From: Arnd Bergmann <arnd@arndb.de>
To: Yury Norov <ynorov@caviumnetworks.com>
Cc: catalin.marinas@arm.com, linux-arm-kernel@lists.infradead.org,
        linux-kernel@vger.kernel.org, pinskia@gmail.com, ddaney.cavm@gmail.com,
        jan.dakinevich@gmail.com, Prasun.Kapoor@caviumnetworks.com,
        christoph.muellner@theobroma-systems.com,
        philipp.tomsich@theobroma-systems.com, broonie@kernel.org,
        andrey.konovalov@linaro.org, Nathan_Lynch@mentor.com, agraf@suse.de,
        bamvor.zhangjian@huawei.com, klimov.linux@gmail.com,
        Andrew Pinski <apinski@cavium.com>,
        Andrew Pinski <Andrew.Pinski@caviumnetworks.com>
Subject: Re: [PATCH v6 13/17] arm64:ilp32: add sys_ilp32.c and a separate table (in entry.S) to use it
Date: Thu, 05 Nov 2015 14:50:10 +0100
Message-ID: <48925827.N0Xfqtb3JD@wuerfel>
User-Agent: KMail/4.11.5 (Linux/3.16.0-10-generic; KDE/4.11.5; x86_64; ; )
In-Reply-To: <1446507046-24604-14-git-send-email-ynorov@caviumnetworks.com>
References: <1446507046-24604-1-git-send-email-ynorov@caviumnetworks.com> <1446507046-24604-14-git-send-email-ynorov@caviumnetworks.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 7Bit
Content-Type: text/plain; charset="us-ascii"
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 1886
Lines: 49

On Tuesday 03 November 2015 02:30:42 Yury Norov wrote:
> From: Andrew Pinski <apinski@cavium.com>
> 
> Add a separate syscall-table for ILP32, which dispatches either to native
> LP64 system call implementation or to compat-syscalls, as appropriate.

The uapi/asm-generic/unistd.h already contains a list of compat syscalls
that should work by default, I think it would be better to use that
list and override only the ones that differ between normal compat
mode and the new mode, e.g. when you require a wrapper or want to
use the native syscall entry.

> +/* We need to make sure the pointer gets copied correctly. */
> +asmlinkage long ilp32_sys_mq_notify(mqd_t mqdes, const struct sigevent __user *u_notification)
> +{
> +	struct sigevent __user *p = NULL;
> +	if (u_notification) {
> +		struct sigevent n;
> +		p = compat_alloc_user_space(sizeof(*p));
> +		if (copy_from_user(&n, u_notification, sizeof(*p)))
> +			return -EFAULT;
> +		if (n.sigev_notify == SIGEV_THREAD)
> +			n.sigev_value.sival_ptr = compat_ptr((uintptr_t)n.sigev_value.sival_ptr);
> +		if (copy_to_user(p, &n, sizeof(*p)))
> +			return -EFAULT;
> +	}
> +	return sys_mq_notify(mqdes, p);
> +}

Could this be avoided by defining sigval_t in a way that is
compatible?

> +/* sigevent contains sigval_t which is now 64bit always
> +   but need special handling due to padding for SIGEV_THREAD.  */
> +#define sys_mq_notify		ilp32_sys_mq_notify
> +
> +/* sigaltstack needs some special handling as the
> +   padding for stack_t might not be non-zero. */
> +long ilp32_sys_sigaltstack(const stack_t __user *uss_ptr,
> +			   stack_t __user *uoss_ptr)

asmlinkage?

	Arnd
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/