Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
Date:   Wed, 26 Oct 2022 06:18:03 +0000
Content-Type: text/plain; charset="utf-8"
Content-Transfer-Encoding: quoted-printable
From:   "Yajun Deng" <yajun.deng@linux.dev>
Message-ID: <b4af84bdc4b9c757c9fea2e63dd98a61@linux.dev>
Subject: Re: [PATCH] memblock: remove repeat round
To:     "Mike Rapoport" <rppt@kernel.org>
Cc:     akpm@linux-foundation.org, linux-mm@kvack.org,
        linux-kernel@vger.kernel.org
In-Reply-To: <Y1jN/KeU6dgJWQIL@kernel.org>
References: <Y1jN/KeU6dgJWQIL@kernel.org> <Y1a6BoawCoDDCo/K@kernel.org>
 <20221019120337.2098298-1-yajun.deng@linux.dev>
 <6cccd4cba4f8407d52815c68f60446fb@linux.dev>
Precedence: bulk

October 26, 2022 2:04 PM, "Mike Rapoport" <rppt@kernel.org> wrote:=0A=0A>=
 On Tue, Oct 25, 2022 at 03:24:58AM +0000, Yajun Deng wrote:=0A> =0A>> Oc=
tober 25, 2022 12:15 AM, "Mike Rapoport" <rppt@kernel.org> wrote:=0A>> =
=0A>> @@ -598,22 +597,6 @@ static int __init_memblock memblock_add_range(=
struct memblock_type *type,=0A>> return 0;=0A>> }=0A>> =0A>> - /*=0A>> - =
* The worst case is when new range overlaps all existing regions,=0A>> - =
* then we'll need type->cnt + 1 empty regions in @type. So if=0A>> - * ty=
pe->cnt * 2 + 1 is less than type->max, we know=0A>> - * that there is en=
ough empty regions in @type, and we can insert=0A>> - * regions directly.=
=0A>> - */=0A>> - if (type->cnt * 2 + 1 < type->max)=0A>> - insert =3D tr=
ue;=0A>> -=0A>> -repeat:=0A>> - /*=0A>> - * The following is executed twi=
ce. Once with %false @insert and=0A>> - * then with %true. The first coun=
ts the number of regions needed=0A>> - * to accommodate the new area. The=
 second actually inserts them.=0A>> - */=0A>> base =3D obase;=0A>> nr_new=
 =3D 0;=0A>> =0A>> I believe nr_new variable is no longer needed, is it?=
=0A>> =0A>> No, nr_new is needed before memblock_merge_regions() for retu=
rn.=0A> =0A> Why?=0A> =0ASorry, nr_new was removed, and added ocnt variab=
le for the original of type->cnt.=0A=0AI already sent another patch with =
'[PATCH v2] memblock: don't run loop in =0Amemblock_add_range() twice' su=
bject, you can see that.=0A=0A> --=0A> Sincerely yours,=0A> Mike.