The increase in number of ingress frame queues means we now risk
depleting the buffer pool before the FQ taildrop kicks in.
Congestion group taildrop allows us to control the number of frames that
can accumulate on a group of Rx frame queues belonging to the same
traffic class. This setting coexists with the frame queue based
taildrop: whichever limit gets hit first triggers the frame drop.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add convenient helper functions that determines whether Rx/Tx pause
frames are enabled based on link state flags received from firmware.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Configure static ingress classification based on VLAN PCP field.
If the DPNI doesn't have enough traffic classes to accommodate all
priority levels, the lowest ones end up on TC 0 (default on miss).
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The firmware reserves for each DPNI a number of RX frame queues
equal to the number of configured flows x number of configured
traffic classes.
Current driver configuration directs all incoming traffic to
FQs corresponding to TC0, leaving all other priority levels unused.
Start adding support for multiple ingress traffic classes, by
configuring the FQs associated with all priority levels, not just
TC0. All settings that are per-TC, such as those related to
hashing and flow steering, are also updated.
Signed-off-by: Ioana Radulescu <ruxandra.radulescu@nxp.com>
Signed-off-by: Ioana Ciornei <ioana.ciornei@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Microchip lan7431 is frequently connected to a phy. However, it
can also be directly connected to a MII remote peer without
any phy in between. For supporting such a phyless hardware setup
in Linux we utilized phylib, which supports a fixed-link
configuration via the device tree. And we added support for
defining the connection type R/GMII in the device tree.
New behavior:
-------------
. The automatic speed and duplex detection of the lan743x silicon
between mac and phy is disabled. Instead phylib is used like in
other typical Linux drivers. The usage of phylib allows to
specify fixed-link parameters in the device tree.
. The device tree entry phy-connection-type is supported now with
the modes RGMII or (G)MII (default).
Development state:
------------------
. Tested with fixed-phy configurations. Not yet tested in normal
configurations with phy. Microchip kindly offered testing
as soon as the Corona measures allow this.
. All review findings of Andrew Lunn are included
Example:
--------
&pcie {
status = "okay";
host@0 {
reg = <0 0 0 0 0>;
#address-cells = <3>;
#size-cells = <2>;
ethernet@0 {
compatible = "weyland-yutani,noscom1", "microchip,lan743x";
status = "okay";
reg = <0 0 0 0 0>;
phy-connection-type = "rgmii";
fixed-link {
speed = <100>;
full-duplex;
};
};
};
};
Signed-off-by: Roelof Berg <rberg@berg-solutions.de>
Signed-off-by: Andrew Lunn <andrew@lunn.ch>
Signed-off-by: David S. Miller <davem@davemloft.net>
In a similar fashion to other control traps, register ACL control traps
with devlink.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In a similar fashion to layer 2 control traps, register layer 3 control
traps with devlink.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In a similar fashion to other traps, register layer 2 control traps with
devlink.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We currently have an Rx listener function for exception traps that marks
received skbs with 'offload_fwd_mark' and injects them to the kernel's
Rx path. The marking is done because all these exceptions occur during
L3 forwarding, after the packets were potentially flooded at L2.
A subsequent patch will add support for control traps. Packets received
via some of these control traps need different handling:
1. Packets might not need to be marked with 'offload_fwd_mark'. For
example, if packet was trapped before L2 forwarding
2. Packets might not need to be injected to the kernel's Rx path. For
example, sampled packets are reported to user space via the psample
module
Factor out a common Rx listener function that only reports trapped
packets to devlink. Call it from mlxsw_sp_rx_no_mark_listener() and
mlxsw_sp_rx_mark_listener() that will inject the packets to the kernel's
Rx path, without and with the marking, respectively.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The layer 3 exceptions are still subject to the same trap policer, so
nothing changes, but user space can choose to assign a different one.
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reviewed-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When enable clk_ref failed, clk_ptp should be disabled rather than
clk_ref itself.
Signed-off-by: Liu Xiang <liuxiang_1999@126.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The removal of mips_swiotlb_ops exposed a problem in octeon_mgmt Ethernet
driver. mips_swiotlb_ops had an mb() after most of the operations and the
removal of the ops had broken the receive functionality of the driver.
My code inspection has shown no other places except
octeon_mgmt_rx_fill_ring() where an explicit barrier would be obviously
missing. The latter function however has to make sure that "ringing the
bell" doesn't happen before RX ring entry is really written.
The patch has been successfully tested on Octeon II.
Fixes: a999933db9 ("MIPS: remove mips_swiotlb_ops")
Cc: stable@vger.kernel.org
Signed-off-by: Alexander Sverdlin <alexander.sverdlin@nokia.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Register ndo callback via flow_indr_dev_register() and
flow_indr_dev_unregister().
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Register ndo callback via flow_indr_dev_register() and
flow_indr_dev_unregister().
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Register ndo callback via flow_indr_dev_register() and
flow_indr_dev_unregister().
No need for mlx5e_rep_indr_clean_block_privs() since flow_block_cb_free()
already releases the internal mapping via ->release callback, which in
this case is mlx5e_rep_indr_tc_block_unbind().
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Shrink the code visually by replacing regmap_update_bits() with
appropriate regmap bit operations where applicable.
Signed-off-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The caller doesn't care about the error codes, they only check for zero
vs non-zero. Still, it's better to preserve the negative error codes
from alloc_uld_rxqs() instead of changing it to 1. We can also return
directly if there is a failure.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
xdp_umem.c had overlapping changes between the 64-bit math fix
for the calculation of npgs and the removal of the zerocopy
memory type which got rid of the chunk_size_nohdr member.
The mlx5 Kconfig conflict is a case where we just take the
net-next copy of the Kconfig entry dependency as it takes on
the ESWITCH dependency by one level of indirection which is
what the 'net' conflicting change is trying to ensure.
Signed-off-by: David S. Miller <davem@davemloft.net>
Until recently, the Micrel KSZ9031 PHY driver ignored any PHY mode
("RGMII-*ID") settings, but used the hardware defaults, augmented by
explicit configuration of individual skew values using the "*-skew-ps"
DT properties. The lack of PHY mode support was compensated by the
EtherAVB MAC driver, which configures TX and/or RX internal delay
itself, based on the PHY mode.
However, now the KSZ9031 driver has gained PHY mode support, delays may
be configured twice, causing regressions. E.g. on the Renesas
Salvator-X board with R-Car M3-W ES1.0, TX performance dropped from ca.
400 Mbps to 0.1-0.3 Mbps, as measured by nuttcp.
As internal delay configuration supported by the KSZ9031 PHY is too
limited for some use cases, the ability to configure MAC internal delay
is deemed useful and necessary. Hence a proper fix would involve
splitting internal delay configuration in two parts, one for the PHY,
and one for the MAC. However, this would require adding new DT
properties, thus breaking DTB backwards-compatibility.
Hence fix the regression in a backwards-compatibility way, by letting
the EtherAVB driver mask the PHY mode when it has inserted a delay, to
avoid the PHY driver adding a second delay. This also fixes messages
like:
Micrel KSZ9031 Gigabit PHY e6800000.ethernet-ffffffff:00: *-skew-ps values should be used only with phy-mode = "rgmii"
as the PHY no longer sees the original RGMII-*ID mode.
Solving the issue by splitting configuration in two parts can be handled
in future patches, and would require retaining a backwards-compatibility
mode anyway.
Fixes: bcf3440c6d ("net: phy: micrel: add phy-mode support for the KSZ9031 PHY")
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Andrew Lunn <andrew@lunn.ch>
Reviewed-by: Florian Fainelli <f.fainelli@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Jeff Kirsher says:
====================
100GbE Intel Wired LAN Driver Updates 2020-05-28
This series contains updates to the ice driver only.
Anirudh (Ani) adds a poll for reset completion before proceeding with
driver initialization when the DDP package fails to load and the firmware
issues a core reset.
Jake cleans up unnecessary code, since ice_set_dflt_vsi_ctx() performs a
memset to clear the info from the context structures. Fixed a potential
double free during probe unrolling after a failure. Also fixed a
potential NULL pointer dereference upon register_netdev() failure.
Tony makes two functions static which are not called outside of their
file.
Brett refactors the ice_ena_vf_mappings(), which was doing the VF's MSIx
and queue mapping in one function which was hard to digest. So create a
new function to handle the enabling MSIx mappings and another function
to handle the enabling of queue mappings. Simplify the code flow in
ice_sriov_configure(). Created a helper function for clearing
VPGEN_VFRTRIG register, as this needs to be done on reset to notify the
VF that we are done resetting it. Fixed the initialization/creation and
reset flows, which was unnecessarily complicated, so separate the two
flows into their own functions. Renamed VF initialization functions to
make it more clear what they do and why. Added functionality to set the
VF trust mode bit on reset. Added helper functions to rebuild the VLAN
and MAC configurations when resetting a VF. Refactored how the VF reset
is handled to prevent VF reset timeouts.
Paul cleaned up code not needed during a CORER/GLOBR reset.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
Simplify handling the power management callbacks.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Functionality for quiescing the chip is spread across different
functions currently. Move it to rtl8169_down().
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move calls that are needed before and after calling rtl8169_hw_reset()
into this function. This requires to move the function in the code.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In places where the indexes have to be reset, we call
rtl8169_init_ring_indexes() anyway after rtl8169_tx_clear().
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We go to runtime-suspend few secs after cable removal. As cable is
removed "physical link up" is the only meaningful WoL source.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Change driver private data type to struct rtl8169_private * to avoid
some overhead.
Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Accumulated cleanup patches and sparse warning fixes for mlx5 driver.
1) sync with mlx5-next branch
2) Eli Cohen declares mpls_entry_encode() helper in mpls.h as suggested
by Jakub Kicinski and David Ahern, and use it in mlx5
3) Jesper Fixes xdp data_meta setup in mlx5
4) Many sparse and build warnings cleanup
-----BEGIN PGP SIGNATURE-----
iQEzBAABCAAdFiEEGhZs6bAKwk/OTgTpSD+KveBX+j4FAl7R3wcACgkQSD+KveBX
+j6/ZQf/QD39naPeImfLjemkRK9L+TKbS4nU6wpUwf1jC33Wdm4HhkhsWEnR6C4l
OwU/Pae3I9EtKP4gRE0W1o8h7zC9h4hY7+IKZOdyQ32iUY55PX/H25oqAiCj1NCM
xzWpXOTwK/vkqmkCedAd+YpNdYlbOhfycr+KVPSsvFdaPqjzfNO1PJcLsUbAbzrX
A+8pYdhUYTtx1N3YHJL5abLN6WzMAKxgwlm9GG8YCXACTJT6CBWWHGebVsC5TDUk
Lj5hJj38mI8/3dcu6vWP0kLGVfRZo0HS/gpPGxbKQFpP+1uBYaRENAQONxkY++6S
GDPix7ccvN+yNMlON893PC/Cogw3Yg==
=WaCJ
-----END PGP SIGNATURE-----
Merge tag 'mlx5-cleanup-2020-05-29' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-cleanup-2020-05-29
Accumulated cleanup patches and sparse warning fixes for mlx5 driver.
1) sync with mlx5-next branch
2) Eli Cohen declares mpls_entry_encode() helper in mpls.h as suggested
by Jakub Kicinski and David Ahern, and use it in mlx5
3) Jesper Fixes xdp data_meta setup in mlx5
4) Many sparse and build warnings cleanup
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
NETIF_F_HW_VLAN_CTAG_FILTER is not set in netdev->hw_feature for
the HNS3 driver, so the handler of NETIF_F_HW_VLAN_CTAG_FILTER
in hns3_nic_set_features() won't be called, remove it.
Reported-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Remove a redundant blank line in hclgevf_cmd_set_promisc_mode(),
and fix a reverse xmas tree coding style issue in
hclgevf_set_rss_tc_mode().
Reported-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
struct hclgevf_dev stands for VF device, its field num_tqps
indicates the number of VF's task queue pairs, so the comment
is incorrect, replace 'PF' with 'VF'.
Reported-by: Jian Shen <shenjian15@huawei.com>
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Macro hclgevf_ring_to_dma_dir and hclgevf_is_csq defined in
hclgevf_cmd.c, but not used, so remove them.
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Macro hclge_is_csq defined in hcgle_cmd.c has not been used,
so remove it.
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Use %d to print int variable 'ret' in hclge_mac_mdio_config().
Signed-off-by: Huazhong Tan <tanhuazhong@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Fix sparse warning:
drivers/net/ethernet/mellanox/mlx5/core/en_dcbnl.c:988:29:
error: symbol 'mlx5e_dcbnl_ops' was not declared. Should it be static?
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Fix some trivial warnings of the type:
warning: incorrect type in initializer (different base types)
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
tls handle and rcd_sn are actually big endian and not in host format.
Fix that.
Fix the following sparse warnings:
drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c:177:21:
warning: cast to restricted __be64
drivers/net/ethernet/mellanox/mlx5/core/en_accel/tls.c:178:52:
warning: incorrect type in argument 2 (different base types)
expected unsigned int [usertype] handle
got restricted __be32 [usertype] handle
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Fix sparse warning:
drivers/net/ethernet/mellanox/mlx5/core/cmd.c:1949:15:
warning: memset with byte count of 271720
mlx5_cmd_stats array is too big to be held inline in mlx5_cmd.
Allocate it separately.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
dr_ste_crc32_calc() calculates crc32 and should return it in HW format.
It is being used to calculate a u32 index, hence we force the return value
of u32 to avoid the sparse warning:
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c:115:16:
warning: incorrect type in return expression (different base types)
expected unsigned int
got restricted __be32 [usertype]
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
raw_ip actual type is __be32 and not u32.
Fix that and get rid of the warning.
drivers/net/ethernet/mellanox/mlx5/core/steering/dr_ste.c:906:31:
warning: cast to restricted __be32
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
HW spec objects should receive a void ptr to work on, the MLX5_SET/GET
macro will know how to handle it.
No need to provide explicit or wrong pointer type in this case.
warning: incorrect type in argument 1 (different base types)
expected unsigned long long const [usertype] *sw_action
got restricted __be64 [usertype] *[assigned] sw_action
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Reviewed-by: Mark Bloch <markb@mellanox.com>
Make use of generic API mpls_entry_encode() to build mpls label and get
rid of local function.
Signed-off-by: Eli Cohen <eli@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Moving the mlx5_ifc_query_qp_out_bits structure on the stack was a bit
excessive and now causes the compiler to complain on 32-bit architectures:
drivers/net/ethernet/mellanox/mlx5/core/debugfs.c: In function 'qp_read_field':
drivers/net/ethernet/mellanox/mlx5/core/debugfs.c:274:1: error: the frame size of 1104 bytes is larger than 1024 bytes [-Werror=frame-larger-than=]
Revert the previous patch partially to use dynamically allocation as
the code did before. Unfortunately there is no good error handling
in case the allocation fails.
Fixes: 57a6c5e992 ("net/mlx5: Replace hand written QP context struct with automatic getters")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Clang warns:
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c:3712:6: warning:
variable 'err' is used uninitialized whenever 'if' condition is false
[-Wsometimes-uninitialized]
if (IS_ERR(d->pkt_reformat)) {
^~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c:3718:6: note:
uninitialized use occurs here
if (err)
^~~
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c:3712:2: note: remove the
'if' if its condition is always true
if (IS_ERR(d->pkt_reformat)) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
drivers/net/ethernet/mellanox/mlx5/core/en_tc.c:3670:9: note: initialize
the variable 'err' to silence this warning
int err;
^
= 0
1 warning generated.
It is not wrong, err is only ever initialized in if statements but this
one is not in one. Initialize err to 0 to fix this.
Fixes: 14e6b038af ("net/mlx5e: Add support for hw decapsulation of MPLS over UDP")
Link: https://github.com/ClangBuiltLinux/linux/issues/1037
Signed-off-by: Nathan Chancellor <natechancellor@gmail.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
The helper function xdp_set_data_meta_invalid() must be called after
setting xdp->data as it depends on it.
The bug was introduced in the cited patch below, and cause the kernel
to crash when using BPF helper bpf_xdp_adjust_head() on mlx5 driver.
Fixes: 39d6443c8d ("mlx5, xsk: Migrate to new MEM_TYPE_XSK_BUFF_POOL")
Reported-by: David Ahern <dsahern@kernel.org>
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Tested-by: David Ahern <dsahern@gmail.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
VNIC protocol version is reported in big-endian format, but it
is not byteswapped before logging. Fix that, and remove version
comparison as only one protocol version exists at this time.
Signed-off-by: Thomas Falcon <tlfalcon@linux.ibm.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A previous refactoring missed some locations the flags were renamed
but not moved from the previous flower_ext_feats to the new flower_en_feats
variable. This lead to the FLOW_MERGE and LAG features not being enabled.
Fixes: e09303d3c4 ("nfp: flower: renaming of feature bits")
Signed-off-by: Louis Peens <louis.peens@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
NXP imx8 family like imx8mp/imx8dxl chips support Synopsys MAC 5.10a IP.
This patch adds settings for NXP imx8 glue layer:
- clocks
- dwmac address width
- phy interface mode selection
- adjust rgmii txclk rate
v2:
- adjust code sequences in order to have reverse christmas
tree local variable ordering.
Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add "snps,dwmac-5.10a" compatible string for 5.10a version that can
avoid to define some plat data in glue layer.
Signed-off-by: Fugang Duan <fugang.duan@nxp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
