[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <DM6PR04MB657586C4A2B162F5B1E7D608FC509@DM6PR04MB6575.namprd04.prod.outlook.com>
Date: Mon, 10 Jan 2022 13:58:33 +0000
From: Avri Altman <Avri.Altman@....com>
To: μνΈμ <hy50.seo@...sung.com>,
"linux-scsi@...r.kernel.org" <linux-scsi@...r.kernel.org>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"alim.akhtar@...sung.com" <alim.akhtar@...sung.com>,
"jejb@...ux.ibm.com" <jejb@...ux.ibm.com>,
"martin.petersen@...cle.com" <martin.petersen@...cle.com>,
"beanhuo@...ron.com" <beanhuo@...ron.com>,
"asutoshd@...eaurora.org" <asutoshd@...eaurora.org>,
"cang@...eaurora.org" <cang@...eaurora.org>,
"bvanassche@....org" <bvanassche@....org>
Subject: RE: [PATCH v2] scsi: ufs: modify Tactive time setting conditions
>
> Hi,
> If register "UFS_DEVICE_QUIRK_HOST_PA_TACTIVATE" quirk,
> using this Tact time function.
> And I mean all Samsung device use UFS_DEVICE_QUIRK_HOST_PA_TACTIVATE
> quirk below this. so I told that.
> static struct ufs_dev_fix ufs_fixups[] = {
> /* UFS cards deviations table */
> UFS_FIX(UFS_VENDOR_MICRON, UFS_ANY_MODEL,
> UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM |
> UFS_DEVICE_QUIRK_SWAP_L2P_ENTRY_FOR_HPB_READ),
> UFS_FIX(UFS_VENDOR_SAMSUNG, UFS_ANY_MODEL,
> UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM |
> UFS_DEVICE_QUIRK_HOST_PA_TACTIVATE |
>
> Other vendors can use it if necessary.
> And for stability, the device tact time must be longer than host tact time.
> (I already check Major device vendors)
> That's reason why use this function.
> But the Host tact time and Device tact time same, stability may not be satisfied.
> So I changed this way
Yes - I agree.
Looks good to me.
Thanks,
Avri
>
> Thanks,
> HOYOUNG.
>
> > -----Original Message-----
> > From: Avri Altman [mailto:Avri.Altman@....com]
> > Sent: Friday, January 7, 2022 7:57 PM
> > To: SEO HOYOUNG; linux-scsi@...r.kernel.org; linux-kernel@...r.kernel.org;
> > alim.akhtar@...sung.com; jejb@...ux.ibm.com;
> martin.petersen@...cle.com;
> > beanhuo@...ron.com; asutoshd@...eaurora.org; cang@...eaurora.org;
> > bvanassche@....org
> > Subject: RE: [PATCH v2] scsi: ufs: modify Tactive time setting conditions
> >
> > > The Tactive time determine the waiting time before burst at hibern8
> > > exit and is determined by H/W at linkup state However, in the case of
> > > samsung devices, guided host's Tactive time
> > > +100us for stability.
> > > If the HCI's Tactive time is equal or greater than the device,
> > > +100us should be set.
> > This way you are changing this for everyone - not just for Samsung.
> > e.g. Qualcomm are using this quirk as well for WDC devices.
> >
> > Thanks,
> > Avri
> >
> > >
> > > Signed-off-by: SEO HOYOUNG <hy50.seo@...sung.com>
> > > ---
> > > drivers/scsi/ufs/ufshcd.c | 2 +-
> > > 1 file changed, 1 insertion(+), 1 deletion(-)
> > >
> > > diff --git a/drivers/scsi/ufs/ufshcd.c b/drivers/scsi/ufs/ufshcd.c
> > > index 1049e41abd5b..460d2b440d2e 100644
> > > --- a/drivers/scsi/ufs/ufshcd.c
> > > +++ b/drivers/scsi/ufs/ufshcd.c
> > > @@ -7815,7 +7815,7 @@ static int
> > > ufshcd_quirk_tune_host_pa_tactivate(struct
> > > ufs_hba *hba)
> > > peer_pa_tactivate_us = peer_pa_tactivate *
> > > gran_to_us_table[peer_granularity - 1];
> > >
> > > - if (pa_tactivate_us > peer_pa_tactivate_us) {
> > > + if (pa_tactivate_us >= peer_pa_tactivate_us) {
> > > u32 new_peer_pa_tactivate;
> > >
> > > new_peer_pa_tactivate = pa_tactivate_us /
> > > --
> > > 2.26.0
>
Powered by blists - more mailing lists