On Mon, 2006-10-30 at 21:19 +0200, Yoshinori K. Okuji wrote: > On Saturday 28 October 2006 20:35, Hollis Blanchard wrote: > > On Sat, 2006-10-28 at 16:16 +0200, Yoshinori K. Okuji wrote: > > > On Friday 27 October 2006 21:19, Hollis Blanchard wrote: > > > > Comments please. Also see next mail to see how the PPC Linux loader > > > > uses it. > > > > > > So this still loads an image to paddr directly, right? > > > > > > I think we need these features: > > > > > > - an ability to pass an address to load an image > > > > > > - an ability to pass the size of a loading area > > > > > > - an ability to get the size of a loaded image > > > > The PowerPC loader patch I sent already does 1 and 3, and you can easily > > see how 2 can be done in the exact same way. Please look at that code, > > in particular 'calcsize' and 'offset_paddr'. > > As for calcsize, I do not think it is a good idea to put this to > architecture-specific code. We always need this feature, so it should be a > part of the API. > > As for offset_paddr, I do not like modifying a member of a header. This > changes a read value, so the state of a header changes after calling > grub_elf*_load. This kind of side effect often makes debugging harder. I > prefer to specify an address more explicitly.
What exactly would you suggest? Using a hook to modify the struct phdr (or veto it) seems quite elegant to me. -Hollis _______________________________________________ Grub-devel mailing list Grub-devel@gnu.org http://lists.gnu.org/mailman/listinfo/grub-devel
