linux-kernel - Re: Should this_cpu

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20181208105220.GF5289@hirez.programming.kicks-ass.net>
Date:   Sat, 8 Dec 2018 11:52:20 +0100
From:   Peter Zijlstra <peterz@...radead.org>
To:     Nadav Amit <nadav.amit@...il.com>
Cc:     Matthew Wilcox <willy@...radead.org>,
        Vlastimil Babka <vbabka@...e.cz>,
        Linux-MM <linux-mm@...ck.org>,
        LKML <linux-kernel@...r.kernel.org>, X86 ML <x86@...nel.org>,
        Ingo Molnar <mingo@...hat.com>,
        Thomas Gleixner <tglx@...utronix.de>
Subject: Re: Should this_cpu_read() be volatile?

On Fri, Dec 07, 2018 at 04:40:52PM -0800, Nadav Amit wrote:

> > I'm actually having difficulty finding the this_cpu_read() in any of the
> > functions you mention, so I cannot make any concrete suggestions other
> > than pointing at the alternative functions available.
> 
> 
> So I got deeper into the code to understand a couple of differences. In the
> case of select_idle_sibling(), the patch (Peter’s) increase the function
> code size by 123 bytes (over the baseline of 986). The per-cpu variable is
> called through the following call chain:
> 
> 	select_idle_sibling()
> 	=> select_idle_cpu()
> 	=> local_clock()
> 	=> raw_smp_processor_id()
> 
> And results in 2 more calls to sched_clock_cpu(), as the compiler assumes
> the processor id changes in between (which obviously wouldn’t happen).

That is the thing with raw_smp_processor_id(), it is allowed to be used
in preemptible context, and there it _obviously_ can change between
subsequent invocations.

So again, this change is actually good.

If we want to fix select_idle_cpu(), we should maybe not use
local_clock() there but use sched_clock_cpu() with a stable argument,
this code runs with IRQs disabled and therefore the CPU number is stable
for us here.

> There may be more changes around, which I didn’t fully analyze. But
> the very least reading the processor id should not get “volatile”.
> 
> As for finish_task_switch(), the impact is only few bytes, but still
> unnecessary. It appears that with your patch preempt_count() causes multiple
> reads of __preempt_count in this code:
> 
>        if (WARN_ONCE(preempt_count() != 2*PREEMPT_DISABLE_OFFSET,
>                      "corrupted preempt_count: %s/%d/0x%x\n",
>                      current->comm, current->pid, preempt_count()))
>                preempt_count_set(FORK_PREEMPT_COUNT);

My patch proposed here:

  https://marc.info/?l=linux-mm&m=154409548410209

would actually fix that one I think, preempt_count() uses
raw_cpu_read_4() which will loose the volatile with that patch.