Skip to contentSkip to conversation

nvme-tcp receive offloads

PATCH Superseded Aurelien Aptel 4 participants 34 messages Feb 3, 2023 – Mar 22, 2023

Subsystems (14)

v11 of 30 Superseded Latest is v30 →

lore
Share & tools
lore
#0 lore raw

v11 cover nvme-tcp receive offloads

Hi,

Here is the next iteration of our nvme-tcp receive offload series.

The main changes are in patch 3 (netlink).

Rebased on top of today net-next
8065c0e13f98 ("Merge branch 'yt8531-support'")

The changes are also available through git:

Repo: https://github.com/aaptel/linux.git branch nvme-rx-offload-v11
Web: https://github.com/aaptel/linux/tree/nvme-rx-offload-v11

The NVMeTCP offload was presented in netdev 0x16 (video now available):
- https://netdevconf.info/0x16/session.html?NVMeTCP-Offload-%E2%80%93-Implementation-and-Performance-Gains
- https://youtu.be/W74TR-SNgi4

From: Aurelien Aptel <[email protected]>
From: Shai Malin <[email protected]>
From: Ben Ben-Ishay <[email protected]>
From: Boris Pismenny <[email protected]>
From: Or Gerlitz <[email protected]>
From: Yoray Zack <[email protected]>

=========================================

This series adds support for NVMe-TCP receive offloads. The method here does not mandate the offload of the network stack to the device. Instead, these work together with TCP to offload:
1. copy from SKB to the block layer buffers.
2. CRC calculation and verification for received PDU.

The series implements these as a generic offload infrastructure for storage protocols, which calls TCP Direct Data Placement and TCP Offload CRC respectively. We use this infrastructure to implement NVMe-TCP offload for copy and CRC.
Future implementations can reuse the same infrastructure for other protocols such as iSCSI.

Note:
These offloads are similar in nature to the packet-based NIC TLS offloads, which are already upstream (see net/tls/tls_device.c).
You can read more about TLS offload here:
https://www.kernel.org/doc/html/latest/networking/tls-offload.html

Queue Level
===========
The offload for IO queues is initialized after the handshake of the NVMe-TCP protocol is finished by calling nvme_tcp_offload_socket with the tcp socket of the nvme_tcp_queue:
This operation sets all relevant hardware contexts in
hardware. If it fails, then the IO queue proceeds as usual with no offload. If it succeeds then nvme_tcp_setup_ddp and nvme_tcp_teardown_ddp may be called to perform copy offload, and crc offload will be used. This initialization does not change the normal operation of NVMe-TCP in any way besides adding the option to call the above mentioned NDO operations.

For the admin queue, NVMe-TCP does not initialize the offload. Instead, NVMe-TCP calls the driver to configure limits for the controller, such as max_hw_sectors and max_segments, these must be limited to accommodate potential HW resource limits, and to improve performance.

If some error occurs, and the IO queue must be closed or reconnected, then offload is teardown and initialized again. Additionally, we handle netdev down events via the existing error recovery flow.

IO Level
========
The NVMe-TCP layer calls the NIC driver to map block layer buffers to CID using nvme_tcp_setup_ddp before sending the read request. When the response is received, then the NIC HW will write the PDU payload directly into the designated buffer, and build an SKB such that it points into the destination buffer. This SKB represents the entire packet received on the wire, but it points to the block layer buffers. Once NVMe-TCP attempts to copy data from this SKB to the block layer buffer it can skip the copy by checking in the copying function: if (src == dst) -> skip copy

Finally, when the PDU has been processed to completion, the NVMe-TCP layer releases the NIC HW context by calling nvme_tcp_teardown_ddp which asynchronously unmaps the buffers from NIC HW.

The NIC must release its mapping between command IDs and the target buffers. This mapping is released when NVMe-TCP calls the NIC
driver (nvme_tcp_offload_socket).
As completing IOs is performance critical, we introduce asynchronous completions for NVMe-TCP, i.e. NVMe-TCP calls the NIC, which will later call NVMe-TCP to complete the IO (nvme_tcp_ddp_teardown_done).

On the IO level, and in order to use the offload only when a clear performance improvement is expected, the offload is used only for IOs which are bigger than io_threshold.

SKB
===
The DDP (zero-copy) and CRC offloads require two additional bits in the SKB. The ddp bit is useful to prevent condensing of SKBs which are targeted for zero-copy. The crc bit is useful to prevent GRO coalescing SKBs with different offload values. This bit is similar in concept to the "decrypted" bit.

After offload is initialized, we use the SKB's crc bit to indicate that:
"there was no problem with the verification of all CRC fields in this packet's payload". The bit is set to zero if there was an error, or if HW skipped offload for some reason. If *any* SKB in a PDU has (crc != 1), then the calling driver must compute the CRC, and check it. We perform this check, and accompanying software fallback at the end of the processing of a received PDU.

Resynchronization flow
======================
The resynchronization flow is performed to reset the hardware tracking of NVMe-TCP PDUs within the TCP stream. The flow consists of a request from the hardware proxied by the driver, regarding a possible location of a PDU header. Followed by a response from the NVMe-TCP driver.

This flow is rare, and it should happen only after packet loss or reordering events that involve NVMe-TCP PDU headers.

CID Mapping
===========
ConnectX-7 assumes linear CID (0...N-1 for queue of size N) where the Linux NVMe driver uses part of the 16 bit CCID for generation counter.
To address that, we use the existing quirk in the NVMe layer when the HW driver advertises that they don't support the full 16 bit CCID range.

Enablement on ConnectX-7
========================
By default, NVMeTCP offload is disabled in the mlx driver and in the nvme-tcp host. In order to enable it:

# Disable CQE compression (specific for ConnectX) ethtool --set-priv-flags <device> rx_cqe_compress off

# Enable the ULP-DDP
ethtool --ulp-ddp <device> nvme-tcp-ddp on nvme-tcp-ddgst-rx-offload on

# Enable ULP offload in nvme-tcp
modprobe nvme-tcp ulp_offload=1

Following the device ULP-DDP enablement, all the IO queues/sockets which are running on the device are offloaded.

This feature requires a patched ethtool that can be obtained from
Web: https://github.com/aaptel/ethtool/tree/ulp-ddp
Git: https://github.com/aaptel/ethtool.git
Branch: ulp-ddp

Performance
===========
With this implementation, using the ConnectX-7 NIC, we were able to demonstrate the following CPU utilization improvement:

Without data digest:
For 64K queued read IOs – up to 32% improvement in the BW/IOPS (111 Gbps vs. 84 Gbps). For 512K queued read IOs – up to 55% improvement in the BW/IOPS (148 Gbps vs. 98 Gbps).

With data digest:
For 64K queued read IOs – up to 107% improvement in the BW/IOPS (111 Gbps vs. 53 Gbps). For 512K queued read IOs – up to 138% improvement in the BW/IOPS (146 Gbps vs. 61 Gbps).

With small IOs we are not expecting that the offload will show a performance gain.

The test configuration:
- fio command: qd=128, jobs=8.
- Server: Intel(R) Xeon(R) Platinum 8380 CPU @ 2.30GHz, 160 cores.

Patches
=======

Patch 1:  Introduce the infrastructure for all ULP DDP and ULP DDP CRC offloads.
Patch 2:  Add ethtool string sets for ULP DDP capability names and stats names.
Patch 3:  Add ethtool operations to get/set ULP DDP capabilities and statistics.
Patch 4:  Documentation of ULP DDP ethtool netlink messages.
Patch 5:  The iov_iter change to skip copy if (src == dst).
Patch 6:  Export the get_netdev_for_sock function from TLS to generic location.
Patch 7:  NVMe-TCP changes to call NIC driver on queue init/teardown and resync.
Patch 8:  NVMe-TCP changes to call NIC driver on IO operation

setup/teardown, and support async completions.
Patch 9: NVMe-TCP changes to support CRC offload on receive
Also, this patch moves CRC calculation to the end of PDU in case offload requires software fallback.
Patch 10: NVMe-TCP handling of netdev events: stop the offload if netdev is
going down.
Patch 11: Add module parameter to the NVMe-TCP control the enable ULP offload. Patch 12: Documentation of ULP DDP offloads.

The rest of the series is the mlx5 implementation of the offload.

Testing
=======
This series was tested on ConnectX-7 HW using various configurations of IO sizes, queue depths, MTUs, and with both the SPDK and kernel NVMe-TCP targets.

Compatibility
=============
* The offload works with bare-metal or SRIOV.
* The HW can support up to 64K connections per device (assuming no other HW accelerations are used). In this series, we will introduce the support for up to 4k connections, and we have plans to increase it.
* SW TLS could not work together with the NVMeTCP offload as the HW will need to track the NVMeTCP headers in the TCP stream.
* The ConnectX HW support HW TLS, but in ConnectX-7 those features could not co-exists (and it is not part of this series).
* The NVMeTCP offload ConnectX 7 HW can support tunneling, but we don’t see the need for this feature yet.
* NVMe poll queues are not in the scope of this series.

Future Work
===========
* NVMeTCP transmit offload.
* NVMeTCP host offloads incremental features.
* NVMeTCP target offload.

Changes since v10:
=================
- Pass extack to drivers for better error reporting in the .set_caps callback (Jakub).
- netlink: use new callbacks, existing macros, padding, fix size add notifications, update specs (Jakub).

Changes since v9:
=================
- Add missing crc checks in tcp_try_coalesce() (Paolo).
- Add missing ifdef guard for socket ops (Paolo).
- Remove verbose netlink format for statistics (Jakub).
- Use regular attributes for statistics (Jakub).
- Expose and document individual stats to uAPI (Jakub).
- Move ethtool ops for caps&stats to netdev_ops->ulp_ddp_ops (Jakub).

Changes since v8:
=================
- Make stats stringset global instead of per-device (Jakub).
- Remove per-queue stats (Jakub).
- Rename ETH_SS_ULP_DDP stringset to ETH_SS_ULP_DDP_CAPS.
- Update & fix kdoc comments.
- Use 80 columns limit for nvme code.

Changes since v7:
=================
- Remove ULP DDP netdev->feature bit (Jakub).
- Expose ULP DDP capabilities to userspace via ethtool netlink messages (Jakub).
- Move ULP DDP stats to dedicated stats group (Jakub).
- Add ethtool_ops operations for setting capabilities and getting stats (Jakub).
- Move ulp_ddp_netdev_ops into net_device_ops (Jakub).
- Use union for protocol-specific struct instances (Jakub).
- Fold netdev_sk_get_lowest_dev() into get_netdev_for_sock (Christoph).
- Rename memcpy skip patch to something more obvious (Christoph).
- Move capabilities from ulp_ddp.h to ulp_ddp_caps.h.
- Add "Compatibility" section on the cover letter (Sagi).

Changes since v6:
=================
- Moved IS_ULP_{DDP,CRC} macros to skb_is_ulp_{ddp,crc} inline functions (Jakub).
- Fix copyright notice (Leon).
- Added missing ifdef to allow build with MLX5_EN_TLS disabled.
- Fix space alignment, indent and long lines (max 99 columns).
- Add missing field documentation in ulp_ddp.h.

Changes since v5:
=================
- Limit the series to RX offloads.
- Added two separated skb indications to avoid wrong flushing of GRO when aggerating offloaded packets.
- Use accessor functions for skb->ddp and skb->crc (Eric D) bits.
- Add kernel-doc for get_netdev_for_sock (Christoph).
- Remove ddp_iter* routines and only modify _copy_to_iter (Al Viro, Christoph).
- Remove consume skb (Sagi).
- Add a knob in the ddp limits struct for the HW driver to advertise if they need the nvme-tcp driver to apply the generation counter quirk. Use this knob for the mlx5 CX7 offload.
- bugfix: use u8 flags instead of bool in mlx5e_nvmeotcp_queue->dgst.
- bugfix: use sg_dma_len(sgl) instead of sgl->length.
- bugfix: remove sgl leak in nvme_tcp_setup_ddp().
- bugfix: remove sgl leak when only using DDGST_RX offload.
- Add error check for dma_map_sg().
- Reduce #ifdef by using dummy macros/functions.
- Remove redundant netdev null check in nvme_tcp_pdu_last_send().
- Rename ULP_DDP_RESYNC_{REQ -> PENDING}.
- Add per-ulp limits struct (Sagi).
- Add ULP DDP capabilities querying (Sagi).
- Simplify RX DDGST logic (Sagi).
- Document resync flow better.
- Add ulp_offload param to nvme-tcp module to enable ULP offload (Sagi).
- Add a revert commit to reintroduce nvme_tcp_queue->queue_size.

Changes since v4:
=================
- Add transmit offload patches.
- Use one feature bit for both receive and transmit offload.

Changes since v3:
=================
- Use DDP_TCP ifdefs in iov_iter and skb iterators to minimize impact when compiled out (Christoph).
- Simplify netdev references and reduce the use of get_netdev_for_sock (Sagi).
- Avoid "static" in it's own line, move it one line down (Christoph)
- Pass (queue, skb, *offset) and retrieve the pdu_seq in nvme_tcp_resync_response (Sagi).
- Add missing assignment of offloading_netdev to null in offload_limits error case (Sagi).
- Set req->offloaded = false once -- the lifetime rules are:
set to false on cmd_setup / set to true when ddp setup succeeds (Sagi).
- Replace pr_info_ratelimited with dev_info_ratelimited (Sagi).
- Add nvme_tcp_complete_request and invoke it from two similar call sites (Sagi).
- Introduce nvme_tcp_req_map_sg earlier in the series (Sagi).
- Add nvme_tcp_consume_skb and put into it a hunk from nvme_tcp_recv_data to handle copy with and without offload.

Changes since v2:
=================
- Use skb->ddp_crc for copy offload to avoid skb_condense.
- Default mellanox driver support to no (experimental feature).
- In iov_iter use non-ddp functions for kvec and iovec.
- Remove typecasting in NVMe-TCP.

Changes since v1:
=================
- Rework iov_iter copy skip if src==dst to be less intrusive (David Ahern).
- Add tcp-ddp documentation (David Ahern).
- Refactor mellanox driver patches into more patches (Saeed Mahameed).
- Avoid pointer casting (David Ahern).
- Rename NVMe-TCP offload flags (Shai Malin).
- Update cover-letter according to the above.

Changes since RFC v1:
=====================
- Split mlx5 driver patches to several commits.
- Fix NVMe-TCP handling of recovery flows. In particular, move queue offload. init/teardown to the start/stop functions.

Aurelien Aptel (6):
net/ethtool: add new stringset ETH_SS_ULP_DDP_{CAPS,STATS} net/ethtool: add ULP_DDP_{GET,SET} operations for caps and stats
Documentation: document netlink ULP_DDP_GET/SET messages net/tls,core: export get_netdev_for_sock
nvme-tcp: Add modparam to control the ULP offload enablement net/mlx5e: NVMEoTCP, statistics

Ben Ben-Ishay (8):
iov_iter: skip copy if src == dst for direct data placement net/mlx5: Add NVMEoTCP caps, HW bits, 128B CQE and enumerations net/mlx5e: NVMEoTCP, offload initialization net/mlx5e: NVMEoTCP, use KLM UMRs for buffer registration net/mlx5e: NVMEoTCP, queue init/teardown
net/mlx5e: NVMEoTCP, ddp setup and resync
net/mlx5e: NVMEoTCP, async ddp invalidation net/mlx5e: NVMEoTCP, data-path for DDP+DDGST offload

Boris Pismenny (4):
net: Introduce direct data placement tcp offload
nvme-tcp: Add DDP offload control path
nvme-tcp: Add DDP data-path
net/mlx5e: TCP flow steering for nvme-tcp acceleration

Or Gerlitz (5):
nvme-tcp: Deal with netdevice DOWN events
net/mlx5e: Rename from tls to transport static params net/mlx5e: Refactor ico sq polling to get budget net/mlx5e: Have mdev pointer directly on the icosq structure net/mlx5e: Refactor doorbell function to allow avoiding a completion

Yoray Zack (2):
nvme-tcp: RX DDGST offload
Documentation: add ULP DDP offload documentation

 Documentation/netlink/specs/ethtool.yaml      |  102 ++
 Documentation/networking/ethtool-netlink.rst  |   92 ++
 Documentation/networking/index.rst            |    1 +
 Documentation/networking/statistics.rst       |    1 +
 Documentation/networking/ulp-ddp-offload.rst  |  376 ++++++
 .../net/ethernet/mellanox/mlx5/core/Kconfig   |   11 +
 .../net/ethernet/mellanox/mlx5/core/Makefile  |    3 +
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |    9 +
 .../net/ethernet/mellanox/mlx5/core/en/fs.h   |    4 +-
 .../ethernet/mellanox/mlx5/core/en/params.c   |   12 +-
 .../ethernet/mellanox/mlx5/core/en/params.h   |    3 +
 .../mellanox/mlx5/core/en/reporter_rx.c       |    4 +-
 .../ethernet/mellanox/mlx5/core/en/rx_res.c   |   28 +
 .../ethernet/mellanox/mlx5/core/en/rx_res.h   |    4 +
 .../net/ethernet/mellanox/mlx5/core/en/tir.c  |   15 +
 .../net/ethernet/mellanox/mlx5/core/en/tir.h  |    2 +
 .../net/ethernet/mellanox/mlx5/core/en/txrx.h |   28 +-
 .../mlx5/core/en_accel/common_utils.h         |   32 +
 .../mellanox/mlx5/core/en_accel/en_accel.h    |    3 +
 .../mellanox/mlx5/core/en_accel/fs_tcp.c      |   12 +-
 .../mellanox/mlx5/core/en_accel/fs_tcp.h      |    2 +-
 .../mellanox/mlx5/core/en_accel/ktls.c        |    2 +-
 .../mellanox/mlx5/core/en_accel/ktls_rx.c     |    8 +-
 .../mellanox/mlx5/core/en_accel/ktls_tx.c     |    8 +-
 .../mellanox/mlx5/core/en_accel/ktls_txrx.c   |   36 +-
 .../mellanox/mlx5/core/en_accel/ktls_utils.h  |   17 +-
 .../mellanox/mlx5/core/en_accel/nvmeotcp.c    | 1116 +++++++++++++++++
 .../mellanox/mlx5/core/en_accel/nvmeotcp.h    |  142 +++
 .../mlx5/core/en_accel/nvmeotcp_rxtx.c        |  325 +++++
 .../mlx5/core/en_accel/nvmeotcp_rxtx.h        |   37 +
 .../mlx5/core/en_accel/nvmeotcp_stats.c       |   66 +
 .../mlx5/core/en_accel/nvmeotcp_utils.h       |   66 +
 .../ethernet/mellanox/mlx5/core/en_ethtool.c  |    6 +
 .../net/ethernet/mellanox/mlx5/core/en_fs.c   |    4 +-
 .../net/ethernet/mellanox/mlx5/core/en_main.c |   31 +-
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   |   67 +-
 .../ethernet/mellanox/mlx5/core/en_stats.h    |    7 +
 .../net/ethernet/mellanox/mlx5/core/en_txrx.c |    4 +-
 drivers/net/ethernet/mellanox/mlx5/core/fw.c  |    6 +
 .../net/ethernet/mellanox/mlx5/core/main.c    |    1 +
 drivers/nvme/host/tcp.c                       |  567 ++++++++-
 include/linux/ethtool.h                       |   32 +
 include/linux/mlx5/device.h                   |   59 +-
 include/linux/mlx5/mlx5_ifc.h                 |   82 +-
 include/linux/mlx5/qp.h                       |    1 +
 include/linux/netdevice.h                     |   18 +-
 include/linux/skbuff.h                        |   24 +
 include/net/inet_connection_sock.h            |    6 +
 include/net/ulp_ddp.h                         |  189 +++
 include/net/ulp_ddp_caps.h                    |   35 +
 include/uapi/linux/ethtool.h                  |    4 +
 include/uapi/linux/ethtool_netlink.h          |   40 +
 lib/iov_iter.c                                |    8 +-
 net/Kconfig                                   |   20 +
 net/core/dev.c                                |   26 +-
 net/core/skbuff.c                             |    3 +-
 net/ethtool/Makefile                          |    2 +-
 net/ethtool/bitset.c                          |   20 +-
 net/ethtool/common.c                          |   23 +
 net/ethtool/common.h                          |    2 +
 net/ethtool/netlink.c                         |   20 +
 net/ethtool/netlink.h                         |    4 +
 net/ethtool/strset.c                          |   11 +
 net/ethtool/ulp_ddp.c                         |  316 +++++
 net/ipv4/tcp_input.c                          |   13 +-
 net/ipv4/tcp_ipv4.c                           |    3 +
 net/ipv4/tcp_offload.c                        |    3 +
 net/tls/tls_device.c                          |   16 -
 68 files changed, 4082 insertions(+), 158 deletions(-)
 create mode 100644 Documentation/networking/ulp-ddp-offload.rst
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/common_utils.h
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.h
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_rxtx.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_rxtx.h
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_stats.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_utils.h
 create mode 100644 include/net/ulp_ddp.h
 create mode 100644 include/net/ulp_ddp_caps.h
 create mode 100644 net/ethtool/ulp_ddp.c
Signature · 1 line
2.31.1
#1 lore raw

PATCH v11 1/25 net: Introduce direct data placement tcp offload

From: Boris Pismenny <[email protected]>

This commit introduces direct data placement (DDP) offload for TCP.

The motivation is saving compute resources/cycles that are spent to copy data from SKBs to the block layer buffers and CRC calculation/verification for received PDUs (Protocol Data Units).

The DDP capability is accompanied by new net_device operations that configure hardware contexts.

There is a context per socket, and a context per DDP operation. Additionally, a resynchronization routine is used to assist hardware handle TCP OOO, and continue the offload. Furthermore, we let the offloading driver advertise what is the max hw sectors/segments.

The interface includes the following net-device ddp operations:

1. sk_add - add offload for the queue represented by socket+config pair
2. sk_del - remove the offload for the socket/queue
3. ddp_setup - request copy offload for buffers associated with an IO
4. ddp_teardown - release offload resources for that IO
5. limits - query NIC driver for quirks and limitations (e.g. max number of scatter gather entries per IO)
6. set_caps - request ULP DDP capabilities enablement
7. get_stats - query NIC driver for ULP DDP stats

Using this interface, the NIC hardware will scatter TCP payload directly to the BIO pages according to the command_id.

To maintain the correctness of the network stack, the driver is expected to construct SKBs that point to the BIO pages.

The SKB passed to the network stack from the driver represents data as it is on the wire, while it is pointing directly to data in destination buffers.

As a result, data from page frags should not be copied out to the linear part. To avoid needless copies, such as when using skb_condense, we mark the skb->ulp_ddp bit.
In addition, the skb->ulp_crc will be used by the upper layers to determine if CRC re-calculation is required. The two separated skb indications are needed to avoid false positives GRO flushing events.

Follow-up patches will use this interface for DDP in NVMe-TCP.

Capability bits stored in net_device allow drivers to report which ULP DDP capabilities a device supports. Control over these capabilities will be exposed to userspace in later patches.

---
 include/linux/netdevice.h          |  15 +++
 include/linux/skbuff.h             |  24 ++++
 include/net/inet_connection_sock.h |   6 +
 include/net/ulp_ddp.h              | 189 +++++++++++++++++++++++++++++
 include/net/ulp_ddp_caps.h         |  35 ++++++
 net/Kconfig                        |  20 +++
 net/core/skbuff.c                  |   3 +-
 net/ipv4/tcp_input.c               |  13 +-
 net/ipv4/tcp_ipv4.c                |   3 +
 net/ipv4/tcp_offload.c             |   3 +
 10 files changed, 309 insertions(+), 2 deletions(-)
 create mode 100644 include/net/ulp_ddp.h
 create mode 100644 include/net/ulp_ddp_caps.h
include/linux/netdevice.h +15 -0
index d5ef4c1fedd2..8861a2f04e00 100644
@@ -52,6 +52,10 @@
 #include <net/net_trackers.h>
 #include <net/net_debug.h>
 
+#ifdef CONFIG_ULP_DDP
+#include <net/ulp_ddp_caps.h>
+#endif
+
 struct netpoll_info;
 struct device;
 struct ethtool_ops;
@@ -1393,6 +1397,8 @@ struct netdev_net_notifier {
  *	Get hardware timestamp based on normal/adjustable time or free running
  *	cycle counter. This function is required if physical clock supports a
  *	free running cycle counter.
+ * struct ulp_ddp_dev_ops *ulp_ddp_ops;
+ *	ULP DDP operations (see include/net/ulp_ddp.h)
  */
 struct net_device_ops {
 	int			(*ndo_init)(struct net_device *dev);
@@ -1617,6 +1623,9 @@ struct net_device_ops {
 	ktime_t			(*ndo_get_tstamp)(struct net_device *dev,
 						  const struct skb_shared_hwtstamps *hwtstamps,
 						  bool cycles);
+#if IS_ENABLED(CONFIG_ULP_DDP)
+	const struct ulp_ddp_dev_ops	*ulp_ddp_ops;
+#endif
 };
 
 struct xdp_metadata_ops {
@@ -1789,6 +1798,9 @@ enum netdev_ml_priv_type {
  *	@mpls_features:	Mask of features inheritable by MPLS
  *	@gso_partial_features: value(s) from NETIF_F_GSO\*
  *
+ *	@ulp_ddp_caps:	Bitflags keeping track of supported and enabled
+ *			ULP DDP capabilities.
+ *
  *	@ifindex:	interface index
  *	@group:		The group the device belongs to
  *
@@ -2083,6 +2095,9 @@ struct net_device {
 	netdev_features_t	mpls_features;
 	netdev_features_t	gso_partial_features;
 
+#ifdef CONFIG_ULP_DDP
+	struct ulp_ddp_netdev_caps ulp_ddp_caps;
+#endif
 	unsigned int		min_mtu;
 	unsigned int		max_mtu;
 	unsigned short		type;
include/linux/skbuff.h +24 -0
index 5ba12185f43e..9d1c35b04e80 100644
@@ -806,6 +806,8 @@ typedef unsigned char *sk_buff_data_t;
  *		delivery_time in mono clock base (i.e. EDT).  Otherwise, the
  *		skb->tstamp has the (rcv) timestamp at ingress and
  *		delivery_time at egress.
+ *	@ulp_ddp: DDP offloaded
+ *	@ulp_crc: CRC offloaded
  *	@napi_id: id of the NAPI struct this skb came from
  *	@sender_cpu: (aka @napi_id) source CPU in XPS
  *	@alloc_cpu: CPU which did the skb allocation.
@@ -978,6 +980,10 @@ struct sk_buff {
 	__u8			slow_gro:1;
 	__u8			csum_not_inet:1;
 	__u8			scm_io_uring:1;
+#ifdef CONFIG_ULP_DDP
+	__u8                    ulp_ddp:1;
+	__u8			ulp_crc:1;
+#endif
 
 #ifdef CONFIG_NET_SCHED
 	__u16			tc_index;	/* traffic control index */
@@ -5048,5 +5054,23 @@ static inline void skb_mark_for_recycle(struct sk_buff *skb)
 }
 #endif
 
+static inline bool skb_is_ulp_ddp(struct sk_buff *skb)
+{
+#ifdef CONFIG_ULP_DDP
+	return skb->ulp_ddp;
+#else
+	return 0;
+#endif
+}
+
+static inline bool skb_is_ulp_crc(struct sk_buff *skb)
+{
+#ifdef CONFIG_ULP_DDP
+	return skb->ulp_crc;
+#else
+	return 0;
+#endif
+}
+
 #endif	/* __KERNEL__ */
 #endif	/* _LINUX_SKBUFF_H */
include/net/inet_connection_sock.h +6 -0
index c2b15f7e5516..b11fbbc95541 100644
@@ -68,6 +68,8 @@ struct inet_connection_sock_af_ops {
  * @icsk_ulp_ops	   Pluggable ULP control hook
  * @icsk_ulp_data	   ULP private data
  * @icsk_clean_acked	   Clean acked data hook
+ * @icsk_ulp_ddp_ops	   Pluggable ULP direct data placement control hook
+ * @icsk_ulp_ddp_data	   ULP direct data placement private data
  * @icsk_ca_state:	   Congestion control state
  * @icsk_retransmits:	   Number of unrecovered [RTO] timeouts
  * @icsk_pending:	   Scheduled timer event
@@ -98,6 +100,10 @@ struct inet_connection_sock {
 	const struct tcp_ulp_ops  *icsk_ulp_ops;
 	void __rcu		  *icsk_ulp_data;
 	void (*icsk_clean_acked)(struct sock *sk, u32 acked_seq);
+#ifdef CONFIG_ULP_DDP
+	const struct ulp_ddp_ulp_ops  *icsk_ulp_ddp_ops;
+	void __rcu		  *icsk_ulp_ddp_data;
+#endif
 	unsigned int		  (*icsk_sync_mss)(struct sock *sk, u32 pmtu);
 	__u8			  icsk_ca_state:5,
 				  icsk_ca_initialized:1,
include/net/ulp_ddp.h new file +189 -0

189 lines not on this page · raw message

include/net/ulp_ddp_caps.h new file +35 -0
new file mode 100644
index 000000000000..8ee964a4df78
@@ -0,0 +1,35 @@
+/* SPDX-License-Identifier: GPL-2.0
+ *
+ * ulp_ddp.h
+ *	Author:	Aurelien Aptel <[email protected]>
+ *	Copyright (c) 2023, NVIDIA CORPORATION & AFFILIATES.  All rights reserved.
+ */
+#ifndef _ULP_DDP_CAPS_H
+#define _ULP_DDP_CAPS_H
+
+#include <linux/types.h>
+
+enum {
+	ULP_DDP_C_NVME_TCP_BIT,
+	ULP_DDP_C_NVME_TCP_DDGST_RX_BIT,
+
+	/* add capabilities above */
+	ULP_DDP_C_COUNT,
+};
+
+struct ulp_ddp_netdev_caps {
+	DECLARE_BITMAP(active, ULP_DDP_C_COUNT);
+	DECLARE_BITMAP(hw, ULP_DDP_C_COUNT);
+};
+
+static inline bool ulp_ddp_cap_turned_on(unsigned long *old, unsigned long *new, int bit_nr)
+{
+	return !test_bit(bit_nr, old) && test_bit(bit_nr, new);
+}
+
+static inline bool ulp_ddp_cap_turned_off(unsigned long *old, unsigned long *new, int bit_nr)
+{
+	return test_bit(bit_nr, old) && !test_bit(bit_nr, new);
+}
+
+#endif
net/Kconfig +20 -0
index 48c33c222199..3c59eba4a438 100644
@@ -471,4 +471,24 @@ config NETDEV_ADDR_LIST_TEST
 	default KUNIT_ALL_TESTS
 	depends on KUNIT
 
+config ULP_DDP
+	bool "ULP direct data placement offload"
+	help
+	  This feature provides a generic infrastructure for Direct
+	  Data Placement (DDP) offload for Upper Layer Protocols (ULP,
+	  such as NVMe-TCP).
+
+	  If the ULP and NIC driver supports it, the ULP code can
+	  request the NIC to place ULP response data directly
+	  into application memory, avoiding a costly copy.
+
+	  This infrastructure also allows for offloading the ULP data
+	  integrity checks (e.g. data digest) that would otherwise
+	  require another costly pass on the data we managed to avoid
+	  copying.
+
+	  For more information, see
+	  <file:Documentation/networking/ulp-ddp-offload.rst>.
+
+
 endif   # if NET
net/core/skbuff.c +2 -1
index 44a19805c355..d6e12b8cc46c 100644
@@ -72,6 +72,7 @@
 #include <net/mptcp.h>
 #include <net/mctp.h>
 #include <net/page_pool.h>
+#include <net/ulp_ddp.h>
 
 #include <linux/uaccess.h>
 #include <trace/events/skb.h>
@@ -6520,7 +6521,7 @@ void skb_condense(struct sk_buff *skb)
 {
 	if (skb->data_len) {
 		if (skb->data_len > skb->end - skb->tail ||
-		    skb_cloned(skb))
+		    skb_cloned(skb) || skb_is_ulp_ddp(skb))
 			return;
 
 		/* Nice, we can free page frag(s) right now */
net/ipv4/tcp_input.c +12 -1
index cc072d2cfcd8..4a7edad5b2b7 100644
@@ -4671,7 +4671,10 @@ static bool tcp_try_coalesce(struct sock *sk,
 	if (from->decrypted != to->decrypted)
 		return false;
 #endif
-
+#ifdef CONFIG_ULP_DDP
+	if (skb_is_ulp_crc(from) != skb_is_ulp_crc(to))
+		return false;
+#endif
 	if (!skb_try_coalesce(to, from, fragstolen, &delta))
 		return false;
 
@@ -5234,6 +5237,10 @@ tcp_collapse(struct sock *sk, struct sk_buff_head *list, struct rb_root *root,
 		memcpy(nskb->cb, skb->cb, sizeof(skb->cb));
 #ifdef CONFIG_TLS_DEVICE
 		nskb->decrypted = skb->decrypted;
+#endif
+#ifdef CONFIG_ULP_DDP
+		nskb->ulp_ddp = skb->ulp_ddp;
+		nskb->ulp_crc = skb->ulp_crc;
 #endif
 		TCP_SKB_CB(nskb)->seq = TCP_SKB_CB(nskb)->end_seq = start;
 		if (list)
@@ -5267,6 +5274,10 @@ tcp_collapse(struct sock *sk, struct sk_buff_head *list, struct rb_root *root,
 #ifdef CONFIG_TLS_DEVICE
 				if (skb->decrypted != nskb->decrypted)
 					goto end;
+#endif
+#ifdef CONFIG_ULP_DDP
+				if (skb_is_ulp_crc(skb) != skb_is_ulp_crc(nskb))
+					goto end;
 #endif
 			}
 		}
net/ipv4/tcp_ipv4.c +3 -0
index ea370afa70ed..5d57aa9f9abd 100644
@@ -1861,6 +1861,9 @@ bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb,
 	      TCP_SKB_CB(skb)->tcp_flags) & (TCPHDR_ECE | TCPHDR_CWR)) ||
 #ifdef CONFIG_TLS_DEVICE
 	    tail->decrypted != skb->decrypted ||
+#endif
+#ifdef CONFIG_ULP_DDP
+	    skb_is_ulp_crc(tail) != skb_is_ulp_crc(skb) ||
 #endif
 	    thtail->doff != th->doff ||
 	    memcmp(thtail + 1, th + 1, hdrlen - sizeof(*th)))
net/ipv4/tcp_offload.c +3 -0
index 45dda7889387..2e62f18e85c0 100644
@@ -268,6 +268,9 @@ struct sk_buff *tcp_gro_receive(struct list_head *head, struct sk_buff *skb)
 #ifdef CONFIG_TLS_DEVICE
 	flush |= p->decrypted ^ skb->decrypted;
 #endif
+#ifdef CONFIG_ULP_DDP
+	flush |= skb_is_ulp_crc(p) ^ skb_is_ulp_crc(skb);
+#endif
 
 	if (flush || skb_gro_receive(p, skb)) {
 		mss = 1;
Signature · 1 line
2.31.1
#2 lore raw

PATCH v11 2/25 net/ethtool: add new stringset ETH_SS_ULP_DDP_{CAPS,STATS}

In reply to Aurelien Aptel

This commit exposes ULP DDP capability and statistics names to userspace via netlink.

In order to support future ULP DDP capabilities and statistics without having to change the netlink protocol (and userspace ethtool) we add new string sets to let userspace dynamically fetch what the kernel supports.

* ETH_SS_ULP_DDP_CAPS stores names of ULP DDP capabilities
* ETH_SS_ULP_DDP_STATS stores names of ULP DDP statistics.

These stringsets will be used in later commits when implementing the new ULP DDP GET/SET netlink messages.

We keep the convention of strset.c of having the static_assert() right after the array declaration, despite the checkpatch warning.

---
 include/linux/ethtool.h              | 32 ++++++++++++++++++++++++++++
 include/uapi/linux/ethtool.h         |  4 ++++
 include/uapi/linux/ethtool_netlink.h | 22 +++++++++++++++++++
 net/ethtool/common.c                 | 23 ++++++++++++++++++++
 net/ethtool/common.h                 |  2 ++
 net/ethtool/strset.c                 | 11 ++++++++++
 6 files changed, 94 insertions(+)
include/linux/ethtool.h +32 -0
index 515c78d8eb7c..6b55db2b8287 100644
@@ -584,6 +584,38 @@ struct ethtool_mm_stats {
 	u64 MACMergeHoldCount;
 };
 
+/**
+ * struct ethtool_ulp_ddp_stats - ULP DDP offload statistics
+ * @rx_nvmeotcp_sk_add: number of sockets successfully prepared for offloading.
+ * @rx_nvmeotcp_sk_add_fail: number of sockets that failed to be prepared for offloading.
+ * @rx_nvmeotcp_sk_del: number of sockets where offloading has been removed.
+ * @rx_nvmeotcp_ddp_setup: number of NVMeTCP PDU successfully prepared for Direct Data Placement.
+ * @rx_nvmeotcp_ddp_setup_fail: number of PDUs that failed DDP preparation.
+ * @rx_nvmeotcp_ddp_teardown: number of PDUs done with DDP.
+ * @rx_nvmeotcp_drop: number of PDUs dropped.
+ * @rx_nvmeotcp_resync: number of resync.
+ * @rx_nvmeotcp_packets: number of offloaded PDUs.
+ * @rx_nvmeotcp_bytes: number of offloaded bytes.
+ */
+struct ethtool_ulp_ddp_stats {
+	u64 rx_nvmeotcp_sk_add;
+	u64 rx_nvmeotcp_sk_add_fail;
+	u64 rx_nvmeotcp_sk_del;
+	u64 rx_nvmeotcp_ddp_setup;
+	u64 rx_nvmeotcp_ddp_setup_fail;
+	u64 rx_nvmeotcp_ddp_teardown;
+	u64 rx_nvmeotcp_drop;
+	u64 rx_nvmeotcp_resync;
+	u64 rx_nvmeotcp_packets;
+	u64 rx_nvmeotcp_bytes;
+
+	/*
+	 * add new stats at the end and keep in sync with
+	 * - ETHTOOL_ULP_DDP_STATS_* enum in uapi
+	 * - ulp_ddp_stats_name stringset
+	 */
+};
+
 /**
  * struct ethtool_ops - optional netdev operations
  * @cap_link_lanes_supported: indicates if the driver supports lanes
include/uapi/linux/ethtool.h +4 -0
index f7fba0dc87e5..8b8585b5fa56 100644
@@ -681,6 +681,8 @@ enum ethtool_link_ext_substate_module {
  * @ETH_SS_STATS_ETH_MAC: names of IEEE 802.3 MAC statistics
  * @ETH_SS_STATS_ETH_CTRL: names of IEEE 802.3 MAC Control statistics
  * @ETH_SS_STATS_RMON: names of RMON statistics
+ * @ETH_SS_ULP_DDP_CAPS: names of ULP DDP capabilities
+ * @ETH_SS_ULP_DDP_STATS: names of ULP DDP statistics
  *
  * @ETH_SS_COUNT: number of defined string sets
  */
@@ -706,6 +708,8 @@ enum ethtool_stringset {
 	ETH_SS_STATS_ETH_MAC,
 	ETH_SS_STATS_ETH_CTRL,
 	ETH_SS_STATS_RMON,
+	ETH_SS_ULP_DDP_CAPS,
+	ETH_SS_ULP_DDP_STATS,
 
 	/* add new constants above here */
 	ETH_SS_COUNT
include/uapi/linux/ethtool_netlink.h +22 -0
index ffb073c0dbb4..1722581b2db6 100644
@@ -972,6 +972,28 @@ enum {
 	ETHTOOL_A_MM_MAX = (__ETHTOOL_A_MM_CNT - 1)
 };
 
+/* ULP DDP */
+
+enum {
+	ETHTOOL_A_ULP_DDP_STATS_UNSPEC,
+	ETHTOOL_A_ULP_DDP_STATS_PAD,
+
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_SK_ADD,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_SK_ADD_FAIL,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_SK_DEL,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DDP_SETUP,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DDP_SETUP_FAIL,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DDP_TEARDOWN,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DROP,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_RESYNC,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_PACKETS,
+	ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_BYTES,
+
+	/* add new constants above here */
+	__ETHTOOL_A_ULP_DDP_STATS_CNT,
+	ETHTOOL_A_ULP_DDP_STATS_MAX = __ETHTOOL_A_ULP_DDP_STATS_CNT - 1
+};
+
 /* generic netlink info */
 #define ETHTOOL_GENL_NAME "ethtool"
 #define ETHTOOL_GENL_VERSION 1
net/ethtool/common.c +23 -0
index 5fb19050991e..751017dca228 100644
@@ -5,6 +5,7 @@
 #include <linux/phy.h>
 #include <linux/rtnetlink.h>
 #include <linux/ptp_clock_kernel.h>
+#include <net/ulp_ddp_caps.h>
 
 #include "common.h"
 
@@ -465,6 +466,28 @@ const char udp_tunnel_type_names[][ETH_GSTRING_LEN] = {
 static_assert(ARRAY_SIZE(udp_tunnel_type_names) ==
 	      __ETHTOOL_UDP_TUNNEL_TYPE_CNT);
 
+const char ulp_ddp_caps_names[][ETH_GSTRING_LEN] = {
+	[ULP_DDP_C_NVME_TCP_BIT]		= "nvme-tcp-ddp",
+	[ULP_DDP_C_NVME_TCP_DDGST_RX_BIT]	= "nvme-tcp-ddgst-rx-offload",
+};
+static_assert(ARRAY_SIZE(ulp_ddp_caps_names) == ULP_DDP_C_COUNT);
+
+const char ulp_ddp_stats_names[][ETH_GSTRING_LEN] = {
+	[ETHTOOL_A_ULP_DDP_STATS_UNSPEC]			= "unspec",
+	[ETHTOOL_A_ULP_DDP_STATS_PAD]				= "pad",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_SK_ADD]		= "rx_nvmeotcp_sk_add",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_SK_ADD_FAIL]	= "rx_nvmeotcp_sk_add_fail",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_SK_DEL]		= "rx_nvmeotcp_sk_del",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DDP_SETUP]		= "rx_nvmeotcp_ddp_setup",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DDP_SETUP_FAIL]	= "rx_nvmeotcp_ddp_setup_fail",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DDP_TEARDOWN]	= "rx_nvmeotcp_ddp_teardown",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_DROP]		= "rx_nvmeotcp_drop",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_RESYNC]		= "rx_nvmeotcp_resync",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_PACKETS]		= "rx_nvmeotcp_packets",
+	[ETHTOOL_A_ULP_DDP_STATS_RX_NVMEOTCP_BYTES]		= "rx_nvmeotcp_bytes",
+};
+static_assert(ARRAY_SIZE(ulp_ddp_stats_names) == __ETHTOOL_A_ULP_DDP_STATS_CNT);
+
 /* return false if legacy contained non-0 deprecated fields
  * maxtxpkt/maxrxpkt. rest of ksettings always updated
  */
net/ethtool/common.h +2 -0
index 28b8aaaf9bcb..ebb0abec04a3 100644
@@ -36,6 +36,8 @@ extern const char sof_timestamping_names[][ETH_GSTRING_LEN];
 extern const char ts_tx_type_names[][ETH_GSTRING_LEN];
 extern const char ts_rx_filter_names[][ETH_GSTRING_LEN];
 extern const char udp_tunnel_type_names[][ETH_GSTRING_LEN];
+extern const char ulp_ddp_caps_names[][ETH_GSTRING_LEN];
+extern const char ulp_ddp_stats_names[][ETH_GSTRING_LEN];
 
 int __ethtool_get_link(struct net_device *dev);
 
net/ethtool/strset.c +11 -0
index 3f7de54d85fb..959a1b61e8e7 100644
@@ -2,6 +2,7 @@
 
 #include <linux/ethtool.h>
 #include <linux/phy.h>
+#include <net/ulp_ddp_caps.h>
 #include "netlink.h"
 #include "common.h"
 
@@ -105,6 +106,16 @@ static const struct strset_info info_template[] = {
 		.count		= __ETHTOOL_A_STATS_RMON_CNT,
 		.strings	= stats_rmon_names,
 	},
+	[ETH_SS_ULP_DDP_CAPS] = {
+		.per_dev	= false,
+		.count		= ULP_DDP_C_COUNT,
+		.strings	= ulp_ddp_caps_names,
+	},
+	[ETH_SS_ULP_DDP_STATS] = {
+		.per_dev	= false,
+		.count		= __ETHTOOL_A_ULP_DDP_STATS_CNT,
+		.strings	= ulp_ddp_stats_names,
+	},
 };
 
 struct strset_req_info {
Signature · 1 line
2.31.1
#4 lore raw

PATCH v11 3/25 net/ethtool: add ULP_DDP_{GET,SET} operations for caps and stats

In reply to Aurelien Aptel

This commit adds:

- 3 new netlink messages:
* ULP_DDP_GET: returns a bitset of supported and active capabilities
* ULP_DDP_SET: tries to activate requested bitset and returns results
* ULP_DDP_NTF: notification for capabilities change

Rename and export bitset_policy for use in ulp_ddp.c.

ULP DDP capabilities handling is similar to netdev features handling.

If a ULP_DDP_GET message has requested statistics via the ETHTOOL_FLAG_STATS header flag, then statistics are returned to userspace.

ULP_DDP_GET request: (header only)
ULP_DDP_GET reply:

      HW             (bitset)
      ACTIVE         (bitset)
      STATS          (nest, optional)
          STATS_xxxx (u64)
          ....

ULP_DDP_SET request:
WANTED (bitset)
ULP_DDP_SET reply:
WANTED (bitset)
ACTIVE (bitset)

---
 include/uapi/linux/ethtool_netlink.h |  18 ++
 net/ethtool/Makefile                 |   2 +-
 net/ethtool/bitset.c                 |  20 +-
 net/ethtool/netlink.c                |  20 ++
 net/ethtool/netlink.h                |   4 +
 net/ethtool/ulp_ddp.c                | 316 +++++++++++++++++++++++++++
 6 files changed, 369 insertions(+), 11 deletions(-)
 create mode 100644 net/ethtool/ulp_ddp.c
include/uapi/linux/ethtool_netlink.h +18 -0
index 1722581b2db6..f6bac0ae15f6 100644
@@ -57,6 +57,8 @@ enum {
 	ETHTOOL_MSG_PLCA_GET_STATUS,
 	ETHTOOL_MSG_MM_GET,
 	ETHTOOL_MSG_MM_SET,
+	ETHTOOL_MSG_ULP_DDP_GET,
+	ETHTOOL_MSG_ULP_DDP_SET,
 
 	/* add new constants above here */
 	__ETHTOOL_MSG_USER_CNT,
@@ -109,6 +111,9 @@ enum {
 	ETHTOOL_MSG_PLCA_NTF,
 	ETHTOOL_MSG_MM_GET_REPLY,
 	ETHTOOL_MSG_MM_NTF,
+	ETHTOOL_MSG_ULP_DDP_GET_REPLY,
+	ETHTOOL_MSG_ULP_DDP_SET_REPLY,
+	ETHTOOL_MSG_ULP_DDP_NTF,
 
 	/* add new constants above here */
 	__ETHTOOL_MSG_KERNEL_CNT,
@@ -974,6 +979,19 @@ enum {
 
 /* ULP DDP */
 
+enum {
+	ETHTOOL_A_ULP_DDP_UNSPEC,
+	ETHTOOL_A_ULP_DDP_HEADER,			/* nest - _A_HEADER_* */
+	ETHTOOL_A_ULP_DDP_HW,				/* bitset */
+	ETHTOOL_A_ULP_DDP_ACTIVE,			/* bitset */
+	ETHTOOL_A_ULP_DDP_WANTED,			/* bitset */
+	ETHTOOL_A_ULP_DDP_STATS,			/* nest - _A_ULP_DDP_STATS_* */
+
+	/* add new constants above here */
+	__ETHTOOL_A_ULP_DDP_CNT,
+	ETHTOOL_A_ULP_DDP_MAX = __ETHTOOL_A_ULP_DDP_CNT - 1
+};
+
 enum {
 	ETHTOOL_A_ULP_DDP_STATS_UNSPEC,
 	ETHTOOL_A_ULP_DDP_STATS_PAD,
net/ethtool/Makefile +1 -1
index 504f954a1b28..a2fdc5ed7655 100644
@@ -8,4 +8,4 @@ ethtool_nl-y	:= netlink.o bitset.o strset.o linkinfo.o linkmodes.o rss.o \
 		   linkstate.o debug.o wol.o features.o privflags.o rings.o \
 		   channels.o coalesce.o pause.o eee.o tsinfo.o cabletest.o \
 		   tunnels.o fec.o eeprom.o stats.o phc_vclocks.o mm.o \
-		   module.o pse-pd.o plca.o mm.o
+		   module.o pse-pd.o plca.o ulp_ddp.o
net/ethtool/bitset.c +10 -10
index 0515d6604b3b..1bef91fcce4b 100644
@@ -302,7 +302,7 @@ int ethnl_put_bitset32(struct sk_buff *skb, int attrtype, const u32 *val,
 	return -EMSGSIZE;
 }
 
-static const struct nla_policy bitset_policy[] = {
+const struct nla_policy ethnl_bitset_policy[] = {
 	[ETHTOOL_A_BITSET_NOMASK]	= { .type = NLA_FLAG },
 	[ETHTOOL_A_BITSET_SIZE]		= NLA_POLICY_MAX(NLA_U32,
 							 ETHNL_MAX_BITSET_SIZE),
@@ -327,11 +327,11 @@ static const struct nla_policy bit_policy[] = {
  */
 int ethnl_bitset_is_compact(const struct nlattr *bitset, bool *compact)
 {
-	struct nlattr *tb[ARRAY_SIZE(bitset_policy)];
+	struct nlattr *tb[ARRAY_SIZE(ethnl_bitset_policy)];
 	int ret;
 
-	ret = nla_parse_nested(tb, ARRAY_SIZE(bitset_policy) - 1, bitset,
-			       bitset_policy, NULL);
+	ret = nla_parse_nested(tb, ARRAY_SIZE(ethnl_bitset_policy) - 1, bitset,
+			       ethnl_bitset_policy, NULL);
 	if (ret < 0)
 		return ret;
 
@@ -553,15 +553,15 @@ int ethnl_update_bitset32(u32 *bitmap, unsigned int nbits,
 			  const struct nlattr *attr, ethnl_string_array_t names,
 			  struct netlink_ext_ack *extack, bool *mod)
 {
-	struct nlattr *tb[ARRAY_SIZE(bitset_policy)];
+	struct nlattr *tb[ARRAY_SIZE(ethnl_bitset_policy)];
 	unsigned int change_bits;
 	bool no_mask;
 	int ret;
 
 	if (!attr)
 		return 0;
-	ret = nla_parse_nested(tb, ARRAY_SIZE(bitset_policy) - 1, attr,
-			       bitset_policy, extack);
+	ret = nla_parse_nested(tb, ARRAY_SIZE(ethnl_bitset_policy) - 1, attr,
+			       ethnl_bitset_policy, extack);
 	if (ret < 0)
 		return ret;
 
@@ -606,7 +606,7 @@ int ethnl_parse_bitset(unsigned long *val, unsigned long *mask,
 		       ethnl_string_array_t names,
 		       struct netlink_ext_ack *extack)
 {
-	struct nlattr *tb[ARRAY_SIZE(bitset_policy)];
+	struct nlattr *tb[ARRAY_SIZE(ethnl_bitset_policy)];
 	const struct nlattr *bit_attr;
 	bool no_mask;
 	int rem;
@@ -614,8 +614,8 @@ int ethnl_parse_bitset(unsigned long *val, unsigned long *mask,
 
 	if (!attr)
 		return 0;
-	ret = nla_parse_nested(tb, ARRAY_SIZE(bitset_policy) - 1, attr,
-			       bitset_policy, extack);
+	ret = nla_parse_nested(tb, ARRAY_SIZE(ethnl_bitset_policy) - 1, attr,
+			       ethnl_bitset_policy, extack);
 	if (ret < 0)
 		return ret;
 	no_mask = tb[ETHTOOL_A_BITSET_NOMASK];
net/ethtool/netlink.c +20 -0
index 08120095cc68..33f1b0992383 100644
@@ -306,6 +306,8 @@ ethnl_default_requests[__ETHTOOL_MSG_USER_CNT] = {
 	[ETHTOOL_MSG_PLCA_GET_STATUS]	= &ethnl_plca_status_request_ops,
 	[ETHTOOL_MSG_MM_GET]		= &ethnl_mm_request_ops,
 	[ETHTOOL_MSG_MM_SET]		= &ethnl_mm_request_ops,
+	[ETHTOOL_MSG_ULP_DDP_GET]	= &ethnl_ulp_ddp_request_ops,
+	[ETHTOOL_MSG_ULP_DDP_SET]	= &ethnl_ulp_ddp_request_ops,
 };
 
 static struct ethnl_dump_ctx *ethnl_dump_context(struct netlink_callback *cb)
@@ -669,6 +671,7 @@ ethnl_default_notify_ops[ETHTOOL_MSG_KERNEL_MAX + 1] = {
 	[ETHTOOL_MSG_MODULE_NTF]	= &ethnl_module_request_ops,
 	[ETHTOOL_MSG_PLCA_NTF]		= &ethnl_plca_cfg_request_ops,
 	[ETHTOOL_MSG_MM_NTF]		= &ethnl_mm_request_ops,
+	[ETHTOOL_MSG_ULP_DDP_NTF]	= &ethnl_ulp_ddp_request_ops,
 };
 
 /* default notification handler */
@@ -764,6 +767,7 @@ static const ethnl_notify_handler_t ethnl_notify_handlers[] = {
 	[ETHTOOL_MSG_MODULE_NTF]	= ethnl_default_notify,
 	[ETHTOOL_MSG_PLCA_NTF]		= ethnl_default_notify,
 	[ETHTOOL_MSG_MM_NTF]		= ethnl_default_notify,
+	[ETHTOOL_MSG_ULP_DDP_NTF]	= ethnl_default_notify,
 };
 
 void ethtool_notify(struct net_device *dev, unsigned int cmd, const void *data)
@@ -1156,6 +1160,22 @@ static const struct genl_ops ethtool_genl_ops[] = {
 		.policy = ethnl_mm_set_policy,
 		.maxattr = ARRAY_SIZE(ethnl_mm_set_policy) - 1,
 	},
+	{
+		.cmd    = ETHTOOL_MSG_ULP_DDP_GET,
+		.doit   = ethnl_default_doit,
+		.start  = ethnl_default_start,
+		.dumpit = ethnl_default_dumpit,
+		.done   = ethnl_default_done,
+		.policy = ethnl_ulp_ddp_get_policy,
+		.maxattr = ARRAY_SIZE(ethnl_ulp_ddp_get_policy) - 1,
+	},
+	{
+		.cmd	= ETHTOOL_MSG_ULP_DDP_SET,
+		.flags	= GENL_UNS_ADMIN_PERM,
+		.doit	= ethnl_default_set_doit,
+		.policy = ethnl_ulp_ddp_set_policy,
+		.maxattr = ARRAY_SIZE(ethnl_ulp_ddp_set_policy) - 1,
+	},
 };
 
 static const struct genl_multicast_group ethtool_nl_mcgrps[] = {
net/ethtool/netlink.h +4 -0
index ae0732460e88..4852f567c0be 100644
@@ -395,10 +395,12 @@ extern const struct ethnl_request_ops ethnl_rss_request_ops;
 extern const struct ethnl_request_ops ethnl_plca_cfg_request_ops;
 extern const struct ethnl_request_ops ethnl_plca_status_request_ops;
 extern const struct ethnl_request_ops ethnl_mm_request_ops;
+extern const struct ethnl_request_ops ethnl_ulp_ddp_request_ops;
 
 extern const struct nla_policy ethnl_header_policy[ETHTOOL_A_HEADER_FLAGS + 1];
 extern const struct nla_policy ethnl_header_policy_stats[ETHTOOL_A_HEADER_FLAGS + 1];
 extern const struct nla_policy ethnl_strset_get_policy[ETHTOOL_A_STRSET_COUNTS_ONLY + 1];
+extern const struct nla_policy ethnl_bitset_policy[ETHTOOL_A_BITSET_MASK + 1];
 extern const struct nla_policy ethnl_linkinfo_get_policy[ETHTOOL_A_LINKINFO_HEADER + 1];
 extern const struct nla_policy ethnl_linkinfo_set_policy[ETHTOOL_A_LINKINFO_TP_MDIX_CTRL + 1];
 extern const struct nla_policy ethnl_linkmodes_get_policy[ETHTOOL_A_LINKMODES_HEADER + 1];
@@ -441,6 +443,8 @@ extern const struct nla_policy ethnl_plca_set_cfg_policy[ETHTOOL_A_PLCA_MAX + 1]
 extern const struct nla_policy ethnl_plca_get_status_policy[ETHTOOL_A_PLCA_HEADER + 1];
 extern const struct nla_policy ethnl_mm_get_policy[ETHTOOL_A_MM_HEADER + 1];
 extern const struct nla_policy ethnl_mm_set_policy[ETHTOOL_A_MM_MAX + 1];
+extern const struct nla_policy ethnl_ulp_ddp_get_policy[ETHTOOL_A_ULP_DDP_HEADER + 1];
+extern const struct nla_policy ethnl_ulp_ddp_set_policy[ETHTOOL_A_ULP_DDP_WANTED + 1];
 
 int ethnl_set_features(struct sk_buff *skb, struct genl_info *info);
 int ethnl_act_cable_test(struct sk_buff *skb, struct genl_info *info);
net/ethtool/ulp_ddp.c new file +316 -0

316 lines not on this page · raw message

Signature · 1 line
2.31.1
#6 lore raw

PATCH v11 4/25 Documentation: document netlink ULP_DDP_GET/SET messages

In reply to Aurelien Aptel

Add detailed documentation about:
- ETHTOOL_MSG_ULP_DDP_GET and ETHTOOL_MSG_ULP_DDP_SET netlink messages
- ETH_SS_ULP_DDP_CAPS and ETH_SS_ULP_DDP_STATS stringsets

ETHTOOL_MSG_ULP_DDP_GET/SET messages are used to configure ULP DDP capabilities and retrieve ULP DDP statistics.

Both statistics and capabilities names can be retrieved dynamically from the kernel via string sets (no need to hardcode them and keep them in sync in ethtool).

---
 Documentation/netlink/specs/ethtool.yaml     | 102 +++++++++++++++++++
 Documentation/networking/ethtool-netlink.rst |  92 +++++++++++++++++
 Documentation/networking/statistics.rst      |   1 +
 3 files changed, 195 insertions(+)
Documentation/netlink/specs/ethtool.yaml +102 -0
index 82f4e6f8ddd3..8932cf310f1a 100644
@@ -236,6 +236,67 @@ attribute-sets:
         name: stats
         type: nest
         nested-attributes: mm-stat
+  -
+    name: ulp-ddp-stat
+    attributes:
+      -
+        name: pad
+        value: 1
+        type: pad
+      -
+        name: rx-nvmeotcp-sk-add
+        type: u64
+      -
+        name: rx-nvmeotcp-sk-add-fail
+        type: u64
+      -
+        name: rx-nvmeotcp-sk-del
+        type: u64
+      -
+        name: rx-nvmeotcp-ddp-setup
+        type: u64
+      -
+        name: rx-nvmeotcp-ddp-setup-fail
+        type: u64
+      -
+        name: rx-nvmeotcp-ddp-teardown
+        type: u64
+      -
+        name: rx-nvmeotcp-ddp-drop
+        type: u64
+      -
+        name: rx-nvmeotcp-ddp-resync
+        type: u64
+      -
+        name: rx-nvmeotcp-ddp-packets
+        type: u64
+      -
+        name: rx-nvmeotcp-ddp-bytes
+        type: u64
+  -
+    name: ulp-ddp
+    attributes:
+      -
+        name: header
+        value: 1
+        type: nest
+        nested-attributes: header
+      -
+        name: hw
+        type: nest
+        nested-attributes: bitset
+      -
+        name: active
+        type: nest
+        nested-attributes: bitset
+      -
+        name: wanted
+        type: nest
+        nested-attributes: bitset
+      -
+        name: stats
+        type: nest
+        nested-attributes: ulp-ddp-stat
 
 operations:
   enum-model: directional
@@ -390,3 +451,44 @@ operations:
       name: mm-ntf
       doc: Notification for change in MAC Merge configuration.
       notify: mm-get
+    -
+      name: ulp-ddp-get
+      doc: Get ULP DDP capabilities and stats
+
+      attribute-set: ulp-ddp
+
+      do: &ulp-ddp-get-op
+        request:
+          value: 44
+          attributes:
+            - header
+        reply:
+          value: 44
+          attributes:
+            - header
+            - hw
+            - active
+            - stats
+      dump: *ulp-ddp-get-op
+    -
+      name: ulp-ddp-set
+      doc: Set ULP DDP capabilities
+
+      attribute-set: ulp-ddp
+
+      do:
+        request:
+          value: 45
+          attributes:
+            - header
+            - wanted
+        reply:
+          value: 45
+          attributes:
+            - header
+            - hw
+            - active
+    -
+      name: ulp-ddp-ntf
+      doc: Notification for change in ULP DDP capabilities
+      notify: ulp-ddp-get
Documentation/networking/ethtool-netlink.rst +92 -0

110 lines not on this page · raw message

Documentation/networking/statistics.rst +1 -0
index 551b3cc29a41..9997c5e8d34e 100644
@@ -172,6 +172,7 @@ statistics are supported in the following commands:
   - `ETHTOOL_MSG_PAUSE_GET`
   - `ETHTOOL_MSG_FEC_GET`
   - `ETHTOOL_MSG_MM_GET`
+  - `ETHTOOL_MSG_ULP_DDP_GET`
 
 debugfs
 -------
Signature · 1 line
2.31.1
#9 lore raw

PATCH v11 5/25 iov_iter: skip copy if src == dst for direct data placement

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

When using direct data placement (DDP) the NIC could write the payload directly into the destination buffer and constructs SKBs such that they point to this data. To skip copies when SKB data already resides in the destination buffer we check if (src == dst), and skip the copy when it's true.

---
 lib/iov_iter.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
lib/iov_iter.c +7 -1
index f9a3ff37ecd1..2df634bb6d27 100644
@@ -526,9 +526,15 @@ size_t _copy_to_iter(const void *addr, size_t bytes, struct iov_iter *i)
 		return copy_pipe_to_iter(addr, bytes, i);
 	if (user_backed_iter(i))
 		might_fault();
+	/*
+	 * When using direct data placement (DDP) the hardware writes
+	 * data directly to the destination buffer, and constructs
+	 * IOVs such that they point to this data.
+	 * Thus, when the src == dst we skip the memcpy.
+	 */
 	iterate_and_advance(i, bytes, base, len, off,
 		copyout(base, addr + off, len),
-		memcpy(base, addr + off, len)
+		(base != addr + off) && memcpy(base, addr + off, len)
 	)
 
 	return bytes;
Signature · 1 line
2.31.1
#10 lore raw

PATCH v11 6/25 net/tls,core: export get_netdev_for_sock

In reply to Aurelien Aptel

* remove netdev_sk_get_lowest_dev() from net/core
* move get_netdev_for_sock() from net/tls to net/core

get_netdev_for_sock() is a utility that is used to obtain the net_device structure from a connected socket.

Later patches will use this for nvme-tcp DDP and DDP DDGST offloads.

---
 include/linux/netdevice.h |  3 +--
 net/core/dev.c            | 26 +++++++++++++-------------
 net/tls/tls_device.c      | 16 ----------------
 3 files changed, 14 insertions(+), 31 deletions(-)
include/linux/netdevice.h +1 -2
index 8861a2f04e00..25a868281594 100644
@@ -3096,8 +3096,7 @@ int init_dummy_netdev(struct net_device *dev);
 struct net_device *netdev_get_xmit_slave(struct net_device *dev,
 					 struct sk_buff *skb,
 					 bool all_slaves);
-struct net_device *netdev_sk_get_lowest_dev(struct net_device *dev,
-					    struct sock *sk);
+struct net_device *get_netdev_for_sock(struct sock *sk);
 struct net_device *dev_get_by_index(struct net *net, int ifindex);
 struct net_device *__dev_get_by_index(struct net *net, int ifindex);
 struct net_device *dev_get_by_index_rcu(struct net *net, int ifindex);
net/core/dev.c +13 -13
index bb42150a38ec..2360b70ea0b1 100644
@@ -8153,27 +8153,27 @@ static struct net_device *netdev_sk_get_lower_dev(struct net_device *dev,
 }
 
 /**
- * netdev_sk_get_lowest_dev - Get the lowest device in chain given device and socket
- * @dev: device
+ * get_netdev_for_sock - Get the lowest device in socket
  * @sk: the socket
  *
- * %NULL is returned if no lower device is found.
+ * Assumes that the socket is already connected.
+ * Returns the lower device or %NULL if no lower device is found.
  */
-
-struct net_device *netdev_sk_get_lowest_dev(struct net_device *dev,
-					    struct sock *sk)
+struct net_device *get_netdev_for_sock(struct sock *sk)
 {
-	struct net_device *lower;
+	struct dst_entry *dst = sk_dst_get(sk);
+	struct net_device *dev, *lower;
 
-	lower = netdev_sk_get_lower_dev(dev, sk);
-	while (lower) {
+	if (unlikely(!dst))
+		return NULL;
+	dev = dst->dev;
+	while ((lower = netdev_sk_get_lower_dev(dev, sk)))
 		dev = lower;
-		lower = netdev_sk_get_lower_dev(dev, sk);
-	}
-
+	dev_hold(dev);
+	dst_release(dst);
 	return dev;
 }
-EXPORT_SYMBOL(netdev_sk_get_lowest_dev);
+EXPORT_SYMBOL_GPL(get_netdev_for_sock);
 
 static void netdev_adjacent_add_links(struct net_device *dev)
 {
net/tls/tls_device.c +0 -16
index 6c593788dc25..3c298dfb77cb 100644
@@ -120,22 +120,6 @@ static void tls_device_queue_ctx_destruction(struct tls_context *ctx)
 		tls_device_free_ctx(ctx);
 }
 
-/* We assume that the socket is already connected */
-static struct net_device *get_netdev_for_sock(struct sock *sk)
-{
-	struct dst_entry *dst = sk_dst_get(sk);
-	struct net_device *netdev = NULL;
-
-	if (likely(dst)) {
-		netdev = netdev_sk_get_lowest_dev(dst->dev, sk);
-		dev_hold(netdev);
-	}
-
-	dst_release(dst);
-
-	return netdev;
-}
-
 static void destroy_record(struct tls_record_info *record)
 {
 	int i;
Signature · 1 line
2.31.1
#11 lore raw

PATCH v11 7/25 nvme-tcp: Add DDP offload control path

In reply to Aurelien Aptel

From: Boris Pismenny <[email protected]>

This commit introduces direct data placement offload to NVME TCP. There is a context per queue, which is established after the handshake using the sk_add/del NDOs.

Additionally, a resynchronization routine is used to assist hardware recovery from TCP OOO, and continue the offload. Resynchronization operates as follows:

1. TCP OOO causes the NIC HW to stop the offload

2. NIC HW identifies a PDU header at some TCP sequence number, and asks NVMe-TCP to confirm it.
This request is delivered from the NIC driver to NVMe-TCP by first finding the socket for the packet that triggered the request, and then finding the nvme_tcp_queue that is used by this routine. Finally, the request is recorded in the nvme_tcp_queue.

3. When NVMe-TCP observes the requested TCP sequence, it will compare it with the PDU header TCP sequence, and report the result to the NIC driver (resync), which will update the HW, and resume offload when all is successful.

Some HW implementation such as ConnectX-7 assume linear CCID (0...N-1 for queue of size N) where the linux nvme driver uses part of the 16 bit CCID for generation counter. To address that, we use the existing quirk in the nvme layer when the HW driver advertises if the device is not supports the full 16 bit CCID range.

Furthermore, we let the offloading driver advertise what is the max hw sectors/segments via ulp_ddp_limits.

A follow-up patch introduces the data-path changes required for this offload.

Socket operations need a netdev reference. This reference is dropped on NETDEV_GOING_DOWN events to allow the device to go down in a follow-up patch.

---
 drivers/nvme/host/tcp.c | 263 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 254 insertions(+), 9 deletions(-)
drivers/nvme/host/tcp.c +254 -9
index 70e273b565de..b350250bf8fb 100644
@@ -16,6 +16,10 @@
 #include <net/busy_poll.h>
 #include <trace/events/sock.h>
 
+#ifdef CONFIG_ULP_DDP
+#include <net/ulp_ddp.h>
+#endif
+
 #include "nvme.h"
 #include "fabrics.h"
 
@@ -104,6 +108,7 @@ enum nvme_tcp_queue_flags {
 	NVME_TCP_Q_ALLOCATED	= 0,
 	NVME_TCP_Q_LIVE		= 1,
 	NVME_TCP_Q_POLLING	= 2,
+	NVME_TCP_Q_OFF_DDP	= 3,
 };
 
 enum nvme_tcp_recv_state {
@@ -131,6 +136,16 @@ struct nvme_tcp_queue {
 	size_t			ddgst_remaining;
 	unsigned int		nr_cqe;
 
+	/*
+	 * resync_req is a speculative PDU header tcp seq number (with
+	 * an additional flag at 32 lower bits) that the HW send to
+	 * the SW, for the SW to verify.
+	 * - The 32 high bits store the seq number
+	 * - The 32 low bits are used as a flag to know if a request
+	 *   is pending (ULP_DDP_RESYNC_PENDING).
+	 */
+	atomic64_t		resync_req;
+
 	/* send state */
 	struct nvme_tcp_request *request;
 
@@ -170,6 +185,9 @@ struct nvme_tcp_ctrl {
 	struct delayed_work	connect_work;
 	struct nvme_tcp_request async_req;
 	u32			io_queues[HCTX_MAX_TYPES];
+
+	struct net_device	*offloading_netdev;
+	u32			offload_io_threshold;
 };
 
 static LIST_HEAD(nvme_tcp_ctrl_list);
@@ -261,6 +279,198 @@ static inline size_t nvme_tcp_pdu_last_send(struct nvme_tcp_request *req,
 	return nvme_tcp_pdu_data_left(req) <= len;
 }
 
+#ifdef CONFIG_ULP_DDP
+
+static inline bool is_netdev_ulp_offload_active(struct net_device *netdev)
+{
+	return test_bit(ULP_DDP_C_NVME_TCP_BIT, netdev->ulp_ddp_caps.active);
+}
+
+static bool nvme_tcp_ddp_query_limits(struct net_device *netdev,
+				      struct ulp_ddp_limits *limits)
+{
+	int ret;
+
+	if (!netdev || !is_netdev_ulp_offload_active(netdev) ||
+	    !netdev->netdev_ops->ulp_ddp_ops->limits)
+		return false;
+
+	limits->type = ULP_DDP_NVME;
+	ret = netdev->netdev_ops->ulp_ddp_ops->limits(netdev, limits);
+	if (ret == -EOPNOTSUPP) {
+		return false;
+	} else if (ret) {
+		WARN_ONCE(ret, "ddp limits failed (ret=%d)", ret);
+		return false;
+	}
+
+	return true;
+}
+
+static bool nvme_tcp_resync_request(struct sock *sk, u32 seq, u32 flags);
+static const struct ulp_ddp_ulp_ops nvme_tcp_ddp_ulp_ops = {
+	.resync_request		= nvme_tcp_resync_request,
+};
+
+static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue)
+{
+	struct net_device *netdev = queue->ctrl->offloading_netdev;
+	struct ulp_ddp_config config = {.type = ULP_DDP_NVME};
+	int ret;
+
+	config.nvmeotcp.pfv = NVME_TCP_PFV_1_0;
+	config.nvmeotcp.cpda = 0;
+	config.nvmeotcp.dgst =
+		queue->hdr_digest ? NVME_TCP_HDR_DIGEST_ENABLE : 0;
+	config.nvmeotcp.dgst |=
+		queue->data_digest ? NVME_TCP_DATA_DIGEST_ENABLE : 0;
+	config.nvmeotcp.queue_size = queue->ctrl->ctrl.sqsize + 1;
+	config.nvmeotcp.queue_id = nvme_tcp_queue_id(queue);
+	config.nvmeotcp.io_cpu = queue->io_cpu;
+
+	/* Socket ops keep a netdev reference. It is put in
+	 * nvme_tcp_unoffload_socket().  This ref is dropped on
+	 * NETDEV_GOING_DOWN events to allow the device to go down
+	 */
+	dev_hold(netdev);
+	ret = netdev->netdev_ops->ulp_ddp_ops->sk_add(netdev,
+						      queue->sock->sk,
+						      &config);
+	if (ret) {
+		dev_put(netdev);
+		return ret;
+	}
+
+	inet_csk(queue->sock->sk)->icsk_ulp_ddp_ops = &nvme_tcp_ddp_ulp_ops;
+	set_bit(NVME_TCP_Q_OFF_DDP, &queue->flags);
+	return 0;
+}
+
+static void nvme_tcp_unoffload_socket(struct nvme_tcp_queue *queue)
+{
+	struct net_device *netdev = queue->ctrl->offloading_netdev;
+
+	if (!netdev) {
+		dev_info_ratelimited(queue->ctrl->ctrl.device, "netdev not found\n");
+		return;
+	}
+
+	clear_bit(NVME_TCP_Q_OFF_DDP, &queue->flags);
+
+	netdev->netdev_ops->ulp_ddp_ops->sk_del(netdev, queue->sock->sk);
+
+	inet_csk(queue->sock->sk)->icsk_ulp_ddp_ops = NULL;
+	dev_put(netdev); /* held by offload_socket */
+}
+
+static void nvme_tcp_offload_limits(struct nvme_tcp_queue *queue,
+				    struct net_device *netdev)
+{
+	struct ulp_ddp_limits limits = {.type = ULP_DDP_NVME };
+
+	if (!nvme_tcp_ddp_query_limits(netdev, &limits)) {
+		queue->ctrl->offloading_netdev = NULL;
+		return;
+	}
+
+	queue->ctrl->offloading_netdev = netdev;
+	dev_dbg_ratelimited(queue->ctrl->ctrl.device,
+			    "netdev %s offload limits: max_ddp_sgl_len %d\n",
+			    netdev->name, limits.max_ddp_sgl_len);
+	queue->ctrl->ctrl.max_segments = limits.max_ddp_sgl_len;
+	queue->ctrl->ctrl.max_hw_sectors =
+		limits.max_ddp_sgl_len << (ilog2(SZ_4K) - 9);
+	queue->ctrl->offload_io_threshold = limits.io_threshold;
+
+	/* offloading HW doesn't support full ccid range, apply the quirk */
+	queue->ctrl->ctrl.quirks |=
+		limits.nvmeotcp.full_ccid_range ? 0 : NVME_QUIRK_SKIP_CID_GEN;
+}
+
+/* In presence of packet drops or network packet reordering, the device may lose
+ * synchronization between the TCP stream and the L5P framing, and require a
+ * resync with the kernel's TCP stack.
+ *
+ * - NIC HW identifies a PDU header at some TCP sequence number,
+ *   and asks NVMe-TCP to confirm it.
+ * - When NVMe-TCP observes the requested TCP sequence, it will compare
+ *   it with the PDU header TCP sequence, and report the result to the
+ *   NIC driver
+ */
+static void nvme_tcp_resync_response(struct nvme_tcp_queue *queue,
+				     struct sk_buff *skb, unsigned int offset)
+{
+	u64 pdu_seq = TCP_SKB_CB(skb)->seq + offset - queue->pdu_offset;
+	struct net_device *netdev = queue->ctrl->offloading_netdev;
+	u64 pdu_val = (pdu_seq << 32) | ULP_DDP_RESYNC_PENDING;
+	u64 resync_val;
+	u32 resync_seq;
+
+	resync_val = atomic64_read(&queue->resync_req);
+	/* Lower 32 bit flags. Check validity of the request */
+	if ((resync_val & ULP_DDP_RESYNC_PENDING) == 0)
+		return;
+
+	/*
+	 * Obtain and check requested sequence number: is this PDU header
+	 * before the request?
+	 */
+	resync_seq = resync_val >> 32;
+	if (before(pdu_seq, resync_seq))
+		return;
+
+	/*
+	 * The atomic operation guarantees that we don't miss any NIC driver
+	 * resync requests submitted after the above checks.
+	 */
+	if (atomic64_cmpxchg(&queue->resync_req, pdu_val,
+			     pdu_val & ~ULP_DDP_RESYNC_PENDING) !=
+			     atomic64_read(&queue->resync_req))
+		netdev->netdev_ops->ulp_ddp_ops->resync(netdev,
+							queue->sock->sk,
+							pdu_seq);
+}
+
+static bool nvme_tcp_resync_request(struct sock *sk, u32 seq, u32 flags)
+{
+	struct nvme_tcp_queue *queue = sk->sk_user_data;
+
+	/*
+	 * "seq" (TCP seq number) is what the HW assumes is the
+	 * beginning of a PDU.  The nvme-tcp layer needs to store the
+	 * number along with the "flags" (ULP_DDP_RESYNC_PENDING) to
+	 * indicate that a request is pending.
+	 */
+	atomic64_set(&queue->resync_req, (((uint64_t)seq << 32) | flags));
+
+	return true;
+}
+
+#else
+
+static inline bool is_netdev_ulp_offload_active(struct net_device *netdev)
+{
+	return false;
+}
+
+static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue)
+{
+	return 0;
+}
+
+static void nvme_tcp_unoffload_socket(struct nvme_tcp_queue *queue)
+{}
+
+static void nvme_tcp_offload_limits(struct nvme_tcp_queue *queue,
+				    struct net_device *netdev)
+{}
+
+static void nvme_tcp_resync_response(struct nvme_tcp_queue *queue,
+				     struct sk_buff *skb, unsigned int offset)
+{}
+
+#endif
+
 static void nvme_tcp_init_iter(struct nvme_tcp_request *req,
 		unsigned int dir)
 {
@@ -703,6 +913,9 @@ static int nvme_tcp_recv_pdu(struct nvme_tcp_queue *queue, struct sk_buff *skb,
 	size_t rcv_len = min_t(size_t, *len, queue->pdu_remaining);
 	int ret;
 
+	if (test_bit(NVME_TCP_Q_OFF_DDP, &queue->flags))
+		nvme_tcp_resync_response(queue, skb, *offset);
+
 	ret = skb_copy_bits(skb, *offset,
 		&pdu[queue->pdu_offset], rcv_len);
 	if (unlikely(ret))
@@ -1660,6 +1873,8 @@ static void __nvme_tcp_stop_queue(struct nvme_tcp_queue *queue)
 	kernel_sock_shutdown(queue->sock, SHUT_RDWR);
 	nvme_tcp_restore_sock_calls(queue);
 	cancel_work_sync(&queue->io_work);
+	if (test_bit(NVME_TCP_Q_OFF_DDP, &queue->flags))
+		nvme_tcp_unoffload_socket(queue);
 }
 
 static void nvme_tcp_stop_queue(struct nvme_ctrl *nctrl, int qid)
@@ -1679,21 +1894,51 @@ static void nvme_tcp_stop_queue(struct nvme_ctrl *nctrl, int qid)
 static int nvme_tcp_start_queue(struct nvme_ctrl *nctrl, int idx)
 {
 	struct nvme_tcp_ctrl *ctrl = to_tcp_ctrl(nctrl);
+	struct net_device *netdev;
 	int ret;
 
-	if (idx)
+	if (idx) {
 		ret = nvmf_connect_io_queue(nctrl, idx);
-	else
+		if (ret)
+			goto err;
+
+		netdev = ctrl->queues[idx].ctrl->offloading_netdev;
+		if (netdev && is_netdev_ulp_offload_active(netdev)) {
+			ret = nvme_tcp_offload_socket(&ctrl->queues[idx]);
+			if (ret) {
+				dev_err(nctrl->device,
+					"failed to setup offload on queue %d ret=%d\n",
+					idx, ret);
+			}
+		}
+	} else {
 		ret = nvmf_connect_admin_queue(nctrl);
+		if (ret)
+			goto err;
 
-	if (!ret) {
-		set_bit(NVME_TCP_Q_LIVE, &ctrl->queues[idx].flags);
-	} else {
-		if (test_bit(NVME_TCP_Q_ALLOCATED, &ctrl->queues[idx].flags))
-			__nvme_tcp_stop_queue(&ctrl->queues[idx]);
-		dev_err(nctrl->device,
-			"failed to connect queue: %d ret=%d\n", idx, ret);
+		netdev = get_netdev_for_sock(ctrl->queues[idx].sock->sk);
+		if (!netdev) {
+			dev_info_ratelimited(ctrl->ctrl.device, "netdev not found\n");
+			ctrl->offloading_netdev = NULL;
+			goto done;
+		}
+		if (is_netdev_ulp_offload_active(netdev))
+			nvme_tcp_offload_limits(&ctrl->queues[idx], netdev);
+		/*
+		 * release the device as no offload context is
+		 * established yet.
+		 */
+		dev_put(netdev);
 	}
+
+done:
+	set_bit(NVME_TCP_Q_LIVE, &ctrl->queues[idx].flags);
+	return 0;
+err:
+	if (test_bit(NVME_TCP_Q_ALLOCATED, &ctrl->queues[idx].flags))
+		__nvme_tcp_stop_queue(&ctrl->queues[idx]);
+	dev_err(nctrl->device,
+		"failed to connect queue: %d ret=%d\n", idx, ret);
 	return ret;
 }
 
Signature · 1 line
2.31.1
#13 lore raw

PATCH v11 9/25 nvme-tcp: RX DDGST offload

In reply to Aurelien Aptel

From: Yoray Zack <[email protected]>

Enable rx side of DDGST offload when supported.

At the end of the capsule, check if all the skb bits are on, and if not recalculate the DDGST in SW and check it.

---
 drivers/nvme/host/tcp.c | 142 ++++++++++++++++++++++++++++++++++++----
 1 file changed, 128 insertions(+), 14 deletions(-)
drivers/nvme/host/tcp.c +128 -14
index 533177971777..7e3feb694e46 100644
@@ -116,6 +116,7 @@ enum nvme_tcp_queue_flags {
 	NVME_TCP_Q_LIVE		= 1,
 	NVME_TCP_Q_POLLING	= 2,
 	NVME_TCP_Q_OFF_DDP	= 3,
+	NVME_TCP_Q_OFF_DDGST_RX = 4,
 };
 
 enum nvme_tcp_recv_state {
@@ -143,6 +144,9 @@ struct nvme_tcp_queue {
 	size_t			ddgst_remaining;
 	unsigned int		nr_cqe;
 
+#ifdef CONFIG_ULP_DDP
+	bool			ddp_ddgst_valid;
+
 	/*
 	 * resync_req is a speculative PDU header tcp seq number (with
 	 * an additional flag at 32 lower bits) that the HW send to
@@ -152,6 +156,7 @@ struct nvme_tcp_queue {
 	 *   is pending (ULP_DDP_RESYNC_PENDING).
 	 */
 	atomic64_t		resync_req;
+#endif
 
 	/* send state */
 	struct nvme_tcp_request *request;
@@ -288,9 +293,22 @@ static inline size_t nvme_tcp_pdu_last_send(struct nvme_tcp_request *req,
 
 #ifdef CONFIG_ULP_DDP
 
-static inline bool is_netdev_ulp_offload_active(struct net_device *netdev)
+static inline bool is_netdev_ulp_offload_active(struct net_device *netdev,
+						struct nvme_tcp_queue *queue)
 {
-	return test_bit(ULP_DDP_C_NVME_TCP_BIT, netdev->ulp_ddp_caps.active);
+	bool ddgst_offload;
+
+	if (test_bit(ULP_DDP_C_NVME_TCP_BIT, netdev->ulp_ddp_caps.active))
+		return true;
+
+	ddgst_offload = test_bit(ULP_DDP_C_NVME_TCP_DDGST_RX_BIT,
+				 netdev->ulp_ddp_caps.active);
+	if (!queue && ddgst_offload)
+		return true;
+	if (queue && queue->data_digest && ddgst_offload)
+		return true;
+
+	return false;
 }
 
 static bool nvme_tcp_ddp_query_limits(struct net_device *netdev,
@@ -298,7 +316,7 @@ static bool nvme_tcp_ddp_query_limits(struct net_device *netdev,
 {
 	int ret;
 
-	if (!netdev || !is_netdev_ulp_offload_active(netdev) ||
+	if (!netdev || !is_netdev_ulp_offload_active(netdev, NULL) ||
 	    !netdev->netdev_ops->ulp_ddp_ops->limits)
 		return false;
 
@@ -314,6 +332,18 @@ static bool nvme_tcp_ddp_query_limits(struct net_device *netdev,
 	return true;
 }
 
+static inline bool nvme_tcp_ddp_ddgst_ok(struct nvme_tcp_queue *queue)
+{
+	return queue->ddp_ddgst_valid;
+}
+
+static inline void nvme_tcp_ddp_ddgst_update(struct nvme_tcp_queue *queue,
+					     struct sk_buff *skb)
+{
+	if (queue->ddp_ddgst_valid)
+		queue->ddp_ddgst_valid = skb_is_ulp_crc(skb);
+}
+
 static int nvme_tcp_req_map_sg(struct nvme_tcp_request *req, struct request *rq)
 {
 	int ret;
@@ -328,6 +358,38 @@ static int nvme_tcp_req_map_sg(struct nvme_tcp_request *req, struct request *rq)
 	return 0;
 }
 
+static void nvme_tcp_ddp_ddgst_recalc(struct ahash_request *hash,
+				      struct request *rq,
+				      __le32 *ddgst)
+{
+	struct nvme_tcp_request *req;
+
+	if (!rq)
+		return;
+
+	req = blk_mq_rq_to_pdu(rq);
+
+	if (!req->offloaded) {
+		/* if we have DDGST_RX offload without DDP the request
+		 * wasn't mapped, so we need to map it here
+		 */
+		if (nvme_tcp_req_map_sg(req, rq))
+			return;
+	}
+
+	req->ddp.sg_table.sgl = req->ddp.first_sgl;
+	ahash_request_set_crypt(hash, req->ddp.sg_table.sgl, (u8 *)ddgst,
+				req->data_len);
+	crypto_ahash_digest(hash);
+
+	if (!req->offloaded) {
+		/* without DDP, ddp_teardown() won't be called, so
+		 * free the table here
+		 */
+		sg_free_table_chained(&req->ddp.sg_table, SG_CHUNK_SIZE);
+	}
+}
+
 static bool nvme_tcp_resync_request(struct sock *sk, u32 seq, u32 flags);
 static void nvme_tcp_ddp_teardown_done(void *ddp_ctx);
 static const struct ulp_ddp_ulp_ops nvme_tcp_ddp_ulp_ops = {
@@ -387,6 +449,10 @@ static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue)
 {
 	struct net_device *netdev = queue->ctrl->offloading_netdev;
 	struct ulp_ddp_config config = {.type = ULP_DDP_NVME};
+	bool offload_ddp = test_bit(ULP_DDP_C_NVME_TCP_BIT,
+				    netdev->ulp_ddp_caps.active);
+	bool offload_ddgst_rx = test_bit(ULP_DDP_C_NVME_TCP_DDGST_RX_BIT,
+					 netdev->ulp_ddp_caps.active);
 	int ret;
 
 	config.nvmeotcp.pfv = NVME_TCP_PFV_1_0;
@@ -413,7 +479,10 @@ static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue)
 	}
 
 	inet_csk(queue->sock->sk)->icsk_ulp_ddp_ops = &nvme_tcp_ddp_ulp_ops;
-	set_bit(NVME_TCP_Q_OFF_DDP, &queue->flags);
+	if (offload_ddp)
+		set_bit(NVME_TCP_Q_OFF_DDP, &queue->flags);
+	if (queue->data_digest && offload_ddgst_rx)
+		set_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags);
 	return 0;
 }
 
@@ -427,6 +496,7 @@ static void nvme_tcp_unoffload_socket(struct nvme_tcp_queue *queue)
 	}
 
 	clear_bit(NVME_TCP_Q_OFF_DDP, &queue->flags);
+	clear_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags);
 
 	netdev->netdev_ops->ulp_ddp_ops->sk_del(netdev, queue->sock->sk);
 
@@ -519,11 +589,26 @@ static bool nvme_tcp_resync_request(struct sock *sk, u32 seq, u32 flags)
 
 #else
 
-static inline bool is_netdev_ulp_offload_active(struct net_device *netdev)
+static inline bool is_netdev_ulp_offload_active(struct net_device *netdev,
+						struct nvme_tcp_queue *queue)
 {
 	return false;
 }
 
+static inline bool nvme_tcp_ddp_ddgst_ok(struct nvme_tcp_queue *queue)
+{
+	return true;
+}
+
+static inline void nvme_tcp_ddp_ddgst_update(struct nvme_tcp_queue *queue,
+					     struct sk_buff *skb)
+{}
+
+static void nvme_tcp_ddp_ddgst_recalc(struct ahash_request *hash,
+				      struct request *rq,
+				      __le32 *ddgst)
+{}
+
 static int nvme_tcp_setup_ddp(struct nvme_tcp_queue *queue, u16 command_id,
 			      struct request *rq)
 {
@@ -806,6 +891,9 @@ static void nvme_tcp_init_recv_ctx(struct nvme_tcp_queue *queue)
 	queue->pdu_offset = 0;
 	queue->data_remaining = -1;
 	queue->ddgst_remaining = 0;
+#ifdef CONFIG_ULP_DDP
+	queue->ddp_ddgst_valid = true;
+#endif
 }
 
 static void nvme_tcp_error_recovery(struct nvme_ctrl *ctrl)
@@ -1009,7 +1097,8 @@ static int nvme_tcp_recv_pdu(struct nvme_tcp_queue *queue, struct sk_buff *skb,
 	size_t rcv_len = min_t(size_t, *len, queue->pdu_remaining);
 	int ret;
 
-	if (test_bit(NVME_TCP_Q_OFF_DDP, &queue->flags))
+	if (test_bit(NVME_TCP_Q_OFF_DDP, &queue->flags) ||
+	    test_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags))
 		nvme_tcp_resync_response(queue, skb, *offset);
 
 	ret = skb_copy_bits(skb, *offset,
@@ -1073,6 +1162,10 @@ static int nvme_tcp_recv_data(struct nvme_tcp_queue *queue, struct sk_buff *skb,
 		nvme_cid_to_rq(nvme_tcp_tagset(queue), pdu->command_id);
 	struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
 
+	if (queue->data_digest &&
+	    test_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags))
+		nvme_tcp_ddp_ddgst_update(queue, skb);
+
 	while (true) {
 		int recv_len, ret;
 
@@ -1101,7 +1194,8 @@ static int nvme_tcp_recv_data(struct nvme_tcp_queue *queue, struct sk_buff *skb,
 		recv_len = min_t(size_t, recv_len,
 				iov_iter_count(&req->iter));
 
-		if (queue->data_digest)
+		if (queue->data_digest &&
+		    !test_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags))
 			ret = skb_copy_and_hash_datagram_iter(skb, *offset,
 				&req->iter, recv_len, queue->rcv_hash);
 		else
@@ -1143,8 +1237,11 @@ static int nvme_tcp_recv_ddgst(struct nvme_tcp_queue *queue,
 	char *ddgst = (char *)&queue->recv_ddgst;
 	size_t recv_len = min_t(size_t, *len, queue->ddgst_remaining);
 	off_t off = NVME_TCP_DIGEST_LENGTH - queue->ddgst_remaining;
+	struct request *rq;
 	int ret;
 
+	if (test_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags))
+		nvme_tcp_ddp_ddgst_update(queue, skb);
 	ret = skb_copy_bits(skb, *offset, &ddgst[off], recv_len);
 	if (unlikely(ret))
 		return ret;
@@ -1155,9 +1252,25 @@ static int nvme_tcp_recv_ddgst(struct nvme_tcp_queue *queue,
 	if (queue->ddgst_remaining)
 		return 0;
 
+	rq = nvme_cid_to_rq(nvme_tcp_tagset(queue),
+			    pdu->command_id);
+
+	if (test_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags)) {
+		/*
+		 * If HW successfully offloaded the digest
+		 * verification, we can skip it
+		 */
+		if (nvme_tcp_ddp_ddgst_ok(queue))
+			goto out;
+		/*
+		 * Otherwise we have to recalculate and verify the
+		 * digest with the software-fallback
+		 */
+		nvme_tcp_ddp_ddgst_recalc(queue->rcv_hash, rq,
+					  &queue->exp_ddgst);
+	}
+
 	if (queue->recv_ddgst != queue->exp_ddgst) {
-		struct request *rq = nvme_cid_to_rq(nvme_tcp_tagset(queue),
-					pdu->command_id);
 		struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
 
 		req->status = cpu_to_le16(NVME_SC_DATA_XFER_ERROR);
@@ -1168,9 +1281,8 @@ static int nvme_tcp_recv_ddgst(struct nvme_tcp_queue *queue,
 			le32_to_cpu(queue->exp_ddgst));
 	}
 
+out:
 	if (pdu->hdr.flags & NVME_TCP_F_DATA_SUCCESS) {
-		struct request *rq = nvme_cid_to_rq(nvme_tcp_tagset(queue),
-					pdu->command_id);
 		struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
 
 		nvme_tcp_end_request(rq, le16_to_cpu(req->status));
@@ -1981,7 +2093,8 @@ static void __nvme_tcp_stop_queue(struct nvme_tcp_queue *queue)
 	kernel_sock_shutdown(queue->sock, SHUT_RDWR);
 	nvme_tcp_restore_sock_calls(queue);
 	cancel_work_sync(&queue->io_work);
-	if (test_bit(NVME_TCP_Q_OFF_DDP, &queue->flags))
+	if (test_bit(NVME_TCP_Q_OFF_DDP, &queue->flags) ||
+	    test_bit(NVME_TCP_Q_OFF_DDGST_RX, &queue->flags))
 		nvme_tcp_unoffload_socket(queue);
 }
 
@@ -2011,7 +2124,8 @@ static int nvme_tcp_start_queue(struct nvme_ctrl *nctrl, int idx)
 			goto err;
 
 		netdev = ctrl->queues[idx].ctrl->offloading_netdev;
-		if (netdev && is_netdev_ulp_offload_active(netdev)) {
+		if (netdev &&
+		    is_netdev_ulp_offload_active(netdev, &ctrl->queues[idx])) {
 			ret = nvme_tcp_offload_socket(&ctrl->queues[idx]);
 			if (ret) {
 				dev_err(nctrl->device,
@@ -2030,7 +2144,7 @@ static int nvme_tcp_start_queue(struct nvme_ctrl *nctrl, int idx)
 			ctrl->offloading_netdev = NULL;
 			goto done;
 		}
-		if (is_netdev_ulp_offload_active(netdev))
+		if (is_netdev_ulp_offload_active(netdev, &ctrl->queues[idx]))
 			nvme_tcp_offload_limits(&ctrl->queues[idx], netdev);
 		/*
 		 * release the device as no offload context is
Signature · 1 line
2.31.1
#12 lore raw

PATCH v11 8/25 nvme-tcp: Add DDP data-path

In reply to Aurelien Aptel

From: Boris Pismenny <[email protected]>

Introduce the NVMe-TCP DDP data-path offload.
Using this interface, the NIC hardware will scatter TCP payload directly to the BIO pages according to the command_id in the PDU.
To maintain the correctness of the network stack, the driver is expected to construct SKBs that point to the BIO pages.

The data-path interface contains two routines: setup/teardown. The setup provides the mapping from command_id to the request buffers, while the teardown removes this mapping.

For efficiency, we introduce an asynchronous nvme completion, which is split between NVMe-TCP and the NIC driver as follows:
NVMe-TCP performs the specific completion, while NIC driver performs the generic mq_blk completion.

---
 drivers/nvme/host/tcp.c | 121 ++++++++++++++++++++++++++++++++++++++--
 1 file changed, 116 insertions(+), 5 deletions(-)
drivers/nvme/host/tcp.c +116 -5
index b350250bf8fb..533177971777 100644
@@ -102,6 +102,13 @@ struct nvme_tcp_request {
 	size_t			offset;
 	size_t			data_sent;
 	enum nvme_tcp_send_state state;
+
+#ifdef CONFIG_ULP_DDP
+	bool			offloaded;
+	struct ulp_ddp_io	ddp;
+	__le16			ddp_status;
+	union nvme_result	result;
+#endif
 };
 
 enum nvme_tcp_queue_flags {
@@ -307,11 +314,75 @@ static bool nvme_tcp_ddp_query_limits(struct net_device *netdev,
 	return true;
 }
 
+static int nvme_tcp_req_map_sg(struct nvme_tcp_request *req, struct request *rq)
+{
+	int ret;
+
+	req->ddp.sg_table.sgl = req->ddp.first_sgl;
+	ret = sg_alloc_table_chained(&req->ddp.sg_table,
+				     blk_rq_nr_phys_segments(rq),
+				     req->ddp.sg_table.sgl, SG_CHUNK_SIZE);
+	if (ret)
+		return -ENOMEM;
+	req->ddp.nents = blk_rq_map_sg(rq->q, rq, req->ddp.sg_table.sgl);
+	return 0;
+}
+
 static bool nvme_tcp_resync_request(struct sock *sk, u32 seq, u32 flags);
+static void nvme_tcp_ddp_teardown_done(void *ddp_ctx);
 static const struct ulp_ddp_ulp_ops nvme_tcp_ddp_ulp_ops = {
 	.resync_request		= nvme_tcp_resync_request,
+	.ddp_teardown_done	= nvme_tcp_ddp_teardown_done,
 };
 
+static void nvme_tcp_teardown_ddp(struct nvme_tcp_queue *queue,
+				  struct request *rq)
+{
+	struct net_device *netdev = queue->ctrl->offloading_netdev;
+	struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
+
+	netdev->netdev_ops->ulp_ddp_ops->teardown(netdev, queue->sock->sk,
+						  &req->ddp, rq);
+	sg_free_table_chained(&req->ddp.sg_table, SG_CHUNK_SIZE);
+}
+
+static void nvme_tcp_ddp_teardown_done(void *ddp_ctx)
+{
+	struct request *rq = ddp_ctx;
+	struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
+
+	if (!nvme_try_complete_req(rq, req->ddp_status, req->result))
+		nvme_complete_rq(rq);
+}
+
+static int nvme_tcp_setup_ddp(struct nvme_tcp_queue *queue, u16 command_id,
+			      struct request *rq)
+{
+	struct net_device *netdev = queue->ctrl->offloading_netdev;
+	struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
+	int ret;
+
+	if (!(rq_data_dir(rq) == READ) ||
+	    queue->ctrl->offload_io_threshold > blk_rq_payload_bytes(rq))
+		return 0;
+
+	req->ddp.command_id = command_id;
+	ret = nvme_tcp_req_map_sg(req, rq);
+	if (ret)
+		return -ENOMEM;
+
+	ret = netdev->netdev_ops->ulp_ddp_ops->setup(netdev, queue->sock->sk,
+						     &req->ddp);
+	if (ret) {
+		sg_free_table_chained(&req->ddp.sg_table, SG_CHUNK_SIZE);
+		return ret;
+	}
+
+	/* if successful, sg table is freed in nvme_tcp_teardown_ddp() */
+	req->offloaded = true;
+	return 0;
+}
+
 static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue)
 {
 	struct net_device *netdev = queue->ctrl->offloading_netdev;
@@ -453,6 +524,12 @@ static inline bool is_netdev_ulp_offload_active(struct net_device *netdev)
 	return false;
 }
 
+static int nvme_tcp_setup_ddp(struct nvme_tcp_queue *queue, u16 command_id,
+			      struct request *rq)
+{
+	return 0;
+}
+
 static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue)
 {
 	return 0;
@@ -740,6 +817,26 @@ static void nvme_tcp_error_recovery(struct nvme_ctrl *ctrl)
 	queue_work(nvme_reset_wq, &to_tcp_ctrl(ctrl)->err_work);
 }
 
+static void nvme_tcp_complete_request(struct request *rq,
+				      __le16 status,
+				      union nvme_result result,
+				      __u16 command_id)
+{
+#ifdef CONFIG_ULP_DDP
+	struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
+
+	if (req->offloaded) {
+		req->ddp_status = status;
+		req->result = result;
+		nvme_tcp_teardown_ddp(req->queue, rq);
+		return;
+	}
+#endif
+
+	if (!nvme_try_complete_req(rq, status, result))
+		nvme_complete_rq(rq);
+}
+
 static int nvme_tcp_process_nvme_cqe(struct nvme_tcp_queue *queue,
 		struct nvme_completion *cqe)
 {
@@ -759,10 +856,9 @@ static int nvme_tcp_process_nvme_cqe(struct nvme_tcp_queue *queue,
 	if (req->status == cpu_to_le16(NVME_SC_SUCCESS))
 		req->status = cqe->status;
 
-	if (!nvme_try_complete_req(rq, req->status, cqe->result))
-		nvme_complete_rq(rq);
+	nvme_tcp_complete_request(rq, req->status, cqe->result,
+				  cqe->command_id);
 	queue->nr_cqe++;
-
 	return 0;
 }
 
@@ -960,10 +1056,13 @@ static int nvme_tcp_recv_pdu(struct nvme_tcp_queue *queue, struct sk_buff *skb,
 
 static inline void nvme_tcp_end_request(struct request *rq, u16 status)
 {
+	struct nvme_tcp_request *req = blk_mq_rq_to_pdu(rq);
+	struct nvme_tcp_queue *queue = req->queue;
+	struct nvme_tcp_data_pdu *pdu = (void *)queue->pdu;
 	union nvme_result res = {};
 
-	if (!nvme_try_complete_req(rq, cpu_to_le16(status << 1), res))
-		nvme_complete_rq(rq);
+	nvme_tcp_complete_request(rq, cpu_to_le16(status << 1), res,
+				  pdu->command_id);
 }
 
 static int nvme_tcp_recv_data(struct nvme_tcp_queue *queue, struct sk_buff *skb,
@@ -1263,6 +1362,15 @@ static int nvme_tcp_try_send_cmd_pdu(struct nvme_tcp_request *req)
 	else
 		flags |= MSG_EOR;
 
+	if (test_bit(NVME_TCP_Q_OFF_DDP, &queue->flags)) {
+		ret = nvme_tcp_setup_ddp(queue, pdu->cmd.common.command_id,
+					 blk_mq_rq_from_pdu(req));
+		WARN_ONCE(ret, "ddp setup failed (queue 0x%x, cid 0x%x, ret=%d)",
+			  nvme_tcp_queue_id(queue),
+			  pdu->cmd.common.command_id,
+			  ret);
+	}
+
 	if (queue->hdr_digest && !req->offset)
 		nvme_tcp_hdgst(queue->snd_hash, pdu, sizeof(*pdu));
 
@@ -2594,6 +2702,9 @@ static blk_status_t nvme_tcp_setup_cmd_pdu(struct nvme_ns *ns,
 	if (ret)
 		return ret;
 
+#ifdef CONFIG_ULP_DDP
+	req->offloaded = false;
+#endif
 	req->state = NVME_TCP_SEND_CMD_PDU;
 	req->status = cpu_to_le16(NVME_SC_SUCCESS);
 	req->offset = 0;
Signature · 1 line
2.31.1
#14 lore raw

PATCH v11 10/25 nvme-tcp: Deal with netdevice DOWN events

In reply to Aurelien Aptel

From: Or Gerlitz <[email protected]>

For ddp setup/teardown and resync, the offloading logic uses HW resources at the NIC driver such as SQ and CQ.

These resources are destroyed when the netdevice does down and hence we must stop using them before the NIC driver destroys them.

Use netdevice notifier for that matter -- offloaded connections are stopped before the stack continues to call the NIC driver close ndo.

We use the existing recovery flow which has the advantage of resuming the offload once the connection is re-set.

This also buys us proper handling for the UNREGISTER event b/c our offloading starts in the UP state, and down is always there between up to unregister.

---
 drivers/nvme/host/tcp.c | 40 ++++++++++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)
drivers/nvme/host/tcp.c +40 -0
index 7e3feb694e46..732f7636a6fc 100644
@@ -204,6 +204,7 @@ struct nvme_tcp_ctrl {
 
 static LIST_HEAD(nvme_tcp_ctrl_list);
 static DEFINE_MUTEX(nvme_tcp_ctrl_mutex);
+static struct notifier_block nvme_tcp_netdevice_nb;
 static struct workqueue_struct *nvme_tcp_wq;
 static const struct blk_mq_ops nvme_tcp_mq_ops;
 static const struct blk_mq_ops nvme_tcp_admin_mq_ops;
@@ -3129,6 +3130,31 @@ static struct nvme_ctrl *nvme_tcp_create_ctrl(struct device *dev,
 	return ERR_PTR(ret);
 }
 
+static int nvme_tcp_netdev_event(struct notifier_block *this,
+				 unsigned long event, void *ptr)
+{
+	struct net_device *ndev = netdev_notifier_info_to_dev(ptr);
+	struct nvme_tcp_ctrl *ctrl;
+
+	switch (event) {
+	case NETDEV_GOING_DOWN:
+		mutex_lock(&nvme_tcp_ctrl_mutex);
+		list_for_each_entry(ctrl, &nvme_tcp_ctrl_list, list) {
+			if (ndev != ctrl->offloading_netdev)
+				continue;
+			nvme_tcp_error_recovery(&ctrl->ctrl);
+		}
+		mutex_unlock(&nvme_tcp_ctrl_mutex);
+		flush_workqueue(nvme_reset_wq);
+		/*
+		 * The associated controllers teardown has completed,
+		 * ddp contexts were also torn down so we should be
+		 * safe to continue...
+		 */
+	}
+	return NOTIFY_DONE;
+}
+
 static struct nvmf_transport_ops nvme_tcp_transport = {
 	.name		= "tcp",
 	.module		= THIS_MODULE,
@@ -3143,13 +3169,26 @@ static struct nvmf_transport_ops nvme_tcp_transport = {
 
 static int __init nvme_tcp_init_module(void)
 {
+	int ret;
+
 	nvme_tcp_wq = alloc_workqueue("nvme_tcp_wq",
 			WQ_MEM_RECLAIM | WQ_HIGHPRI, 0);
 	if (!nvme_tcp_wq)
 		return -ENOMEM;
 
+	nvme_tcp_netdevice_nb.notifier_call = nvme_tcp_netdev_event;
+	ret = register_netdevice_notifier(&nvme_tcp_netdevice_nb);
+	if (ret) {
+		pr_err("failed to register netdev notifier\n");
+		goto out_free_workqueue;
+	}
+
 	nvmf_register_transport(&nvme_tcp_transport);
 	return 0;
+
+out_free_workqueue:
+	destroy_workqueue(nvme_tcp_wq);
+	return ret;
 }
 
 static void __exit nvme_tcp_cleanup_module(void)
@@ -3157,6 +3196,7 @@ static void __exit nvme_tcp_cleanup_module(void)
 	struct nvme_tcp_ctrl *ctrl;
 
 	nvmf_unregister_transport(&nvme_tcp_transport);
+	unregister_netdevice_notifier(&nvme_tcp_netdevice_nb);
 
 	mutex_lock(&nvme_tcp_ctrl_mutex);
 	list_for_each_entry(ctrl, &nvme_tcp_ctrl_list, list)
Signature · 1 line
2.31.1
#15 lore raw

PATCH v11 11/25 nvme-tcp: Add modparam to control the ULP offload enablement

In reply to Aurelien Aptel

Add ulp_offload module parameter to the nvme-tcp module to control ULP offload at the NVMe-TCP layer.

Turn ULP offload off be default, regardless of the NIC driver support.

Overall, in order to enable ULP offload:
- nvme-tcp ulp_offload modparam must be set to 1
- netdev->ulp_ddp_caps.active must have ULP_DDP_C_NVME_TCP and/or ULP_DDP_C_NVME_TCP_DDGST_RX capabilities flag set.

---
 drivers/nvme/host/tcp.c | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)
drivers/nvme/host/tcp.c +19 -0
index 732f7636a6fc..77422f49fc76 100644
@@ -35,6 +35,16 @@ static int so_priority;
 module_param(so_priority, int, 0644);
 MODULE_PARM_DESC(so_priority, "nvme tcp socket optimize priority");
 
+#ifdef CONFIG_ULP_DDP
+/* NVMeTCP direct data placement and data digest offload will not
+ * happen if this parameter false (default), regardless of what the
+ * underlying netdev capabilities are.
+ */
+static bool ulp_offload;
+module_param(ulp_offload, bool, 0644);
+MODULE_PARM_DESC(ulp_offload, "Enable or disable NVMeTCP ULP support");
+#endif
+
 #ifdef CONFIG_DEBUG_LOCK_ALLOC
 /* lockdep can detect a circular dependency of the form
  *   sk_lock -> mmap_lock (page fault) -> fs locks -> sk_lock
@@ -317,6 +327,9 @@ static bool nvme_tcp_ddp_query_limits(struct net_device *netdev,
 {
 	int ret;
 
+	if (!ulp_offload)
+		return false;
+
 	if (!netdev || !is_netdev_ulp_offload_active(netdev, NULL) ||
 	    !netdev->netdev_ops->ulp_ddp_ops->limits)
 		return false;
@@ -456,6 +469,9 @@ static int nvme_tcp_offload_socket(struct nvme_tcp_queue *queue)
 					 netdev->ulp_ddp_caps.active);
 	int ret;
 
+	if (!ulp_offload)
+		return 0;
+
 	config.nvmeotcp.pfv = NVME_TCP_PFV_1_0;
 	config.nvmeotcp.cpda = 0;
 	config.nvmeotcp.dgst =
@@ -510,6 +526,9 @@ static void nvme_tcp_offload_limits(struct nvme_tcp_queue *queue,
 {
 	struct ulp_ddp_limits limits = {.type = ULP_DDP_NVME };
 
+	if (!ulp_offload)
+		return;
+
 	if (!nvme_tcp_ddp_query_limits(netdev, &limits)) {
 		queue->ctrl->offloading_netdev = NULL;
 		return;
Signature · 1 line
2.31.1
#16 lore raw

PATCH v11 12/25 Documentation: add ULP DDP offload documentation

In reply to Aurelien Aptel

From: Yoray Zack <[email protected]>

Document the new ULP DDP API and add it under "networking". Use NVMe-TCP implementation as an example.

---
 Documentation/netlink/specs/ethtool.yaml     |   6 +-
 Documentation/networking/index.rst           |   1 +
 Documentation/networking/ulp-ddp-offload.rst | 376 +++++++++++++++++++
 3 files changed, 380 insertions(+), 3 deletions(-)
 create mode 100644 Documentation/networking/ulp-ddp-offload.rst
Documentation/netlink/specs/ethtool.yaml +3 -3
index 8932cf310f1a..6c3b32572028 100644
@@ -453,7 +453,7 @@ operations:
       notify: mm-get
     -
       name: ulp-ddp-get
-      doc: Get ULP DDP capabilities and stats
+      doc: Get ULP DDP capabilities and stats.
 
       attribute-set: ulp-ddp
 
@@ -472,7 +472,7 @@ operations:
       dump: *ulp-ddp-get-op
     -
       name: ulp-ddp-set
-      doc: Set ULP DDP capabilities
+      doc: Set ULP DDP capabilities.
 
       attribute-set: ulp-ddp
 
@@ -490,5 +490,5 @@ operations:
             - active
     -
       name: ulp-ddp-ntf
-      doc: Notification for change in ULP DDP capabilities
+      doc: Notification for change in ULP DDP capabilities.
       notify: ulp-ddp-get
Documentation/networking/index.rst +1 -0
index 4ddcae33c336..bfb57df4b583 100644
@@ -106,6 +106,7 @@ Contents:
    tc-actions-env-rules
    tc-queue-filters
    tcp-thin
+   ulp-ddp-offload
    team
    timestamping
    tipc
Documentation/networking/ulp-ddp-offload.rst new file +376 -0

376 lines not on this page · raw message

Signature · 1 line
2.31.1
#17 lore raw

PATCH v11 13/25 net/mlx5e: Rename from tls to transport static params

In reply to Aurelien Aptel

From: Or Gerlitz <[email protected]>

The static params structure is used in TLS but also in other transports we're offloading like nvmeotcp:

- Rename the relevant structures/fields
- Create common file for appropriate transports
- Apply changes in the TLS code

No functional change here.

---
 .../mlx5/core/en_accel/common_utils.h         | 32 +++++++++++++++++
 .../mellanox/mlx5/core/en_accel/ktls.c        |  2 +-
 .../mellanox/mlx5/core/en_accel/ktls_rx.c     |  6 ++--
 .../mellanox/mlx5/core/en_accel/ktls_tx.c     |  8 ++---
 .../mellanox/mlx5/core/en_accel/ktls_txrx.c   | 36 ++++++++-----------
 .../mellanox/mlx5/core/en_accel/ktls_utils.h  | 17 ++-------
 include/linux/mlx5/device.h                   |  8 ++---
 include/linux/mlx5/mlx5_ifc.h                 |  8 +++--
 8 files changed, 67 insertions(+), 50 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/common_utils.h
drivers/net/ethernet/mellanox/mlx5/core/en_accel/common_utils.h new file +32 -0
new file mode 100644
index 000000000000..efdf48125848
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2023, NVIDIA CORPORATION & AFFILIATES. */
+#ifndef __MLX5E_COMMON_UTILS_H__
+#define __MLX5E_COMMON_UTILS_H__
+
+#include "en.h"
+
+struct mlx5e_set_transport_static_params_wqe {
+	struct mlx5_wqe_ctrl_seg ctrl;
+	struct mlx5_wqe_umr_ctrl_seg uctrl;
+	struct mlx5_mkey_seg mkc;
+	struct mlx5_wqe_transport_static_params_seg params;
+};
+
+/* macros for transport_static_params handling */
+#define MLX5E_TRANSPORT_SET_STATIC_PARAMS_WQEBBS \
+	(DIV_ROUND_UP(sizeof(struct mlx5e_set_transport_static_params_wqe), MLX5_SEND_WQE_BB))
+
+#define MLX5E_TRANSPORT_FETCH_SET_STATIC_PARAMS_WQE(sq, pi) \
+	((struct mlx5e_set_transport_static_params_wqe *)\
+	 mlx5e_fetch_wqe(&(sq)->wq, pi, sizeof(struct mlx5e_set_transport_static_params_wqe)))
+
+#define MLX5E_TRANSPORT_STATIC_PARAMS_WQE_SZ \
+	(sizeof(struct mlx5e_set_transport_static_params_wqe))
+
+#define MLX5E_TRANSPORT_STATIC_PARAMS_DS_CNT \
+	(DIV_ROUND_UP(MLX5E_TRANSPORT_STATIC_PARAMS_WQE_SZ, MLX5_SEND_WQE_DS))
+
+#define MLX5E_TRANSPORT_STATIC_PARAMS_OCTWORD_SIZE \
+	(MLX5_ST_SZ_BYTES(transport_static_params) / MLX5_SEND_WQE_DS)
+
+#endif /* __MLX5E_COMMON_UTILS_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls.c +1 -1
index cf704f106b7c..f80fdc7a7a5a 100644
@@ -100,7 +100,7 @@ bool mlx5e_is_ktls_rx(struct mlx5_core_dev *mdev)
 		return false;
 
 	/* Check the possibility to post the required ICOSQ WQEs. */
-	if (WARN_ON_ONCE(max_sq_wqebbs < MLX5E_TLS_SET_STATIC_PARAMS_WQEBBS))
+	if (WARN_ON_ONCE(max_sq_wqebbs < MLX5E_TRANSPORT_SET_STATIC_PARAMS_WQEBBS))
 		return false;
 	if (WARN_ON_ONCE(max_sq_wqebbs < MLX5E_TLS_SET_PROGRESS_PARAMS_WQEBBS))
 		return false;
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c +3 -3
index 4be770443b0c..f4c8f71cbcaa 100644
@@ -136,16 +136,16 @@ static struct mlx5_wqe_ctrl_seg *
 post_static_params(struct mlx5e_icosq *sq,
 		   struct mlx5e_ktls_offload_context_rx *priv_rx)
 {
-	struct mlx5e_set_tls_static_params_wqe *wqe;
+	struct mlx5e_set_transport_static_params_wqe *wqe;
 	struct mlx5e_icosq_wqe_info wi;
 	u16 pi, num_wqebbs;
 
-	num_wqebbs = MLX5E_TLS_SET_STATIC_PARAMS_WQEBBS;
+	num_wqebbs = MLX5E_TRANSPORT_SET_STATIC_PARAMS_WQEBBS;
 	if (unlikely(!mlx5e_icosq_can_post_wqe(sq, num_wqebbs)))
 		return ERR_PTR(-ENOSPC);
 
 	pi = mlx5e_icosq_get_next_pi(sq, num_wqebbs);
-	wqe = MLX5E_TLS_FETCH_SET_STATIC_PARAMS_WQE(sq, pi);
+	wqe = MLX5E_TRANSPORT_FETCH_SET_STATIC_PARAMS_WQE(sq, pi);
 	mlx5e_ktls_build_static_params(wqe, sq->pc, sq->sqn, &priv_rx->crypto_info,
 				       mlx5e_tir_get_tirn(&priv_rx->tir),
 				       mlx5_crypto_dek_get_id(priv_rx->dek),
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_tx.c +4 -4
index e80b43b7aac9..8e7ebc9d7ed4 100644
@@ -33,7 +33,7 @@ u16 mlx5e_ktls_get_stop_room(struct mlx5_core_dev *mdev, struct mlx5e_params *pa
 
 	num_dumps = mlx5e_ktls_dumps_num_wqes(params, MAX_SKB_FRAGS, TLS_MAX_PAYLOAD_SIZE);
 
-	stop_room += mlx5e_stop_room_for_wqe(mdev, MLX5E_TLS_SET_STATIC_PARAMS_WQEBBS);
+	stop_room += mlx5e_stop_room_for_wqe(mdev, MLX5E_TRANSPORT_SET_STATIC_PARAMS_WQEBBS);
 	stop_room += mlx5e_stop_room_for_wqe(mdev, MLX5E_TLS_SET_PROGRESS_PARAMS_WQEBBS);
 	stop_room += num_dumps * mlx5e_stop_room_for_wqe(mdev, MLX5E_KTLS_DUMP_WQEBBS);
 	stop_room += 1; /* fence nop */
@@ -548,12 +548,12 @@ post_static_params(struct mlx5e_txqsq *sq,
 		   struct mlx5e_ktls_offload_context_tx *priv_tx,
 		   bool fence)
 {
-	struct mlx5e_set_tls_static_params_wqe *wqe;
+	struct mlx5e_set_transport_static_params_wqe *wqe;
 	u16 pi, num_wqebbs;
 
-	num_wqebbs = MLX5E_TLS_SET_STATIC_PARAMS_WQEBBS;
+	num_wqebbs = MLX5E_TRANSPORT_SET_STATIC_PARAMS_WQEBBS;
 	pi = mlx5e_txqsq_get_next_pi(sq, num_wqebbs);
-	wqe = MLX5E_TLS_FETCH_SET_STATIC_PARAMS_WQE(sq, pi);
+	wqe = MLX5E_TRANSPORT_FETCH_SET_STATIC_PARAMS_WQE(sq, pi);
 	mlx5e_ktls_build_static_params(wqe, sq->pc, sq->sqn, &priv_tx->crypto_info,
 				       priv_tx->tisn,
 				       mlx5_crypto_dek_get_id(priv_tx->dek),
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_txrx.c +15 -21
index 570a912dd6fa..8abea6fe6cd9 100644
@@ -8,10 +8,6 @@ enum {
 	MLX5E_STATIC_PARAMS_CONTEXT_TLS_1_2 = 0x2,
 };
 
-enum {
-	MLX5E_ENCRYPTION_STANDARD_TLS = 0x1,
-};
-
 #define EXTRACT_INFO_FIELDS do { \
 	salt    = info->salt;    \
 	rec_seq = info->rec_seq; \
@@ -20,7 +16,7 @@ enum {
 } while (0)
 
 static void
-fill_static_params(struct mlx5_wqe_tls_static_params_seg *params,
+fill_static_params(struct mlx5_wqe_transport_static_params_seg *params,
 		   union mlx5e_crypto_info *crypto_info,
 		   u32 key_id, u32 resync_tcp_sn)
 {
@@ -53,25 +49,25 @@ fill_static_params(struct mlx5_wqe_tls_static_params_seg *params,
 		return;
 	}
 
-	gcm_iv      = MLX5_ADDR_OF(tls_static_params, ctx, gcm_iv);
-	initial_rn  = MLX5_ADDR_OF(tls_static_params, ctx, initial_record_number);
+	gcm_iv      = MLX5_ADDR_OF(transport_static_params, ctx, gcm_iv);
+	initial_rn  = MLX5_ADDR_OF(transport_static_params, ctx, initial_record_number);
 
 	memcpy(gcm_iv,      salt,    salt_sz);
 	memcpy(initial_rn,  rec_seq, rec_seq_sz);
 
 	tls_version = MLX5E_STATIC_PARAMS_CONTEXT_TLS_1_2;
 
-	MLX5_SET(tls_static_params, ctx, tls_version, tls_version);
-	MLX5_SET(tls_static_params, ctx, const_1, 1);
-	MLX5_SET(tls_static_params, ctx, const_2, 2);
-	MLX5_SET(tls_static_params, ctx, encryption_standard,
-		 MLX5E_ENCRYPTION_STANDARD_TLS);
-	MLX5_SET(tls_static_params, ctx, resync_tcp_sn, resync_tcp_sn);
-	MLX5_SET(tls_static_params, ctx, dek_index, key_id);
+	MLX5_SET(transport_static_params, ctx, tls_version, tls_version);
+	MLX5_SET(transport_static_params, ctx, const_1, 1);
+	MLX5_SET(transport_static_params, ctx, const_2, 2);
+	MLX5_SET(transport_static_params, ctx, acc_type,
+		 MLX5_TRANSPORT_STATIC_PARAMS_ACC_TYPE_TLS);
+	MLX5_SET(transport_static_params, ctx, resync_tcp_sn, resync_tcp_sn);
+	MLX5_SET(transport_static_params, ctx, dek_index, key_id);
 }
 
 void
-mlx5e_ktls_build_static_params(struct mlx5e_set_tls_static_params_wqe *wqe,
+mlx5e_ktls_build_static_params(struct mlx5e_set_transport_static_params_wqe *wqe,
 			       u16 pc, u32 sqn,
 			       union mlx5e_crypto_info *crypto_info,
 			       u32 tis_tir_num, u32 key_id, u32 resync_tcp_sn,
@@ -80,19 +76,17 @@ mlx5e_ktls_build_static_params(struct mlx5e_set_tls_static_params_wqe *wqe,
 	struct mlx5_wqe_umr_ctrl_seg *ucseg = &wqe->uctrl;
 	struct mlx5_wqe_ctrl_seg     *cseg  = &wqe->ctrl;
 	u8 opmod = direction == TLS_OFFLOAD_CTX_DIR_TX ?
-		MLX5_OPC_MOD_TLS_TIS_STATIC_PARAMS :
-		MLX5_OPC_MOD_TLS_TIR_STATIC_PARAMS;
-
-#define STATIC_PARAMS_DS_CNT DIV_ROUND_UP(sizeof(*wqe), MLX5_SEND_WQE_DS)
+		MLX5_OPC_MOD_TRANSPORT_TIS_STATIC_PARAMS :
+		MLX5_OPC_MOD_TRANSPORT_TIR_STATIC_PARAMS;
 
 	cseg->opmod_idx_opcode = cpu_to_be32((pc << 8) | MLX5_OPCODE_UMR | (opmod << 24));
 	cseg->qpn_ds           = cpu_to_be32((sqn << MLX5_WQE_CTRL_QPN_SHIFT) |
-					     STATIC_PARAMS_DS_CNT);
+					     MLX5E_TRANSPORT_STATIC_PARAMS_DS_CNT);
 	cseg->fm_ce_se         = fence ? MLX5_FENCE_MODE_INITIATOR_SMALL : 0;
 	cseg->tis_tir_num      = cpu_to_be32(tis_tir_num << 8);
 
 	ucseg->flags = MLX5_UMR_INLINE;
-	ucseg->bsf_octowords = cpu_to_be16(MLX5_ST_SZ_BYTES(tls_static_params) / 16);
+	ucseg->bsf_octowords = cpu_to_be16(MLX5E_TRANSPORT_STATIC_PARAMS_OCTWORD_SIZE);
 
 	fill_static_params(&wqe->params, crypto_info, key_id, resync_tcp_sn);
 }
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_utils.h +2 -15
index 3d79cd379890..5e2d186778aa 100644
@@ -6,6 +6,7 @@
 
 #include <net/tls.h>
 #include "en.h"
+#include "en_accel/common_utils.h"
 
 enum {
 	MLX5E_TLS_PROGRESS_PARAMS_AUTH_STATE_NO_OFFLOAD     = 0,
@@ -33,13 +34,6 @@ union mlx5e_crypto_info {
 	struct tls12_crypto_info_aes_gcm_256 crypto_info_256;
 };
 
-struct mlx5e_set_tls_static_params_wqe {
-	struct mlx5_wqe_ctrl_seg ctrl;
-	struct mlx5_wqe_umr_ctrl_seg uctrl;
-	struct mlx5_mkey_seg mkc;
-	struct mlx5_wqe_tls_static_params_seg params;
-};
-
 struct mlx5e_set_tls_progress_params_wqe {
 	struct mlx5_wqe_ctrl_seg ctrl;
 	struct mlx5_wqe_tls_progress_params_seg params;
@@ -50,19 +44,12 @@ struct mlx5e_get_tls_progress_params_wqe {
 	struct mlx5_seg_get_psv  psv;
 };
 
-#define MLX5E_TLS_SET_STATIC_PARAMS_WQEBBS \
-	(DIV_ROUND_UP(sizeof(struct mlx5e_set_tls_static_params_wqe), MLX5_SEND_WQE_BB))
-
 #define MLX5E_TLS_SET_PROGRESS_PARAMS_WQEBBS \
 	(DIV_ROUND_UP(sizeof(struct mlx5e_set_tls_progress_params_wqe), MLX5_SEND_WQE_BB))
 
 #define MLX5E_KTLS_GET_PROGRESS_WQEBBS \
 	(DIV_ROUND_UP(sizeof(struct mlx5e_get_tls_progress_params_wqe), MLX5_SEND_WQE_BB))
 
-#define MLX5E_TLS_FETCH_SET_STATIC_PARAMS_WQE(sq, pi) \
-	((struct mlx5e_set_tls_static_params_wqe *)\
-	 mlx5e_fetch_wqe(&(sq)->wq, pi, sizeof(struct mlx5e_set_tls_static_params_wqe)))
-
 #define MLX5E_TLS_FETCH_SET_PROGRESS_PARAMS_WQE(sq, pi) \
 	((struct mlx5e_set_tls_progress_params_wqe *)\
 	 mlx5e_fetch_wqe(&(sq)->wq, pi, sizeof(struct mlx5e_set_tls_progress_params_wqe)))
@@ -76,7 +63,7 @@ struct mlx5e_get_tls_progress_params_wqe {
 	 mlx5e_fetch_wqe(&(sq)->wq, pi, sizeof(struct mlx5e_dump_wqe)))
 
 void
-mlx5e_ktls_build_static_params(struct mlx5e_set_tls_static_params_wqe *wqe,
+mlx5e_ktls_build_static_params(struct mlx5e_set_transport_static_params_wqe *wqe,
 			       u16 pc, u32 sqn,
 			       union mlx5e_crypto_info *crypto_info,
 			       u32 tis_tir_num, u32 key_id, u32 resync_tcp_sn,
include/linux/mlx5/device.h +4 -4
index bc531bd9804f..9e1671506067 100644
@@ -444,8 +444,8 @@ enum {
 };
 
 enum {
-	MLX5_OPC_MOD_TLS_TIS_STATIC_PARAMS = 0x1,
-	MLX5_OPC_MOD_TLS_TIR_STATIC_PARAMS = 0x2,
+	MLX5_OPC_MOD_TRANSPORT_TIS_STATIC_PARAMS = 0x1,
+	MLX5_OPC_MOD_TRANSPORT_TIR_STATIC_PARAMS = 0x2,
 };
 
 enum {
@@ -453,8 +453,8 @@ enum {
 	MLX5_OPC_MOD_TLS_TIR_PROGRESS_PARAMS = 0x2,
 };
 
-struct mlx5_wqe_tls_static_params_seg {
-	u8     ctx[MLX5_ST_SZ_BYTES(tls_static_params)];
+struct mlx5_wqe_transport_static_params_seg {
+	u8     ctx[MLX5_ST_SZ_BYTES(transport_static_params)];
 };
 
 struct mlx5_wqe_tls_progress_params_seg {
include/linux/mlx5/mlx5_ifc.h +6 -2
index 1b6201bb04c1..170df4fd60de 100644
@@ -12140,12 +12140,16 @@ enum {
 	MLX5_GENERAL_OBJECT_TYPE_ENCRYPTION_KEY_PURPOSE_MACSEC = 0x4,
 };
 
-struct mlx5_ifc_tls_static_params_bits {
+enum {
+	MLX5_TRANSPORT_STATIC_PARAMS_ACC_TYPE_TLS               = 0x1,
+};
+
+struct mlx5_ifc_transport_static_params_bits {
 	u8         const_2[0x2];
 	u8         tls_version[0x4];
 	u8         const_1[0x2];
 	u8         reserved_at_8[0x14];
-	u8         encryption_standard[0x4];
+	u8         acc_type[0x4];
 
 	u8         reserved_at_20[0x20];
 
Signature · 1 line
2.31.1
#24 lore raw

PATCH v11 20/25 net/mlx5e: NVMEoTCP, use KLM UMRs for buffer registration

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

NVMEoTCP offload uses buffer registration for ddp operation. Every request comprises from SG list that might consist from elements with multiple combination sizes, thus the appropriate way to perform buffer registration is with KLM UMRs.

UMR stands for user-mode memory registration, it is a mechanism to alter address translation properties of MKEY by posting WorkQueueElement aka WQE on send queue.

MKEY stands for memory key, MKEY are used to describe a region in memory that can be later used by HW.

KLM stands for {Key, Length, MemVa}, KLM_MKEY is indirect MKEY that enables to map multiple memory spaces with different sizes in unified MKEY. KLM UMR is a UMR that use to update a KLM_MKEY.

---
 .../net/ethernet/mellanox/mlx5/core/en/txrx.h |   3 +
 .../mellanox/mlx5/core/en_accel/nvmeotcp.c    | 123 ++++++++++++++++++
 .../mlx5/core/en_accel/nvmeotcp_utils.h       |  25 ++++
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   |   4 +
 4 files changed, 155 insertions(+)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_utils.h
drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h +3 -0
index 37217e4cc118..58be28376ca5 100644
@@ -50,6 +50,9 @@ enum mlx5e_icosq_wqe_type {
 	MLX5E_ICOSQ_WQE_SET_PSV_TLS,
 	MLX5E_ICOSQ_WQE_GET_PSV_TLS,
 #endif
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+	MLX5E_ICOSQ_WQE_UMR_NVMEOTCP,
+#endif
 };
 
 /* General */
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c +123 -0

141 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_utils.h new file +25 -0
new file mode 100644
index 000000000000..6ef92679c5d0
@@ -0,0 +1,25 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2023, NVIDIA CORPORATION & AFFILIATES. */
+#ifndef __MLX5E_NVMEOTCP_UTILS_H__
+#define __MLX5E_NVMEOTCP_UTILS_H__
+
+#include "en.h"
+
+#define MLX5E_NVMEOTCP_FETCH_KLM_WQE(sq, pi) \
+	((struct mlx5e_umr_wqe *)\
+	 mlx5e_fetch_wqe(&(sq)->wq, pi, sizeof(struct mlx5e_umr_wqe)))
+
+#define MLX5_CTRL_SEGMENT_OPC_MOD_UMR_NVMEOTCP_TIR_PROGRESS_PARAMS 0x4
+
+#define MLX5_CTRL_SEGMENT_OPC_MOD_UMR_TIR_PARAMS 0x2
+#define MLX5_CTRL_SEGMENT_OPC_MOD_UMR_UMR 0x0
+
+enum wqe_type {
+	KLM_UMR,
+	BSF_KLM_UMR,
+	SET_PSV_UMR,
+	BSF_UMR,
+	KLM_INV_UMR,
+};
+
+#endif /* __MLX5E_NVMEOTCP_UTILS_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +4 -0
index d3fc3724a39f..3ae2c3f09476 100644
@@ -981,6 +981,10 @@ int mlx5e_poll_ico_cq(struct mlx5e_cq *cq, int budget)
 			case MLX5E_ICOSQ_WQE_GET_PSV_TLS:
 				mlx5e_ktls_handle_get_psv_completion(wi, sq);
 				break;
+#endif
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+			case MLX5E_ICOSQ_WQE_UMR_NVMEOTCP:
+				break;
 #endif
 			default:
 				netdev_WARN_ONCE(cq->netdev,
Signature · 1 line
2.31.1
#18 lore raw

PATCH v11 14/25 net/mlx5e: Refactor ico sq polling to get budget

In reply to Aurelien Aptel

From: Or Gerlitz <[email protected]>

The mlx5e driver uses ICO SQs for internal control operations which are not visible to the network stack, such as UMR mapping for striding RQ (MPWQ) and etc more cases.

The upcoming nvmeotcp offload uses ico sq for umr mapping as part of the offload. As a pre-step for nvmeotcp ico sqs which have their own napi and need to comply with budget, add the budget as parameter to the polling of cqs related to ico sqs.

The polling already stops after a limit is reached, so just have the caller to provide this limit as the budget.

No functional change here.

---
 drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h | 2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_rx.c   | 4 ++--
 drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c | 4 ++--
 3 files changed, 5 insertions(+), 5 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h +1 -1
index c10c6ab2e7bc..92a71d1a7605 100644
@@ -62,7 +62,7 @@ void mlx5e_trigger_irq(struct mlx5e_icosq *sq);
 void mlx5e_completion_event(struct mlx5_core_cq *mcq, struct mlx5_eqe *eqe);
 void mlx5e_cq_error_event(struct mlx5_core_cq *mcq, enum mlx5_event event);
 int mlx5e_napi_poll(struct napi_struct *napi, int budget);
-int mlx5e_poll_ico_cq(struct mlx5e_cq *cq);
+int mlx5e_poll_ico_cq(struct mlx5e_cq *cq, int budget);
 
 /* RX */
 void mlx5e_page_dma_unmap(struct mlx5e_rq *rq, struct page *page);
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +2 -2
index a9473a51edc1..d3fc3724a39f 100644
@@ -913,7 +913,7 @@ static void mlx5e_handle_shampo_hd_umr(struct mlx5e_shampo_umr umr,
 	shampo->ci = (shampo->ci + umr.len) & (shampo->hd_per_wq - 1);
 }
 
-int mlx5e_poll_ico_cq(struct mlx5e_cq *cq)
+int mlx5e_poll_ico_cq(struct mlx5e_cq *cq, int budget)
 {
 	struct mlx5e_icosq *sq = container_of(cq, struct mlx5e_icosq, cq);
 	struct mlx5_cqe64 *cqe;
@@ -988,7 +988,7 @@ int mlx5e_poll_ico_cq(struct mlx5e_cq *cq)
 						 wi->wqe_type);
 			}
 		} while (!last_wqe);
-	} while ((++i < MLX5E_TX_CQ_POLL_BUDGET) && (cqe = mlx5_cqwq_get_cqe(&cq->wq)));
+	} while ((++i < budget) && (cqe = mlx5_cqwq_get_cqe(&cq->wq)));
 
 	sq->cc = sqcc;
 
drivers/net/ethernet/mellanox/mlx5/core/en_txrx.c +2 -2
index 9a458a5d9853..9ddacb5e1bf4 100644
@@ -176,8 +176,8 @@ int mlx5e_napi_poll(struct napi_struct *napi, int budget)
 		busy |= work_done == budget;
 	}
 
-	mlx5e_poll_ico_cq(&c->icosq.cq);
-	if (mlx5e_poll_ico_cq(&c->async_icosq.cq))
+	mlx5e_poll_ico_cq(&c->icosq.cq, MLX5E_TX_CQ_POLL_BUDGET);
+	if (mlx5e_poll_ico_cq(&c->async_icosq.cq, MLX5E_TX_CQ_POLL_BUDGET))
 		/* Don't clear the flag if nothing was polled to prevent
 		 * queueing more WQEs and overflowing the async ICOSQ.
 		 */
Signature · 1 line
2.31.1
#19 lore raw

PATCH v11 15/25 net/mlx5e: Have mdev pointer directly on the icosq structure

In reply to Aurelien Aptel

From: Or Gerlitz <[email protected]>

This provides better separation between channels to ICO SQs for use-cases where they are not tightly coupled (such as the upcoming nvmeotcp code).

No functional change here.

---
 drivers/net/ethernet/mellanox/mlx5/core/en.h               | 1 +
 drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c   | 4 ++--
 drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c | 2 +-
 drivers/net/ethernet/mellanox/mlx5/core/en_main.c          | 5 ++---
 4 files changed, 6 insertions(+), 6 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en.h +1 -0
index 6f8723cc6874..dd68f373516b 100644
@@ -592,6 +592,7 @@ struct mlx5e_icosq {
 	/* control path */
 	struct mlx5_wq_ctrl        wq_ctrl;
 	struct mlx5e_channel      *channel;
+	struct mlx5_core_dev      *mdev;
 
 	struct work_struct         recover_work;
 } ____cacheline_aligned_in_smp;
drivers/net/ethernet/mellanox/mlx5/core/en/reporter_rx.c +2 -2
index 95edab4a1732..bfac3f09f964 100644
@@ -33,7 +33,7 @@ static int mlx5e_query_rq_state(struct mlx5_core_dev *dev, u32 rqn, u8 *state)
 
 static int mlx5e_wait_for_icosq_flush(struct mlx5e_icosq *icosq)
 {
-	struct mlx5_core_dev *dev = icosq->channel->mdev;
+	struct mlx5_core_dev *dev = icosq->mdev;
 	unsigned long exp_time;
 
 	exp_time = jiffies + msecs_to_jiffies(mlx5_tout_ms(dev, FLUSH_ON_ERROR));
@@ -78,7 +78,7 @@ static int mlx5e_rx_reporter_err_icosq_cqe_recover(void *ctx)
 	rq = &icosq->channel->rq;
 	if (test_bit(MLX5E_RQ_STATE_ENABLED, &icosq->channel->xskrq.state))
 		xskrq = &icosq->channel->xskrq;
-	mdev = icosq->channel->mdev;
+	mdev = icosq->mdev;
 	dev = icosq->channel->netdev;
 	err = mlx5_core_query_sq_state(mdev, icosq->sqn, &state);
 	if (err) {
drivers/net/ethernet/mellanox/mlx5/core/en_accel/ktls_rx.c +1 -1
index f4c8f71cbcaa..e7866e88c599 100644
@@ -267,7 +267,7 @@ resync_post_get_progress_params(struct mlx5e_icosq *sq,
 		goto err_out;
 	}
 
-	pdev = mlx5_core_dma_dev(sq->channel->priv->mdev);
+	pdev = mlx5_core_dma_dev(sq->mdev);
 	buf->dma_addr = dma_map_single(pdev, &buf->progress,
 				       PROGRESS_PARAMS_PADDED_SIZE, DMA_FROM_DEVICE);
 	if (unlikely(dma_mapping_error(pdev, buf->dma_addr))) {
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +2 -3
index 0e87432ec6f1..43a3ff2d2d90 100644
@@ -1389,6 +1389,7 @@ static int mlx5e_alloc_icosq(struct mlx5e_channel *c,
 	int err;
 
 	sq->channel   = c;
+	sq->mdev      = mdev;
 	sq->uar_map   = mdev->mlx5e_res.hw_objs.bfreg.map;
 	sq->reserved_room = param->stop_room;
 
@@ -1786,11 +1787,9 @@ void mlx5e_deactivate_icosq(struct mlx5e_icosq *icosq)
 
 static void mlx5e_close_icosq(struct mlx5e_icosq *sq)
 {
-	struct mlx5e_channel *c = sq->channel;
-
 	if (sq->ktls_resync)
 		mlx5e_ktls_rx_resync_destroy_resp_list(sq->ktls_resync);
-	mlx5e_destroy_sq(c->mdev, sq->sqn);
+	mlx5e_destroy_sq(sq->mdev, sq->sqn);
 	mlx5e_free_icosq_descs(sq);
 	mlx5e_free_icosq(sq);
 }
Signature · 1 line
2.31.1
#28 lore raw

PATCH v11 24/25 net/mlx5e: NVMEoTCP, data-path for DDP+DDGST offload

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

This patch implements the data-path for direct data placement (DDP) and DDGST offloads. NVMEoTCP DDP constructs an SKB from each CQE, while pointing at NVME destination buffers. In turn, this enables the offload, as the NVMe-TCP layer will skip the copy when src == dst.

Additionally, this patch adds support for DDGST (CRC32) offload. HW will report DDGST offload only if it has not encountered an error in the received packet. We pass this indication in skb->ulp_crc up the stack to NVMe-TCP to skip computing the DDGST if all corresponding SKBs were verified by HW.

This patch also handles context resynchronization requests made by NIC HW. The resync request is passed to the NVMe-TCP layer to be handled at a later point in time.

Finally, we also use the skb->ulp_ddp bit to avoid skb_condense. This is critical as every SKB that uses DDP has a hole that fits perfectly with skb_condense's policy, but filling this hole is counter-productive as the data there already resides in its destination buffer.

This work has been done on pre-silicon functional simulator, and hence data-path performance numbers are not provided.

---
 .../net/ethernet/mellanox/mlx5/core/Makefile  |   2 +-
 .../mlx5/core/en_accel/nvmeotcp_rxtx.c        | 316 ++++++++++++++++++
 .../mlx5/core/en_accel/nvmeotcp_rxtx.h        |  37 ++
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   |  40 ++-
 4 files changed, 384 insertions(+), 11 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_rxtx.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_rxtx.h
drivers/net/ethernet/mellanox/mlx5/core/Makefile +1 -1
index 9df9999047d1..9804bd086bf4 100644
@@ -103,7 +103,7 @@ mlx5_core-$(CONFIG_MLX5_EN_TLS) += en_accel/ktls_stats.o \
 				   en_accel/fs_tcp.o en_accel/ktls.o en_accel/ktls_txrx.o \
 				   en_accel/ktls_tx.o en_accel/ktls_rx.o
 
-mlx5_core-$(CONFIG_MLX5_EN_NVMEOTCP) += en_accel/fs_tcp.o en_accel/nvmeotcp.o
+mlx5_core-$(CONFIG_MLX5_EN_NVMEOTCP) += en_accel/fs_tcp.o en_accel/nvmeotcp.o en_accel/nvmeotcp_rxtx.o
 
 mlx5_core-$(CONFIG_MLX5_SW_STEERING) += steering/dr_domain.o steering/dr_table.o \
 					steering/dr_matcher.o steering/dr_rule.o \
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_rxtx.c new file +316 -0

316 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_rxtx.h new file +37 -0
new file mode 100644
index 000000000000..a8ca8a53bac6
@@ -0,0 +1,37 @@
+/* SPDX-License-Identifier: GPL-2.0 OR Linux-OpenIB */
+/* Copyright (c) 2023, NVIDIA CORPORATION & AFFILIATES. */
+#ifndef __MLX5E_NVMEOTCP_RXTX_H__
+#define __MLX5E_NVMEOTCP_RXTX_H__
+
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+
+#include <linux/skbuff.h>
+#include "en_accel/nvmeotcp.h"
+
+bool
+mlx5e_nvmeotcp_rebuild_rx_skb(struct mlx5e_rq *rq, struct sk_buff *skb,
+			      struct mlx5_cqe64 *cqe, u32 cqe_bcnt);
+
+static inline int mlx5_nvmeotcp_get_headlen(struct mlx5_cqe64 *cqe, u32 cqe_bcnt)
+{
+	struct mlx5e_cqe128 *cqe128;
+
+	if (!cqe_is_nvmeotcp_zc(cqe))
+		return cqe_bcnt;
+
+	cqe128 = container_of(cqe, struct mlx5e_cqe128, cqe64);
+	return be16_to_cpu(cqe128->hlen);
+}
+
+#else
+
+static inline bool
+mlx5e_nvmeotcp_rebuild_rx_skb(struct mlx5e_rq *rq, struct sk_buff *skb,
+			      struct mlx5_cqe64 *cqe, u32 cqe_bcnt)
+{ return true; }
+
+static inline int mlx5_nvmeotcp_get_headlen(struct mlx5_cqe64 *cqe, u32 cqe_bcnt)
+{ return cqe_bcnt; }
+
+#endif /* CONFIG_MLX5_EN_NVMEOTCP */
+#endif /* __MLX5E_NVMEOTCP_RXTX_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +30 -10
index 71f73cd5d36e..1a6786448d75 100644
@@ -53,7 +53,7 @@
 #include "en_accel/macsec.h"
 #include "en_accel/ipsec_rxtx.h"
 #include "en_accel/ktls_txrx.h"
-#include "en_accel/nvmeotcp.h"
+#include "en_accel/nvmeotcp_rxtx.h"
 #include "en/xdp.h"
 #include "en/xsk/rx.h"
 #include "en/health.h"
@@ -1481,7 +1481,7 @@ static inline void mlx5e_handle_csum(struct net_device *netdev,
 
 #define MLX5E_CE_BIT_MASK 0x80
 
-static inline void mlx5e_build_rx_skb(struct mlx5_cqe64 *cqe,
+static inline bool mlx5e_build_rx_skb(struct mlx5_cqe64 *cqe,
 				      u32 cqe_bcnt,
 				      struct mlx5e_rq *rq,
 				      struct sk_buff *skb)
@@ -1492,6 +1492,13 @@ static inline void mlx5e_build_rx_skb(struct mlx5_cqe64 *cqe,
 
 	skb->mac_len = ETH_HLEN;
 
+	if (IS_ENABLED(CONFIG_MLX5_EN_NVMEOTCP) && cqe_is_nvmeotcp(cqe)) {
+		bool ret = mlx5e_nvmeotcp_rebuild_rx_skb(rq, skb, cqe, cqe_bcnt);
+
+		if (unlikely(!ret))
+			return ret;
+	}
+
 	if (unlikely(get_cqe_tls_offload(cqe)))
 		mlx5e_ktls_handle_rx_skb(rq, skb, cqe, &cqe_bcnt);
 
@@ -1537,6 +1544,8 @@ static inline void mlx5e_build_rx_skb(struct mlx5_cqe64 *cqe,
 
 	if (unlikely(mlx5e_skb_is_multicast(skb)))
 		stats->mcast_packets++;
+
+	return true;
 }
 
 static void mlx5e_shampo_complete_rx_cqe(struct mlx5e_rq *rq,
@@ -1560,7 +1569,7 @@ static void mlx5e_shampo_complete_rx_cqe(struct mlx5e_rq *rq,
 	}
 }
 
-static inline void mlx5e_complete_rx_cqe(struct mlx5e_rq *rq,
+static inline bool mlx5e_complete_rx_cqe(struct mlx5e_rq *rq,
 					 struct mlx5_cqe64 *cqe,
 					 u32 cqe_bcnt,
 					 struct sk_buff *skb)
@@ -1569,7 +1578,7 @@ static inline void mlx5e_complete_rx_cqe(struct mlx5e_rq *rq,
 
 	stats->packets++;
 	stats->bytes += cqe_bcnt;
-	mlx5e_build_rx_skb(cqe, cqe_bcnt, rq, skb);
+	return mlx5e_build_rx_skb(cqe, cqe_bcnt, rq, skb);
 }
 
 static inline
@@ -1810,7 +1819,8 @@ static void mlx5e_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 		goto free_wqe;
 	}
 
-	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	if (unlikely(!mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb)))
+		goto free_wqe;
 
 	if (mlx5e_cqe_regb_chain(cqe))
 		if (!mlx5e_tc_update_skb(cqe, skb)) {
@@ -1863,7 +1873,8 @@ static void mlx5e_handle_rx_cqe_rep(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe)
 		goto free_wqe;
 	}
 
-	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	if (unlikely(!mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb)))
+		goto free_wqe;
 
 	if (rep->vlan && skb_vlan_tag_present(skb))
 		skb_vlan_pop(skb);
@@ -1914,7 +1925,8 @@ static void mlx5e_handle_rx_cqe_mpwrq_rep(struct mlx5e_rq *rq, struct mlx5_cqe64
 	if (!skb)
 		goto mpwrq_cqe_out;
 
-	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	if (unlikely(!mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb)))
+		goto mpwrq_cqe_out;
 
 	mlx5e_rep_tc_receive(cqe, rq, skb);
 
@@ -1959,13 +1971,18 @@ mlx5e_fill_skb_data(struct sk_buff *skb, struct mlx5e_rq *rq,
 	}
 }
 
+static inline u16 mlx5e_get_headlen_hint(struct mlx5_cqe64 *cqe, u32 cqe_bcnt)
+{
+	return min_t(u32, MLX5E_RX_MAX_HEAD, mlx5_nvmeotcp_get_headlen(cqe, cqe_bcnt));
+}
+
 static struct sk_buff *
 mlx5e_skb_from_cqe_mpwrq_nonlinear(struct mlx5e_rq *rq, struct mlx5e_mpw_info *wi,
 				   struct mlx5_cqe64 *cqe, u16 cqe_bcnt, u32 head_offset,
 				   u32 page_idx)
 {
 	union mlx5e_alloc_unit *au = &wi->alloc_units[page_idx];
-	u16 headlen = min_t(u16, MLX5E_RX_MAX_HEAD, cqe_bcnt);
+	u16 headlen        = mlx5e_get_headlen_hint(cqe, cqe_bcnt);
 	u32 frag_offset    = head_offset + headlen;
 	u32 byte_cnt       = cqe_bcnt - headlen;
 	union mlx5e_alloc_unit *head_au = au;
@@ -2277,7 +2294,8 @@ static void mlx5e_handle_rx_cqe_mpwrq(struct mlx5e_rq *rq, struct mlx5_cqe64 *cq
 	if (!skb)
 		goto mpwrq_cqe_out;
 
-	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	if (unlikely(!mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb)))
+		goto mpwrq_cqe_out;
 
 	if (mlx5e_cqe_regb_chain(cqe))
 		if (!mlx5e_tc_update_skb(cqe, skb)) {
@@ -2614,7 +2632,9 @@ static void mlx5e_trap_handle_rx_cqe(struct mlx5e_rq *rq, struct mlx5_cqe64 *cqe
 	if (!skb)
 		goto free_wqe;
 
-	mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb);
+	if (unlikely(!mlx5e_complete_rx_cqe(rq, cqe, cqe_bcnt, skb)))
+		goto free_wqe;
+
 	skb_push(skb, ETH_HLEN);
 
 	dl_port = mlx5e_devlink_get_dl_port(priv);
Signature · 1 line
2.31.1
#25 lore raw

PATCH v11 21/25 net/mlx5e: NVMEoTCP, queue init/teardown

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

Adds the ddp ops of sk_add, sk_del and offload limits.

When nvme-tcp establishes new queue/connection, the sk_add op is called. We allocate a hardware context to offload operations for this queue:
- use a steering rule based on the connection 5-tuple to mark packets of this queue/connection with a flow-tag in their completion (cqe)
- use a dedicated TIR to identify the queue and maintain the HW context
- use a dedicated ICOSQ to maintain the HW context by UMR postings
- use a dedicated tag buffer for buffer registration
- maintain static and progress HW contexts by posting the proper WQEs.

When nvme-tcp teardowns a queue/connection, the sk_del op is called. We teardown the queue and free the corresponding contexts.

The offload limits we advertise deal with the max SG supported.

[Re-enabled calling open/close icosq out of en_main.c]

---
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |   4 +
 .../ethernet/mellanox/mlx5/core/en/rx_res.c   |  28 +
 .../ethernet/mellanox/mlx5/core/en/rx_res.h   |   4 +
 .../net/ethernet/mellanox/mlx5/core/en/tir.c  |  15 +
 .../net/ethernet/mellanox/mlx5/core/en/tir.h  |   2 +
 .../net/ethernet/mellanox/mlx5/core/en/txrx.h |   6 +
 .../mellanox/mlx5/core/en_accel/nvmeotcp.c    | 562 +++++++++++++++++-
 .../mellanox/mlx5/core/en_accel/nvmeotcp.h    |   4 +
 .../mlx5/core/en_accel/nvmeotcp_utils.h       |  41 ++
 .../net/ethernet/mellanox/mlx5/core/en_main.c |   8 +-
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   |  15 +-
 11 files changed, 679 insertions(+), 10 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en.h +4 -0
index 1f8cf297464d..46bdea1a9b0a 100644
@@ -1059,6 +1059,10 @@ int mlx5e_create_rq(struct mlx5e_rq *rq, struct mlx5e_rq_param *param);
 void mlx5e_destroy_rq(struct mlx5e_rq *rq);
 
 struct mlx5e_sq_param;
+int mlx5e_open_icosq(struct mlx5e_channel *c, struct mlx5e_params *params,
+		     struct mlx5e_sq_param *param, struct mlx5e_icosq *sq,
+		     work_func_t recover_work_func);
+void mlx5e_close_icosq(struct mlx5e_icosq *sq);
 int mlx5e_open_xdpsq(struct mlx5e_channel *c, struct mlx5e_params *params,
 		     struct mlx5e_sq_param *param, struct xsk_buff_pool *xsk_pool,
 		     struct mlx5e_xdpsq *sq, bool is_redirect);
drivers/net/ethernet/mellanox/mlx5/core/en/rx_res.c +28 -0
index e1095bc36543..4a88b675a02c 100644
@@ -611,6 +611,34 @@ struct mlx5e_rss_params_hash mlx5e_rx_res_get_current_hash(struct mlx5e_rx_res *
 	return mlx5e_rss_get_hash(res->rss[0]);
 }
 
+int mlx5e_rx_res_nvmeotcp_tir_create(struct mlx5e_rx_res *res, unsigned int rxq, bool crc_rx,
+				     u32 tag_buf_id, struct mlx5e_tir *tir)
+{
+	bool inner_ft_support = res->features & MLX5E_RX_RES_FEATURE_INNER_FT;
+	struct mlx5e_tir_builder *builder;
+	u32 rqtn;
+	int err;
+
+	builder = mlx5e_tir_builder_alloc(false);
+	if (!builder)
+		return -ENOMEM;
+
+	rqtn = mlx5e_rx_res_get_rqtn_direct(res, rxq);
+
+	mlx5e_tir_builder_build_rqt(builder, res->mdev->mlx5e_res.hw_objs.td.tdn, rqtn,
+				    inner_ft_support);
+	mlx5e_tir_builder_build_direct(builder);
+	mlx5e_tir_builder_build_nvmeotcp(builder, crc_rx, tag_buf_id);
+	down_read(&res->pkt_merge_param_sem);
+	mlx5e_tir_builder_build_packet_merge(builder, &res->pkt_merge_param);
+	err = mlx5e_tir_init(tir, builder, res->mdev, false);
+	up_read(&res->pkt_merge_param_sem);
+
+	mlx5e_tir_builder_free(builder);
+
+	return err;
+}
+
 int mlx5e_rx_res_tls_tir_create(struct mlx5e_rx_res *res, unsigned int rxq,
 				struct mlx5e_tir *tir)
 {
drivers/net/ethernet/mellanox/mlx5/core/en/rx_res.h +4 -0
index 5d5f64fab60f..59c22cac9ef4 100644
@@ -66,4 +66,8 @@ struct mlx5e_rss_params_hash mlx5e_rx_res_get_current_hash(struct mlx5e_rx_res *
 /* Accel TIRs */
 int mlx5e_rx_res_tls_tir_create(struct mlx5e_rx_res *res, unsigned int rxq,
 				struct mlx5e_tir *tir);
+
+int mlx5e_rx_res_nvmeotcp_tir_create(struct mlx5e_rx_res *res, unsigned int rxq, bool crc_rx,
+				     u32 tag_buf_id, struct mlx5e_tir *tir);
+
 #endif /* __MLX5_EN_RX_RES_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en/tir.c +15 -0
index d4239e3b3c88..8bdf74cbd8cd 100644
@@ -143,6 +143,21 @@ void mlx5e_tir_builder_build_direct(struct mlx5e_tir_builder *builder)
 	MLX5_SET(tirc, tirc, rx_hash_fn, MLX5_RX_HASH_FN_INVERTED_XOR8);
 }
 
+void mlx5e_tir_builder_build_nvmeotcp(struct mlx5e_tir_builder *builder, bool crc_rx,
+				      u32 tag_buf_id)
+{
+	void *tirc = mlx5e_tir_builder_get_tirc(builder);
+
+	WARN_ON(builder->modify);
+
+	MLX5_SET(tirc, tirc, nvmeotcp_zero_copy_en, 1);
+	MLX5_SET(tirc, tirc, nvmeotcp_tag_buffer_table_id, tag_buf_id);
+	MLX5_SET(tirc, tirc, nvmeotcp_crc_en, !!crc_rx);
+	MLX5_SET(tirc, tirc, self_lb_block,
+		 MLX5_TIRC_SELF_LB_BLOCK_BLOCK_UNICAST |
+		 MLX5_TIRC_SELF_LB_BLOCK_BLOCK_MULTICAST);
+}
+
 void mlx5e_tir_builder_build_tls(struct mlx5e_tir_builder *builder)
 {
 	void *tirc = mlx5e_tir_builder_get_tirc(builder);
drivers/net/ethernet/mellanox/mlx5/core/en/tir.h +2 -0
index 857a84bcd53a..bdec6931444b 100644
@@ -35,6 +35,8 @@ void mlx5e_tir_builder_build_rss(struct mlx5e_tir_builder *builder,
 				 bool inner);
 void mlx5e_tir_builder_build_direct(struct mlx5e_tir_builder *builder);
 void mlx5e_tir_builder_build_tls(struct mlx5e_tir_builder *builder);
+void mlx5e_tir_builder_build_nvmeotcp(struct mlx5e_tir_builder *builder, bool crc_rx,
+				      u32 tag_buf_id);
 
 struct mlx5_core_dev;
 
drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h +6 -0
index 58be28376ca5..31e240610e48 100644
@@ -52,6 +52,7 @@ enum mlx5e_icosq_wqe_type {
 #endif
 #ifdef CONFIG_MLX5_EN_NVMEOTCP
 	MLX5E_ICOSQ_WQE_UMR_NVMEOTCP,
+	MLX5E_ICOSQ_WQE_SET_PSV_NVMEOTCP,
 #endif
 };
 
@@ -211,6 +212,11 @@ struct mlx5e_icosq_wqe_info {
 		struct {
 			struct mlx5e_ktls_rx_resync_buf *buf;
 		} tls_get_params;
+#endif
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+		struct {
+			struct mlx5e_nvmeotcp_queue *queue;
+		} nvmeotcp_q;
 #endif
 	};
 };
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c +558 -4

638 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.h +4 -0
index a665b7a72bc2..555f3ed7e2e2 100644
@@ -106,6 +106,10 @@ void mlx5e_nvmeotcp_build_netdev(struct mlx5e_priv *priv);
 int mlx5e_nvmeotcp_init(struct mlx5e_priv *priv);
 int set_ulp_ddp_nvme_tcp(struct net_device *netdev, bool enable);
 void mlx5e_nvmeotcp_cleanup(struct mlx5e_priv *priv);
+struct mlx5e_nvmeotcp_queue *
+mlx5e_nvmeotcp_get_queue(struct mlx5e_nvmeotcp *nvmeotcp, int id);
+void mlx5e_nvmeotcp_put_queue(struct mlx5e_nvmeotcp_queue *queue);
+void mlx5e_nvmeotcp_ctx_complete(struct mlx5e_icosq_wqe_info *wi);
 static inline void mlx5e_nvmeotcp_init_rx(struct mlx5e_priv *priv) {}
 void mlx5e_nvmeotcp_cleanup_rx(struct mlx5e_priv *priv);
 extern const struct ulp_ddp_dev_ops mlx5e_nvmeotcp_ops;
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_utils.h +41 -0

57 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_main.c +4 -4
index e582e97e7c44..f227dd9f85ac 100644
@@ -1740,9 +1740,9 @@ void mlx5e_tx_err_cqe_work(struct work_struct *recover_work)
 	mlx5e_reporter_tx_err_cqe(sq);
 }
 
-static int mlx5e_open_icosq(struct mlx5e_channel *c, struct mlx5e_params *params,
-			    struct mlx5e_sq_param *param, struct mlx5e_icosq *sq,
-			    work_func_t recover_work_func)
+int mlx5e_open_icosq(struct mlx5e_channel *c, struct mlx5e_params *params,
+		     struct mlx5e_sq_param *param, struct mlx5e_icosq *sq,
+		     work_func_t recover_work_func)
 {
 	struct mlx5e_create_sq_param csp = {};
 	int err;
@@ -1786,7 +1786,7 @@ void mlx5e_deactivate_icosq(struct mlx5e_icosq *icosq)
 	synchronize_net(); /* Sync with NAPI. */
 }
 
-static void mlx5e_close_icosq(struct mlx5e_icosq *sq)
+void mlx5e_close_icosq(struct mlx5e_icosq *sq)
 {
 	if (sq->ktls_resync)
 		mlx5e_ktls_rx_resync_destroy_resp_list(sq->ktls_resync);
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +13 -2
index 3ae2c3f09476..253aa3f59f20 100644
@@ -53,6 +53,7 @@
 #include "en_accel/macsec.h"
 #include "en_accel/ipsec_rxtx.h"
 #include "en_accel/ktls_txrx.h"
+#include "en_accel/nvmeotcp.h"
 #include "en/xdp.h"
 #include "en/xsk/rx.h"
 #include "en/health.h"
@@ -877,16 +878,23 @@ void mlx5e_free_icosq_descs(struct mlx5e_icosq *sq)
 		ci = mlx5_wq_cyc_ctr2ix(&sq->wq, sqcc);
 		wi = &sq->db.wqe_info[ci];
 		sqcc += wi->num_wqebbs;
-#ifdef CONFIG_MLX5_EN_TLS
 		switch (wi->wqe_type) {
+#ifdef CONFIG_MLX5_EN_TLS
 		case MLX5E_ICOSQ_WQE_SET_PSV_TLS:
 			mlx5e_ktls_handle_ctx_completion(wi);
 			break;
 		case MLX5E_ICOSQ_WQE_GET_PSV_TLS:
 			mlx5e_ktls_handle_get_psv_completion(wi, sq);
 			break;
-		}
 #endif
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+		case MLX5E_ICOSQ_WQE_SET_PSV_NVMEOTCP:
+			mlx5e_nvmeotcp_ctx_complete(wi);
+			break;
+#endif
+		default:
+			break;
+		}
 	}
 	sq->cc = sqcc;
 }
@@ -985,6 +993,9 @@ int mlx5e_poll_ico_cq(struct mlx5e_cq *cq, int budget)
 #ifdef CONFIG_MLX5_EN_NVMEOTCP
 			case MLX5E_ICOSQ_WQE_UMR_NVMEOTCP:
 				break;
+			case MLX5E_ICOSQ_WQE_SET_PSV_NVMEOTCP:
+				mlx5e_nvmeotcp_ctx_complete(wi);
+				break;
 #endif
 			default:
 				netdev_WARN_ONCE(cq->netdev,
Signature · 1 line
2.31.1
#20 lore raw

PATCH v11 16/25 net/mlx5e: Refactor doorbell function to allow avoiding a completion

In reply to Aurelien Aptel

From: Or Gerlitz <[email protected]>

Currently the doorbell function always asks for completion to be generated.

Refactor things such that all existing call sites are untouched and no branching is added. This is done using inner function which can be invoked directly in cases completion is not desired (as done in downstream patch).

No functional change here.

---
 drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h | 13 ++++++++++---
 1 file changed, 10 insertions(+), 3 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h +10 -3
index 92a71d1a7605..37217e4cc118 100644
@@ -243,10 +243,10 @@ static inline u16 mlx5e_icosq_get_next_pi(struct mlx5e_icosq *sq, u16 size)
 }
 
 static inline void
-mlx5e_notify_hw(struct mlx5_wq_cyc *wq, u16 pc, void __iomem *uar_map,
-		struct mlx5_wqe_ctrl_seg *ctrl)
+__mlx5e_notify_hw(struct mlx5_wq_cyc *wq, u16 pc, void __iomem *uar_map,
+		  struct mlx5_wqe_ctrl_seg *ctrl, u8 cq_update)
 {
-	ctrl->fm_ce_se |= MLX5_WQE_CTRL_CQ_UPDATE;
+	ctrl->fm_ce_se |= cq_update;
 	/* ensure wqe is visible to device before updating doorbell record */
 	dma_wmb();
 
@@ -260,6 +260,13 @@ mlx5e_notify_hw(struct mlx5_wq_cyc *wq, u16 pc, void __iomem *uar_map,
 	mlx5_write64((__be32 *)ctrl, uar_map);
 }
 
+static inline void
+mlx5e_notify_hw(struct mlx5_wq_cyc *wq, u16 pc, void __iomem *uar_map,
+		struct mlx5_wqe_ctrl_seg *ctrl)
+{
+	__mlx5e_notify_hw(wq, pc, uar_map, ctrl, MLX5_WQE_CTRL_CQ_UPDATE);
+}
+
 static inline void mlx5e_cq_arm(struct mlx5e_cq *cq)
 {
 	struct mlx5_core_cq *mcq;
Signature · 1 line
2.31.1
#21 lore raw

PATCH v11 17/25 net/mlx5: Add NVMEoTCP caps, HW bits, 128B CQE and enumerations

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

Add the necessary infrastructure for NVMEoTCP offload:
- Create mlx5_cqe128 structure for NVMEoTCP offload. The new structure consist from the regular mlx5_cqe64 + NVMEoTCP data information for offloaded packets.
- Add nvmetcp field to mlx5_cqe64, this field define the type of the data that the additional NVMEoTCP part represents.
- Add nvmeotcp_zero_copy_en + nvmeotcp_crc_en bit to the TIR, for identify NVMEoTCP offload flow and tag_buffer_id that will be used by the connected nvmeotcp_queues.
- Add new capability to HCA_CAP that represents the NVMEoTCP offload ability.

---
 drivers/net/ethernet/mellanox/mlx5/core/fw.c |  6 ++
 include/linux/mlx5/device.h                  | 51 +++++++++++++-
 include/linux/mlx5/mlx5_ifc.h                | 74 ++++++++++++++++++--
 include/linux/mlx5/qp.h                      |  1 +
 4 files changed, 127 insertions(+), 5 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/fw.c +6 -0
index 7bb7be01225a..5f87f5c82485 100644
@@ -292,6 +292,12 @@ int mlx5_query_hca_caps(struct mlx5_core_dev *dev)
 			return err;
 	}
 
+	if (MLX5_CAP_GEN(dev, nvmeotcp)) {
+		err = mlx5_core_get_caps(dev, MLX5_CAP_DEV_NVMEOTCP);
+		if (err)
+			return err;
+	}
+
 	return 0;
 }
 
include/linux/mlx5/device.h +50 -1

87 lines not on this page · raw message

include/linux/mlx5/mlx5_ifc.h +70 -4

139 lines not on this page · raw message

include/linux/mlx5/qp.h +1 -0
index 4657d5c54abe..bda53b241d71 100644
@@ -227,6 +227,7 @@ struct mlx5_wqe_ctrl_seg {
 #define MLX5_WQE_CTRL_OPCODE_MASK 0xff
 #define MLX5_WQE_CTRL_WQE_INDEX_MASK 0x00ffff00
 #define MLX5_WQE_CTRL_WQE_INDEX_SHIFT 8
+#define MLX5_WQE_CTRL_TIR_TIS_INDEX_SHIFT 8
 
 enum {
 	MLX5_ETH_WQE_L3_INNER_CSUM      = 1 << 4,
Signature · 1 line
2.31.1
#29 lore raw

PATCH v11 25/25 net/mlx5e: NVMEoTCP, statistics

In reply to Aurelien Aptel

NVMEoTCP offload statistics include both control and data path
statistic: counters for the netdev ddp ops, offloaded packets/bytes, resync and dropped packets.

Expose the statistics using ulp_ddp_ops->get_stats() instead of the regular statistics flow.

---
 .../net/ethernet/mellanox/mlx5/core/Makefile  |  3 +-
 .../mellanox/mlx5/core/en_accel/nvmeotcp.c    | 53 ++++++++++++---
 .../mellanox/mlx5/core/en_accel/nvmeotcp.h    | 16 +++++
 .../mlx5/core/en_accel/nvmeotcp_rxtx.c        | 11 +++-
 .../mlx5/core/en_accel/nvmeotcp_stats.c       | 66 +++++++++++++++++++
 .../ethernet/mellanox/mlx5/core/en_stats.h    |  7 ++
 6 files changed, 145 insertions(+), 11 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_stats.c
drivers/net/ethernet/mellanox/mlx5/core/Makefile +2 -1
index 9804bd086bf4..d48bde4ca8de 100644
@@ -103,7 +103,8 @@ mlx5_core-$(CONFIG_MLX5_EN_TLS) += en_accel/ktls_stats.o \
 				   en_accel/fs_tcp.o en_accel/ktls.o en_accel/ktls_txrx.o \
 				   en_accel/ktls_tx.o en_accel/ktls_rx.o
 
-mlx5_core-$(CONFIG_MLX5_EN_NVMEOTCP) += en_accel/fs_tcp.o en_accel/nvmeotcp.o en_accel/nvmeotcp_rxtx.o
+mlx5_core-$(CONFIG_MLX5_EN_NVMEOTCP) += en_accel/fs_tcp.o en_accel/nvmeotcp.o \
+					en_accel/nvmeotcp_rxtx.o en_accel/nvmeotcp_stats.o
 
 mlx5_core-$(CONFIG_MLX5_SW_STEERING) += steering/dr_domain.o steering/dr_table.o \
 					steering/dr_matcher.o steering/dr_rule.o \
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c +44 -9

139 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.h +16 -0
index a5cfd9e31be7..f2a7f3cc945d 100644
@@ -9,6 +9,15 @@
 #include "en.h"
 #include "en/params.h"
 
+struct mlx5e_nvmeotcp_sw_stats {
+	atomic64_t rx_nvmeotcp_sk_add;
+	atomic64_t rx_nvmeotcp_sk_add_fail;
+	atomic64_t rx_nvmeotcp_sk_del;
+	atomic64_t rx_nvmeotcp_ddp_setup;
+	atomic64_t rx_nvmeotcp_ddp_setup_fail;
+	atomic64_t rx_nvmeotcp_ddp_teardown;
+};
+
 struct mlx5e_nvmeotcp_queue_entry {
 	struct mlx5e_nvmeotcp_queue *queue;
 	u32 sgl_length;
@@ -52,6 +61,7 @@ struct mlx5e_nvmeotcp_queue_handler {
  *	@sk: The socket used by the NVMe-TCP queue
  *	@crc_rx: CRC Rx offload indication for this queue
  *	@priv: mlx5e netdev priv
+ *	@sw_stats: Global software statistics for nvmeotcp offload
  *	@static_params_done: Async completion structure for the initial umr mapping
  *	synchronization
  *	@sq_lock: Spin lock for the icosq
@@ -88,6 +98,7 @@ struct mlx5e_nvmeotcp_queue {
 	u8 crc_rx:1;
 	/* for ddp invalidate flow */
 	struct mlx5e_priv *priv;
+	struct mlx5e_nvmeotcp_sw_stats *sw_stats;
 	/* end of data-path section */
 
 	struct completion static_params_done;
@@ -97,6 +108,7 @@ struct mlx5e_nvmeotcp_queue {
 };
 
 struct mlx5e_nvmeotcp {
+	struct mlx5e_nvmeotcp_sw_stats sw_stats;
 	struct ida queue_ids;
 	struct rhashtable queue_hash;
 	bool enabled;
@@ -113,6 +125,7 @@ void mlx5e_nvmeotcp_ddp_inv_done(struct mlx5e_icosq_wqe_info *wi);
 void mlx5e_nvmeotcp_ctx_complete(struct mlx5e_icosq_wqe_info *wi);
 static inline void mlx5e_nvmeotcp_init_rx(struct mlx5e_priv *priv) {}
 void mlx5e_nvmeotcp_cleanup_rx(struct mlx5e_priv *priv);
+int mlx5e_nvmeotcp_get_stats(struct mlx5e_priv *priv, struct ethtool_ulp_ddp_stats *stats);
 extern const struct ulp_ddp_dev_ops mlx5e_nvmeotcp_ops;
 #else
 
@@ -122,5 +135,8 @@ static inline void mlx5e_nvmeotcp_cleanup(struct mlx5e_priv *priv) {}
 static inline int set_ulp_ddp_nvme_tcp(struct net_device *dev, bool en) { return -EOPNOTSUPP; }
 static inline void mlx5e_nvmeotcp_init_rx(struct mlx5e_priv *priv) {}
 static inline void mlx5e_nvmeotcp_cleanup_rx(struct mlx5e_priv *priv) {}
+static inline int mlx5e_nvmeotcp_get_stats(struct mlx5e_priv *priv,
+					   struct ethtool_ulp_ddp_stats *stats)
+{ return 0; }
 #endif
 #endif /* __MLX5E_NVMEOTCP_H__ */
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_rxtx.c +10 -1
index 4c7dab28ef56..d703ca3f0155 100644
@@ -111,6 +111,7 @@ mlx5e_nvmeotcp_rebuild_rx_skb_nonlinear(struct mlx5e_rq *rq, struct sk_buff *skb
 	int ccoff, cclen, hlen, ccid, remaining, fragsz, to_copy = 0;
 	struct net_device *netdev = rq->netdev;
 	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_rq_stats *stats = rq->stats;
 	struct mlx5e_nvmeotcp_queue_entry *nqe;
 	skb_frag_t org_frags[MAX_SKB_FRAGS];
 	struct mlx5e_nvmeotcp_queue *queue;
@@ -122,12 +123,14 @@ mlx5e_nvmeotcp_rebuild_rx_skb_nonlinear(struct mlx5e_rq *rq, struct sk_buff *skb
 	queue = mlx5e_nvmeotcp_get_queue(priv->nvmeotcp, queue_id);
 	if (unlikely(!queue)) {
 		dev_kfree_skb_any(skb);
+		stats->nvmeotcp_drop++;
 		return false;
 	}
 
 	cqe128 = container_of(cqe, struct mlx5e_cqe128, cqe64);
 	if (cqe_is_nvmeotcp_resync(cqe)) {
 		nvmeotcp_update_resync(queue, cqe128);
+		stats->nvmeotcp_resync++;
 		mlx5e_nvmeotcp_put_queue(queue);
 		return true;
 	}
@@ -201,7 +204,8 @@ mlx5e_nvmeotcp_rebuild_rx_skb_nonlinear(struct mlx5e_rq *rq, struct sk_buff *skb
 						 org_nr_frags,
 						 frag_index);
 	}
-
+	stats->nvmeotcp_packets++;
+	stats->nvmeotcp_bytes += cclen;
 	mlx5e_nvmeotcp_put_queue(queue);
 	return true;
 }
@@ -213,6 +217,7 @@ mlx5e_nvmeotcp_rebuild_rx_skb_linear(struct mlx5e_rq *rq, struct sk_buff *skb,
 	int ccoff, cclen, hlen, ccid, remaining, fragsz, to_copy = 0;
 	struct net_device *netdev = rq->netdev;
 	struct mlx5e_priv *priv = netdev_priv(netdev);
+	struct mlx5e_rq_stats *stats = rq->stats;
 	struct mlx5e_nvmeotcp_queue_entry *nqe;
 	struct mlx5e_nvmeotcp_queue *queue;
 	struct mlx5e_cqe128 *cqe128;
@@ -222,12 +227,14 @@ mlx5e_nvmeotcp_rebuild_rx_skb_linear(struct mlx5e_rq *rq, struct sk_buff *skb,
 	queue = mlx5e_nvmeotcp_get_queue(priv->nvmeotcp, queue_id);
 	if (unlikely(!queue)) {
 		dev_kfree_skb_any(skb);
+		stats->nvmeotcp_drop++;
 		return false;
 	}
 
 	cqe128 = container_of(cqe, struct mlx5e_cqe128, cqe64);
 	if (cqe_is_nvmeotcp_resync(cqe)) {
 		nvmeotcp_update_resync(queue, cqe128);
+		stats->nvmeotcp_resync++;
 		mlx5e_nvmeotcp_put_queue(queue);
 		return true;
 	}
@@ -301,6 +308,8 @@ mlx5e_nvmeotcp_rebuild_rx_skb_linear(struct mlx5e_rq *rq, struct sk_buff *skb,
 				       hlen + cclen, remaining);
 	}
 
+	stats->nvmeotcp_packets++;
+	stats->nvmeotcp_bytes += cclen;
 	mlx5e_nvmeotcp_put_queue(queue);
 	return true;
 }
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp_stats.c new file +66 -0

66 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_stats.h +7 -0
index 375752d6546d..8b3d43518d05 100644
@@ -128,6 +128,7 @@ void mlx5e_stats_rmon_get(struct mlx5e_priv *priv,
 			  const struct ethtool_rmon_hist_range **ranges);
 void mlx5e_get_link_ext_stats(struct net_device *dev,
 			      struct ethtool_link_ext_stats *stats);
+void mlx5e_stats_ulp_ddp_get(struct mlx5e_priv *priv, struct ethtool_ulp_ddp_stats *stats);
 
 /* Concrete NIC Stats */
 
@@ -395,6 +396,12 @@ struct mlx5e_rq_stats {
 	u64 tls_resync_res_skip;
 	u64 tls_err;
 #endif
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+	u64 nvmeotcp_drop;
+	u64 nvmeotcp_resync;
+	u64 nvmeotcp_packets;
+	u64 nvmeotcp_bytes;
+#endif
 };
 
 struct mlx5e_sq_stats {
Signature · 1 line
2.31.1
#26 lore raw

PATCH v11 22/25 net/mlx5e: NVMEoTCP, ddp setup and resync

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

NVMEoTCP offload uses buffer registration for every NVME request to perform direct data placement. This is achieved by creating a NIC HW mapping between the CCID (command capsule ID) to the set of buffers that compose the request. The registration is implemented via MKEY for which we do fast/async mapping using KLM UMR WQE.

The buffer registration takes place when the ULP calls the ddp_setup op which is done before they send their corresponding request to the other side (e.g nvmf target). We don't wait for the completion of the registration before returning back to the ulp. The reason being that the HW mapping should be in place fast enough vs the RTT it would take for the request to be responded. If this doesn't happen, some IO may not be ddp-offloaded, but that doesn't stop the overall offloading session.

When the offloading HW gets out of sync with the protocol session, a hardware/software handshake takes place to resync. The ddp_resync op is the part of the handshake where the SW confirms to the HW that a indeed they identified correctly a PDU header at a certain TCP sequence number. This allows the HW to resume the offload.

The 1st part of the handshake is when the HW identifies such sequence number in an arriving packet. A special mark is made on the completion (cqe) and then the mlx5 driver invokes the ddp resync_request callback advertised by the ULP in the ddp context - this is in downstream patch.

---
 .../mellanox/mlx5/core/en_accel/nvmeotcp.c    | 146 +++++++++++++++++-
 1 file changed, 144 insertions(+), 2 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c +144 -2
index 64e4b3d2936a..2b9d69254ead 100644
@@ -682,19 +682,156 @@ mlx5e_nvmeotcp_queue_teardown(struct net_device *netdev,
 	mlx5e_nvmeotcp_put_queue(queue);
 }
 
+static bool
+mlx5e_nvmeotcp_validate_small_sgl_suffix(struct scatterlist *sg, int sg_len, int mtu)
+{
+	int i, hole_size, hole_len, chunk_size = 0;
+
+	for (i = 1; i < sg_len; i++)
+		chunk_size += sg_dma_len(&sg[i]);
+
+	if (chunk_size >= mtu)
+		return true;
+
+	hole_size = mtu - chunk_size - 1;
+	hole_len = DIV_ROUND_UP(hole_size, PAGE_SIZE);
+
+	if (sg_len + hole_len > MAX_SKB_FRAGS)
+		return false;
+
+	return true;
+}
+
+static bool
+mlx5e_nvmeotcp_validate_big_sgl_suffix(struct scatterlist *sg, int sg_len, int mtu)
+{
+	int i, j, last_elem, window_idx, window_size = MAX_SKB_FRAGS - 1;
+	int chunk_size = 0;
+
+	last_elem = sg_len - window_size;
+	window_idx = window_size;
+
+	for (j = 1; j < window_size; j++)
+		chunk_size += sg_dma_len(&sg[j]);
+
+	for (i = 1; i <= last_elem; i++, window_idx++) {
+		chunk_size += sg_dma_len(&sg[window_idx]);
+		if (chunk_size < mtu - 1)
+			return false;
+
+		chunk_size -= sg_dma_len(&sg[i]);
+	}
+
+	return true;
+}
+
+/* This function makes sure that the middle/suffix of a PDU SGL meets the
+ * restriction of MAX_SKB_FRAGS. There are two cases here:
+ * 1. sg_len < MAX_SKB_FRAGS - the extreme case here is a packet that consists
+ * of one byte from the first SG element + the rest of the SGL and the remaining
+ * space of the packet will be scattered to the WQE and will be pointed by
+ * SKB frags.
+ * 2. sg_len => MAX_SKB_FRAGS - the extreme case here is a packet that consists
+ * of one byte from middle SG element + 15 continuous SG elements + one byte
+ * from a sequential SG element or the rest of the packet.
+ */
+static bool
+mlx5e_nvmeotcp_validate_sgl_suffix(struct scatterlist *sg, int sg_len, int mtu)
+{
+	int ret;
+
+	if (sg_len < MAX_SKB_FRAGS)
+		ret = mlx5e_nvmeotcp_validate_small_sgl_suffix(sg, sg_len, mtu);
+	else
+		ret = mlx5e_nvmeotcp_validate_big_sgl_suffix(sg, sg_len, mtu);
+
+	return ret;
+}
+
+static bool
+mlx5e_nvmeotcp_validate_sgl_prefix(struct scatterlist *sg, int sg_len, int mtu)
+{
+	int i, hole_size, hole_len, tmp_len, chunk_size = 0;
+
+	tmp_len = min_t(int, sg_len, MAX_SKB_FRAGS);
+
+	for (i = 0; i < tmp_len; i++)
+		chunk_size += sg_dma_len(&sg[i]);
+
+	if (chunk_size >= mtu)
+		return true;
+
+	hole_size = mtu - chunk_size;
+	hole_len = DIV_ROUND_UP(hole_size, PAGE_SIZE);
+
+	if (tmp_len + hole_len > MAX_SKB_FRAGS)
+		return false;
+
+	return true;
+}
+
+/* This function is responsible to ensure that a PDU could be offloaded.
+ * PDU is offloaded by building a non-linear SKB such that each SGL element is
+ * placed in frag, thus this function should ensure that all packets that
+ * represent part of the PDU won't exaggerate from MAX_SKB_FRAGS SGL.
+ * In addition NVMEoTCP offload has one PDU offload for packet restriction.
+ * Packet could start with a new PDU and then we should check that the prefix
+ * of the PDU meets the requirement or a packet can start in the middle of SG
+ * element and then we should check that the suffix of PDU meets the requirement.
+ */
+static bool
+mlx5e_nvmeotcp_validate_sgl(struct scatterlist *sg, int sg_len, int mtu)
+{
+	int max_hole_frags;
+
+	max_hole_frags = DIV_ROUND_UP(mtu, PAGE_SIZE);
+	if (sg_len + max_hole_frags <= MAX_SKB_FRAGS)
+		return true;
+
+	if (!mlx5e_nvmeotcp_validate_sgl_prefix(sg, sg_len, mtu) ||
+	    !mlx5e_nvmeotcp_validate_sgl_suffix(sg, sg_len, mtu))
+		return false;
+
+	return true;
+}
+
 static int
 mlx5e_nvmeotcp_ddp_setup(struct net_device *netdev,
 			 struct sock *sk,
 			 struct ulp_ddp_io *ddp)
 {
+	struct scatterlist *sg = ddp->sg_table.sgl;
+	struct mlx5e_nvmeotcp_queue_entry *nvqt;
 	struct mlx5e_nvmeotcp_queue *queue;
+	struct mlx5_core_dev *mdev;
+	int i, size = 0, count = 0;
 
 	queue = container_of(ulp_ddp_get_ctx(sk),
 			     struct mlx5e_nvmeotcp_queue, ulp_ddp_ctx);
+	mdev = queue->priv->mdev;
+	count = dma_map_sg(mdev->device, ddp->sg_table.sgl, ddp->nents,
+			   DMA_FROM_DEVICE);
+
+	if (count <= 0)
+		return -EINVAL;
 
-	/* Placeholder - map_sg and initializing the count */
+	if (WARN_ON(count > mlx5e_get_max_sgl(mdev)))
+		return -ENOSPC;
+
+	if (!mlx5e_nvmeotcp_validate_sgl(sg, count, READ_ONCE(netdev->mtu)))
+		return -EOPNOTSUPP;
+
+	for (i = 0; i < count; i++)
+		size += sg_dma_len(&sg[i]);
+
+	nvqt = &queue->ccid_table[ddp->command_id];
+	nvqt->size = size;
+	nvqt->ddp = ddp;
+	nvqt->sgl = sg;
+	nvqt->ccid_gen++;
+	nvqt->sgl_length = count;
+	mlx5e_nvmeotcp_post_klm_wqe(queue, KLM_UMR, ddp->command_id, count);
 
-	mlx5e_nvmeotcp_post_klm_wqe(queue, KLM_UMR, ddp->command_id, 0);
 	return 0;
 }
 
@@ -717,6 +854,11 @@ static void
 mlx5e_nvmeotcp_ddp_resync(struct net_device *netdev,
 			  struct sock *sk, u32 seq)
 {
+	struct mlx5e_nvmeotcp_queue *queue =
+		container_of(ulp_ddp_get_ctx(sk), struct mlx5e_nvmeotcp_queue, ulp_ddp_ctx);
+
+	queue->after_resync_cqe = 1;
+	mlx5e_nvmeotcp_rx_post_static_params_wqe(queue, seq);
 }
 
 struct mlx5e_nvmeotcp_queue *
Signature · 1 line
2.31.1
#22 lore raw

PATCH v11 18/25 net/mlx5e: NVMEoTCP, offload initialization

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

This commit introduces the driver structures and initialization blocks for NVMEoTCP offload. The mlx5 nvmeotcp structures are:

- queue (mlx5e_nvmeotcp_queue) - pairs 1:1 with nvmeotcp driver queues and deals with the offloading parts. The mlx5e queue is accessed in the ddp
ops: initialized on sk_add, used in ddp setup,teardown,resync and in the fast path when dealing with packets, destroyed in the sk_del op.

- queue entry (nvmeotcp_queue_entry) - pairs 1:1 with offloaded IO from that queue. Keeps pointers to the SG elements describing the buffers used for the IO and the ddp context of it.

- queue handler (mlx5e_nvmeotcp_queue_handler) - we use icosq per NVME-TCP queue for UMR mapping as part of the ddp offload. Those dedicated SQs are unique in the sense that they are driven directly by the NVME-TCP layer to submit and invalidate ddp requests.
Since the life-cycle of these icosqs is not tied to the channels, we create dedicated napi contexts for polling them such that channels can be re-created during offloading. The queue handler has pointer to the cq associated with the queue's sq and napi context.

- main offload context (mlx5e_nvmeotcp) - has ida and hash table instances. Each offloaded queue gets an ID from the ida instance and the <id, queue> pairs are kept in the hash table. The id is programmed as flow tag to be set by HW on the completion (cqe) of all packets related to this queue (by 5-tuple steering). The fast path which deals with packets uses the flow tag to access the hash table and retrieve the queue for the processing.

We query nvmeotcp capabilities to see if the offload can be supported and use 128B CQEs when this happens. By default, the offload is off but can be enabled with ethtool --ulp-ddp <device> nvme-tcp-ddp on.

---
 .../net/ethernet/mellanox/mlx5/core/Kconfig   |  11 +
 .../net/ethernet/mellanox/mlx5/core/Makefile  |   2 +
 drivers/net/ethernet/mellanox/mlx5/core/en.h  |   4 +
 .../net/ethernet/mellanox/mlx5/core/en/fs.h   |   4 +-
 .../ethernet/mellanox/mlx5/core/en/params.c   |  12 +-
 .../ethernet/mellanox/mlx5/core/en/params.h   |   3 +
 .../mellanox/mlx5/core/en_accel/en_accel.h    |   3 +
 .../mellanox/mlx5/core/en_accel/fs_tcp.h      |   2 +-
 .../mellanox/mlx5/core/en_accel/nvmeotcp.c    | 216 ++++++++++++++++++
 .../mellanox/mlx5/core/en_accel/nvmeotcp.h    | 121 ++++++++++
 .../ethernet/mellanox/mlx5/core/en_ethtool.c  |   6 +
 .../net/ethernet/mellanox/mlx5/core/en_fs.c   |   4 +-
 .../net/ethernet/mellanox/mlx5/core/en_main.c |  18 ++
 .../net/ethernet/mellanox/mlx5/core/main.c    |   1 +
 14 files changed, 398 insertions(+), 9 deletions(-)
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c
 create mode 100644 drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.h
drivers/net/ethernet/mellanox/mlx5/core/Kconfig +11 -0
index 26685fd0fdaa..0c790952fdf7 100644
@@ -163,6 +163,17 @@ config MLX5_EN_TLS
 	help
 	Build support for TLS cryptography-offload acceleration in the NIC.
 
+config MLX5_EN_NVMEOTCP
+	bool "NVMEoTCP acceleration"
+	depends on ULP_DDP
+	depends on MLX5_CORE_EN
+	default y
+	help
+	Build support for NVMEoTCP acceleration in the NIC.
+	This includes Direct Data Placement and CRC offload.
+	Note: Support for hardware with this capability needs to be selected
+	for this option to become available.
+
 config MLX5_SW_STEERING
 	bool "Mellanox Technologies software-managed steering"
 	depends on MLX5_CORE_EN && MLX5_ESWITCH
drivers/net/ethernet/mellanox/mlx5/core/Makefile +2 -0
index cd4a1ab0ea78..9df9999047d1 100644
@@ -103,6 +103,8 @@ mlx5_core-$(CONFIG_MLX5_EN_TLS) += en_accel/ktls_stats.o \
 				   en_accel/fs_tcp.o en_accel/ktls.o en_accel/ktls_txrx.o \
 				   en_accel/ktls_tx.o en_accel/ktls_rx.o
 
+mlx5_core-$(CONFIG_MLX5_EN_NVMEOTCP) += en_accel/fs_tcp.o en_accel/nvmeotcp.o
+
 mlx5_core-$(CONFIG_MLX5_SW_STEERING) += steering/dr_domain.o steering/dr_table.o \
 					steering/dr_matcher.o steering/dr_rule.o \
 					steering/dr_icm_pool.o steering/dr_buddy.o \
drivers/net/ethernet/mellanox/mlx5/core/en.h +4 -0
index dd68f373516b..1f8cf297464d 100644
@@ -328,6 +328,7 @@ struct mlx5e_params {
 	unsigned int sw_mtu;
 	int hard_mtu;
 	bool ptp_rx;
+	bool nvmeotcp;
 };
 
 static inline u8 mlx5e_get_dcb_num_tc(struct mlx5e_params *params)
@@ -960,6 +961,9 @@ struct mlx5e_priv {
 #endif
 #ifdef CONFIG_MLX5_EN_TLS
 	struct mlx5e_tls          *tls;
+#endif
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+	struct mlx5e_nvmeotcp      *nvmeotcp;
 #endif
 	struct devlink_health_reporter *tx_reporter;
 	struct devlink_health_reporter *rx_reporter;
drivers/net/ethernet/mellanox/mlx5/core/en/fs.h +2 -2
index 5233d4daca41..c3598c2e05d7 100644
@@ -77,7 +77,7 @@ enum {
 	MLX5E_INNER_TTC_FT_LEVEL,
 	MLX5E_FS_TT_UDP_FT_LEVEL = MLX5E_INNER_TTC_FT_LEVEL + 1,
 	MLX5E_FS_TT_ANY_FT_LEVEL = MLX5E_INNER_TTC_FT_LEVEL + 1,
-#ifdef CONFIG_MLX5_EN_TLS
+#if defined(CONFIG_MLX5_EN_TLS) || defined(CONFIG_MLX5_EN_NVMEOTCP)
 	MLX5E_ACCEL_FS_TCP_FT_LEVEL = MLX5E_INNER_TTC_FT_LEVEL + 1,
 #endif
 #ifdef CONFIG_MLX5_EN_ARFS
@@ -169,7 +169,7 @@ struct mlx5e_fs_any *mlx5e_fs_get_any(struct mlx5e_flow_steering *fs);
 void mlx5e_fs_set_any(struct mlx5e_flow_steering *fs, struct mlx5e_fs_any *any);
 struct mlx5e_fs_udp *mlx5e_fs_get_udp(struct mlx5e_flow_steering *fs);
 void mlx5e_fs_set_udp(struct mlx5e_flow_steering *fs, struct mlx5e_fs_udp *udp);
-#ifdef CONFIG_MLX5_EN_TLS
+#if defined(CONFIG_MLX5_EN_TLS) || defined(CONFIG_MLX5_EN_NVMEOTCP)
 struct mlx5e_accel_fs_tcp *mlx5e_fs_get_accel_tcp(struct mlx5e_flow_steering *fs);
 void mlx5e_fs_set_accel_tcp(struct mlx5e_flow_steering *fs, struct mlx5e_accel_fs_tcp *accel_tcp);
 #endif
drivers/net/ethernet/mellanox/mlx5/core/en/params.c +8 -4
index a21bd1179477..b34518eaaa1b 100644
@@ -806,7 +806,8 @@ static void mlx5e_build_common_cq_param(struct mlx5_core_dev *mdev,
 	void *cqc = param->cqc;
 
 	MLX5_SET(cqc, cqc, uar_page, mdev->priv.uar->index);
-	if (MLX5_CAP_GEN(mdev, cqe_128_always) && cache_line_size() >= 128)
+	if (MLX5_CAP_GEN(mdev, cqe_128_always) &&
+	    (cache_line_size() >= 128 || param->force_cqe128))
 		MLX5_SET(cqc, cqc, cqe_sz, CQE_STRIDE_128_PAD);
 }
 
@@ -836,6 +837,9 @@ static void mlx5e_build_rx_cq_param(struct mlx5_core_dev *mdev,
 	void *cqc = param->cqc;
 	u8 log_cq_size;
 
+	/* nvme-tcp offload mandates 128 byte cqes */
+	param->force_cqe128 |= IS_ENABLED(CONFIG_MLX5_EN_NVMEOTCP) && params->nvmeotcp;
+
 	switch (params->rq_wq_type) {
 	case MLX5_WQ_TYPE_LINKED_LIST_STRIDING_RQ:
 		hw_stridx = MLX5_CAP_GEN(mdev, mini_cqe_resp_stride_index);
@@ -1175,9 +1179,9 @@ static u8 mlx5e_build_async_icosq_log_wq_sz(struct mlx5_core_dev *mdev)
 	return MLX5E_PARAMS_MINIMUM_LOG_SQ_SIZE;
 }
 
-static void mlx5e_build_icosq_param(struct mlx5_core_dev *mdev,
-				    u8 log_wq_size,
-				    struct mlx5e_sq_param *param)
+void mlx5e_build_icosq_param(struct mlx5_core_dev *mdev,
+			     u8 log_wq_size,
+			     struct mlx5e_sq_param *param)
 {
 	void *sqc = param->sqc;
 	void *wq = MLX5_ADDR_OF(sqc, sqc, wq);
drivers/net/ethernet/mellanox/mlx5/core/en/params.h +3 -0
index c9be6eb88012..d5b3455c4875 100644
@@ -17,6 +17,7 @@ struct mlx5e_cq_param {
 	struct mlx5_wq_param       wq;
 	u16                        eq_ix;
 	u8                         cq_period_mode;
+	bool                       force_cqe128;
 };
 
 struct mlx5e_rq_param {
@@ -146,6 +147,8 @@ void mlx5e_build_xdpsq_param(struct mlx5_core_dev *mdev,
 			     struct mlx5e_params *params,
 			     struct mlx5e_xsk_param *xsk,
 			     struct mlx5e_sq_param *param);
+void mlx5e_build_icosq_param(struct mlx5_core_dev *mdev,
+			     u8 log_wq_size, struct mlx5e_sq_param *param);
 int mlx5e_build_channel_param(struct mlx5_core_dev *mdev,
 			      struct mlx5e_params *params,
 			      u16 q_counter,
drivers/net/ethernet/mellanox/mlx5/core/en_accel/en_accel.h +3 -0
index 07187028f0d3..e38656229399 100644
@@ -40,6 +40,7 @@
 #include "en_accel/ktls.h"
 #include "en_accel/ktls_txrx.h"
 #include <en_accel/macsec.h>
+#include "en_accel/nvmeotcp.h"
 #include "en.h"
 #include "en/txrx.h"
 
@@ -202,11 +203,13 @@ static inline void mlx5e_accel_tx_finish(struct mlx5e_txqsq *sq,
 
 static inline int mlx5e_accel_init_rx(struct mlx5e_priv *priv)
 {
+	mlx5e_nvmeotcp_init_rx(priv);
 	return mlx5e_ktls_init_rx(priv);
 }
 
 static inline void mlx5e_accel_cleanup_rx(struct mlx5e_priv *priv)
 {
+	mlx5e_nvmeotcp_cleanup_rx(priv);
 	mlx5e_ktls_cleanup_rx(priv);
 }
 
drivers/net/ethernet/mellanox/mlx5/core/en_accel/fs_tcp.h +1 -1
index a032bff482a6..d907e352ffae 100644
@@ -6,7 +6,7 @@
 
 #include "en/fs.h"
 
-#ifdef CONFIG_MLX5_EN_TLS
+#if defined(CONFIG_MLX5_EN_TLS) || defined(CONFIG_MLX5_EN_NVMEOTCP)
 int mlx5e_accel_fs_tcp_create(struct mlx5e_flow_steering *fs);
 void mlx5e_accel_fs_tcp_destroy(struct mlx5e_flow_steering *fs);
 struct mlx5_flow_handle *mlx5e_accel_fs_add_sk(struct mlx5e_flow_steering *fs,
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c new file +216 -0

216 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.h new file +121 -0

121 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_ethtool.c +6 -0
index 7708acc9b2ab..cb323a400e6e 100644
@@ -38,6 +38,7 @@
 #include "en/ptp.h"
 #include "lib/clock.h"
 #include "en/fs_ethtool.h"
+#include "en_accel/nvmeotcp.h"
 
 void mlx5e_ethtool_get_drvinfo(struct mlx5e_priv *priv,
 			       struct ethtool_drvinfo *drvinfo)
@@ -1939,6 +1940,11 @@ int mlx5e_modify_rx_cqe_compression_locked(struct mlx5e_priv *priv, bool new_val
 		return -EINVAL;
 	}
 
+	if (priv->channels.params.nvmeotcp) {
+		netdev_warn(priv->netdev, "Can't set CQE compression after ULP DDP NVMe-TCP offload\n");
+		return -EINVAL;
+	}
+
 	new_params = priv->channels.params;
 	MLX5E_SET_PFLAG(&new_params, MLX5E_PFLAG_RX_CQE_COMPRESS, new_val);
 	if (rx_filter)
drivers/net/ethernet/mellanox/mlx5/core/en_fs.c +2 -2
index 7298fe782e9e..63df22d9ad99 100644
@@ -62,7 +62,7 @@ struct mlx5e_flow_steering {
 #ifdef CONFIG_MLX5_EN_ARFS
 	struct mlx5e_arfs_tables       *arfs;
 #endif
-#ifdef CONFIG_MLX5_EN_TLS
+#if defined(CONFIG_MLX5_EN_TLS) || defined(CONFIG_MLX5_EN_NVMEOTCP)
 	struct mlx5e_accel_fs_tcp      *accel_tcp;
 #endif
 	struct mlx5e_fs_udp            *udp;
@@ -1560,7 +1560,7 @@ void mlx5e_fs_set_any(struct mlx5e_flow_steering *fs, struct mlx5e_fs_any *any)
 	fs->any = any;
 }
 
-#ifdef CONFIG_MLX5_EN_TLS
+#if defined(CONFIG_MLX5_EN_TLS) || defined(CONFIG_MLX5_EN_NVMEOTCP)
 struct mlx5e_accel_fs_tcp *mlx5e_fs_get_accel_tcp(struct mlx5e_flow_steering *fs)
 {
 	return fs->accel_tcp;
drivers/net/ethernet/mellanox/mlx5/core/en_main.c +18 -0
index 43a3ff2d2d90..e582e97e7c44 100644
@@ -49,6 +49,7 @@
 #include "en_accel/macsec.h"
 #include "en_accel/en_accel.h"
 #include "en_accel/ktls.h"
+#include "en_accel/nvmeotcp.h"
 #include "lib/vxlan.h"
 #include "lib/clock.h"
 #include "en/port.h"
@@ -4146,6 +4147,13 @@ static netdev_features_t mlx5e_fix_features(struct net_device *netdev,
 		}
 	}
 
+	if (features & (NETIF_F_LRO | NETIF_F_GRO_HW)) {
+		if (params->nvmeotcp) {
+			netdev_warn(netdev, "Disabling HW-GRO/LRO, not supported after ULP DDP NVMe-TCP offload\n");
+			features &= ~(NETIF_F_LRO | NETIF_F_GRO_HW);
+		}
+	}
+
 	if (mlx5e_is_uplink_rep(priv))
 		features = mlx5e_fix_uplink_rep_features(netdev, features);
 
@@ -4914,6 +4922,9 @@ const struct net_device_ops mlx5e_netdev_ops = {
 	.ndo_has_offload_stats   = mlx5e_has_offload_stats,
 	.ndo_get_offload_stats   = mlx5e_get_offload_stats,
 #endif
+#ifdef CONFIG_MLX5_EN_NVMEOTCP
+	.ulp_ddp_ops             = &mlx5e_nvmeotcp_ops,
+#endif
 };
 
 static u32 mlx5e_choose_lro_timeout(struct mlx5_core_dev *mdev, u32 wanted_timeout)
@@ -5181,6 +5192,7 @@ static void mlx5e_build_nic_netdev(struct net_device *netdev)
 	mlx5e_macsec_build_netdev(priv);
 	mlx5e_ipsec_build_netdev(priv);
 	mlx5e_ktls_build_netdev(priv);
+	mlx5e_nvmeotcp_build_netdev(priv);
 }
 
 void mlx5e_create_q_counters(struct mlx5e_priv *priv)
@@ -5247,13 +5259,19 @@ static int mlx5e_nic_init(struct mlx5_core_dev *mdev,
 	if (err)
 		mlx5_core_err(mdev, "TLS initialization failed, %d\n", err);
 
+	err = mlx5e_nvmeotcp_init(priv);
+	if (err)
+		mlx5_core_err(mdev, "NVMEoTCP initialization failed, %d\n", err);
+
 	mlx5e_health_create_reporters(priv);
+
 	return 0;
 }
 
 static void mlx5e_nic_cleanup(struct mlx5e_priv *priv)
 {
 	mlx5e_health_destroy_reporters(priv);
+	mlx5e_nvmeotcp_cleanup(priv);
 	mlx5e_ktls_cleanup(priv);
 	mlx5e_fs_cleanup(priv->fs);
 }
drivers/net/ethernet/mellanox/mlx5/core/main.c +1 -0
index 9441588ac524..10df78b4f35c 100644
@@ -1556,6 +1556,7 @@ static const int types[] = {
 	MLX5_CAP_MACSEC,
 	MLX5_CAP_ADV_VIRTUALIZATION,
 	MLX5_CAP_CRYPTO,
+	MLX5_CAP_DEV_NVMEOTCP,
 };
 
 static void mlx5_hca_caps_free(struct mlx5_core_dev *dev)
Signature · 1 line
2.31.1
#23 lore raw

PATCH v11 19/25 net/mlx5e: TCP flow steering for nvme-tcp acceleration

In reply to Aurelien Aptel

From: Boris Pismenny <[email protected]>

Both nvme-tcp and tls acceleration require tcp flow steering. Add reference counter to share TCP flow steering structure.

---
 .../ethernet/mellanox/mlx5/core/en_accel/fs_tcp.c    | 12 +++++++++++-
 1 file changed, 11 insertions(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/en_accel/fs_tcp.c +11 -1
index 88a5aed9d678..29152d6e80d2 100644
@@ -14,6 +14,7 @@ enum accel_fs_tcp_type {
 struct mlx5e_accel_fs_tcp {
 	struct mlx5e_flow_table tables[ACCEL_FS_TCP_NUM_TYPES];
 	struct mlx5_flow_handle *default_rules[ACCEL_FS_TCP_NUM_TYPES];
+	refcount_t user_count;
 };
 
 static enum mlx5_traffic_types fs_accel2tt(enum accel_fs_tcp_type i)
@@ -360,6 +361,9 @@ void mlx5e_accel_fs_tcp_destroy(struct mlx5e_flow_steering *fs)
 	if (!accel_tcp)
 		return;
 
+	if (!refcount_dec_and_test(&accel_tcp->user_count))
+		return;
+
 	accel_fs_tcp_disable(fs);
 
 	for (i = 0; i < ACCEL_FS_TCP_NUM_TYPES; i++)
@@ -371,12 +375,17 @@ void mlx5e_accel_fs_tcp_destroy(struct mlx5e_flow_steering *fs)
 
 int mlx5e_accel_fs_tcp_create(struct mlx5e_flow_steering *fs)
 {
-	struct mlx5e_accel_fs_tcp *accel_tcp;
+	struct mlx5e_accel_fs_tcp *accel_tcp = mlx5e_fs_get_accel_tcp(fs);
 	int i, err;
 
 	if (!MLX5_CAP_FLOWTABLE_NIC_RX(mlx5e_fs_get_mdev(fs), ft_field_support.outer_ip_version))
 		return -EOPNOTSUPP;
 
+	if (accel_tcp) {
+		refcount_inc(&accel_tcp->user_count);
+		return 0;
+	}
+
 	accel_tcp = kzalloc(sizeof(*accel_tcp), GFP_KERNEL);
 	if (!accel_tcp)
 		return -ENOMEM;
@@ -392,6 +401,7 @@ int mlx5e_accel_fs_tcp_create(struct mlx5e_flow_steering *fs)
 	if (err)
 		goto err_destroy_tables;
 
+	refcount_set(&accel_tcp->user_count, 1);
 	return 0;
 
 err_destroy_tables:
Signature · 1 line
2.31.1
#27 lore raw

PATCH v11 23/25 net/mlx5e: NVMEoTCP, async ddp invalidation

In reply to Aurelien Aptel

From: Ben Ben-Ishay <[email protected]>

After the ULP consumed the buffers of the offloaded request, it calls the ddp_teardown op to release the NIC mapping for them and allow the NIC to reuse the HW contexts associated with offloading this IO. We do a fast/async un-mapping via UMR WQE. In this case, the ULP does holds off with completing the request towards the upper/application layers until the HW unmapping is done.

When the corresponding CQE is received, a notification is done via the the teardown_done ddp callback advertised by the ULP in the ddp context.

---
 .../net/ethernet/mellanox/mlx5/core/en/txrx.h |  4 ++
 .../mellanox/mlx5/core/en_accel/nvmeotcp.c    | 66 ++++++++++++++++---
 .../mellanox/mlx5/core/en_accel/nvmeotcp.h    |  1 +
 .../net/ethernet/mellanox/mlx5/core/en_rx.c   |  6 ++
 4 files changed, 67 insertions(+), 10 deletions(-)
drivers/net/ethernet/mellanox/mlx5/core/en/txrx.h +4 -0
index 31e240610e48..0bec8c535dc4 100644
@@ -52,6 +52,7 @@ enum mlx5e_icosq_wqe_type {
 #endif
 #ifdef CONFIG_MLX5_EN_NVMEOTCP
 	MLX5E_ICOSQ_WQE_UMR_NVMEOTCP,
+	MLX5E_ICOSQ_WQE_UMR_NVMEOTCP_INVALIDATE,
 	MLX5E_ICOSQ_WQE_SET_PSV_NVMEOTCP,
 #endif
 };
@@ -217,6 +218,9 @@ struct mlx5e_icosq_wqe_info {
 		struct {
 			struct mlx5e_nvmeotcp_queue *queue;
 		} nvmeotcp_q;
+		struct {
+			struct mlx5e_nvmeotcp_queue_entry *entry;
+		} nvmeotcp_qe;
 #endif
 	};
 };
drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.c +56 -10

126 lines not on this page · raw message

drivers/net/ethernet/mellanox/mlx5/core/en_accel/nvmeotcp.h +1 -0
index 555f3ed7e2e2..a5cfd9e31be7 100644
@@ -109,6 +109,7 @@ void mlx5e_nvmeotcp_cleanup(struct mlx5e_priv *priv);
 struct mlx5e_nvmeotcp_queue *
 mlx5e_nvmeotcp_get_queue(struct mlx5e_nvmeotcp *nvmeotcp, int id);
 void mlx5e_nvmeotcp_put_queue(struct mlx5e_nvmeotcp_queue *queue);
+void mlx5e_nvmeotcp_ddp_inv_done(struct mlx5e_icosq_wqe_info *wi);
 void mlx5e_nvmeotcp_ctx_complete(struct mlx5e_icosq_wqe_info *wi);
 static inline void mlx5e_nvmeotcp_init_rx(struct mlx5e_priv *priv) {}
 void mlx5e_nvmeotcp_cleanup_rx(struct mlx5e_priv *priv);
drivers/net/ethernet/mellanox/mlx5/core/en_rx.c +6 -0
index 253aa3f59f20..71f73cd5d36e 100644
@@ -888,6 +888,9 @@ void mlx5e_free_icosq_descs(struct mlx5e_icosq *sq)
 			break;
 #endif
 #ifdef CONFIG_MLX5_EN_NVMEOTCP
+		case MLX5E_ICOSQ_WQE_UMR_NVMEOTCP_INVALIDATE:
+			mlx5e_nvmeotcp_ddp_inv_done(wi);
+			break;
 		case MLX5E_ICOSQ_WQE_SET_PSV_NVMEOTCP:
 			mlx5e_nvmeotcp_ctx_complete(wi);
 			break;
@@ -993,6 +996,9 @@ int mlx5e_poll_ico_cq(struct mlx5e_cq *cq, int budget)
 #ifdef CONFIG_MLX5_EN_NVMEOTCP
 			case MLX5E_ICOSQ_WQE_UMR_NVMEOTCP:
 				break;
+			case MLX5E_ICOSQ_WQE_UMR_NVMEOTCP_INVALIDATE:
+				mlx5e_nvmeotcp_ddp_inv_done(wi);
+				break;
 			case MLX5E_ICOSQ_WQE_SET_PSV_NVMEOTCP:
 				mlx5e_nvmeotcp_ctx_complete(wi);
 				break;
Signature · 1 line
2.31.1
maintainer Acked-by on v11 4/25
#7 lore raw

Re: [PATCH v11 04/25] Documentation: document netlink ULP_DDP_GET/SET messages

In reply to Aurelien Aptel
Quoted 21 lines · On Fri, 3 Feb 2023 15:26:44 +0200 Aurelien Aptel wrote:
Documentation/netlink/specs/ethtool.yaml +21 -0
+      do: &ulp-ddp-get-op
+        request:
+          value: 44
+          attributes:
+            - header
+        reply:
+          value: 44
+          attributes:
+            - header
+            - hw
+            - active
+            - stats
+      dump: *ulp-ddp-get-op
+    -
+      name: ulp-ddp-set
+      doc: Set ULP DDP capabilities
+
+      attribute-set: ulp-ddp
+
+      do:
+        request:

Documentation/netlink/specs/ethtool.yaml operations: · line 490

Quoted 10 lines
Documentation/netlink/specs/ethtool.yaml +10 -0
+          value: 45
+          attributes:
+            - header
+            - wanted
+        reply:
+          value: 45
+          attributes:
+            - header
+            - hw
+            - active

If you need to respin for other reasons - you can drop the value:
entries here. I had to add it for the previous command because of the discontinuity in IDs but if you're just adding subsequent entries
- the value being previous + 1 is implied.

maintainer Reviewed-by on v11 2/25
#3 lore raw

Re: [PATCH v11 02/25] net/ethtool: add new stringset ETH_SS_ULP_DDP_{CAPS,STATS}

In reply to Aurelien Aptel
Quoted 8 lines · On Fri, 3 Feb 2023 15:26:42 +0200 Aurelien Aptel wrote:

This commit exposes ULP DDP capability and statistics names to userspace via netlink.

In order to support future ULP DDP capabilities and statistics without having to change the netlink protocol (and userspace ethtool) we add new string sets to let userspace dynamically fetch what the kernel supports.

Quoted 8 lines

* ETH_SS_ULP_DDP_CAPS stores names of ULP DDP capabilities
* ETH_SS_ULP_DDP_STATS stores names of ULP DDP statistics.

These stringsets will be used in later commits when implementing the new ULP DDP GET/SET netlink messages.

We keep the convention of strset.c of having the static_assert() right after the array declaration, despite the checkpatch warning.

maintainer Acked-by on v11 3/25
#5 lore raw

Re: [PATCH v11 03/25] net/ethtool: add ULP_DDP_{GET,SET} operations for caps and stats

In reply to Aurelien Aptel
Quoted 7 lines · On Fri, 3 Feb 2023 15:26:43 +0200 Aurelien Aptel wrote:

This commit adds:

- 3 new netlink messages:
* ULP_DDP_GET: returns a bitset of supported and active capabilities
* ULP_DDP_SET: tries to activate requested bitset and returns results
* ULP_DDP_NTF: notification for capabilities change

Quoted 8 lines

Rename and export bitset_policy for use in ulp_ddp.c.

ULP DDP capabilities handling is similar to netdev features handling.

If a ULP_DDP_GET message has requested statistics via the ETHTOOL_FLAG_STATS header flag, then statistics are returned to userspace.

#8 lore raw

RE: [PATCH v11 04/25] Documentation: document netlink ULP_DDP_GET/SET messages

In reply to Jakub Kicinski
Quoted 4 lines

If you need to respin for other reasons - you can drop the value:
entries here. I had to add it for the previous command because of the discontinuity in IDs but if you're just adding subsequent entries
- the value being previous + 1 is implied.

Ok

Quoted 1 line

Thanks Jakub.

maintainer
#30 lore raw
In reply to Aurelien Aptel
Quoted 19 lines

Hi,

Here is the next iteration of our nvme-tcp receive offload series.

The main changes are in patch 3 (netlink).

Rebased on top of today net-next 8065c0e13f98 ("Merge branch 'yt8531-support'")

The changes are also available through git:

Repo: https://github.com/aaptel/linux.git branch nvme-rx-offload-v11
Web: https://github.com/aaptel/linux/tree/nvme-rx-offload-v11

The NVMeTCP offload was presented in netdev 0x16 (video now available):
- https://netdevconf.info/0x16/session.html?NVMeTCP-Offload-%E2%80%93-Implementation-and-Performance-Gains
- https://youtu.be/W74TR-SNgi4

Quoted 6 lines

From: Aurelien Aptel <[email protected]>
From: Shai Malin <[email protected]>
From: Ben Ben-Ishay <[email protected]>
From: Boris Pismenny <[email protected]>
From: Or Gerlitz <[email protected]>
From: Yoray Zack <[email protected]>

Hey Aurelien and Co,

I've spent some time today looking at the last iteration of this, What I cannot understand, is how will this ever be used outside of the kernel nvme-tcp host driver?

It seems that the interface is diesigned to fit only a kernel consumer, and a very specific one.

Have you considered using a more standard interfaces to use this such that spdk or an io_uring based initiator can use it?

To me it appears that:
- ddp limits can be obtained via getsockopt
- sk_add/sk_del can be done via setsockopt
- offloaded DDGST crc can be obtained via something like msghdr.msg_control
- Perhaps for setting up the offload per IO, recvmsg would be the vehicle with a new msg flag MSG_RCV_DDP or something, that would hide all the details of what the HW needs (the command_id would be set somewhere in the msghdr).
- And all of the resync flow would be something that a separate ulp socket provider would take care of. Similar to how TLS presents itself to a tcp application. So the application does not need to be aware of it.

I'm not sure that such interface could cover everything that is needed, but what I'm trying to convey, is that the current interface limits the usability for almost anything else. Please correct me if I'm wrong.
Is this designed to also cater anything else outside of the kernel
nvme-tcp host driver?

Quoted 8 lines

Compatibility
=============
* The offload works with bare-metal or SRIOV.
* The HW can support up to 64K connections per device (assuming no other HW accelerations are used). In this series, we will introduce the support for up to 4k connections, and we have plans to increase it.
* SW TLS could not work together with the NVMeTCP offload as the HW will need to track the NVMeTCP headers in the TCP stream.

Can't say I like that.

Quoted 5 lines

* The ConnectX HW support HW TLS, but in ConnectX-7 those features could not co-exists (and it is not part of this series).
* The NVMeTCP offload ConnectX 7 HW can support tunneling, but we don’t see the need for this feature yet.
* NVMe poll queues are not in the scope of this series.

bonding/teaming?

Quoted 6 lines

Future Work
===========
* NVMeTCP transmit offload.
* NVMeTCP host offloads incremental features.
* NVMeTCP target offload.

Which target? which host?

#31 lore raw

Hi Sagi,

Quoted 8 lines · Sagi Grimberg

Hey Aurelien and Co,

I've spent some time today looking at the last iteration of this, What I cannot understand, is how will this ever be used outside of the kernel nvme-tcp host driver?

It seems that the interface is diesigned to fit only a kernel consumer, and a very specific one.

As part of this series, we are only covering the kernel nvme-tcp host driver. The ULP layer we are introducing was designed also for other kernel drivers such as the kernel nvme-tcp target and iSCSI.

Quoted 3 lines

Have you considered using a more standard interfaces to use this such that spdk or an io_uring based initiator can use it?

The main problem which I will explain in more detail (under the digest, teardown and resync flows) is that in order to use a more standard interface that will hide what the HW needs it will require duplicating the NVMeTCP logic of tracking the PDUs in the TCP stream – this seems wrong to us.

I can add that we are also working on an end-to-end user-space design for SPDK and we don’t see how the two designs could use the same socket APIs without impacting the performance gain of both cases.

Quoted 6 lines

To me it appears that:
- ddp limits can be obtained via getsockopt
- sk_add/sk_del can be done via setsockopt
- offloaded DDGST crc can be obtained via something like msghdr.msg_control

If we wish to hide it from the NVMeTCP driver it will require us to duplicate the NVMeTCP logic of tracking the PDUs in the TCP stream.
In the positive case, when there are no DDGST errors, nothing is needed to be done in the NVMeTCP driver, but in the case of errors (or just in the case of out of order packet in the middle of the PDU), the DDGST will need to be calculated in SW and it seems wrong to us to duplicate it outside of the NVMeTCP driver.

Quoted 4 lines

- Perhaps for setting up the offload per IO, recvmsg would be the vehicle with a new msg flag MSG_RCV_DDP or something, that would hide all the details of what the HW needs (the command_id would be set somewhere in the msghdr).

Our design includes the following steps per IO:
- ddp_setup (register the command id and buffer to the HW)
- The existing flow which sends the command
- The existing flow of read_sock()
- ddp_teardown (for the per IO HW teardown, before posting the NVMe completion)

Using the recvmsg will only replace the read_sock() but this part in the NVMeTCP driver is not impacted by the offload design.
The ddp_setup is needed in order to set the command id and the buffer, and this needs to be done before or as part of the sending of command and prior to the receive flow.
In addition, without duplicating the tracking of the PDUs in the TCP stream, it is not possible to hide the teardown flow from the NVMeTCP driver.

Quoted 4 lines

- And all of the resync flow would be something that a separate ulp socket provider would take care of. Similar to how TLS presents itself to a tcp application. So the application does not need to be aware of it.

The resync flow requires awareness of TCP sequence numbers and NVMe PDUs boundaries. If we hide it from the NVMeTCP driver we would have to again duplicate NVMe PDU tracking code.
TLS is a stream protocol and maps cleanly to TCP socket operations. NVMeTCP on the other hand is a request-response protocol.
On the data path, it's not comparable.
While it could be done by contorting recvmsg() with new flags, adding input fields to the msghdr struct and changing the behaviour of uAPI, it will add a lot more friction than a separate ULP DDP API specifically made for this purpose.

Quoted 6 lines

I'm not sure that such interface could cover everything that is needed, but what I'm trying to convey, is that the current interface limits the usability for almost anything else. Please correct me if I'm wrong. Is this designed to also cater anything else outside of the kernel nvme-tcp host driver?

As part of this series, we are targeting the kernel nvme-tcp host driver, and later we are planning to add support for the kernel nvme-tcp target driver.

The ULP layer was designed to be generic for other request-response based protocols.

Quoted 11 lines
Quoted 9 lines

Compatibility
=============
* The offload works with bare-metal or SRIOV.
* The HW can support up to 64K connections per device (assuming no other HW accelerations are used). In this series, we will introduce the support for up to 4k connections, and we have plans to increase it.
* SW TLS could not work together with the NVMeTCP offload as the HW will need to track the NVMeTCP headers in the TCP stream.

Can't say I like that.

The answer should be to support both TLS offload and NVMeTCP offload, which is a HW limit in our current generation.

Quoted 8 lines
Quoted 6 lines

* The ConnectX HW support HW TLS, but in ConnectX-7 those features could not co-exists (and it is not part of this series).
* The NVMeTCP offload ConnectX 7 HW can support tunneling, but we don’t see the need for this feature yet.
* NVMe poll queues are not in the scope of this series.

bonding/teaming?

We are planning to add it as part of the "incremental feature". It will follow the existing design of the mlx HW TLS.

Quoted 9 lines
Quoted 7 lines

Future Work
===========
* NVMeTCP transmit offload.
* NVMeTCP host offloads incremental features.
* NVMeTCP target offload.

Which target? which host?

Kernel nvme-tcp host driver and kernel nvme-tcp target driver.

maintainer
#32 lore raw
Quoted 2 lines

Hi Sagi,

Hey Shai,

Quoted 7 lines · Sagi Grimberg
Quoted 6 lines

Hey Aurelien and Co,

I've spent some time today looking at the last iteration of this, What I cannot understand, is how will this ever be used outside of the kernel nvme-tcp host driver?

Quoted 6 lines
Quoted 2 lines

It seems that the interface is diesigned to fit only a kernel consumer, and a very specific one.

As part of this series, we are only covering the kernel nvme-tcp host driver. The ULP layer we are introducing was designed also for other kernel drivers such as the kernel nvme-tcp target and iSCSI.

Yes, I can see how this can fit iscsi, but my question was more for userspace.

Quoted 7 lines
Quoted 2 lines

Have you considered using a more standard interfaces to use this such that spdk or an io_uring based initiator can use it?

The main problem which I will explain in more detail (under the digest, teardown and resync flows) is that in order to use a more standard interface that will hide what the HW needs it will require duplicating the NVMeTCP logic of tracking the PDUs in the TCP stream – this seems wrong to us.

Hmm. Its not really the entire logic, its only the pdu/data boundaries. Every data coming to the host is structured, with a standard C2HData PDU.

If we assume for the sake of the discussion that such a parser exist, how would the interface to the ulp look like?

Quoted 3 lines

I can add that we are also working on an end-to-end user-space design for SPDK and we don’t see how the two designs could use the same socket APIs without impacting the performance gain of both cases.

Can you share a bit more how this will be done for spdk? spdk runs on normal sockets today, hence I'd love to know how you plan to introduce it there.

Quoted 8 lines
Quoted 7 lines

To me it appears that:
- ddp limits can be obtained via getsockopt
- sk_add/sk_del can be done via setsockopt
- offloaded DDGST crc can be obtained via something like msghdr.msg_control

Quoted 7 lines

If we wish to hide it from the NVMeTCP driver it will require us to duplicate the NVMeTCP logic of tracking the PDUs in the TCP stream.
In the positive case, when there are no DDGST errors, nothing is needed to be done in the NVMeTCP driver, but in the case of errors (or just in the case of out of order packet in the middle of the PDU), the DDGST will need to be calculated in SW and it seems wrong to us to duplicate it outside of the NVMeTCP driver.

I didn't suggest that the recalc would be hidden, if the calculation failed the ULP can do it.

Something like:

	ddp_ddgst = CMSG_DATA(CMSG_FIRSTHDR(msg))
	if (ddp_ddgst->valid) {
		great...
	} else {
		recalc_ddgst
	}
Quoted 12 lines
Quoted 5 lines

- Perhaps for setting up the offload per IO, recvmsg would be the vehicle with a new msg flag MSG_RCV_DDP or something, that would hide all the details of what the HW needs (the command_id would be set somewhere in the msghdr).

Our design includes the following steps per IO:
- ddp_setup (register the command id and buffer to the HW)
- The existing flow which sends the command
- The existing flow of read_sock()
- ddp_teardown (for the per IO HW teardown, before posting the NVMe completion)

Quoted 5 lines

Using the recvmsg will only replace the read_sock() but this part in the NVMeTCP driver is not impacted by the offload design.
The ddp_setup is needed in order to set the command id and the buffer, and this needs to be done before or as part of the sending of command and prior to the receive flow.

I agree, I didn't suggest that replacing the entire read_sock flow (although that can change to recvmsg as well).

I was thinking more along the lines of MSG_PEEK that doesn't actually reads anything, where the driver prior to sending the command, would do:

	// do a DDP recvmsg to setup direct data placement
	msg = { .msg_flags = MSG_DDP_SETUP };
	iov_iter_bvec(&msg.msg_iter, ITER_DEST, vec, nr_bvec, size);
	sock_recvmsg(queue->sock, &msg, msg.msg_flags);

	...

	// now send the cmd to the controller
	nvme_tcp_setup_cmd_pdu
	nvme_tcp_try_send_cmd_pdu
	...

That is what I was thinking about.

Quoted 2 lines

In addition, without duplicating the tracking of the PDUs in the TCP stream, it is not possible to hide the teardown flow from the NVMeTCP driver.

The teardown flow is a bummer because it delays the nvme completion. Is it ok to dma_unmap and then schedule a ddp teardown async without delaying the nvme completion?

How does TLS offload handles the async teardown?

Quoted 9 lines
Quoted 5 lines

- And all of the resync flow would be something that a separate ulp socket provider would take care of. Similar to how TLS presents itself to a tcp application. So the application does not need to be aware of it.

The resync flow requires awareness of TCP sequence numbers and NVMe PDUs boundaries. If we hide it from the NVMeTCP driver we would have to again duplicate NVMe PDU tracking code.

Its really a pretty trivial part of the pdu tracking. Just when a C2HData PDU starts...

Quoted 3 lines

TLS is a stream protocol and maps cleanly to TCP socket operations. NVMeTCP on the other hand is a request-response protocol. On the data path, it's not comparable.

Is it? looks similar to me in the sense that the input stream needs to know when a msg starts and ends. The only difference is that the ULP needs to timely provide the context to the offload.

Quoted 4 lines

While it could be done by contorting recvmsg() with new flags, adding input fields to the msghdr struct and changing the behaviour of uAPI, it will add a lot more friction than a separate ULP DDP API specifically made for this purpose.

This may be true. I was merely asking if this was a path that was genuinely explored and ruled out. To me, it sounds like it can work, and also allow userspace to automatically gain from it.

Quoted 11 lines
Quoted 5 lines

I'm not sure that such interface could cover everything that is needed, but what I'm trying to convey, is that the current interface limits the usability for almost anything else. Please correct me if I'm wrong. Is this designed to also cater anything else outside of the kernel nvme-tcp host driver?

As part of this series, we are targeting the kernel nvme-tcp host driver, and later we are planning to add support for the kernel nvme-tcp target driver.

The ULP layer was designed to be generic for other request-response based protocols.

I understand.

Quoted 11 lines
Quoted 11 lines
Quoted 10 lines

Compatibility
=============
* The offload works with bare-metal or SRIOV.
* The HW can support up to 64K connections per device (assuming no other HW accelerations are used). In this series, we will introduce the support for up to 4k connections, and we have plans to increase it.
* SW TLS could not work together with the NVMeTCP offload as the HW will need to track the NVMeTCP headers in the TCP stream.

Quoted 4 lines
Quoted 1 line

Can't say I like that.

The answer should be to support both TLS offload and NVMeTCP offload, which is a HW limit in our current generation.

I can't complain about this because we don't have TLS supported in nvme-tcp. But _can_ they work together on future HW?

Quoted 8 lines
Quoted 8 lines
Quoted 7 lines

* The ConnectX HW support HW TLS, but in ConnectX-7 those features could not co-exists (and it is not part of this series).
* The NVMeTCP offload ConnectX 7 HW can support tunneling, but we don’t see the need for this feature yet.
* NVMe poll queues are not in the scope of this series.

Quoted 4 lines
Quoted 1 line

bonding/teaming?

We are planning to add it as part of the "incremental feature". It will follow the existing design of the mlx HW TLS.

OK.

Quoted 9 lines
Quoted 7 lines
Quoted 5 lines

Future Work
===========
* NVMeTCP transmit offload.
* NVMeTCP host offloads incremental features.
* NVMeTCP target offload.

Which target? which host?

Kernel nvme-tcp host driver and kernel nvme-tcp target driver.

Look, I did spend time looking into this patchset several times in the past, and also provided feedback. The reason why I'm questioning the interface now, is because I'm wandering if it can be less intrusive and use the common socket interface instead of inventing a new one.

#33 lore raw

Hi Sagi

Quoted 11 lines · On Tue, 7 Mar 2023 at 12:11, Sagi Grimberg [email protected] wrote:
Quoted 7 lines
Quoted 2 lines

Have you considered using a more standard interfaces to use this such that spdk or an io_uring based initiator can use it?

The main problem which I will explain in more detail (under the digest, teardown and resync flows) is that in order to use a more standard interface that will hide what the HW needs it will require duplicating the NVMeTCP logic of tracking the PDUs in the TCP stream – this seems wrong to us.

Hmm. Its not really the entire logic, its only the pdu/data boundaries. Every data coming to the host is structured, with a standard C2HData PDU.

In order to track the PDUs we will need to duplicate the parsing and tracking for all the common headers of all the PDU types, that is doable, but it will also need to be trustworthy, which means duplicating the header digest verification if enabled.

Quoted 3 lines

If we assume for the sake of the discussion that such a parser exist, how would the interface to the ulp look like?

Enablement could be done via setsockopt(). We would add a new SOL_NVME_TCP flag and store its state in new socket struct fields. We would also set the resync callbacks from ipvX/tcp.c (explained further down).

Per IO setup:
-------------
The ddp_setup() call would be moved to ipv4/ipv6 implementations of TCP, in sock_recvmsg().
As you suggested, we would add a new MSG_NVME_DDP flag to register an expected NVMeTCP data response PDU.
Adding a command_id field to the msghdr struct would break userspace ABI. Instead we would use msghdr ancillary fields and add a new cmsg_level (SOL_NVME_TCP) and a new cmsg_type (NVME_TCP_COMMAND_ID).

      struct msghdr msg = {.flags = MSG_NVME_DDP};
      int cmsg_len = sizeof(u32); // command_id size
      struct cmsghdr *cmsg;
      char buf[CMSG_SPACE(cmsg_len)];
 
      msg.msg_control = buf;
      msg.msg_controllen = sizeof(buf);
      cmsg = CMSG_FIRSTHDR(&msg);
      cmsg->cmsg_level = SOL_NVME_TCP;
      cmsg->cmsg_type = NVME_TCP_COMMAND_ID;
      cmsg->cmsg_len = CMSG_LEN(cmsg_len);
      *CMSG_DATA(cmsg) = 0x1; // command_id
      msg.msg_controllen = cmsg->cmsg_len;
 
      iov_iter_bvec(&msg.msg_iter, ITER_DEST, vec, nr_bvec, size);
      sock_recvmsg(queue->sock, &msg, msg.msg_flags);

Per IO teardown:
----------------
The ddp_teardown will need to remain as a separate API, as explained before, it is needed in order to verify that the HW teardown flow is complete before posting the NVMe completion.

It’s not possible to move the teardown into sock_recvmsg() as it adds a significant delay (~1 usec) and will not allow us to process other PDUs in the meantime.

The challenge here is that we haven’t found a good and inexpensive reference in the socket API for such IO notifications (I assume we don’t want to use signals).
This might block the socket API approach.

Data Digest offload:
--------------------
Since recvmsg() would only read a single PDU, your suggested approach could work:

          ddp_ddgst = CMSG_DATA(CMSG_FIRSTHDR(msg))
          if (ddp_ddgst->valid) {
                 great...
          } else {
                 recalc_ddgst
          }

Receive flow:
-------------
Using recvmsg() doesn't have the callback mechanism that sock->read_sock() provides so the logic in the NVMeTCP driver code would have to be changed. Instead of parsing and copying the PDU in chunks as it comes via the callback, it would have to process the entire PDU upon returning from recvmsg() in one go.

To handle re-ordered NVMe-TCP responses, the ULP would duplicate the tracking of PDUs and maintain a list of completed responses yet to be received by the NVMeTCP driver.
The NVMeTCP driver will call the recvmsg(s, &ddp_buf, cccid) only for C2H PDUs, and based on the PDU header type and command_id.

If placement was successful, the data is already in ddp_buf. If not, the kernel would perform the regular software copy. It would be transparent for the NVMeTCP driver.

As mentioned before, in order to safely track the PDUs, we will need to duplicate the header digest verification if enabled.

In summary the flow for NVMeTCP driver could be something like:

// Enable nvme-tcp offload

                setsockopt(s, SOL_NVME_TCP, ...);

...

// Command sending:

                // register memory
                msg.flags = MSG_NVME_DDP;
                msg.iov_iter = resp_buf;
 
                // set 1 msg ancillary field with the command id
                cmsg->cmsg_level = SOL_NVME_TCP;
                cmsg->cmsg_type = NVME_TCP_COMMAND_ID;
                cmsg->cmsg_len = CMSG_LEN(sizeof(u32));
                *CMSG_DATA(cmsg) = 0x1; // command id
                recvmsg(s, msg, msg.flags);
 
                // send the request
                send(s, req_buf, ...);

...

// C2H receive:

We will need to modify nvme_tcp_recv_pdu() to read only the next PDU header and according to the pdu type and command_id perform the following:

      ddp_buf = get_cid_buffer(cid);
      msg.iov_iter = ddp_buf;
      // set 1 msg ancillary field with the command id
      cmsg->cmsg_level = SOL_NVME_TCP;
      cmsg->cmsg_type = NVME_TCP_COMMAND_ID;
      cmsg->cmsg_len = CMSG_LEN(sizeof(u32));
      *CMSG_DATA(cmsg) = cid;
      recvmsg(s, msg, msg.flags);
      
      // Teardown (for this we don't have a good solution)
 
      // Data digest check
      ddp_ddgst = CMSG_DATA(CMSG_FIRSTHDR(msg))
      if (ddp_ddgst->valid) {
            great...
      } else {
            recalc_ddgst
      }    

Resync:
-------
This API could be removed from the NVMeTCP driver, and we could move the processing of resync request to ipv4/ipv6 tcp code.
The resync pending state would be moved to the socket struct. But also in this case, in order to safely track the PDUs, we will need to parse all PDUs and to also calculate the header digest if enabled.

User space:
-----------
Regarding the user space design, since issuing 3 syscalls per IO (recvmsg + sendmsg + recvmsg) would be too expensive, we could add support for io_uring so it will wrap the above APIs.
For example, similarly to recvmsg(), we would submit a IORING_OP_RECVMSG sqe before each IO (with the MSG_NVME_DDP flag) to register memory for DDP.

Quoted 8 lines
Quoted 4 lines

I can add that we are also working on an end-to-end user-space design for SPDK and we don’t see how the two designs could use the same socket APIs without impacting the performance gain of both cases.

Can you share a bit more how this will be done for spdk? spdk runs on normal sockets today, hence I'd love to know how you plan to introduce it there.

The SPDK developed at Nvidia can use XLIO
https://docs.nvidia.com/networking/display/XLIOv214 as the TCP stack. It will be out of the context of this discussion as it is not following the normal socket operations.

Quoted 20 lines
Quoted 9 lines
Quoted 8 lines

To me it appears that:
- ddp limits can be obtained via getsockopt
- sk_add/sk_del can be done via setsockopt
- offloaded DDGST crc can be obtained via something like msghdr.msg_control

Quoted 7 lines

If we wish to hide it from the NVMeTCP driver it will require us to duplicate the NVMeTCP logic of tracking the PDUs in the TCP stream.
In the positive case, when there are no DDGST errors, nothing is needed to be done in the NVMeTCP driver, but in the case of errors (or just in the case of out of order packet in the middle of the PDU), the DDGST will need to be calculated in SW and it seems wrong to us to duplicate it outside of the NVMeTCP driver.

I didn't suggest that the recalc would be hidden, if the calculation failed the ULP can do it.

Quoted 7 lines

Something like:

        ddp_ddgst = CMSG_DATA(CMSG_FIRSTHDR(msg))
        if (ddp_ddgst->valid) {
                great...
        } else {
                recalc_ddgst
        }

Yes, we agree, as explained before.

Quoted 37 lines
Quoted 13 lines
Quoted 6 lines

- Perhaps for setting up the offload per IO, recvmsg would be the vehicle with a new msg flag MSG_RCV_DDP or something, that would hide all the details of what the HW needs (the command_id would be set somewhere in the msghdr).

Our design includes the following steps per IO:
- ddp_setup (register the command id and buffer to the HW)
- The existing flow which sends the command
- The existing flow of read_sock()
- ddp_teardown (for the per IO HW teardown, before posting the NVMe completion)

Quoted 5 lines

Using the recvmsg will only replace the read_sock() but this part in the NVMeTCP driver is not impacted by the offload design.
The ddp_setup is needed in order to set the command id and the buffer, and this needs to be done before or as part of the sending of command and prior to the receive flow.

I agree, I didn't suggest that replacing the entire read_sock flow (although that can change to recvmsg as well).

I was thinking more along the lines of MSG_PEEK that doesn't actually reads anything, where the driver prior to sending the command, would do:

        // do a DDP recvmsg to setup direct data placement
        msg = { .msg_flags = MSG_DDP_SETUP };
        iov_iter_bvec(&msg.msg_iter, ITER_DEST, vec, nr_bvec, size);
        sock_recvmsg(queue->sock, &msg, msg.msg_flags);

        ...

        // now send the cmd to the controller
        nvme_tcp_setup_cmd_pdu
        nvme_tcp_try_send_cmd_pdu
        ...
Quoted 1 line

That is what I was thinking about.

Was answered as part of the “receive flow” section above.

Quoted 7 lines
Quoted 3 lines

In addition, without duplicating the tracking of the PDUs in the TCP stream, it is not possible to hide the teardown flow from the NVMeTCP driver.

The teardown flow is a bummer because it delays the nvme completion. Is it ok to dma_unmap and then schedule a ddp teardown async without delaying the nvme completion?

With the existing HW it will not be possible. The teardown is required in order to guarantee that the HW will not try accessing the HW resources or the IO buffer for which nvme completion was already called.

In a future HW, the teardown flow could improve.

Quoted 2 lines

How does TLS offload handles the async teardown?

In TLS we don’t have such operation per IO. Only the resync is similar.

Quoted 7 lines
Quoted 7 lines
Quoted 6 lines

- And all of the resync flow would be something that a separate ulp socket provider would take care of. Similar to how TLS presents itself to a tcp application. So the application does not need to be aware of it.

Quoted 6 lines
Quoted 3 lines

The resync flow requires awareness of TCP sequence numbers and NVMe PDUs boundaries. If we hide it from the NVMeTCP driver we would have to again duplicate NVMe PDU tracking code.

Its really a pretty trivial part of the pdu tracking. Just when a C2HData PDU starts...

The resync is not only in the context of the C2HData PDUs,
As explained before, it will require to parse the common header and validate the header digest in order to verify the resync of the HW.
We really don’t like this duplication.

Quoted 8 lines
Quoted 4 lines

TLS is a stream protocol and maps cleanly to TCP socket operations. NVMeTCP on the other hand is a request-response protocol. On the data path, it's not comparable.

Is it? looks similar to me in the sense that the input stream needs to know when a msg starts and ends. The only difference is that the ULP needs to timely provide the context to the offload.

It's a big difference, the app that is using the TLS doesn’t need to handle the IO path operations that NVMeTCP offload requires (ddp_setup, ddp_treadown).

Quoted 9 lines
Quoted 5 lines

While it could be done by contorting recvmsg() with new flags, adding input fields to the msghdr struct and changing the behaviour of uAPI, it will add a lot more friction than a separate ULP DDP API specifically made for this purpose.

This may be true. I was merely asking if this was a path that was genuinely explored and ruled out. To me, it sounds like it can work, and also allow userspace to automatically gain from it.

We have explored different approaches but ultimately picked this one. While it might be more intrusive, it offers the better compromise with performance and cleanliness (avoiding code duplication and major refactoring).

The socket API:
1. Won’t simplify the nvme-tcp code: it will arguably make it more complex. That is, assuming we resolve all the issues mentioned earlier.

2. Its only real improvement would be to have the feature available from userspace which we doubt can be as performant.
(side note – based on our experience, managing the entire TCP stack from userspace will achieve much better performance but this is again out of the scope of this discussion)

Based on that, our request would be to review and consider merging this series with our original design.

Quoted 27 lines
Quoted 11 lines
Quoted 6 lines

I'm not sure that such interface could cover everything that is needed, but what I'm trying to convey, is that the current interface limits the usability for almost anything else. Please correct me if I'm wrong. Is this designed to also cater anything else outside of the kernel nvme-tcp host driver?

As part of this series, we are targeting the kernel nvme-tcp host driver, and later we are planning to add support for the kernel nvme-tcp target driver.

The ULP layer was designed to be generic for other request-response

based

Quoted 1 line

protocols.

I understand.

Quoted 11 lines
Quoted 11 lines
Quoted 10 lines

Compatibility
=============
* The offload works with bare-metal or SRIOV.
* The HW can support up to 64K connections per device (assuming no other HW accelerations are used). In this series, we will introduce the support for up to 4k connections, and we have plans to increase it.
* SW TLS could not work together with the NVMeTCP offload as the HW will need to track the NVMeTCP headers in the TCP stream.

Quoted 7 lines
Quoted 4 lines
Quoted 1 line

Can't say I like that.

The answer should be to support both TLS offload and NVMeTCP offload, which is a HW limit in our current generation.

I can't complain about this because we don't have TLS supported in nvme-tcp. But _can_ they work together on future HW?

Yes. With our original design the effort to support both offloads is incremental.

But, with the new socket API design discussed here, it will be more challenging to stack ULPs, and hence doing TLS + NVMe-TCP will be harder with this design.

Quoted 22 lines
Quoted 9 lines
Quoted 9 lines
Quoted 8 lines

* The ConnectX HW support HW TLS, but in ConnectX-7 those features could not co-exists (and it is not part of this series).
* The NVMeTCP offload ConnectX 7 HW can support tunneling, but we don’t see the need for this feature yet.
* NVMe poll queues are not in the scope of this series.

Quoted 4 lines
Quoted 1 line

bonding/teaming?

We are planning to add it as part of the "incremental feature". It will follow the existing design of the mlx HW TLS.

OK.

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

Future Work
===========
* NVMeTCP transmit offload.
* NVMeTCP host offloads incremental features.
* NVMeTCP target offload.

Quoted 8 lines
Quoted 3 lines
Quoted 1 line

Which target? which host?

Kernel nvme-tcp host driver and kernel nvme-tcp target driver.

Look, I did spend time looking into this patchset several times in the past, and also provided feedback. The reason why I'm questioning the interface now, is because I'm wandering if it can be less intrusive and use the common socket interface instead of inventing a new one.

Thanks again for your time looking into this and providing your feedback. It’s a valid question but we think the current design is the better compromise.