[CAULDRON] Topics for the Toolchain and Linux kernel BoF

[Jose E. Marchesi]

> Xi Ruoyao <xry111@xry111.site> writes:
>
>> On Mon, 2024-09-23 at 08:02 +0100, Sam James wrote:
>>> Xi Ruoyao <xry111@xry111.site> writes:
>>> 
>>> > On Thu, 2024-09-12 at 18:13 +0200, Jose E. Marchesi wrote:
>>> > > - "noreturn" and jump tables run-time hints
>>> > > 
>>> > >   It has been expressed on the kernel side the desire of having the C compiler
>>> > >   emit run-time hints marking functions that are not supposed to return and
>>> > >   also to provide annotations on jump tables.  This is for the benefit of
>>> > >   objtool in arm64, see references below.
>>> > > 
>>> > >   Goal of the discussion:
>>> > > 
>>> > >   Collect and assess the requirements of these features, discuss their
>>> > >   pertinence and the way it could be best implemented.  The outcome of the
>>> > >   discussion will then be used to continue the discussion with the clang/llvm
>>> > >   and kernel hackers at LPC.
>>> > > 
>>> > >   References:
>>> > >   https://lore.kernel.org/linux-arm-kernel/YyLmhUxTUaNzaieC@hirez.programming.kicks-ass.net/
>>> > 
>>> > FWIW we are having very similar issues for using objtool in loongarch. 
>>> > And we also hit a similar issue with table-based computed goto.  I'm not
>>> > sure if the computed goto is an issue for arm64 either.
>>> 
>>> I brought it up briefly on your behalf.
>>> 
>>> > 
>>> > Some related materials:
>>> > 
>>> > - https://gcc.gnu.org/PR82365#c13
>>> > - https://gcc.gnu.org/PR116483
>>> > - https://gcc.gnu.org/pipermail/gcc-patches/2024-July/657641.html
>>> > (Nacked off-line)
>>
>> Is there's something conclusive from the meeting?  I was busy on
>> $RANDOM_MESSY_THINGS last week and had not noticed it before.
>
> Sorry, this fell out of my inbox. I don't remember the details -- it was
> only a quick interjection and the discussion was moving on through the
> remaining topics. I _think_ we ended up talking a bit about SFrames and
> about how there's a kind of tangential need for this sort of thing there
> too but I might be making that up.
>
> Jose, do you have any notes about it? (Or maybe Indu?)

As far as I remember:

Maciej suggested to look at the MIPS GCC backend, because he remembers
doing something to mark jump tables.  Me and Josh looked after the
meeting, and sure enough, the MIPS backend does that when it emits
compact instructions.  It uses a couple of symbols, so there is a
precedent.

Other people suggested to use a GCC plugin to do this.  Jakub followed
up on this idea in a thread in gcc@gcc.gnu.og post-conference. See
https://gcc.gnu.org/pipermail/gcc/2024-September/244796.html.