Re: [PATCH] Add "riscv" as an alias for "riscv32"

[Palmer Dabbelt]

I don't really disagree with you here, as the original rationale for naming the 
tuples "riscv32-*" and "riscv64-*" was to avoid leaving anything ambiguous.  
Unfortunately the internet is full of people with strong opinions on these 
sorts of things, and some members of the RISC-V community have strong enough 
opinions that the tuples should be named "riscv-*" that they've gone and 
patched the toolchain sources to accept these names and then distributed 
binaries with the non-standard names.

I've actually gone to pretty long lengths to avoid sending this patch, but it's 
been a year since I first attempted to get people to stop wasting time with 
non-standard tuple names and nothing has changed.  I finally caved and 
submitted this patch so we could at least contain the name at config.sub to 
avoid it ending up in everyone's build scripts.

So I'm in the odd situation of submitting a patch I actually disagree with :).

The one saving grace here is that most of the RISC-V toolchains support 
generating code for all targets -- in other words "riscv32-*-gcc -march=rv64gc" 
is a valid thing to do, and will generate 64-bit code (and links against a 
64-bit libc).  As a result, none of this actually matters.

On Wed, 20 Jun 2018 14:40:01 PDT (-0700), address@hidden wrote:
> Why does a user who doesn't know what *bitwidth* they want pass an
> explicit config anyways? IMO it's a shame legacy ones need have both
> explicit and implicit variants on so many axes---one has to know the
> fully history to tell what's going on which might be fine for greybeards
> but is hostile to new programmers. The required bitwidth for new arches
> like wasm and riscv restores the symmetry, and so is pleasing to the
> eye, clearer to the beginner, and simpler for the implementation. I am
> strongly against this change.
>
> John
>
> On 06/20/18 13:24, Palmer Dabbelt wrote:
> > Most RISC-V toolchains can generate code for all RISC-V targets, but are
> > currently named "riscv32-*" and "riscv64-*".  Users have suggested that
> > these names are unintuiative and would like tuples to be "riscv-*".
> >
> > This patch enables "riscv-*" tuples by adding a config.sub rule that
> > replaces "riscv-*" with "riscv32-*".  The first ISA described by the
> > RISC-V specification is RV32I, a 32-bit ISA, so making "riscv-*" mean
> > 32-bit seems like the sane way to go.  There are also handful of tests
> > added, which all pass for me.
> >
> > This has been discussed on various mailing lists a half dozen times, but
> > I never got around to submitting a patch