public inbox for libc-alpha@sourceware.org
 help / color / mirror / Atom feed
From: Palmer Dabbelt <palmer@rivosinc.com>
To: adhemerval.zanella@linaro.org
Cc: libc-alpha@sourceware.org
Subject: Re: [PATCH 3/5] riscv: Add the clone3 wrapper
Date: Mon, 29 May 2023 07:04:50 -0700 (PDT)	[thread overview]
Message-ID: <mhng-4a9d49cc-5719-4d1e-9cc4-f6db86c9377d@palmer-ri-x1c9> (raw)
In-Reply-To: <20230203171237.1220878-4-adhemerval.zanella@linaro.org>

On Fri, 03 Feb 2023 09:12:35 PST (-0800), adhemerval.zanella@linaro.org wrote:
> It follows the internal signature:
>
>   extern int clone3 (struct clone_args *__cl_args, size_t __size,
>  int (*__func) (void *__arg), void *__arg);
>
> Checked on riscv64-linux-gnu-rv64imafdc-lp64d.

Due to a bug we only had clone3 on rv64 until 59a4e0d5511b ("RISC-V: 
Include clone3() on rv32"), which landed in 5.15 and was then backported 
all the way to 5.4.

> ---
>  sysdeps/unix/sysv/linux/riscv/clone3.S | 80 ++++++++++++++++++++++++++
>  sysdeps/unix/sysv/linux/riscv/sysdep.h |  1 +
>  2 files changed, 81 insertions(+)
>  create mode 100644 sysdeps/unix/sysv/linux/riscv/clone3.S
>
> diff --git a/sysdeps/unix/sysv/linux/riscv/clone3.S b/sysdeps/unix/sysv/linux/riscv/clone3.S
> new file mode 100644
> index 0000000000..ee5780ee2f
> --- /dev/null
> +++ b/sysdeps/unix/sysv/linux/riscv/clone3.S
> @@ -0,0 +1,80 @@
> +/* The clone3 syscall wrapper.  Linux/RISC-V version.
> +   Copyright (C) 2023 Free Software Foundation, Inc.
> +
> +   This file is part of the GNU C Library.
> +
> +   The GNU C Library is free software; you can redistribute it and/or
> +   modify it under the terms of the GNU Lesser General Public
> +   License as published by the Free Software Foundation; either
> +   version 2.1 of the License, or (at your option) any later version.
> +
> +   The GNU C Library is distributed in the hope that it will be useful,
> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
> +   Lesser General Public License for more details.
> +
> +   You should have received a copy of the GNU Lesser General Public
> +   License along with the GNU C Library; if not, see
> +   <https://www.gnu.org/licenses/>.  */
> +
> +#include <asm/errno.h>
> +#include <sys/asm.h>
> +#include <sysdep.h>
> +
> +/* The userland implementation is:
> +   int clone3 (struct clone_args *cl_args, size_t size,
> +               int (*func)(void *arg), void *arg);
> +
> +   the kernel entry is:
> +   int clone3 (struct clone_args *cl_args, size_t size);
> +
> +   The parameters are passed in registers from userland:
> +   a0: cl_args
> +   a1: size
> +   a2: func
> +   a3: arg  */
> +
> +        .text
> +ENTRY(__clone3)
> +	/* Sanity check args.  */
> +	beqz	a0, L(invalid)	/* No NULL cl_args pointer.  */
> +	beqz	a2, L(invalid)  /* No NULL function pointer.  */
> +
> +	/* Do the system call, the kernel expects:
> +	   a7: system call number
> +	   a0: cl_args
> +	   a1: size  */
> +	li	a7, __NR_clone3

IIUC this only builds with kernel headers 5.15 or newer on rv32.  I'm 
not sure if that's a problem: unless I'm missing something we support 
building for those systems now, but rv32 is pretty esoteric so I'm not 
sure anyone cares -- maybe someone uses 5.15 on rv32 (some rv64 hardware 
is still being released with 5.10-based vendor kernels), but 5.4 is 
pretty old for us.

> +	scall
> +
> +	bltz	a0, L(error)
> +	beqz	a0, L(thread_start)
> +
> +	ret
> +
> +L(invalid):
> +	li	a0, -EINVAL
> +L(error):
> +	tail	__syscall_error
> +END (__clone3)
> +
> +ENTRY(__thread_start_clone3)
> +L(thread_start):
> +	/* Terminate call stack by noting ra is undefined.  Use a dummy
> +	   .cfi_label to force starting the FDE.  */
> +	.cfi_label .Ldummy
> +	cfi_undefined (ra)
> +
> +	/* Restore the arg for user's function and call the user's
> +	   function.  */
> +	mv		a1, a2	/* Function pointer.  */

I think we don't need that mv, we can just call a2 directly?  Unless I'm 
missing some reason the thread main needs to see a pointer to itself?

> +	mv		a0, a3	/* Argument pointer.  */
> +	jalr		a1
> +
> +	/* Call exit with the function's return value.  */
> +	li		a7, __NR_exit
> +	scall
> +END(__thread_start_clone3)
> +
> +libc_hidden_def (__clone3)
> +weak_alias (__clone3, clone3)
> diff --git a/sysdeps/unix/sysv/linux/riscv/sysdep.h b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> index 4af5fe5dbc..e96a930409 100644
> --- a/sysdeps/unix/sysv/linux/riscv/sysdep.h
> +++ b/sysdeps/unix/sysv/linux/riscv/sysdep.h
> @@ -151,6 +151,7 @@
>
>  /* RV32 does not support the gettime VDSO syscalls.  */
>  # endif
> +# define HAVE_CLONE3_WRAPPER		1
>
>  /* List of system calls which are supported as vsyscalls (for RV32 and
>     RV64).  */

That's just some minor comments though, so

Reviewed-by: Palmer Dabbelt <palmer@rivosinc.com>

Thanks! (and sorry I missed it for a bit)

  parent reply	other threads:[~2023-05-29 14:04 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-02-03 17:12 [PATCH 0/5] Add clone3 support for multiple architectures Adhemerval Zanella
2023-02-03 17:12 ` [PATCH 1/5] powerpc64: Add the clone3 wrapper Adhemerval Zanella
2023-02-03 21:13   ` Paul E Murphy
2023-02-03 17:12 ` [PATCH 2/5] s390x: " Adhemerval Zanella
2023-05-25 17:36   ` Adhemerval Zanella Netto
2023-02-03 17:12 ` [PATCH 3/5] riscv: " Adhemerval Zanella
2023-05-25 17:36   ` Adhemerval Zanella Netto
2023-05-29 14:04   ` Palmer Dabbelt [this message]
2023-05-29 20:38     ` Adhemerval Zanella Netto
2023-02-03 17:12 ` [PATCH 4/5] arm: " Adhemerval Zanella
2023-05-25 17:36   ` Adhemerval Zanella Netto
2023-02-03 17:12 ` [PATCH 5/5] mips: " Adhemerval Zanella
2023-05-25 17:36   ` Adhemerval Zanella Netto

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-4a9d49cc-5719-4d1e-9cc4-f6db86c9377d@palmer-ri-x1c9 \
    --to=palmer@rivosinc.com \
    --cc=adhemerval.zanella@linaro.org \
    --cc=libc-alpha@sourceware.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: link
Be 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 read-only IMAP folder(s) and NNTP newsgroup(s).