Public bug reported:

[Impact]

Google has requested a mainline backport patchset to support Mellanox
ConnectX-8 RDMA in 6.8 and later kernels. Since our current interim
release has a kernel version of 6.14, which contains these patches
already, we only need to backport to 6.8.

[Fix]

The requested patches are:
e738e3550452 ("net/mlx5: Enable 4 ports multiport E-switch")
4f4db190893f ("IB/mlx5: Adjust mlx5 rate mapping to support 800Gb")
b28ad32442be ("IB/mlx5: Rename 400G_8X speed to comply to naming convention")
948f0bf5ad6a ("IB/mlx5: Add support for 800G_8X lane speed")
561b4a3ac655 ("IB/mlx5: Expose XDR speed through MAD")
703289ce43f7 ("IB/core: Add support for XDR link speed")
8c54c89ad45a ("net/mlx5e: Add support for 800Gbps link modes")
e0e6adfe8c20 ("net/mlx5: Enable 8 ports LAG")
af48f95492dc ("RDMA/core: Introduce "name_assign_type" for an IB device")
7a2210a57d42 ("RDMA/mlx5: Support per-plane port IB counters by querying PPCNT 
register")
c6b6677d85d4 ("net/mlx5: mlx5_ifc update for accessing ppcnt register of plane 
ports")
3b43399b297c ("RDMA/mlx5: Add plane index support when querying PTYS registers")
294424839b5e ("RDMA/nldev: Add support to dump device type and parent device if 
exists")
060c642b2ab8 ("RDMA/nldev: Add support to add/delete a sub IB device through 
netlink")
026a425990af ("RDMA/mlx5: Support plane device and driver APIs to add and 
delete it")
a9e0facacfd1 ("RDMA/core: Create GSI QP only when CM is supported")
bca51197620a ("RDMA/core: Support IB sub device with type "SMI"")
2a5db20fa532 ("RDMA/mlx5: Add support to multi-plane device and port")
65528cfb21fd ("net/mlx5: mlx5_ifc update for multi-plane support")
50660c5197f5 ("RDMA/core: Create "issm*" device nodes only when SMI is 
supported")
c77aec65e828 ("RDMA/mlx5: Consider the query_vuid cap for data_direct")
ec7ad6530909 ("RDMA/mlx5: Introduce GET_DATA_DIRECT_SYSFS_PATH ioctl")
de8f847a5114 ("RDMA/mlx5: Add support for DMABUF MR registrations with 
Data-direct")
3aa73c6b795b ("RDMA: Pass uverbs_attr_bundle as part of '.reg_user_mr_dmabuf' 
API")
253c61dc256b ("RDMA/umem: Introduce an option to revoke DMABUF umem")
682358fd35de ("RDMA/umem: Add support for creating pinned DMABUF umem with a 
given dma device")
2e8e631d7a41 ("RDMA/mlx5: Add the initialization flow to utilize the 'data 
direct' device")
6910e3660d86 ("RDMA/mlx5: Introduce the 'data direct' driver")
c772a2c69018 ("net/mlx5: Add IFC related stuff for data direct")
47c8846a49ba ("PCI: Extend ACS configurability")
8b36f7c3c661 ("RDMA/mlx5: Support OOO RX WQE consumption")
8439662f6f16 ("Introduce mlx5 data direct placement (DDP)")
eb3d354efb39 ("RDMA/mlx5: Support querying per-plane IB PortCounters")
45d339fefaa3 ("RDMA/mlx5: Enable multiplane mode only when it is supported")
fe4bf8d0b671 ("vfio/pci: Properly hide first-in-list PCIe extended capability")
4d617b57574f ("net/mlx5: DR, add support for ConnectX-8 steering")
aa90a30804a5 ("net/mlx5: DR, expand SWS STE callbacks and consolidate common 
structs")
e799ac9dd3c4 ("net/mlx5: Add ConnectX-8 device to ifc")

[Test plan]

SSH into a machine using a ConnectX-8 NIC with this kernel installed.

[Regression potential]

Most of these patches are specific to RDMA and mlx5, but there are some
that modify behavior in the core PCI system. Because this patchset is
large, and does span multiple modules, but primarily is contained to
feature additions and driver code, regression potential is moderate.
Some amount of stress testing with the new hardware is recommended but
not strictly required.

[Other]

SF: 00407979

** Affects: linux-gcp (Ubuntu)
     Importance: Undecided
         Status: New

** Affects: linux-gcp (Ubuntu Noble)
     Importance: Undecided
         Status: New

** Also affects: linux-gcp (Ubuntu Noble)
   Importance: Undecided
       Status: New

-- 
You received this bug notification because you are a member of Kernel
Packages, which is subscribed to linux-gcp in Ubuntu.
https://bugs.launchpad.net/bugs/2109806

Title:
  Mellanox ConnectX-8 RDMA support

Status in linux-gcp package in Ubuntu:
  New
Status in linux-gcp source package in Noble:
  New

Bug description:
  [Impact]

  Google has requested a mainline backport patchset to support Mellanox
  ConnectX-8 RDMA in 6.8 and later kernels. Since our current interim
  release has a kernel version of 6.14, which contains these patches
  already, we only need to backport to 6.8.

  [Fix]

  The requested patches are:
  e738e3550452 ("net/mlx5: Enable 4 ports multiport E-switch")
  4f4db190893f ("IB/mlx5: Adjust mlx5 rate mapping to support 800Gb")
  b28ad32442be ("IB/mlx5: Rename 400G_8X speed to comply to naming convention")
  948f0bf5ad6a ("IB/mlx5: Add support for 800G_8X lane speed")
  561b4a3ac655 ("IB/mlx5: Expose XDR speed through MAD")
  703289ce43f7 ("IB/core: Add support for XDR link speed")
  8c54c89ad45a ("net/mlx5e: Add support for 800Gbps link modes")
  e0e6adfe8c20 ("net/mlx5: Enable 8 ports LAG")
  af48f95492dc ("RDMA/core: Introduce "name_assign_type" for an IB device")
  7a2210a57d42 ("RDMA/mlx5: Support per-plane port IB counters by querying 
PPCNT register")
  c6b6677d85d4 ("net/mlx5: mlx5_ifc update for accessing ppcnt register of 
plane ports")
  3b43399b297c ("RDMA/mlx5: Add plane index support when querying PTYS 
registers")
  294424839b5e ("RDMA/nldev: Add support to dump device type and parent device 
if exists")
  060c642b2ab8 ("RDMA/nldev: Add support to add/delete a sub IB device through 
netlink")
  026a425990af ("RDMA/mlx5: Support plane device and driver APIs to add and 
delete it")
  a9e0facacfd1 ("RDMA/core: Create GSI QP only when CM is supported")
  bca51197620a ("RDMA/core: Support IB sub device with type "SMI"")
  2a5db20fa532 ("RDMA/mlx5: Add support to multi-plane device and port")
  65528cfb21fd ("net/mlx5: mlx5_ifc update for multi-plane support")
  50660c5197f5 ("RDMA/core: Create "issm*" device nodes only when SMI is 
supported")
  c77aec65e828 ("RDMA/mlx5: Consider the query_vuid cap for data_direct")
  ec7ad6530909 ("RDMA/mlx5: Introduce GET_DATA_DIRECT_SYSFS_PATH ioctl")
  de8f847a5114 ("RDMA/mlx5: Add support for DMABUF MR registrations with 
Data-direct")
  3aa73c6b795b ("RDMA: Pass uverbs_attr_bundle as part of '.reg_user_mr_dmabuf' 
API")
  253c61dc256b ("RDMA/umem: Introduce an option to revoke DMABUF umem")
  682358fd35de ("RDMA/umem: Add support for creating pinned DMABUF umem with a 
given dma device")
  2e8e631d7a41 ("RDMA/mlx5: Add the initialization flow to utilize the 'data 
direct' device")
  6910e3660d86 ("RDMA/mlx5: Introduce the 'data direct' driver")
  c772a2c69018 ("net/mlx5: Add IFC related stuff for data direct")
  47c8846a49ba ("PCI: Extend ACS configurability")
  8b36f7c3c661 ("RDMA/mlx5: Support OOO RX WQE consumption")
  8439662f6f16 ("Introduce mlx5 data direct placement (DDP)")
  eb3d354efb39 ("RDMA/mlx5: Support querying per-plane IB PortCounters")
  45d339fefaa3 ("RDMA/mlx5: Enable multiplane mode only when it is supported")
  fe4bf8d0b671 ("vfio/pci: Properly hide first-in-list PCIe extended 
capability")
  4d617b57574f ("net/mlx5: DR, add support for ConnectX-8 steering")
  aa90a30804a5 ("net/mlx5: DR, expand SWS STE callbacks and consolidate common 
structs")
  e799ac9dd3c4 ("net/mlx5: Add ConnectX-8 device to ifc")

  [Test plan]

  SSH into a machine using a ConnectX-8 NIC with this kernel installed.

  [Regression potential]

  Most of these patches are specific to RDMA and mlx5, but there are
  some that modify behavior in the core PCI system. Because this
  patchset is large, and does span multiple modules, but primarily is
  contained to feature additions and driver code, regression potential
  is moderate. Some amount of stress testing with the new hardware is
  recommended but not strictly required.

  [Other]

  SF: 00407979

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux-gcp/+bug/2109806/+subscriptions


-- 
Mailing list: https://launchpad.net/~kernel-packages
Post to     : kernel-packages@lists.launchpad.net
Unsubscribe : https://launchpad.net/~kernel-packages
More help   : https://help.launchpad.net/ListHelp

Reply via email to