| Age | Commit message (Collapse) | Author |
|---|
|
The buffers mapping of the Multi-Packet WQEs (of Striding RQ)
is done via UMR posts, one UMR WQE per an RX MPWQE.
A single MPWQE is capable of serving many incoming packets,
usually larger than the budget of a single napi cycle.
Hence, posting a single UMR WQE per napi cycle (and handling its
completion in the next cycle) works fine in many common cases,
but not always.
When an XDP program is loaded, every MPWQE is capable of serving less
packets, to satisfy the packet-per-page requirement.
Thus, for the same number of packets more MPWQEs (and UMR posts)
are needed (twice as much for the default MTU), giving less latency
room for the UMR completions.
In this patch, we add support for multiple outstanding UMR posts,
to allow faster gap closure between consuming MPWQEs and reposting
them back into the WQ.
For better SW and HW locality, we combine the UMR posts in bulks of
(at least) two.
This is expected to improve packet rate in high CPU scale.
Performance test:
As expected, huge improvement in large-scale (48 cores).
xdp_redirect_map, 64B UDP multi-stream.
Redirect from ConnectX-5 100Gbps to ConnectX-6 100Gbps.
CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz.
Before: Unstable, 7 to 30 Mpps
After: Stable, at 70.5 Mpps
No degradation in other tested scenarios.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Writing 64B CQEs to 128B cache lines results in a RMW operation. Padding
the CQEs to 128B if possible improves performance on 128B cache line
systems like PPC.
Testing on PPC showed up to a 24% improvement in small packet throughput
vs the default behavior, depending on the workload and system topology.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
The function is only used to retrieve CQEs, use the proper type as the
return value.
Signed-off-by: Daniel Jurgens <danielj@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
mlx5e netdevice used to calculate fragment edges by a call to
mlx5_wq_cyc_get_frag_size(). This calculation did not give the correct
indication for queues smaller than a PAGE_SIZE, (broken by default on
PowerPC, where PAGE_SIZE == 64KB). Here it is replaced by the correct new
calls/API.
Since (TX/RX) Work Queues buffers are fragmented, here we introduce
changes to the API in core driver, so that it gets a stride index and
returns the index of last stride on same fragment, and an additional
wrapping function that returns the number of physically contiguous
strides that can be written contiguously to the work queue.
This obsoletes the following API functions, and their buggy
usage in EN driver:
* mlx5_wq_cyc_get_frag_size()
* mlx5_wq_cyc_ctr2fragix()
The new API improves modularity and hides the details of such
calculation for mlx5e netdevice and mlx5_ib rdma drivers.
New calculation is also more efficient, and improves performance
as follows:
Packet rate test: pktgen, UDP / IPv4, 64byte, single ring, 8K ring size.
Before: 16,477,619 pps
After: 17,085,793 pps
3.7% improvement
Fixes: 3a2f70331226 ("net/mlx5: Use order-0 allocations for all WQ types")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Minimal stride size is 16.
Hence, the number of strides in a fragment (of PAGE_SIZE)
is <= PAGE_SIZE / 16 <= 4K.
u16 is sufficient to represent this.
Fixes: 388ca8be0037 ("IB/mlx5: Implement fragmented completion queue (CQ)")
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reviewed-by: Eran Ben Elisha <eranbe@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Add per-ring and global ethtool counters for congested UMR requests.
These events indicate congestion in UMR handlers in HW.
Such event is concluded when there's an outstanding UMR post,
yet the SW consumed at least two additional MPWQEs in the meanwhile.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Now that LRO is not supported for Legacy RQ, there is no source of
out-of-order completions in the WQ, and we can use a cyclic one.
This has multiple advantages:
- reduces the WQE size (smaller PCI transactions).
- lower overhead in datapath (no handling of 'next' pointers).
- no reserved WQE for the WQ head (was need in linked-list).
- allows using a constant map between frag and dma_info struct, in downstream patch.
Performance tests:
ConnectX-4, single core, single RX ring.
Major gain in packet rate of single ring XDP drop.
Bottleneck is shifted form HW (at 16Mpps) to SW (at 20Mpps).
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Complete the transition of all WQ types to use fragmented
order-0 coherent memory instead of high-order allocations.
CQ-WQ already uses order-0.
Here we do the same for cyclic and linked-list WQs.
This allows the driver to load cleanly on systems with a highly
fragmented coherent memory.
Performance tests:
ConnectX-5 100Gbps, CPU: Intel(R) Xeon(R) CPU E5-2680 v3 @ 2.50GHz
Packet rate of 64B packets, single transmit ring, size 8K.
No degradation is sensed.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Use the WQ API to get the WQ size, and to map a counter
into a WQ entry index.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Remove the 'linear' field from struct mlx5_wq_param.
It is redundant, set but never read.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
The current implementation of create CQ requires contiguous
memory, such requirement is problematic once the memory is
fragmented or the system is low in memory, it causes for
failures in dma_zalloc_coherent().
This patch implements new scheme of fragmented CQ to overcome
this issue by introducing new type: 'struct mlx5_frag_buf_ctrl'
to allocate fragmented buffers, rather than contiguous ones.
Base the Completion Queues (CQs) on this new fragmented buffer.
It fixes following crashes:
kworker/29:0: page allocation failure: order:6, mode:0x80d0
CPU: 29 PID: 8374 Comm: kworker/29:0 Tainted: G OE 3.10.0
Workqueue: ib_cm cm_work_handler [ib_cm]
Call Trace:
[<>] dump_stack+0x19/0x1b
[<>] warn_alloc_failed+0x110/0x180
[<>] __alloc_pages_slowpath+0x6b7/0x725
[<>] __alloc_pages_nodemask+0x405/0x420
[<>] dma_generic_alloc_coherent+0x8f/0x140
[<>] x86_swiotlb_alloc_coherent+0x21/0x50
[<>] mlx5_dma_zalloc_coherent_node+0xad/0x110 [mlx5_core]
[<>] ? mlx5_db_alloc_node+0x69/0x1b0 [mlx5_core]
[<>] mlx5_buf_alloc_node+0x3e/0xa0 [mlx5_core]
[<>] mlx5_buf_alloc+0x14/0x20 [mlx5_core]
[<>] create_cq_kernel+0x90/0x1f0 [mlx5_ib]
[<>] mlx5_ib_create_cq+0x3b0/0x4e0 [mlx5_ib]
Signed-off-by: Yonatan Cohen <yonatanc@mellanox.com>
Reviewed-by: Tariq Toukan <tariqt@mellanox.com>
Signed-off-by: Leon Romanovsky <leon@kernel.org>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
A QP in ConnectX is a concatenation of RQ and SQ which share a QP-number
and work together.
Add support for allocating and managing the work-queue buffer for a QP, in
a similar way to how SQs and RQs are already supported.
Signed-off-by: Ilan Tayari <ilant@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Move mlx5e_get_cqe routine to wq.h and rename it to
mlx5_cqwq_get_cqe.
This allows it to be used by other CQ users outside of the
ethernet driver code.
A later patch in this patchset will make use of it from
FPGA code for the FPGA high-speed connection.
Signed-off-by: Ilan Tayari <ilant@mellanox.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
|
|
Add new type of struct mlx5_frag_buf which is used to allocate fragmented
buffers rather than contiguous, and make the Completion Queues (CQs) use
it as they are big (default of 2MB per CQ in Striding RQ).
This fixes the failures of type:
"mlx5e_open_locked: mlx5e_open_channels failed, -12"
due to dma_zalloc_coherent insufficient contiguous coherent memory to
satisfy the driver's request when the user tries to setup more or larger
rings.
Signed-off-by: Tariq Toukan <tariqt@mellanox.com>
Reported-by: Sebastian Ott <sebott@linux.vnet.ibm.com>
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
By affinity hints and XPS, each mlx5e channel is assigned a CPU
core.
Channel DMA coherent memory that is written by the NIC and read
by SW (e.g CQ buffer) is allocated on the NUMA node of the CPU
core assigned for the channel.
Channel DMA coherent memory that is written by SW and read by the
NIC (e.g SQ/RQ buffer) is allocated on the NUMA node of the NIC.
Doorbell record (written by SW and read by the NIC) is an
exception since it is accessed by SW more frequently.
Signed-off-by: Saeed Mahameed <saeedm@mellanox.com>
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
en_[rt]x.c contains the data path related code specific to tx or rx.
en_txrx.c contains data path code which is common for both the rx and
tx, this is mainly napi related code.
Below are the objects that are being used by the hardware and the driver
in the data path:
Channel - one channel per IRQ. Every channel object contains:
RQ - describes the rx queue
TIR - One TIR (Transport Interface Receive) object per flow type. TIR
contains attributes for a type of rx flow (e.g IPv4, IPv6 etc).
A flow is defined in the Flow Table.
Currently TIR describes the RSS hash parameters if exists and LRO
attributes.
SQ - describes the a tx queue. There is one SQ (Send Queue) per
TC (traffic class).
TIS - There is one TIS (Transport Interface Send) per TC. It
describes the TC and may later be extended to describe more
transport properties.
Both RQ and SQ inherit from the object WQ (work queue). This common code
to describe the layout of CQE's WQE's in memory is in the files wq.[cj]
For every channel there is one NAPI context that is used for RX and
for TX.
Driver is using netdev_alloc_skb() to allocate skb's.
Signed-off-by: Amir Vadai <amirv@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
Tariq Toukan
Daniel Jurgens
Yonatan Cohen
Ilan Tayari
Saeed Mahameed
Amir Vadai