[PATCH 0/3] add fixup_dev_quirks vops for ufs-qcom

[PATCH 0/3] add fixup_dev_quirks vops for ufs-qcom

[Manish Pandey]

Add fixup_dev_quirk vops in QCOM UFS platforms and provide
an initial vendor-specific device quirk table.

Micron and Skhynix UFS device needs DELAY_BEFORE_LPM quirk to have a delay
before VCC is powered off. So add Micron and Skhynix vendors ID and this
quirk for both.

Toshiba UFS devices require delay after VCC power rail is turned-off in QCOM
platforms. Hence add Toshiba vendor ID and DELAY_AFTER_LPM quirk for Toshiba
UFS devices in QCOM platforms.

Manish Pandey (3):
  scsi: ufs: ufs-qcom: add fixup_dev_quirks vops
  scsi: ufs: ufs-qcom: Add DELAY_BEFORE_LPM quirk for Micron and Skhynix
  scsi: ufs: ufs-qcom: Apply DELAY_AFTER_LPM quirk for Toshiba devices

 drivers/ufs/host/ufs-qcom.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

-- 
2.17.1

[PATCH 1/3] scsi: ufs: ufs-qcom: add fixup_dev_quirks vops

[Manish Pandey]

Add fixup_dev_quirk vops in QCOM UFS platforms and provide an initial
vendor-specific device quirk table to add UFS device specific quirks
which are enabled only for specified UFS devices.

Signed-off-by: Manish Pandey <quic_mapa@quicinc.com>
---
 drivers/ufs/host/ufs-qcom.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/ufs/host/ufs-qcom.c b/drivers/ufs/host/ufs-qcom.c
index 810e637047d0..290558843ca5 100644
--- a/drivers/ufs/host/ufs-qcom.c
+++ b/drivers/ufs/host/ufs-qcom.c
@@ -834,6 +834,16 @@ static int ufs_qcom_apply_dev_quirks(struct ufs_hba *hba)
 	return err;
 }
 
+static struct ufs_dev_quirk ufs_qcom_dev_fixups[] = {
+	/* add UFS device specific quirks */
+	{}
+};
+
+static void ufs_qcom_fixup_dev_quirks(struct ufs_hba *hba)
+{
+	ufshcd_fixup_dev_quirks(hba, ufs_qcom_dev_fixups);
+}
+
 static u32 ufs_qcom_get_ufs_hci_version(struct ufs_hba *hba)
 {
 	return ufshci_version(2, 0);
@@ -1798,6 +1808,7 @@ static const struct ufs_hba_variant_ops ufs_hba_qcom_vops = {
 	.link_startup_notify    = ufs_qcom_link_startup_notify,
 	.pwr_change_notify	= ufs_qcom_pwr_change_notify,
 	.apply_dev_quirks	= ufs_qcom_apply_dev_quirks,
+	.fixup_dev_quirks       = ufs_qcom_fixup_dev_quirks,
 	.suspend		= ufs_qcom_suspend,
 	.resume			= ufs_qcom_resume,
 	.dbg_register_dump	= ufs_qcom_dump_dbg_regs,
-- 
2.17.1

[PATCH 2/3] scsi: ufs: ufs-qcom: Add DELAY_BEFORE_LPM quirk for Micron and Skhynix

[Manish Pandey]

Micron and Skhynix UFS device needs DELAY_BEFORE_LPM quirk to have a
delay before VCC is powered off. So add Micron and SKHYNIX vendors ID
and this quirk for both devices in QCOM platforms.

Signed-off-by: Manish Pandey <quic_mapa@quicinc.com>
---
 drivers/ufs/host/ufs-qcom.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/ufs/host/ufs-qcom.c b/drivers/ufs/host/ufs-qcom.c
index 290558843ca5..6d2622e79d3f 100644
--- a/drivers/ufs/host/ufs-qcom.c
+++ b/drivers/ufs/host/ufs-qcom.c
@@ -836,6 +836,12 @@ static int ufs_qcom_apply_dev_quirks(struct ufs_hba *hba)
 
 static struct ufs_dev_quirk ufs_qcom_dev_fixups[] = {
 	/* add UFS device specific quirks */
+	{ .wmanufacturerid = UFS_VENDOR_MICRON,
+	  .model = UFS_ANY_MODEL,
+	  .quirk = UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM },
+	{ .wmanufacturerid = UFS_VENDOR_SKHYNIX,
+	  .model = UFS_ANY_MODEL,
+	  .quirk = UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM },
 	{}
 };
 
-- 
2.17.1

[PATCH 3/3] scsi: ufs: ufs-qcom: Apply DELAY_AFTER_LPM quirk for Toshiba devices

[Manish Pandey]

Toshiba UFS devices require delay after VCC power rail is turned-off
in QCOM platforms. Hence add Toshiba vendor ID and DELAY_AFTER_LPM
quirk for Toshiba UFS devices in QCOM platforms.

Signed-off-by: Manish Pandey <quic_mapa@quicinc.com>
---
 drivers/ufs/host/ufs-qcom.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/ufs/host/ufs-qcom.c b/drivers/ufs/host/ufs-qcom.c
index 6d2622e79d3f..61f8c3f22ab9 100644
--- a/drivers/ufs/host/ufs-qcom.c
+++ b/drivers/ufs/host/ufs-qcom.c
@@ -842,6 +842,9 @@ static struct ufs_dev_quirk ufs_qcom_dev_fixups[] = {
 	{ .wmanufacturerid = UFS_VENDOR_SKHYNIX,
 	  .model = UFS_ANY_MODEL,
 	  .quirk = UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM },
+	{ .wmanufacturerid = UFS_VENDOR_TOSHIBA,
+	  .model = UFS_ANY_MODEL,
+	  .quirk = UFS_DEVICE_QUIRK_DELAY_AFTER_LPM },
 	{}
 };
 
-- 
2.17.1

Re: [PATCH 1/3] scsi: ufs: ufs-qcom: add fixup_dev_quirks vops

[Bart Van Assche]

On 8/20/24 5:37 AM, Manish Pandey wrote:
> +static struct ufs_dev_quirk ufs_qcom_dev_fixups[] = {
> +	/* add UFS device specific quirks */
> +	{}
> +};

Please change the comment into /* UFS device-specific quirks */ and move
it above the array definition. I think that's the style followed by
kernel code.

Thanks,

Bart.

Re: [PATCH 2/3] scsi: ufs: ufs-qcom: Add DELAY_BEFORE_LPM quirk for Micron and Skhynix

[Bart Van Assche]

On 8/20/24 5:37 AM, Manish Pandey wrote:
>   static struct ufs_dev_quirk ufs_qcom_dev_fixups[] = {
>   	/* add UFS device specific quirks */
> +	{ .wmanufacturerid = UFS_VENDOR_MICRON,
> +	  .model = UFS_ANY_MODEL,
> +	  .quirk = UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM },
> +	{ .wmanufacturerid = UFS_VENDOR_SKHYNIX,
> +	  .model = UFS_ANY_MODEL,
> +	  .quirk = UFS_DEVICE_QUIRK_DELAY_BEFORE_LPM },
>   	{}
>   };

What makes these quirks specific for Qualcomm controllers? Are these
quirks perhaps required for all UFSHCI controllers?

Thanks,

Bart.

Re: [PATCH 3/3] scsi: ufs: ufs-qcom: Apply DELAY_AFTER_LPM quirk for Toshiba devices

[Bart Van Assche]

On 8/20/24 5:37 AM, Manish Pandey wrote:
> +	{ .wmanufacturerid = UFS_VENDOR_TOSHIBA,
> +	  .model = UFS_ANY_MODEL,
> +	  .quirk = UFS_DEVICE_QUIRK_DELAY_AFTER_LPM },

Isn't three patches a bit much for these changes? I think all three
patches can be combined into a single patch without making it harder for
reviewers to understand what is going on.

Thanks,

Bart.

Re: [PATCH 3/3] scsi: ufs: ufs-qcom: Apply DELAY_AFTER_LPM quirk for Toshiba devices

[MANISH PANDEY]

On 8/21/2024 3:05 AM, Bart Van Assche wrote:
> On 8/20/24 5:37 AM, Manish Pandey wrote:
>> +    { .wmanufacturerid = UFS_VENDOR_TOSHIBA,
>> +      .model = UFS_ANY_MODEL,
>> +      .quirk = UFS_DEVICE_QUIRK_DELAY_AFTER_LPM },
> 
> Isn't three patches a bit much for these changes? I think all three
> patches can be combined into a single patch without making it harder for
> reviewers to understand what is going on.
> 
> Thanks,
> 
> Bart.
> 
> 

Thanks Bart for quick review.
I will merge all 3 changes to a single change in next patch set [V2].

Regards
Manish