[PATCH v2 1/1] objdump, as: Add callx support for BPF CPU v1
Will Hawkins
hawkinsw@obs.cr
Thu Feb 15 21:44:08 GMT 2024
On Thu, Feb 15, 2024 at 10:32 AM Jose E. Marchesi
<jose.marchesi@oracle.com> wrote:
>
>
> > On Wed, Feb 14, 2024 at 11:14 AM Jose E. Marchesi
> > <jose.marchesi@oracle.com> wrote:
> >>
> >>
> >> > On Wed, Feb 14, 2024 at 5:58 AM Jose E. Marchesi
> >> > <jose.marchesi@oracle.com> wrote:
> >> >>
> >> >>
> >> >> > First, as I said yesterday in my direct message to you, thank you for
> >> >> > making binutils such a pleasant place to contribute to FOSS. You have
> >> >> > no idea what that means to contributors like me!
> >> >> >
> >> >> > Second, thank you for this helpful critique. I really appreciated
> >> >> > reading your feedback and will reply inline below (including with an
> >> >> > offer for a patch unrelated to callx that may clean up some of the
> >> >> > non-constant uses throughout the bpf-specific code).
> >> >> >
> >> >> > Third, there is good news: The heavy lifting of this patch is largely
> >> >> > "overcome by events" -- clang/llvm developers are changing their
> >> >> > encoding of the callx instruction to more closely match what gcc does.
> >> >> > In fact, v3 of this patch will look much more like v1 than v2.
> >> >>
> >> >> Hi Will.
> >> >>
> >> >> It seems to me that all we need binutils-wise is to enable the callx
> >> >> instruction with BPF >= v1. No other changes are necessary as far as I
> >> >> can see, other than adjusting the testsuite accordingly.
> >> >
> >> > It should be making its way across the Internet to you now!
> >>
> >> Thanks :)
> >>
> >> > The only thing ... it will probably cause a build error until a
> >> > corresponding patch in the simulator lands that takes into
> >> > consideration the new enum value.
> >> >
> >> > I am happy to handle that however you like!
> >>
> >> No problem. I will tackle the simulator patch in the GDB list and apply
> >> both at almost the same time so the buildbots do not get (too) upset.
> >
> > Thanks!! It's been fun working on this. And, as I offered to Nick last
> > night, I am more than happy to take on a little project to do a
> > cleanup of the code. Just let me know if you would be open to that!
>
> By all means, and than you.
> Improvements are always very welcome :)
>
> BTW, I would ask you to not include a separated cover letter for single
> patches. This makes it easier to reply and quote to both the
> description of the patch and its contents.
Thank you! That's great feedback! I will definitely do that from now on!
Will
>
> >
> > Will
> >
> >>
> >> > Sorry for the delay!
> >> > Will
> >> >
> >> >
> >> >>
> >> >> Thanks!
More information about the Binutils
mailing list