Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Feedback-ID: i56a14606:Fastmail
User-Agent: Cyrus-JMAP/3.7.0-alpha0-1185-g841157300a-fm-20221208.002-g84115730
Mime-Version: 1.0
Message-Id: <7c747897-bb5b-46b0-a342-aefe30079c06@app.fastmail.com>
In-Reply-To: <20230103164359.24347-13-ysionneau@kalray.eu>
References: <20230103164359.24347-1-ysionneau@kalray.eu>
 <20230103164359.24347-13-ysionneau@kalray.eu>
Date:   Wed, 04 Jan 2023 16:07:32 +0100
From:   "Arnd Bergmann" <arnd@arndb.de>
To:     "Yann Sionneau" <ysionneau@kalray.eu>
Cc:     "Christian Brauner" <brauner@kernel.org>,
        "Paul Walmsley" <paul.walmsley@sifive.com>,
        "Palmer Dabbelt" <palmer@dabbelt.com>,
        "Albert Ou" <aou@eecs.berkeley.edu>, linux-kernel@vger.kernel.org,
        linux-riscv@lists.infradead.org,
        "Clement Leger" <clement.leger@bootlin.com>,
        "Guillaume Thouvenin" <gthouvenin@kalray.eu>,
        "Julian Vetter" <jvetter@kalray.eu>,
        "Julien Villette" <jvillette@kalray.eu>,
        "Marius Gligor" <mgligor@kalray.eu>
Subject: Re: [RFC PATCH 12/25] kvx: Add system call support
Content-Type: text/plain
Precedence: bulk

On Tue, Jan 3, 2023, at 17:43, Yann Sionneau wrote:

> +
> +/*
> + * Ensure that the range [addr, addr+size) is within the process's
> + * address space
> + */
> +static inline int __access_ok(unsigned long addr, unsigned long size)
> +{
> +	return size <= TASK_SIZE && addr <= TASK_SIZE - size;
> +}

This is the same as the generic version, so just use that instead.

> +#define HAVE_GET_KERNEL_NOFAULT
> +
> +#define __get_kernel_nofault(dst, src, type, err_label)			\
> +do {									\
> +	long __kr_err;							\
> +									\
> +	__get_user_nocheck(*((type *)(dst)), (type *)(src), __kr_err);	\
> +	if (unlikely(__kr_err))						\
> +		goto err_label;						\
> +} while (0)
> +
> +#define __put_kernel_nofault(dst, src, type, err_label)			\
> +do {									\
> +	long __kr_err;							\
> +									\
> +	__put_user_nocheck(*((type *)(src)), (type *)(dst), __kr_err);	\
> +	if (unlikely(__kr_err))						\
> +		goto err_label;						\
> +} while (0)

The wrapper around __get_user_nocheck/__put_user_nocheck
is not ideal here. Since I think you only support new
compilers anyway, you can use the asm-goto-with-output feature
to define the asm to branch to the label directly, and use
the same thing to build __get_user()/__put_user().

> +++ b/arch/kvx/include/uapi/asm/unistd.h
> @@ -0,0 +1,16 @@

> +
> +#define __ARCH_WANT_RENAMEAT
> +#define __ARCH_WANT_NEW_STAT
> +#define __ARCH_WANT_SET_GET_RLIMIT
> +#define __ARCH_WANT_SYS_CLONE3

It's good to have clone3() but the other three sets
of syscalls should no longer be defined for new architectures,
so please remove those.

      Arnd