Upgrading 24.1.10 to 24.7: kernel panic and reset (SYS-E300-9A-4C test setup)

[franco]

> Have you checked how much your fork differs from upstream's sys/kernel? I don't think expecting users to cherry pick commits (or go through your cherry picking history) is a realistic approach.

I think you have that logic backwards.


Cheers,
Franco

[blblblb]

I wasn't sure if I was reading gonzopancho or you Franco 

No, really. Users can do free QA, are you also expecting them to be developers (kernel developers at that! -which you are not, also-)  for the foundations of your commercial product? That is really all that usually happens with OPNsense: the free users are guinea pigs for the commercial offering. In this aspect, pfSense is doing the same.

No need to be conceited or abrasive about it.

[doktornotor]

Gonzo, is that you?

Let's clarify this - those Netgate devs with FreeBSD commit bit seem to use every FreeBSD user as a guinea pig for their experiments lately. Such as the recent "stateful ICMP" SNAFU.

BTW, "blb" means a dumba*s in Czech. Nomen omen, it seems.

[blblblb]

Gonzo, is that you?

Let's clarify this - those Netgate devs with FreeBSD commit bit seem to use every FreeBSD user as a guinea pig for their experiments lately. Such as the recent "stateful ICMP" SNAFU.

BTW, "blb" means a dumba*s in Czech. Nomen omen, it seems.

Anything productive to add to the discussion? Those "Netgate devs with commit privileges" were FreeBSD developers before working for Netgate. Judging by your post history and attitude, are you sure you are not talking about yourself? You seem to have a problem grasping a whole lot of things, and engage strangers online aggressively. Copium much?

[doktornotor]

Anything productive to add to the discussion?

Nah, not really... GIGO.

[chemlud]

pfsense fanboy trash currently on high level here :-D

[franco]

Boy, that escalated quickly. Let's give blblblb an opportunity to breathe.


Cheers,
Franco

[doktornotor]

Well, I already pledged to do my best to avoid future interactions with the Netgate-affiliated folks Will gladly extend that to this forum as well. 

[franco]

And to be fair I'll explain:

>> Have you checked how much your fork differs from upstream's sys/kernel? I don't think expecting users to cherry pick commits (or go through your cherry picking history) is a realistic approach.
>
> I think you have that logic backwards.

The reason why I said this is because I asked for a panic capture but received nothing so far.

If this is a problem introduced in FreeBSD 14.1 I will take my chances that the issues lies with FreeBSD 14.1 (vs. 13.2), not OPNsense 24.7 (vs 24.1). We have several points of reference for this stance already in the short time we are on FreeBSD 14.1 and our kernel is already more stable than FreeBSD 14.1 ever will be although we did help to bring some of these things to FreeBSD for later releases like 14.2 and 15.0

And given a panic that could be caused by us we would assume a fix could be written... is that such a novel idea?


Cheers,
Franco

[ScottD]

I ran into the same panic and investigated further, with some help from Franco. I've raised FreeBSD bug and proposed a patch in the cxgb driver to fix the issue. That fix has been accepted by the upstream maintainer, and committed to FreeBSD 15.0, with a planned backport to 14.x.
As Franco suggested in OPNSense issue 224, I've submitted a pull request to OPNSense with the same fix, so we may see it sooner in OPNSense 24.7 than in FreeBSD 14.

[franco]

Yes, Scott's fix will be part of 24.7.8 tomorrow.


Cheers,
Franco