User Details
- User Since
- Jan 4 2016, 7:08 PM (567 w, 3 d)
Wed, Nov 18
I'm sorry for keeping you waiting. Let's go with current version to get the lock issue fixed. Any changes resulting from the UX and firmware link-management discussion can be handled in a follow-up patch.
Thu, Nov 12
Tue, Nov 10
LGTM assuming commit message is fixed, as pointed out by kbowling.
Mon, Nov 9
instead of:
Fri, Nov 6
Thank you for looking into this! In general it looks good to me. I'm only wondering if tunneled packets are handled correctly. With ICE_RSS_ANY_HEADERS set HW uses inner header fields for hash calculation, but ice_ptype_to_hash uses outer ip_ver and inner protocol so e.g for IPv6 --> GRE/NAT -> IPv4 -> UDP we will get M_HASHTYPE_RSS_IPV6. Does it make sense?
Thu, Nov 5
Tue, Nov 3
Ouch :( my apologies for this oversight. ixgbe_apply_ldo_link_state is guarded by !sc->ldo_valid so it should be fine.
Mon, Nov 2
Fri, Oct 30
Thu, Oct 29
Wed, Oct 21
Oct 15 2026
The ICE version of the sparse check does the right thing. Both the X700 and E800 generations of adapters have the same limits: a header can span at most three descriptors, and a single frame can span up to eight. I guess this wasn't backported from ICE to IXL (and IAVF?) due to my oversight.
Oct 12 2026
I'm seeing this kernel panic when unloading IAVF. A queue interrupt firing after iflib_stop() shouldn't be possible, so I'm investigating the issue in the driver itself. However, I think this patch makes sense on its own.
Oct 6 2026
- Implement fernape@ suggestion.
- Revert the sizeof change as new version was less readable.
Oct 5 2026
Sep 29 2026
- Collapse the bootp.c TAG_ROOTPATH branch to a plain strlcpy
Sep 28 2026
Sep 25 2026
- Address @imp comment.
Sep 24 2026
Sep 18 2026
Address @kbowling comment.
Sep 10 2026
@kbowling - This should fix both issues you have found: https://reviews.freebsd.org/D59578
Sep 9 2026
The problem comes from https://reviews.freebsd.org/D58149. I don't know how it slipped through internal testing, but I think I found a root cause. I'll submit a fix for review this week.
Sep 7 2026
This is odd. I can reproduce this also on E830 QSFP adapters, which I used quite recently without any link issues. I'm looking into which patch might have caused the regression.
Sep 2 2026
Sep 1 2026
Aug 21 2026
Thanks a lot for all your fixes. I'm building an image for our validation team to run some regression tests. I'll get back to you as soon as I get any feedback.
Aug 20 2026
D54096 resolves that at iflib level.
Aug 18 2026
Aug 17 2026
Aug 12 2026
This is great! Both as a patch and as an advertisement for the fail(9). Thank you!
Jul 10 2026
Jul 2 2026
Jun 29 2026
Jun 26 2026
Jun 24 2026
I'm under impression that we use only TC0, but I need to refresh my memory. Do you actually see non-zero values in registers MPC[1-8]?
Jun 12 2026
Shouldn't this be actually handled by this condition? (https://cgit.freebsd.org/src/tree/sys/net/iflib.c#n2969)
/* was this only a completion queue message? */ if (__predict_false(ri.iri_nfrags == 0)) continue;
I mean if a packet is discarded does it need to use any iri_frags?
@cperciva Could you please take a look?
May 29 2026
May 21 2026
It's submitted also to Forgejo: https://ron-dev.freebsd.org/FreeBSD/src/pulls/32, but I guess Phabricator is still preferred.
Apr 21 2026
Apr 9 2026
I'm testing that with GENERIC so with no RSS option. Without the patch both locally generated (iperf3) and forwarded (pkt-gen) UDP traffic is send using TX 0. I did a quick run on vanilla CURRENT with RSS enabled and it looks that in that case also only one TX queue is used, though it's not always queue 0. I haven't tried IPv6 yet.
Mar 27 2026
Mar 25 2026
- Add FW logging to features supported and enabled
on E610 adapters.
