On Fri, 19 Apr 2019 at 22:04, Kinney, Michael D <michael.d.kin...@intel.com> wrote: > > Ard, > > Saw your patch to cache the GUID table. The ESRT Table > is not that much bigger, so the algorithm may be simpler > if you just make a copy of the ESRT table with the > active entries. > > * 16+24 bytes per ESRT entry. > * 16 bytes/entry for just the GUID. > > The only advantage of checking against the ESRT GUID at > RT is to reject capsules that will be rejected when the > capsule is processed later. This can prevent extra reboots > if an OS agent is sending capsules that do not really > apply to the current system. I expect OS agent to only > send capsules that are in the ESRT. >
Yeah. The only problem is that some capsules have no header at all, and so without a GUID check, we might end up interpreting random data as the capsule's HeaderSize field, which would be bad. In general, I think we should be robust against random junk being passed into the QueryCapsuleCapabilities() runtime service. -=-=-=-=-=-=-=-=-=-=-=- Groups.io Links: You receive all messages sent to this group. View/Reply Online (#39344): https://edk2.groups.io/g/devel/message/39344 Mute This Topic: https://groups.io/mt/31245843/21656 Group Owner: devel+ow...@edk2.groups.io Unsubscribe: https://edk2.groups.io/g/devel/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
