FreeBSD-head-riscv64-build - Build #9623 (r336573) - Failure

[Konstantin Belousov]

On Sun, Jul 22, 2018 at 10:39:08PM +0100, Li-Wen Hsu wrote:
> On Sun, Jul 22, 2018 at 10:27 PM Konstantin Belousov
> <kostikbel at gmail.com> wrote:
> >
> > On Sun, Jul 22, 2018 at 06:16:02PM +0100, Li-Wen Hsu wrote:
> > > On Sun, Jul 22, 2018 at 18:45:05 +0300, Konstantin Belousov wrote:
> > > > On Sun, Jul 22, 2018 at 01:01:16AM +0000, Li-Wen Hsu wrote:
> > > > Well, the arch(7) manpage documents __riscv__.   Compilers typically
> > > > provide both __XXX__ and __XXX, while FreeBSD traditionally uses
> > > > the __XXX__ form.
> > >
> > > Please check r322168, __riscv__ is replaced by __riscv and
> > > __riscv64 is replaced by (__riscv && __riscv_xlen == 64).  Details are
> > > in the commit message.
> > >
> > > Alghough I grep'd sys/ and there are some __riscv__ still existing:
> > >
> > > sys/vm/vm_unix.c:72:#if !defined(__aarch64__) && !defined(__riscv__)
> > > sys/vm/vm_unix.c:81:#else /* defined(__aarch64__) || defined(__riscv__) */
> > > sys/vm/vm_unix.c:83:#endif /* defined(__aarch64__) || defined(__riscv__) */
> > >
> > > I guess those also need changing, as well as arch(7)
> > >
> > >
> > > > With that change, I think that your patch should go in regardless of
> > > > the second issue below.
> > >
> > > Thanks, please commit or approve it.
> > Why do you need an approval ?  I already said that your patch looks fine.
> 
> Oh, I did not realize that means a green light.  Also I am not sure
> about your opinion of __riscv and __riscv__.  Does my original patch
> look OK to you?

The change to __riscv looks strange.  At least arch(7) should be updated,
but this also contradicts the usual syntax.

Anyway, I do not have an opinion there, perhaps use __riscv since this
apparently is what Ruslan wants to use.