Date: Sat, 16 Nov 2024 07:26:22 -0700 From: Alan Somers <asomers@freebsd.org> To: Tom Jones <thj@freebsd.org> Cc: Alan Somers <asomers@freebsd.org>, freebsd-riscv@freebsd.org Subject: Re: Help wanted improving Rust support for RISC-V Message-ID: <CAOtMX2g838YdPnfL2Jo31jHNTnoPC_RpxC6M2de5pvquO=cnCA@mail.gmail.com> In-Reply-To: <fca8a77e-99b2-4615-a7d3-1dae4d2d3565@app.fastmail.com> References: <CAOtMX2iUfMktVkYrnc5ufQ460PpnqG2YP2DYmavAoDQz%2BOWk9g@mail.gmail.com> <fca8a77e-99b2-4615-a7d3-1dae4d2d3565@app.fastmail.com>
next in thread | previous in thread | raw e-mail | index | archive | help
--0000000000006ee1c606270878d8 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable On Sat, Nov 16, 2024, 2:21=E2=80=AFAM Tom Jones <thj@freebsd.org> wrote: > > > On Fri, Nov 15, 2024, at 23:47, Alan Somers wrote: > > I want to improve the state of Rust on FreeBSD in general, and RISC-V > > in particular. My plan is: > > > > * Assemble a corpus of crates heavy on OS stuff (like Nix and Tokio) > > and ensure that their tests pass on Riscv FreeBSD. > > * Update libc's ABI to FreeBSD 12, finally eliminating the need for > > COMPAT_11 on riscv to run Rust programs. > > * Ensure that the corpus of crates still passes their tests. > > * Convince the libc maintainers to use the FreeBSD 12 ABI, for riscv > only. > > * After a full Rust release cycle (at least 6 weeks), and maybe two, > > if there are no reports of breakage, and if Poudriere looks good, use > > that as justification to raise libc's ABI for all architectures. > > > > But in order to do any of this, I need access to riscv hardware. A > > virtual instance at Scaleway[^1][^2] would be ideal, but I don't know > > if FreeBSD can run there. Has anybody tried it? Failing that, I > > would need to either acquire some hardware myself, or borrow access to > > somebody else's. Or, convince some other developer to do the work > > himself. But I won't be able to help very much if I don't have direct > > access. > > > > Can anybody on this list help me get some access? > > I can=E2=80=99t help with scaleway, it sounds like an interesting choice. > > I have no idea of the current state of easily and affordable available > hardware. The visonfive 2 might be a good choice, there has been a lot of > work in the year since I last tried it. > > https://adventurist.me/posts/00315 > > It=E2=80=99s probably still the easiest to get hold of if a little underw= helming. > > You don=E2=80=99t mention qemu, is there some reason you can=E2=80=99t st= art there? > > - Tom > I haven't had good luck with using qemu-user-static to emulate armv7. > --0000000000006ee1c606270878d8 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable <div dir=3D"auto"><div><br><br><div class=3D"gmail_quote"><div dir=3D"ltr" = class=3D"gmail_attr">On Sat, Nov 16, 2024, 2:21=E2=80=AFAM Tom Jones <<a= href=3D"mailto:thj@freebsd.org">thj@freebsd.org</a>> wrote:<br></div><b= lockquote class=3D"gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-le= ft:1px solid rgb(204,204,204);padding-left:1ex"><br> <br> On Fri, Nov 15, 2024, at 23:47, Alan Somers wrote:<br> > I want to improve the state of Rust on FreeBSD in general, and RISC-V<= br> > in particular.=C2=A0 My plan is:<br> ><br> > * Assemble a corpus of crates heavy on OS stuff (like Nix and Tokio)<b= r> > and ensure that their tests pass on Riscv FreeBSD.<br> > * Update libc's ABI to FreeBSD 12, finally eliminating the need fo= r<br> > COMPAT_11 on riscv to run Rust programs.<br> > * Ensure that the corpus of crates still passes their tests.<br> > * Convince the libc maintainers to use the FreeBSD 12 ABI, for riscv o= nly.<br> > * After a full Rust release cycle (at least 6 weeks), and maybe two,<b= r> > if there are no reports of breakage, and if Poudriere looks good, use<= br> > that as justification to raise libc's ABI for all architectures.<b= r> ><br> > But in order to do any of this, I need access to riscv hardware.=C2=A0= A<br> > virtual instance at Scaleway[^1][^2] would be ideal, but I don't k= now<br> > if FreeBSD can run there.=C2=A0 Has anybody tried it?=C2=A0 Failing th= at, I<br> > would need to either acquire some hardware myself, or borrow access to= <br> > somebody else's.=C2=A0 Or, convince some other developer to do the= work<br> > himself.=C2=A0 But I won't be able to help very much if I don'= t have direct<br> > access.<br> ><br> > Can anybody on this list help me get some access?<br> <br> I can=E2=80=99t help with scaleway, it sounds like an interesting choice. <= br> <br> I have no idea of the current state of easily and affordable available hard= ware. The visonfive 2 might be a good choice, there has been a lot of work = in the year since I last tried it. <br> <br> <a href=3D"https://adventurist.me/posts/00315" rel=3D"noreferrer noreferrer= " target=3D"_blank">https://adventurist.me/posts/00315</a><br>; <br> It=E2=80=99s probably still the easiest to get hold of if a little underwhe= lming. <br> <br> You don=E2=80=99t mention qemu, is there some reason you can=E2=80=99t star= t there?<br> <br> - Tom<br></blockquote></div></div><div dir=3D"auto"><br></div><div dir=3D"a= uto">I haven't had good luck with using qemu-user-static to emulate arm= v7.</div><div dir=3D"auto"><div class=3D"gmail_quote"><blockquote class=3D"= gmail_quote" style=3D"margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(20= 4,204,204);padding-left:1ex"> </blockquote></div></div></div> --0000000000006ee1c606270878d8--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CAOtMX2g838YdPnfL2Jo31jHNTnoPC_RpxC6M2de5pvquO=cnCA>