On Mon, Feb 10, 2025 at 07:02:04PM +0100, Eric Dumazet wrote:

On Mon, Feb 10, 2025 at 6:59???PM Shradha Gupta
[off-list ref] wrote:

quoted

On Mon, Feb 10, 2025 at 09:57:53AM -0800, Shradha Gupta wrote:

quoted

On Mon, Feb 10, 2025 at 04:55:54PM +0100, Eric Dumazet wrote:

quoted

On Mon, Feb 10, 2025 at 5:40???AM Shradha Gupta
[off-list ref] wrote:

quoted

Allow the max aggregated pkt size to go up-to GSO_MAX_SIZE for MANA NIC.
This patch only increases the max allowable gso/gro pkt size for MANA
devices and does not change the defaults.
Following are the perf benefits by increasing the pkt aggregate size from
legacy gso_max_size value(64K) to newer one(up-to 511K)

for i in {1..10}; do netperf -t TCP_RR  -H 10.0.0.5 -p50000 -- -r80000,80000
-O MIN_LATENCY,P90_LATENCY,P99_LATENCY,THROUGHPUT|tail -1; done

Was this tested with IPv6 ?

Hi Eric,
yes, sanity and functional testing where performed(manually) and passed on azure
VMs with IPv6.

Forgot to mention that the tests were performed on both IPv4 and IPv6
and these numbers are from IPv4 testing

Where is the IPv6 jumbo header removed ?

I think this is missed in this patchset. In our IPv6 tests, patch sanity was
performed without changing the default values.
I will add this support, thorughly test IPv6 and send out another
version with complete IPv6 support and numbers.

Thanks for the pointers Eric.

Regards,
Shradha Gupta.