David Cantrell wrote on 2011-07-14: > Getting tboot included in the default package set is a FESCo decision. > If they agree, it would be added to the comps database accordingly. > Actually, it should be added to that database anyway, it's just > whether or not it would be installed by default or not. Does the default package set mean those packages included in the install iso image or those by default installed packages? So far, we could already yum install tboot for f15, does it mean it has already been added to the comps database right? I could not find it in f15 install iso image yet. > For the changes necessary in anaconda, I have the following questions: > > 1) How do we know that setting up tboot is appropriate for the system? > Is there something we can examine in /proc or /sys that tells us > whether or not tboot should be configured? Or (and this is what would > be really > nice) can we always set up tboot if the package is installed and then > tboot will either do something related to TXT or just fall through and > boot up normally on systems that lack TXT support? Correct. We can always set up tboot if the package is installed. Tboot will fall through and just boot up kernel on systems that lack TXT support. > 2) tboot is something that sits on top of grub, correct? Not > something that we use in place of grub on the appropriate systems? Correct. Tboot sits on top of grub. It just takes advantage of the multiboot capability of grub. > 3) The Fedora feature page you link to mentions possibly having to > provide a patch to grubby to handle tboot entries in grub.conf. This > isn't really an optional thing, grubby will need to be modified to > handle tboot settings so that they do not get lost in upgrades. The grubby already support updates for single level multiboot cases such as xen+linux or tboot+linux. What it lacks of is support for updating two level multiboot case such as tboot+xen+linux. The two level multiboot may not be a must to support in grubby, right? But anyway, I would try to make it after anaconda support done. > Thinking about it from an implementation standpoint, I do not feel > like this is too bad for anaconda. Assuming we can get answers to > these questions, it feels like an easy extension to the x86 boot loader class. Yes, I agree. I will try to make a patch after we get answers to those implementation questions. After all, thanks for the timely response. Jimmy _______________________________________________ Anaconda-devel-list mailing list Anaconda-devel-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/anaconda-devel-list
