Re: [edk2-devel] [PATCH 2/2] MdePkg: Update code to be more C11 compliant by using __func__

["Ard Biesheuvel" <ardb@kernel.org>]

On Fri, 10 Feb 2023 at 02:21, Rebecca Cran <rebecca@bsdio.com> wrote:
>
> Mike,
>
> Yes, it passed CI - see https://github.com/tianocore/edk2/pull/4022 .
>
> I wasn't sure if we cared about EBC builds these days. Given the link on
> https://www.intel.com/content/www/us/en/architecture-and-technology/unified-extensible-firmware-interface/efi-homepage-general-technology.html
> to the compiler goes to a generic page that doesn't mention it, is it
> possible to get a copy if I want to verify if changes I'm making are
> compatible?
>

We tried to get access to this at Linaro, but I don't think it is
still available. Leif may remember better.

> I know there was a project to add support for building EBC using LLVM,
> but I'm not sure if anything came of it.
>

No that never went anywhere afaik. The problem was that, in EBC,
sizeof() is not a compile time constant, and this breaks too many
assumptions in the existing frontends, so this would only be feasible
if we restricted ourselves to a specialized version of EBC that
targets 64-bit architectures only, but this seems like a lot of effort
for questionable benefits.

Given that X64 code can run on arm64 under emulation without too many
problems [0], and that the 32-bit ARM procedure call standard is
fundamentally incompatible with EBC, the issue lost its urgency
somewhat, at least from the side of ARM at the time.

[0] https://github.com/ardbiesheuvel/X86EmulatorPkg

I wonder whether we should simply drop EBC support? It is unlikely
that all the changes that went into EDK2 over the years have not
caused a single EBC regression in the library components that are
needed to build UEFI drivers that target EBC.

Or is this still being used/tested internally at Intel?