linux-kernel - Re: [PATCH 2/4] block/part_stat: use __this_cpu_add() instead of access by smp_processor

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]

Date:   Mon, 4 May 2020 18:02:28 +0300
From:   Konstantin Khlebnikov <khlebnikov@...dex-team.ru>
To:     Christoph Hellwig <hch@...radead.org>
Cc:     linux-kernel@...r.kernel.org, linux-block@...r.kernel.org,
        Jens Axboe <axboe@...nel.dk>, Christoph Lameter <cl@...ux.com>,
        Dennis Zhou <dennis@...nel.org>, Tejun Heo <tj@...nel.org>
Subject: Re: [PATCH 2/4] block/part_stat: use __this_cpu_add() instead of
 access by smp_processor_id()

On 04/05/2020 17.03, Christoph Hellwig wrote:
>> +#define __part_stat_add(part, field, addnd)				\
>> +	(part_stat_get(part, field) += (addnd))
> 
> Just open coding part_stat_get for the UP side would seems a little
> easier to read.

If rewrite filed definition as

#ifdef	CONFIG_SMP
	struct disk_stats __percpu *dkstats;
#else
	struct disk_stats __percpu dkstats[1];
#endif

Then all per-cpu macro should work as is for UP case too.
Surprisingly arrow operator (struct->filed) works for arrays too =)


Inlining per-cpu structures should be good for tiny UP systems.
Especially if this could be done by few macro only in three places:
definition, allocating and freeing.


But sparse still complains.

./include/linux/part_stat.h:45:24: warning: incorrect type in initializer (different address spaces)
./include/linux/part_stat.h:45:24:    expected void const [noderef] <asn:3> *__vpp_verify
./include/linux/part_stat.h:45:24:    got struct disk_stats [noderef] *

Looks like it cannot assign different address-space to the filed.


> 
> Otherwise this looks good:
> 
> Reviewed-by: Christoph Hellwig <hch@....de>
>