On Wed, Jun 21, 2023 at 10:54:00AM +0200, Borislav Petkov wrote: > On Tue, Jun 20, 2023 at 03:43:15PM -0500, Michael Roth wrote: > > Basically, arch/x86/coco/Makefile is never processed if arch/x86/Kbuild > > indicates that CONFIG_HAS_CC_PLATFORM is not set. So if we want to have > > stuff in arch/x86/coco/Makefile that build for !CONFIG_HAS_CC_PLATFORM, > > like SNP host support, which does not rely on CONFIG_HAS_CC_PLATFORM > > being set, that check needs to be moved down into arch/x86/coco/Makefile. > > Ok, so if you put SNP host support into arch/x86/virt/svm/sev.c, that > should work too and won't have any relation to CONFIG_HAS_CC_PLATFORM, > right? Right, that works out just as well, and ends up being a bit more straightforward. I have it implemented here: https://github.com/mdroth/linux/commits/snp-host-latest-v9b https://github.com/mdroth/linux/commit/a889a2dd64b62d9c3bf74cf02e7d8d71c7061667 and dropped the patch that reworks arch/x86/coco/Makefile. Thanks, Mike > > The CC_PLATFORM thing is a way to check for confidential computing guest > features by abstracting the capabilities so that you don't have to check > *each* and *every* conf guest type in the conditionals and thus go nuts. > > Thx. > > -- > Regards/Gruss, > Boris. > > https://people.kernel.org/tglx/notes-about-netiquette
