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;
Message-ID: <d67b8969c8d2097377d9dbb414481f11958a9c9f.camel@gmail.com>
Subject: Re: [PATCH 1/1] scsi: ufs: remove redundant HPB unmap
From:   Bean Huo <huobean@gmail.com>
To:     Peter Wang <peter.wang@mediatek.com>,
        Po-Wen Kao <powen.kao@mediatek.com>,
        Alim Akhtar <alim.akhtar@samsung.com>,
        Avri Altman <avri.altman@wdc.com>,
        "James E.J. Bottomley" <jejb@linux.ibm.com>,
        "Martin K. Petersen" <martin.petersen@oracle.com>,
        Matthias Brugger <matthias.bgg@gmail.com>
Cc:     wsd_upstream@mediatek.com, stanley.chu@mediatek.com,
        alice.chao@mediatek.com, chun-hung.wu@mediatek.com,
        cc.chou@mediatek.com, chaotian.jing@mediatek.com,
        jiajie.hao@mediatek.com, yohan.joung@sk.com, jason.li@sk.com,
        linux-scsi@vger.kernel.org, linux-kernel@vger.kernel.org,
        linux-arm-kernel@lists.infradead.org,
        linux-mediatek@lists.infradead.org
Date:   Wed, 13 Apr 2022 14:43:38 +0200
In-Reply-To: <7e979863-21e5-ba63-862a-8ba6a63a8ae4@mediatek.com>
References: <20220412073131.10644-1-powen.kao@mediatek.com>
         <781bfe33efb1e44d22f71e6711da7d15be9d3ab4.camel@gmail.com>
         <7550da4043ae45ffd58399f1ee696238a5689f25.camel@gmail.com>
         <7e979863-21e5-ba63-862a-8ba6a63a8ae4@mediatek.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
User-Agent: Evolution 3.44.0-1 
MIME-Version: 1.0
Precedence: bulk

Hi Peter,


On Wed, 2022-04-13 at 11:53 +0800, Peter Wang wrote:
> Hi Bean,
>=20
> How about if reset flow not clear hpb mapping?

UFS will go through normal read, because regions are not active in the
device. The responses mabye different among different vendors. for
example, UFS device will raise a response to host to update the region
(in case of host mode) or inactivate it in case of device mode.

The best way is that inactivate the active regions in the host side in
the case the regions are not active in the device.

>=20
> After device is reset by RST_N pin, if the hpb mapping table still
> exist?

no, the regions in the device will be inactivated.

>=20
> BTW, Host seem not clear region mapping table in reset flow.
>=20

Yes, you are right. There will be a performance impact. I will look at
it, then back to you.

By the way, there is one issue in my patch, ignore that patch.


Kind regards,
Bean