[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20150515065417.GB29973@gmail.com>
Date: Fri, 15 May 2015 08:54:17 +0200
From: Ingo Molnar <mingo@...nel.org>
To: "H. Peter Anvin" <hpa@...or.com>
Cc: Borislav Petkov <bp@...en8.de>, Gu Zheng <guz.fnst@...fujitsu.com>,
Thomas Gleixner <tglx@...utronix.de>,
linux-kernel@...r.kernel.org, x86@...nel.org
Subject: Re: [RFC PATCH] x86, espfix: use spin_lock rather than mutex
* H. Peter Anvin <hpa@...or.com> wrote:
> On 05/14/2015 02:27 PM, Borislav Petkov wrote:
> > Remove stable@ from CC.
> >
> > On Thu, May 14, 2015 at 08:29:55PM +0200, Ingo Molnar wrote:
> >> We could allocate them on the boot CPU side and hand them over to
> >> the secondary CPU.
> >
> > Yeah, something along those lines. I mean, they're allocated and
> > in-use during the complete system lifetime, we might just as well
> > allocate them all in one go. Btw, what's our allocator that early,
> > memblock?
> >
> > Still, what I find strange is why are we seeing this only now? Is
> > it because it had to be a big box (cpu >= 128) or something else
> > changed...?
> >
>
> Quite probable. You don't really want to allocate them until you
> know if a CPU at least exists, though.
>
> I like Ingo's suggestion of allocating them before CPU bringup on
> the initiating CPU.
The only slightly subtle detail with that is to use alloc_pages_node()
with the secondary CPU's node, to make sure the espfix stack is
NUMA-local to the CPU that is going to use it.
Thanks,
Ingo
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists