From: palmer@sifive.com (Palmer Dabbelt) To: linux-riscv@lists.infradead.org Subject: [PATCH] riscv: add asm/unistd.h UAPI header Date: Tue, 06 Nov 2018 16:08:12 -0800 (PST) [thread overview] Message-ID: <mhng-3c26b2fb-0c11-42f2-9026-1113b200a7b2@palmer-si-x1c4> (raw) In-Reply-To: <CAK8P3a3JcztCSYBuu5ZJXaV83VNLv_1v3v-EKMR+tC1oOsQesw@mail.gmail.com> On Mon, 05 Nov 2018 12:56:15 PST (-0800), Arnd Bergmann wrote: > On 11/5/18, David Abdurachmanov <david.abdurachmanov@gmail.com> wrote: >> Marcin Juszkiewicz reported issues while generating syscall table for riscv >> using 4.20-rc1. The patch refactors our unistd.h files to match some other >> architectures. >> >> - Add asm/unistd.h UAPI header, which has __ARCH_WANT_NEW_STAT >> - Remove asm/syscalls.h UAPI header and merge to asm/unistd.h >> - Adjust kernel asm/unistd.h >> >> So now asm/unistd.h UAPI header should show all syscalls for riscv. >> >> Before this, Makefile simply put `#include <asm-generic/unistd.h>` into >> generated asm/unistd.h UAPI header thus user didn't see: >> >> - __NR_riscv_flush_icache >> - __NR_newfstatat >> - __NR_fstat >> >> which are supported by riscv kernel. >> >> Signed-off-by: David Abdurachmanov <david.abdurachmanov@gmail.com> >> Cc: Arnd Bergmann <arnd@arndb.de> >> Cc: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org> >> Cc: Guenter Roeck <linux@roeck-us.net> > > Thanks for addressing this, your patch correctly fixes riscv64, and > I should have noticed the mistake when I originally merged the > broken patch. > > However, looking closer I found another problem with the original > patch that your fix does not address: > > __ARCH_WANT_NEW_STAT should only be set on 64-bit > architectures. > > For a 32-bit architecture, we only want __ARCH_WANT_STAT64 if > any. For 64-bit architectures with compat mode, we still need to > set __ARCH_WANT_STAT64 from the non-uapi file so we get > the syscall implementation. > > If we don't care about the riscv32 ABI changing yet, we can > decide to leave out __ARCH_WANT_STAT64 here, and require > glibc to implement it using statx() like any new architecture. > stat64 is not y2038 safe, and statx replaces it because of that. Thanks for pointing this out. A while ago we decided the rv32 ABI was "slushy": it can change if it has a good reason to. Right now the only planned changes are the y2038 changes, which I consider this a part of. For some reason I thought we'd already done this, but since we haven't then I think it should go in sooner rather than later -- that will help the glibc guys get everything lined up. The target is still the next glibc release (Feb 1st) for a stable RV32I ABI. That's progressing well, with one last blocking issue related to some of our floating-point emulation routines before we can submit the port. This should give us ample time to line up the ABIs correctly so everything works. So I think the correct answer here is to drop __ARCH_WANT_STAT64 from RISC-V. >> Fixes: 67314ec7b025 > > That line should be formatted as > > Fixes: 67314ec7b025 ("RISC-V: Request newstat syscalls") Yep, and I have [pretty] fixes = Fixes: %h (\"%s\") to make that slightly easier for me to remember :).
WARNING: multiple messages have this Message-ID (diff)
From: Palmer Dabbelt <palmer@sifive.com> To: Arnd Bergmann <arnd@arndb.de> Cc: aou@eecs.berkeley.edu, david.abdurachmanov@gmail.com, linux-kernel@vger.kernel.org, marcin.juszkiewicz@linaro.org, linux-riscv@lists.infradead.org, linux@roeck-us.net Subject: Re: [PATCH] riscv: add asm/unistd.h UAPI header Date: Tue, 06 Nov 2018 16:08:12 -0800 (PST) [thread overview] Message-ID: <mhng-3c26b2fb-0c11-42f2-9026-1113b200a7b2@palmer-si-x1c4> (raw) Message-ID: <20181107000812.Wboy3ZGvobpT2itlROOhX_jlyUOuQ2mmtMAbMM7k0ew@z> (raw) In-Reply-To: <CAK8P3a3JcztCSYBuu5ZJXaV83VNLv_1v3v-EKMR+tC1oOsQesw@mail.gmail.com> On Mon, 05 Nov 2018 12:56:15 PST (-0800), Arnd Bergmann wrote: > On 11/5/18, David Abdurachmanov <david.abdurachmanov@gmail.com> wrote: >> Marcin Juszkiewicz reported issues while generating syscall table for riscv >> using 4.20-rc1. The patch refactors our unistd.h files to match some other >> architectures. >> >> - Add asm/unistd.h UAPI header, which has __ARCH_WANT_NEW_STAT >> - Remove asm/syscalls.h UAPI header and merge to asm/unistd.h >> - Adjust kernel asm/unistd.h >> >> So now asm/unistd.h UAPI header should show all syscalls for riscv. >> >> Before this, Makefile simply put `#include <asm-generic/unistd.h>` into >> generated asm/unistd.h UAPI header thus user didn't see: >> >> - __NR_riscv_flush_icache >> - __NR_newfstatat >> - __NR_fstat >> >> which are supported by riscv kernel. >> >> Signed-off-by: David Abdurachmanov <david.abdurachmanov@gmail.com> >> Cc: Arnd Bergmann <arnd@arndb.de> >> Cc: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org> >> Cc: Guenter Roeck <linux@roeck-us.net> > > Thanks for addressing this, your patch correctly fixes riscv64, and > I should have noticed the mistake when I originally merged the > broken patch. > > However, looking closer I found another problem with the original > patch that your fix does not address: > > __ARCH_WANT_NEW_STAT should only be set on 64-bit > architectures. > > For a 32-bit architecture, we only want __ARCH_WANT_STAT64 if > any. For 64-bit architectures with compat mode, we still need to > set __ARCH_WANT_STAT64 from the non-uapi file so we get > the syscall implementation. > > If we don't care about the riscv32 ABI changing yet, we can > decide to leave out __ARCH_WANT_STAT64 here, and require > glibc to implement it using statx() like any new architecture. > stat64 is not y2038 safe, and statx replaces it because of that. Thanks for pointing this out. A while ago we decided the rv32 ABI was "slushy": it can change if it has a good reason to. Right now the only planned changes are the y2038 changes, which I consider this a part of. For some reason I thought we'd already done this, but since we haven't then I think it should go in sooner rather than later -- that will help the glibc guys get everything lined up. The target is still the next glibc release (Feb 1st) for a stable RV32I ABI. That's progressing well, with one last blocking issue related to some of our floating-point emulation routines before we can submit the port. This should give us ample time to line up the ABIs correctly so everything works. So I think the correct answer here is to drop __ARCH_WANT_STAT64 from RISC-V. >> Fixes: 67314ec7b025 > > That line should be formatted as > > Fixes: 67314ec7b025 ("RISC-V: Request newstat syscalls") Yep, and I have [pretty] fixes = Fixes: %h (\"%s\") to make that slightly easier for me to remember :). _______________________________________________ linux-riscv mailing list linux-riscv@lists.infradead.org http://lists.infradead.org/mailman/listinfo/linux-riscv
next prev parent reply other threads:[~2018-11-07 0:08 UTC|newest] Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top 2018-11-05 14:26 [PATCH] riscv: add asm/unistd.h UAPI header David Abdurachmanov 2018-11-05 14:26 ` David Abdurachmanov 2018-11-05 14:43 ` David Abdurachmanov 2018-11-05 14:43 ` David Abdurachmanov 2018-11-05 16:02 ` Marcin Juszkiewicz 2018-11-05 16:02 ` Marcin Juszkiewicz 2018-11-05 20:56 ` Arnd Bergmann 2018-11-05 20:56 ` Arnd Bergmann 2018-11-07 0:08 ` Palmer Dabbelt [this message] 2018-11-07 0:08 ` Palmer Dabbelt 2018-11-07 18:30 ` David Abdurachmanov 2018-11-07 18:30 ` David Abdurachmanov 2018-11-07 21:09 ` Arnd Bergmann 2018-11-07 21:09 ` Arnd Bergmann 2018-11-08 2:10 ` Palmer Dabbelt 2018-11-08 2:10 ` Palmer Dabbelt 2018-11-08 10:30 ` Arnd Bergmann 2018-11-08 10:30 ` Arnd Bergmann 2018-11-08 16:57 ` Palmer Dabbelt 2018-11-08 16:57 ` Palmer Dabbelt 2018-11-08 10:38 ` David Abdurachmanov 2018-11-08 10:38 ` David Abdurachmanov 2018-11-08 16:57 ` Palmer Dabbelt 2018-11-08 16:57 ` Palmer Dabbelt
Reply instructions: You may reply publicly to this message via plain-text email using any one of the following methods: * Save the following mbox file, import it into your mail client, and reply-to-all from there: mbox Avoid top-posting and favor interleaved quoting: https://en.wikipedia.org/wiki/Posting_style#Interleaved_style * Reply using the --to, --cc, and --in-reply-to switches of git-send-email(1): git send-email \ --in-reply-to=mhng-3c26b2fb-0c11-42f2-9026-1113b200a7b2@palmer-si-x1c4 \ --to=palmer@sifive.com \ --cc=linux-riscv@lists.infradead.org \ /path/to/YOUR_REPLY https://kernel.org/pub/software/scm/git/docs/git-send-email.html * If your mail client supports setting the In-Reply-To header via mailto: links, try the mailto: linkBe sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).