Cmnt: [SRU][J:linux-bluefield][PATCH v3 0/9] revert SAUCE patches for mlxbf-tmfifo driver

David Thompson davthompson at nvidia.com
Mon Mar 24 20:36:54 UTC 2025


> -----Original Message-----
> From: Massimiliano Pellizzer <massimiliano.pellizzer at canonical.com>
> Sent: Monday, March 24, 2025 9:44 AM
> To: kernel-team at lists.ubuntu.com
> Cc: Haifei Luo <haifeil at nvidia.com>; David Thompson
> <davthompson at nvidia.com>
> Subject: Cmnt: [SRU][J:linux-bluefield][PATCH v3 0/9] revert SAUCE patches for
> mlxbf-tmfifo driver
> 
> On Mon, 24 Mar 2025 at 03:00, Haifei Luo <haifeil at nvidia.com> wrote:
> >
> > BugLink: https://bugs.launchpad.net/bugs/2103386
> >
> > SRU Justification
> >
> > [Description]
> > The following SAUCE patches should be reverted and replaced with upstream
> commits:
> > 7789adb5fe90 UBUNTU: SAUCE: mlxbf-tmfifo: fix potential race
> > 3d8185022ec5 UBUNTU: SAUCE: platform/mellanox: mlxbf-tmfifo: Add
> > BlueField-3 support
> >
> > [Impact]
> > Align with upstream commits.
> >
> > [Fix]
> > Other two commits are depended on the commits and need reverted. The
> reverted commits are:
> > fe4595c4d14e platform/mellanox: mlxbf-tmfifo: Drop Tx network packet
> > when Tx TmFIFO is full
> > 7789adb5fe90 UBUNTU: SAUCE: mlxbf-tmfifo: fix potential race
> > 3d77627b0ada UBUNTU: SAUCE: platform/mellanox: Add ctrl message and
> > MAC configuration
> > 3d8185022ec5 UBUNTU: SAUCE: platform/mellanox: mlxbf-tmfifo: Add
> > BlueField-3 support
> >
> > Upstream ones are using ACPI API and this one should be cherry-picked too.
> > 5db72fd ACPI: utils: Add acpi_dev_uid_to_integer() helper to get _UID
> > as integer
> >
> > [Test Case]
> > 1. Get the iperf data with current mlxbf_tmfifo module.
> > 2. Unload mlxbf_tmfifo and load the newly built module. Get the iperf data.
> > 3. Make sure it works and no performance degradation.
> >
> > Andy Shevchenko (1):
> >   ACPI: utils: Add acpi_dev_uid_to_integer() helper to get _UID as
> >     integer
> >
> > Haifei Luo (4):
> >   Revert "platform/mellanox: mlxbf-tmfifo: Drop Tx network packet when
> >     Tx TmFIFO is full"
> >   Revert "UBUNTU: SAUCE: mlxbf-tmfifo: fix potential race"
> >   Revert "UBUNTU: SAUCE: platform/mellanox: Add ctrl message and MAC
> >     configuration"
> >   Revert "UBUNTU: SAUCE: platform/mellanox: mlxbf-tmfifo: Add
> >     BlueField-3 support"
> >
> > Liming Sun (4):
> >   platform/mellanox: Add BlueField-3 support in the tmfifo driver
> >   UBUNTU: SAUCE: platform/mellanox: Add ctrl message and MAC
> >     configuration
> >   platform/mellanox: fix potential race in mlxbf-tmfifo driver
> >   platform/mellanox: mlxbf-tmfifo: Drop Tx network packet when Tx TmFIFO
> >     is full
> >
> >  drivers/acpi/utils.c                          |  24 ++++
> >  drivers/platform/mellanox/mlxbf-tmfifo-regs.h |  16 +--
> >  drivers/platform/mellanox/mlxbf-tmfifo.c      | 107 +++++++++---------
> >  include/acpi/acpi_bus.h                       |   1 +
> >  include/linux/acpi.h                          |   5 +
> >  5 files changed, 92 insertions(+), 61 deletions(-)
> >
> > --
> > 2.34.1
> >
> >
> > --
> > kernel-team mailing list
> > kernel-team at lists.ubuntu.com
> > https://lists.ubuntu.com/mailman/listinfo/kernel-team
> 
> Thanks for sending a v3.
> The patchset looks good, however I noticed that the commit:
> - bc05ea63b394 platform/mellanox: Add BlueField-3 support in the tmfifo driver
> has a follow-up in mainline, which fixes some kernel-doc warnings:
> - cbf92564100c platform/mellanox: tmfifo: fix kernel-doc warnings Do you think
> it's worth it to include the follow-up in the patchset?
> 

Hi Massimiliano,

I believe there are other upstream patches for mlxbf-tmfifo driver that are not yet
in the Jammy kernel.  With that in mind, I would prefer to keep this patch series as-is,
and we can make a note to include the "fix kernel-doc warnings" upstream patch
next time.  We are hoping to get this mlxbf-tmfifo driver patch series into the 
April SRU kernel and would prefer not to add another patch unless necessary.

Is that ok with you?

 - Dave

> Thanks.
> 
> --
> Massimiliano Pellizzer


More information about the kernel-team mailing list