linux-kernel - Re: [PATCH] x86, cpu: Fix node state for whether it contains 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: <20160829133612.GQ10153@twins.programming.kicks-ass.net>
Date:   Mon, 29 Aug 2016 15:36:12 +0200
From:   Peter Zijlstra <peterz@...radead.org>
To:     Tim Chen <tim.c.chen@...ux.intel.com>
Cc:     Andrew Morton <akpm@...ux-foundation.org>,
        Ingo Molnar <mingo@...e.hu>, "H. Peter Anvin" <hpa@...or.com>,
        "Huang, Ying" <ying.huang@...el.com>,
        Andi Kleen <andi@...stfloor.org>,
        Dave Hansen <dave.hansen@...el.com>,
        Dan Williams <dan.j.williams.intel.com@...ux.intel.com>,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH] x86, cpu: Fix node state for whether it contains CPU

On Wed, Aug 24, 2016 at 04:26:49PM -0700, Tim Chen wrote:
> In current kernel code, we only call node_set_state(cpu_to_node(cpu),
> N_CPU) when a cpu is hot plugged.  But we do not set the node state for
> N_CPU when the cpus are brought online during boot.
> 
> So this could lead to failure when we check to see
> if a node contains cpu with node_state(node_id, N_CPU).
> 
> One use case is in the node_reclaime function:
> 
>         /*
>          * Only run node reclaim on the local node or on nodes that do
>          * not
>          * have associated processors. This will favor the local
>          * processor
>          * over remote processors and spread off node memory allocations
>          * as wide as possible.
>          */
>         if (node_state(pgdat->node_id, N_CPU) && pgdat->node_id !=
> 		numa_node_id())
>                 return NODE_RECLAIM_NOSCAN;
> 
> I instrumented the kernel to call this function after boot and it
> always returns 0 on a x86 desktop machine until I apply
> the attached patch.
> 
> static int num_cpu_node(void)
> {
>        int i, nr_cpu_nodes = 0;
> 
>        for_each_node(i) {
>                if (node_state(i, N_CPU))
>                        ++ nr_cpu_nodes;
>        }
> 
>        return nr_cpu_nodes;
> }
> 
> I have not tested other architectues but they are likely
> to have similar issue.
> 
> Signed-off-by: Tim Chen <tim.c.chen@...ux.intel.com>
> ---
>  arch/x86/kernel/smpboot.c | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/arch/x86/kernel/smpboot.c b/arch/x86/kernel/smpboot.c
> index d8f7d01..04c0574 100644
> --- a/arch/x86/kernel/smpboot.c
> +++ b/arch/x86/kernel/smpboot.c
> @@ -259,6 +259,7 @@ static void notrace start_secondary(void *unused)
>  	lock_vector_lock();
>  	setup_vector_irq(smp_processor_id());
>  	set_cpu_online(smp_processor_id(), true);
> +	node_set_state(cpu_to_node(smp_processor_id()), N_CPU);
>  	unlock_vector_lock();
>  	cpu_set_state_online(smp_processor_id());
>  	x86_platform.nmi_init();

Would it not be easier to register the vmstat_notifier earlier, before
SMP bringup? Because with this change, we need to go fix all
architectures.