On 03/22/16 06:59, Jordan Justen wrote: > Why not a guid based event, similar to gIdleLoopEventGuid?
Two reasons: - specifically, it keeps the current structure of the AcpiPlatformDxe driver (i.e., we stick to the way the PCI bus driver installs gEfiPciEnumerationCompleteProtocolGuid, we just use a different GUID and tie it to a different circumstance) - (I vaguely think I might have mentioned this earlier) generally, I prefer protocol installations to events, in this kind of situation. The protocol stays there permanently, and in general it allows the "consuming" driver to load & start after the protocol is installed. The generic reason doesn't apply in this case, but I preferred to stick with the pattern. Do you have a strong reason to dislike it? Thanks Laszlo > On 2016-03-21 03:19:06, Laszlo Ersek wrote: >> On 03/21/16 10:58, Ard Biesheuvel wrote: >> >>> But by all means, let's decouple this from the review of this series. >> >> Thanks. Two recent Protocol/*.h additions were a48d0a3d72f9 and >> 78741ce91e12, those also added the extern declarations. I agree the >> declarations shouldn't be duplicated, but then we should all start >> following this new rule. >> > > Maybe one of you should post an email with a more noticable subject to > prompt discussion of this topic more globally on edk2-devel? > > -Jordan > _______________________________________________ edk2-devel mailing list edk2-devel@lists.01.org https://lists.01.org/mailman/listinfo/edk2-devel