Hi Sunil, On Mon, Jul 3, 2023 at 12:22 PM Sunil V L <sunilvl@xxxxxxxxxxxxxxxx> wrote: > > Hi Yunhui Cui, > On Sun, Jul 02, 2023 at 02:47:41PM +0100, Conor Dooley wrote: > > Hey, > > %subject: riscv: obtain ACPI RSDP from FFI. > > > > This subject is a bit unhelpful because FFI would commonly mean "foreign > > function interface" & you have not yet introduced it. It seems like it > > would be better to do s/FFI/devicetree/ or similar. > > Please also drop the full stop from the commit messages ;) > > > > Please use a cover letter for multi-patch series & include changelogs. > > > > +CC Sunil, Alex: > > > > Can you guys please take a look at this & see if it is something that we > > want to do (ACPI without EFI)? > > > > We have supported ACPI only with UEFI. The current booting contract > between firmware and OS is to pass only one of DT or ACPI, not both. > This approach brings another booting contract for Linux mixing ACPI and > DT which affects RVI specs. As per policy and since it can affect > multiple OSs, a frozen RVI spec is required for taking this patch into > linux. So, could you please bring this topic for discussion in [1] and > get agreement? > > Isn't it simpler to provide a minimum UEFI configuration table and > stubbed BS/RS? > > Have you done a PoC? I am curious how do you handle EFI memory map > dependencies. Yes, Poc has been completed. a memory node in DTS can solve it. > > In case this is approved, I am wondering why do we need new FFI? > > [1] - https://lists.riscv.org/g/tech-brs We have discussed with Ard and Ron many times about the series of questions you mentioned above, and reached a consensus. Please see the v1: https://patches.linaro.org/project/linux-acpi/patch/20230426034001.16-1-cuiyunhui@xxxxxxxxxxxxx/ Thanks, Yunhui
