All of lore.kernel.org
 help / color / mirror / Atom feed
From: Dan Streetman <ddstreet@ieee.org>
To: Michal Hocko <mhocko@kernel.org>
Cc: Yu Zhao <yuzhao@google.com>, Seth Jennings <sjenning@redhat.com>,
	Linux-MM <linux-mm@kvack.org>,
	linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2] zswap: only use CPU notifier when HOTPLUG_CPU=y
Date: Fri, 2 Dec 2016 09:24:35 -0500	[thread overview]
Message-ID: <CALZtONBhvHNpGW4u1a8pVQeHx_8dX17vnFS52rrYbWA5dOtQ8w@mail.gmail.com> (raw)
In-Reply-To: <20161202134604.GA6837@dhcp22.suse.cz>

On Fri, Dec 2, 2016 at 8:46 AM, Michal Hocko <mhocko@kernel.org> wrote:
> On Wed 30-11-16 13:15:16, Yu Zhao wrote:
>> __unregister_cpu_notifier() only removes registered notifier from its
>> linked list when CPU hotplug is configured. If we free registered CPU
>> notifier when HOTPLUG_CPU=n, we corrupt the linked list.
>>
>> To fix the problem, we can either use a static CPU notifier that walks
>> through each pool or just simply disable CPU notifier when CPU hotplug
>> is not configured (which is perfectly safe because the code in question
>> is called after all possible CPUs are online and will remain online
>> until power off).
>>
>> v2: #ifdef for cpu_notifier_register_done during cleanup.
>
> this ifedfery is just ugly as hell. I am also wondering whether it is
> really needed. __register_cpu_notifier and __unregister_cpu_notifier are
> noops for CONFIG_HOTPLUG_CPU=n. So what's exactly that is broken here?

hmm, that's interesting, __unregister_cpu_notifier is always a noop if
HOTPLUG_CPU=n, but __register_cpu_notifier is only a noop if
HOTPLUG_CPU=n *and* MODULE.  If !MODULE, __register_cpu_notifier does
actually register!  This was added by commit
47e627bc8c9a70392d2049e6af5bd55fae61fe53 ('hotplug: Allow modules to
use the cpu hotplug notifiers even if !CONFIG_HOTPLUG_CPU') and looks
like it's to allow built-ins to register so they can notice during
boot when cpus are initialized.

IMHO, that is the real problem - sure, without HOTPLUG_CPU, nobody
should ever get a notification that a cpu is dying, but that doesn't
mean builtins that register notifiers will never unregister their
notifiers and then free them.

Changing zswap is only working around the symptom; instead, hotplug
should be changed to provide unregister_cpu_notifier in all cases
where register_cpu_notifier is provided.

>
>> Signe-off-by: Yu Zhao <yuzhao@google.com>
>> ---
>>  mm/zswap.c | 14 ++++++++++++++
>>  1 file changed, 14 insertions(+)
>>
>> diff --git a/mm/zswap.c b/mm/zswap.c
>> index 275b22c..2915f44 100644
>> --- a/mm/zswap.c
>> +++ b/mm/zswap.c
>> @@ -118,7 +118,9 @@ struct zswap_pool {
>>       struct kref kref;
>>       struct list_head list;
>>       struct work_struct work;
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       struct notifier_block notifier;
>> +#endif
>>       char tfm_name[CRYPTO_MAX_ALG_NAME];
>>  };
>>
>> @@ -448,6 +450,7 @@ static int __zswap_cpu_comp_notifier(struct zswap_pool *pool,
>>       return NOTIFY_OK;
>>  }
>>
>> +#ifdef CONFIG_HOTPLUG_CPU
>>  static int zswap_cpu_comp_notifier(struct notifier_block *nb,
>>                                  unsigned long action, void *pcpu)
>>  {
>> @@ -456,27 +459,34 @@ static int zswap_cpu_comp_notifier(struct notifier_block *nb,
>>
>>       return __zswap_cpu_comp_notifier(pool, action, cpu);
>>  }
>> +#endif
>>
>>  static int zswap_cpu_comp_init(struct zswap_pool *pool)
>>  {
>>       unsigned long cpu;
>>
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       memset(&pool->notifier, 0, sizeof(pool->notifier));
>>       pool->notifier.notifier_call = zswap_cpu_comp_notifier;
>>
>>       cpu_notifier_register_begin();
>> +#endif
>>       for_each_online_cpu(cpu)
>>               if (__zswap_cpu_comp_notifier(pool, CPU_UP_PREPARE, cpu) ==
>>                   NOTIFY_BAD)
>>                       goto cleanup;
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       __register_cpu_notifier(&pool->notifier);
>>       cpu_notifier_register_done();
>> +#endif
>>       return 0;
>>
>>  cleanup:
>>       for_each_online_cpu(cpu)
>>               __zswap_cpu_comp_notifier(pool, CPU_UP_CANCELED, cpu);
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       cpu_notifier_register_done();
>> +#endif
>>       return -ENOMEM;
>>  }
>>
>> @@ -484,11 +494,15 @@ static void zswap_cpu_comp_destroy(struct zswap_pool *pool)
>>  {
>>       unsigned long cpu;
>>
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       cpu_notifier_register_begin();
>> +#endif
>>       for_each_online_cpu(cpu)
>>               __zswap_cpu_comp_notifier(pool, CPU_UP_CANCELED, cpu);
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       __unregister_cpu_notifier(&pool->notifier);
>>       cpu_notifier_register_done();
>> +#endif
>>  }
>>
>>  /*********************************
>> --
>> 2.8.0.rc3.226.g39d4020
>>
>> --
>> To unsubscribe, send a message with 'unsubscribe linux-mm' in
>> the body to majordomo@kvack.org.  For more info on Linux MM,
>> see: http://www.linux-mm.org/ .
>> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
>
> --
> Michal Hocko
> SUSE Labs

WARNING: multiple messages have this Message-ID (diff)
From: Dan Streetman <ddstreet@ieee.org>
To: Michal Hocko <mhocko@kernel.org>
Cc: Yu Zhao <yuzhao@google.com>, Seth Jennings <sjenning@redhat.com>,
	Linux-MM <linux-mm@kvack.org>,
	linux-kernel <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH v2] zswap: only use CPU notifier when HOTPLUG_CPU=y
Date: Fri, 2 Dec 2016 09:24:35 -0500	[thread overview]
Message-ID: <CALZtONBhvHNpGW4u1a8pVQeHx_8dX17vnFS52rrYbWA5dOtQ8w@mail.gmail.com> (raw)
In-Reply-To: <20161202134604.GA6837@dhcp22.suse.cz>

On Fri, Dec 2, 2016 at 8:46 AM, Michal Hocko <mhocko@kernel.org> wrote:
> On Wed 30-11-16 13:15:16, Yu Zhao wrote:
>> __unregister_cpu_notifier() only removes registered notifier from its
>> linked list when CPU hotplug is configured. If we free registered CPU
>> notifier when HOTPLUG_CPU=n, we corrupt the linked list.
>>
>> To fix the problem, we can either use a static CPU notifier that walks
>> through each pool or just simply disable CPU notifier when CPU hotplug
>> is not configured (which is perfectly safe because the code in question
>> is called after all possible CPUs are online and will remain online
>> until power off).
>>
>> v2: #ifdef for cpu_notifier_register_done during cleanup.
>
> this ifedfery is just ugly as hell. I am also wondering whether it is
> really needed. __register_cpu_notifier and __unregister_cpu_notifier are
> noops for CONFIG_HOTPLUG_CPU=n. So what's exactly that is broken here?

hmm, that's interesting, __unregister_cpu_notifier is always a noop if
HOTPLUG_CPU=n, but __register_cpu_notifier is only a noop if
HOTPLUG_CPU=n *and* MODULE.  If !MODULE, __register_cpu_notifier does
actually register!  This was added by commit
47e627bc8c9a70392d2049e6af5bd55fae61fe53 ('hotplug: Allow modules to
use the cpu hotplug notifiers even if !CONFIG_HOTPLUG_CPU') and looks
like it's to allow built-ins to register so they can notice during
boot when cpus are initialized.

IMHO, that is the real problem - sure, without HOTPLUG_CPU, nobody
should ever get a notification that a cpu is dying, but that doesn't
mean builtins that register notifiers will never unregister their
notifiers and then free them.

Changing zswap is only working around the symptom; instead, hotplug
should be changed to provide unregister_cpu_notifier in all cases
where register_cpu_notifier is provided.

>
>> Signe-off-by: Yu Zhao <yuzhao@google.com>
>> ---
>>  mm/zswap.c | 14 ++++++++++++++
>>  1 file changed, 14 insertions(+)
>>
>> diff --git a/mm/zswap.c b/mm/zswap.c
>> index 275b22c..2915f44 100644
>> --- a/mm/zswap.c
>> +++ b/mm/zswap.c
>> @@ -118,7 +118,9 @@ struct zswap_pool {
>>       struct kref kref;
>>       struct list_head list;
>>       struct work_struct work;
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       struct notifier_block notifier;
>> +#endif
>>       char tfm_name[CRYPTO_MAX_ALG_NAME];
>>  };
>>
>> @@ -448,6 +450,7 @@ static int __zswap_cpu_comp_notifier(struct zswap_pool *pool,
>>       return NOTIFY_OK;
>>  }
>>
>> +#ifdef CONFIG_HOTPLUG_CPU
>>  static int zswap_cpu_comp_notifier(struct notifier_block *nb,
>>                                  unsigned long action, void *pcpu)
>>  {
>> @@ -456,27 +459,34 @@ static int zswap_cpu_comp_notifier(struct notifier_block *nb,
>>
>>       return __zswap_cpu_comp_notifier(pool, action, cpu);
>>  }
>> +#endif
>>
>>  static int zswap_cpu_comp_init(struct zswap_pool *pool)
>>  {
>>       unsigned long cpu;
>>
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       memset(&pool->notifier, 0, sizeof(pool->notifier));
>>       pool->notifier.notifier_call = zswap_cpu_comp_notifier;
>>
>>       cpu_notifier_register_begin();
>> +#endif
>>       for_each_online_cpu(cpu)
>>               if (__zswap_cpu_comp_notifier(pool, CPU_UP_PREPARE, cpu) ==
>>                   NOTIFY_BAD)
>>                       goto cleanup;
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       __register_cpu_notifier(&pool->notifier);
>>       cpu_notifier_register_done();
>> +#endif
>>       return 0;
>>
>>  cleanup:
>>       for_each_online_cpu(cpu)
>>               __zswap_cpu_comp_notifier(pool, CPU_UP_CANCELED, cpu);
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       cpu_notifier_register_done();
>> +#endif
>>       return -ENOMEM;
>>  }
>>
>> @@ -484,11 +494,15 @@ static void zswap_cpu_comp_destroy(struct zswap_pool *pool)
>>  {
>>       unsigned long cpu;
>>
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       cpu_notifier_register_begin();
>> +#endif
>>       for_each_online_cpu(cpu)
>>               __zswap_cpu_comp_notifier(pool, CPU_UP_CANCELED, cpu);
>> +#ifdef CONFIG_HOTPLUG_CPU
>>       __unregister_cpu_notifier(&pool->notifier);
>>       cpu_notifier_register_done();
>> +#endif
>>  }
>>
>>  /*********************************
>> --
>> 2.8.0.rc3.226.g39d4020
>>
>> --
>> To unsubscribe, send a message with 'unsubscribe linux-mm' in
>> the body to majordomo@kvack.org.  For more info on Linux MM,
>> see: http://www.linux-mm.org/ .
>> Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
>
> --
> Michal Hocko
> SUSE Labs

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

  reply	other threads:[~2016-12-02 14:44 UTC|newest]

Thread overview: 24+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-11-30 21:15 [PATCH v2] zswap: only use CPU notifier when HOTPLUG_CPU=y Yu Zhao
2016-11-30 21:15 ` Yu Zhao
2016-12-02 13:46 ` Michal Hocko
2016-12-02 13:46   ` Michal Hocko
2016-12-02 14:24   ` Dan Streetman [this message]
2016-12-02 14:24     ` Dan Streetman
2016-12-02 14:38     ` Michal Hocko
2016-12-02 14:38       ` Michal Hocko
2016-12-02 14:44       ` Michal Hocko
2016-12-02 14:44         ` Michal Hocko
2016-12-02 14:56         ` Dan Streetman
2016-12-02 14:56           ` Dan Streetman
2016-12-02 15:19           ` [PATCH] hotplug: make register and unregister notifier API symmetric Michal Hocko
2016-12-02 15:19             ` Michal Hocko
2016-12-03  5:15             ` kbuild test robot
2016-12-05  6:08               ` Michal Hocko
2016-12-05  6:08                 ` Michal Hocko
2016-12-03  7:18             ` kbuild test robot
2016-12-05 20:59             ` Yu Zhao
2016-12-05 20:59               ` Yu Zhao
2016-12-06  9:30               ` Michal Hocko
2016-12-06  9:30                 ` Michal Hocko
2016-12-05 21:11   ` [PATCH v2] zswap: only use CPU notifier when HOTPLUG_CPU=y Yu Zhao
2016-12-05 21:11     ` Yu Zhao

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=CALZtONBhvHNpGW4u1a8pVQeHx_8dX17vnFS52rrYbWA5dOtQ8w@mail.gmail.com \
    --to=ddstreet@ieee.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-mm@kvack.org \
    --cc=mhocko@kernel.org \
    --cc=sjenning@redhat.com \
    --cc=yuzhao@google.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.