if the code is ever moved to another kernel
version, or if the definition of ktime_t ever changes again in the
future.
Signed-off-by: Jacob Keller <jacob.e.keller@intel.com>
Acked-by: Vinicius Costa Gomes <vinicius.gomes@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
This patch contains fix for a problem with command:
'ethtool -m <dev>'
which breaks functionality of:
'ethtool <dev>'
when called on X722 NIC
Disallowed update of link phy_types on X722 NIC
Currently correct value cannot be obtained from FW
Previously wrong value returned by FW was used and was
a root cause for incorrect output of 'ethtool <dev>' command
Signed-off-by: Arkadiusz Kubalewski <arkadiusz.kubalewski@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Since commit 92418fb147 ("i40e/i40evf: Use usec value instead of reg
value for ITR defines") the driver tracks the interrupt throttling
intervals in single usec units, although the actual ITRN registers are
programmed in 2 usec units. Most register programming flows in the driver
correctly handle the conversion, although it is currently not applied when
the registers are initialized to their default values. Most of the time
this doesn't present a problem since the default values are usually
immediately overwritten through the standard adaptive throttling mechanism,
or updated manually by the user, but if adaptive throttling is disabled and
the interval values are left alone then the incorrect value will persist.
Since the intended default interval of 50 usecs (vs. 100 usecs as
programmed) performs better for most traffic workloads, this can lead to
performance regressions.
This patch adds the correct conversion when writing the initial values to
the ITRN registers.
Signed-off-by: Nicholas Nunley <nicholas.d.nunley@intel.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Currently the for-loop counter i is a u8 however it is being checked
against a maximum value hw->num_tx_sched_layers which is a u16. Hence
there is a potential wrap-around of counter i back to zero if
hw->num_tx_sched_layers is greater than 255. Fix this by making i
a u16.
Addresses-Coverity: ("Infinite loop")
Fixes: b36c598c99 ("ice: Updates to Tx scheduler code")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Tested-by: Andrew Bowers <andrewx.bowers@intel.com>
Signed-off-by: Jeff Kirsher <jeffrey.t.kirsher@intel.com>
Octeon's input ring-buffer entry has 14 bits-wide size field, so to account
for second possible VLAN header max_mtu must be further reduced.
Fixes: 109cc16526 ("ethernet/cavium: use core min/max MTU checking")
Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch fixes the problem of the spin locks, originally
meant for the netpoll path of hns driver, causing deadlock in
the normal NAPI poll path. The issue happened due to the presence
of the stray leftover spin lock code related to the netpoll,
whose support was earlier removed from the HNS[1], got activated
due to enabling of NET_POLL_CONTROLLER switch.
Earlier background:
The netpoll handling code originally had this bug(as identified
by Marc Zyngier[2]) of wrong spin lock API being used which did
not disable the interrupts and hence could cause locking issues.
i.e. if the lock were first acquired in context to thread like
'ip' util and this lock if ever got later acquired again in
context to the interrupt context like TX/RX (Interrupts could
always pre-empt the lock holding task and acquire the lock again)
and hence could cause deadlock.
Proposed Solution:
1. If the netpoll was enabled in the HNS driver, which is not
right now, we could have simply used spin_[un]lock_irqsave()
2. But as netpoll is disabled, therefore, it is best to get rid
of the existing locks and stray code for now. This should
solve the problem reported by Marc.
[1] https://git.kernel.org/torvalds/c/4bd2c03be7
[2] https://patchwork.ozlabs.org/patch/1189139/
Fixes: 4bd2c03be7 ("net: hns: remove ndo_poll_controller")
Cc: lipeng <lipeng321@huawei.com>
Cc: Yisen Zhuang <yisen.zhuang@huawei.com>
Cc: Eric Dumazet <edumazet@google.com>
Cc: David S. Miller <davem@davemloft.net>
Reported-by: Marc Zyngier <maz@kernel.org>
Acked-by: Marc Zyngier <maz@kernel.org>
Tested-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Salil Mehta <salil.mehta@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A malicious device may give half an answer when asked
for its MTU. The driver will proceed after this with
a garbage MTU. Anything but a complete answer must be treated
as an error.
V2: used sizeof as request by Alexander
Reported-and-tested-by: syzbot+0631d878823ce2411636@syzkaller.appspotmail.com
Signed-off-by: Oliver Neukum <oneukum@suse.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since old firmware does not support HCLGE_OPC_PF_RST_DONE, it will
return -EOPNOTSUPP to the driver when received this command. So
for this case, it should just print a warning and return success
to the caller.
Fixes: 72e2fb0799 ("net: hns3: clear reset interrupt status in hclge_irq_handle()")
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEGhZs6bAKwk/OTgTpSD+KveBX+j4FAl3DQ0sACgkQSD+KveBX
+j4nKgf/XnUZGR4PHGcEfalc/p67J7jmi7SN5Xuts8uyUTgqMGni/SkWRl+S1mQy
8dl5kZK/JxBhAzzlVIcXpcmkfeXydcrm1NHMPCOMsnUrAWAEP0lE1lfLBkX1zWPi
OHb3+nziJHc0s4AI/Haca7JOj8HjBSOPpv3Udg6f53leizo31pK2FSiBASc5bRoX
EUNMD3eFAap7fjBMVLzO/JeQXxgjaqb9qU3AKopy34cg+AInZ+ZH2S++aoT7RG3J
MCPOnrSxLmhA4BUu5I81vfgFNxxPP65vI7NPIhSMY7+pOQJNPIc1p08XvIpQ9X2o
5o/rQ9tVxq2NZqcMB+b6tiHVFYu7Kw==
=OmLZ
-----END PGP SIGNATURE-----
Merge tag 'mlx5-fixes-2019-11-06' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahamees says:
====================
Mellanox, mlx5 fixes 2019-11-06
This series introduces some fixes to mlx5 driver.
Please pull and let me know if there is any problem.
No -stable this time.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Functions like phy_modify_paged() read the current page, on Realtek
PHY's this means reading the value of register 0x1f. Add special
handling for reading this register, similar to what we do already
in r8168g_mdio_write(). Currently we read a random value that by
chance seems to be 0 always.
Fixes: a2928d2864 ("r8169: use paged versions of phylib MDIO access functions")
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
IOC bit must be only set in the last descriptor. Move the logic up a
little bit to make sure it's set in the correct descriptor.
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When using addressing > 32 bits the TSO first descriptor only has the
header so we can't set the payload field for this descriptor. Let's
reset the variable so that buffer 2 value is zero.
Fixes: a993db88d1 ("net: stmmac: Enable support for > 32 Bits addressing in XGMAC")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, stmmac_rx() is counting the number of descriptors but it
should count the number of packets as specified by the NAPI limit.
Fix this.
Fixes: ec222003bd ("net: stmmac: Prepare to add Split Header support")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
MMC interrupts were being enabled, which is not what we want because it
will lead to a storm of interrupts that are not handled at all. Fix it
by disabling all MMC interrupts for XGMAC.
Fixes: b6cdf09f51 ("net: stmmac: xgmac: Implement MMC counters")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When in AVB mode we need to disable flow control to prevent MAC from
pausing in TX side.
Fixes: ec6ea8e3ee ("net: stmmac: Add CBS support in XGMAC2")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix incorrect precedence of operators. For reference: AV implies AV
Feature but RAV implies only RX side AV Feature. As we want full AV
features we need to check RAV.
Fixes: c2b69474d6 ("net: stmmac: xgmac: Correct RAVSEL field interpretation")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we change between Transmission Scheduling Algorithms, we need to
clear previous values so that the new chosen algorithm is correctly
selected.
Fixes: ec6ea8e3ee ("net: stmmac: Add CBS support in XGMAC2")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Split Header length is only available when L34T == 0. Fix this by
correctly checking if L34T is zero before trying to get Header length.
Fixes: 67afd6d1cf ("net: stmmac: Add Split Header support and enable it in XGMAC cores")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In L2 tests that filter packets by destination MAC address we need to
prevent false positives that can occur if we add an address that
collides with the existing ones.
To fix this, lets manually check if the new address to be added is
already present in the NIC and use a different one if so. For Hash
filtering this also envolves converting the address to the hash.
Fixes: 091810dbde ("net: stmmac: Introduce selftests support")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The bitrev32 function returns an u32 var, not an int. Fix it.
Fixes: 0efedbf11f ("net: stmmac: xgmac: Fix XGMAC selftests")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The bitrev32 function returns an u32 var, not an int. Fix it.
Fixes: 477286b53f ("stmmac: add GMAC4 core support")
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The "read-modify-write register index" function is declared with a
confusing prototype: the "mask" and "reg" arguments are swapped.
Fortunately, this does not affect callers so far. Both arguments are
u32, and the wrapper macros (ocelot_rmw_ix etc) have the arguments in
the correct order (the one from ocelot_io.c).
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
lag_upper_info may be NULL on slave removal.
Fixes: dc96ee3730 ("net: mscc: ocelot: add bonding support")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The check that the event is actually for this device should be moved
from the "port" handler to the net device handler.
Otherwise the port handler will deny bonding configuration for other
net devices in the same system (like enetc in the LS1028A) that don't
have the lag_upper_info->tx_type restriction that ocelot has.
Fixes: dc96ee3730 ("net: mscc: ocelot: add bonding support")
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com>
Signed-off-by: Vladimir Oltean <olteanv@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
For vlan push action, if eswitch flow source capability is enabled, flow
source value compared with MLX5_VPORT_UPLINK enum, to determine uplink
port. This lead to syndrome in dmesg if try to add vlan push action.
For example:
$ tc filter add dev vxlan0 ingress protocol ip prio 1 flower \
enc_dst_port 4789 \
action tunnel_key unset pipe \
action vlan push id 20 pipe \
action mirred egress redirect dev ens1f0_0
$ dmesg
...
[ 2456.883693] mlx5_core 0000:82:00.0: mlx5_cmd_check:756:(pid 5273): SET_FLOW_TABLE_ENTRY(0x936) op_mod(0x0) failed, status bad parameter(0x3), syndrome (0xa9c090)
Use the correct enum value MLX5_FLOW_CONTEXT_FLOW_SOURCE_UPLINK.
Fixes: bb204dcf39fe ("net/mlx5e: Determine source port properly for vlan push action")
Signed-off-by: Dmytro Linkin <dmitrolin@mellanox.com>
Reviewed-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The rewrite data was no freed.
Fixes: 9db810ed2d ("net/mlx5: DR, Expose steering action functionality")
Signed-off-by: Alex Vesker <valex@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The value is already the calculation so remove the log prefix.
Fixes: e52c280240 ("net/mlx5: E-Switch, Add chains and priorities")
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Eli Britstein <elibr@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The phy_init_hw() function may reset the PHY to a configuration
that does not match manual network settings stored in the phydev
structure. If the phy state machine is polled rather than event
driven this can create a timing hazard where the phy state machine
might alter the settings stored in the phydev structure from the
value read from the BMCR.
This commit follows invocations of phy_init_hw() by the bcmgenet
driver with invocations of the genphy_config_aneg() function to
ensure that the BMCR is written to match the settings held in the
phydev structure. This prevents the risk of manual settings being
accidentally altered.
Fixes: 1c1008c793 ("net: bcmgenet: add main driver file")
Signed-off-by: Doug Berger <opendmb@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This reverts commit 1f51548627.
This commit improved the chances of the umac resetting cleanly by
ensuring that the PHY was restored to its normal operation prior
to resetting the umac. However, there were still cases when the
PHY might not be driving a Tx clock to the umac during this window
(e.g. when the PHY detects no link).
The previous commit now ensures that the unimac receives clocks
from the MAC during its reset window so this commit is no longer
needed. This commit also has an unintended negative impact on the
MDIO performance of the UniMAC MDIO interface because it is used
before the MDIO interrupts are reenabled, so it should be removed.
Signed-off-by: Doug Berger <opendmb@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As noted in commit 28c2d1a7a0 ("net: bcmgenet: enable loopback
during UniMAC sw_reset") the UniMAC must be clocked while sw_reset
is asserted for its state machines to reset cleanly.
The transmit and receive clocks used by the UniMAC are derived from
the signals used on its PHY interface. The bcmgenet MAC can be
configured to work with different PHY interfaces including MII,
GMII, RGMII, and Reverse MII on internal and external interfaces.
Unfortunately for the UniMAC, when configured for MII the Tx clock
is always driven from the PHY which places it outside of the direct
control of the MAC.
The earlier commit enabled a local loopback mode within the UniMAC
so that the receive clock would be derived from the transmit clock
which addressed the observed issue with an external GPHY disabling
it's Rx clock. However, when a Tx clock is not available this
loopback is insufficient.
This commit implements a workaround that leverages the fact that
the MAC can reliably generate all of its necessary clocking by
enterring the external GPHY RGMII interface mode with the UniMAC in
local loopback during the sw_reset interval. Unfortunately, this
has the undesirable side efect of the RGMII GTXCLK signal being
driven during the same window.
In most configurations this is a benign side effect as the signal
is either not routed to a pin or is already expected to drive the
pin. The one exception is when an external MII PHY is expected to
drive the same pin with its TX_CLK output creating output driver
contention.
This commit exploits the IEEE 802.3 clause 22 standard defined
isolate mode to force an external MII PHY to present a high
impedance on its TX_CLK output during the window to prevent any
contention at the pin.
The MII interface is used internally with the 40nm internal EPHY
which agressively disables its clocks for power savings leading to
incomplete resets of the UniMAC and many instabilities observed
over the years. The workaround of this commit is expected to put
an end to those problems.
Fixes: 1c1008c793 ("net: bcmgenet: add main driver file")
Signed-off-by: Doug Berger <opendmb@gmail.com>
Acked-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
During the exit/unregistration process of the RmNet driver, the function
rmnet_unregister_real_device() is called to handle freeing the driver's
internal state and removing the RX handler on the underlying physical
device. However, the order of operations this function performs is wrong
and can lead to a use after free of the rmnet_port structure.
Before calling netdev_rx_handler_unregister(), this port structure is
freed with kfree(). If packets are received on any RmNet devices before
synchronize_net() completes, they will attempt to use this already-freed
port structure when processing the packet. As such, before cleaning up any
other internal state, the RX handler must be unregistered in order to
guarantee that no further packets will arrive on the device.
Fixes: ceed73a2cf ("drivers: net: ethernet: qualcomm: rmnet: Initial implementation")
Signed-off-by: Sean Tranchetti <stranche@codeaurora.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
The reason for the pre-allocation of one CQE is to enable resizing of
the CQ.
Fix comment accordingly.
Signed-off-by: Dotan Barak <dotanb@dev.mellanox.co.il>
Signed-off-by: Eli Cohen <eli@mellanox.co.il>
Signed-off-by: Vladimir Sokolovsky <vlad@mellanox.com>
Signed-off-by: Yuval Shaia <yuval.shaia@oracle.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
With the DSA core doing the call to dsa_port_disable() we do not need to
do that within the driver itself. This could cause an use after free
since past dsa_unregister_switch() we should not be accessing any
dsa_switch internal structures.
Fixes: 0394a63acf ("net: dsa: enable and disable all ports")
Signed-off-by: Florian Fainelli <f.fainelli@gmail.com>
Reviewed-by: Vivien Didelot <vivien.didelot@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch corrects the SPDX License Identifier style in
header files related to Hisilicon network devices. For C header files
Documentation/process/license-rules.rst mandates C-like comments
(opposed to C source files where C++ style should be used)
Changes made by using a script provided by Joe Perches here:
https://lkml.org/lkml/2019/2/7/46.
Suggested-by: Joe Perches <joe@perches.com>
Signed-off-by: Nishad Kamdar <nishadkamdar@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since de77ecd4ef ("bonding: improve link-status update in
mii-monitoring"), the bonding driver has utilized two separate variables
to indicate the next link state a particular slave should transition to.
Each is used to communicate to a different portion of the link state
change commit logic; one to the bond_miimon_commit function itself, and
another to the state transition logic.
Unfortunately, the two variables can become unsynchronized,
resulting in incorrect link state transitions within bonding. This can
cause slaves to become stuck in an incorrect link state until a
subsequent carrier state transition.
The issue occurs when a special case in bond_slave_netdev_event
sets slave->link directly to BOND_LINK_FAIL. On the next pass through
bond_miimon_inspect after the slave goes carrier up, the BOND_LINK_FAIL
case will set the proposed next state (link_new_state) to BOND_LINK_UP,
but the new_link to BOND_LINK_DOWN. The setting of the final link state
from new_link comes after that from link_new_state, and so the slave
will end up incorrectly in _DOWN state.
Resolve this by combining the two variables into one.
Reported-by: Aleksei Zakharov <zakharov.a.g@yandex.ru>
Reported-by: Sha Zhang <zhangsha.zhang@huawei.com>
Cc: Mahesh Bandewar <maheshb@google.com>
Fixes: de77ecd4ef ("bonding: improve link-status update in mii-monitoring")
Signed-off-by: Jay Vosburgh <jay.vosburgh@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In mcp251x_restart_work_handler() the variable to stop the interrupt
handler (priv->force_quit) is reset after the chip is restarted and thus
a interrupt might occur.
This patch fixes the potential race condition by resetting force_quit
before enabling interrupts.
Signed-off-by: Timo Schlüßler <schluessler@krause.de>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
While the ti_hecc has interrupts to report when the error counters increase
to a certain level and which change state it doesn't handle the case that
the error counters go down again, so the reported state can actually be
wrong. Since there is no interrupt for that, do update state based on the
error counters, when the state is not error active and goes down again.
Signed-off-by: Jeroen Hofstee <jhofstee@victronenergy.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The HECC_CANES register handles the flags specially, it only updates
the flags after a one is written to them. Since the interrupt for
frame errors is not enabled an old error can hence been seen when a
state interrupt arrives. For example if the device is not connected
to the CAN-bus the error warning interrupt will have HECC_CANES
indicating there is no ack. The error passive interrupt thereafter
will have HECC_CANES flagging that there is a warning level. And if
thereafter there is a message successfully send HECC_CANES points to
an error passive event, while in reality it became error warning
again. In summary, the state is not always reported correctly.
So handle the state changes and frame errors separately. The state
changes are now based on the interrupt flags and handled directly
when they occur. The reporting of the frame errors is still done as
before, as a side effect of another interrupt.
note: the hecc_clear_bit will do a read, modify, write. So it will
not only clear the bit, but also reset all other bits being set as
a side affect, hence it is replaced with only clearing the flags.
note: The HECC_CANMC_CCR is no longer cleared in the state change
interrupt, it is completely unrelated.
And use net_ratelimit to make checkpatch happy.
Signed-off-by: Jeroen Hofstee <jhofstee@victronenergy.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
When the rx FIFO overflows the ti_hecc would silently drop them since
the overwrite protection is enabled for all mailboxes. So disable it for
the lowest priority mailbox and return a proper error value when receive
message lost is set. Drop the message itself in that case, since it
might be partially updated.
Signed-off-by: Jeroen Hofstee <jhofstee@victronenergy.com>
Acked-by: Jeroen Hofstee <jhofstee@victronenergy.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Release the mailbox after reading it, so it can be reused a bit earlier.
Since "can: rx-offload: continue on error" all pending message bits are
cleared directly, so remove clearing them in ti_hecc.
Suggested-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Jeroen Hofstee <jhofstee@victronenergy.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The HECC_CANMIM is set in the xmit path and cleared in the interrupt.
Since this is done with a read, modify, write action the register might
end up with some more MIM enabled then intended, since it is not
protected. That doesn't matter at all, since the tx interrupt disables
the mailbox with HECC_CANME (while holding a spinlock). So lets just
always keep MIM set.
While at it, since the mailbox direction never changes, don't set it
every time a message is send, ti_hecc_reset() already sets them to tx.
Signed-off-by: Jeroen Hofstee <jhofstee@victronenergy.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
When the interface goes down, the CPK should no longer take an active
part in the CAN-bus communication, like sending acks and error frames.
So enable configuration mode in ti_hecc_stop, so the CPK is no longer
active.
When a transceiver switch is present the acks and errors don't make it
to the bus, but disabling the CPK then does prevent oddities, like
ti_hecc_reset() failing, since the CPK can become bus-off and starts
counting the 11 bit recessive bits, which seems to block the reset. It
can also cause invalid interrupts and disrupt the CAN-bus, since
transmission can be stopped in the middle of a message, by disabling the
tranceiver while the CPK is sending.
Since the CPK is disabled after normal power on, it is typically only
seen when the interface is restarted.
Signed-off-by: Jeroen Hofstee <jhofstee@victronenergy.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The call to can_rx_offload_queue_sorted() may fail and return an error
(in the current implementation due to resource shortage). The passed skb
is consumed.
This patch adds incrementing of the appropriate error counters to let
the device statistics reflect that there's a problem.
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
The call to can_rx_offload_queue_sorted() may fail and return an error
(in the current implementation due to resource shortage). The passed skb
is consumed.
This patch adds incrementing of the appropriate error counters to let
the device statistics reflect that there's a problem.
Reported-by: Martin Hundebøll <martin@geanix.com>
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
