[PATCH net-next v5 0/3] net: stmmac: Introduce hw queue priority offload
Daniel Thompson
daniel at riscstar.com
Tue Oct 6 09:18:14 PDT 2026
On Tue, Oct 06, 2026 at 02:21:28PM +0200, Lorenzo Bianconi wrote:
> > Hi!
> >
> > This is an automated message. This series looks like a fix, but its
> > commit messages seem to be missing some information:
> >
> > - Whether the issue was actually triggered, or is only theoretical
> > (e.g. found by code inspection). If it was triggered please include
> > the symptoms, like the stack trace or error messages.
>
> This is a theoretical issue found during code inspection.
I haven't triggered anything specific for the real_num_tx_queues changes
but the overall with/without mqprio offload changes can (and have) been
reproduced with netperf:
netperf -H 192.168.10.2 -s 1 -l 12 -y 7 -t UDP_STREAM -- -m 1472 & \
netperf -H 192.168.10.2 -s 2 -l 10 -y 5 -t UDP_STREAM -- -m 1472 & \
wait
With the patchset applied (and the traffic class 7 mapped to a higher
priority queue and TC 5) then the split of bandwidth massively favours
TC 7 (with a 2.5G link. TC 7 measure ~2G versus TC 5 at ~2M).
Without the patchset both streams get ~1G.
I guess FWIW the whole series is:
Tested-by: Daniel Thompson <daniel at riscstar.com>
Daniel.
More information about the linux-arm-kernel
mailing list