https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=248652
--- Comment #29 from Sylvain Galliano ---
(In reply to Vincenzo Maffione from comment #28)
Yes, this is much better:
6 queues, nm_tx_tmr_us=5:
983.492185 main_thread [2639] 37.907 Mpps (37.945 Mpkts 18.196 Gbps in 1001000
usec) 512.00 avg
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=248652
--- Comment #28 from Vincenzo Maffione ---
(In reply to Sylvain Galliano from comment #26)
Thanks.
I just figured out that there may be a major flaw introduced by the porting of
ixl to iflib. This flaw should couse too many writebacks from
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=248652
Vincenzo Maffione changed:
What|Removed |Added
Attachment #218866|0 |1
is obsolete|
Hi,
I was recently doing some tests and found out that on the iflib e1000
driver I can do much less media status checks per second compared to the
non-iflib driver. I found out that this is caused by the definition of
safe_pause_us() which calls:
pause("e1000_delay", max(1, x/(100/hz)));
If
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=250366
Michael Tuexen changed:
What|Removed |Added
CC||n...@freebsd.org,
https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=248652
--- Comment #26 from Sylvain Galliano ---
(In reply to Vincenzo Maffione from comment #25)
you are correct, using 1 queue on both ix/ixl NIC + tuning new sysctl, we have
same result as FreeBSD 11
Regarding avg_batch values on ixl, they ar