On 30.01.2025 22:14, Stefano Stabellini wrote: > On Thu, 30 Jan 2025, Jan Beulich wrote: >> On 26.12.2024 17:57, Daniel P. Smith wrote: >>> Look for a subnode of type `multiboot,kernel` within a domain node. If >>> found, >>> process the reg property for the MB1 module index. If the bootargs property >>> is >>> present and there was not an MB1 string, then use the command line from the >>> device tree definition. >> >> While multiboot is apparently the first x86-specific part (as far as Xen >> goes) >> to be put under domain-builder/, I wonder: >> - Wouldn't looking for "multiboot,kernel" simply yield nothing on non-x86, >> so having the code under common/ would still be okay? > > One small clarification: multiboot,kernel is actually common between > both ARM and x86. It is "module-index" which is x86-specific and would > "simply yield nothing on non-x86", as you wrote. > > I'll let Dan address your point that "having the code under common/ > would still be okay". > > >> - What's "multiboot" describing here? The origin of the module? (What other >> origins would then be possible? How would MB1 and MB2 be distinguished? >> What about a native xen.efi boot?) A property of the kernel (when Linux >> doesn't use MB)? > > Each device tree node has a compatible string to qualify what kind of > information the node is describing. The compatible string for device > tree nodes describing a kernel binary or a ramdisk previously loaded > into memory by a bootloader have a "multiboot," prefix. See > docs/misc/arm/device-tree/booting.txt. This is unrelated to the binary > multiboot protocol Grub uses on x86 to boot Xen. > > A distinction between MB1 and MB2 is not needed in device tree, that > information is retrieved via the Grub multiboot protocol as usual. The > only thing needed here in device tree is the location of the kernel, > either by RAM address, or by Grub multiboot module index. This last > option (Grub multiboot module index) is the "module-index" property I > mentioned above.
Hmm, then I'm afraid I can't make sense of the mentioning of MB1 in the description. Yet that's a point more towards Daniel than you. Jan
