On Thu, 2020-05-14 at 13:31 +0200, Ard Biesheuvel wrote: > On Thu, 14 May 2020 at 13:28, Loïc Yhuel <loic.yhuel@xxxxxxxxx> wrote: > > Le jeu. 14 mai 2020 à 12:54, Jarkko Sakkinen > > <jarkko.sakkinen@xxxxxxxxxxxxxxx> a écrit : > > > So it is clear that "pr_warn(FW_BUG ..." would be a sane to have there. > > So only to tell the UEFI might have logged events the kernel can't read ? > > There is no warning if the table is missing, which would have the same result. > > > > I can try to dump it, perhaps it is using the SHA-1 log format. > > If so, would a patch to support this non-standard behavior be accepted ? > > That is why I was asking the question: what exact condition should we > warn about? And at which point? Always when final table is missing there should be some sort of notification because the event log is incomplete. I.e. it misses PCR5 extends from GetEventLog(). No additional info, just a note that we don't have the tail. I'm fine with info level message too. /Jarkko
