Skip to contentSkip to conversation

scsi: ufs: ufs-mediatek: Add UFS host support for MT8195 SoC

PATCH Applied Macpaul Lin 7 participants 25 messages Jul 22, 2025 – Oct 20, 2025

Subsystems (6)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/4 scsi: ufs: ufs-mediatek: Add UFS host support for MT8195 SoC

Add "mediatek,mt8195-ufshci" to the of_device_id table to enable support for MediaTek MT8195/MT8395 UFS host controller. This matches the device node entry in the MT8195/MT8395 device tree and allows proper driver binding.

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

Changes vor f2:
- No change.

drivers/ufs/host/ufs-mediatek.c +1 -0
index 182f58d0c9db..e1dbf0231c5e 100644
@@ -50,6 +50,7 @@ static const struct ufs_dev_quirk ufs_mtk_dev_fixups[] = {
 
 static const struct of_device_id ufs_mtk_of_match[] = {
 	{ .compatible = "mediatek,mt8183-ufshci" },
+	{ .compatible = "mediatek,mt8195-ufshci" },
 	{},
 };
 MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
Signature · 1 line
2.45.2
#1 lore raw

PATCH v2 2/4 dt-bindings: ufs: mediatek,ufs: add ufs-disable-mcq flag for UFS host

Add the 'mediatek,ufs-disable-mcq' property to the UFS device-tree bindings. This flag corresponds to the UFS_MTK_CAP_DISABLE_MCQ host capability recently introduced in the UFS host driver, allowing it to disable the Multiple Circular Queue (MCQ) feature when present. The binding schema has also been updated to resolve DTBS check errors.

Fixes 46bd3e31d74b ("scsi: ufs: mediatek: Add UFS_MTK_CAP_DISABLE_MCQ")
---
 Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml | 4 ++++
 1 file changed, 4 insertions(+)

Changes for v2:
- Split new property from the origin patch.
- Add dependency description. Since the code in ufs-mediatek.c has been backport to stable tree. The dt-bindings should be backport to the same stable tree as well.

Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +4 -0
index 32fd535a514a..20f341d25ebc 100644
@@ -33,6 +33,10 @@ properties:
 
   vcc-supply: true
 
+  mediatek,ufs-disable-mcq:
+    $ref: /schemas/types.yaml#/definitions/flag
+    description: The mask to disable MCQ (Multi-Circular Queue) for UFS host.
+
 required:
   - compatible
   - clocks
Signature · 1 line
2.45.2
maintainer Reviewed-by on v2 2/4
#2 lore raw

Re: [PATCH v2 2/4] dt-bindings: ufs: mediatek,ufs: add ufs-disable-mcq flag for UFS host

Quoted 13 lines · Macpaul Lin

Add the 'mediatek,ufs-disable-mcq' property to the UFS device-tree bindings. This flag corresponds to the UFS_MTK_CAP_DISABLE_MCQ host capability recently introduced in the UFS host driver, allowing it to disable the Multiple Circular Queue (MCQ) feature when present. The binding schema has also been updated to resolve DTBS check errors.

Fixes 46bd3e31d74b ("scsi: ufs: mediatek: Add UFS_MTK_CAP_DISABLE_MCQ")
---
 Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml | 4 ++++
 1 file changed, 4 insertions(+)
Quoted 6 lines

Changes for v2:
- Split new property from the origin patch.
- Add dependency description. Since the code in ufs-mediatek.c has been backport to stable tree. The dt-bindings should be backport to the same stable tree as well.

maintainer Reviewed-by on v2 2/4
#3 lore raw

Re: [PATCH v2 2/4] dt-bindings: ufs: mediatek,ufs: add ufs-disable-mcq flag for UFS host

In reply to Macpaul Lin
Quoted 31 lines · Macpaul Lin

Add the 'mediatek,ufs-disable-mcq' property to the UFS device-tree bindings. This flag corresponds to the UFS_MTK_CAP_DISABLE_MCQ host capability recently introduced in the UFS host driver, allowing it to disable the Multiple Circular Queue (MCQ) feature when present. The binding schema has also been updated to resolve DTBS check errors.

Fixes 46bd3e31d74b ("scsi: ufs: mediatek: Add

UFS_MTK_CAP_DISABLE_MCQ")

---
 Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml | 4 ++++

 1 file changed, 4 insertions(+)

Changes for v2:
 - Split new property from the origin patch.
 - Add dependency description. Since the code in ufs-mediatek.c has been backport to stable tree. The dt-bindings should be backport
   to the same stable tree as well.

Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +0 -0
index 32fd535a514a..20f341d25ebc 100644
@@ -33,6 +33,10 @@ properties:


   vcc-supply: true

Quoted 8 lines

+  mediatek,ufs-disable-mcq:
+    $ref: /schemas/types.yaml#/definitions/flag
+    description: The mask to disable MCQ (Multi-Circular Queue) for UFS host.
+
 required:
   - compatible
   - clocks

#4 lore raw

PATCH v2 3/4 dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

In reply to Macpaul Lin

Add MT8195 UFSHCI compatible string.
Relax the schema to allow between one to eight clocks/clock-names entries for all MediaTek UFS nodes. Legacy platforms may only need a few clocks, whereas newer devices such as the MT8195 require additional clock-gating domains. For MT8195 specifically, enforce exactly eight clocks and clock-names entries to satisfy its hardware requirements.

---
 .../devicetree/bindings/ufs/mediatek,ufs.yaml | 42 ++++++++++++++++---
 1 file changed, 36 insertions(+), 6 deletions(-)

Changes for v2:
- Remove duplicate minItems and maxItems as suggested in the review.
- Add a description of how the MT8195 hardware differs from earlier platforms.

Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +36 -6
index 20f341d25ebc..1dec54fb00f3 100644
@@ -9,21 +9,20 @@ title: Mediatek Universal Flash Storage (UFS) Controller
 maintainers:
   - Stanley Chu <[email protected]>
 
-allOf:
-  - $ref: ufs-common.yaml
-
 properties:
   compatible:
     enum:
       - mediatek,mt8183-ufshci
       - mediatek,mt8192-ufshci
+      - mediatek,mt8195-ufshci
 
   clocks:
-    maxItems: 1
+    minItems: 1
+    maxItems: 8
 
   clock-names:
-    items:
-      - const: ufs
+    minItems: 1
+    maxItems: 8
 
   phys:
     maxItems: 1
@@ -47,6 +46,37 @@ required:
 
 unevaluatedProperties: false
 
+allOf:
+  - $ref: ufs-common.yaml
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - mediatek,mt8195-ufshci
+    then:
+      properties:
+        clocks:
+          minItems: 8
+        clock-names:
+          items:
+            - const: ufs
+            - const: ufs_aes
+            - const: ufs_tick
+            - const: unipro_sysclk
+            - const: unipro_tick
+            - const: unipro_mp_bclk
+            - const: ufs_tx_symbol
+            - const: ufs_mem_sub
+    else:
+      properties:
+        clocks:
+          maxItems: 1
+        clock-names:
+          items:
+            - const: ufs
+
 examples:
   - |
     #include <dt-bindings/clock/mt8183-clk.h>
Signature · 1 line
2.45.2
#5 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Il 22/07/25 10:57, Macpaul Lin ha scritto:

Quoted 64 lines

Add MT8195 UFSHCI compatible string.
Relax the schema to allow between one to eight clocks/clock-names entries for all MediaTek UFS nodes. Legacy platforms may only need a few clocks, whereas newer devices such as the MT8195 require additional clock-gating domains. For MT8195 specifically, enforce exactly eight clocks and clock-names entries to satisfy its hardware requirements.

---
  .../devicetree/bindings/ufs/mediatek,ufs.yaml | 42 ++++++++++++++++---
  1 file changed, 36 insertions(+), 6 deletions(-)

Changes for v2:
- Remove duplicate minItems and maxItems as suggested in the review.
- Add a description of how the MT8195 hardware differs from earlier platforms.

Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +16 -6
index 20f341d25ebc..1dec54fb00f3 100644
@@ -9,21 +9,20 @@ title: Mediatek Universal Flash Storage (UFS) Controller
  maintainers:
    - Stanley Chu <[email protected]>
 
-allOf:
-  - $ref: ufs-common.yaml
-
  properties:
    compatible:
      enum:
        - mediatek,mt8183-ufshci
        - mediatek,mt8192-ufshci
+      - mediatek,mt8195-ufshci
 
    clocks:
-    maxItems: 1
+    minItems: 1
+    maxItems: 8
 
    clock-names:
-    items:
-      - const: ufs
+    minItems: 1
+    maxItems: 8
 
    phys:
      maxItems: 1
@@ -47,6 +46,37 @@ required:
 
  unevaluatedProperties: false
 
+allOf:
+  - $ref: ufs-common.yaml
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - mediatek,mt8195-ufshci
+    then:
+      properties:

Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml required: · line 69

Quoted 10 lines
Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +10 -0
+        clocks:
+          minItems: 8
+        clock-names:
+          items:
+            - const: ufs
+            - const: ufs_aes
+            - const: ufs_tick
+            - const: unipro_sysclk
+            - const: unipro_tick
+            - const: unipro_mp_bclk

The unipro mp_bclk really is the ufs-sap clock; besides, the standard has clocks for both TX and RX symbols - and also MT8195 (and also MT6991, MT8196, and others) UFS controller do have both TX and RX symbol clocks.

Besides, you're also missing the crypto clocks for UFS, which brings the count to 12 total clocks for MT8195.

Please, look at my old submission, which actually fixes the compatibles other than adding the right clocks for all UFS controllers in MediaTek platforms.

https://lore.kernel.org/all/[email protected]/

I want to take the occasion to remind everyone that my fixes were discarded because the MediaTek UFS driver maintainer wants to keep the low quality of the driver in favor of easier downstream porting - which is *not* in any way adhering to quality standards that the Linux community deserves.

Cheers,
Angelo

Quoted 13 lines
Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +10 -0
+            - const: ufs_tx_symbol
+            - const: ufs_mem_sub
+    else:
+      properties:
+        clocks:
+          maxItems: 1
+        clock-names:
+          items:
+            - const: ufs
+
  examples:
    - |
      #include <dt-bindings/clock/mt8183-clk.h>
maintainer
#6 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 80 lines · AngeloGioacchino Del Regno

Il 22/07/25 10:57, Macpaul Lin ha scritto:

Quoted 58 lines

Add MT8195 UFSHCI compatible string.
Relax the schema to allow between one to eight clocks/clock-names entries for all MediaTek UFS nodes. Legacy platforms may only need a few clocks, whereas newer devices such as the MT8195 require additional clock-gating domains. For MT8195 specifically, enforce exactly eight clocks and clock-names entries to satisfy its hardware requirements.

---
  .../devicetree/bindings/ufs/mediatek,ufs.yaml | 42 ++++++++++++++++---
  1 file changed, 36 insertions(+), 6 deletions(-)

Changes for v2:
- Remove duplicate minItems and maxItems as suggested in the review.
- Add a description of how the MT8195 hardware differs from earlier platforms.

Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +16 -6
index 20f341d25ebc..1dec54fb00f3 100644
@@ -9,21 +9,20 @@ title: Mediatek Universal Flash Storage (UFS) Controller
  maintainers:
    - Stanley Chu <[email protected]>
-allOf:
-  - $ref: ufs-common.yaml
-
  properties:
    compatible:
      enum:
        - mediatek,mt8183-ufshci
        - mediatek,mt8192-ufshci
+      - mediatek,mt8195-ufshci
    clocks:
-    maxItems: 1
+    minItems: 1
+    maxItems: 8
    clock-names:
-    items:
-      - const: ufs
+    minItems: 1
+    maxItems: 8
    phys:
      maxItems: 1
@@ -47,6 +46,37 @@ required:
  unevaluatedProperties: false
+allOf:
+  - $ref: ufs-common.yaml
+
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - mediatek,mt8195-ufshci
+    then:
+      properties:

Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml required: · line 67

Quoted 10 lines
Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml +10 -0
+        clocks:
+          minItems: 8
+        clock-names:
+          items:
+            - const: ufs
+            - const: ufs_aes
+            - const: ufs_tick
+            - const: unipro_sysclk
+            - const: unipro_tick
+            - const: unipro_mp_bclk

The unipro mp_bclk really is the ufs-sap clock; besides, the standard has clocks for both TX and RX symbols - and also MT8195 (and also MT6991, MT8196, and others) UFS controller do have both TX and RX symbol clocks.

Besides, you're also missing the crypto clocks for UFS, which brings the count to 12 total clocks for MT8195.

Please, look at my old submission, which actually fixes the compatibles other than adding the right clocks for all UFS controllers in MediaTek platforms.

Quoted 6 lines

https://lore.kernel.org/all/[email protected]/

I want to take the occasion to remind everyone that my fixes were discarded because the MediaTek UFS driver maintainer wants to keep the low quality of the driver in favor of easier downstream porting - which is *not* in any way adhering to quality standards that the Linux community deserves.

Sounds like we need a new maintainer then. They clearly don't understand that downstream doesn't exist.

Rob

#7 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

In reply to AngeloGioacchino Del Regno
Quoted 4 lines · AngeloGioacchino Del Regno

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

[snip...]

Quoted 17 lines

The unipro mp_bclk really is the ufs-sap clock; besides, the standard has clocks
for both TX and RX symbols - and also MT8195 (and also MT6991, MT8196, and others)
UFS controller do have both TX and RX symbol clocks.

Besides, you're also missing the crypto clocks for UFS, which brings the count to
12 total clocks for MT8195.

Please, look at my old submission, which actually fixes the compatibles other than
adding the right clocks for all UFS controllers in MediaTek platforms.

https://lore.kernel.org/all/[email protected]/

Quoted 8 lines

I want to take the occasion to remind everyone that my fixes were discarded because
the MediaTek UFS driver maintainer wants to keep the low quality of the driver in
favor of easier downstream porting - which is *not* in any way adhering to quality
standards that the Linux community deserves.

Oops, it is sad to hear that.
MediaTek will raise an internal discussion and hope we could get the more specific requirements back to you soon.

Quoted 2 lines

Cheers,
Angelo

#8 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Il 23/07/25 09:33, Macpaul Lin (林智斌) ha scritto:

Quoted 34 lines · AngeloGioacchino Del Regno
Quoted 4 lines

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

[snip...]

Quoted 17 lines

The unipro mp_bclk really is the ufs-sap clock; besides, the standard has clocks
for both TX and RX symbols - and also MT8195 (and also MT6991, MT8196, and others)
UFS controller do have both TX and RX symbol clocks.

Besides, you're also missing the crypto clocks for UFS, which brings the count to
12 total clocks for MT8195.

Please, look at my old submission, which actually fixes the compatibles other than
adding the right clocks for all UFS controllers in MediaTek platforms.

https://lore.kernel.org/all/[email protected]/

Quoted 8 lines

I want to take the occasion to remind everyone that my fixes were discarded because
the MediaTek UFS driver maintainer wants to keep the low quality of the driver in
favor of easier downstream porting - which is *not* in any way adhering to quality
standards that the Linux community deserves.

Quoted 4 lines

Oops, it is sad to hear that.
MediaTek will raise an internal discussion and hope we could get the more specific requirements back to you soon.

Thanks for that.

Cheers,
Angelo

#9 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

In reply to AngeloGioacchino Del Regno
Quoted 11 lines · AngeloGioacchino Del Regno

The unipro mp_bclk really is the ufs-sap clock; besides, the standard has clocks
for both TX and RX symbols - and also MT8195 (and also MT6991, MT8196, and others)
UFS controller do have both TX and RX symbol clocks.

Besides, you're also missing the crypto clocks for UFS, which brings the count to
12 total clocks for MT8195.

Quoted 7 lines

Please, look at my old submission, which actually fixes the compatibles other than
adding the right clocks for all UFS controllers in MediaTek platforms.

https://lore.kernel.org/all/[email protected]/

Hi Angelo,

The clock architecture may vary depending on the platform. These clock patch look good to me.

Quoted 10 lines

I want to take the occasion to remind everyone that my fixes were discarded because
the MediaTek UFS driver maintainer wants to keep the low quality of the driver in
favor of easier downstream porting - which is *not* in any way adhering to quality
standards that the Linux community deserves.

Cheers,
Angelo

I want to clarify that I am not opposing this in order to keep the low quality of the driver for the sake of easier downstream porting. My objection is purely due to the quality of this patch:
https://lore.kernel.org/all/[email protected]/
Originally, this could have been a simple matter handled by a single DTS setting,
but this patch requires checking a bunch of DTS voltage settings. This increases the complexity of the boot process, and I don't see any benefit from it.

Even though I think the other patches look good to me, I haven't seen any new ones uploaded.
I would like to reiterate that I welcome any upstream contributions related to Linux UFS on MediaTek platforms.

Thanks
Peter

#10 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 22 lines · Peter Wang (王信友)
Quoted 7 lines

Please, look at my old submission, which actually fixes the compatibles other than
adding the right clocks for all UFS controllers in MediaTek platforms.

https://lore.kernel.org/all/[email protected]/

Hi Angelo,

The clock architecture may vary depending on the platform. These clock patch look good to me.

Quoted 8 lines

I want to take the occasion to remind everyone that my fixes were discarded because
the MediaTek UFS driver maintainer wants to keep the low quality of the driver in
favor of easier downstream porting - which is *not* in any way adhering to quality
standards that the Linux community deserves.

Quoted 5 lines
Quoted 2 lines

Cheers,
Angelo

I want to clarify that I am not opposing this in order to keep the low quality of the driver for the sake of easier downstream porting.

You did.

You wrote very clearly here:
https://lore.kernel.org/all/[email protected]/

"In addition, it will require MediaTek to put in extra effort to migrate the kernel. "

Also you wrote:
"The role of MediaTek UFS maintainer is not suitable to be handed over to someone outside of MediaTek."

https://lore.kernel.org/all/[email protected]/

Holy molly, you really wrote this!

That's completely unacceptable. You don't understand how upstream development works and you push your downstream narrative which for us does not matter. You also object community led efforts, because you apparently want to control the upstream process.

That is red flag.

I think you should step down from maintainer position and find more suitable person, who is willing to work with the community, or rethink how upstream process works and understand that your downstream goals do not matter completely.

I will be watching closely this and if situation does not improve, I believe we should mark the driver orphaned until we find maintainer caring about community, not about corporate goals.

Best regards,
Krzysztof

#11 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 9 lines · Krzysztof Kozlowski

You did.

You wrote very clearly here:
https://lore.kernel.org/all/[email protected]/

"In addition, it will require MediaTek to put in extra effort to migrate the kernel. "

Hi Krzysztof Kozlowski,

The main reason for my objection was also clearly stated:
"removing these DTS settings will make what was originally a simple task more complicated."
I’m not sure if you are quoting only the "In addition" part to take it out of context?

Quoted 10 lines

Also you wrote:
"The role of MediaTek UFS maintainer is not suitable to be handed over
to someone outside of MediaTek."

https://lore.kernel.org/all/[email protected]/

Holy molly, you really wrote this!

"The role of MediaTek UFS maintainer is not suitable to be handed over to someone outside of MediaTek."
My main point is that MediaTek’s internal personnel certainly have a better understanding of the SoC architecture than external parties.
Wouldn’t it be more appropriate for maintainers to be internal staff?

Quoted 5 lines

That's completely unacceptable. You don't understand how upstream development works and you push your downstream narrative which for us does not matter. You also object community led efforts, because you apparently want to control the upstream process.

I don’t see how this relates to upstream/downstream? Aren’t you reading too much into this? My objection is purely because I don’t want to complicate a simple matter, not because I object to community-led efforts.
Please don’t misunderstand my intention.

Quoted 9 lines

That is red flag.

I think you should step down from maintainer position and find more suitable person, who is willing to work with the community, or rethink
how upstream process works and understand that your downstream goals do
not matter completely.

Quoted 6 lines

I will be watching closely this and if situation does not improve, I believe we should mark the driver orphaned until we find maintainer caring about community, not about corporate goals.

Best regards,
Krzysztof

Mediatek will add a few more maintainers internally,

Thanks
Peter

#12 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 11 lines · Peter Wang (王信友)
Quoted 9 lines · Krzysztof Kozlowski

You did.

You wrote very clearly here:
https://lore.kernel.org/all/[email protected]/

"In addition, it will require MediaTek to put in extra effort to migrate the kernel. "

Quoted 7 lines

Hi Krzysztof Kozlowski,

The main reason for my objection was also clearly stated:
"removing these DTS settings will make what was originally a simple task more complicated."
I’m not sure if you are quoting only the "In addition" part to take it out of context?

It is not out of context. It was the statement on its own.

Quoted 14 lines
Quoted 4 lines
Quoted 9 lines

Also you wrote:
"The role of MediaTek UFS maintainer is not suitable to be handed over
to someone outside of MediaTek."

https://lore.kernel.org/all/[email protected]/

Holy molly, you really wrote this!

Quoted 6 lines

"The role of MediaTek UFS maintainer is not suitable to be handed over to someone outside of MediaTek."
My main point is that MediaTek’s internal personnel certainly have a better understanding of the SoC architecture than external parties.
Wouldn’t it be more appropriate for maintainers to be internal staff?

You denied community to participate and now you twist the argument like you want Mediatek people to be involved. No one denied Mediatek to be maintainer.

It is you who denied community to join the maintainers.

This is not acceptable and you still do not understand why.

Quoted 9 lines
Quoted 8 lines

That's completely unacceptable. You don't understand how upstream development works and you push your downstream narrative which for us does not matter. You also object community led efforts, because you apparently want to control the upstream process.

Quoted 5 lines

I don’t see how this relates to upstream/downstream? Aren’t you reading too much into this? My objection is purely because I don’t want to complicate a simple matter, not because I object to community-led efforts.
Please don’t misunderstand my intention.

You could apologize and explain your mistakes, but instead you push same narrative.

Still a red flag. I will not accept such vendor-like behaviors, because they significantly harm the community.

I am very surprised that UFS maintainers did not object to it. This should be clearly ostracized.

Quoted 16 lines
Quoted 16 lines

That is red flag.

I think you should step down from maintainer position and find more suitable person, who is willing to work with the community, or rethink
how upstream process works and understand that your downstream goals do
not matter completely.

I will be watching closely this and if situation does not improve, I believe we should mark the driver orphaned until we find maintainer caring about community, not about corporate goals.

Quoted 5 lines
Quoted 2 lines

Best regards,
Krzysztof

Mediatek will add a few more maintainers internally,

Consider stepping down and choosing them if they better understand how upstream works.

As Rob wrote earlier:

"Sounds like we need a new maintainer then. They clearly don't understand that downstream doesn't exist."

Best regards,
Krzysztof

#13 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 11 lines · Krzysztof Kozlowski
Quoted 9 lines

Hi Krzysztof Kozlowski,

The main reason for my objection was also clearly stated:
"removing these DTS settings will make what was originally a simple task more complicated."
I’m not sure if you are quoting only the "In addition" part to take it out of context?

It is not out of context. It was the statement on its own.

Hi Krzysztof Kozlowski,

However, you haven’t addressed the main reason for my objection. "removing these DTS settings will make what was originally a simple task more complicated."

Quoted 11 lines

You denied community to participate and now you twist the argument like
you want Mediatek people to be involved. No one denied Mediatek to be maintainer.

It is you who denied community to join the maintainers.

This is not acceptable and you still do not understand why.

I think I understand your point now, you believe that opposing this patch means opposing community participation and support, right? But it’s clear that you haven’t carefully considered the main reason for my objection?

Quoted 6 lines

You could apologize and explain your mistakes, but instead you push same
narrative.

Quoted 8 lines

Still a red flag. I will not accept such vendor-like behaviors, because
they significantly harm the community.

I am very surprised that UFS maintainers did not object to it. This should be clearly ostracized.

Sorry, I still don’t quite understand why having people who know the SoC better maintain the SoC driver would be considered harmful to the community?
Also, I still don’t see the relation between opposing the unnecessary complication patch and the downstream? Can you tell me why I should accept code that I feel is unreasonable to be merged? Instead of repeatedly telling me that my objection is about downstream or something like that.

Quoted 6 lines

Consider stepping down and choosing them if they better understand how
upstream works.

Quoted 8 lines

As Rob wrote earlier:

"Sounds like we need a new maintainer then. They clearly don't understand that downstream doesn't exist."

Best regards,
Krzysztof

I must reiterate that I do not oppose patches that are beneficial to the community; I only object to patches that are not helpful.

Thanks
Peter

#14 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 11 lines · Peter Wang (王信友)
Quoted 10 lines · Krzysztof Kozlowski
Quoted 9 lines

Hi Krzysztof Kozlowski,

The main reason for my objection was also clearly stated:
"removing these DTS settings will make what was originally a simple task more complicated."
I’m not sure if you are quoting only the "In addition" part to take it out of context?

Quoted 7 lines
Quoted 1 line

It is not out of context. It was the statement on its own.

Hi Krzysztof Kozlowski,

However, you haven’t addressed the main reason for my objection. "removing these DTS settings will make what was originally a simple task more complicated."

You did not object in technical matter at all here:
https://lore.kernel.org/all/[email protected]/

Look at this patch.

You said nothing about actual change, except blocking the community maintainer. You did not raise any other concerns so what are you speaking about "other main concerns"?

Even if such existed, they did not matter, because YOU WROTE ONLY:

"The role of MediaTek UFS maintainer is not suitable to be handed over to someone outside of MediaTek."

This is what we discuss here.

Quoted 9 lines
Quoted 9 lines

You denied community to participate and now you twist the argument like
you want Mediatek people to be involved. No one denied Mediatek to be maintainer.

Quoted 6 lines
Quoted 3 lines

It is you who denied community to join the maintainers.

This is not acceptable and you still do not understand why.

I think I understand your point now, you believe that opposing this patch means opposing community participation and support, right?

Do you even read your own comments and where did you place them? Do you understand that we discuss emails, not some unsaid or other threads?

Look at this:

https://lore.kernel.org/all/[email protected]/

Quoted 2 lines

But it’s clear that you haven’t carefully considered the main reason for my objection?

Main reason for objection? What?

Quoted 13 lines
Quoted 13 lines

You could apologize and explain your mistakes, but instead you push same
narrative.

Still a red flag. I will not accept such vendor-like behaviors, because
they significantly harm the community.

Quoted 8 lines
Quoted 4 lines

I am very surprised that UFS maintainers did not object to it. This should be clearly ostracized.

Sorry, I still don’t quite understand why having people who know the SoC better maintain the SoC driver would be considered harmful to the community?

You are twisting the problem, like anyone denied you being the maintainer.

YOU DENIED OTHER PEOPLE!

I finish the discussion here, I am considering your explanations intentionally twisting the point thus I find it still harmful behavior.

Best regards,
Krzysztof

#15 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 45 lines · Krzysztof Kozlowski
Quoted 11 lines · Peter Wang (王信友)
Quoted 10 lines · Krzysztof Kozlowski
Quoted 9 lines

Hi Krzysztof Kozlowski,

The main reason for my objection was also clearly stated:
"removing these DTS settings will make what was originally a simple task more complicated."
I’m not sure if you are quoting only the "In addition" part to take it out of context?

Quoted 7 lines
Quoted 1 line

It is not out of context. It was the statement on its own.

Hi Krzysztof Kozlowski,

However, you haven’t addressed the main reason for my objection. "removing these DTS settings will make what was originally a simple task more complicated."

You did not object in technical matter at all here:
https://lore.kernel.org/all/[email protected]/

Look at this patch.

You said nothing about actual change, except blocking the community maintainer. You did not raise any other concerns so what are you speaking about "other main concerns"?

Even if such existed, they did not matter, because YOU WROTE ONLY:

"The role of MediaTek UFS maintainer is not suitable to be handed over
to someone outside of MediaTek."

This is what we discuss here.

Do you even read your own comments and where did you place them? Do you
understand that we discuss emails, not some unsaid or other threads?

Look at this:

Quoted 8 lines

https://lore.kernel.org/all/[email protected]/

Quoted 2 lines

But it’s clear that you haven’t carefully considered the main reason for my objection?

Main reason for objection? What?

Hi Krzysztof Kozlowski,

I think you misunderstood—these are different patches. This one only changes the maintainer. What I was referring to is another patch that removes parts of the DTS setting.
https://lore.kernel.org/all/[email protected]/

I don’t know who AngeloGioacchino is, so isn’t it reasonable for me to oppose directly changing the maintainer?
Or do you think everyone should know who AngeloGioacchino is and just accept this change?

Let’s put it this way: if a strager you don’t know suddenly comes to your home and says they’re now the maintainer of your house, would you be comfortable with that?

Quoted 5 lines

You are twisting the problem, like anyone denied you being the maintainer.

Quoted 8 lines

YOU DENIED OTHER PEOPLE!

I finish the discussion here, I am considering your explanations intentionally twisting the point thus I find it still harmful behavior.

Best regards,
Krzysztof

I think you’re the one twisting my words.
What I said was that I oppose people OUTSIDE of MediaTek becoming maintainers, not that I oppose other people in GENERAL. In fact, I also mentioned that other MediaTek maintainers would be joining.

Thanks
Peter

#16 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Il 20/10/25 12:46, Peter Wang (王信友) ha scritto:

Quoted 57 lines · Krzysztof Kozlowski
Quoted 45 lines · Peter Wang (王信友)
Quoted 11 lines · Krzysztof Kozlowski
Quoted 10 lines
Quoted 9 lines

Hi Krzysztof Kozlowski,

The main reason for my objection was also clearly stated:
"removing these DTS settings will make what was originally a simple task more complicated."
I’m not sure if you are quoting only the "In addition" part to take it out of context?

Quoted 7 lines
Quoted 1 line

It is not out of context. It was the statement on its own.

Hi Krzysztof Kozlowski,

However, you haven’t addressed the main reason for my objection. "removing these DTS settings will make what was originally a simple task more complicated."

You did not object in technical matter at all here:
https://lore.kernel.org/all/[email protected]/

Look at this patch.

You said nothing about actual change, except blocking the community maintainer. You did not raise any other concerns so what are you speaking about "other main concerns"?

Even if such existed, they did not matter, because YOU WROTE ONLY:

"The role of MediaTek UFS maintainer is not suitable to be handed over
to someone outside of MediaTek."

This is what we discuss here.

Do you even read your own comments and where did you place them? Do you
understand that we discuss emails, not some unsaid or other threads?

Look at this:

Quoted 8 lines

https://lore.kernel.org/all/[email protected]/

Quoted 2 lines

But it’s clear that you haven’t carefully considered the main reason for my objection?

Main reason for objection? What?

Hi Krzysztof Kozlowski,

Quoted 7 lines

I think you misunderstood—these are different patches. This one only changes the maintainer. What I was referring to is another patch that removes parts of the DTS setting.
https://lore.kernel.org/all/[email protected]/

I don’t know who AngeloGioacchino is,

Sorry Peter, but a 10 seconds research on your side would have made you aware of who I am.

https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/MAINTAINERS?h=v6.6#n2346

...then a 60 seconds research would reveal way more than just that about me, and also your colleagues know me quite a bit :-)

Besides, you don't really need to know who somebody is to make an upstream review:
this is a community, and a good patch may come from old and recognized contributors as much as from new ones sending their first patch upstream.

Quoted 6 lines

so isn’t it reasonable for
me to oppose directly changing the maintainer?
Or do you think everyone should know who AngeloGioacchino is and just accept this change?

Quoted 16 lines

Let’s put it this way: if a strager you don’t know suddenly comes to your home and says they’re now the maintainer of your house, would you be comfortable with that?

Quoted 10 lines

You are twisting the problem, like anyone denied you being the maintainer.

YOU DENIED OTHER PEOPLE!

I finish the discussion here, I am considering your explanations intentionally twisting the point thus I find it still harmful behavior.

Krzysztof, many thanks for taking time to defend the community.

Regards,
Angelo

Quoted 15 lines
Quoted 3 lines

Best regards,
Krzysztof

I think you’re the one twisting my words.
What I said was that I oppose people OUTSIDE of MediaTek becoming maintainers, not that I oppose other people in GENERAL. In fact, I also mentioned that other MediaTek maintainers would be joining.

Thanks
Peter

#17 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 11 lines · AngeloGioacchino Del Regno

Sorry Peter, but a 10 seconds research on your side would have made you aware of
who I am.

https://urldefense.com/v3/__https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/MAINTAINERS?h=v6.6*n2346__;Iw!!CTRNKA9wMg0ARbw!j46OdEicZdlCQk4DoUc9bC32inkXJnyn4mTFypGGCK_bH1CJlW4Et75egmmKPBOgf3SHcPJTdX_vt7w_hp82-lkypdUcP32s$

...then a 60 seconds research would reveal way more than just that about me,
and also your colleagues know me quite a bit :-)

Hi AngeloGioacchino,

Yes, however, I was not aware of this when I responded last year. There have been internal discussions since then, and we have decided that new MediaTek maintainers will be added. This position is still not appropriate for someone external to MediaTek. I hope for your understanding.

Quoted 9 lines

Besides, you don't really need to know who somebody is to make an upstream review:
this is a community, and a good patch may come from old and recognized contributors
as much as from new ones sending their first patch upstream.

I agree with your point, perhaps we simply have different viewpoints, but our goal is the same—to make this community better. If there are any shortcomings in our ideas, we can discuss them and focus on the patch itself, rather than resorting to blanket criticisms such as labeling downstream as low quality or upstream as inherently superior.
Such remarks do not contribute to the improvement of the community.

Thanks
Peter

#18 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

In reply to Peter Wang (王信友)
Quoted 11 lines · Peter Wang (王信友)
Quoted 11 lines

Consider stepping down and choosing them if they better understand how
upstream works.

As Rob wrote earlier:

"Sounds like we need a new maintainer then. They clearly don't understand that downstream doesn't exist."

Quoted 6 lines
Quoted 2 lines

Best regards,
Krzysztof

I must reiterate that I do not oppose patches that are beneficial to the community; I only object to patches that are not helpful.

Let's quote you again:

"*In addition*, it will require MediaTek to put in extra effort to migrate the kernel. "

This is ADDITIONAL argument you used. This is what you wrote, this is what you claimed to be ADDITIONAL argument.

In your opinion ADDITIONAL argument is downstream and you still do not understand why such argument is instant NAK for you as reviewer.

Best regards,
Krzysztof

#19 lore raw

Re: [PATCH v2 3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes

Quoted 9 lines · Krzysztof Kozlowski

Let's quote you again:

"*In addition*, it will require MediaTek to put in extra effort to migrate the kernel. "

This is ADDITIONAL argument you used. This is what you wrote, this is what you claimed to be ADDITIONAL argument.

Quoted 7 lines

In your opinion ADDITIONAL argument is downstream and you still do not
understand why such argument is instant NAK for you as reviewer.

Best regards,
Krzysztof

Hi Krzysztof Kozlowski,

I acknowledge that this was indeed a lack of consideration on my part, and I must apologize for making such statements. Nevertheless, I have also clarified that this was not the main reason for my objection, the actual reason is "removing these DTS settings will make what was originally a simple task more complicated."
And then, when I asked you about the main reason and its relation to downstream, you just kept repeating “ADDITIONAL.” Why not address the issue directly and tell me what the benefit is of making the code more complicated?

Thanks
Peter

#20 lore raw

PATCH v2 4/4 arm64: dts: mediatek: mt8195: add UFSHCI node

In reply to Macpaul Lin

From: Rice Lee <[email protected]>

Add a UFS host controller interface (UFSHCI) node to mt8195.dtsi. Introduce the 'mediatek,ufs-disable-mcq' property to allow disabling Multiple Circular Queue (MCQ) support.

---
 arch/arm64/boot/dts/mediatek/mt8195.dtsi | 25 ++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

Changes for v2:
- No change.

arch/arm64/boot/dts/mediatek/mt8195.dtsi +25 -0
index dd065b1bf94a..8877953ce292 100644
@@ -1430,6 +1430,31 @@ mmc2: mmc@11250000 {
 			status = "disabled";
 		};
 
+		ufshci: ufshci@11270000 {
+			compatible = "mediatek,mt8195-ufshci";
+			reg = <0 0x11270000 0 0x2300>;
+			interrupts = <GIC_SPI 137 IRQ_TYPE_LEVEL_HIGH 0>;
+			phys = <&ufsphy>;
+			clocks = <&infracfg_ao CLK_INFRA_AO_AES_UFSFDE>,
+				 <&infracfg_ao CLK_INFRA_AO_AES>,
+				 <&infracfg_ao CLK_INFRA_AO_UFS_TICK>,
+				 <&infracfg_ao CLK_INFRA_AO_UNIPRO_SYS>,
+				 <&infracfg_ao CLK_INFRA_AO_UNIPRO_TICK>,
+				 <&infracfg_ao CLK_INFRA_AO_UFS_MP_SAP_B>,
+				 <&infracfg_ao CLK_INFRA_AO_UFS_TX_SYMBOL>,
+				 <&infracfg_ao CLK_INFRA_AO_PERI_UFS_MEM_SUB>;
+			clock-names = "ufs", "ufs_aes", "ufs_tick",
+					"unipro_sysclk", "unipro_tick",
+					"unipro_mp_bclk", "ufs_tx_symbol",
+					"ufs_mem_sub";
+			freq-table-hz = <0 0>, <0 0>, <0 0>,
+					<0 0>, <0 0>, <0 0>,
+					<0 0>, <0 0>;
+
+			mediatek,ufs-disable-mcq;
+			status = "disabled";
+		};
+
 		lvts_mcu: thermal-sensor@11278000 {
 			compatible = "mediatek,mt8195-lvts-mcu";
 			reg = <0 0x11278000 0 0x1000>;
Signature · 1 line
2.45.2
maintainer Reviewed-by on v2 4/4
#21 lore raw

Re: [PATCH v2 4/4] arm64: dts: mediatek: mt8195: add UFSHCI node

Quoted 61 lines · Macpaul Lin

From: Rice Lee <[email protected]>

Add a UFS host controller interface (UFSHCI) node to mt8195.dtsi. Introduce the 'mediatek,ufs-disable-mcq' property to allow disabling Multiple Circular Queue (MCQ) support.

---
 arch/arm64/boot/dts/mediatek/mt8195.dtsi | 25
++++++++++++++++++++++++

 1 file changed, 25 insertions(+)

Changes for v2:
 - No change.

arch/arm64/boot/dts/mediatek/mt8195.dtsi +0 -0
index dd065b1bf94a..8877953ce292 100644
@@ -1430,6 +1430,31 @@ mmc2: mmc@11250000 {
  			status = "disabled";
  		};

arch/arm64/boot/dts/mediatek/mt8195.dtsi +4 -0
+		ufshci: ufshci@11270000 {
+			compatible = "mediatek,mt8195-ufshci";
+			reg = <0 0x11270000 0 0x2300>;
+			interrupts = <GIC_SPI 137

IRQ_TYPE_LEVEL_HIGH 0>;

arch/arm64/boot/dts/mediatek/mt8195.dtsi +2 -0
+			phys = <&ufsphy>;
+			clocks = <&infracfg_ao

CLK_INFRA_AO_AES_UFSFDE>,

arch/arm64/boot/dts/mediatek/mt8195.dtsi +2 -0
+				 <&infracfg_ao CLK_INFRA_AO_AES>,
+				 <&infracfg_ao

CLK_INFRA_AO_UFS_TICK>,

arch/arm64/boot/dts/mediatek/mt8195.dtsi +1 -0
+				 <&infracfg_ao

CLK_INFRA_AO_UNIPRO_SYS>,

arch/arm64/boot/dts/mediatek/mt8195.dtsi +1 -0
+				 <&infracfg_ao

CLK_INFRA_AO_UNIPRO_TICK>,

arch/arm64/boot/dts/mediatek/mt8195.dtsi +1 -0
+				 <&infracfg_ao

CLK_INFRA_AO_UFS_MP_SAP_B>,

arch/arm64/boot/dts/mediatek/mt8195.dtsi +1 -0
+				 <&infracfg_ao

CLK_INFRA_AO_UFS_TX_SYMBOL>,

arch/arm64/boot/dts/mediatek/mt8195.dtsi +1 -0
+				 <&infracfg_ao

CLK_INFRA_AO_PERI_UFS_MEM_SUB>;

arch/arm64/boot/dts/mediatek/mt8195.dtsi +2 -0
+			clock-names = "ufs", "ufs_aes", "ufs_tick",
+					"unipro_sysclk",

"unipro_tick",

arch/arm64/boot/dts/mediatek/mt8195.dtsi +1 -0
+					"unipro_mp_bclk",

"ufs_tx_symbol",

arch/arm64/boot/dts/mediatek/mt8195.dtsi +9 -0
+					"ufs_mem_sub";
+			freq-table-hz = <0 0>, <0 0>, <0 0>,
+					<0 0>, <0 0>, <0 0>,
+					<0 0>, <0 0>;
+
+			mediatek,ufs-disable-mcq;
+			status = "disabled";
+		};
+
Quoted 3 lines

  lvts_mcu: thermal-sensor@11278000 { compatible = "mediatek,mt8195-lvts-mcu"; reg = <0 0x11278000 0 0x1000>;

maintainer Reviewed-by on v2 1/4
#22 lore raw
In reply to Macpaul Lin
Quoted 26 lines · Macpaul Lin

Add "mediatek,mt8195-ufshci" to the of_device_id table to enable support for MediaTek MT8195/MT8395 UFS host controller. This matches the
device node entry in the MT8195/MT8395 device tree and allows proper driver
binding.

---
 drivers/ufs/host/ufs-mediatek.c | 1 +

 1 file changed, 1 insertion(+)

Changes vor f2:
 - No change.

drivers/ufs/host/ufs- renamed from drivers/ufs/host/ufs-mediatek.c +0 -0

mediatek.c
index 182f58d0c9db..e1dbf0231c5e 100644

drivers/ufs/host/ufs-mediatek.c +0 -0
@@ -50,6 +50,7 @@ static const struct ufs_dev_quirk
 ufs_mtk_dev_fixups[] = {


 static const struct of_device_id ufs_mtk_of_match[] = { { .compatible = "mediatek,mt8183-ufshci" },

drivers/ufs/host/ufs-mediatek.c +1 -0
+	{ .compatible = "mediatek,mt8195-ufshci" },
Quoted 3 lines

  {},
 };
 MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);

#23 lore raw
In reply to Macpaul Lin
Quoted 6 lines · Macpaul Lin

Add "mediatek,mt8195-ufshci" to the of_device_id table to enable support for MediaTek MT8195/MT8395 UFS host controller. This matches the device node entry in the MT8195/MT8395 device tree and allows proper driver binding.

Applied to 6.17/scsi-queue, thanks!

[1/4] scsi: ufs: ufs-mediatek: Add UFS host support for MT8195 SoC

      https://git.kernel.org/mkp/scsi/c/6f1fd3e0279f
[2/4] dt-bindings: ufs: mediatek,ufs: add ufs-disable-mcq flag for UFS host
      https://git.kernel.org/mkp/scsi/c/794ff7a0a6e7
[3/4] dt-bindings: ufs: mediatek,ufs: add MT8195 compatible and update clock nodes
      https://git.kernel.org/mkp/scsi/c/d01cfeac89e9
[4/4] arm64: dts: mediatek: mt8195: add UFSHCI node
      https://git.kernel.org/mkp/scsi/c/a28f98103890
Signature · 1 line
Martin K. Petersen	Oracle Linux Engineering
#24 lore raw
In reply to Macpaul Lin
Quoted 4 lines · Macpaul Lin

Add "mediatek,mt8195-ufshci" to the of_device_id table to enable support for MediaTek MT8195/MT8395 UFS host controller. This matches the device node entry in the MT8195/MT8395 device tree and allows proper driver binding.

Please always include a cover letter when posting a patch series. The cover letter generated by git includes helpful information like a diffstat.

Thanks,

Bart.