linux-kernel - Re: [PATCH 1/3] mm/zsmalloc: avoid unregister a NOT-registered zsmalloc zpool driver

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:	Fri, 14 Nov 2014 07:38:16 +0800
From:	Ganesh Mahendran <opensource.ganesh@...il.com>
To:	Sergey Senozhatsky <sergey.senozhatsky@...il.com>
Cc:	Minchan Kim <minchan@...nel.org>, Nitin Gupta <ngupta@...are.org>,
	ddstreet@...e.org, linux-mm@...ck.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 1/3] mm/zsmalloc: avoid unregister a NOT-registered
 zsmalloc zpool driver

Hello

2014-11-13 23:22 GMT+08:00 Sergey Senozhatsky <sergey.senozhatsky@...il.com>:
> On (11/13/14 21:37), Mahendran Ganesh wrote:
>> Now zsmalloc can be registered as a zpool driver into zpool when
>> CONFIG_ZPOOL is enabled. During the init of zsmalloc, when error happens,
>> we need to do cleanup. But in current code, it will unregister a not yet
>> registered zsmalloc zpool driver(*zs_zpool_driver*).
>>
>> This patch puts the cleanup in zs_init() instead of calling zs_exit()
>> where it will unregister a not-registered zpool driver.
>>
>> Signed-off-by: Mahendran Ganesh <opensource.ganesh@...il.com>
>> ---
>>  mm/zsmalloc.c |   12 ++++++++----
>>  1 file changed, 8 insertions(+), 4 deletions(-)
>>
>> diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c
>> index 839a48c..3d2bb36 100644
>> --- a/mm/zsmalloc.c
>> +++ b/mm/zsmalloc.c
>> @@ -907,10 +907,8 @@ static int zs_init(void)
>>       __register_cpu_notifier(&zs_cpu_nb);
>>       for_each_online_cpu(cpu) {
>>               ret = zs_cpu_notifier(NULL, CPU_UP_PREPARE, (void *)(long)cpu);
>> -             if (notifier_to_errno(ret)) {
>> -                     cpu_notifier_register_done();
>> +             if (notifier_to_errno(ret))
>>                       goto fail;
>> -             }
>>       }
>>
>>       cpu_notifier_register_done();
>> @@ -920,8 +918,14 @@ static int zs_init(void)
>>  #endif
>>
>>       return 0;
>> +
>>  fail:
>> -     zs_exit();
>> +     for_each_online_cpu(cpu)
>> +             zs_cpu_notifier(NULL, CPU_UP_CANCELED, (void *)(long)cpu);
>> +     __unregister_cpu_notifier(&zs_cpu_nb);
>> +
>> +     cpu_notifier_register_done();
>> +
>>       return notifier_to_errno(ret);
>
> so we duplicate same code, and there is a bit confusing part
> now: zs_cpu_notifier(CPU_UP_CANCELED) and zs_cpu_notifier(CPU_DEAD)
> calls.

Thanks for your review.

CPU_UP_CANCELED and CPU_DEAD have different scenario.
CPU_UP_CANCELED is better for fail to complete initialization. While
CPU_DEAD is better for exit(such as removing module)

Thanks.

>
> how about something like this?
>
> Factor out zsmalloc cpu notifier unregistration code and call
> it from both zs_exit() and zs_init() error path.
>
> Signed-off-by: Sergey Senozhatsky <sergey.senozhatsky@...il.com>
>
> ---
>
>  mm/zsmalloc.c | 20 +++++++++++---------
>  1 file changed, 11 insertions(+), 9 deletions(-)
>
> diff --git a/mm/zsmalloc.c b/mm/zsmalloc.c
> index b3b57ef..c4d2d60 100644
> --- a/mm/zsmalloc.c
> +++ b/mm/zsmalloc.c
> @@ -881,14 +881,10 @@ static struct notifier_block zs_cpu_nb = {
>         .notifier_call = zs_cpu_notifier
>  };
>
> -static void zs_exit(void)
> +static inline void zs_unregister_cpu_notifier()
>  {
>         int cpu;
>
> -#ifdef CONFIG_ZPOOL
> -       zpool_unregister_driver(&zs_zpool_driver);
> -#endif
> -
>         cpu_notifier_register_begin();
>
>         for_each_online_cpu(cpu)
> @@ -898,6 +894,14 @@ static void zs_exit(void)
>         cpu_notifier_register_done();
>  }
>
> +static void zs_exit(void)
> +{
> +#ifdef CONFIG_ZPOOL
> +       zpool_unregister_driver(&zs_zpool_driver);
> +#endif
> +       zs_unregister_cpu_notifier();
> +}
> +
>  static int zs_init(void)
>  {
>         int cpu, ret;
> @@ -907,10 +911,8 @@ static int zs_init(void)
>         __register_cpu_notifier(&zs_cpu_nb);
>         for_each_online_cpu(cpu) {
>                 ret = zs_cpu_notifier(NULL, CPU_UP_PREPARE, (void *)(long)cpu);
> -               if (notifier_to_errno(ret)) {
> -                       cpu_notifier_register_done();
> +               if (notifier_to_errno(ret))
>                         goto fail;
> -               }
>         }
>
>         cpu_notifier_register_done();
> @@ -921,7 +923,7 @@ static int zs_init(void)
>
>         return 0;
>  fail:
> -       zs_exit();
> +       zs_unregister_cpu_notifier();
>         return notifier_to_errno(ret);
>  }
>
--
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/