Details to be added after investigation; submitting this as a placeholder to track the issue.
On what architecture? HardenedBSD has been using lld as the default linker on amd64 since March of 2017 (commit f38b51668efcd53b8146789010611a4632cafade). I've been using VIMAGE for multiple years. I haven't had a single issue.
A commit references this bug: Author: emaste Date: Thu Jan 11 19:08:43 UTC 2018 New revision: 327840 URL: https://svnweb.freebsd.org/changeset/base/327840 Log: Temporarily disable VIMAGE on i386 An lld-linked i386 kernel hangs on boot, after em(4) starts. This seems similar to the issue that prompted us to disable VIMAGE on arm64 in r326179. Disable VIMAGE on i386 for now while we investigate. PR: 225077 Sponsored by: The FreeBSD Foundation Changes: head/sys/i386/conf/GENERIC
A commit references this bug: Author: emaste Date: Sun Jan 14 16:04:52 UTC 2018 New revision: 327969 URL: https://svnweb.freebsd.org/changeset/base/327969 Log: Enable VIMAGE in i386 GENERIC (revert r327840) We've switched back to ld.bfd on i386 for now. PR: 225077 Sponsored by: The FreeBSD Foundation Changes: head/sys/i386/conf/GENERIC
(In reply to Shawn Webb from comment #1) On i386, as in the summary line above.
I'm not able to reproduce this using bhyve with e1000 emulation. This is on HEAD as of about two days ago. I built the world using WITH_LLD_IS_LD= and WITH_LLD_BOOTSTRAP=. Is anyone else able to confirm that the bug is still present?
(In reply to Mark Johnston from comment #5) have you tried with epair built as module and loaded as module?
(In reply to Bjoern A. Zeeb from comment #6) Loaded from loader.conf? I tried that, still no hang.
Still no luck with this one. I tested on a laptop as well with a variety of kernel configs. Is it possible that this has been fixed in one of the LLVM updates since the PR was submitted? Are there any special steps needed to reproduce the issue?
It looks like the change to enable lld as the bootstrap linker on i386, and the revert thereof, happened before clang/llvm/lld 6.0 came to head. I am trying to reproduce locally now.
I cannot reproduce this problem now.