[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <d09641c7-c266-4f0a-a0e3-56f63d8c9ce3@quicinc.com>
Date: Tue, 8 Apr 2025 11:07:58 +0530
From: MANISH PANDEY <quic_mapa@...cinc.com>
To: Manivannan Sadhasivam <manivannan.sadhasivam@...aro.org>
CC: "James E.J. Bottomley" <James.Bottomley@...senpartnership.com>,
"Martin K.
Petersen" <martin.petersen@...cle.com>,
Alim Akhtar
<alim.akhtar@...sung.com>,
Avri Altman <avri.altman@....com>,
Bart Van Assche
<bvanassche@....org>, <linux-scsi@...r.kernel.org>,
<linux-kernel@...r.kernel.org>, <linux-arm-msm@...r.kernel.org>,
<quic_nitirawa@...cinc.com>, <quic_bhaskarv@...cinc.com>,
<quic_rampraka@...cinc.com>, <quic_cang@...cinc.com>,
<quic_nguyenb@...cinc.com>
Subject: Re: [PATCH V2 2/2] scsi: ufs: introduce quirk to extend
PA_HIBERN8TIME for UFS devices
On 4/7/2025 12:05 AM, Manivannan Sadhasivam wrote:
> On Fri, Apr 04, 2025 at 11:15:39PM +0530, Manish Pandey wrote:
>> Some UFS devices need additional time in hibern8 mode before exiting,
>> beyond the negotiated handshaking phase between the host and device.
>> Introduce a quirk to increase the PA_HIBERN8TIME parameter by 100 µs
>> to ensure proper hibernation process.
>>
>
> This commit message didn't mention the UFS device for which this quirk is being
> applied.
>
Since it's a quirk and may be applicable to other vendors also in
future, so i thought to keep it general.
Will update in next patch set if required.
>> Signed-off-by: Manish Pandey <quic_mapa@...cinc.com>
>> ---
>> drivers/ufs/core/ufshcd.c | 31 +++++++++++++++++++++++++++++++
>> include/ufs/ufs_quirks.h | 6 ++++++
>> 2 files changed, 37 insertions(+)
>>
>> diff --git a/drivers/ufs/core/ufshcd.c b/drivers/ufs/core/ufshcd.c
>> index 464f13da259a..2b8203fe7b8c 100644
>> --- a/drivers/ufs/core/ufshcd.c
>> +++ b/drivers/ufs/core/ufshcd.c
>> @@ -278,6 +278,7 @@ static const struct ufs_dev_quirk ufs_fixups[] = {
>> .model = UFS_ANY_MODEL,
>> .quirk = UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM |
>> UFS_DEVICE_QUIRK_HOST_PA_TACTIVATE |
>> + UFS_DEVICE_QUIRK_PA_HIBER8TIME |
>> UFS_DEVICE_QUIRK_RECOVERY_FROM_DL_NAC_ERRORS },
>> { .wmanufacturerid = UFS_VENDOR_SKHYNIX,
>> .model = UFS_ANY_MODEL,
>> @@ -8384,6 +8385,33 @@ static int ufshcd_quirk_tune_host_pa_tactivate(struct ufs_hba *hba)
>> return ret;
>> }
>>
>> +/**
>> + * ufshcd_quirk_override_pa_h8time - Ensures proper adjustment of PA_HIBERN8TIME.
>> + * @hba: per-adapter instance
>> + *
>> + * Some UFS devices require specific adjustments to the PA_HIBERN8TIME parameter
>> + * to ensure proper hibernation timing. This function retrieves the current
>> + * PA_HIBERN8TIME value and increments it by 100us.
>> + */
>> +static void ufshcd_quirk_override_pa_h8time(struct ufs_hba *hba)
>> +{
>> + u32 pa_h8time = 0;
>
> Why do you need to initialize it?
>
Agree.. Not needed, will update.>> + int ret;
>> +
>> + ret = ufshcd_dme_get(hba, UIC_ARG_MIB(PA_HIBERN8TIME),
>> + &pa_h8time);
>> + if (ret) {
>> + dev_err(hba->dev, "Failed to get PA_HIBERN8TIME: %d\n", ret);
>> + return;
>> + }
>> +
>> + /* Increment by 1 to increase hibernation time by 100 µs */
>
> From where the value of 100us adjustment is coming from?
>
> - Mani
>
These values are derived from experiments on Qualcomm SoCs.
However this is also matching with ufs-exynos.c
fsd_ufs_post_link() {
ufshcd_dme_get(hba,UIC_ARG_MIB(PA_HIBERN8TIME),
&max_rx_hibern8_time_cap);
.......
ufshcd_dme_set(hba, UIC_ARG_MIB(PA_HIBERN8TIME),
max_rx_hibern8_time_cap + 1);
...
}
Thanks
Manish
Powered by blists - more mailing lists