Re: [PATCH v2] RISC-V: Show accurate per-hart isa in /proc/cpuinfo

[Evan Green]

On Thu, Jul 6, 2023 at 12:45 PM Palmer Dabbelt <palmer@xxxxxxxxxxx> wrote:
>
> On Thu, 06 Jul 2023 12:34:30 PDT (-0700), Evan Green wrote:
> > On Thu, Jul 6, 2023 at 3:53 AM Conor Dooley <conor.dooley@xxxxxxxxxxxxx> wrote:
> >>
> >> On Thu, Jul 06, 2023 at 10:01:31AM +0200, Andrew Jones wrote:
> >> > On Wed, Jul 05, 2023 at 10:29:31AM -0700, Evan Green wrote:
> >> > > In /proc/cpuinfo, most of the information we show for each processor is
> >> > > specific to that hart: marchid, mvendorid, mimpid, processor, hart,
> >> > > compatible, and the mmu size. But the ISA string gets filtered through a
> >> > > lowest common denominator mask, so that if one CPU is missing an ISA
> >> > > extension, no CPUs will show it.
> >> > >
> >> > > Now that we track the ISA extensions for each hart, let's report ISA
> >> > > extension info accurately per-hart in /proc/cpuinfo. We cannot change
> >> > > the "isa:" line, as usermode may be relying on that line to show only
> >> > > the common set of extensions supported across all harts. Add a new "hart
> >> > > isa" line instead, which reports the true set of extensions for that
> >> > > hart. This matches what is returned in riscv_hwprobe() when querying a
> >> > > given hart.
> >> > >
> >> > > Signed-off-by: Evan Green <evan@xxxxxxxxxxxx>
> >> > >
> >> > > ---
> >> > >
> >> > > Changes in v2:
> >> > >  - Added new "hart isa" line rather than altering behavior of existing
> >> > >    "isa" line (Conor, Palmer)
> >> > >
> >> > >
> >> > > I based this series on top of Conor's riscv-extensions-strings branch
> >> > > from July 3rd, since otherwise this change gets hopelessly entangled
> >> > > with that series.
> >> > >
> >> > > I was unsure if I could snuggle the new "hart isa" line in just below
> >> > > "isa". Aesthetically it would be quite pleasing, but it runs the risk of
> >> > > breaking brittle usermode parsers that are assuming a specific line
> >> > > order. So I put it at the end.
> >> >
> >> > Actually, they're probably only aesthetically pleasing when they match. If
> >> > there are differences, then I'd guess having them side by side, almost the
> >> > same, but different, would make them even harder to look at then they
> >> > already are. So I think I'll be happier with them separated by a few lines
> >> > anyway.
> >>
> >> This list is eventually going to be so big that I don't think doing
> >> by-eye anything is going to be useful, so aesthetics be damned.
> >> That said, a parser that relies on the order of individual lines like
> >> that might deserve to be broken ;)
> >>
> >> Anyway, change looks good to me:
> >> Reviewed-by: Conor Dooley <conor.dooley@xxxxxxxxxxxxx>
> >>
> >> I was thinking the uabi doc might need an update - should we add to it
> >> that "isa" means the common set & "hart isa"?
> >
> > I did do a little search for documentation to update, but failed to
> > come up with any. Which doc are you thinking of?
>
> Should be somewhere like Documentation/arch/riscv/uabi.rst

Thanks, I'll spin this patch and add something here.
-Evan