Oliver Upton <[email protected]> writes:
On Fri, Jun 27, 2025 at 08:45:38PM +0000, Colton Lewis wrote:
Hi Ben. Thanks for the review.
Ben Horgan <[email protected]> writes:
> Hi Colton,
> On 6/26/25 21:04, Colton Lewis wrote:
> > There's no reason Enums shouldn't be equivalent to UnsignedEnums and
> > explicitly specify they are unsigned. This will avoid the annoyance I
> > had with HPMN0.
> An Enum can be annotated with the field's sign by updating it to
> UnsignedEnum or SignedEnum. This is explained in [1].
> With this change ID_AA64PFR1_EL1.MTE_frac would be marked as unsigned
> when it should really be considered signed.
> Enum 43:40 MTE_frac
> 0b0000 ASYNC
> 0b1111 NI
> EndEnum
Thanks for the explanation. I made this a separate commit because I
considered people might object and HPMN0 is already an UnsignedEnum in
my previous commit.
Do you think it would be a good idea to make plain Enums signed by
default or should I just remove this commit from the series?
It is presumptive to associate a sign with an enumeration. Generally
speaking, the only fields that can do signed / unsigned comparisons are
the Feature ID register fields.
So please drop this and only keep the change for HPMN0.
Done.
Thanks,
Oliver
> > Signed-off-by: Colton Lewis <[email protected]>
> > ---
> > arch/arm64/tools/gen-sysreg.awk | 1 +
> > 1 file changed, 1 insertion(+)
> > diff --git a/arch/arm64/tools/gen-sysreg.awk
> > b/arch/arm64/tools/gen-sysreg.awk
> > index f2a1732cb1f6..fa21a632d9b7 100755
> > --- a/arch/arm64/tools/gen-sysreg.awk
> > +++ b/arch/arm64/tools/gen-sysreg.awk
> > @@ -308,6 +308,7 @@ $1 == "Enum" && (block_current() == "Sysreg" ||
> > block_current() == "SysregFields
> > parse_bitdef(reg, field, $2)
> > define_field(reg, field, msb, lsb)
> > + define_field_sign(reg, field, "false")
> > next
> > }
> Thanks,
> Ben
> [1]
>
https://lore.kernel.org/all/[email protected]/
