From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-oa1-x2d.google.com (mail-oa1-x2d.google.com [IPv6:2001:4860:4864:20::2d]) by sourceware.org (Postfix) with ESMTPS id 1C4F93858412 for ; Mon, 29 May 2023 20:38:55 +0000 (GMT) DMARC-Filter: OpenDMARC Filter v1.4.2 sourceware.org 1C4F93858412 Authentication-Results: sourceware.org; dmarc=pass (p=none dis=none) header.from=linaro.org Authentication-Results: sourceware.org; spf=pass smtp.mailfrom=linaro.org Received: by mail-oa1-x2d.google.com with SMTP id 586e51a60fabf-19e7008a20aso1523226fac.1 for ; Mon, 29 May 2023 13:38:55 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linaro.org; s=google; t=1685392734; x=1687984734; h=content-transfer-encoding:in-reply-to:organization:from:references :cc:to:content-language:subject:user-agent:mime-version:date :message-id:from:to:cc:subject:date:message-id:reply-to; bh=oN8BscsmRQj3pGnHfFhx67TS6CgdoB+8mwd/aLs+Bgw=; b=HknmiGslsa6XIAq2HxN7UpCc1r0z4Ja+QlLgViYfjlwuJezdhTVRQrwJThXKgCQUjf k1MxCcUD5r/NFBXIYIZ9dhCPSrpumhrnOgSsQ9Z+ywW8SRYc0cMmzaMX07Wtw12RiMRm cmd62zActKNM2BsxBl9Y5ZQB+OptFk2l//kIhasTks+3Wgzf4kgNCzvgJm38w26Aj/T6 ZraifLqAS7bAT4oxSSkJxDXHifXGrMxFGb+Kz5tz/dlMp/KQj64yf79XkhEH4DixSUQF IPsuVhA9MyiyNl1X2Xp5xCrr/0PMki6qqTWy6JVek6unR7jXlt7zdulGgQNtSVXE2vcA 2Lhg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20221208; t=1685392734; x=1687984734; h=content-transfer-encoding:in-reply-to:organization:from:references :cc:to:content-language:subject:user-agent:mime-version:date :message-id:x-gm-message-state:from:to:cc:subject:date:message-id :reply-to; bh=oN8BscsmRQj3pGnHfFhx67TS6CgdoB+8mwd/aLs+Bgw=; b=Eh/a56+Y8W4bCwh3Vt8by4++G19JxsUyjTOIfng5VNlMFeLfnPpYWYTplqrdmDspqB ZNgyuRDTjBN5wxaA+C++QfcsBlnuwpxy6W4Waxa7qIu3JpgiXPGnMvpQEn+8pJDXDzTs 6SSBxZarxWrJJBk9wUwOMVTFTkGCv5Ez2DkXCJeKWFp0BfhwyxSzZI+/WGcBa97VLpcv b77BZoGDwiRkViCTyoAbp60/Yd/MM009Yu9OvuwLVoB8jBjB+t/m8PmqZBgwsI0kmvjs 1RgOs8y3peTcDzY2cLxgpNCreR17RPyUmZr3c4l61chMpEZnqONM4ENBpLzTnq5Vq9lr vwpQ== X-Gm-Message-State: AC+VfDyGV8wEbI6k7am7a2tCUuJgR0cGjyKPoIkMf8BWo0oJgODVg7lO qkxD9txhMYfnYlEt3fSiYQKXgrmv/L01pU32yDBjzA== X-Google-Smtp-Source: ACHHUZ6k5JgtjEnXo7DD3w+ayK/AVwp2QAWjpcs0a6GyTQxdw65oPjAdVis8t0y41IdvQdLEeRyaqA== X-Received: by 2002:a05:6870:5403:b0:187:e563:77b9 with SMTP id d3-20020a056870540300b00187e56377b9mr6181663oan.45.1685392734435; Mon, 29 May 2023 13:38:54 -0700 (PDT) Received: from ?IPV6:2804:1b3:a7c1:4dd5:b058:c94a:90a7:2c43? ([2804:1b3:a7c1:4dd5:b058:c94a:90a7:2c43]) by smtp.gmail.com with ESMTPSA id m2-20020a056870a40200b0019f6751fc26sm1377366oal.7.2023.05.29.13.38.52 (version=TLS1_3 cipher=TLS_AES_128_GCM_SHA256 bits=128/128); Mon, 29 May 2023 13:38:53 -0700 (PDT) Message-ID: <9864c8ad-c2ef-56fc-796d-13f20de0e94f@linaro.org> Date: Mon, 29 May 2023 17:38:50 -0300 MIME-Version: 1.0 User-Agent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10.15; rv:102.0) Gecko/20100101 Thunderbird/102.11.1 Subject: Re: [PATCH 3/5] riscv: Add the clone3 wrapper Content-Language: en-US To: Palmer Dabbelt Cc: libc-alpha@sourceware.org References: From: Adhemerval Zanella Netto Organization: Linaro In-Reply-To: Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-Spam-Status: No, score=-10.8 required=5.0 tests=BAYES_00,BODY_8BITS,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,DKIM_VALID_EF,GIT_PATCH_0,KAM_SHORT,NICE_REPLY_A,RCVD_IN_DNSWL_NONE,SPF_HELO_NONE,SPF_PASS,TXREP,T_SCC_BODY_TEXT_LINE autolearn=ham autolearn_force=no version=3.4.6 X-Spam-Checker-Version: SpamAssassin 3.4.6 (2021-04-09) on server2.sourceware.org List-Id: On 29/05/23 11:04, Palmer Dabbelt wrote: > 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 >> +   .  */ >> + >> +#include >> +#include >> +#include >> + >> +/* 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. The __NR_XXX macros are now provided by arch-syscall.h, which in turn is generated by the kernel header. And we already have clone3 for rv32 since 3387c40a8bb. > >> +    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? We can, I will change it. > >> +    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 Thanks. > > Thanks! (and sorry I missed it for a bit)