Re: [RFC][PATCH 6/8] x86/mm: add vsyscall address helper

From: Jann Horn <jannh@google.com>
To: Dave Hansen <dave.hansen@linux.intel.com>
Cc: kernel list <linux-kernel@vger.kernel.org>,
	sean.j.christopherson@intel.com,
	Peter Zijlstra <peterz@infradead.org>,
	Thomas Gleixner <tglx@linutronix.de>,
	"the arch/x86 maintainers" <x86@kernel.org>,
	Andy Lutomirski <luto@kernel.org>
Subject: Re: [RFC][PATCH 6/8] x86/mm: add vsyscall address helper
Date: Sat, 8 Sep 2018 02:46:35 +0200	[thread overview]
Message-ID: <CAG48ez0FAizJH-4LivpMPSnZqDUeUq0XHf6K8cA4xCqkiPMNYw@mail.gmail.com> (raw)
In-Reply-To: <20180907194901.C54A0101@viggo.jf.intel.com>

On Sat, Sep 8, 2018 at 2:25 AM Dave Hansen <dave.hansen@linux.intel.com> wrote:
> We will shortly be using this check in two locations.  Put it in
> a helper before we do so.
[...]
> +/*
> + * The (legacy) vsyscall page is the long page in the kernel portion
> + * of the address space that has user-accessible permissions.
> + */
> +static bool is_vsyscall_vaddr(unsigned long vaddr)
> +{
> +       return (vaddr & ~0xfff) == VSYSCALL_ADDR;
> +}

<bikeshed>
Since you're touching this code anyway: Would it make sense to change
that constant to a more explicit "~0xfffUL" (or alternatively
PAGE_MASK)? I tend to end up staring at code like this forever, trying
to figure out whether the upper 32 bits of the constant end up being
set or clear. As a reader, looking at the current code, it's quite
annoying to see what actually happens - first there's a signed 32-bit
literal 0xfff, then a 32-bit negation happens, and then the number is
converted to 64 bits with sign extension.
</bikeshed>