From nobody Sat Apr 20 02:05:17 2024 X-Original-To: freebsd-hackers@mlmmj.nyi.freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2610:1c1:1:606c::19:1]) by mlmmj.nyi.freebsd.org (Postfix) with ESMTP id 4VLvww0Plhz5GkW1 for ; Sat, 20 Apr 2024 02:05:20 +0000 (UTC) (envelope-from jeffpc@josefsipek.net) Received: from smtp.jeffnet.31bits.net (josefsipek.net [71.174.62.3]) by mx1.freebsd.org (Postfix) with ESMTP id 4VLvwv5KMdz4mkH for ; Sat, 20 Apr 2024 02:05:19 +0000 (UTC) (envelope-from jeffpc@josefsipek.net) Authentication-Results: mx1.freebsd.org; none Received: from satis (satis [172.27.0.85]) by smtp.jeffnet.31bits.net (Postfix) with ESMTPSA id 0B6172BE8A; Sat, 20 Apr 2024 02:05:19 +0000 (UTC) Date: Fri, 19 Apr 2024 22:05:17 -0400 From: Josef 'Jeff' Sipek To: Warner Losh Cc: FreeBSD Hackers Subject: Re: Upgrading -RELEASE to -CURRENT Message-ID: References: List-Id: Technical discussions relating to FreeBSD List-Archive: https://lists.freebsd.org/archives/freebsd-hackers List-Help: List-Post: List-Subscribe: List-Unsubscribe: Sender: owner-freebsd-hackers@FreeBSD.org MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: X-Spamd-Bar: ---- X-Rspamd-Pre-Result: action=no action; module=replies; Message is reply to one we originated X-Spamd-Result: default: False [-4.00 / 15.00]; REPLY(-4.00)[]; ASN(0.00)[asn:701, ipnet:71.174.0.0/16, country:US] X-Rspamd-Queue-Id: 4VLvwv5KMdz4mkH On Fri, Apr 19, 2024 at 18:15:59 -0600, Warner Losh wrote: > On Fri, Apr 19, 2024, 6:08 PM void wrote: > > > On Fri, Apr 19, 2024 at 06:46:04PM -0400, Josef 'Jeff' Sipek wrote: > > > > >Is upgrading from 14.0-RELEASE directly to -CURRENT supposed to work? I > > >didn't see anything in the docs that says it shouldn't. > > > > I think the general principle is "upgrade to latest version of thing > > before upgrading major version" so for 14.0 that would involve building & > > installing 14-stable and then 15-current. > > > > > Last release to current almost always works. So 14.0 to current is doable. > I did 14.0 beta1 -> current 3 weeks ago. > > Etcupdate pre mode, Buildworld, buildkernel, installkernel, reboot, > installworld, etcupdate, reboot -r > > But OP said they did that twice, once for stale/14 and once for main so I'm > not sure what is going on. Need error message i think. Ok, trying again: 0. install from ISO & freebsd-update to 14.0-RELEASE-p6; install git from pkg 1. git clone main (1bd4f769caf) into /usr/src 2. cd /usr/src 3. make buildworld -j48 First build I get a fairly prompt failure: sh: llvm-min-tblgen: not found Some flailing around with 'make clean' and retrying the same buildworld gets past it. At this point, I have llvm-min-tblgen{,.debug} in /usr/obj/... (I'm assuming that there is some missing dependency in tools build.) Weirdly enough, from this point on, if blow everything away (git clean -fdx; rm -rf /usr/obj/*), the subsequent 'make buildworld -j48' is fine. So, a. the first time(?) I try buildworld, it fails because of missing llvm-min-tblgen b. subsequent buildworlds complete fine even if there is no OBJDIR and the tree is clean 4. make buildkernel -j48 5. etcupdate -p 6. make installkernel -j48 7. reboot (into single user mode) 8. zfs set readonly=off zroot/ROOT/default 9. zfs mount -a 10. cd /usr/src 11. make installworld -j48 12. etcupdate -B This prints "Failed to build new tree." and exits with 1 13. cc ld-elf.so.1: /lib/libcxxrt.so.1: version CXXABI_1.3.11 required by /lib/libc++.so.1 not found 14. reboot -r (sort of pointless to mention given that the system is already in a broken state) (At this point, cc, and other C++-based programs fail to run.) So, it looks like 'etcupdate -B' is upset for whatever reason. I'll try to poke at various things, but I've seen these two failures (llvm-min-tblgen and CXXABI) about half a dozen times when I try to go from 14.0-RELEASE to -CURRENT, but *zero* times if I make a detour through -STABLE. Thanks, Jeff. P.S. I've seen the same exact failures on my "antique" 4-core Nehalem. So, I don't think it has to do with the high parallelism (-j48) in the above steps.