Does the loader have enough information? Does it for instance know if NFS support is compiled into the kernel? Maybe set vfs.root.mountfrom should be an ordered list set by the loader. On Wed, 27 Oct 2004 08:55:48 +0200, Poul-Henning Kamp <phk_at_phk.freebsd.dk> wrote: > In message <200410261704.49182.jhb_at_FreeBSD.org>, John Baldwin writes: > > > >On Thursday 21 October 2004 05:24 pm, Poul-Henning Kamp wrote: > >> In message <200410211505.27635.jhb_at_FreeBSD.org>, John Baldwin writes: > >> >> Doesn't the search terminate if [0] is NULL ? > >> > > >> >Perhaps there should be a small wrapper function to register a candidate > >> > root with an associated priority and that wrapper function could then > >> > manage the rootdevnames[] array and keep them sorted based on the > >> > priority? > >> > >> No, the entire thing should be rototiled. > > > >Well, the interface I'm thinking of is something like > >'rootdevice_add(const char *name, int priority)' or maybe > >'rootdevice_add(dev_t dev, int priority)' (which won't work for NFS, so I > >guess back to the char * version) and not having the array or linked list or > >whatever of names visible outside vfs_mount.c (or where the root mount code > >lives). > > After looking again I think I'll change my position: This entire thing > should be yanked out and we should rely on the loader to set the right > variables instead. > > -- > > > Poul-Henning Kamp | UNIX since Zilog Zeus 3.20 > phk_at_FreeBSD.ORG | TCP/IP since RFC 956 > FreeBSD committer | BSD since 4.3-tahoe > Never attribute to malice what can adequately be explained by incompetence. >Received on Wed Oct 27 2004 - 14:24:36 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:38:19 UTC