Bernd Walter wrote: > On Tue, Jul 15, 2003 at 12:22:31PM -0700, Tim Kientzle wrote: > > Harald Schmalzbauer wrote: > > >The problem is solved. It was stupid, but I thought why should > I have to > > >set > > >/ in /etc/fstab when the filesystem isn't mounted yet, so the > file can't be > > >read. > > >But it seems the kernel reads this file "loader-like" *before* the > > >filesystem is mounted. > > > > I believe that the loader actually reads this file (it > > has to be able to locate and read files from UFS anyway > > in order to load the kernel) and passes the root f/s > > information to the kernel when it boots. This should > > probably be documented in fstab(5) > > You can pass a filesystem at the loader prompt: > vfs.root.mountfrom="" > > > >Considering the above to be correct I can't understand the > ability to enter > > >e. g. ufs:/dev/ad0a at "mountroot>" when it doesn't work. > > > > I've also been stung by the fact that the > > "mountroot>" prompt is broken. <sigh> > > I looked briefly at the code, but the > > bug is not particularly obvious. > > At least it wasn't broken a while ago. > What happens exactly after entering the correct device? The machine rebooted. No matter if I did "?" or any "ufs:xxYz". It's behaviour was like "empty line". That was mine dunno about Tim's Regarding loader reades /etc/fstab: Of course loader and even boot can read from ufs but I verified that current device was disk0s1 (where my 165-ID was) and also something like root-mount was set to disk0s1. Next thing is that boot0 could'nt boot my kernel (5.1-release). Some time ago I built a little accesspoint with 4.6 and there was no problem booting the kernel with only boot0, no loader. Has this feature been intentionally removed? Best regards, -Harry > > -- > B.Walter BWCT http://www.bwct.de > ticso_at_bwct.de info_at_bwct.de >Received on Tue Jul 15 2003 - 14:48:32 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:37:15 UTC