Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   =?UTF-8?B?7ISc7Zi47JiB?= <hy50.seo@samsung.com>
To:     <linux-scsi@vger.kernel.org>, <linux-kernel@vger.kernel.org>,
        <alim.akhtar@samsung.com>, <avri.altman@wdc.com>,
        <jejb@linux.ibm.com>, <martin.petersen@oracle.com>,
        <beanhuo@micron.com>, <asutoshd@codeaurora.org>,
        <cang@codeaurora.org>, <bvanassche@acm.org>,
        <bhoon95.kim@samsung.com>, <kwmad.kim@samsung.com>
Cc:     "'kernel test robot'" <lkp@intel.com>
In-Reply-To: <40986ecb6c81812a1e1ab24d93e46eda75974c4e.camel@mediatek.com>
Subject: RE: [PATCH v2] scsi: ufs: disable auto hibern8 while entering
 suspend
Date:   Wed, 26 Jan 2022 14:34:51 +0900
Message-ID: <101901d81276$70d9ea40$528dbec0$@samsung.com>
MIME-Version: 1.0
Content-Transfer-Encoding: quoted-printable
Thread-Index: AQIlzz+JKCbQP4e9lbbSHSqpu8nHvgHyie3oAmFa8EWrtq73AA==
Content-Language: ko
Content-Type: text/plain; charset="utf-8"
CMS-TYPE: 102P
DLP-Filter: Pass
References: <CGME20220125062155epcas2p15da28303164091b1bf5a00dcf99fe59b@epcas2p1.samsung.com>
        <20220124180637.160524-1-hy50.seo@samsung.com>
        <40986ecb6c81812a1e1ab24d93e46eda75974c4e.camel@mediatek.com>
Precedence: bulk

Hi,
I think content is lacking in the UFSHCI spec.
In the process, AH8 is in operation, and it seems that sending the command
hibern8 by manual has a defeat.
I don't know what the all hci vendor's hardware design will be, but there i=
s a possibility that ah8 and manual hibern8 may overlap.
So if is operating in ah8, it is thought that it will be safer to disable a=
h8 before sending hibern8 command.

> -----Original Message-----
> From: Stanley Chu =5Bmailto:stanley.chu=40mediatek.com=5D
> Sent: Wednesday, January 26, 2022 10:22 AM
> To: SEO HOYOUNG; linux-scsi=40vger.kernel.org; linux-kernel=40vger.kernel=
.org;
> alim.akhtar=40samsung.com; avri.altman=40wdc.com; jejb=40linux.ibm.com;
> martin.petersen=40oracle.com; beanhuo=40micron.com; asutoshd=40codeaurora=
.org;
> cang=40codeaurora.org; bvanassche=40acm.org; bhoon95.kim=40samsung.com;
> kwmad.kim=40samsung.com
> Cc: kernel test robot; peter.wang=40mediatek.com
> Subject: Re: =5BPATCH v2=5D scsi: ufs: disable auto hibern8 while enterin=
g
> suspend
>=20
> Hi Hoyoung,
>=20
> On Tue, 2022-01-25 at 03:06 +0900, SEO HOYOUNG wrote:
> > v1-> v2: fixed no previous prototype warning
> > Reported-by: kernel test robot <lkp=40intel.com>
> >
> > All warnings (new ones prefixed by >>):
> > > > drivers/scsi/ufs/ufshcd.c:4207:6: warning: no previous prototype
> >
> > for 'ufshcd_auto_hibern8_disable' =5B-Wmissing-prototypes=5D
> > 4207 =7C void ufshcd_auto_hibern8_disable(struct ufs_hba *hba)
> >
> > If using auto hibern8 mode, need to disable auto hibern8 while
> > entering suspend.
> > When using auto hibern8 mode, it does not seem right to send a uic
> > command
>=20
> The UFSHCI spec does not mention the above rule.
> Why would you need to disable AH8 before using UIC command to enter H8?
>=20
> > for entry into hibern8 in the next
> > line(ufshcd_lik_state_transition(..))
> > It seem right to send after disable auto hibern8.
> >
> > In addition, if the auto hibern8 mode supported, it is enabled in
> > resume.
> > So it seems that it will be paired only when auto hibern8 is disabled
> > while entering suspend.
> >
> > Signed-off-by: SEO HOYOUNG <hy50.seo=40samsung.com>
> > ---
> >  drivers/scsi/ufs/ufshcd.c =7C 14 ++++++++++++++
> >  1 file changed, 14 insertions(+)
> >
> > diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> > index 460d2b440d2e..a6edbbd8ca2c 100644
> > --- a/drivers/scsi/ufs/ufshcd.c
> > +++ b/drivers/scsi/ufs/ufshcd.c
> > =40=40 -254,6 +254,7 =40=40 static void
> > ufshcd_wb_toggle_flush_during_h8(struct ufs_hba *hba, bool set);
> > static inline void ufshcd_wb_toggle_flush(struct ufs_hba *hba, bool
> > enable);  static void ufshcd_hba_vreg_set_lpm(struct ufs_hba *hba);
> > static void ufshcd_hba_vreg_set_hpm(struct ufs_hba *hba);
> > +static void ufshcd_auto_hibern8_disable(struct ufs_hba *hba);
> >
> >  static inline void ufshcd_enable_irq(struct ufs_hba *hba)  =7B =40=40
> > -4204,6 +4205,18 =40=40 void ufshcd_auto_hibern8_update(struct ufs_hba
> > *hba, u32 ahit)  =7D  EXPORT_SYMBOL_GPL(ufshcd_auto_hibern8_update);
> >
> > +static void ufshcd_auto_hibern8_disable(struct ufs_hba *hba) =7B
> > +	unsigned long flags;
> > +
> > +	if (=21ufshcd_is_auto_hibern8_supported(hba))
> > +		return;
> > +
> > +	spin_lock_irqsave(hba->host->host_lock, flags);
> > +	ufshcd_writel(hba, 0, REG_AUTO_HIBERNATE_IDLE_TIMER);
> > +	spin_unlock_irqrestore(hba->host->host_lock, flags); =7D
> > +
> >  void ufshcd_auto_hibern8_enable(struct ufs_hba *hba)  =7B
> >  	unsigned long flags;
> > =40=40 -8925,6 +8938,7 =40=40 static int __ufshcd_wl_suspend(struct ufs=
_hba
> > *hba, enum ufs_pm_op pm_op)
> >  	 * with the link off, so do not check for bkops.
> >  	 */
> >  	check_for_bkops =3D =21ufshcd_is_ufs_dev_deepsleep(hba);
> > +	ufshcd_auto_hibern8_disable(hba);
> >  	ret =3D ufshcd_link_state_transition(hba, req_link_state,
> > check_for_bkops);
> >  	if (ret)
> >  		goto set_dev_active;