On Wed, 2021-01-13 at 13:40 +0000, David Howells wrote: > Hi Linus, > > Are you willing to take this between merge windows - or does it need > to wait for the next merge window? It's not technically a bug fix to > the kernel, but it does have a CVE attached to it. > > Note that I've also updated Jarkko's address in his Reviewed-by since > his Intel address no longer works.
Sorry, late to the party. I suppose I lost the argument that we shouldn't really be trusting any certs from db when shim is in operation because they're all EFI binary signing ones and will usually simply be the microsoft certificate and possibly an OEM platform one and we're usually pivoting the root of trust to the certificates in the MokList. However, if we are going to do this, we should also be blacklisting the certificates in MokListX which the OS sees through MokListXRT. Since MokListX is an essential piece of our revocation infrastructure it should have been mentioned in the CVE but wasn't for some reason. James