zhengbin [Tue, 12 Nov 2019 07:08:40 +0000 (15:08 +0800)]
cxgb4: make function 'cxgb4_mqprio_free_hw_resources' static
Fix sparse warnings:
drivers/net/ethernet/chelsio/cxgb4/cxgb4_tc_mqprio.c:242:6: warning: symbol 'cxgb4_mqprio_free_hw_resources' was not declared. Should it be static?
Reported-by: Hulk Robot <hulkci@huawei.com> Fixes: 2d0cb84dd973 ("cxgb4: add ETHOFLD hardware queue support") Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
zhengbin [Tue, 12 Nov 2019 06:59:42 +0000 (14:59 +0800)]
net: atlantic: make function 'aq_ethtool_get_priv_flags', 'aq_ethtool_set_priv_flags' static
Fix sparse warnings:
drivers/net/ethernet/aquantia/atlantic/aq_ethtool.c:706:5: warning: symbol 'aq_ethtool_get_priv_flags' was not declared. Should it be static?
drivers/net/ethernet/aquantia/atlantic/aq_ethtool.c:713:5: warning: symbol 'aq_ethtool_set_priv_flags' was not declared. Should it be static?
Reported-by: Hulk Robot <hulkci@huawei.com> Fixes: ea4b4d7fc106 ("net: atlantic: loopback tests via private flags") Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
zhengbin [Tue, 12 Nov 2019 06:59:41 +0000 (14:59 +0800)]
net: atlantic: make symbol 'aq_pm_ops' static
Fix sparse warnings:
drivers/net/ethernet/aquantia/atlantic/aq_pci_func.c:426:25: warning: symbol 'aq_pm_ops' was not declared. Should it be static?
Reported-by: Hulk Robot <hulkci@huawei.com> Fixes: 8aaa112a57c1 ("net: atlantic: refactoring pm logic") Signed-off-by: zhengbin <zhengbin13@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
David S. Miller [Tue, 12 Nov 2019 18:54:02 +0000 (10:54 -0800)]
Merge branch 'mlxsw-Add-extended-ACK-for-EMADs'
Ido Schimmel says:
====================
mlxsw: Add extended ACK for EMADs
Shalom says:
Ethernet Management Datagrams (EMADs) are Ethernet packets sent between
the driver and device's firmware. They are used to pass various
configurations to the device, but also to get events (e.g., port up)
from it. After the Ethernet header, these packets are built in a TLV
format.
Up until now, whenever the driver issued an erroneous register access it
only got an error code indicating a bad parameter was used. This patch
set adds a new TLV (string TLV) that can be used by the firmware to
encode a 128 character string describing the error. The new TLV is
allocated by the driver and set to zeros. In case of error, the driver
will check the length of the string in the response and report it using
devlink hwerr tracepoint.
Example:
$ perf record -a -q -e devlink:devlink_hwerr &
$ pkill -2 perf
$ perf script -F trace:event,trace | grep hwerr
devlink:devlink_hwerr: bus_name=pci dev_name=0000:03:00.0 driver_name=mlxsw_spectrum err=7 (tid=9913892d00001593,reg_id=8018(rauhtd)) bad parameter (inside er_rauhtd_write_query(), num_rec=32 is over the maximum number of records supported)
Patch #1 parses the offsets of the different TLVs in incoming EMADs and
stores them in the skb's control block. This makes it easier to later
add new TLVs.
Patches #2-#3 remove deprecated TLVs and add string TLV definition.
Patches #4-#7 gradually add support for the new string TLV.
v2:
* Use existing devlink hwerr tracepoint to report the error string,
instead of printing it to kernel log
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Shalom Toledo [Tue, 12 Nov 2019 06:48:24 +0000 (08:48 +0200)]
mlxsw: core: Parse TLVs' offsets of incoming EMADs
Until now the code assumes a fixed structure which makes it difficult to
support EMADs with and without new TLVs.
Make it more generic by parsing the TLVs when the EMADs are received and
store the offset to the different TLVs in the control block. Using these
offsets to extract information from the EMADs without relying on a specific
structure.
Signed-off-by: Shalom Toledo <shalomt@mellanox.com> Acked-by: Jiri Pirko <jiri@mellanox.com> Signed-off-by: Ido Schimmel <idosch@mellanox.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Mao Wenan [Tue, 12 Nov 2019 06:33:58 +0000 (14:33 +0800)]
net: ethernet: ti: Add dependency for TI_DAVINCI_EMAC
If TI_DAVINCI_EMAC=y and GENERIC_ALLOCATOR is not set,
below erros can be seen:
drivers/net/ethernet/ti/davinci_cpdma.o: In function `cpdma_desc_pool_destroy.isra.14':
davinci_cpdma.c:(.text+0x359): undefined reference to `gen_pool_size'
davinci_cpdma.c:(.text+0x365): undefined reference to `gen_pool_avail'
davinci_cpdma.c:(.text+0x373): undefined reference to `gen_pool_avail'
davinci_cpdma.c:(.text+0x37f): undefined reference to `gen_pool_size'
drivers/net/ethernet/ti/davinci_cpdma.o: In function `__cpdma_chan_free':
davinci_cpdma.c:(.text+0x4a2): undefined reference to `gen_pool_free_owner'
drivers/net/ethernet/ti/davinci_cpdma.o: In function `cpdma_chan_submit_si':
davinci_cpdma.c:(.text+0x66c): undefined reference to `gen_pool_alloc_algo_owner'
davinci_cpdma.c:(.text+0x805): undefined reference to `gen_pool_free_owner'
drivers/net/ethernet/ti/davinci_cpdma.o: In function `cpdma_ctlr_create':
davinci_cpdma.c:(.text+0xabd): undefined reference to `devm_gen_pool_create'
davinci_cpdma.c:(.text+0xb79): undefined reference to `gen_pool_add_owner'
drivers/net/ethernet/ti/davinci_cpdma.o: In function `cpdma_check_free_tx_desc':
davinci_cpdma.c:(.text+0x16c6): undefined reference to `gen_pool_avail'
This patch mades TI_DAVINCI_EMAC select GENERIC_ALLOCATOR.
Fixes: 99f629718272 ("net: ethernet: ti: cpsw: drop TI_DAVINCI_CPDMA config option") Signed-off-by: Mao Wenan <maowenan@huawei.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jose Abreu [Mon, 11 Nov 2019 14:42:38 +0000 (15:42 +0100)]
net: stmmac: Rework stmmac_rx()
This looks over-engineered. Let's use some helpers to get the buffer
length and hereby simplify the stmmac_rx() function. No performance drop
was seen with the new implementation.
Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jose Abreu [Mon, 11 Nov 2019 14:42:34 +0000 (15:42 +0100)]
net: stmmac: Fix sparse warning
The VID is converted to le16 so the variable must be __le16 type.
Reported-by: kbuild test robot <lkp@intel.com> Fixes: c7ab0b8088d7 ("net: stmmac: Fallback to VLAN Perfect filtering if HASH is not available") Signed-off-by: Jose Abreu <Jose.Abreu@synopsys.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Colin Ian King [Mon, 11 Nov 2019 12:33:34 +0000 (12:33 +0000)]
tipc: fix update of the uninitialized variable err
Variable err is not uninitialized and hence can potentially contain
any garbage value. This may cause an error when logical or'ing the
return values from the calls to functions crypto_aead_setauthsize or
crypto_aead_setkey. Fix this by setting err to the return of
crypto_aead_setauthsize rather than or'ing in the return into the
uninitialized variable
Addresses-Coverity: ("Uninitialized scalar variable") Fixes: fc1b6d6de220 ("tipc: introduce TIPC encryption & authentication") Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Side effect of some kbuild changes resulted in breaking the
documented way to build samples/bpf/.
This patch change the samples/bpf/Makefile to work again, when
invoking make from the subdir samples/bpf/. Also update the
documentation in README.rst, to reflect the new way to build.
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com> Acked-by: Daniel Borkmann <daniel@iogearbox.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Heiner Kallweit [Sun, 10 Nov 2019 15:13:35 +0000 (16:13 +0100)]
r8169: add support for RTL8117
Add support for chip version RTL8117. Settings have been copied from
Realtek's r8168 driver, there however chip ID 54a belongs to a chip
version called RTL8168FP. It was confirmed that RTL8117 works with
Realtek's driver, so both chip versions seem to be the same or at
least compatible.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
sfp: Allow slow to initialise GPON modules to work
Some GPON modules take longer than the SFF MSA specified time to
initialise and respond to transactions on the I2C bus for either
both 0x50 and 0x51, or 0x51 bus addresses. Technically these modules
are non-compliant with the SFP Multi-Source Agreement, they have
been around for some time, so are difficult to just ignore.
Most of the patch series is restructuring the code to make it more
readable, and split various things into separate functions.
We split the three state machines into three separate functions, and
re-arrange them to start probing the module as soon as a module has
been detected (without waiting for the network device.) We try to
read the module's EEPROM, retrying quickly for the first second, and
then once every five seconds for about a minute until we have read
the EEPROM. So that the kernel isn't entirely silent, we print a
message indicating that we're waiting for the module to respond after
the first second, or when all retries have expired.
Once the module ID has been read, we kick off a delayed work queue
which attempts to register the hwmon, retrying for up to a minute if
the monitoring parameters are unreadable; this allows us to proceed
with module initialisation independently of the hwmon state.
With high-power modules, we wait for the netdev to be attached before
switching the module power mode, and retry this in a similar way to
before until we have successfully read and written the EEPROM at 0x51.
We also move the handling of the TX_DISABLE signal entirely to the main
state machine, and avoid probing any on-board PHY while TX_FAULT is
set.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:07:35 +0000 (14:07 +0000)]
net: sfp: allow modules with slow diagnostics to probe
When a module is inserted, we attempt to read read the ID from address
0x50. Once we are able to read the ID, we immediately attempt to
initialise the hwmon support by reading from address 0x51. If this
fails, then we fall into error state, and assume that the module is
not usable.
Modules such as the ALCATELLUCENT 3FE46541AA use a real EEPROM for
I2C address 0x50, which responds immediately. However, address 0x51
is an emulated, which only becomes available once the on-board firmware
has booted. This prompts us to fall into the error state.
Since the module may be usable without diagnostics, arrange for the
hwmon probe independent of the rest of the SFP itself, retrying every
5s for up to about 60s for the monitoring to become available, and
print an error message if it doesn't become available.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:07:30 +0000 (14:07 +0000)]
net: sfp: allow sfp to probe slow to initialise GPON modules
Some GPON modules (e.g. Huawei MA5671A) take a significant amount of
time to start responding on the I2C bus, contary to the SFF
specifications.
Work around this by implementing a two-level timeout strategy, where
we initially quickly retry for the module, and then use a slower retry
after we exceed a maximum number of quick attempts.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:07:25 +0000 (14:07 +0000)]
net: sfp: move module insert reporting out of probe
Move the module insertion reporting out of the probe handling, but
after we have detected that the upstream has attached (since that is
whom we are reporting insertion to.)
Only report module removal if we had previously reported a module
insertion.
This gives cleaner semantics, and means we can probe the module before
we have an upstream attached.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:07:20 +0000 (14:07 +0000)]
net: sfp: split power mode switching from probe
Switch the power mode switching from the probe, so that we don't
repeatedly re-probe the SFP device if there is a problem accessing
the registers at I2C address 0x51.
In splitting this out, we can also fix a bug where we leave the module
in high-power mode when the upstream device is detached but the module
is still inserted.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:07:14 +0000 (14:07 +0000)]
net: sfp: track upstream's attachment state in state machine
Track the upstream's attachment state in the state machine rather than
maintaining a boolean, which ensures that we have a strict order of
ATTACH followed by an UP event - we can never believe that a newly
attached upstream will be anything but down.
Rearrange the order of state machines so we run the module state
machine after the upstream device's state machine, so the module state
machine can check the current state of the device and take action to
e.g. reset back to empty state when the upstream is detached.
This is to allow the module detection to run independently of the
network device becoming available.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:07:09 +0000 (14:07 +0000)]
net: sfp: ensure TX_FAULT has deasserted before probing the PHY
TX_FAULT should be deasserted to indicate that the module has completed
its initialisation. This may include the on-board PHY, so wait until
the module has deasserted TX_FAULT before probing the PHY.
This means that we need an extra state to handle a TX_FAULT that
remains set for longer than t_init, since using the existing handling
state would bypass the PHY probe.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:59 +0000 (14:06 +0000)]
net: sfp: eliminate mdelay() from PHY probe
Rather than using mdelay() to wait before probing the PHY (which holds
several locks, including the rtnl lock), add an extra wait state to
the state machine to introduce the 50ms delay without holding any
locks.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:54 +0000 (14:06 +0000)]
net: sfp: split the PHY probe from sfp_sm_mod_init()
Move the PHY probe into a separate function, splitting it from
sfp_sm_mod_init(). This will allow us to eliminate the 50ms mdelay()
inside the state machine.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:49 +0000 (14:06 +0000)]
net: sfp: control TX_DISABLE and phy only from main state machine
We initialise TX_DISABLE when the sfp cage is probed, and then
maintain its state in the main state machine. However, the module
state machine:
- negates it when detecting a newly inserted module when it's already
guaranteed to be negated.
- negates it when the module is removed, but the main state machine
will do this anyway.
Make TX_DISABLE entirely controlled by the main state machine.
The main state machine also probes the module for a PHY, and removes
the PHY when the the module is removed. Hence, removing the PHY in
sfp_sm_module_remove() is also redundant, and is a left-over from
when we tried to probe for the PHY from the module state machine.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:44 +0000 (14:06 +0000)]
net: sfp: avoid power switch on address-change modules
If the module indicates that it requires an address change sequence to
switch between address 0x50 and 0x51, which we don't support, we can't
write to the register that controls the power mode to switch to high
power mode. Warn the user that the module may not be functional in
this case, and don't try to change the power mode.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:39 +0000 (14:06 +0000)]
net: sfp: parse SFP power requirement earlier
Parse the SFP power requirement earlier, in preparation for moving the
power level setup code.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:33 +0000 (14:06 +0000)]
net: sfp: rename T_PROBE_WAIT to T_SERIAL
SFF-8472 rev 12.2 defines the time for the serial bus to become ready
using t_serial. Use this as our identifier for this timeout to make
it clear what we are referring to.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:28 +0000 (14:06 +0000)]
net: sfp: handle module remove outside state machine
Removing a module resets the module state machine back to its initial
state. Rather than explicitly handling this in every state, handle it
early on outside of the state machine.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:23 +0000 (14:06 +0000)]
net: sfp: rename sfp_sm_ins_next() as sfp_sm_mod_next()
sfp_sm_ins_next() modifies the module state machine. Change it's name
to reflect this.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:18 +0000 (14:06 +0000)]
net: sfp: move tx disable on device down to main state machine
Move the tx disable assertion on device down to the main state
machine.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:06:13 +0000 (14:06 +0000)]
net: sfp: move sfp sub-state machines into separate functions
Move the SFP sub-state machines out of the main state machine function,
in preparation for it doing a bit more with the device state. By doing
so, we ensure that our debug after the main state machine is always
printed.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Sun, 10 Nov 2019 14:04:11 +0000 (14:04 +0000)]
net: sfp: fix sfp_bus_put() kernel documentation
The kbuild test robot found a problem with htmldocs with the recent
change to the SFP interfaces. Fix the kernel documentation for
sfp_bus_put() which was missing an '@' before the argument name
description.
Fixes: 727b3668b730 ("net: sfp: rework upstream interface") Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Heiner Kallweit [Sun, 10 Nov 2019 13:44:54 +0000 (14:44 +0100)]
r8169: respect EEE user setting when restarting network
Currently, if network is re-started, we advertise all supported EEE
modes, thus potentially overriding a manual adjustment the user made
e.g. via ethtool. Be friendly to the user and preserve a manual
setting on network re-start.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Xin Long [Sun, 10 Nov 2019 04:26:21 +0000 (12:26 +0800)]
lwtunnel: ignore any TUNNEL_OPTIONS_PRESENT flags set by users
TUNNEL_OPTIONS_PRESENT (TUNNEL_GENEVE_OPT|TUNNEL_VXLAN_OPT|
TUNNEL_ERSPAN_OPT) flags should be set only according to
tb[LWTUNNEL_IP_OPTS], which is done in ip_tun_parse_opts().
When setting info key.tun_flags, the TUNNEL_OPTIONS_PRESENT
bits in tb[LWTUNNEL_IP(6)_FLAGS] passed from users should
be ignored.
While at it, replace all (TUNNEL_GENEVE_OPT|TUNNEL_VXLAN_OPT|
TUNNEL_ERSPAN_OPT) with 'TUNNEL_OPTIONS_PRESENT'.
Fixes: 3093fbe7ff4b ("route: Per route IP tunnel metadata via lightweight tunnel") Fixes: 32a2b002ce61 ("ipv6: route: per route IP tunnel metadata via lightweight tunnel") Signed-off-by: Xin Long <lucien.xin@gmail.com> Reviewed-by: Simon Horman <simon.horman@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Xin Long [Sun, 10 Nov 2019 04:21:18 +0000 (12:21 +0800)]
lwtunnel: get nlsize for erspan options properly
erspan v1 has OPT_ERSPAN_INDEX while erspan v2 has OPT_ERSPAN_DIR and
OPT_ERSPAN_HWID attributes, and they require different nlsize when
dumping.
So this patch is to get nlsize for erspan options properly according
to erspan version.
Fixes: b0a21810bd5e ("lwtunnel: add options setting and dumping for erspan") Signed-off-by: Xin Long <lucien.xin@gmail.com> Reviewed-by: Simon Horman <simon.horman@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Xin Long [Sun, 10 Nov 2019 04:16:22 +0000 (12:16 +0800)]
lwtunnel: change to use nla_parse_nested on new options
As the new options added in kernel, all should always use strict
parsing from the beginning with nla_parse_nested(), instead of
nla_parse_nested_deprecated().
Fixes: b0a21810bd5e ("lwtunnel: add options setting and dumping for erspan") Fixes: edf31cbb1502 ("lwtunnel: add options setting and dumping for vxlan") Fixes: 4ece47787077 ("lwtunnel: add options setting and dumping for geneve") Signed-off-by: Xin Long <lucien.xin@gmail.com> Reviewed-by: Simon Horman <simon.horman@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
Accomodate DSA front-end into Ocelot
After the nice "change-my-mind" discussion about Ocelot, Felix and
LS1028A (which can be read here: https://lkml.org/lkml/2019/6/21/630),
we have decided to take the route of reworking the Ocelot implementation
in a way that is DSA-compatible.
This is a large series, but hopefully is easy enough to digest, since it
contains mostly code refactoring. What needs to be changed:
- The struct net_device, phy_device needs to be isolated from Ocelot
private structures (struct ocelot, struct ocelot_port). These will
live as 1-to-1 equivalents to struct dsa_switch and struct dsa_port.
- The function prototypes need to be compatible with DSA (of course,
struct dsa_switch will become struct ocelot).
- The CPU port needs to be assigned via a higher-level API, not
hardcoded in the driver.
What is going to be interesting is that the new DSA front-end of Ocelot
will need to have features in lockstep with the DSA core itself. At the
moment, some more advanced tc offloading features of Ocelot (tc-flower,
etc) are not available in the DSA front-end due to lack of API in the
DSA core. It also means that Ocelot practically re-implements large
parts of DSA (although it is not a DSA switch per se) - see the FDB API
for example.
The code has been only compile-tested on Ocelot, since I don't have
access to any VSC7514 hardware. It was proven to work on NXP LS1028A,
which instantiates a DSA derivative of Ocelot. So I would like to ask
Alex Belloni if you could confirm this series causes no regression on
the Ocelot MIPS SoC.
The goal is to get this rework upstream as quickly as possible,
precisely because it is a large volume of code that risks gaining merge
conflicts if we keep it for too long.
This is but the first chunk of the LS1028A Felix DSA driver upstreaming.
For those who are interested, the concept can be seen on my private
Github repo, the user of this reworked Ocelot driver living under
drivers/net/dsa/vitesse/:
https://github.com/vladimiroltean/ls1028ardb-linux
====================
Acked-by: Horatiu Vultur <horatiu.vultur@microchip.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 13:03:01 +0000 (15:03 +0200)]
net: mscc: ocelot: don't hardcode the number of the CPU port
VSC7514 is a 10-port switch with 2 extra "CPU ports" (targets in the
queuing subsystem for terminating traffic locally).
There are 2 issues with hardcoding the CPU port as #10:
- It is not clear which snippets of the code are configuring something
for one of the CPU ports, and which snippets are just doing something
related to the number of physical ports.
- Actually any physical port can act as a CPU port connected to an
external CPU (in addition to the local CPU). This is called NPI mode
(Node Processor Interface) and is the way that the 6-port VSC9959
(Felix) switch is integrated inside NXP LS1028A (the "local management
CPU" functionality is not used there).
This patch makes it clear that the ocelot_bridge_stp_state_set function
operates on the CPU port (by making it an implicit member of the
bridging domain), and at the same time adds logic for the NPI port (aka
a physical port) to play the role of a CPU port (it shouldn't be part of
bridge_fwd_mask, as it's not explicitly enslaved to a bridge).
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 13:03:00 +0000 (15:03 +0200)]
net: mscc: ocelot: split assignment of the cpu port into a separate function
Now that the places that configure routing destinations for the CPU port
have been marked as such, allow callers to specify their own CPU port
that is different than ocelot->num_phys_ports. A user will be the Felix
DSA driver, where the CPU port is one of the physical ports (NPI mode).
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Claudiu Manoil [Sat, 9 Nov 2019 13:02:58 +0000 (15:02 +0200)]
net: mscc: ocelot: initialize list of multicast addresses in common code
This is just common path code that belongs to ocelot_init,
it has nothing to do with a specific SoC/board instance.
Signed-off-by: Claudiu Manoil <claudiu.manoil@nxp.com> Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 13:02:55 +0000 (15:02 +0200)]
net: mscc: ocelot: limit vlan ingress filtering to actual number of ports
The VSC7514 switch (Ocelot) is a 10-port device, while VSC9959 (Felix)
is 6-port. Therefore the VLAN filtering mask would be out of bounds when
calling for this new switch. Fix that.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 13:02:53 +0000 (15:02 +0200)]
net: mscc: ocelot: separate net_device related items out of ocelot_port
The ocelot and ocelot_port structures will be used by a new DSA driver,
so the ocelot_board.c file will have to allocate and work with a private
structure (ocelot_port_private), which embeds the generic struct
ocelot_port. This is because in DSA, at least one interface does not
have a net_device, and the DSA driver API does not interact with that
anyway.
The ocelot_port structure is equivalent to dsa_port, and ocelot to
dsa_switch. The members of ocelot_port which have an equivalent in
dsa_port (such as dp->vlan_filtering) have been moved to
ocelot_port_private.
We want to enforce the coding convention that "ocelot_port" refers to
the structure, and "port" refers to the integer index. One can retrieve
the structure at any time from ocelot->ports[port].
The patch is large but only contains variable renaming and mechanical
movement of fields from one structure to another.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 13:02:51 +0000 (15:02 +0200)]
net: mscc: ocelot: change prototypes of switchdev port attribute handlers
This is needed so that the Felix DSA front-end can call the Ocelot
implementations.
The implementation of the "mc_disabled" switchdev attribute has also
been simplified by using the read-modify-write macro instead of
open-coding that operation.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 13:02:49 +0000 (15:02 +0200)]
net: mscc: ocelot: break out fdb operations into abstract implementations
To be able to implement a DSA front-end over ocelot_fdb_add,
ocelot_fdb_del, ocelot_fdb_dump, these need to have a simple function
prototype that is independent of struct net_device, netlink skb, etc.
So rename the ndo ops of the ocelot driver into
ocelot_port_fdb_{add,del,dump}, and have them all call the abstract
implementations. At the same time, refactor ocelot_port_fdb_do_dump into
a function whose prototype is compatible with dsa_fdb_dump_cb_t, so that
the do_dump implementations can live together and be called by the
ocelot_fdb_dump through a function pointer.
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
These functions have a prototype that is better aligned to the DSA API.
The function also had some static initialization (TPID, drop frames with
multicast source MAC) which was not being changed from any place, so
that was just moved to ocelot_probe_port (one of the 6 callers of
ocelot_vlan_port_apply).
Signed-off-by: Vladimir Oltean <vladimir.oltean@nxp.com> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
net: dsa: mv88e6xxx: Add support for port mirroring
This patch series add support for port mirroring in the mv88e6xx switch driver.
The first patch changes the set_egress_port function to allow different egress
ports for egress and ingress traffic. The second patch adds the actual code for
port mirroring support.
tc qdisc add dev wan0 clsact
tc filter add dev wan0 ingress matchall skip_sw \
action mirred egress mirror dev lan2
tc filter add dev wan0 egress matchall skip_sw \
action mirred egress mirror dev lan3
Changes in v3
- Use enum for egress traffic direction
- Keep track of egress ports on mv88e6390
- Move booleans in struct for better structure packing
Changes in v2
- Support mirroring egress and ingress traffic to different ports
- Check for invalid configurations when multiple ports are mirrored
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Iwan R Timmer [Thu, 7 Nov 2019 21:11:14 +0000 (22:11 +0100)]
net: dsa: mv88e6xxx: Add support for port mirroring
Add support for configuring port mirroring through the cls_matchall
classifier. We do a full ingress and/or egress capture towards a
capture port. It allows setting a different capture port for ingress
and egress traffic.
It keeps track of the mirrored ports and the destination ports to
prevent changes to the capture port while other ports are being
mirrored.
Signed-off-by: Iwan R Timmer <irtimmer@gmail.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
Iwan R Timmer [Thu, 7 Nov 2019 21:11:13 +0000 (22:11 +0100)]
net: dsa: mv88e6xxx: Split monitor port configuration
Separate the configuration of the egress and ingress monitor port.
This allows the port mirror functionality to do ingress and egress
port mirroring to separate ports.
Signed-off-by: Iwan R Timmer <irtimmer@gmail.com> Reviewed-by: Andrew Lunn <andrew@lunn.ch> Signed-off-by: David S. Miller <davem@davemloft.net>
The LAN743x Ethernet controller provides two independent PTP event
channels. Each one can be used to generate a periodic output from
the PTP clock. The output can be routed to any one of the available
GPIO pins on the device.
The PTP clock API can now be used to:
- select any LAN743x GPIO pin to function as a periodic output
- select either LAN743x PTP event channel to generate the output
The LAN7430 has 4 GPIO pins that are multiplexed with its internal
PHY LED control signals. A pin assigned to the LED control function
will be assigned to the GPIO function if selected for PTP periodic
output.
Signed-off-by: John Efstathiades <john.efstathiades@pebblebay.com> Signed-off-by: David S. Miller <davem@davemloft.net>
====================
Unlock new potential in SJA1105 with PTP system timestamping
The SJA1105 being an automotive switch means it is designed to live in a
set-and-forget environment, far from the configure-at-runtime nature of
Linux. Frequently resetting the switch to change its static config means
it loses track of its PTP time, which is not good.
This patch series implements PTP system timestamping for this switch
(using the API introduced for SPI here:
https://www.mail-archive.com/netdev@vger.kernel.org/msg316725.html),
adding the following benefits to the driver:
- When under control of a user space PTP servo loop (ptp4l, phc2sys),
the loss of sync during a switch reset is much more manageable, and
the switch still remains in the s2 (locked servo) state.
- When synchronizing the switch using the software technique (based on
reading clock A and writing the value to clock B, as opposed to
relying on hardware timestamping), e.g. by using phc2sys, the sync
accuracy is vastly improved due to the fact that the actual switch PTP
time can now be more precisely correlated with something of better
precision (CLOCK_REALTIME). The issue is that SPI transfers are
inherently bad for measuring time with low jitter, but the newly
introduced API aims to alleviate that issue somewhat.
This series is also a requirement for a future patch set that adds full
time-aware scheduling offload support for the switch.
====================
Acked-by: Richard Cochran <richardcochran@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 11:32:24 +0000 (13:32 +0200)]
net: dsa: sja1105: Disallow management xmit during switch reset
The purpose here is to avoid ptp4l fail due to this condition:
timed out while polling for tx timestamp
increasing tx_timestamp_timeout may correct this issue, but it is likely caused by a driver bug
port 1: send peer delay request failed
So either reset the switch before the management frame was sent, or
after it was timestamped as well, but not in the middle.
The condition may arise either due to a true timeout (i.e. because
re-uploading the static config takes time), or due to the TX timestamp
actually getting lost due to reset. For the former we can increase
tx_timestamp_timeout in userspace, for the latter we need this patch.
Locking all traffic during switch reset does not make sense at all,
though. Forcing all CPU-originated traffic to potentially block waiting
for a sleepable context to send > 800 bytes over SPI is not a good idea.
Flows that are autonomously forwarded by the switch will get dropped
anyway during switch reset no matter what. So just let all other
CPU-originated traffic be dropped as well.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 11:32:23 +0000 (13:32 +0200)]
net: dsa: sja1105: Restore PTP time after switch reset
The PTP time of the switch is not preserved when uploading a new static
configuration. Work around this hardware oddity by reading its PTP time
before a static config upload, and restoring it afterwards.
Static config changes are expected to occur at runtime even in scenarios
directly related to PTP, i.e. the Time-Aware Scheduler of the switch is
programmed in this way.
Perhaps the larger implication of this patch is that the PTP .gettimex64
and .settime functions need to be exposed to sja1105_main.c, where the
PTP lock needs to be held during this entire process. So their core
implementation needs to move to some common functions which get exposed
in sja1105_ptp.h.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Vladimir Oltean [Sat, 9 Nov 2019 11:32:22 +0000 (13:32 +0200)]
net: dsa: sja1105: Implement the .gettimex64 system call for PTP
Through the PTP_SYS_OFFSET_EXTENDED ioctl, it is possible for userspace
applications (i.e. phc2sys) to compensate for the delays incurred while
reading the PHC's time.
The task itself of taking the software timestamp is delegated to the SPI
subsystem, through the newly introduced API in struct spi_transfer. The
goal is to cross-timestamp I/O operations on the switch's PTP clock with
values in the local system clock (CLOCK_REALTIME). For that we need to
understand a bit of the hardware internals.
The 'read PTP time' message is a 12 byte structure, first 4 bytes of
which represent the SPI header, and the last 8 bytes represent the
64-bit PTP time. The switch itself starts processing the command
immediately after receiving the last bit of the address, i.e. at the
middle of byte 3 (last byte of header). The PTP time is shadowed to a
buffer register in the switch, and retrieved atomically during the
subsequent SPI frames.
A similar thing goes on for the 'write PTP time' message, although in
that case the switch waits until the 64-bit PTP time becomes fully
available before taking any action. So the byte that needs to be
software-timestamped is byte 11 (last) of the transfer.
The patch creates a common (and local) sja1105_xfer implementation for
the SPI I/O, and offers 3 front-ends:
- sja1105_xfer_u32 and sja1105_xfer_u64: these are capable of optionally
requesting a PTP timestamp
- sja1105_xfer_buf: this is for large transfers (e.g. the static config
buffer) and other misc data, and there is no point in giving
timestamping capabilities to this.
Signed-off-by: Vladimir Oltean <olteanv@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Heiner Kallweit [Sat, 9 Nov 2019 21:01:36 +0000 (22:01 +0100)]
r8169: add helper r8168d_modify_extpage
Certain integrated PHY's from RTL8168d support extended pages. On page
0x0007 the number of the extended page is written to register 0x1e,
then the registers on the extended page can be accessed. Add a helper
for this to improve readability and simplify the code.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Heiner Kallweit [Sat, 9 Nov 2019 21:00:13 +0000 (22:00 +0100)]
r8169: add helper r8168d_phy_param
Integrated PHY's from RTL8168d support an indirect access method for
PHY parameters. On page 0x0005 parameter number is written to register
0x05, then the parameter can be accessed via register 0x06.
Add a helper for this to improve readability and simplify the code.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Heiner Kallweit [Sat, 9 Nov 2019 20:59:43 +0000 (21:59 +0100)]
r8169: add helper r8168g_phy_param
Integrated PHY's from RTL8168g support an indirect access method for
PHY parameters. On page 0x0a43 parameter number is written to register
0x13, then the parameter can be accessed via register 0x14.
Add a helper for this to improve readability and simplify the code.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Russell King [Fri, 8 Nov 2019 17:39:29 +0000 (17:39 +0000)]
net: sfp: rework upstream interface
The current upstream interface is an all-or-nothing, which is
sub-optimal for future changes, as it doesn't allow the upstream driver
to prepare for the SFP module becoming available, as it is at boot.
Switch to a find-sfp-bus, add-upstream, del-upstream, put-sfp-bus
interface structure instead, which allows the upstream driver to
prepare for a module being available as soon as add-upstream is called.
Signed-off-by: Russell King <rmk+kernel@armlinux.org.uk> Signed-off-by: David S. Miller <davem@davemloft.net>
2) Fix powerpc bpf tail call implementation, from Eric Dumazet.
3) DCCP leaks jiffies on the wire, fix also from Eric Dumazet.
4) Fix crash in ebtables when using dnat target, from Florian Westphal.
5) Fix port disable handling whne removing bcm_sf2 driver, from Florian
Fainelli.
6) Fix kTLS sk_msg trim on fallback to copy mode, from Jakub Kicinski.
7) Various KCSAN fixes all over the networking, from Eric Dumazet.
8) Memory leaks in mlx5 driver, from Alex Vesker.
9) SMC interface refcounting fix, from Ursula Braun.
10) TSO descriptor handling fixes in stmmac driver, from Jose Abreu.
11) Add a TX lock to synchonize the kTLS TX path properly with crypto
operations. From Jakub Kicinski.
12) Sock refcount during shutdown fix in vsock/virtio code, from Stefano
Garzarella.
13) Infinite loop in Intel ice driver, from Colin Ian King.
* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net: (108 commits)
ixgbe: need_wakeup flag might not be set for Tx
i40e: need_wakeup flag might not be set for Tx
igb/igc: use ktime accessors for skb->tstamp
i40e: Fix for ethtool -m issue on X722 NIC
iavf: initialize ITRN registers with correct values
ice: fix potential infinite loop because loop counter being too small
qede: fix NULL pointer deref in __qede_remove()
net: fix data-race in neigh_event_send()
vsock/virtio: fix sock refcnt holding during the shutdown
net: ethernet: octeon_mgmt: Account for second possible VLAN header
mac80211: fix station inactive_time shortly after boot
net/fq_impl: Switch to kvmalloc() for memory allocation
mac80211: fix ieee80211_txq_setup_flows() failure path
ipv4: Fix table id reference in fib_sync_down_addr
ipv6: fixes rt6_probe() and fib6_nh->last_probe init
net: hns: Fix the stray netpoll locks causing deadlock in NAPI path
net: usb: qmi_wwan: add support for DW5821e with eSIM support
CDC-NCM: handle incomplete transfer of MTU
nfc: netlink: fix double device reference drop
NFC: st21nfca: fix double free
...
Linus Torvalds [Sat, 9 Nov 2019 02:15:55 +0000 (18:15 -0800)]
Merge tag 'for-linus-2019-11-08' of git://git.kernel.dk/linux-block
Pull block fixes from Jens Axboe:
- Two NVMe device removal crash fixes, and a compat fixup for for an
ioctl that was introduced in this release (Anton, Charles, Max - via
Keith)
- Missing error path mutex unlock for drbd (Dan)
- cgroup writeback fixup on dead memcg (Tejun)
- blkcg online stats print fix (Tejun)
* tag 'for-linus-2019-11-08' of git://git.kernel.dk/linux-block:
cgroup,writeback: don't switch wbs immediately on dead wbs if the memcg is dead
block: drbd: remove a stray unlock in __drbd_send_protocol()
blkcg: make blkcg_print_stat() print stats only for online blkgs
nvme: change nvme_passthru_cmd64 to explicitly mark rsvd
nvme-multipath: fix crash in nvme_mpath_clear_ctrl_paths
nvme-rdma: fix a segmentation fault during module unload
David S. Miller [Sat, 9 Nov 2019 00:50:14 +0000 (16:50 -0800)]
Merge branch '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/jkirsher/net-queue
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>
Magnus Karlsson [Fri, 8 Nov 2019 19:58:10 +0000 (20:58 +0100)]
ixgbe: need_wakeup flag might not be set for Tx
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>
Magnus Karlsson [Fri, 8 Nov 2019 19:58:09 +0000 (20:58 +0100)]
i40e: need_wakeup flag might not be set for Tx
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>
Jacob Keller [Wed, 6 Nov 2019 17:18:23 +0000 (09:18 -0800)]
igb/igc: use ktime accessors for skb->tstamp
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>
Nicholas Nunley [Tue, 5 Nov 2019 12:22:14 +0000 (04:22 -0800)]
iavf: initialize ITRN registers with correct values
Since commit 92418fb14750 ("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>
Colin Ian King [Fri, 1 Nov 2019 14:00:17 +0000 (14:00 +0000)]
ice: fix potential infinite loop because loop counter being too small
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: b36c598c999c ("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>
Xin Long [Fri, 8 Nov 2019 05:20:36 +0000 (13:20 +0800)]
sctp: add SCTP_PEER_ADDR_THLDS_V2 sockopt
Section 7.2 of rfc7829: "Peer Address Thresholds (SCTP_PEER_ADDR_THLDS)
Socket Option" extends 'struct sctp_paddrthlds' with 'spt_pathcpthld'
added to allow a user to change ps_retrans per sock/asoc/transport, as
other 2 paddrthlds: pf_retrans, pathmaxrxt.
Note: to not break the user's program, here to support pf_retrans dump
and setting by adding a new sockopt SCTP_PEER_ADDR_THLDS_V2, and a new
structure sctp_paddrthlds_v2 instead of extending sctp_paddrthlds.
Also, when setting ps_retrans, the value is not allowed to be greater
than pf_retrans.
v1->v2:
- use SCTP_PEER_ADDR_THLDS_V2 to set/get pf_retrans instead,
as Marcelo and David Laight suggested.
Signed-off-by: Xin Long <lucien.xin@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Xin Long [Fri, 8 Nov 2019 05:20:35 +0000 (13:20 +0800)]
sctp: add support for Primary Path Switchover
This is a new feature defined in section 5 of rfc7829: "Primary Path
Switchover". By introducing a new tunable parameter:
Primary.Switchover.Max.Retrans (PSMR)
The primary path will be changed to another active path when the path
error counter on the old primary path exceeds PSMR, so that "the SCTP
sender is allowed to continue data transmission on a new working path
even when the old primary destination address becomes active again".
This patch is to add this tunable parameter, 'ps_retrans' per netns,
sock, asoc and transport. It also allows a user to change ps_retrans
per netns by sysctl, and ps_retrans per sock/asoc/transport will be
initialized with it.
The check will be done in sctp_do_8_2_transport_strike() when this
feature is enabled.
Note this feature is disabled by initializing 'ps_retrans' per netns
as 0xffff by default, and its value can't be less than 'pf_retrans'
when changing by sysctl.
v3->v4:
- add define SCTP_PS_RETRANS_MAX 0xffff, and use it on extra2 of
sysctl 'ps_retrans'.
- add a new entry for ps_retrans on ip-sysctl.txt.
Signed-off-by: Xin Long <lucien.xin@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
This is a sockopt defined in section 7.3 of rfc7829: "Exposing
the Potentially Failed Path State", by which users can change
pf_expose per sock and asoc.
The new sockopt SCTP_EXPOSE_POTENTIALLY_FAILED_STATE is also
known as SCTP_EXPOSE_PF_STATE for short.
SCTP Quick failover draft section 5.1, point 5 has been removed
from rfc7829. Instead, "the sender SHOULD (i) notify the Upper
Layer Protocol (ULP) about this state transition", as said in
section 3.2, point 8.
So this patch is to add SCTP_ADDR_POTENTIALLY_FAILED, defined
in section 7.1, "which is reported if the affected address
becomes PF". Also remove transport cwnd's update when moving
from PF back to ACTIVE , which is no longer in rfc7829 either.
Note that ulp_notify will be set to false if asoc->expose is
not 'enabled', according to last patch.
v2->v3:
- define SCTP_ADDR_PF SCTP_ADDR_POTENTIALLY_FAILED.
v3->v4:
- initialize spc_state with SCTP_ADDR_AVAILABLE, as Marcelo suggested.
- check asoc->pf_expose in sctp_assoc_control_transport(), as Marcelo
suggested.
Signed-off-by: Xin Long <lucien.xin@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Xin Long [Fri, 8 Nov 2019 05:20:32 +0000 (13:20 +0800)]
sctp: add pf_expose per netns and sock and asoc
As said in rfc7829, section 3, point 12:
The SCTP stack SHOULD expose the PF state of its destination
addresses to the ULP as well as provide the means to notify the
ULP of state transitions of its destination addresses from
active to PF, and vice versa. However, it is recommended that
an SCTP stack implementing SCTP-PF also allows for the ULP to be
kept ignorant of the PF state of its destinations and the
associated state transitions, thus allowing for retention of the
simpler state transition model of [RFC4960] in the ULP.
Not only does it allow to expose the PF state to ULP, but also
allow to ignore sctp-pf to ULP.
So this patch is to add pf_expose per netns, sock and asoc. And in
sctp_assoc_control_transport(), ulp_notify will be set to false if
asoc->expose is not 'enabled' in next patch.
It also allows a user to change pf_expose per netns by sysctl, and
pf_expose per sock and asoc will be initialized with it.
Note that pf_expose also works for SCTP_GET_PEER_ADDR_INFO sockopt,
to not allow a user to query the state of a sctp-pf peer address
when pf_expose is 'disabled', as said in section 7.3.
v1->v2:
- Fix a build warning noticed by Nathan Chancellor.
v2->v3:
- set pf_expose to UNUSED by default to keep compatible with old
applications.
v3->v4:
- add a new entry for pf_expose on ip-sysctl.txt, as Marcelo suggested.
- change this patch to 1/5, and move sctp_assoc_control_transport
change into 2/5, as Marcelo suggested.
- use SCTP_PF_EXPOSE_UNSET instead of SCTP_PF_EXPOSE_UNUSED, and
set SCTP_PF_EXPOSE_UNSET to 0 in enum, as Marcelo suggested.
Signed-off-by: Xin Long <lucien.xin@gmail.com> Acked-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Jiri Pirko [Fri, 8 Nov 2019 20:42:43 +0000 (21:42 +0100)]
devlink: disallow reload operation during device cleanup
There is a race between driver code that does setup/cleanup of device
and devlink reload operation that in some drivers works with the same
code. Use after free could we easily obtained by running:
while true; do
echo 10 > /sys/bus/netdevsim/new_device
devlink dev reload netdevsim/netdevsim10 &
echo 10 > /sys/bus/netdevsim/del_device
done
Fix this by enabling reload only after setup of device is complete and
disabling it at the beginning of the cleanup process.
Reported-by: Ido Schimmel <idosch@mellanox.com> Fixes: 2d8dc5bbf4e7 ("devlink: Add support for reload") Signed-off-by: Jiri Pirko <jiri@mellanox.com> Acked-by: Jakub Kicinski <jakub.kicinski@netronome.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Manish Chopra [Fri, 8 Nov 2019 10:42:30 +0000 (02:42 -0800)]
qede: fix NULL pointer deref in __qede_remove()
While rebooting the system with SR-IOV vfs enabled leads
to below crash due to recurrence of __qede_remove() on the VF
devices (first from .shutdown() flow of the VF itself and
another from PF's .shutdown() flow executing pci_disable_sriov())
This patch adds a safeguard in __qede_remove() flow to fix this,
so that driver doesn't attempt to remove "already removed" devices.
projects / mirror_ubuntu-kernels.git / log