Date: Mon, 26 Sep 2016 22:21:48 +0800 From: Stephan CHEDLIVILI <stephan@theched.org> To: "freebsd-virtualization@freebsd.org" <freebsd-virtualization@freebsd.org> Subject: Re: UEFI bhyve and EFI shell at boot Message-ID: <3257247.Qm6eVUophb@panda.test.me> In-Reply-To: <20160926123157.GO97879@e-new.0x20.net> References: <9276239.oRRSstIAVb@panda.test.me> <CY4PR12MB14478911B5A62D9C8FB8F86FA3CD0@CY4PR12MB1447.namprd12.prod.outlook.com> <20160926123157.GO97879@e-new.0x20.net>
next in thread | previous in thread | raw e-mail | index | archive | help
On Monday 26 September 2016 14:31:57 Lars Engels wrote: > On Mon, Sep 26, 2016 at 05:56:59AM +0000, Justin Holcomb wrote: > > > From: owner-freebsd-virtualization@freebsd.org > > > <owner-freebsd-virtualization@freebsd.org> on behalf of Stephan > > > CHEDLIVILI <stephan@theched.org> Sent: Saturday, September 24, 2016 > > > 12:50 AM > > > To: freebsd-virtualization@freebsd.org > > > Subject: UEFI bhyve and EFI shell at boot > > > > > > Hi gents, > > > > > > I was giving a try to the UEFI-GOP on a FreeBSD 11.0-RC3. Launching the > > > install of, let's say a Debian works fine and I can attach a VNC viewer > > > for the progress. > > > > > > All is fine , even rebooting after the installation is finished I can > > > log in Debian. > > > > > > However, when I do a bhyvectl --destroy --vm=xxxxxxx and I try to reboot > > > the VM and it greets me with the error message "Boot failed, EFI > > > Harddrive" at boot and sends me to the EFI shell. > > > > > > I then have to manually use the shell menu to launch the boot via the > > > ad-hoc file (/boot/efi/efi/debian/grubx64.efi) and it boot flawlessly. > > > > > > And of course, the same error happens after I reboot the FreeBSD host > > > machine > > > > > > Is there somethign I am missing here ? > > > > > > Thanks for this admirable piece of work ! > > > > > > -Stephan > > > > Stephan, > > > > I have also experienced this as well. My scriptable work around was to > > start the guest with a rEFInd ISO[1] instead a 'null.iso'. rEFInd sees > > the Debian installation on the image/volume and will boot from it after > > the 15 seconds timeout elapses. > > > > As for the why... my rudimentary understanding is the Debian installation > > creates and relies on the UEFI boot entry it creates during installation. > > However that entry is forgotten once the guest's VMM resources are > > reclaimed as the UEFI environment is not saved and is reloaded exactly > > from the UEFI ROM file (not from the previous state). > > > > -Justin D Holcomb > > > > [1] http://www.rodsbooks.com/refind/getting.html > > That's also true for Ubuntu 16.04 But this problem does not happen with Fedora 4.7
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?3257247.Qm6eVUophb>