-----Original Message-----
From: Eric Biggers <ebiggers@kernel.org>
Sent: Thursday, May 15, 2025 10:13 PM
To: Bart Van Assche <bvanassche@acm.org>
Cc: netdev@vger.kernel.org; Bernard Metzler <redacted>; linux-
nvme@lists.infradead.org; linux-sctp@vger.kernel.org; linux-
rdma@vger.kernel.org; linux-kernel@vger.kernel.org; Daniel Borkmann
[off-list ref]; Marcelo Ricardo Leitner
[off-list ref]; Sagi Grimberg [off-list ref]; Ard
Biesheuvel [off-list ref]
Subject: [EXTERNAL] Re: [PATCH net-next 04/10] RDMA/siw: use skb_crc32c()
instead of __skb_checksum()

On Thu, May 15, 2025 at 01:02:20PM -0700, Bart Van Assche wrote:

quoted

On 5/10/25 5:41 PM, Eric Biggers wrote:

quoted

Instead of calling __skb_checksum() with a skb_checksum_ops struct that
does CRC32C, just call the new function skb_crc32c().  This is faster
and simpler.

Bernard, since you are the owner and author of the siw driver, please

help

quoted

with reviewing this patch.

Eric, do you already have a test case for the siw driver? If not,
multiple tests in the blktests framework use this driver intensively,
including the SRP tests that I wrote myself. See also
https% 

3A__github.com_osandov_blktests&d=DwIBAg&c=BSDicqBQBDjDI9RkVyTcHQ&r=4ynb4Sj
_4MUcZXbhvovE4tYSbqxyOwdSiLedP4yO55g&m=42RdxLC4OJSKYtu0EOtGxON-
daxZpwd8xG9briOxYyywxa7RWrL-
KukLFCucHn_c&s=dtz3xdRBYawI7d1FpWaWUV_QhDeExTyHGbPMZLxpzPQ&e= .

No.  I'll try that out when I have a chance.

If the developers/maintainers of the driver could help test it, that would
be a
lot easier though.  I've been cleaning up the CRC calls across the whole
kernel,
and it gets time-consuming when individual subsystems insist on me running
their
custom test suite(s) and providing subsystem-specific benchmarks.

- Eric

I'll definitively test it. Sorry for the delay. You will hear from me
next days. If it comes to things like checksums, I always try to get
a setup where I can test against RMDA hardware (Chelsio). Hope to
find time over the WE.

Thanks,
Bernard.