Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Content-Type: text/plain;
        charset=utf-8
Mime-Version: 1.0 (Mac OS X Mail 12.4 \(3445.104.11\))
Subject: Re: [PATCH v8 4/5] locking/qspinlock: Introduce starvation avoidance
 into CNA
From:   Alex Kogan <alex.kogan@oracle.com>
In-Reply-To: <edc53126-bfe4-5102-d2eb-2332bf3a68e5@redhat.com>
Date:   Fri, 24 Jan 2020 16:27:32 -0500
Cc:     Peter Zijlstra <peterz@infradead.org>, linux@armlinux.org.uk,
        Ingo Molnar <mingo@redhat.com>,
        Will Deacon <will.deacon@arm.com>,
        Arnd Bergmann <arnd@arndb.de>, linux-arch@vger.kernel.org,
        linux-arm-kernel <linux-arm-kernel@lists.infradead.org>,
        linux-kernel@vger.kernel.org, Thomas Gleixner <tglx@linutronix.de>,
        Borislav Petkov <bp@alien8.de>, hpa@zytor.com, x86@kernel.org,
        Hanjun Guo <guohanjun@huawei.com>,
        Jan Glauber <jglauber@marvell.com>,
        Steven Sistare <steven.sistare@oracle.com>,
        Daniel Jordan <daniel.m.jordan@oracle.com>,
        dave.dice@oracle.com
Content-Transfer-Encoding: quoted-printable
Message-Id: <D39064BF-6EF3-4C13-B2D1-34C282A20F5E@oracle.com>
References: <20191230194042.67789-1-alex.kogan@oracle.com>
 <20191230194042.67789-5-alex.kogan@oracle.com>
 <20200121132949.GL14914@hirez.programming.kicks-ass.net>
 <cfdf635d-be2e-9d4b-c4ca-6bcbddc6868f@redhat.com>
 <3862F8A1-FF9B-40AD-A88E-2C0BA7AF6F58@oracle.com>
 <20200124075235.GX14914@hirez.programming.kicks-ass.net>
 <2c6741c5-d89d-4b2c-cebe-a7c7f6eed884@redhat.com>
 <48ce49e5-98a7-23cd-09f4-8290a65abbb5@redhat.com>
 <8D3AFB47-B595-418C-9568-08780DDC58FF@oracle.com>
 <714892cd-d96f-4d41-ae8b-d7b7642a6e3c@redhat.com>
 <1669BFDE-A1A5-4ED8-B586-035460BBF68A@oracle.com>
 <45660873-731a-a810-8c57-1a5a19d266b4@redhat.com>
 <693E6287-E37C-4C5D-BE33-B3D813BE505D@oracle.com>
 <edc53126-bfe4-5102-d2eb-2332bf3a68e5@redhat.com>
To:     Waiman Long <longman@redhat.com>
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk


> On Jan 24, 2020, at 4:12 PM, Waiman Long <longman@redhat.com> wrote:
>=20
> On 1/24/20 3:09 PM, Alex Kogan wrote:
>>>> We also probably do not want those =E2=80=9Cprioritized=E2=80=9D =
threads to disrupt
>>>> normal
>>>> CNA operation. E.g., if the main queue looks like T1_1, P2_1, T1_2,
>>>> =E2=80=A6, where
>>>> T1_x is a thread running on node 1 and P2_1 is a prioritized thread
>>>> running
>>>> on node 2, we want to pass the lock from T1_1 to P2_1 and then to =
T1_2
>>>> (rather than have P2_1 to scan for another thread on node 2).
>>>>=20
>>>> There is a way to achieve that =E2=80=94 when we pass the lock to =
P2_1,
>>>> we can set its numa node field to 1. This means that we need to
>>>> reset the numa
>>>> node field in cna_init_node(), but AFAICT this is relatively cheap.
>>>> The rest
>>>> of the CNA logic should not change.
>>>=20
>>> I won't recommend doing that. If the lock cacheline has been moved
>>> from node 1 to 2, I will say it is better to stick with node 2 =
rather
>>> than switching back to node 1. That will mean that the secondary
>>> queue may contain lock waiters from the same nodes, but they will
>>> eventually be flushed back to the primary queue.
>>>=20
>> That=E2=80=99s right, assuming we do not reset intra_node count when
>> transferring the
>> lock to a prioritized thread from another node. Otherwise, we may =
starve
>> waiters in the secondary queue.=20
>>=20
>> Still, that can make the lock even less fair to non-prioritized
>> threads. When
>> you flush the secondary queue, the preference may remain with the =
same
>> node. This will not happen in the current form of CNA, as we never =
get=20
>> threads from the preferred node in the secondary queue.
>=20
> That is true.
>=20
> However, it is no different from the current scheme that a waiter from
> another node may have to wait for 64k other waiters to go first before
> it has a chance to get it. Now that waiter can be from the same node =
as
> well.

The difference is that in the current form of CNA, the preferred node =
_will
change after 64k lock transitions. In the change you propose, this is no
longer the case. It may take another ~64k transitions for that to =
happen.
More generally, I think this makes the analysis of the lock behavior =
more
convoluted.

I think we should treat those prioritized threads as =E2=80=9Cwild=E2=80=9D=
 cards, passing the=20
lock through them, but keeping the preferred node intact. This will =
potentially
cost one extra lock migration, but will make reasoning about the lock
behavior easier.

Regards,
=E2=80=94 Alex=