[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <20191127094313.GA11668@infradead.org>
Date: Wed, 27 Nov 2019 01:43:13 -0800
From: Christoph Hellwig <hch@...radead.org>
To: Wang ShaoBo <bobo.shaobowang@...wei.com>
Cc: cj.chengjian@...wei.com, mark.rutland@....com,
huawei.libin@...wei.com, guohanjun@...wei.com, xiexiuqi@...wei.com,
wcohen@...hat.com, linux-kernel@...r.kernel.org,
mtk.manpages@...il.com, wezhang@...hat.com
Subject: Re: [PATCH] sys_personality: Streamline code in sys_personality()
On Tue, Nov 26, 2019 at 05:40:45PM +0800, Wang ShaoBo wrote:
> SYSCALL_DEFINE1 in kernel/exec_domain.c looks like verbose,
> ksys_personality() can make it more concise.
What do you try to say with this sentence? I can't really parse it.
> --- a/kernel/exec_domain.c
> +++ b/kernel/exec_domain.c
> @@ -37,10 +37,5 @@ module_init(proc_execdomains_init);
>
> SYSCALL_DEFINE1(personality, unsigned int, personality)
> {
> - unsigned int old = current->personality;
> -
> - if (personality != 0xffffffff)
> - set_personality(personality);
> -
> - return old;
> + return ksys_personality(personality);
This looks ok, but I'd much rather just kill ksys_personality and add
an optional arch hook to reject personality settings for the arm64
special case.
Powered by blists - more mailing lists