| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue
Tony Nguyen says:
====================
Intel Wired LAN Driver Updates 2023-08-16 (iavf, i40e)
This series contains updates to iavf and i40e drivers.
Piotr adds checks for unsupported Flow Director rules on iavf.
Andrii replaces incorrect 'write' messaging on read operations for i40e.
* '40GbE' of git://git.kernel.org/pub/scm/linux/kernel/git/tnguy/net-queue:
i40e: fix misleading debug logs
iavf: fix FDIR rule fields masks validation
====================
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Mitigate race between registering on wait list and receiving
AQ Response from FW.
ice_aq_prep_for_event() should be called before sending AQ command,
ice_aq_wait_for_event() should be called after sending AQ command,
to wait for AQ Response.
Please note, that this was found by reading the code,
an actual race has not yet materialized.
Reviewed-by: Jacob Keller <[email protected]>
Signed-off-by: Przemek Kitszel <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Tested-by: Pucha Himasekhar Reddy <[email protected]> (A Contingent worker at Intel)
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Expose struct ice_aq_task to callers,
what takes burden of memory ownership out from AQ-wait family of functions,
and reduces need for heap-based allocations.
Embed struct ice_rq_event_info event into struct ice_aq_task
(instead of it being a ptr) to remove some more code from the callers.
Subsequent commit will improve more based on this one.
Reviewed-by: Jacob Keller <[email protected]>
Signed-off-by: Przemek Kitszel <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Tested-by: Pucha Himasekhar Reddy <[email protected]> (A Contingent worker at Intel)
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Allow task's event buffer to be filled also in the case that it's size
is exactly the size of the message.
Fixes: d69ea414c9b4 ("ice: implement device flash update via devlink")
Reviewed-by: Jacob Keller <[email protected]>
Signed-off-by: Przemek Kitszel <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Tested-by: Pucha Himasekhar Reddy <[email protected]> (A Contingent worker at Intel)
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Drop @num_entries and @cd params, latter of which was always NULL.
Number of entities to alloc is passed in internal buffer, the outer layer
(that @num_entries was assigned to) meaning is closer to "the number of
requests", which was =1 in all cases.
ice_free_hw_res() was always called with 1 as its @num arg.
Signed-off-by: Przemek Kitszel <[email protected]>
Tested-by: Pucha Himasekhar Reddy <[email protected]> (A Contingent worker at Intel)
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Convert list_for_each() to list_for_each_entry() where applicable.
No functional changed.
Signed-off-by: Yang Yingliang <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Tested-by: Sujai Buvaneswaran <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Remove a call to disable VLAN stripping on switchdev control plane VSI, as
it is disabled by default.
Signed-off-by: Marcin Szycik <[email protected]>
Tested-by: Sujai Buvaneswaran <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Commit 885fe6932a11 ("ice: Add support for SMA control multiplexer")
accidentally placed all of the E810T SMA control functions in the middle of
the device agnostic functions section of ice_ptp_hw.c
This works fine, but makes it harder for readers to follow. The
ice_ptp_hw.c file is laid out such that each hardware family has the
specific functions in one block, with the access functions placed at the
end of the file.
Move the E810T functions so that they are in a block just after the E810
functions. Also move the ice_get_phy_tx_tstamp_ready_e810 which got added
at the end of the E810T block.
This keeps the functions laid out in a logical order and avoids intermixing
the generic access functions with the device specific implementations.
Signed-off-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Tested-by: Arpana Arland <[email protected]> (A Contingent worker at Intel)
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
As this method became static, and is already called
with check for vsi being non-null, an unnecessary check along
with superfluous parentheses is removed.
Signed-off-by: Jan Sokolowski <[email protected]>
Reviewed-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
As following methods are not used outside ice_ptp_hw,
they can be made static:
ice_read_phy_reg_e822
ice_write_phy_reg_e822
ice_ptp_prep_port_adj_e822
Signed-off-by: Jan Sokolowski <[email protected]>
Reviewed-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
As ice_sched_set_node_bw_lmt_per_tc is not used
outside of ice_sched, it can be made static.
Signed-off-by: Jan Sokolowski <[email protected]>
Reviewed-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
The if/else check for bit setting can be replaced by using the
assign_bit() helper so do so.
Suggested-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5 fixes 2023-08-16
This series provides bug fixes to mlx5 driver.
* tag 'mlx5-fixes-2023-08-16' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
net/mlx5: Fix mlx5_cmd_update_root_ft() error flow
net/mlx5e: XDP, Fix fifo overrun on XDP_REDIRECT
====================
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
ADQ and switchdev are not supported simultaneously. Enabling both at the
same time can result in nullptr dereference.
To prevent this, check if ADQ is active when changing devlink mode to
switchdev mode, and check if switchdev is active when enabling ADQ.
Fixes: fbc7b27af0f9 ("ice: enable ndo_setup_tc support for mqprio_qdisc")
Signed-off-by: Marcin Szycik <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Tested-by: Sujai Buvaneswaran <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
While performing certain power-off sequences, PCI drivers are
called to suspend and resume their underlying devices through
PCI PM (power management) interface. However this NIC hardware
does not support PCI PM suspend/resume operations so system wide
suspend/resume leads to bad MFW (management firmware) state which
causes various follow-up errors in driver when communicating with
the device/firmware afterwards.
To fix this driver implements PCI PM suspend handler to indicate
unsupported operation to the PCI subsystem explicitly, thus avoiding
system to go into suspended/standby mode.
Without this fix device/firmware does not recover unless system
is power cycled.
Fixes: 2950219d87b0 ("qede: Add basic network device support")
Signed-off-by: Manish Chopra <[email protected]>
Signed-off-by: Alok Prasad <[email protected]>
Reviewed-by: John Meneghini <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Existing comment in the source explains why we don't want efx_init_tc()
failure to be fatal. Cited commit erroneously consolidated failure
paths causing the probe to be failed in this case.
Fixes: 7e056e2360d9 ("sfc: obtain device mac address based on firmware handle for ef100")
Reviewed-by: Martin Habets <[email protected]>
Signed-off-by: Edward Cree <[email protected]>
Link: https://lore.kernel.org/r/aa7f589dd6028bd1ad49f0a85f37ab33c09b2b45.1692114888.git.ecree.xilinx@gmail.com
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
In efx_init_tc(), move the setting of efx->tc->up after the
flow_indr_dev_register() call, so that if it fails, efx_fini_tc()
won't call flow_indr_dev_unregister().
Fixes: 5b2e12d51bd8 ("sfc: bind indirect blocks for TC offload on EF100")
Suggested-by: Pieter Jansen van Vuuren <[email protected]>
Reviewed-by: Martin Habets <[email protected]>
Signed-off-by: Edward Cree <[email protected]>
Link: https://lore.kernel.org/r/a81284d7013aba74005277bd81104e4cfbea3f6f.1692114888.git.ecree.xilinx@gmail.com
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
As following methods are not used outside ice_vf_lib,
they can be made static:
ice_vf_rebuild_host_vlan_cfg
ice_vf_rebuild_host_tx_rate_cfg
ice_vf_set_host_trust_cfg
ice_vf_rebuild_host_mac_cfg
ice_vf_rebuild_aggregator_node_cfg
ice_vf_rebuild_host_cfg
ice_set_vf_state_qs_dis
ice_vf_set_initialized
In order to achieve that, the order in which these
were defined was reorganized.
Signed-off-by: Jan Sokolowski <[email protected]>
Reviewed-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
As following methods are not used outside of ice_lib,
they can be made static:
ice_vsi_is_vlan_pruning_ena
ice_vsi_cfg_frame_size
Signed-off-by: Jan Sokolowski <[email protected]>
Reviewed-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
As following methods are not used outside of ice_ddp,
they can be made static:
ice_verify_pgk
ice_pkg_val_buf
ice_aq_download_pkg
ice_aq_update_pkg
ice_find_seg_in_pkg
Signed-off-by: Jan Sokolowski <[email protected]>
Reviewed-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Following methods were found to no longer be in use:
ice_is_pca9575_present
ice_mac_fltr_exist
ice_napi_del
Remove them.
Signed-off-by: Jan Sokolowski <[email protected]>
Reviewed-by: Jacob Keller <[email protected]>
Reviewed-by: Przemek Kitszel <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
If the switch is reset during active EEPROM transactions, as in
just after an SoC reset after power up, the I2C bus transaction
may be cut short leaving the EEPROM internal I2C state machine
in the wrong state. When the switch is reset again, the bad
state machine state may result in data being read from the wrong
memory location causing the switch to enter unexpected mode
rendering it inoperational.
Fixes: a3dcb3e7e70c ("net: dsa: mv88e6xxx: Wait for EEPROM done after HW reset")
Signed-off-by: Alfred Lee <[email protected]>
Reviewed-by: Andrew Lunn <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
This reverts commit 90bc21aaef4adaefceda2d385756138fc247c0c2.
Patch was merged too hastily, Vladimir requested changes in:
https://lore.kernel.org/all/20230816121305.5dio5tk3chge2ndh@skbuf/
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
The cited patch change mlx5_cmd_update_root_ft() to work with multiple
peer devices. However, it didn't align the error flow as well.
Hence, Fix the error code to work with multiple peer devices.
Fixes: 222dd185833e ("{net/RDMA}/mlx5: introduce lag_for_each_peer")
Signed-off-by: Shay Drory <[email protected]>
Reviewed-by: Roi Dayan <[email protected]>
Signed-off-by: Saeed Mahameed <[email protected]>
|
|
Before this fix, running high rate traffic through XDP_REDIRECT
with multibuf could overrun the fifo used to release the
xdp frames after tx completion. This resulted in corrupted data
being consumed on the free side.
The culplirt was a miscalculation of the fifo size: the maximum ratio
between fifo entries / data segments was incorrect. This ratio serves to
calculate the max fifo size for a full sq where each packet uses the
worst case number of entries in the fifo.
This patch fixes the formula and names the constant. It also makes sure
that future values will use a power of 2 number of entries for the fifo
mask to work.
Signed-off-by: Dragos Tatulea <[email protected]>
Fixes: 3f734b8c594b ("net/mlx5e: XDP, Use multiple single-entry objects in xdpi_fifo")
Reviewed-by: Tariq Toukan <[email protected]>
Signed-off-by: Saeed Mahameed <[email protected]>
|
|
Pass a pointer to the PF's private data structure rather than
bouncing in and out of the PF's PCI function address.
Signed-off-by: Shannon Nelson <[email protected]>
Signed-off-by: Brett Creeley <[email protected]>
Reviewed-by: Kevin Tian <[email protected]>
Reviewed-by: Shameer Kolothum <[email protected]>
Reviewed-by: Jason Gunthorpe <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Alex Williamson <[email protected]>
|
|
Finally, fix 3 structures which are allocated technically correctly,
i.e. the calculated size equals to the one that struct_size() would
return, except for sizeof(). For &virtchnl_vlan_filter_list_v2, use
the same approach when there are no enough space as taken previously
for &virtchnl_vlan_filter_list, i.e. let the maximum size be calculated
automatically instead of trying to guestimate it using maths.
Signed-off-by: Alexander Lobakin <[email protected]>
Reviewed-by: Kees Cook <[email protected]>
Tested-by: Rafal Romanowski <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
There are five virtchnl structures, which are allocated and checked in
the code as `nents + 1`, meaning that they always have memory for one
excessive element regardless of their actual number. This comes from
that their sizeof() includes space for 1 element and then they get
allocated via struct_size() or its open-coded equivalents, passing
the actual number of elements.
Expand virtchnl_struct_size() to handle such structures and replace
those 1-elem arrays with proper flex ones. Also fix several places
which open-code %IAVF_VIRTCHNL_VF_RESOURCE_SIZE. Finally, let the
virtchnl_ether_addr_list size be computed automatically when there's
no enough space for the whole list, otherwise we have to open-code
reverse struct_size() logics.
Signed-off-by: Alexander Lobakin <[email protected]>
Reviewed-by: Kees Cook <[email protected]>
Tested-by: Rafal Romanowski <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
The two most problematic virtchnl structures are virtchnl_rss_key and
virtchnl_rss_lut. Their "flex" arrays have the type of u8, thus, when
allocating / checking, the actual size is calculated as `sizeof +
nents - 1 byte`. But their sizeof() is not 1 byte larger than the size
of such structure with proper flex array, it's two bytes larger due to
the padding. That said, their size is always 1 byte larger unless
there are no tail elements -- then it's +2 bytes.
Add virtchnl_struct_size() macro which will handle this case (and later
other cases as well). Make its calling conv the same as we call
struct_size() to allow it to be drop-in, even though it's unlikely to
become possible to switch to generic API. The macro will calculate a
proper size of a structure with a flex array at the end, so that it
becomes transparent for the compilers, but add the difference from the
old values, so that the real size of sorta-ABI-messages doesn't change.
Use it on the allocation side in IAVF and the receiving side (defined
as static inline in virtchnl.h) for the mentioned two structures.
Signed-off-by: Alexander Lobakin <[email protected]>
Reviewed-by: Kees Cook <[email protected]>
Tested-by: Rafal Romanowski <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Change "write" into the actual "read" word.
Change parameters description.
Fixes: 7073f46e443e ("i40e: Add AQ commands for NVM Update for X722")
Signed-off-by: Aleksandr Loktionov <[email protected]>
Signed-off-by: Andrii Staikov <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
Return an error if a field's mask is neither full nor empty. When a mask
is only partial the field is not being used for rule programming but it
gives a wrong impression it is used. Fix by returning an error on any
partial mask to make it clear they are not supported.
The ip_ver assignment is moved earlier in code to allow using it in
iavf_validate_fdir_fltr_masks.
Fixes: 527691bf0682 ("iavf: Support IPv4 Flow Director filters")
Fixes: e90cbc257a6f ("iavf: Support IPv6 Flow Director filters")
Signed-off-by: Piotr Gardocki <[email protected]>
Tested-by: Rafal Romanowski <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
|
|
SIU_INT_IRQ1 is not used anywhere and __IO_BASE is defined in
asm/io.h
Remove m82xx_pci.h
Then the only thing remaining in mpc8260.h is MPC82XX_BCR_PLDP
Move MPC82XX_BCR_PLDP into asm/cpm2.h then remove mpc8260.h
Signed-off-by: Christophe Leroy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Link: https://msgid.link/afe23bf3624c389ff17e9789884c78c124b7b202.1691474658.git.christophe.leroy@csgroup.eu
|
|
tqm8xx_setup.c and fs_enet.h don't use any items provided by fs_pd.h
Remove unneeded #include <asm/fs_pd.h>
Signed-off-by: Christophe Leroy <[email protected]>
Signed-off-by: Michael Ellerman <[email protected]>
Link: https://msgid.link/b056c4e986a4a7707fc1994304c34f7bd15d6871.1691474658.git.christophe.leroy@csgroup.eu
|
|
On some I219 devices, ethernet cable plugging detection only works once
from PCI D3 state. Subsequent cable plugging does set PME bit correctly,
but device still doesn't get woken up.
Since I219 connects to the root complex directly, it relies on platform
firmware (ACPI) to wake it up. In this case, the GPE from _PRW only
works for first cable plugging but fails to notify the driver for
subsequent plugging events.
The issue was originally found on CNP, but the same issue can be found
on ADL too. So workaround the issue by continuing use PME poll after
first ACPI wake. As PME poll is always used, the runtime suspend
restriction for CNP can also be removed.
Signed-off-by: Kai-Heng Feng <[email protected]>
Tested-by: Naama Meir <[email protected]>
Acked-by: Sasha Neftin <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: Tony Nguyen <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
This patch adds MQPRIO Qdisc offload in full 'channel' mode which allows
not only setting up pri:tc mapping, but also configuring TX shapers on
external port FIFOs. The K3 CPSW MQPRIO Qdisc offload is expected to work
with VLAN/priority tagged packets. Non-tagged packets have to be mapped
only to TC0.
- TX traffic classes must be rated starting from TC that has highest
priority and with no gaps
- Traffic classes are used starting from 0, that has highest priority
- min_rate defines Committed Information Rate (guaranteed)
- max_rate defines Excess Information Rate (non guaranteed) and offloaded
as (max_rate[i] - tcX_min_rate[i])
- VLAN/priority tagged packets mapped to TC0 will exit switch with VLAN tag
priority 0
The configuration example:
ethtool -L eth1 tx 5
ethtool --set-priv-flags eth1 p0-rx-ptype-rrobin off
tc qdisc add dev eth1 parent root handle 100: mqprio num_tc 3 \
map 0 0 1 2 0 0 0 0 0 0 0 0 0 0 0 0 \
queues 1@0 1@1 1@2 hw 1 mode channel \
shaper bw_rlimit min_rate 0 100mbit 200mbit max_rate 0 101mbit 202mbit
tc qdisc replace dev eth2 handle 100: parent root mqprio num_tc 1 \
map 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 queues 1@0 hw 1
ip link add link eth1 name eth1.100 type vlan id 100
ip link set eth1.100 type vlan egress 0:0 1:1 2:2 3:3 4:4 5:5 6:6 7:7
In the above example two ports share the same TX CPPI queue 0 for low
priority traffic. 3 traffic classes are defined for eth1 and mapped to:
TC0 - low priority, TX CPPI queue 0 -> ext Port 1 fifo0, no rate limit
TC1 - prio 2, TX CPPI queue 1 -> ext Port 1 fifo1, CIR=100Mbit/s, EIR=1Mbit/s
TC2 - prio 3, TX CPPI queue 2 -> ext Port 1 fifo2, CIR=200Mbit/s, EIR=2Mbit/s
Signed-off-by: Grygorii Strashko <[email protected]>
Signed-off-by: Roger Quadros <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The driver depends on CONFIG_OF, it is not necessary to use
of_match_ptr() here.
Signed-off-by: Ruan Jinjie <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The driver depends on CONFIG_OF, it is not necessary to use
of_match_ptr() here.
Signed-off-by: Ruan Jinjie <[email protected]>
Reviewed-by: Simon Horman <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The driver depends on CONFIG_OF, it is not necessary to use
of_match_ptr() here.
Signed-off-by: Ruan Jinjie <[email protected]>
Acked-by: Linus Walleij <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The driver depends on CONFIG_OF, it is not necessary to use
of_match_ptr() here.
Signed-off-by: Ruan Jinjie <[email protected]>
Reviewed-by: Vladimir Oltean <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The driver depends on CONFIG_OF, it is not necessary to use
of_match_ptr() here.
Signed-off-by: Ruan Jinjie <[email protected]>
Acked-by: Linus Walleij <[email protected]>
Reviewed-by: Vladimir Oltean <[email protected]>
Reviewed-by: Linus Walleij <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
In the original RPU query command, the status register values of
multiple RPU tunnels are accumulated by default, which is unreasonable.
This patch Fix it by querying the specified tunnel ID.
The tunnel number of the device can be obtained from firmware
during initialization.
Fixes: ddb54554fa51 ("net: hns3: add DFX registers information for ethtool -d")
Signed-off-by: Jijie Shao <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The dump register function is being refactored.
The third step in refactoring is to support tlv info in regs data for
HNS3 PF driver.
Currently, if we use "ethtool -d" to dump regs value,
the output is as follows:
offset1: 00 01 02 03 04 05 ...
offset2：10 11 12 13 14 15 ...
......
We can't get the value of a register directly.
This patch deletes the original separator information and
add tag_len_value information in regs data.
ethtool can parse register data in key-value format by -d command.
a patch will be added to the ethtool to parse regs data
in the following format:
reg1 : value2
reg2 : value2
......
Signed-off-by: Jijie Shao <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The dump register function is being refactored.
The second step in refactoring is to support tlv info in regs data for
HNS3 PF driver.
Currently, if we use "ethtool -d" to dump regs value,
the output is as follows:
offset1: 00 01 02 03 04 05 ...
offset2：10 11 12 13 14 15 ...
......
We can't get the value of a register directly.
This patch deletes the original separator information and
add tag_len_value information in regs data.
ethtool can parse register data in key-value format by -d command.
a patch will be added to the ethtool to parse regs data
in the following format:
reg1 : value2
reg2 : value2
......
Signed-off-by: Jijie Shao <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The dump register function is being refactored.
The first step in refactoring is put the dump regs function
into a separate file.
Signed-off-by: Jijie Shao <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
As suggested by Jesper and Alexander, we can avoid converting xdp_buff
to xdp_frame in case of XDP_TX to save a bunch of CPU cycles, so that
we can further improve the XDP_TX performance.
Before this patch on i.MX8MP-EVK board, the performance shows as follows.
root@imx8mpevk:~# ./xdp2 eth0
proto 17: 353918 pkt/s
proto 17: 352923 pkt/s
proto 17: 353900 pkt/s
proto 17: 352672 pkt/s
proto 17: 353912 pkt/s
proto 17: 354219 pkt/s
After applying this patch, the performance is improved.
root@imx8mpevk:~# ./xdp2 eth0
proto 17: 369261 pkt/s
proto 17: 369267 pkt/s
proto 17: 369206 pkt/s
proto 17: 369214 pkt/s
proto 17: 369126 pkt/s
proto 17: 369272 pkt/s
Signed-off-by: Wei Fang <[email protected]>
Suggested-by: Alexander Lobakin <[email protected]>
Suggested-by: Jesper Dangaard Brouer <[email protected]>
Reviewed-by: Jesper Dangaard Brouer <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
The XDP_TX feature is not supported before, and all the frames
which are deemed to do XDP_TX action actually do the XDP_DROP
action. So this patch adds the XDP_TX support to FEC driver.
I tested the performance of XDP_TX in XDP_DRV mode and XDP_SKB
mode respectively on i.MX8MP-EVK platform, and as suggested by
Jesper, I also tested the performance of XDP_REDIRECT on the
same platform. And the test steps and results are as follows.
XDP_TX test:
Step 1: One board is used as generator and connects to switch,and
the FEC port of DUT also connects to the switch. Both boards with
flow control off. Then the generator runs the
pktgen_sample03_burst_single_flow.sh script to generate and send
burst traffic to DUT. Note that the size of packet was set to 64
bytes and the procotol of packet was UDP in my test scenario. In
addition, the SMAC of the packet need to be different from the MAC
of the generator, because the xdp2 program will swap the DMAC and
SMAC of the packet and send it back to the generator. If the SMAC
of the generated packet is the MAC of the generator, the generator
will receive the returned traffic which increase the CPU loading
and significantly degrade the transmit speed of the generator, and
finally it affects the test of XDP_TX performance.
Step 2: The DUT runs the xdp2 program to transmit received UDP
packets back out on the same port where they were received.
root@imx8mpevk:~# ./xdp2 eth0
proto 17: 353918 pkt/s
proto 17: 352923 pkt/s
proto 17: 353900 pkt/s
proto 17: 352672 pkt/s
proto 17: 353912 pkt/s
proto 17: 354219 pkt/s
root@imx8mpevk:~# ./xdp2 -S eth0
proto 17: 160604 pkt/s
proto 17: 160708 pkt/s
proto 17: 160564 pkt/s
proto 17: 160684 pkt/s
proto 17: 160640 pkt/s
proto 17: 160720 pkt/s
The above results show that the XDP_TX performance of XDP_DRV mode
is much better than XDP_SKB mode, more than twice that of XDP_SKB
mode, which is in line with our expectation.
XDP_REDIRECT test:
Step1: Both the generator and the FEC port of the DUT connet to the
switch port. All the ports with flow control off, then the generator
runs the pktgen script to generate and send burst traffic to DUT.
Note that the size of packet was set to 64 bytes and the procotol of
packet was UDP in my test scenario.
Step2: The DUT runs the xdp_redirect program to redirect the traffic
from the FEC port to the FEC port itself.
root@imx8mpevk:~# ./xdp_redirect eth0 eth0
Redirecting from eth0 (ifindex 2; driver fec) to eth0
(ifindex 2; driver fec)
Summary 232,302 rx/s 0 err,drop/s 232,344 xmit/s
Summary 234,579 rx/s 0 err,drop/s 234,577 xmit/s
Summary 235,548 rx/s 0 err,drop/s 235,549 xmit/s
Summary 234,704 rx/s 0 err,drop/s 234,703 xmit/s
Summary 235,504 rx/s 0 err,drop/s 235,504 xmit/s
Summary 235,223 rx/s 0 err,drop/s 235,224 xmit/s
Summary 234,509 rx/s 0 err,drop/s 234,507 xmit/s
Summary 235,481 rx/s 0 err,drop/s 235,482 xmit/s
Summary 234,684 rx/s 0 err,drop/s 234,683 xmit/s
Summary 235,520 rx/s 0 err,drop/s 235,520 xmit/s
Summary 235,461 rx/s 0 err,drop/s 235,461 xmit/s
Summary 234,627 rx/s 0 err,drop/s 234,627 xmit/s
Summary 235,611 rx/s 0 err,drop/s 235,611 xmit/s
Packets received : 3,053,753
Average packets/s : 234,904
Packets transmitted : 3,053,792
Average transmit/s : 234,907
Compared the performance of XDP_TX with XDP_REDIRECT, XDP_TX is also
much better than XDP_REDIRECT. It's also in line with our expectation.
Signed-off-by: Wei Fang <[email protected]>
Suggested-by: Jesper Dangaard Brouer <[email protected]>
Suggested-by: Jakub Kicinski <[email protected]>
Reviewed-by: Larysa Zaremba <[email protected]>
Acked-by: Jesper Dangaard Brouer <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
Return result of b44_writephy() instead of zero to
deal with possible error.
Found by Linux Verification Center (linuxtesting.org) with SVACE.
Signed-off-by: Artem Chernyshev <[email protected]>
Reviewed-by: Leon Romanovsky <[email protected]>
Signed-off-by: David S. Miller <[email protected]>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux
Saeed Mahameed says:
====================
mlx5-updates-2023-08-14
1) Handle PTP out of order CQEs issue
2) Check FW status before determining reset successful
3) Expose maximum supported SFs via devlink resource
4) MISC cleanups
* tag 'mlx5-updates-2023-08-14' of git://git.kernel.org/pub/scm/linux/kernel/git/saeed/linux:
net/mlx5: Don't query MAX caps twice
net/mlx5: Remove unused MAX HCA capabilities
net/mlx5: Remove unused CAPs
net/mlx5: Fix error message in mlx5_sf_dev_state_change_handler()
net/mlx5: Remove redundant check of mlx5_vhca_event_supported()
net/mlx5: Use mlx5_sf_start_function_id() helper instead of directly calling MLX5_CAP_GEN()
net/mlx5: Remove redundant SF supported check from mlx5_sf_hw_table_init()
net/mlx5: Use auxiliary_device_uninit() instead of device_put()
net/mlx5: E-switch, Add checking for flow rule destinations
net/mlx5: Check with FW that sync reset completed successfully
net/mlx5: Expose max possible SFs via devlink resource
net/mlx5e: Add recovery flow for tx devlink health reporter for unhealthy PTP SQ
net/mlx5e: Make tx_port_ts logic resilient to out-of-order CQEs
net/mlx5: Consolidate devlink documentation in devlink/mlx5.rst
====================
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Mario reports that loading r8152 on his system leads to a:
netif_napi_add_weight() called with weight 256
warning getting printed. We don't have any solid data
on why such high budget was chosen, and it may cause
stalls in processing other softirqs and rt threads.
So try to switch back to the default (64) weight.
If this slows down someone's system we should investigate
which part of stopping starting the NAPI poll in this
driver are expensive.
Reported-by: Mario Limonciello <[email protected]>
Link: https://lore.kernel.org/all/[email protected]/
Acked-by: Hayes Wang <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|
|
Commit bdfe2da6aefd ("e1000e: cosmetic move of function prototypes to the new mac.h")
declared but never implemented them.
Signed-off-by: Yue Haibing <[email protected]>
Reviewed-by: Tony Nguyen <[email protected]>
Link: https://lore.kernel.org/r/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>
|