Date: Tue, 28 Jul 2020 16:54:25 -0400 From: Vivek Goyal Message-ID: <20200728205425.GE78409@redhat.com> References: <20200727190223.422280-1-stefanha@redhat.com> <20200727190223.422280-4-stefanha@redhat.com> <20200728131250.GB78409@redhat.com> <20200728155233.GC3443476@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Disposition: inline Content-Transfer-Encoding: quoted-printable In-Reply-To: <20200728155233.GC3443476@redhat.com> Subject: Re: [Virtio-fs] [PATCH v2 3/3] virtiofsd: probe unshare(CLONE_FS) and print an error List-Id: Development discussions about virtio-fs List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Daniel =?iso-8859-1?Q?P=2E_Berrang=E9?= Cc: "vromanso@redhat.com" , "mpatel@redhat.com" , "qemu-devel@nongnu.org" , "virtio-fs@redhat.com" , "misono.tomohiro@fujitsu.com" , Roman Mohr On Tue, Jul 28, 2020 at 04:52:33PM +0100, Daniel P. Berrang=E9 wrote: > On Tue, Jul 28, 2020 at 09:12:50AM -0400, Vivek Goyal wrote: > > On Tue, Jul 28, 2020 at 12:00:20PM +0200, Roman Mohr wrote: > > > On Tue, Jul 28, 2020 at 3:07 AM misono.tomohiro@fujitsu.com < > > > misono.tomohiro@fujitsu.com> wrote: > > >=20 > > > > > Subject: [PATCH v2 3/3] virtiofsd: probe unshare(CLONE_FS) and pr= int an > > > > error > > > > > > > > > > An assertion failure is raised during request processing if > > > > > unshare(CLONE_FS) fails. Implement a probe at startup so the prob= lem can > > > > > be detected right away. > > > > > > > > > > Unfortunately Docker/Moby does not include unshare in the seccomp= .json > > > > > list unless CAP_SYS_ADMIN is given. Other seccomp.json lists alwa= ys > > > > > include unshare (e.g. podman is unaffected): > > > > > > > > > https://raw.githubusercontent.com/seccomp/containers-golang/master/= seccomp.json > > > > > > > > > > Use "docker run --security-opt seccomp=3Dpath/to/seccomp.json ...= " if the > > > > > default seccomp.json is missing unshare. > > > > > > > > Hi, sorry for a bit late. > > > > > > > > unshare() was added to fix xattr problem: > > > > > > > > https://github.com/qemu/qemu/commit/bdfd66788349acc43cd3f1298718ad4= 91663cfcc# > > > > In theory we don't need to call unshare if xattr is disabled, but i= t is > > > > hard to get to know > > > > if xattr is enabled or disabled in fv_queue_worker(), right? > > > > > > > > > > > In kubevirt we want to run virtiofsd in containers. We would already = not > > > have xattr support for e.g. overlayfs in the VM after this patch seri= es (an > > > acceptable con at least for us right now). > > > If we can get rid of the unshare (and potentially of needing root) th= at > > > would be great. We always assume that everything which we run in cont= ainers > > > should work for cri-o and docker. > >=20 > > But cri-o and docker containers run as root, isn't it? (or atleast have > > the capability to run as root). Havind said that, it will be nice to be= able > > to run virtiofsd without root.=20 > >=20 > > There are few hurdles though. > >=20 > > - For file creation, we switch uid/gid (seteuid/setegid) and that seems > > to require root. If we were to run unpriviliged, probably all files > > on host will have to be owned by unpriviliged user and guest visible > > uid/gid will have to be stored in xattrs. I think virtfs supports > > something similar. >=20 > I think I've mentioned before, 9p virtfs supports different modes, > passthrough, squashed or remapped. >=20 > passthrough should be reasonably straightforward to support in virtiofs. > The guest sees all the host UID/GIDs ownership as normal, and can read > any files the host user can read, but are obviously restricted to write > to only the files that host user can write too. No DAC-OVERRIDE facility > in essence. You'll just get EPERM, which is fine. This simple passthrough > scenario would be just what's desired for a typical desktop virt use > cases, where you want to share part/all of your home dir with a guest for > easy file access. Personally this is the mode I'd be most interested in > seeing provided for unprivileged virtiofsd usage. Interesting. So passthrough will have two sub modes. priviliged and unpriviliged. As of now we support priviliged passthrough.=20 I guess it does make sense to look into unpriviliged passthrough and see what other operations will not be allowed. Thanks Vivek >=20 > squash is similar to passthrough, except the guest sees everything > as owned by the same user. This can be surprising as the guest might > see a file owned by them, but not be able to write to it, as on the > host its actually owned by some other user. Fairly niche use case > I think. >=20 > remapping would be needed for a more general purpose use cases > allowing the guest to do arbitrary UID/GID changes, but on the host > everything is still stored as one user and remapped somehow. >=20 > The main challenge for all the unprivileged scenarios is safety of > the sandbox, to avoid risk of guests escaping to access files outside > of the exported dir via symlink attacks or similar. >=20 >=20 >=20 > Regards, > Daniel > --=20 > |: https://berrange.com -o- https://www.flickr.com/photos/dberran= ge :| > |: https://libvirt.org -o- https://fstop138.berrange.c= om :| > |: https://entangle-photo.org -o- https://www.instagram.com/dberran= ge :|