Received-SPF: pass (google.com: domain of linux-kernel+bounces-60252-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Content-Type: text/plain; charset=utf-8
Content-Transfer-Encoding: quoted-printable
Precedence: bulk
Mime-Version: 1.0
Subject: Re: [PATCH 1/1] hugetlb: fix CONFIG_PADATA dependency for non-SMP system
From: Muchun Song <muchun.song@linux.dev>
In-Reply-To: <86B96EE5-A51C-4F06-BB18-0B3B12DBCCEC@linux.dev>
Cc: Gang Li <ligang.bdlg@bytedance.com>,
 Andrew Morton <akpm@linux-foundation.org>, linux-kernel@vger.kernel.org,
 linux-mm@kvack.org, Randy Dunlap <rdunlap@infradead.org>,
 kernel test robot <lkp@intel.com>
Date: Sat, 10 Feb 2024 10:45:35 +0800
Message-Id: <2B59E906-4CC9-4926-A893-4954587D3BCD@linux.dev>
References: <86B96EE5-A51C-4F06-BB18-0B3B12DBCCEC@linux.dev>
To: Gang Li <gang.li@linux.dev>, jane.chu@oracle.com,
 Daniel Jordan <daniel.m.jordan@oracle.com>, steffen.klassert@secunet.com


CC Jane, Daniel Jordan, Steffen as well in this thread.

> On Feb 8, 2024, at 10:39, Muchun Song <muchun.song@linux.dev> wrote:
> =EF=BB=BF
>=20
>> On Feb 5, 2024, at 16:08, Gang Li <gang.li@linux.dev> wrote:
>>=20
>> =EF=BB=BFOn 2024/2/5 15:37, Muchun Song wrote:
>>> Actually, I did not get it. Why the above code cannot work? The above
>>> code already make it serialized in one call, right? What do I miss here?=

>>> Thanks.
>>=20
>> PADATA consists of two distinct functionality:
>>=20
>> One part is `padata_do_multithreaded`, which disregards
>> order and simply divides tasks into several groups for parallel
>> execution. My patch use `padata_do_multithreaded`.
>=20
> OK. Since all users of PADATA of non-SMP case currently only use padata_do=
_multithreaded, it=E2=80=99s possible and easy to implement a variant of tha=
t in non-SMP case. It is not necessary to implement another functionality un=
less the only user of crypto/pcrypt.c does not depend on SMP in the future, a=
t least now it=E2=80=99s dependent on SMP.
>=20
> Thanks.
>=20
>>=20
>> The other part is composed of a set of APIs that, while handling data in
>> an out-of-order parallel manner, can eventually return the data with
>> ordered sequence. Only `crypto/pcrypt.c` use them. I guess these APIs
>> are designed specifically for `crypto/pcrypt.c`.
>> ```
>> padata_alloc
>> padata_alloc_shell
>> padata_do_parallel
>> padata_do_serial
>> padata_free_shell
>> padata_free
>> ```