Skip to contentSkip to conversation

ufs: mediatek: Fix probe failure on MT8395 SoC

PATCH Applied Mary Guillemard 5 participants 11 messages Aug 18, 2024 – Aug 29, 2024

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

cover ufs: mediatek: Fix probe failure on MT8395 SoC

Hello,

This series fixes probe failure on MT8395 SoC caused by LSDBS field in the CAP register being set while the controller only support UFSHCI 2.1.

This is based on 6.11/scsi-fixes as it requires the LSDBS quirk from [1].

This was tested on a Radxa NIO 12L with [2] and appropriate dt changes.

[1]https://lore.kernel.org/linux-scsi/[email protected]/ [2]https://lore.kernel.org/all/[email protected]/

Mary Guillemard (1):
scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

 drivers/ufs/host/ufs-mediatek.c | 3 +++
 1 file changed, 3 insertions(+)

base-commit: cbaac68987b8699397df29413b33bd51f5255255

Signature · 1 line
2.46.0
#1 lore raw

PATCH 1/1 scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in the reserved part for the Legacy Single Doorbell Support (LSDBS) capability.

This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is explicitly disabled, allowing the device to be properly registered.

---
 drivers/ufs/host/ufs-mediatek.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/ufs/host/ufs-mediatek.c +3 -0
index 02c9064284e1..9a5919434c4e 100644
@@ -1026,6 +1026,9 @@ static int ufs_mtk_init(struct ufs_hba *hba)
 	if (host->caps & UFS_MTK_CAP_DISABLE_AH8)
 		hba->caps |= UFSHCD_CAP_HIBERN8_WITH_CLK_GATING;
 
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)
+		hba->quirks |= UFSHCD_QUIRK_BROKEN_LSDBS_CAP;
+
 	ufs_mtk_init_clocks(hba);
 
 	/*
Signature · 1 line
2.46.0
#2 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

Quoted 3 lines · Mary Guillemard

MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in the reserved part for the Legacy Single Doorbell Support (LSDBS) capability.

Wow... I never thought that this quirk will be used outside of Qcom SoCs...

Quoted 13 lines

This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is explicitly disabled, allowing the device to be properly registered.

---
 drivers/ufs/host/ufs-mediatek.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/ufs/host/ufs-mediatek.c +0 -0
index 02c9064284e1..9a5919434c4e 100644

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba *hba) · line 1029

Quoted 4 lines
drivers/ufs/host/ufs-mediatek.c +1 -0
@@ -1026,6 +1026,9 @@ static int ufs_mtk_init(struct ufs_hba *hba)
 	if (host->caps & UFS_MTK_CAP_DISABLE_AH8)
 		hba->caps |= UFSHCD_CAP_HIBERN8_WITH_CLK_GATING;
 
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)

How can this be the deciding factor? You said above that the issue is with MT8183 SoC. So why not just use the quirk only for that platform?

- Mani

Quoted 9 lines
drivers/ufs/host/ufs-mediatek.c +2 -0
+		hba->quirks |= UFSHCD_QUIRK_BROKEN_LSDBS_CAP;
+
 	ufs_mtk_init_clocks(hba);
 
 	/*
Signature · 1 line
2.46.0
Signature · 1 line
மணிவண்ணன் சதாசிவம்
#3 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

Quoted 7 lines · Manivannan Sadhasivam
Quoted 3 lines · Mary Guillemard

MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in the reserved part for the Legacy Single Doorbell Support (LSDBS) capability.

Wow... I never thought that this quirk will be used outside of Qcom SoCs...

Yeah I found that by trial and error some weeks ago and noticed your serie while looking to upstream this change, quite funny to see other vendors having the same quirk here.

Quoted 16 lines
Quoted 16 lines

This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is explicitly disabled, allowing the device to be properly registered.

---
 drivers/ufs/host/ufs-mediatek.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/ufs/host/ufs-mediatek.c +0 -0
index 02c9064284e1..9a5919434c4e 100644
@@ -1026,6 +1026,9 @@ static int ufs_mtk_init(struct ufs_hba *hba)
 	if (host->caps & UFS_MTK_CAP_DISABLE_AH8)
 		hba->caps |= UFSHCD_CAP_HIBERN8_WITH_CLK_GATING;
 
Quoted 7 lines

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba *hba) · line 1029

Quoted 1 line
drivers/ufs/host/ufs-mediatek.c +1 -0
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)

How can this be the deciding factor? You said above that the issue is with MT8183 SoC. So why not just use the quirk only for that platform?

- Mani

So my current assumption is that it also affect other Mediatek SoCs that are also based on UFS 2.1 spec but I cannot check this.

Instead, we know that if MCQ isn't supported, we must fallback to LSDB as there is no other ways to drive the device.

UFS_MTK_CAP_DISABLE_MCQ (mediatek,ufs-disable-mcq) being unused upstream, I think that's an acceptable fix.

Another way to handle this would be to add a new dt property and add it to ufs_mtk_host_caps but I feel that my approach should be enough.

Quoted 4 lines
Quoted 4 lines
drivers/ufs/host/ufs-mediatek.c +2 -0
+		hba->quirks |= UFSHCD_QUIRK_BROKEN_LSDBS_CAP;
+
 	ufs_mtk_init_clocks(hba);
 
Quoted 8 lines
Quoted 5 lines
drivers/ufs/host/ufs-mediatek.c +0 -0
 	/*
Signature · 1 line
2.46.0
Signature · 1 line
மணிவண்ணன் சதாசிவம்

- Mary

#4 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

Quoted 55 lines · Mary Guillemard

External email : Please do not click links or open attachments until you have verified the sender or the content.

Quoted 2 lines · Manivannan Sadhasivam
Quoted 1 line · Mary Guillemard

MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in

the

Quoted 1 line
Quoted 1 line

reserved part for the Legacy Single Doorbell Support (LSDBS)

capability.

Quoted 3 lines
Quoted 1 line

Wow... I never thought that this quirk will be used outside of Qcom

SoCs...

Quoted 1 line

Yeah I found that by trial and error some weeks ago and noticed your serie while looking to upstream this change, quite funny to see other vendors having the same quirk here.

Quoted 1 line
Quoted 1 line

This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is

explicitly

Quoted 8 lines

Review

Quoted 8 lines

disabled, allowing the device to be properly registered.

---
 drivers/ufs/host/ufs-mediatek.c | 3 +++
 1 file changed, 3 insertions(+)
(quoted diff) +0 -0

b/drivers/ufs/host/ufs-mediatek.c

Quoted 4 lines

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba

Quoted 4 lines

index 02c9064284e1..9a5919434c4e 100644

drivers/ufs/host/ufs-mediatek.c +0 -0
@@ -1026,6 +1026,9 @@ static int ufs_mtk_init(struct ufs_hba

*hba)

Quoted 6 lines

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba *hba) · line 1029

Quoted 4 lines
drivers/ufs/host/ufs-mediatek.c +1 -0
 if (host->caps & UFS_MTK_CAP_DISABLE_AH8)
 hba->caps |= UFSHCD_CAP_HIBERN8_WITH_CLK_GATING;
 
+if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)

How can this be the deciding factor? You said above that the issue

is with

Quoted 4 lines

MT8183 SoC. So why not just use the quirk only for that platform?

- Mani

So my current assumption is that it also affect other Mediatek SoCs that are also based on UFS 2.1 spec but I cannot check this.

Instead, we know that if MCQ isn't supported, we must fallback to LSDB
as there is no other ways to drive the device.

Quoted 8 lines

UFS_MTK_CAP_DISABLE_MCQ (mediatek,ufs-disable-mcq) being unused upstream,
I think that's an acceptable fix.

Another way to handle this would be to add a new dt property and add it
to ufs_mtk_host_caps but I feel that my approach should be enough.

Hi Mary,

Yes, the MT8395 indeed requires the LSDBS flag,
but not every MediaTek legacy chip does.
So setting the LSDBS flag here is appropriate.

Thanks.
Peter

Quoted 15 lines
Quoted 4 lines
Quoted 4 lines
drivers/ufs/host/ufs-mediatek.c +2 -0
+hba->quirks |= UFSHCD_QUIRK_BROKEN_LSDBS_CAP;
+
 ufs_mtk_init_clocks(hba);
 
Quoted 8 lines
Quoted 5 lines
drivers/ufs/host/ufs-mediatek.c +0 -0
 /*
Signature · 1 line
2.46.0
Signature · 1 line
மணிவண்ணன் சதாசிவம்

- Mary

#5 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

In reply to Mary Guillemard
Quoted 37 lines · Mary Guillemard
Quoted 7 lines · Manivannan Sadhasivam
Quoted 3 lines · Mary Guillemard

MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in the reserved part for the Legacy Single Doorbell Support (LSDBS) capability.

Wow... I never thought that this quirk will be used outside of Qcom SoCs...

Yeah I found that by trial and error some weeks ago and noticed your serie while looking to upstream this change, quite funny to see other vendors having the same quirk here.

Quoted 16 lines
Quoted 16 lines

This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is explicitly disabled, allowing the device to be properly registered.

---
 drivers/ufs/host/ufs-mediatek.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/ufs/host/ufs-mediatek.c +0 -0
index 02c9064284e1..9a5919434c4e 100644
@@ -1026,6 +1026,9 @@ static int ufs_mtk_init(struct ufs_hba *hba)
 	if (host->caps & UFS_MTK_CAP_DISABLE_AH8)
 		hba->caps |= UFSHCD_CAP_HIBERN8_WITH_CLK_GATING;
 
Quoted 7 lines

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba *hba) · line 1029

Quoted 1 line
drivers/ufs/host/ufs-mediatek.c +1 -0
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)

How can this be the deciding factor? You said above that the issue is with MT8183 SoC. So why not just use the quirk only for that platform?

- Mani

Quoted 9 lines

So my current assumption is that it also affect other Mediatek SoCs that are also based on UFS 2.1 spec but I cannot check this.

Instead, we know that if MCQ isn't supported, we must fallback to LSDB as there is no other ways to drive the device.

UFS_MTK_CAP_DISABLE_MCQ (mediatek,ufs-disable-mcq) being unused upstream, I think that's an acceptable fix.

If you use this quirk, then you need to use the corresponding DT property. But using the 'mediatek,ufs-disable-mcq' property for 2.1 controller doesn't make sense as MCQ is for controllers >= 4.0.

Quoted 3 lines

Another way to handle this would be to add a new dt property and add it to ufs_mtk_host_caps but I feel that my approach should be enough.

No need to add a DT property. Just use the SoC specific compatible as I did for SM8550 SoC.

- Mani

Signature · 1 line
மணிவண்ணன் சதாசிவம்
#6 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

Quoted 22 lines · Manivannan Sadhasivam
Quoted 7 lines · Mary Guillemard
Quoted 5 lines · Manivannan Sadhasivam

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba *hba) · line 1029

Quoted 1 line · Mary Guillemard
drivers/ufs/host/ufs-mediatek.c +1 -0
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)

How can this be the deciding factor? You said above that the issue is with MT8183 SoC. So why not just use the quirk only for that platform?

Quoted 9 lines

So my current assumption is that it also affect other Mediatek SoCs that are also based on UFS 2.1 spec but I cannot check this.

Instead, we know that if MCQ isn't supported, we must fallback to LSDB as there is no other ways to drive the device.

UFS_MTK_CAP_DISABLE_MCQ (mediatek,ufs-disable-mcq) being unused upstream, I think that's an acceptable fix.

If you use this quirk, then you need to use the corresponding DT property. But using the 'mediatek,ufs-disable-mcq' property for 2.1 controller doesn't make sense as MCQ is for controllers >= 4.0.

Quoted 6 lines
Quoted 3 lines

Another way to handle this would be to add a new dt property and add it to ufs_mtk_host_caps but I feel that my approach should be enough.

No need to add a DT property. Just use the SoC specific compatible as I did for SM8550 SoC.

Mary, do you plan to implement Manivannan's feedback?

Thanks,

Bart.

#7 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

Quoted 27 lines · Bart Van Assche
Quoted 26 lines · Manivannan Sadhasivam
Quoted 7 lines · Mary Guillemard
Quoted 5 lines · Manivannan Sadhasivam

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba *hba) · line 1029

Quoted 1 line · Mary Guillemard
drivers/ufs/host/ufs-mediatek.c +1 -0
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)

How can this be the deciding factor? You said above that the issue is with MT8183 SoC. So why not just use the quirk only for that platform?

Quoted 9 lines

So my current assumption is that it also affect other Mediatek SoCs that are also based on UFS 2.1 spec but I cannot check this.

Instead, we know that if MCQ isn't supported, we must fallback to LSDB as there is no other ways to drive the device.

UFS_MTK_CAP_DISABLE_MCQ (mediatek,ufs-disable-mcq) being unused upstream, I think that's an acceptable fix.

If you use this quirk, then you need to use the corresponding DT property. But using the 'mediatek,ufs-disable-mcq' property for 2.1 controller doesn't make sense as MCQ is for controllers >= 4.0.

Quoted 3 lines

Another way to handle this would be to add a new dt property and add it to ufs_mtk_host_caps but I feel that my approach should be enough.

Quoted 9 lines
Quoted 2 lines

No need to add a DT property. Just use the SoC specific compatible as I did for SM8550 SoC.

Mary, do you plan to implement Manivannan's feedback?

Thanks,

Bart.

Hello Bart,

I think that considering Peter's reply, explicitly checking for the MT8183 controller isn't required.

I also think it could be required for at least the MT8192 and MT8195 considering they are apparently also based on UFS 2.1 spec [1].

However, if you want me to add an explicit check, I will happily send a v2.

Thanks,

Mary.

[1]https://corp.mediatek.com/news-events/press-releases/mediatek-announces-new-mt8192-and-mt8195-chipsets-designed-for-next-generation-of-chromebooks

#8 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

Quoted 38 lines · Mary Guillemard
Quoted 27 lines · Bart Van Assche
Quoted 26 lines · Manivannan Sadhasivam
Quoted 7 lines · Mary Guillemard
Quoted 5 lines · Manivannan Sadhasivam

drivers/ufs/host/ufs-mediatek.c static int ufs_mtk_init(struct ufs_hba *hba) · line 1029

Quoted 1 line · Mary Guillemard
drivers/ufs/host/ufs-mediatek.c +1 -0
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)

How can this be the deciding factor? You said above that the issue is with MT8183 SoC. So why not just use the quirk only for that platform?

Quoted 9 lines

So my current assumption is that it also affect other Mediatek SoCs that are also based on UFS 2.1 spec but I cannot check this.

Instead, we know that if MCQ isn't supported, we must fallback to LSDB as there is no other ways to drive the device.

UFS_MTK_CAP_DISABLE_MCQ (mediatek,ufs-disable-mcq) being unused upstream, I think that's an acceptable fix.

If you use this quirk, then you need to use the corresponding DT property. But using the 'mediatek,ufs-disable-mcq' property for 2.1 controller doesn't make sense as MCQ is for controllers >= 4.0.

Quoted 3 lines

Another way to handle this would be to add a new dt property and add it to ufs_mtk_host_caps but I feel that my approach should be enough.

Quoted 9 lines
Quoted 2 lines

No need to add a DT property. Just use the SoC specific compatible as I did for SM8550 SoC.

Mary, do you plan to implement Manivannan's feedback?

Thanks,

Bart.

Quoted 8 lines

Hello Bart,

I think that considering Peter's reply, explicitly checking for the MT8183 controller isn't required.

I also think it could be required for at least the MT8192 and MT8195 considering they are apparently also based on UFS 2.1 spec [1].

How can you add a quirk that is specifically meant for 4.x controllers to 2.1 controllers? It doesn't make sense.

Also it is weird that the existing DT files doesn't have ufshc nodes for any SoCs, but the SoCs are supporting UFSHC.

- Mani

Signature · 1 line
மணிவண்ணன் சதாசிவம்
maintainer Reviewed-by on v1 1/1
#9 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

In reply to Mary Guillemard
Quoted 33 lines · Mary Guillemard

External email : Please do not click links or open attachments until you have verified the sender or the content.
MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in
the
reserved part for the Legacy Single Doorbell Support (LSDBS) capability.

This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is explicitly disabled, allowing the device to be properly registered.

---
 drivers/ufs/host/ufs-mediatek.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/ufs/host/ufs- renamed from drivers/ufs/host/ufs-mediatek.c +0 -0

mediatek.c
index 02c9064284e1..9a5919434c4e 100644

drivers/ufs/host/ufs-mediatek.c +3 -0
@@ -1026,6 +1026,9 @@ static int ufs_mtk_init(struct ufs_hba *hba)
 	if (host->caps & UFS_MTK_CAP_DISABLE_AH8)
 		hba->caps |= UFSHCD_CAP_HIBERN8_WITH_CLK_GATING;
 
+	if (host->caps & UFS_MTK_CAP_DISABLE_MCQ)
+		hba->quirks |= UFSHCD_QUIRK_BROKEN_LSDBS_CAP;
+
 	ufs_mtk_init_clocks(hba);
 
 	/*
Signature · 1 line
2.46.0
#10 lore raw

Re: [PATCH 1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

In reply to Mary Guillemard
Quoted 8 lines · Mary Guillemard

MT8183 supports UFSHCI 2.1 spec, but report a bogus value of 1 in the reserved part for the Legacy Single Doorbell Support (LSDBS) capability.

This set UFSHCD_QUIRK_BROKEN_LSDBS_CAP when MCQ support is explicitly disabled, allowing the device to be properly registered.

[...]

Applied to 6.11/scsi-fixes, thanks!

[1/1] scsi: ufs-mediatek: Add UFSHCD_QUIRK_BROKEN_LSDBS_CAP

      https://git.kernel.org/mkp/scsi/c/0f9592ae26ff
Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering