perf tools build broken for RISCV 32 bit

[Emiliano Ingrassia]

Hi Arnd,

On Wed, Jan 20, 2021 at 05:03:30PM +0100, Arnd Bergmann wrote:
> On Wed, Jan 20, 2021 at 4:45 PM Emiliano Ingrassia
> <ingrassia at epigenesys.com> wrote:
> > > > > rv32 is in the unfortunate position of being the first one that has a glibc
> > > > > port, so it hits a lot of problems that other architectures do not.
> > > > > If the applications get fixed properly, then at least they will also
> > > > > work with musl-1.2 and a (will work-in-progress) glibc that will
> > > > > eventually allow building with 64-bit time_t.
> > > >
> > > > Ok, so in case of perf we probably should limit the needed patch to
> > > > riscv 32 bit architecture, unless we want to solve the problem for all
> > > > 32 bit architecture with time_t 64 bit support.
> > >
> > > No, whatever you do should be written to work on all 32-bit
> > > architectures. It would be crazy to add a special case for one
> > > architecture when it's easier to fix it properly.
> > >
> > > What was wrong with the helper function I suggested?
> >
> > Sorry, I was just trying to have a more clear picture of the problem.
> >
> > I'm totally fine with the helper function you proposed. About that,
> > should such approach be used for every time64 syscall for which is known
> > that there is no libc wrapper?
>
> We have previously discussed adding generated wrappers for every
> syscall, but my suggestion to have an inline function in one of the kernel
> headers was not very popular at the time.
>
> I think futex() is very much a special case here because this is the
> only[1] thing that gets called directly by a number of applications through
> syscall(). All the other time32 system calls have wrappers that are
> provided by any C library, which will just use the correct syscall number
> to match the arguments, or do a conversion if needed.
>
> Most of the packages that use the other __NR_* macros need them
> for something other than calling them (e.g. strace), so a wrapper does
> not help.
>
>       Arnd
>
> [1] technically, there is also io_{p,}getevents, but that is usually abstracted
> through libaio, and only that needs to be fixed. I checked
> https://pagure.io/libaio/commits/master, which seems to be the upstream
> but I do not see a fix in there, so this version is currently broken on 32-bit
> architectures with musl-1.2 as well as on rv32 with glibc. Distros may
> have already fixed this in their downstream copies.
>
>         Arnd

That's OK with me.

Please, when you'll propose the patch cc me so I can test and give you my review.
About this, will you also patch perf to use __kernel_futex()?

Thank you, best regards.

Emiliano