On Fri, 28 Apr 2023 at 17:09, ron minnich <rminnich@xxxxxxxxx> wrote: > > There is lots of text in the preceding notes :-), which is nice because we're clearly looking at something that matters! > > But, note, ARM Chromebooks run Linux, and I checked with the firmware team just now: > "Right. We're not using UEFI or ACPI or SMBIOS or DMI or any of that on Arm. Just the Device Tree." > > So I do not agree that we need UEFI tables due to some presumed semantics that they implement, because: several tens of millions of ARM chromebooks running Linux show otherwise. > > We've got a chance here to move to self describing data, and I think we need to take it. It will be a long time before we get this chance again. > I'm not sure what you mean by self-describing: device tree is definitely not self-describing, and we maintain a huge collection of DT bindings (which are documented in separate YAML files) in the kernel tree that specify in detail how a device tree must be constructed in order to comply with the device tree based boot protocol. However, introducing such a binding for SMBIOS is perfectly reasonable, although I would suggest that we don't copy the SMBIOS/SMBIOS3 entry point address into the device tree (as this patch does), but properly describe the memory region that contains the actual SMBIOS structured data directly, along with its version. This might be reused by other DT based platforms as well. Doing the same for ACPI is where we'll get into trouble, given that we'd end up with two conflicting hardware descriptions and unfulfilled dependencies on EFI specific data structures, and it is not the kernel's job to reason about which h/w description should take precedence, or to make guesses about memory types. So I fully agree with Ron that moving to device tree is a much better choice here - that way, you can avoid ACPI and UEFI altogether > On Thu, Apr 27, 2023 at 8:18 PM 运辉崔 <cuiyunhui@xxxxxxxxxxxxx> wrote: >> >> Hi Ard, Mark, >> >> On Thu, Apr 27, 2023 at 8:52 PM Ard Biesheuvel <ardb@xxxxxxxxxx> wrote: >> >> > and the only combination we do not support is ACPI without EFI, as >> > ACPI on arm64 depends on the EFI memory map. >> > >> > What this patch seems to be proposing is a combination of all of >> > these, i.e., doing a pseudo-EFI direct kernel boot where the EFI >> > dependencies of ACPI are being fulfilled by ad-hoc descriptions passed >> > in via DT. >> > >> > I am concerned that this will result in a maintenance burden for Linux >> > with very little gain, so I feel we should not go down this road. >> >> Judging from the current kernel, getting acpi smbios, memmap tables is >> not just a way to have EFI, right? >> smbios:SMBIOS_ENTRY_POINT_SCAN_START >> acpi:CONFIG_ACPI_LEGACY_TABLES_LOOKUP >> memmap: e820 >> >> Our current situation is that coreboot does not support EFI, but supports fdt, >> but we need to support ACPI, and riscv does not have a reserved >> address segment >> like x86 that can be used, so our current solution is to pass acpi and >> other tables through fdt. >> >> Based on this, do you have a better suggestion ? >> >> Thanks， >> Yunhui
