On Thu, Sep 18, 2025 at 10:01:23AM +0300, Tariq Toukan wrote:

From: Patrisious Haddad <redacted>

Write combining is an optimization feature in CPUs that is frequently
used by modern devices to generate 32 or 64 byte TLPs at the PCIe level.
These large TLPs allow certain optimizations in the driver to HW
communication that improve performance. As WC is unpredictable and
optional the HW designs all tolerate cases where combining doesn't
happen and simply experience a performance degradation.

...

Signed-off-by: Patrisious Haddad <redacted>
Reviewed-by: Michael Guralnik <redacted>
Reviewed-by: Moshe Shemesh <redacted>
Signed-off-by: Tariq Toukan <tariqt@nvidia.com>
---
 drivers/net/ethernet/mellanox/mlx5/core/wc.c | 28 ++++++++++++++++++--
 1 file changed, 26 insertions(+), 2 deletions(-)

Find V2 here:
https://lore.kernel.org/all/1757925308-614943-1-git-send-email-tariqt@nvidia.com/ (local)

V3:
- Move the new copy assembly code to be inline, within the same file it
  is used.
- Use ".arch_extension simd;\n\t" to avoid the need for separate file
  and special compilation flags.

Hi Tariq, Patrisious, all,

This is not a full review - although I've been following this patch with
much interest, I don't feel qualified to provide one.

But what I can do is be the poor soul to report that unfortunately
the patch doesn't compile against net-next. So at some point it will
need to be rebased and reposted.

Thanks for your persistence on working on this, it is indeed a nettlesome issue.

-- 
pw-bot: changes-requested