All of lore.kernel.org
 help / color / mirror / Atom feed
From: Haggai Eran <haggaie-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
To: Parav Pandit <pandit.parav-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
Cc: cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-doc-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org,
	Tejun Heo <tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org>,
	lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org,
	Johannes Weiner <hannes-druUgvl0LCNAfugRpC6u6w@public.gmane.org>,
	Doug Ledford <dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org>,
	Liran Liss <liranl-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>,
	"Hefty,
	Sean" <sean.hefty-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>,
	Jason Gunthorpe
	<jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org>,
	Jonathan Corbet <corbet-T1hC0tSOHrs@public.gmane.org>,
	james.l.morris-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org,
	serge-A9i7LUbDfNHQT0dZR+AlfA@public.gmane.org,
	Or Gerlitz <ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>,
	Matan Barak <matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>,
	raindel-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org,
	akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org,
	linux-security-module-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
Subject: Re: [PATCHv6 1/3] rdmacg: Added rdma cgroup controller
Date: Thu, 25 Feb 2016 14:03:13 +0200	[thread overview]
Message-ID: <56CEED81.7010507@mellanox.com> (raw)
In-Reply-To: <CAG53R5Uof+Ve7CndWy=BrgtxxCisQpzP_Ls0kw=Q270DhoEsZw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>

On 24/02/2016 18:16, Parav Pandit wrote:
>>> +     struct rdmacg_resource_pool *rpool;
>>> +     struct rdmacg_pool_info *pool_info = &device->pool_info;
>>> +
>>> +     spin_lock(&cg->rpool_list_lock);
>>> +     rpool = find_cg_rpool_locked(cg, device);
>> Is it possible for rpool to be NULL?
>>
> Unlikely, unless we have but in cgroup implementation.
> It may be worth to add WARN_ON and return from here to avoid kernel crash.
Sounds good.

>>> +static int charge_cg_resource(struct rdma_cgroup *cg,
>>> +                           struct rdmacg_device *device,
>>> +                           int index, int num)
>>> +{
>>> +     struct rdmacg_resource_pool *rpool;
>>> +     s64 new;
>>> +     int ret = 0;
>>> +
>>> +retry:
>>> +     spin_lock(&cg->rpool_list_lock);
>>> +     rpool = find_cg_rpool_locked(cg, device);
>>> +     if (!rpool) {
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +             ret = alloc_cg_rpool(cg, device);
>>> +             if (ret)
>>> +                     goto err;
>>> +             else
>>> +                     goto retry;
>> Instead of retrying after allocation of a new rpool, why not just return the
>> newly allocated rpool (or the existing one) from alloc_cg_rpool?
> 
> It can be done, but locking semantics just becomes difficult to
> review/maintain with that where alloc_cg_rpool will unlock and lock
> conditionally later on.
Maybe I'm missing something, but couldn't you simply lock rpool_list_lock
inside alloc_cg_rpool()? It already does that around its call to 
find_cg_rpool_locked() and the insertion to cg_list.

> This path will be hit anyway on first allocation typically. Once
> application is warm up, it will be unlikely to enter here.
> I should change if(!rpool) to if (unlikely(!rpool)).
Theoretically the new allocated rpool can be released again by the time you
get to the second call to find_cg_rpool_locked().

>>> +     spin_lock(&cg->rpool_list_lock);
>>> +     rpool = find_cg_rpool_locked(cg, device);
>>> +     if (!rpool) {
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +             ret = alloc_cg_rpool(cg, device);
>>> +             if (ret)
>>> +                     goto opt_err;
>>> +             else
>>> +                     goto retry;
>> You can avoid the retry here too. Perhaps this can go into a function.
>>
> In v5 I had wrapper around code which used to similar hiding using
> get_cg_rpool and put_cg_rpool helper functions.
> But Tejun was of opinion that I should have locks outside of all those
> functions. With that approach, this is done.
> So I think its ok. to have it this way.
I thought that was about functions that only locked the lock, called the 
find function, and released the lock. What I'm suggesting is to have one
function that does "lock + find + allocate if needed + unlock", and another
function that does (under caller's lock) "check ref count + check max count +
release rpool".

>>> +     }
>>> +
>>> +     /* now set the new limits of the rpool */
>>> +     while (enables) {
>>> +             /* if user set the limit, enables bit is set */
>>> +             if (enables & BIT(i)) {
>>> +                     enables &= ~BIT(i);
>>> +                     set_resource_limit(rpool, i, new_limits[i]);
>>> +             }
>>> +             i++;
>>> +     }
>>> +     if (rpool->refcnt == 0 &&
>>> +         rpool->num_max_cnt == pool_info->table_len) {
>>> +             /*
>>> +              * No user of the rpool and all entries are
>>> +              * set to max, so safe to delete this rpool.
>>> +              */
>>> +             list_del(&rpool->cg_list);
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +             free_cg_rpool(rpool);
>>> +     } else {
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +     }
>> You should consider putting this piece of code in a function (the
>> check of the reference counts and release of the rpool).
>>
> Yes. I did. Same as above comment. Also this function will have to
> unlock. Its usually better to lock/unlock from same function level,
> instead of locking at one level and unlocking from inside the
> function.
> Or
> I should have
> cg_rpool_cond_free_unlock() for above code (check of the reference
> counts and release of the rpool)?
It is confusing to lock and unlock in different contexts. Why not lock
in the caller context? free_cg_rpool() can be called under rpool_list_lock,
couldn't it? It locks device->rpool_lock, but uncharge_cg_resource() also
locks both in the same order.

Thanks,
Haggai
--
To unsubscribe from this list: send the line "unsubscribe linux-rdma" in
the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

WARNING: multiple messages have this Message-ID (diff)
From: Haggai Eran <haggaie@mellanox.com>
To: Parav Pandit <pandit.parav@gmail.com>
Cc: <cgroups@vger.kernel.org>, <linux-doc@vger.kernel.org>,
	<linux-kernel@vger.kernel.org>, <linux-rdma@vger.kernel.org>,
	Tejun Heo <tj@kernel.org>, <lizefan@huawei.com>,
	Johannes Weiner <hannes@cmpxchg.org>,
	Doug Ledford <dledford@redhat.com>,
	Liran Liss <liranl@mellanox.com>,
	"Hefty, Sean" <sean.hefty@intel.com>,
	Jason Gunthorpe <jgunthorpe@obsidianresearch.com>,
	Jonathan Corbet <corbet@lwn.net>, <james.l.morris@oracle.com>,
	<serge@hallyn.com>, Or Gerlitz <ogerlitz@mellanox.com>,
	Matan Barak <matanb@mellanox.com>, <raindel@mellanox.com>,
	<akpm@linux-foundation.org>,
	<linux-security-module@vger.kernel.org>
Subject: Re: [PATCHv6 1/3] rdmacg: Added rdma cgroup controller
Date: Thu, 25 Feb 2016 14:03:13 +0200	[thread overview]
Message-ID: <56CEED81.7010507@mellanox.com> (raw)
In-Reply-To: <CAG53R5Uof+Ve7CndWy=BrgtxxCisQpzP_Ls0kw=Q270DhoEsZw@mail.gmail.com>

On 24/02/2016 18:16, Parav Pandit wrote:
>>> +     struct rdmacg_resource_pool *rpool;
>>> +     struct rdmacg_pool_info *pool_info = &device->pool_info;
>>> +
>>> +     spin_lock(&cg->rpool_list_lock);
>>> +     rpool = find_cg_rpool_locked(cg, device);
>> Is it possible for rpool to be NULL?
>>
> Unlikely, unless we have but in cgroup implementation.
> It may be worth to add WARN_ON and return from here to avoid kernel crash.
Sounds good.

>>> +static int charge_cg_resource(struct rdma_cgroup *cg,
>>> +                           struct rdmacg_device *device,
>>> +                           int index, int num)
>>> +{
>>> +     struct rdmacg_resource_pool *rpool;
>>> +     s64 new;
>>> +     int ret = 0;
>>> +
>>> +retry:
>>> +     spin_lock(&cg->rpool_list_lock);
>>> +     rpool = find_cg_rpool_locked(cg, device);
>>> +     if (!rpool) {
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +             ret = alloc_cg_rpool(cg, device);
>>> +             if (ret)
>>> +                     goto err;
>>> +             else
>>> +                     goto retry;
>> Instead of retrying after allocation of a new rpool, why not just return the
>> newly allocated rpool (or the existing one) from alloc_cg_rpool?
> 
> It can be done, but locking semantics just becomes difficult to
> review/maintain with that where alloc_cg_rpool will unlock and lock
> conditionally later on.
Maybe I'm missing something, but couldn't you simply lock rpool_list_lock
inside alloc_cg_rpool()? It already does that around its call to 
find_cg_rpool_locked() and the insertion to cg_list.

> This path will be hit anyway on first allocation typically. Once
> application is warm up, it will be unlikely to enter here.
> I should change if(!rpool) to if (unlikely(!rpool)).
Theoretically the new allocated rpool can be released again by the time you
get to the second call to find_cg_rpool_locked().

>>> +     spin_lock(&cg->rpool_list_lock);
>>> +     rpool = find_cg_rpool_locked(cg, device);
>>> +     if (!rpool) {
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +             ret = alloc_cg_rpool(cg, device);
>>> +             if (ret)
>>> +                     goto opt_err;
>>> +             else
>>> +                     goto retry;
>> You can avoid the retry here too. Perhaps this can go into a function.
>>
> In v5 I had wrapper around code which used to similar hiding using
> get_cg_rpool and put_cg_rpool helper functions.
> But Tejun was of opinion that I should have locks outside of all those
> functions. With that approach, this is done.
> So I think its ok. to have it this way.
I thought that was about functions that only locked the lock, called the 
find function, and released the lock. What I'm suggesting is to have one
function that does "lock + find + allocate if needed + unlock", and another
function that does (under caller's lock) "check ref count + check max count +
release rpool".

>>> +     }
>>> +
>>> +     /* now set the new limits of the rpool */
>>> +     while (enables) {
>>> +             /* if user set the limit, enables bit is set */
>>> +             if (enables & BIT(i)) {
>>> +                     enables &= ~BIT(i);
>>> +                     set_resource_limit(rpool, i, new_limits[i]);
>>> +             }
>>> +             i++;
>>> +     }
>>> +     if (rpool->refcnt == 0 &&
>>> +         rpool->num_max_cnt == pool_info->table_len) {
>>> +             /*
>>> +              * No user of the rpool and all entries are
>>> +              * set to max, so safe to delete this rpool.
>>> +              */
>>> +             list_del(&rpool->cg_list);
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +             free_cg_rpool(rpool);
>>> +     } else {
>>> +             spin_unlock(&cg->rpool_list_lock);
>>> +     }
>> You should consider putting this piece of code in a function (the
>> check of the reference counts and release of the rpool).
>>
> Yes. I did. Same as above comment. Also this function will have to
> unlock. Its usually better to lock/unlock from same function level,
> instead of locking at one level and unlocking from inside the
> function.
> Or
> I should have
> cg_rpool_cond_free_unlock() for above code (check of the reference
> counts and release of the rpool)?
It is confusing to lock and unlock in different contexts. Why not lock
in the caller context? free_cg_rpool() can be called under rpool_list_lock,
couldn't it? It locks device->rpool_lock, but uncharge_cg_resource() also
locks both in the same order.

Thanks,
Haggai

  parent reply	other threads:[~2016-02-25 12:03 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2016-02-20 11:00 [PATCHv6 0/3] rdmacg: IB/core: rdma controller support Parav Pandit
2016-02-20 11:00 ` Parav Pandit
2016-02-20 11:00 ` [PATCHv6 2/3] IB/core: added support to use rdma cgroup controller Parav Pandit
     [not found]   ` <1455966006-13774-3-git-send-email-pandit.parav-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2016-02-24 13:43     ` Haggai Eran
2016-02-24 13:43       ` Haggai Eran
2016-02-24 16:05       ` Parav Pandit
     [not found] ` <1455966006-13774-1-git-send-email-pandit.parav-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2016-02-20 11:00   ` [PATCHv6 1/3] rdmacg: Added " Parav Pandit
2016-02-20 11:00     ` Parav Pandit
2016-02-21  7:43     ` Leon Romanovsky
2016-02-21 11:33       ` Parav Pandit
     [not found]         ` <CAG53R5XGJESb+_-FtjLsK+jHoEmbePLd==G2knBMOcrfdPm62Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-02-21 13:45           ` Leon Romanovsky
2016-02-21 13:45             ` Leon Romanovsky
     [not found]             ` <20160221134518.GM30450-2ukJVAZIZ/Y@public.gmane.org>
2016-02-21 14:11               ` Parav Pandit
2016-02-21 14:11                 ` Parav Pandit
2016-02-21 15:09                 ` Leon Romanovsky
2016-02-21 15:15                   ` Parav Pandit
2016-02-24 13:13     ` Haggai Eran
2016-02-24 13:13       ` Haggai Eran
     [not found]       ` <56CDAC7A.6030206-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-02-24 16:16         ` Parav Pandit
2016-02-24 16:16           ` Parav Pandit
     [not found]           ` <CAG53R5Uof+Ve7CndWy=BrgtxxCisQpzP_Ls0kw=Q270DhoEsZw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-02-25 12:03             ` Haggai Eran [this message]
2016-02-25 12:03               ` Haggai Eran
     [not found]               ` <56CEED81.7010507-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org>
2016-02-25 13:34                 ` Parav Pandit
2016-02-25 13:34                   ` Parav Pandit
2016-02-25 14:26                   ` Parav Pandit
     [not found]                     ` <CAG53R5WxXYZof4QzuJndakzvG1+t388pDRXL2O466eEDkYJ+bw-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-02-25 14:42                       ` Haggai Eran
2016-02-25 14:42                         ` Haggai Eran
     [not found]                   ` <CAG53R5UZb=9WR7zk2b5C_FuKmt+WdNkbcrVbW+g1-oAj6J=w_Q-JsoAwUIsXosN+BqQ9rBEUg@public.gmane.org>
2016-02-25 14:30                     ` Haggai Eran
2016-02-25 14:30                       ` Haggai Eran
2016-02-20 11:00   ` [PATCHv6 3/3] rdmacg: Added documentation for rdmacg Parav Pandit
2016-02-20 11:00     ` Parav Pandit
2016-02-20 11:00     ` Parav Pandit
     [not found]     ` <1455966006-13774-4-git-send-email-pandit.parav-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org>
2016-02-24 14:26       ` Haggai Eran
2016-02-24 14:26         ` Haggai Eran
2016-02-24 15:21         ` Parav Pandit
2016-02-28  8:55           ` Haggai Eran
2016-02-28  8:55             ` Haggai Eran
2016-02-28  9:02             ` Parav Pandit
2016-02-22  4:59 ` [PATCHv6 0/3] rdmacg: IB/core: rdma controller support Parav Pandit

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=56CEED81.7010507@mellanox.com \
    --to=haggaie-vpraknaxozvwk0htik3j/w@public.gmane.org \
    --cc=akpm-de/tnXTf+JLsfHDXvbKv3WD2FQJk+8+b@public.gmane.org \
    --cc=cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=corbet-T1hC0tSOHrs@public.gmane.org \
    --cc=dledford-H+wXaHxf7aLQT0dZR+AlfA@public.gmane.org \
    --cc=hannes-druUgvl0LCNAfugRpC6u6w@public.gmane.org \
    --cc=james.l.morris-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org \
    --cc=jgunthorpe-ePGOBjL8dl3ta4EC/59zMFaTQe2KTcn/@public.gmane.org \
    --cc=linux-doc-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=linux-kernel-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=linux-rdma-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=linux-security-module-u79uwXL29TY76Z2rM5mHXA@public.gmane.org \
    --cc=liranl-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=lizefan-hv44wF8Li93QT0dZR+AlfA@public.gmane.org \
    --cc=matanb-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=ogerlitz-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=pandit.parav-Re5JQEeQqe8AvxtiuMwx3w@public.gmane.org \
    --cc=raindel-VPRAkNaXOzVWk0Htik3J/w@public.gmane.org \
    --cc=sean.hefty-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org \
    --cc=serge-A9i7LUbDfNHQT0dZR+AlfA@public.gmane.org \
    --cc=tj-DgEjT+Ai2ygdnm+yROfE0A@public.gmane.org \
    /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.