Date: Thu, 5 Nov 2020 09:34:45 +0500 From: "Eugene M. Zheganin" <emz@norma.perm.ru> To: freebsd-stable <freebsd-stable@freebsd.org> Cc: freebsd-virt <freebsd-virt@freebsd.org> Subject: FreeBSD 12.x, virtio and alicloud (aliyun.com) Message-ID: <b31d221c-0763-b8b0-57d9-28799f5fc49b@norma.perm.ru>
next in thread | raw e-mail | index | archive | help
Hello, Guys, does anyone have VM running in AliCloud Chinese provider (one of the biggest, if not the biggest one) ? They seem to provide stock FreeBSD 11.x images on some Redhat-based Linux with VirtIO which run just fine (at least I take a look at their kernel and it seem to be a stock GENERIC), but after source uprgading to 12.2 it cannot mountroot, because literally no disks are found after thee kernel booting stage. This, in turn, is cause by a bunch of repeatable virtio errors, which looks like (screenshot provided in the link): virtio_pci1: cannot map I/O space device_attach: virtio_pci1 attach returned 6 (https://enazadev.ru/stub-data/freebsd12-alicloud-cannot-map-io.png) So not only vtbd0 cannot be attached to, but also a network adater. Surprisingly, virtio_console and memory baloon device seems to be working. I've took a look at various VirtIO cases in the bug tracker and compiled a kernel without netmap (yeah, after some consideration this could help only with virtio_net part), but this doesn't help. Is this some sort of regression that needs to be reported ? Is there some kind of known workaround ? I also have a running 11.3 on a second VM, so I can provide any necessary details if needed. Thanks. Eugene.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?b31d221c-0763-b8b0-57d9-28799f5fc49b>