* [PATCH v10 01/26] dt-bindings: phy: Add mediatek,mt8196-ufsphy variant
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
@ 2026-07-30 9:48 ` Louis-Alexis Eyraud
2026-07-30 9:48 ` [PATCH v10 02/26] dt-bindings: ufs: mediatek,ufs: Complete the binding Louis-Alexis Eyraud
` (24 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:48 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Conor Dooley,
Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The MediaTek MT8196 SoC includes an M-PHY compatible with the already
existing mt8183 binding.
However, one omission from the original binding was that all of these
variants may have an optional reset.
Add the new compatible, and also the resets property, with an example.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Acked-by: Conor Dooley <conor.dooley@microchip.com>
Acked-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
.../devicetree/bindings/phy/mediatek,ufs-phy.yaml | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/Documentation/devicetree/bindings/phy/mediatek,ufs-phy.yaml b/Documentation/devicetree/bindings/phy/mediatek,ufs-phy.yaml
index 6e2edd43fc2a..ee71dfa4e0c0 100644
--- a/Documentation/devicetree/bindings/phy/mediatek,ufs-phy.yaml
+++ b/Documentation/devicetree/bindings/phy/mediatek,ufs-phy.yaml
@@ -27,6 +27,7 @@ properties:
- items:
- enum:
- mediatek,mt8195-ufsphy
+ - mediatek,mt8196-ufsphy
- const: mediatek,mt8183-ufsphy
- const: mediatek,mt8183-ufsphy
@@ -43,6 +44,10 @@ properties:
- const: unipro
- const: mp
+ resets:
+ items:
+ - description: Optional UFS M-PHY reset.
+
"#phy-cells":
const: 0
@@ -66,5 +71,16 @@ examples:
clock-names = "unipro", "mp";
#phy-cells = <0>;
};
+ - |
+ #include <dt-bindings/reset/mediatek,mt8196-resets.h>
+ ufs-phy@16800000 {
+ compatible = "mediatek,mt8196-ufsphy", "mediatek,mt8183-ufsphy";
+ reg = <0x16800000 0x10000>;
+ clocks = <&ufs_ao_clk 3>,
+ <&ufs_ao_clk 5>;
+ clock-names = "unipro", "mp";
+ resets = <&ufs_ao_clk MT8196_UFSAO_RST0_UFS_MPHY>;
+ #phy-cells = <0>;
+ };
...
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 02/26] dt-bindings: ufs: mediatek,ufs: Complete the binding
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
2026-07-30 9:48 ` [PATCH v10 01/26] dt-bindings: phy: Add mediatek,mt8196-ufsphy variant Louis-Alexis Eyraud
@ 2026-07-30 9:48 ` Louis-Alexis Eyraud
2026-07-30 9:57 ` sashiko-bot
2026-07-30 9:48 ` [PATCH v10 03/26] dt-bindings: ufs: mediatek,ufs: Add mt8196 variant Louis-Alexis Eyraud
` (23 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:48 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
As it stands, the mediatek,ufs.yaml binding is startlingly incomplete.
Its one example, which is the only real "user" of this binding in
mainline, uses the deprecated freq-table-hz property.
The resets, of which there are three optional ones, are completely
absent.
The clock description for MT8195 is incomplete, as is the one for
MT8192. It's not known if the one clock binding for MT8183 is even
correct, but I do not have access to the necessary code and
documentation to find this out myself.
The power supply situation is not much better; the binding describes one
required power supply, but it's the UFS card supply, not any of the
supplies feeding the controller silicon.
No second example is present in the binding, making verification
difficult.
Disallow freq-table-hz and move to operating-points-v2. It's fine to
break compatibility here, as the binding is currently unused and would
be impossible to correctly use in its current state.
Add the three resets and the corresponding reset-names property. These
resets appear to be optional, i.e. not required for the functioning of
the device.
Move the list of clock names out of the if condition, and expand it for
the confirmed clocks I could find by cross-referencing several clock
drivers. For MT8195, increase the minimum number of clocks to include
the rx_symbol ones, as they're internal to the SoC and should always
be present, and should therefore not be omitted.
MT8192 gets to have at least 3 clocks, as these were the ones I could
quickly confirm from a glance at various trees. I can't say this was an
exhaustive search though, but it's better than the current situation.
Properly document all supplies, with which pin name on the SoCs they
supply. Complete the example with them.
Also add a MT8195 example to the binding, using supply labels that I am
pretty sure would be the right ones for e.g. the Radxa NIO 12L.
Finally, remove the 'ufs_' prefix from all clock names containing it
and rename 'ufs' clock to 'main'.
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
.../devicetree/bindings/ufs/mediatek,ufs.yaml | 115 +++++++++++++++++----
1 file changed, 96 insertions(+), 19 deletions(-)
diff --git a/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml b/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
index 15c347f5e660..e7e360c59ba5 100644
--- a/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
+++ b/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
@@ -19,11 +19,25 @@ properties:
clocks:
minItems: 1
- maxItems: 8
+ maxItems: 10
clock-names:
minItems: 1
- maxItems: 8
+ items:
+ - const: main
+ - const: aes
+ - const: tick
+ - const: unipro_sysclk
+ - const: unipro_tick
+ - const: unipro_mp_bclk
+ - const: tx_symbol
+ - const: mem_sub
+ - const: rx_symbol0
+ - const: rx_symbol1
+
+ operating-points-v2: true
+
+ freq-table-hz: false
phys:
maxItems: 1
@@ -31,8 +45,36 @@ properties:
reg:
maxItems: 1
+ resets:
+ items:
+ - description: reset for the UniPro layer
+ - description: reset for the cryptography engine
+ - description: reset for the host controller
+
+ reset-names:
+ items:
+ - const: unipro
+ - const: crypto
+ - const: hci
+
+ avdd09-supply:
+ description: Phandle to the 0.9V supply powering the AVDD09_UFS pin
+
+ avdd12-supply:
+ description: Phandle to the 1.2V supply powering the AVDD12_UFS pin
+
+ avdd12-ckbuf-supply:
+ description: Phandle to the 1.2V supply powering the AVDD12_CKBUF_UFS pin
+
+ avdd18-supply:
+ description: Phandle to the 1.8V supply powering the AVDD18_UFS pin
+
vcc-supply: true
+ vccq-supply: true
+
+ vccq2-supply: true
+
mediatek,ufs-disable-mcq:
$ref: /schemas/types.yaml#/definitions/flag
description: The mask to disable MCQ (Multi-Circular Queue) for UFS host.
@@ -54,29 +96,41 @@ allOf:
properties:
compatible:
contains:
- enum:
- - mediatek,mt8195-ufshci
+ const: mediatek,mt8183-ufshci
then:
properties:
clocks:
- minItems: 8
+ maxItems: 1
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:
+ - const: main
+ avdd12-ckbuf-supply: false
+ - if:
+ properties:
+ compatible:
+ contains:
+ const: mediatek,mt8192-ufshci
+ then:
properties:
clocks:
- maxItems: 1
+ minItems: 3
+ maxItems: 3
+ clocks-names:
+ minItems: 3
+ maxItems: 3
+ avdd09-supply: false
+ - if:
+ properties:
+ compatible:
+ contains:
+ const: mediatek,mt8195-ufshci
+ then:
+ properties:
+ clocks:
+ minItems: 10
clock-names:
- items:
- - const: ufs
+ minItems: 10
+ avdd09-supply: false
examples:
- |
@@ -94,9 +148,32 @@ examples:
phys = <&ufsphy>;
clocks = <&infracfg_ao CLK_INFRA_UFS>;
- clock-names = "ufs";
- freq-table-hz = <0 0>;
+ clock-names = "main";
vcc-supply = <&mt_pmic_vemc_ldo_reg>;
};
};
+ - |
+ ufshci@11270000 {
+ compatible = "mediatek,mt8195-ufshci";
+ reg = <0x11270000 0x2300>;
+ interrupts = <GIC_SPI 137 IRQ_TYPE_LEVEL_HIGH>;
+ phys = <&ufsphy>;
+ clocks = <&infracfg_ao 63>, <&infracfg_ao 64>, <&infracfg_ao 65>,
+ <&infracfg_ao 54>, <&infracfg_ao 55>, <&infracfg_ao 56>,
+ <&infracfg_ao 90>, <&infracfg_ao 93>, <&topckgen 212>,
+ <&topckgen 215>;
+ clock-names = "main", "aes", "tick", "unipro_sysclk", "unipro_tick",
+ "unipro_mp_bclk", "tx_symbol", "mem_sub", "rx_symbol0",
+ "rx_symbol1";
+
+ operating-points-v2 = <&ufs_opp_table>;
+
+ avdd12-supply = <&mt6359_vrf12_ldo_reg>;
+ avdd12-ckbuf-supply = <&mt6359_vbbck_ldo_reg>;
+ avdd18-supply = <&mt6359_vio18_ldo_reg>;
+ vcc-supply = <&mt6359_vemc_1_ldo_reg>;
+ vccq2-supply = <&mt6359_vufs_ldo_reg>;
+
+ mediatek,ufs-disable-mcq;
+ };
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 02/26] dt-bindings: ufs: mediatek,ufs: Complete the binding
2026-07-30 9:48 ` [PATCH v10 02/26] dt-bindings: ufs: mediatek,ufs: Complete the binding Louis-Alexis Eyraud
@ 2026-07-30 9:57 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 9:57 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: conor+dt, neil.armstrong, devicetree, linux-phy, olteanv, robh,
vkoul, linux-scsi
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] The property `clock-names` is misspelled as `clocks-names` in the MT8192 conditional block, leaving the actual property unconstrained.
--
--- Patch [2]: [PATCH v10 02/26] dt-bindings: ufs: mediatek,ufs: Complete the binding ---
commit 515312405d3e2237cad578f6e9bc085ee3772431
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
dt-bindings: ufs: mediatek,ufs: Complete the binding
This commit updates the MediaTek UFS binding to replace deprecated frequency
tables with operating-points-v2 and adds missing optional resets. It also
corrects the clock definitions for different SoC variants and documents the
required power supplies.
> diff --git a/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml b/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
> index 15c347f5e6604..e7e360c59ba59 100644
> --- a/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
> +++ b/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
[ ... ]
> @@ -54,29 +96,41 @@ allOf:
[ ... ]
> + - if:
> + properties:
> + compatible:
> + contains:
> + const: mediatek,mt8192-ufshci
> + then:
> properties:
> clocks:
> - maxItems: 1
> + minItems: 3
> + maxItems: 3
> + clocks-names:
^^^^^^^^^^^^
[Severity: Low]
Is there a typo in this property name? It appears to say clocks-names instead
of clock-names. This might leave the actual clock-names property unconstrained
for the MT8192 hardware variant and cause schema validation to miss incorrect
device tree configurations.
> + minItems: 3
> + maxItems: 3
> + avdd09-supply: false
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=2
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 03/26] dt-bindings: ufs: mediatek,ufs: Add mt8196 variant
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
2026-07-30 9:48 ` [PATCH v10 01/26] dt-bindings: phy: Add mediatek,mt8196-ufsphy variant Louis-Alexis Eyraud
2026-07-30 9:48 ` [PATCH v10 02/26] dt-bindings: ufs: mediatek,ufs: Complete the binding Louis-Alexis Eyraud
@ 2026-07-30 9:48 ` Louis-Alexis Eyraud
2026-07-30 9:59 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 04/26] scsi: ufs: mediatek: Move MTK_SIP_UFS_CONTROL to mtk_sip_svc.h Louis-Alexis Eyraud
` (22 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:48 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The MediaTek MT8196 SoC's UFS controller uses a different set of supplies.
It is therefore not compatible with the MT8195.
While it does have a AVDD09_UFS_1 pin in addition to the AVDD09_UFS pin,
it appears that these two pins are commoned together, as the board
schematic I have access to uses the same supply for both, and the
downstream driver does not distinguish between the two supplies either.
Add a compatible for it, and modify the binding correspondingly.
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
.../devicetree/bindings/ufs/mediatek,ufs.yaml | 45 ++++++++++++++++++++++
1 file changed, 45 insertions(+)
diff --git a/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml b/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
index e7e360c59ba5..70164e4e54c5 100644
--- a/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
+++ b/Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml
@@ -16,6 +16,7 @@ properties:
- mediatek,mt8183-ufshci
- mediatek,mt8192-ufshci
- mediatek,mt8195-ufshci
+ - mediatek,mt8196-ufshci
clocks:
minItems: 1
@@ -131,6 +132,20 @@ allOf:
clock-names:
minItems: 10
avdd09-supply: false
+ - if:
+ properties:
+ compatible:
+ contains:
+ const: mediatek,mt8196-ufshci
+ then:
+ properties:
+ clocks:
+ minItems: 10
+ clock-names:
+ minItems: 10
+ avdd18-supply: false
+ required:
+ - operating-points-v2
examples:
- |
@@ -177,3 +192,33 @@ examples:
mediatek,ufs-disable-mcq;
};
+ - |
+ #include <dt-bindings/reset/mediatek,mt8196-resets.h>
+ #include <dt-bindings/interrupt-controller/arm-gic.h>
+
+ ufshci@16810000 {
+ compatible = "mediatek,mt8196-ufshci";
+ reg = <0x16810000 0x2a00>;
+ interrupts = <GIC_SPI 320 IRQ_TYPE_LEVEL_HIGH>;
+
+ clocks = <&ufs_ao_clk 6>, <&ufs_ao_clk 7>, <&clk26m>, <&ufs_ao_clk 3>,
+ <&clk26m>, <&ufs_ao_clk 4>, <&ufs_ao_clk 0>,
+ <&topckgen 7>, <&ufs_ao_clk 1>, <&ufs_ao_clk 2>;
+ clock-names = "main", "aes", "tick", "unipro_sysclk",
+ "unipro_tick", "unipro_mp_bclk", "tx_symbol",
+ "mem_sub", "rx_symbol0", "rx_symbol1";
+
+ operating-points-v2 = <&ufs_opp_table>;
+
+ phys = <&ufsphy>;
+
+ avdd09-supply = <&mt6363_vsram_modem>;
+ vcc-supply = <&mt6363_vemc>;
+ vccq-supply = <&mt6363_vufs12>;
+
+ resets = <&ufs_ao_clk MT8196_UFSAO_RST1_UFS_UNIPRO>,
+ <&ufs_ao_clk MT8196_UFSAO_RST1_UFS_CRYPTO>,
+ <&ufs_ao_clk MT8196_UFSAO_RST1_UFSHCI>;
+ reset-names = "unipro", "crypto", "hci";
+ mediatek,ufs-disable-mcq;
+ };
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 03/26] dt-bindings: ufs: mediatek,ufs: Add mt8196 variant
2026-07-30 9:48 ` [PATCH v10 03/26] dt-bindings: ufs: mediatek,ufs: Add mt8196 variant Louis-Alexis Eyraud
@ 2026-07-30 9:59 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 9:59 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: devicetree, linux-phy, conor+dt, neil.armstrong, robh, vkoul,
olteanv, linux-scsi
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Medium] A typo in the `mediatek,mt8192-ufshci` conditional block uses `clocks-names` instead of `clock-names`, bypassing array length validation.
- [Low] The binding omits a supply property for the `AVDD09_UFS_1` pin, tying it to `AVDD09_UFS`, which violates the principle that Device Tree describes the SoC hardware rather than a specific board's power routing.
--
--- Patch [3]: [PATCH v10 03/26] dt-bindings: ufs: mediatek,ufs: Add mt8196 variant ---
commit e0eeeaec009272d40ddbc2fee8902782c15e776c
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
dt-bindings: ufs: mediatek,ufs: Add mt8196 variant
The MT8196 SoC's UFS controller uses a different set of supplies, making it
incompatible with MT8195. This commit adds a compatible string for mt8196
and updates the binding. It also notes that the AVDD09_UFS and AVDD09_UFS_1
pins are commoned together on the referenced board.
> While it does have a AVDD09_UFS_1 pin in addition to the AVDD09_UFS pin,
> it appears that these two pins are commoned together, as the board
> schematic I have access to uses the same supply for both, and the
> downstream driver does not distinguish between the two supplies either.
[Severity: Low]
Does omitting a separate supply property for the AVDD09_UFS_1 pin violate the
principle that Device Tree bindings should describe the SoC hardware rather
than a specific board's power routing?
If a future board design supplies the AVDD09_UFS and AVDD09_UFS_1 pins from
separate voltage regulators, the current binding wouldn't be able to accurately
describe that hardware.
While reviewing the file touched by this patch, I noticed a regression in the
conditional block for the mt8192-ufshci compatible string:
Documentation/devicetree/bindings/ufs/mediatek,ufs.yaml:allOf
> - if:
> properties:
> compatible:
> contains:
> const: mediatek,mt8192-ufshci
> then:
> properties:
> clocks:
> minItems: 3
> maxItems: 3
> clocks-names:
> minItems: 3
> maxItems: 3
[Severity: Medium]
Is the "clocks-names" property a typo here? It seems like this should be
"clock-names".
Because of the typo, the length of the clock-names array is not properly
constrained for mt8192-ufshci. An invalid device tree with 3 clocks but a
mismatched number of clock-names could silently pass dt_binding_check validation.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=3
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 04/26] scsi: ufs: mediatek: Move MTK_SIP_UFS_CONTROL to mtk_sip_svc.h
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (2 preceding siblings ...)
2026-07-30 9:48 ` [PATCH v10 03/26] dt-bindings: ufs: mediatek,ufs: Add mt8196 variant Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 05/26] phy: mediatek: ufs: Add support for resets Louis-Alexis Eyraud
` (21 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
SMC commands used by multiple drivers need to live in a shared header
file somewhere to avoid code duplication. In order to rework the MPHY
reset control to be in the phy-mtk-ufs.c driver, both ufs-mediatek and
the phy driver need access to this command.
Move it to mtk_sip_svc.h, where other such command definitions already
live.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek-sip.h | 1 -
include/linux/soc/mediatek/mtk_sip_svc.h | 3 +++
2 files changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/ufs/host/ufs-mediatek-sip.h b/drivers/ufs/host/ufs-mediatek-sip.h
index 7d17aedf6fb8..d627dfb4a766 100644
--- a/drivers/ufs/host/ufs-mediatek-sip.h
+++ b/drivers/ufs/host/ufs-mediatek-sip.h
@@ -11,7 +11,6 @@
/*
* SiP (Slicon Partner) commands
*/
-#define MTK_SIP_UFS_CONTROL MTK_SIP_SMC_CMD(0x276)
#define UFS_MTK_SIP_VA09_PWR_CTRL BIT(0)
#define UFS_MTK_SIP_DEVICE_RESET BIT(1)
#define UFS_MTK_SIP_CRYPTO_CTRL BIT(2)
diff --git a/include/linux/soc/mediatek/mtk_sip_svc.h b/include/linux/soc/mediatek/mtk_sip_svc.h
index 6c95a29b79fa..36ab3153fe44 100644
--- a/include/linux/soc/mediatek/mtk_sip_svc.h
+++ b/include/linux/soc/mediatek/mtk_sip_svc.h
@@ -22,6 +22,9 @@
ARM_SMCCC_CALL_VAL(ARM_SMCCC_FAST_CALL, MTK_SIP_SMC_CONVENTION, \
ARM_SMCCC_OWNER_SIP, fn_id)
+/* UFS related SMC call */
+#define MTK_SIP_UFS_CONTROL MTK_SIP_SMC_CMD(0x276)
+
/* Modem related SMC call */
#define MTK_SIP_KERNEL_CCCI_CONTROL MTK_SIP_SMC_CMD(0x505)
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 05/26] phy: mediatek: ufs: Add support for resets
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (3 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 04/26] scsi: ufs: mediatek: Move MTK_SIP_UFS_CONTROL to mtk_sip_svc.h Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 06/26] scsi: ufs: mediatek: Rework resets Louis-Alexis Eyraud
` (20 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The MediaTek UFS PHY supports PHY resets. Until now, they've been
implemented in the UFS host driver. Since they were never documented in
the UFS HCI node's DT bindings, and no mainline DT uses it, it's fine if
it's moved to the correct location, which is the PHY driver.
Implement the MPHY reset logic in this driver and expose it through the
phy subsystem's reset op. The reset itself is optional, as judging by
other mainline devices that use this hardware, it's not required for the
device to function.
If no reset is present, the reset op returns -EOPNOTSUPP, which means
that the ufshci driver can detect it's present and not double sleep in
its own reset function, where it will call the phy reset.
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Acked-by: Vinod Koul <vkoul@kernel.org>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/phy/mediatek/phy-mtk-ufs.c | 71 ++++++++++++++++++++++++++++++++++++++
1 file changed, 71 insertions(+)
diff --git a/drivers/phy/mediatek/phy-mtk-ufs.c b/drivers/phy/mediatek/phy-mtk-ufs.c
index fc19e0fa8ed5..4439a9eb6b7a 100644
--- a/drivers/phy/mediatek/phy-mtk-ufs.c
+++ b/drivers/phy/mediatek/phy-mtk-ufs.c
@@ -4,12 +4,15 @@
* Author: Stanley Chu <stanley.chu@mediatek.com>
*/
+#include <linux/arm-smccc.h>
#include <linux/clk.h>
#include <linux/delay.h>
#include <linux/io.h>
#include <linux/module.h>
#include <linux/phy/phy.h>
#include <linux/platform_device.h>
+#include <linux/reset.h>
+#include <linux/soc/mediatek/mtk_sip_svc.h>
#include "phy-mtk-io.h"
@@ -35,9 +38,17 @@
#define UFSPHY_CLKS_CNT 2
+#define UFS_MTK_SIP_MPHY_CTRL BIT(8)
+
+enum ufs_mtk_mphy_op {
+ UFS_MPHY_BACKUP = 0,
+ UFS_MPHY_RESTORE
+};
+
struct ufs_mtk_phy {
struct device *dev;
void __iomem *mmio;
+ struct reset_control *reset;
struct clk_bulk_data clks[UFSPHY_CLKS_CNT];
};
@@ -140,9 +151,59 @@ static int ufs_mtk_phy_power_off(struct phy *generic_phy)
return 0;
}
+static int ufs_mtk_phy_ctrl(struct ufs_mtk_phy *phy, enum ufs_mtk_mphy_op op)
+{
+ struct arm_smccc_res res;
+
+ arm_smccc_smc(MTK_SIP_UFS_CONTROL, UFS_MTK_SIP_MPHY_CTRL, op,
+ 0, 0, 0, 0, 0, &res);
+
+ switch (res.a0) {
+ case SMCCC_RET_NOT_SUPPORTED:
+ return -EOPNOTSUPP;
+ case SMCCC_RET_INVALID_PARAMETER:
+ return -EINVAL;
+ default:
+ return 0;
+ }
+}
+
+static int ufs_mtk_phy_reset(struct phy *generic_phy)
+{
+ struct ufs_mtk_phy *phy = get_ufs_mtk_phy(generic_phy);
+ int ret;
+
+ if (!phy->reset)
+ return -EOPNOTSUPP;
+
+ ret = reset_control_assert(phy->reset);
+ if (ret)
+ return ret;
+
+ usleep_range(100, 110);
+
+ ret = reset_control_deassert(phy->reset);
+ if (ret)
+ return ret;
+
+ /*
+ * To avoid double-sleep and other unintended side-effects in the ufshci
+ * driver, don't return the phy_ctrl retval here, but just return -EPROTO.
+ */
+ ret = ufs_mtk_phy_ctrl(phy, UFS_MPHY_RESTORE);
+ if (ret) {
+ dev_err(phy->dev, "UFS_MPHY_RESTORE SMC command failed: %pe\n",
+ ERR_PTR(ret));
+ return -EPROTO;
+ }
+
+ return 0;
+}
+
static const struct phy_ops ufs_mtk_phy_ops = {
.power_on = ufs_mtk_phy_power_on,
.power_off = ufs_mtk_phy_power_off,
+ .reset = ufs_mtk_phy_reset,
.owner = THIS_MODULE,
};
@@ -162,8 +223,18 @@ static int ufs_mtk_phy_probe(struct platform_device *pdev)
if (IS_ERR(phy->mmio))
return PTR_ERR(phy->mmio);
+ phy->reset = devm_reset_control_get_optional_exclusive(dev, NULL);
+ if (IS_ERR(phy->reset))
+ return dev_err_probe(dev, PTR_ERR(phy->reset), "Failed to get reset\n");
+
phy->dev = dev;
+ if (phy->reset) {
+ ret = ufs_mtk_phy_ctrl(phy, UFS_MPHY_BACKUP);
+ if (ret)
+ return dev_err_probe(dev, ret, "Failed to back up MPHY\n");
+ }
+
ret = ufs_mtk_phy_clk_init(phy);
if (ret)
return ret;
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 06/26] scsi: ufs: mediatek: Rework resets
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (4 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 05/26] phy: mediatek: ufs: Add support for resets Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:29 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 07/26] scsi: ufs: mediatek: Rework 0.9V regulator Louis-Alexis Eyraud
` (19 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Rework the reset control getting in the driver's probe function to use
the bulk reset APIs. Use the optional variant instead of defaulting to
NULL if the resets fail, so that absent resets can be distinguished from
erroneous resets.
Also remove all remnants of the MPHY reset ever having lived in this
driver.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Reviewed-by: Philipp Zabel <p.zabel@pengutronix.de>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek-sip.h | 8 ----
drivers/ufs/host/ufs-mediatek.c | 78 ++++++++++++++++++-------------------
drivers/ufs/host/ufs-mediatek.h | 7 ++--
3 files changed, 42 insertions(+), 51 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek-sip.h b/drivers/ufs/host/ufs-mediatek-sip.h
index d627dfb4a766..256598cc3b5b 100644
--- a/drivers/ufs/host/ufs-mediatek-sip.h
+++ b/drivers/ufs/host/ufs-mediatek-sip.h
@@ -31,11 +31,6 @@ enum ufs_mtk_vcc_num {
UFS_VCC_MAX
};
-enum ufs_mtk_mphy_op {
- UFS_MPHY_BACKUP = 0,
- UFS_MPHY_RESTORE
-};
-
/*
* SMC call wrapper function
*/
@@ -84,9 +79,6 @@ static inline void _ufs_mtk_smc(struct ufs_mtk_smc_arg s)
#define ufs_mtk_device_pwr_ctrl(on, ufs_version, res) \
ufs_mtk_smc(UFS_MTK_SIP_DEVICE_PWR_CTRL, &(res), on, ufs_version)
-#define ufs_mtk_mphy_ctrl(op, res) \
- ufs_mtk_smc(UFS_MTK_SIP_MPHY_CTRL, &(res), op)
-
#define ufs_mtk_mtcmos_ctrl(op, res) \
ufs_mtk_smc(UFS_MTK_SIP_MTCMOS_CTRL, &(res), op)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 814c1b7343b9..7d4fb9d067de 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -93,6 +93,12 @@ static const char *const ufs_uic_dl_err_str[] = {
"PA_INIT"
};
+static const char *const ufs_reset_names[] = {
+ "unipro",
+ "crypto",
+ "hci",
+};
+
static bool ufs_mtk_is_boost_crypt_enabled(struct ufs_hba *hba)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
@@ -203,49 +209,45 @@ static void ufs_mtk_crypto_enable(struct ufs_hba *hba)
static void ufs_mtk_host_reset(struct ufs_hba *hba)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
- struct arm_smccc_res res;
-
- reset_control_assert(host->hci_reset);
- reset_control_assert(host->crypto_reset);
- reset_control_assert(host->unipro_reset);
- reset_control_assert(host->mphy_reset);
-
- usleep_range(100, 110);
+ int ret;
- reset_control_deassert(host->unipro_reset);
- reset_control_deassert(host->crypto_reset);
- reset_control_deassert(host->hci_reset);
- reset_control_deassert(host->mphy_reset);
+ ret = reset_control_bulk_assert(MTK_UFS_NUM_RESETS, host->resets);
+ if (ret)
+ dev_warn(hba->dev, "Host reset assert failed: %pe\n", ERR_PTR(ret));
- /* restore mphy setting aftre mphy reset */
- if (host->mphy_reset)
- ufs_mtk_mphy_ctrl(UFS_MPHY_RESTORE, res);
-}
+ ret = phy_reset(host->mphy);
-static void ufs_mtk_init_reset_control(struct ufs_hba *hba,
- struct reset_control **rc,
- char *str)
-{
- *rc = devm_reset_control_get(hba->dev, str);
- if (IS_ERR(*rc)) {
- dev_info(hba->dev, "Failed to get reset control %s: %ld\n",
- str, PTR_ERR(*rc));
- *rc = NULL;
+ /*
+ * Only sleep if MPHY doesn't have a reset implemented (which already
+ * sleeps) or the PHY reset function failed somehow, just to be safe
+ */
+ if (ret) {
+ usleep_range(100, 110);
+ if (ret != -EOPNOTSUPP)
+ dev_warn(hba->dev, "PHY reset failed: %pe\n", ERR_PTR(ret));
}
+
+ ret = reset_control_bulk_deassert(MTK_UFS_NUM_RESETS, host->resets);
+ if (ret)
+ dev_warn(hba->dev, "Host reset deassert failed: %pe\n", ERR_PTR(ret));
}
-static void ufs_mtk_init_reset(struct ufs_hba *hba)
+static int ufs_mtk_init_reset(struct ufs_hba *hba)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
+ int ret, i;
+
+ for (i = 0; i < MTK_UFS_NUM_RESETS; i++)
+ host->resets[i].id = ufs_reset_names[i];
- ufs_mtk_init_reset_control(hba, &host->hci_reset,
- "hci_rst");
- ufs_mtk_init_reset_control(hba, &host->unipro_reset,
- "unipro_rst");
- ufs_mtk_init_reset_control(hba, &host->crypto_reset,
- "crypto_rst");
- ufs_mtk_init_reset_control(hba, &host->mphy_reset,
- "mphy_rst");
+ ret = devm_reset_control_bulk_get_optional_exclusive(hba->dev, MTK_UFS_NUM_RESETS,
+ host->resets);
+ if (ret) {
+ dev_err(hba->dev, "Failed to get resets: %pe\n", ERR_PTR(ret));
+ return ret;
+ }
+
+ return 0;
}
static int ufs_mtk_hce_enable_notify(struct ufs_hba *hba,
@@ -1247,11 +1249,9 @@ static int ufs_mtk_init(struct ufs_hba *hba)
if (err)
goto out_variant_clear;
- ufs_mtk_init_reset(hba);
-
- /* backup mphy setting if mphy can reset */
- if (host->mphy_reset)
- ufs_mtk_mphy_ctrl(UFS_MPHY_BACKUP, res);
+ err = ufs_mtk_init_reset(hba);
+ if (err)
+ goto out_variant_clear;
/* Enable runtime autosuspend */
hba->caps |= UFSHCD_CAP_RPM_AUTOSUSPEND;
diff --git a/drivers/ufs/host/ufs-mediatek.h b/drivers/ufs/host/ufs-mediatek.h
index 73cdc726f290..f95f12eb9926 100644
--- a/drivers/ufs/host/ufs-mediatek.h
+++ b/drivers/ufs/host/ufs-mediatek.h
@@ -7,12 +7,14 @@
#define _UFS_MEDIATEK_H
#include <linux/bitops.h>
+#include <linux/reset.h>
/*
* MCQ define and struct
*/
#define UFSHCD_MAX_Q_NR 8
#define MTK_MCQ_INVALID_IRQ 0xFFFF
+#define MTK_UFS_NUM_RESETS 3
/* REG_UFS_MMIO_OPT_CTRL_0 160h */
#define EHS_EN BIT(0)
@@ -175,10 +177,7 @@ struct ufs_mtk_mcq_intr_info {
struct ufs_mtk_host {
struct phy *mphy;
struct regulator *reg_va09;
- struct reset_control *hci_reset;
- struct reset_control *unipro_reset;
- struct reset_control *crypto_reset;
- struct reset_control *mphy_reset;
+ struct reset_control_bulk_data resets[MTK_UFS_NUM_RESETS];
struct ufs_hba *hba;
struct ufs_mtk_crypt_cfg *crypt;
struct ufs_mtk_clk mclk;
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 06/26] scsi: ufs: mediatek: Rework resets
2026-07-30 9:49 ` [PATCH v10 06/26] scsi: ufs: mediatek: Rework resets Louis-Alexis Eyraud
@ 2026-07-30 10:29 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:29 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: conor+dt, vkoul, neil.armstrong, devicetree, olteanv, linux-phy,
linux-scsi, robh
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [High] Hardware Reset Sequence Violation due to inverted bulk reset array order and improper PHY synchronization.
- [High] Missing mandatory hardware reset delay when UFS MPHY is unbound.
--
--- Patch [6]: [PATCH v10 06/26] scsi: ufs: mediatek: Rework resets ---
commit 6e1eee5b92733c43e2ae7add5906da2c59fbe389
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Rework resets
This commit reworks the reset control getting in the driver's probe
function to use the bulk reset APIs instead of getting resets individually.
It also removes all remnants of the MPHY reset from this driver.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
> @@ -93,6 +93,12 @@ static const char *const ufs_uic_dl_err_str[] = {
> "PA_INIT"
> };
>
> +static const char *const ufs_reset_names[] = {
> + "unipro",
> + "crypto",
> + "hci",
> +};
[Severity: High]
Does this array ordering accidentally reverse the original hardware reset
sequence?
The original code explicitly asserted resets in the order: HCI, Crypto,
UniPro, and deasserted them in the order: UniPro, Crypto, HCI.
Because reset_control_bulk_assert() iterates forwards and
reset_control_bulk_deassert() iterates backwards, this new array ordering
will assert UniPro, Crypto, HCI, and deassert HCI, Crypto, UniPro.
[ ... ]
> @@ -203,49 +209,45 @@ static void ufs_mtk_crypto_enable(struct ufs_hba *hba)
> static void ufs_mtk_host_reset(struct ufs_hba *hba)
> {
> struct ufs_mtk_host *host = ufshcd_get_variant(hba);
> - struct arm_smccc_res res;
> -
> - reset_control_assert(host->hci_reset);
> - reset_control_assert(host->crypto_reset);
> - reset_control_assert(host->unipro_reset);
> - reset_control_assert(host->mphy_reset);
> -
> - usleep_range(100, 110);
>
> - reset_control_deassert(host->unipro_reset);
> - reset_control_deassert(host->crypto_reset);
> - reset_control_deassert(host->hci_reset);
> - reset_control_deassert(host->mphy_reset);
> + int ret;
>
> - /* restore mphy setting aftre mphy reset */
> - if (host->mphy_reset)
> - ufs_mtk_mphy_ctrl(UFS_MPHY_RESTORE, res);
> -}
> + ret = reset_control_bulk_assert(MTK_UFS_NUM_RESETS, host->resets);
> + if (ret)
> + dev_warn(hba->dev, "Host reset assert failed: %pe\n", ERR_PTR(ret));
>
> + ret = phy_reset(host->mphy);
[Severity: High]
Are we restoring the PHY while the MAC/host controller blocks are still held
in reset?
In the original code, the MPHY SMC restore was called after all controller
blocks were deasserted. Here, phy_reset() executes the full MPHY reset and
restore before reset_control_bulk_deassert() is called.
> +
> + /*
> + * Only sleep if MPHY doesn't have a reset implemented (which already
> + * sleeps) or the PHY reset function failed somehow, just to be safe
> + */
> + if (ret) {
> + usleep_range(100, 110);
> + if (ret != -EOPNOTSUPP)
> + dev_warn(hba->dev, "PHY reset failed: %pe\n", ERR_PTR(ret));
> }
[Severity: High]
Does this skip the mandatory hardware reset delay when the UFS MPHY is
unbound?
If host->mphy is NULL, phy_reset(NULL) will return 0. This causes the if (ret)
condition to evaluate to false, bypassing the usleep_range(100, 110)
fallback.
This would result in the MAC/HCI resets being asserted and deasserted
instantaneously back-to-back, which can cause the controller IP block to
fail to reset.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=6
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 07/26] scsi: ufs: mediatek: Rework 0.9V regulator
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (5 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 06/26] scsi: ufs: mediatek: Rework resets Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:13 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 08/26] scsi: ufs: mediatek: Rework init function Louis-Alexis Eyraud
` (18 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The mediatek UFS host driver does some pretty bad stuff with regards to
the 0.9V regulator. Instead of just checking for the presence of the
regulator, it adds a cap if it's there, and then checks for the cap. It
also sleeps to stabilise the supply after enabling the regulator, which
is something that should be done by the regulator framework with the
appropriate delay properties in the DTS instead of random sleeps in the
driver code.
Rework this code and rename it to the avdd09 name I've chosen in the
binding for this supply name, instead of the downstream "va09" name that
isn't used by the datasheets for any of these chips.
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 141 ++++++++++++++++++++++++++--------------
drivers/ufs/host/ufs-mediatek.h | 3 +-
2 files changed, 95 insertions(+), 49 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 7d4fb9d067de..f6859cd0b1d4 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -38,6 +38,10 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up);
#define MAX_SUPP_MAC 64
#define MCQ_QUEUE_OFFSET(c) ((((c) >> 16) & 0xFF) * 0x200)
+struct ufs_mtk_soc_data {
+ bool has_avdd09;
+};
+
static const struct ufs_dev_quirk ufs_mtk_dev_fixups[] = {
{ .wmanufacturerid = UFS_ANY_VENDOR,
.model = UFS_ANY_MODEL,
@@ -48,8 +52,13 @@ static const struct ufs_dev_quirk ufs_mtk_dev_fixups[] = {
{}
};
+static const struct ufs_mtk_soc_data mt8183_data = {
+ .has_avdd09 = true,
+};
+
static const struct of_device_id ufs_mtk_of_match[] = {
- { .compatible = "mediatek,mt8183-ufshci" },
+ { .compatible = "mediatek,mt8183-ufshci", .data = &mt8183_data },
+ { .compatible = "mediatek,mt8192-ufshci" },
{ .compatible = "mediatek,mt8195-ufshci" },
{},
};
@@ -106,13 +115,6 @@ static bool ufs_mtk_is_boost_crypt_enabled(struct ufs_hba *hba)
return host->caps & UFS_MTK_CAP_BOOST_CRYPT_ENGINE;
}
-static bool ufs_mtk_is_va09_supported(struct ufs_hba *hba)
-{
- struct ufs_mtk_host *host = ufshcd_get_variant(hba);
-
- return host->caps & UFS_MTK_CAP_VA09_PWR_CTRL;
-}
-
static bool ufs_mtk_is_broken_vcc(struct ufs_hba *hba)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
@@ -506,44 +508,70 @@ static int ufs_mtk_wait_link_state(struct ufs_hba *hba, u32 state,
return -ETIMEDOUT;
}
+static int ufs_mtk_09v_off(struct ufs_mtk_host *host)
+{
+ struct arm_smccc_res res;
+ int ret;
+
+ if (!host->reg_avdd09)
+ return 0;
+
+ ufs_mtk_va09_pwr_ctrl(res, 0);
+ ret = regulator_disable(host->reg_avdd09);
+ if (ret) {
+ dev_err(host->hba->dev, "Failed to disable avdd09-supply: %pe\n",
+ ERR_PTR(ret));
+ ufs_mtk_va09_pwr_ctrl(res, 1);
+ return ret;
+ }
+
+ return 0;
+}
+
+static int ufs_mtk_09v_on(struct ufs_mtk_host *host)
+{
+ struct arm_smccc_res res;
+ int ret;
+
+ if (!host->reg_avdd09)
+ return 0;
+
+ ret = regulator_enable(host->reg_avdd09);
+ if (ret) {
+ dev_err(host->hba->dev, "Failed to enable avdd09-supply: %pe\n",
+ ERR_PTR(ret));
+ return ret;
+ }
+
+ ufs_mtk_va09_pwr_ctrl(res, 1);
+
+ return 0;
+}
+
static int ufs_mtk_mphy_power_on(struct ufs_hba *hba, bool on)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
struct phy *mphy = host->mphy;
- struct arm_smccc_res res;
- int ret = 0;
+ int ret;
- if (!mphy || !(on ^ host->mphy_powered_on))
+ if (!mphy || on == host->mphy_powered_on)
return 0;
if (on) {
- if (ufs_mtk_is_va09_supported(hba)) {
- ret = regulator_enable(host->reg_va09);
- if (ret < 0)
- goto out;
- /* wait 200 us to stablize VA09 */
- usleep_range(200, 210);
- ufs_mtk_va09_pwr_ctrl(res, 1);
- }
+ ret = ufs_mtk_09v_on(host);
+ if (ret)
+ return ret;
phy_power_on(mphy);
} else {
phy_power_off(mphy);
- if (ufs_mtk_is_va09_supported(hba)) {
- ufs_mtk_va09_pwr_ctrl(res, 0);
- ret = regulator_disable(host->reg_va09);
- }
- }
-out:
- if (ret) {
- dev_info(hba->dev,
- "failed to %s va09: %d\n",
- on ? "enable" : "disable",
- ret);
- } else {
- host->mphy_powered_on = on;
+ ret = ufs_mtk_09v_off(host);
+ if (ret)
+ return ret;
}
- return ret;
+ host->mphy_powered_on = on;
+
+ return 0;
}
static int ufs_mtk_get_host_clk(struct device *dev, const char *name,
@@ -678,17 +706,6 @@ static void ufs_mtk_init_boost_crypt(struct ufs_hba *hba)
return;
}
-static void ufs_mtk_init_va09_pwr_ctrl(struct ufs_hba *hba)
-{
- struct ufs_mtk_host *host = ufshcd_get_variant(hba);
-
- host->reg_va09 = regulator_get(hba->dev, "va09");
- if (IS_ERR(host->reg_va09))
- dev_info(hba->dev, "failed to get va09");
- else
- host->caps |= UFS_MTK_CAP_VA09_PWR_CTRL;
-}
-
static void ufs_mtk_init_host_caps(struct ufs_hba *hba)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
@@ -697,9 +714,6 @@ static void ufs_mtk_init_host_caps(struct ufs_hba *hba)
if (of_property_read_bool(np, "mediatek,ufs-boost-crypt"))
ufs_mtk_init_boost_crypt(hba);
- if (of_property_read_bool(np, "mediatek,ufs-support-va09"))
- ufs_mtk_init_va09_pwr_ctrl(hba);
-
if (of_property_read_bool(np, "mediatek,ufs-disable-ah8"))
host->caps |= UFS_MTK_CAP_DISABLE_AH8;
@@ -1205,6 +1219,35 @@ static void ufs_mtk_init_mcq_irq(struct ufs_hba *hba)
host->mcq_nr_intr = 0;
}
+/**
+ * ufs_mtk_get_supplies - acquire variant-specific supplies
+ * @host: pointer to driver's private &struct ufs_mtk_host instance
+ *
+ * Returns 0 on success, negative errno on error.
+ */
+static int ufs_mtk_get_supplies(struct ufs_mtk_host *host)
+{
+ struct device *dev = host->hba->dev;
+ const struct ufs_mtk_soc_data *data = of_device_get_match_data(dev);
+
+ if (!data || !data->has_avdd09)
+ return 0;
+
+ host->reg_avdd09 = devm_regulator_get_optional(dev, "avdd09");
+ if (IS_ERR(host->reg_avdd09)) {
+ if (PTR_ERR(host->reg_avdd09) == -ENODEV) {
+ host->reg_avdd09 = NULL;
+ return 0;
+ }
+
+ dev_err(dev, "Failed to get avdd09 regulator: %pe\n",
+ host->reg_avdd09);
+ return PTR_ERR(host->reg_avdd09);
+ }
+
+ return 0;
+}
+
/**
* ufs_mtk_init - find other essential mmio bases
* @hba: host controller instance
@@ -1288,6 +1331,10 @@ static int ufs_mtk_init(struct ufs_hba *hba)
ufs_mtk_init_clocks(hba);
+ err = ufs_mtk_get_supplies(host);
+ if (err)
+ goto out_variant_clear;
+
/*
* ufshcd_vops_init() is invoked after
* ufshcd_setup_clock(true) in ufshcd_hba_init() thus
diff --git a/drivers/ufs/host/ufs-mediatek.h b/drivers/ufs/host/ufs-mediatek.h
index f95f12eb9926..c63740b98e85 100644
--- a/drivers/ufs/host/ufs-mediatek.h
+++ b/drivers/ufs/host/ufs-mediatek.h
@@ -125,7 +125,6 @@ enum {
*/
enum ufs_mtk_host_caps {
UFS_MTK_CAP_BOOST_CRYPT_ENGINE = 1 << 0,
- UFS_MTK_CAP_VA09_PWR_CTRL = 1 << 1,
UFS_MTK_CAP_DISABLE_AH8 = 1 << 2,
UFS_MTK_CAP_BROKEN_VCC = 1 << 3,
@@ -176,7 +175,7 @@ struct ufs_mtk_mcq_intr_info {
struct ufs_mtk_host {
struct phy *mphy;
- struct regulator *reg_va09;
+ struct regulator *reg_avdd09;
struct reset_control_bulk_data resets[MTK_UFS_NUM_RESETS];
struct ufs_hba *hba;
struct ufs_mtk_crypt_cfg *crypt;
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 07/26] scsi: ufs: mediatek: Rework 0.9V regulator
2026-07-30 9:49 ` [PATCH v10 07/26] scsi: ufs: mediatek: Rework 0.9V regulator Louis-Alexis Eyraud
@ 2026-07-30 10:13 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:13 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: robh, devicetree, vkoul, linux-scsi, linux-phy, neil.armstrong,
conor+dt, olteanv
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
New issues:
- [Medium] `ufs_mtk_get_supplies()` uses `dev_err()` instead of `dev_err_probe()`, resulting in excessive log spam when the regulator probe defers.
Pre-existing issues:
- [High] Missing error handling in `ufs_mtk_mphy_power_on()` leads to PHY framework reference count underflow and state desynchronization.
--
--- Patch [7]: [PATCH v10 07/26] scsi: ufs: mediatek: Rework 0.9V regulator ---
Note: The format of this report has been altered due to recitation restrictions. The original patch diff is not quoted, and findings are provided as a free-form summary.
commit 53fa70c9d3d751e82ebb61d97d0c2282f65046c5
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Rework 0.9V regulator
This commit reworks the 0.9V regulator handling in the Mediatek UFS host
driver. It removes the custom downstream naming, uses the standard regulator
API with the avdd09 supply name, and eliminates manual stabilization delays.
[Severity: High]
This is a pre-existing issue, but does ufs_mtk_mphy_power_on() handle
failures correctly during PHY power transitions?
When calling phy_power_on(), the return value is not checked. If hardware
initialization fails, host->mphy_powered_on is still set to true.
Additionally, on the power-off path, if the newly added ufs_mtk_09v_off()
fails, the function returns early without setting host->mphy_powered_on
to false.
Could this desynchronize the driver's state from the PHY hardware state?
If the driver incorrectly records the power state as true, subsequent
suspend or error handling paths might blindly call phy_power_off() again,
which would lead to power_count underflows in the PHY framework and
permanently break the PHY until a system reboot.
[Severity: Medium]
In the newly added ufs_mtk_get_supplies(), dev_err() is used to log
failures from devm_regulator_get_optional() for the avdd09 regulator.
If the regulator returns -EPROBE_DEFER, will this dev_err() call cause
excessive log spam during the device probe phase?
Would it make sense to use dev_err_probe() here instead to properly handle
and suppress expected probe deferral messages?
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=7
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 08/26] scsi: ufs: mediatek: Rework init function
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (6 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 07/26] scsi: ufs: mediatek: Rework 0.9V regulator Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:07 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 09/26] scsi: ufs: mediatek: Rework the crypt-boost stuff Louis-Alexis Eyraud
` (17 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Printing an error message on ENOMEM is pointless. The print will not
work because there is no memory.
Adding an of_match_device to the init function is pointless. Why would a
different device with a different probe function ever use the same init
function? Get rid of it.
zero-initialising an error variable just so you can then goto a bare
return statement with that error variable to signal success is also
pointless, just return directly, there's no unwind being done.
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 43 +++++++++++++++--------------------------
1 file changed, 16 insertions(+), 27 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index f6859cd0b1d4..50e7701d7943 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -52,18 +52,6 @@ static const struct ufs_dev_quirk ufs_mtk_dev_fixups[] = {
{}
};
-static const struct ufs_mtk_soc_data mt8183_data = {
- .has_avdd09 = true,
-};
-
-static const struct of_device_id ufs_mtk_of_match[] = {
- { .compatible = "mediatek,mt8183-ufshci", .data = &mt8183_data },
- { .compatible = "mediatek,mt8192-ufshci" },
- { .compatible = "mediatek,mt8195-ufshci" },
- {},
-};
-MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
-
/*
* Details of UIC Errors
*/
@@ -1260,29 +1248,19 @@ static int ufs_mtk_get_supplies(struct ufs_mtk_host *host)
*/
static int ufs_mtk_init(struct ufs_hba *hba)
{
- const struct of_device_id *id;
struct device *dev = hba->dev;
struct ufs_mtk_host *host;
struct Scsi_Host *shost = hba->host;
- int err = 0;
+ int err;
struct arm_smccc_res res;
host = devm_kzalloc(dev, sizeof(*host), GFP_KERNEL);
- if (!host) {
- err = -ENOMEM;
- dev_info(dev, "%s: no memory for mtk ufs host\n", __func__);
- goto out;
- }
+ if (!host)
+ return -ENOMEM;
host->hba = hba;
ufshcd_set_variant(hba, host);
- id = of_match_device(ufs_mtk_of_match, dev);
- if (!id) {
- err = -EINVAL;
- goto out;
- }
-
/* Initialize host capability */
ufs_mtk_init_host_caps(hba);
@@ -1356,11 +1334,10 @@ static int ufs_mtk_init(struct ufs_hba *hba)
ufs_mtk_get_hw_ip_version(hba);
- goto out;
+ return 0;
out_variant_clear:
ufshcd_set_variant(hba, NULL);
-out:
return err;
}
@@ -2402,6 +2379,18 @@ static const struct ufs_hba_variant_ops ufs_hba_mtk_vops = {
.config_scsi_dev = ufs_mtk_config_scsi_dev,
};
+static const struct ufs_mtk_soc_data mt8183_data = {
+ .has_avdd09 = true,
+};
+
+static const struct of_device_id ufs_mtk_of_match[] = {
+ { .compatible = "mediatek,mt8183-ufshci", .data = &mt8183_data },
+ { .compatible = "mediatek,mt8192-ufshci" },
+ { .compatible = "mediatek,mt8195-ufshci" },
+ {},
+};
+MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
+
/**
* ufs_mtk_probe - probe routine of the driver
* @pdev: pointer to Platform device handle
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 08/26] scsi: ufs: mediatek: Rework init function
2026-07-30 9:49 ` [PATCH v10 08/26] scsi: ufs: mediatek: Rework init function Louis-Alexis Eyraud
@ 2026-07-30 10:07 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:07 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: neil.armstrong, devicetree, vkoul, conor+dt, olteanv, robh,
linux-phy, linux-scsi
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
Pre-existing issues:
- [High] The return value of `ufs_mtk_mphy_power_on()` is ignored during host initialization.
- [High] Use-After-Free risk due to saving a released device pointer in `host->phy_dev`.
--
--- Patch [8]: [PATCH v10 08/26] scsi: ufs: mediatek: Rework init function ---
commit 5d79268ac1559dd4c6c615601ae1b9838e4b77bc
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Rework init function
This commit reworks the initialization function for the MediaTek UFS driver.
It removes unnecessary error prints for memory allocation failures and removes
redundant device tree matching. It also simplifies the error handling path
by removing unnecessary goto statements and returning directly on success.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index f6859cd0b1d4d..50e7701d7943b 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
> @@ -1356,11 +1334,10 @@ static int ufs_mtk_init(struct ufs_hba *hba)
>
> ufs_mtk_get_hw_ip_version(hba);
>
> - goto out;
> + return 0;
>
> out_variant_clear:
> ufshcd_set_variant(hba, NULL);
> -out:
> return err;
> }
[Severity: High]
This isn't a bug introduced by this patch, but does this code ignore the
return value of ufs_mtk_mphy_power_on() earlier in ufs_mtk_init()?
Looking at ufs_mtk_init():
err = ufs_mtk_get_supplies(host);
if (err)
goto out_variant_clear;
/*
* ufshcd_vops_init() is invoked after
* ufshcd_setup_clock(true) in ufshcd_hba_init() thus
* phy clock setup is skipped.
*
* Enable phy clocks specifically here.
*/
ufs_mtk_mphy_power_on(hba, true);
if (ufs_mtk_is_rtff_mtcmos(hba)) {
If ufs_mtk_mphy_power_on() fails (e.g. regulator_enable fails in
ufs_mtk_09v_on), the driver will proceed as if the PHY powered on
successfully. Should this check the return value and handle the error?
[ ... ]
> @@ -2402,6 +2379,18 @@ static const struct ufs_hba_variant_ops ufs_hba_mtk_vops = {
> .config_scsi_dev = ufs_mtk_config_scsi_dev,
> };
>
> +static const struct ufs_mtk_soc_data mt8183_data = {
> + .has_avdd09 = true,
> +};
> +
> +static const struct of_device_id ufs_mtk_of_match[] = {
> + { .compatible = "mediatek,mt8183-ufshci", .data = &mt8183_data },
> + { .compatible = "mediatek,mt8192-ufshci" },
> + { .compatible = "mediatek,mt8195-ufshci" },
> + {},
> +};
> +MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
> +
> /**
> * ufs_mtk_probe - probe routine of the driver
> * @pdev: pointer to Platform device handle
[Severity: High]
This is a pre-existing issue, but is there a use-after-free risk in
ufs_mtk_probe() when saving the released device pointer?
Looking at ufs_mtk_probe():
if (phy_node) {
phy_pdev = of_find_device_by_node(phy_node);
if (!phy_pdev)
goto skip_phy;
phy_dev = &phy_pdev->dev;
...
put_device(phy_dev);
}
...
skip_phy:
err = ufshcd_pltfrm_init(pdev, &ufs_hba_mtk_vops);
...
if (phy_node && phy_dev) {
host = ufshcd_get_variant(hba);
host->phy_dev = phy_dev;
}
of_find_device_by_node() acquires a reference to the phy device. This
reference is unconditionally dropped via put_device(phy_dev), but the
unreferenced pointer is then stored in host->phy_dev.
If the phy device is removed or unbound and the memory is freed, the UFS
driver continues to dereference it during runtime PM operations:
drivers/ufs/host/ufs-mediatek.c:ufs_mtk_runtime_suspend() {
...
if (host->phy_dev)
pm_runtime_put_sync(host->phy_dev);
...
}
Should the reference be held until the UFS host device is unbound?
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=8
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 09/26] scsi: ufs: mediatek: Rework the crypt-boost stuff
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (7 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 08/26] scsi: ufs: mediatek: Rework init function Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:16 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 10/26] scsi: ufs: mediatek: Handle misc host voltage regulators Louis-Alexis Eyraud
` (16 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
I don't know whether the crypt-boost functionality as it is currently
implemented is even appropriate for mainline. It might be better done in
some generic way. But what I do know is that I can rework the code to
make it less obtuse.
Prefix the boost stuff with the appropriate vendor prefix, remove the
pointless clock wrappers, and rework the function.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang (王信友) <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 89 ++++++++++++++---------------------------
1 file changed, 30 insertions(+), 59 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 50e7701d7943..55701172bcf1 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -562,21 +562,6 @@ static int ufs_mtk_mphy_power_on(struct ufs_hba *hba, bool on)
return 0;
}
-static int ufs_mtk_get_host_clk(struct device *dev, const char *name,
- struct clk **clk_out)
-{
- struct clk *clk;
- int err = 0;
-
- clk = devm_clk_get(dev, name);
- if (IS_ERR(clk))
- err = PTR_ERR(clk);
- else
- *clk_out = clk;
-
- return err;
-}
-
static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
@@ -633,65 +618,51 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
clk_disable_unprepare(cfg->clk_crypt_mux);
}
-static int ufs_mtk_init_host_clk(struct ufs_hba *hba, const char *name,
- struct clk **clk)
-{
- int ret;
-
- ret = ufs_mtk_get_host_clk(hba->dev, name, clk);
- if (ret) {
- dev_info(hba->dev, "%s: failed to get %s: %d", __func__,
- name, ret);
- }
-
- return ret;
-}
-
static void ufs_mtk_init_boost_crypt(struct ufs_hba *hba)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
struct ufs_mtk_crypt_cfg *cfg;
struct device *dev = hba->dev;
- struct regulator *reg;
- u32 volt;
+ int ret;
- host->crypt = devm_kzalloc(dev, sizeof(*(host->crypt)),
- GFP_KERNEL);
- if (!host->crypt)
- goto disable_caps;
+ cfg = devm_kzalloc(dev, sizeof(*cfg), GFP_KERNEL);
+ if (!cfg)
+ return;
- reg = devm_regulator_get_optional(dev, "dvfsrc-vcore");
- if (IS_ERR(reg)) {
- dev_info(dev, "failed to get dvfsrc-vcore: %ld",
- PTR_ERR(reg));
- goto disable_caps;
+ cfg->reg_vcore = devm_regulator_get_optional(dev, "dvfsrc-vcore");
+ if (IS_ERR(cfg->reg_vcore)) {
+ dev_err(dev, "Failed to get dvfsrc-vcore: %pe", cfg->reg_vcore);
+ return;
}
- if (of_property_read_u32(dev->of_node, "boost-crypt-vcore-min",
- &volt)) {
- dev_info(dev, "failed to get boost-crypt-vcore-min");
- goto disable_caps;
+ ret = of_property_read_u32(dev->of_node, "mediatek,boost-crypt-vcore-min",
+ &cfg->vcore_volt);
+ if (ret) {
+ dev_err(dev, "Failed to get mediatek,boost-crypt-vcore-min: %pe\n",
+ ERR_PTR(ret));
+ return;
}
- cfg = host->crypt;
- if (ufs_mtk_init_host_clk(hba, "crypt_mux",
- &cfg->clk_crypt_mux))
- goto disable_caps;
+ cfg->clk_crypt_mux = devm_clk_get(dev, "crypt_mux");
+ if (IS_ERR(cfg->clk_crypt_mux)) {
+ dev_err(dev, "Failed to get clock crypt_mux: %pe\n", cfg->clk_crypt_mux);
+ return;
+ }
- if (ufs_mtk_init_host_clk(hba, "crypt_lp",
- &cfg->clk_crypt_lp))
- goto disable_caps;
+ cfg->clk_crypt_lp = devm_clk_get(dev, "crypt_lp");
+ if (IS_ERR(cfg->clk_crypt_lp)) {
+ dev_err(dev, "Failed to get clock crypt_lp: %pe\n", cfg->clk_crypt_lp);
+ return;
+ }
- if (ufs_mtk_init_host_clk(hba, "crypt_perf",
- &cfg->clk_crypt_perf))
- goto disable_caps;
+ cfg->clk_crypt_perf = devm_clk_get(dev, "crypt_perf");
+ if (IS_ERR(cfg->clk_crypt_perf)) {
+ dev_err(dev, "Failed to get clock crypt_perf: %pe\n", cfg->clk_crypt_perf);
+ return;
+ }
- cfg->reg_vcore = reg;
- cfg->vcore_volt = volt;
+ host->crypt = cfg;
host->caps |= UFS_MTK_CAP_BOOST_CRYPT_ENGINE;
-
-disable_caps:
- return;
}
static void ufs_mtk_init_host_caps(struct ufs_hba *hba)
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 09/26] scsi: ufs: mediatek: Rework the crypt-boost stuff
2026-07-30 9:49 ` [PATCH v10 09/26] scsi: ufs: mediatek: Rework the crypt-boost stuff Louis-Alexis Eyraud
@ 2026-07-30 10:16 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:16 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: devicetree, vkoul, neil.armstrong, linux-phy, conor+dt, robh,
olteanv, linux-scsi
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [Low] Missing newline character `\n` in `dev_err` format string.
- [Low] Incompatible pointer type passed to `of_property_read_u32`.
--
--- Patch [9]: [PATCH v10 09/26] scsi: ufs: mediatek: Rework the crypt-boost stuff ---
commit cbee34751328446d33c164e90e794b44773af5eb
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Rework the crypt-boost stuff
This commit reworks the crypt-boost functionality by adding vendor
prefixes to the device tree properties, removing unnecessary clock
wrappers, and restructuring the initialization code.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index 50e7701d7943b..55701172bcf15 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
> @@ -633,65 +618,51 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
[ ... ]
> + cfg->reg_vcore = devm_regulator_get_optional(dev, "dvfsrc-vcore");
> + if (IS_ERR(cfg->reg_vcore)) {
> + dev_err(dev, "Failed to get dvfsrc-vcore: %pe", cfg->reg_vcore);
[Severity: Low]
This isn't a bug, but should this error message include a trailing newline?
Without it, the next printed kernel message might be appended to the same
line, resulting in garbled logs.
> + return;
> }
>
> - if (of_property_read_u32(dev->of_node, "boost-crypt-vcore-min",
> - &volt)) {
> - dev_info(dev, "failed to get boost-crypt-vcore-min");
> - goto disable_caps;
> + ret = of_property_read_u32(dev->of_node, "mediatek,boost-crypt-vcore-min",
> + &cfg->vcore_volt);
[Severity: Low]
Does passing &cfg->vcore_volt to of_property_read_u32() cause a compiler
warning here?
Looking at struct ufs_mtk_crypt_cfg, vcore_volt is declared as an int, but
of_property_read_u32() expects a pointer to a u32. This might trigger an
incompatible pointer type or pointer sign warning on strict build
configurations.
> + if (ret) {
> + dev_err(dev, "Failed to get mediatek,boost-crypt-vcore-min: %pe\n",
> + ERR_PTR(ret));
> + return;
> }
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=9
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 10/26] scsi: ufs: mediatek: Handle misc host voltage regulators
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (8 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 09/26] scsi: ufs: mediatek: Rework the crypt-boost stuff Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:29 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 11/26] scsi: ufs: mediatek: Remove undocumented downstream reset cruft Louis-Alexis Eyraud
` (15 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
MediaTek SoCs handled by this driver contain a per-SoC specific set of
miscellaneous supplies. These feed parts of the UFS controller silicon
inside the SoC, as opposed to the UFS card.
Add the necessary driver code to acquire these supplies using the
regulator bulk API. They should be kept on during suspend, so enable
them when acquiring.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 37 ++++++++++++++++++++++++++++++++++---
1 file changed, 34 insertions(+), 3 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 55701172bcf1..65036c839031 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -40,6 +40,8 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up);
struct ufs_mtk_soc_data {
bool has_avdd09;
+ u8 num_reg_names;
+ const char *const *reg_names;
};
static const struct ufs_dev_quirk ufs_mtk_dev_fixups[] = {
@@ -1188,8 +1190,21 @@ static int ufs_mtk_get_supplies(struct ufs_mtk_host *host)
{
struct device *dev = host->hba->dev;
const struct ufs_mtk_soc_data *data = of_device_get_match_data(dev);
+ int ret;
+
+ if (!data)
+ return 0;
+
+ if (data->num_reg_names) {
+ ret = devm_regulator_bulk_get_enable(dev, data->num_reg_names,
+ data->reg_names);
+ if (ret) {
+ dev_err(dev, "Failed to get misc regulators: %pe\n", ERR_PTR(ret));
+ return ret;
+ }
+ }
- if (!data || !data->has_avdd09)
+ if (!data->has_avdd09)
return 0;
host->reg_avdd09 = devm_regulator_get_optional(dev, "avdd09");
@@ -2350,14 +2365,30 @@ static const struct ufs_hba_variant_ops ufs_hba_mtk_vops = {
.config_scsi_dev = ufs_mtk_config_scsi_dev,
};
+static const char *const ufs_mtk_regs_avdd12_avdd18[] = {
+ "avdd12", "avdd18"
+};
+
+static const char *const ufs_mtk_regs_avdd12_ckbuf_avdd18[] = {
+ "avdd12", "avdd12-ckbuf", "avdd18"
+};
+
static const struct ufs_mtk_soc_data mt8183_data = {
.has_avdd09 = true,
+ .reg_names = ufs_mtk_regs_avdd12_avdd18,
+ .num_reg_names = ARRAY_SIZE(ufs_mtk_regs_avdd12_avdd18),
+};
+
+static const struct ufs_mtk_soc_data mt8192_8195_data = {
+ .has_avdd09 = false,
+ .reg_names = ufs_mtk_regs_avdd12_ckbuf_avdd18,
+ .num_reg_names = ARRAY_SIZE(ufs_mtk_regs_avdd12_ckbuf_avdd18),
};
static const struct of_device_id ufs_mtk_of_match[] = {
{ .compatible = "mediatek,mt8183-ufshci", .data = &mt8183_data },
- { .compatible = "mediatek,mt8192-ufshci" },
- { .compatible = "mediatek,mt8195-ufshci" },
+ { .compatible = "mediatek,mt8192-ufshci", .data = &mt8192_8195_data },
+ { .compatible = "mediatek,mt8195-ufshci", .data = &mt8192_8195_data },
{},
};
MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 10/26] scsi: ufs: mediatek: Handle misc host voltage regulators
2026-07-30 9:49 ` [PATCH v10 10/26] scsi: ufs: mediatek: Handle misc host voltage regulators Louis-Alexis Eyraud
@ 2026-07-30 10:29 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:29 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: devicetree, linux-phy, linux-scsi, conor+dt, vkoul,
neil.armstrong, robh, olteanv
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
Pre-existing issues:
- [High] The UFS driver caches a dangling pointer to the PHY device after prematurely dropping its reference, leading to a Use-After-Free.
- [High] The driver ignores the return value of `ufs_mtk_mphy_power_on()`, potentially leading to hardware hangs if the PHY fails to power on.
--
--- Patch [10]: [PATCH v10 10/26] scsi: ufs: mediatek: Handle misc host voltage regulators ---
commit 8242d1302e10cd1046e43a3ff8124f21223143fe
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Handle misc host voltage regulators
This commit adds support for acquiring miscellaneous host voltage regulators
using the regulator bulk API in the MediaTek UFS driver. It handles per-SoC
specific regulator names and enables them during acquisition so they are
kept on during suspend.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index 55701172bcf15..65036c8390318 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
> @@ -1188,8 +1190,21 @@ static int ufs_mtk_get_supplies(struct ufs_mtk_host *host)
> {
> struct device *dev = host->hba->dev;
> const struct ufs_mtk_soc_data *data = of_device_get_match_data(dev);
> + int ret;
> +
> + if (!data)
> + return 0;
> +
> + if (data->num_reg_names) {
> + ret = devm_regulator_bulk_get_enable(dev, data->num_reg_names,
> + data->reg_names);
> + if (ret) {
> + dev_err(dev, "Failed to get misc regulators: %pe\n", ERR_PTR(ret));
> + return ret;
> + }
> + }
>
> - if (!data || !data->has_avdd09)
> + if (!data->has_avdd09)
> return 0;
>
> host->reg_avdd09 = devm_regulator_get_optional(dev, "avdd09");
[ ... ]
[Severity: High]
This isn't a bug introduced by this patch, but does this driver ignore the
return value of ufs_mtk_mphy_power_on() when initializing?
In ufs_mtk_init(), the function is called without checking if it fails:
drivers/ufs/host/ufs-mediatek.c:ufs_mtk_init() {
...
/*
* ufshcd_vops_init() is invoked after
* ufshcd_setup_clock(true) in ufshcd_hba_init() thus
* phy clock setup is skipped.
*
* Enable phy clocks specifically here.
*/
ufs_mtk_mphy_power_on(hba, true);
if (ufs_mtk_is_rtff_mtcmos(hba)) {
...
}
If the underlying regulator enablement fails, the PHY could remain unpowered.
Could this cause the UFS host controller driver to assume the PHY is
functional, which might lead to timeouts or unrecoverable hardware hangs
during initialization?
[Severity: High]
This is also a pre-existing issue, but does this driver cache a dangling
pointer to the PHY device?
Looking at ufs_mtk_probe(), the driver acquires a device reference to the
PHY node but immediately drops it:
drivers/ufs/host/ufs-mediatek.c:ufs_mtk_probe() {
...
if (phy_node) {
phy_dev = of_find_device_by_node(phy_node);
of_node_put(phy_node);
if (!phy_dev) {
dev_err(dev, "phy_dev not found\n");
goto skip_phy;
}
put_device(phy_dev);
dev_info(dev, "phys node found\n");
}
...
skip_phy:
...
if (phy_node && phy_dev) {
host = ufshcd_get_variant(hba);
host->phy_dev = phy_dev;
}
...
}
Because put_device() is called earlier, phy_dev could potentially be freed if
the PHY device is unbound.
If the PHY device is later freed, will accessing pm_runtime_put_sync() or
pm_runtime_get_sync() on host->phy_dev during suspend and resume callbacks
trigger a use-after-free?
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=10
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 11/26] scsi: ufs: mediatek: Remove undocumented downstream reset cruft
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (9 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 10/26] scsi: ufs: mediatek: Handle misc host voltage regulators Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:23 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 12/26] scsi: ufs: mediatek: Remove vendor kernel quirks cruft Louis-Alexis Eyraud
` (14 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The MediaTek UFS host driver's probe function allows using a
ti,syscon-reset as a reset, without going through the appropriate
abstractions, or by documenting this in the binding at all.
Remove this, it's downstream code and does not belong here.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 33 +++------------------------------
1 file changed, 3 insertions(+), 30 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 65036c839031..a2749e79514c 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -2402,38 +2402,12 @@ MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
static int ufs_mtk_probe(struct platform_device *pdev)
{
int err;
- struct device *dev = &pdev->dev, *phy_dev = NULL;
- struct device_node *reset_node, *phy_node = NULL;
- struct platform_device *reset_pdev, *phy_pdev = NULL;
- struct device_link *link;
struct ufs_hba *hba;
+ struct platform_device *phy_pdev = NULL;
+ struct device *dev = &pdev->dev, *phy_dev = NULL;
+ struct device_node *phy_node = NULL;
struct ufs_mtk_host *host;
- reset_node = of_find_compatible_node(NULL, NULL,
- "ti,syscon-reset");
- if (!reset_node) {
- dev_notice(dev, "find ti,syscon-reset fail\n");
- goto skip_reset;
- }
- reset_pdev = of_find_device_by_node(reset_node);
- if (!reset_pdev) {
- dev_notice(dev, "find reset_pdev fail\n");
- goto skip_reset;
- }
- link = device_link_add(dev, &reset_pdev->dev,
- DL_FLAG_AUTOPROBE_CONSUMER);
- put_device(&reset_pdev->dev);
- if (!link) {
- dev_notice(dev, "add reset device_link fail\n");
- goto skip_reset;
- }
- /* supplier is not probed */
- if (link->status == DL_STATE_DORMANT) {
- err = -EPROBE_DEFER;
- goto out;
- }
-
-skip_reset:
/* find phy node */
phy_node = of_parse_phandle(dev->of_node, "phys", 0);
@@ -2479,7 +2453,6 @@ static int ufs_mtk_probe(struct platform_device *pdev)
out:
of_node_put(phy_node);
- of_node_put(reset_node);
return err;
}
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 11/26] scsi: ufs: mediatek: Remove undocumented downstream reset cruft
2026-07-30 9:49 ` [PATCH v10 11/26] scsi: ufs: mediatek: Remove undocumented downstream reset cruft Louis-Alexis Eyraud
@ 2026-07-30 10:23 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:23 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: robh, neil.armstrong, olteanv, devicetree, linux-scsi, linux-phy,
vkoul, conor+dt
Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
Pre-existing issues:
- [High] The kobject reference for `phy_dev` is dropped too early during probe, leaving a dangling pointer stored in `host->phy_dev` that leads to a use-after-free in runtime PM callbacks.
- [High] The runtime PM usage count and enablement state for `phy_dev` are permanently leaked if the probe fails, and the PM state is never restored on driver removal.
--
--- Patch [11]: [PATCH v10 11/26] scsi: ufs: mediatek: Remove undocumented downstream reset cruft ---
commit 2e47ec1b9626d31e1288701fe0c4a25c195e38d4
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Remove undocumented downstream reset cruft
The MediaTek UFS host driver's probe function allows using a ti,syscon-reset
as a reset, without going through the appropriate abstractions or documenting
this in the binding. This commit removes this downstream code as it does not
belong here.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index 65036c8390318..a2749e79514c4 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
> @@ -2402,38 +2402,12 @@ MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
> static int ufs_mtk_probe(struct platform_device *pdev)
> {
> int err;
> - struct device *dev = &pdev->dev, *phy_dev = NULL;
> - struct device_node *reset_node, *phy_node = NULL;
> - struct platform_device *reset_pdev, *phy_pdev = NULL;
> - struct device_link *link;
> struct ufs_hba *hba;
> + struct platform_device *phy_pdev = NULL;
> + struct device *dev = &pdev->dev, *phy_dev = NULL;
> + struct device_node *phy_node = NULL;
> struct ufs_mtk_host *host;
>
> - reset_node = of_find_compatible_node(NULL, NULL,
> - "ti,syscon-reset");
> - if (!reset_node) {
> - dev_notice(dev, "find ti,syscon-reset fail\n");
> - goto skip_reset;
> - }
> - reset_pdev = of_find_device_by_node(reset_node);
> - if (!reset_pdev) {
> - dev_notice(dev, "find reset_pdev fail\n");
> - goto skip_reset;
> - }
> - link = device_link_add(dev, &reset_pdev->dev,
> - DL_FLAG_AUTOPROBE_CONSUMER);
> - put_device(&reset_pdev->dev);
> - if (!link) {
> - dev_notice(dev, "add reset device_link fail\n");
> - goto skip_reset;
> - }
> - /* supplier is not probed */
> - if (link->status == DL_STATE_DORMANT) {
> - err = -EPROBE_DEFER;
> - goto out;
> - }
> -
> -skip_reset:
> /* find phy node */
> phy_node = of_parse_phandle(dev->of_node, "phys", 0);
>
> @@ -2479,7 +2453,6 @@ static int ufs_mtk_probe(struct platform_device *pdev)
>
> out:
> of_node_put(phy_node);
> - of_node_put(reset_node);
> return err;
> }
[Severity: High]
This isn't a bug introduced by this patch, but could this lead to a
use-after-free of phy_dev?
In ufs_mtk_probe(), after looking up the PHY device, the kobject reference
is dropped almost immediately:
drivers/ufs/host/ufs-mediatek.c:ufs_mtk_probe() {
phy_pdev = of_find_device_by_node(phy_node);
if (!phy_pdev)
goto skip_phy;
phy_dev = &phy_pdev->dev;
...
put_device(phy_dev);
...
}
However, the pointer is later saved into the host structure for later use:
drivers/ufs/host/ufs-mediatek.c:ufs_mtk_probe() {
...
if (phy_node && phy_dev) {
host = ufshcd_get_variant(hba);
host->phy_dev = phy_dev;
}
...
}
If the PHY device is later unbound via sysfs and its memory freed, would the
UFS host's runtime PM callbacks, such as ufs_mtk_runtime_suspend(),
dereference this dangling host->phy_dev pointer when calling
pm_runtime_put_sync()?
[Severity: High]
This is also a pre-existing issue, but is there a PM runtime usage count
leak in ufs_mtk_probe() on error paths?
If a failure occurs during initialization (e.g., ufshcd_pltfrm_init() fails),
the function jumps to the out label without calling pm_runtime_put_sync() or
pm_runtime_disable() for phy_dev:
drivers/ufs/host/ufs-mediatek.c:ufs_mtk_probe() {
if (phy_node) {
...
pm_runtime_enable(phy_dev);
pm_runtime_get_sync(phy_dev);
...
}
...
err = ufshcd_pltfrm_init(pdev, &ufs_hba_mtk_vops);
if (err)
goto out;
...
out:
of_node_put(phy_node);
return err;
}
Additionally, ufs_mtk_remove() doesn't appear to tear down this state.
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=11
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 12/26] scsi: ufs: mediatek: Remove vendor kernel quirks cruft
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (10 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 11/26] scsi: ufs: mediatek: Remove undocumented downstream reset cruft Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:30 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 13/26] scsi: ufs: mediatek: Use the common PHY framework Louis-Alexis Eyraud
` (13 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Krzysztof Kozlowski,
Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Both ufs_mtk_vreg_fix_vcc and ufs_mtk_vreg_fix_vccqx look like they are
vendor kernel hacks to work around existing downstream device trees.
Mainline does not need or want them, so remove them.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski@oss.qualcomm.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 69 -----------------------------------------
1 file changed, 69 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index a2749e79514c..e5527969a5e6 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -1017,73 +1017,6 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
}
}
-#define MAX_VCC_NAME 30
-static int ufs_mtk_vreg_fix_vcc(struct ufs_hba *hba)
-{
- struct ufs_vreg_info *info = &hba->vreg_info;
- struct device_node *np = hba->dev->of_node;
- struct device *dev = hba->dev;
- char vcc_name[MAX_VCC_NAME];
- struct arm_smccc_res res;
- int err, ver;
-
- if (info->vcc)
- return 0;
-
- if (of_property_read_bool(np, "mediatek,ufs-vcc-by-num")) {
- ufs_mtk_get_vcc_num(res);
- if (res.a1 > UFS_VCC_NONE && res.a1 < UFS_VCC_MAX)
- snprintf(vcc_name, MAX_VCC_NAME, "vcc-opt%lu", res.a1);
- else
- return -ENODEV;
- } else if (of_property_read_bool(np, "mediatek,ufs-vcc-by-ver")) {
- ver = (hba->dev_info.wspecversion & 0xF00) >> 8;
- snprintf(vcc_name, MAX_VCC_NAME, "vcc-ufs%u", ver);
- } else {
- return 0;
- }
-
- err = ufshcd_populate_vreg(dev, vcc_name, &info->vcc, false);
- if (err)
- return err;
-
- err = ufshcd_get_vreg(dev, info->vcc);
- if (err)
- return err;
-
- err = regulator_enable(info->vcc->reg);
- if (!err) {
- info->vcc->enabled = true;
- dev_info(dev, "%s: %s enabled\n", __func__, vcc_name);
- }
-
- return err;
-}
-
-static void ufs_mtk_vreg_fix_vccqx(struct ufs_hba *hba)
-{
- struct ufs_vreg_info *info = &hba->vreg_info;
- struct ufs_vreg **vreg_on, **vreg_off;
-
- if (hba->dev_info.wspecversion >= 0x0300) {
- vreg_on = &info->vccq;
- vreg_off = &info->vccq2;
- } else {
- vreg_on = &info->vccq2;
- vreg_off = &info->vccq;
- }
-
- if (*vreg_on)
- (*vreg_on)->always_on = true;
-
- if (*vreg_off) {
- regulator_disable((*vreg_off)->reg);
- devm_kfree(hba->dev, (*vreg_off)->name);
- devm_kfree(hba->dev, *vreg_off);
- *vreg_off = NULL;
- }
-}
-
static void ufs_mtk_setup_clk_gating(struct ufs_hba *hba)
{
unsigned long flags;
@@ -1991,8 +1924,6 @@ static void ufs_mtk_fixup_dev_quirks(struct ufs_hba *hba)
host->ip_ver == IP_VER_MT6995_B0))
hba->quirks |= UFSHCD_QUIRK_VCC_ON_DELAY;
- ufs_mtk_vreg_fix_vcc(hba);
- ufs_mtk_vreg_fix_vccqx(hba);
ufs_mtk_fix_ahit(hba);
ufs_mtk_fix_clock_scaling(hba);
}
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 12/26] scsi: ufs: mediatek: Remove vendor kernel quirks cruft
2026-07-30 9:49 ` [PATCH v10 12/26] scsi: ufs: mediatek: Remove vendor kernel quirks cruft Louis-Alexis Eyraud
@ 2026-07-30 10:30 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:30 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: linux-scsi, neil.armstrong, devicetree, linux-phy, robh, olteanv,
vkoul, conor+dt
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] Dangling pointer to `phy_dev` in `ufs_mtk_probe()` due to premature `put_device()`.
--
--- Patch [12]: [PATCH v10 12/26] scsi: ufs: mediatek: Remove vendor kernel quirks cruft ---
commit 6a21968869510eaa0eea1b9f7b65e382357ae809
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Remove vendor kernel quirks cruft
Both ufs_mtk_vreg_fix_vcc and ufs_mtk_vreg_fix_vccqx look like they are
vendor kernel hacks to work around existing downstream device trees.
Mainline does not need or want them, so remove them.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index a2749e79514c4..e5527969a5e67 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
[Severity: High]
This isn't a bug introduced by this patch, but does dropping the kobject
reference to phy_dev with put_device() leave host->phy_dev as a dangling
pointer?
drivers/ufs/host/ufs-mediatek.c:ufs_mtk_probe() {
...
pm_runtime_set_active(phy_dev);
pm_runtime_enable(phy_dev);
pm_runtime_get_sync(phy_dev);
put_device(phy_dev);
dev_info(dev, "phys node found\n");
...
if (phy_node && phy_dev) {
host = ufshcd_get_variant(hba);
host->phy_dev = phy_dev;
}
...
}
If the PHY platform device is unregistered (for example, via DT overlay removal
or manual sysfs unbind) between the UFS probe and runtime PM transitions,
would dereferencing this pointer later in ufs_mtk_runtime_suspend() or
ufs_mtk_runtime_resume() cause a use-after-free?
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=12
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 13/26] scsi: ufs: mediatek: Use the common PHY framework
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (11 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 12/26] scsi: ufs: mediatek: Remove vendor kernel quirks cruft Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:40 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 14/26] scsi: ufs: mediatek: Remove mediatek,ufs-broken-rtc property Louis-Alexis Eyraud
` (12 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
There is no need to reinvent the PHY framework, especially not its OF
parsing.
Change the code to simply use the PHY framework to acquire the device's
PHY in the ufshcd init, so that it's device linked to the right device.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 134 +++++++++++++---------------------------
drivers/ufs/host/ufs-mediatek.h | 1 -
2 files changed, 42 insertions(+), 93 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index e5527969a5e6..385f0180b230 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -293,44 +293,6 @@ static int ufs_mtk_hce_enable_notify(struct ufs_hba *hba,
return 0;
}
-static int ufs_mtk_bind_mphy(struct ufs_hba *hba)
-{
- struct ufs_mtk_host *host = ufshcd_get_variant(hba);
- struct device *dev = hba->dev;
- struct device_node *np = dev->of_node;
- int err = 0;
-
- host->mphy = devm_of_phy_get_by_index(dev, np, 0);
-
- if (host->mphy == ERR_PTR(-EPROBE_DEFER)) {
- /*
- * UFS driver might be probed before the phy driver does.
- * In that case we would like to return EPROBE_DEFER code.
- */
- err = -EPROBE_DEFER;
- dev_info(dev,
- "%s: required phy hasn't probed yet. err = %d\n",
- __func__, err);
- } else if (IS_ERR(host->mphy)) {
- err = PTR_ERR(host->mphy);
- if (err != -ENODEV) {
- dev_info(dev, "%s: PHY get failed %d\n", __func__,
- err);
- }
- }
-
- if (err)
- host->mphy = NULL;
- /*
- * Allow unbound mphy because not every platform needs specific
- * mphy control.
- */
- if (err == -ENODEV)
- err = 0;
-
- return err;
-}
-
static int ufs_mtk_setup_ref_clk(struct ufs_hba *hba, bool on)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
@@ -1185,13 +1147,21 @@ static int ufs_mtk_init(struct ufs_hba *hba)
ufs_mtk_init_mcq_irq(hba);
- err = ufs_mtk_bind_mphy(hba);
- if (err)
+ host->mphy = devm_phy_get(dev, NULL);
+ if (IS_ERR(host->mphy)) {
+ err = dev_err_probe(dev, PTR_ERR(host->mphy), "Failed to get PHY\n");
+ goto out_variant_clear;
+ }
+
+ err = phy_init(host->mphy);
+ if (err) {
+ dev_err_probe(dev, err, "Failed to initialize PHY\n");
goto out_variant_clear;
+ }
err = ufs_mtk_init_reset(hba);
if (err)
- goto out_variant_clear;
+ goto out_phy_exit;
/* Enable runtime autosuspend */
hba->caps |= UFSHCD_CAP_RPM_AUTOSUSPEND;
@@ -1230,7 +1200,7 @@ static int ufs_mtk_init(struct ufs_hba *hba)
err = ufs_mtk_get_supplies(host);
if (err)
- goto out_variant_clear;
+ goto out_phy_exit;
/*
* ufshcd_vops_init() is invoked after
@@ -1255,11 +1225,22 @@ static int ufs_mtk_init(struct ufs_hba *hba)
return 0;
+out_phy_exit:
+ phy_exit(host->mphy);
out_variant_clear:
ufshcd_set_variant(hba, NULL);
return err;
}
+static void ufs_mtk_exit(struct ufs_hba *hba)
+{
+ struct ufs_mtk_host *host = ufshcd_get_variant(hba);
+
+ ufs_mtk_mphy_power_on(hba, false);
+
+ phy_exit(host->mphy);
+}
+
static int ufs_mtk_negotiate_pwr_mode(struct ufs_hba *hba,
const struct ufs_pa_layer_attr *dev_max_params,
struct ufs_pa_layer_attr *dev_req_params)
@@ -2272,6 +2253,7 @@ static void ufs_mtk_config_scsi_dev(struct scsi_device *sdev)
static const struct ufs_hba_variant_ops ufs_hba_mtk_vops = {
.name = "mediatek.ufshci",
.init = ufs_mtk_init,
+ .exit = ufs_mtk_exit,
.get_ufs_hci_version = ufs_mtk_get_ufs_hci_version,
.setup_clocks = ufs_mtk_setup_clocks,
.hce_enable_notify = ufs_mtk_hce_enable_notify,
@@ -2332,48 +2314,15 @@ MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
*/
static int ufs_mtk_probe(struct platform_device *pdev)
{
- int err;
+ struct device *dev = &pdev->dev;
struct ufs_hba *hba;
- struct platform_device *phy_pdev = NULL;
- struct device *dev = &pdev->dev, *phy_dev = NULL;
- struct device_node *phy_node = NULL;
- struct ufs_mtk_host *host;
-
- /* find phy node */
- phy_node = of_parse_phandle(dev->of_node, "phys", 0);
-
- if (phy_node) {
- phy_pdev = of_find_device_by_node(phy_node);
- if (!phy_pdev)
- goto skip_phy;
- phy_dev = &phy_pdev->dev;
-
- pm_runtime_set_active(phy_dev);
- pm_runtime_enable(phy_dev);
- pm_runtime_get_sync(phy_dev);
-
- put_device(phy_dev);
- dev_info(dev, "phys node found\n");
- } else {
- dev_notice(dev, "phys node not found\n");
- }
+ int ret;
-skip_phy:
- /* perform generic probe */
- err = ufshcd_pltfrm_init(pdev, &ufs_hba_mtk_vops);
- if (err) {
- dev_err(dev, "probe failed %d\n", err);
- goto out;
- }
+ ret = ufshcd_pltfrm_init(pdev, &ufs_hba_mtk_vops);
+ if (ret)
+ return dev_err_probe(dev, ret, "Generic platform probe failed\n");
hba = platform_get_drvdata(pdev);
- if (!hba)
- goto out;
-
- if (phy_node && phy_dev) {
- host = ufshcd_get_variant(hba);
- host->phy_dev = phy_dev;
- }
/*
* Because the default power setting of VSx (the upper layer of
@@ -2382,16 +2331,12 @@ static int ufs_mtk_probe(struct platform_device *pdev)
*/
ufs_mtk_dev_vreg_set_lpm(hba, false);
-out:
- of_node_put(phy_node);
- return err;
+ return 0;
}
/**
* ufs_mtk_remove - set driver_data of the device to NULL
* @pdev: pointer to platform device handle
- *
- * Always return 0
*/
static void ufs_mtk_remove(struct platform_device *pdev)
{
@@ -2448,9 +2393,8 @@ static int ufs_mtk_system_resume(struct device *dev)
static int ufs_mtk_runtime_suspend(struct device *dev)
{
struct ufs_hba *hba = dev_get_drvdata(dev);
- struct ufs_mtk_host *host = ufshcd_get_variant(hba);
struct arm_smccc_res res;
- int ret = 0;
+ int ret;
ret = ufshcd_runtime_suspend(dev);
if (ret)
@@ -2461,8 +2405,11 @@ static int ufs_mtk_runtime_suspend(struct device *dev)
if (ufs_mtk_is_rtff_mtcmos(hba))
ufs_mtk_mtcmos_ctrl(false, res);
- if (host->phy_dev)
- pm_runtime_put_sync(host->phy_dev);
+ ret = ufs_mtk_mphy_power_on(hba, false);
+ if (ret) {
+ dev_err(dev, "Failed to power off PHY: %pe\n", ERR_PTR(ret));
+ return ret;
+ }
return 0;
}
@@ -2470,14 +2417,17 @@ static int ufs_mtk_runtime_suspend(struct device *dev)
static int ufs_mtk_runtime_resume(struct device *dev)
{
struct ufs_hba *hba = dev_get_drvdata(dev);
- struct ufs_mtk_host *host = ufshcd_get_variant(hba);
struct arm_smccc_res res;
+ int ret;
if (ufs_mtk_is_rtff_mtcmos(hba))
ufs_mtk_mtcmos_ctrl(true, res);
- if (host->phy_dev)
- pm_runtime_get_sync(host->phy_dev);
+ ret = ufs_mtk_mphy_power_on(hba, true);
+ if (ret) {
+ dev_err(dev, "Failed to power on PHY: %pe\n", ERR_PTR(ret));
+ return ret;
+ }
ufs_mtk_dev_vreg_set_lpm(hba, false);
diff --git a/drivers/ufs/host/ufs-mediatek.h b/drivers/ufs/host/ufs-mediatek.h
index c63740b98e85..8b3cd17e6227 100644
--- a/drivers/ufs/host/ufs-mediatek.h
+++ b/drivers/ufs/host/ufs-mediatek.h
@@ -195,7 +195,6 @@ struct ufs_mtk_host {
bool is_mcq_intr_enabled;
int mcq_nr_intr;
struct ufs_mtk_mcq_intr_info mcq_intr_info[UFSHCD_MAX_Q_NR];
- struct device *phy_dev;
};
/* MTK delay of autosuspend: 500 ms */
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 13/26] scsi: ufs: mediatek: Use the common PHY framework
2026-07-30 9:49 ` [PATCH v10 13/26] scsi: ufs: mediatek: Use the common PHY framework Louis-Alexis Eyraud
@ 2026-07-30 10:40 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:40 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: linux-phy, neil.armstrong, linux-scsi, conor+dt, robh, vkoul,
olteanv, devicetree
Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
New issues:
- [High] The driver now fails to probe on platforms where the UFS PHY is not defined in the device tree, breaking backward compatibility.
- [High] Returning an error late in PM runtime callbacks leaves the device in an inconsistent state and breaks PM state synchronization.
Pre-existing issues:
- [High] The return values of `phy_power_on()` and `phy_power_off()` are ignored, which can lead to unbalanced PM runtime counts and hardware state corruption.
--
--- Patch [13]: [PATCH v10 13/26] scsi: ufs: mediatek: Use the common PHY framework ---
commit 114da5212062d5f11b7f9274ba9701891ef82788
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Use the common PHY framework
This patch refactors the MediaTek UFS driver to use the common PHY framework
to acquire the device's PHY during ufshcd init, ensuring it is device-linked
to the correct device.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index e5527969a5e67..385f0180b2309 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
> @@ -1185,13 +1147,21 @@ static int ufs_mtk_init(struct ufs_hba *hba)
>
> ufs_mtk_init_mcq_irq(hba);
>
> - err = ufs_mtk_bind_mphy(hba);
> - if (err)
> + host->mphy = devm_phy_get(dev, NULL);
> + if (IS_ERR(host->mphy)) {
> + err = dev_err_probe(dev, PTR_ERR(host->mphy), "Failed to get PHY\n");
[Severity: High]
Will this cause the driver probe to fail on platforms where the UFS PHY
is not defined in the device tree?
The original ufs_mtk_bind_mphy() tolerated -ENODEV to allow for unbound
mphys since not all platforms require specific mphy control. By switching
to devm_phy_get() instead of devm_phy_optional_get(), won't this return
-ENODEV and fail the probe if the "phys" property is missing, breaking
backward compatibility with older device trees?
> + goto out_variant_clear;
> + }
> +
> + err = phy_init(host->mphy);
> + if (err) {
> + dev_err_probe(dev, err, "Failed to initialize PHY\n");
> goto out_variant_clear;
> + }
[ ... ]
> @@ -2448,9 +2393,8 @@ static int ufs_mtk_system_resume(struct device *dev)
> static int ufs_mtk_runtime_suspend(struct device *dev)
> {
> struct ufs_hba *hba = dev_get_drvdata(dev);
> - struct ufs_mtk_host *host = ufshcd_get_variant(hba);
> struct arm_smccc_res res;
> - int ret = 0;
> + int ret;
>
> ret = ufshcd_runtime_suspend(dev);
> if (ret)
> @@ -2461,8 +2405,11 @@ static int ufs_mtk_runtime_suspend(struct device *dev)
> if (ufs_mtk_is_rtff_mtcmos(hba))
> ufs_mtk_mtcmos_ctrl(false, res);
>
> - if (host->phy_dev)
> - pm_runtime_put_sync(host->phy_dev);
> + ret = ufs_mtk_mphy_power_on(hba, false);
> + if (ret) {
> + dev_err(dev, "Failed to power off PHY: %pe\n", ERR_PTR(ret));
> + return ret;
> + }
[Severity: High]
Does returning an error here leave the device in an inconsistent state?
If ufshcd_runtime_suspend() succeeds, the main UFS controller hardware is
fully suspended (clocks gated, link hibern8). If ufs_mtk_mphy_power_on()
fails afterward, this returns the error to the PM core.
Since the PM core intercepts the non-zero return as a runtime suspend
failure, it leaves the software state as RPM_ACTIVE. If the software
thinks the device is active while the hardware is actually suspended,
won't the next hardware access skip runtime_resume and cause a system
hang or bus error?
[Severity: High]
This is a pre-existing issue, but does ufs_mtk_mphy_power_on() safely
handle phy_power_on() failures?
Looking at ufs_mtk_mphy_power_on() in drivers/ufs/host/ufs-mediatek.c:
static int ufs_mtk_mphy_power_on(struct ufs_hba *hba, bool on)
{
...
if (on) {
ret = ufs_mtk_09v_on(host);
if (ret)
return ret;
phy_power_on(mphy);
} else {
phy_power_off(mphy);
...
host->mphy_powered_on = on;
return 0;
}
If phy_power_on() fails, it cleans up its runtime PM and does not
increment the power_count, but ufs_mtk_mphy_power_on() ignores the return
value and still sets host->mphy_powered_on = true.
When ufs_mtk_mphy_power_on(hba, false) is called later, it will blindly
call phy_power_off(). Won't this decrement the PHY's power_count
(potentially underflowing it), corrupting the PHY driver's PM usage
counter and causing an unbalanced PM state?
>
> return 0;
> }
> @@ -2470,14 +2417,17 @@ static int ufs_mtk_runtime_suspend(struct device *dev)
> static int ufs_mtk_runtime_resume(struct device *dev)
> {
> struct ufs_hba *hba = dev_get_drvdata(dev);
> - struct ufs_mtk_host *host = ufshcd_get_variant(hba);
> struct arm_smccc_res res;
> + int ret;
>
> if (ufs_mtk_is_rtff_mtcmos(hba))
> ufs_mtk_mtcmos_ctrl(true, res);
>
> - if (host->phy_dev)
> - pm_runtime_get_sync(host->phy_dev);
> + ret = ufs_mtk_mphy_power_on(hba, true);
> + if (ret) {
> + dev_err(dev, "Failed to power on PHY: %pe\n", ERR_PTR(ret));
> + return ret;
> + }
[Severity: High]
Could a failure here cause similar PM state inconsistencies?
If ufs_mtk_mphy_power_on() fails during runtime resume, returning an
error without rolling back the prior partial initialization (such as
ufs_mtk_mtcmos_ctrl) might confuse the PM core. Does this need to clean
up previously resumed components before returning?
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=13
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 14/26] scsi: ufs: mediatek: Remove mediatek,ufs-broken-rtc property
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (12 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 13/26] scsi: ufs: mediatek: Use the common PHY framework Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 15/26] scsi: ufs: mediatek: Rework _ufs_mtk_clk_scale error paths Louis-Alexis Eyraud
` (11 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
This flag property was never described in the binding, and its
capability wrapper seems pointless.
If one of the MediaTek SoCs needs the ufshcd quirk applied, then this
can be done per-compatible, without needing to give the device tree
author the option to forget to set it.
Remove it and the associated capability flag wrapping code.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 5 -----
drivers/ufs/host/ufs-mediatek.h | 2 --
2 files changed, 7 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 385f0180b230..b4867c9c2743 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -655,9 +655,6 @@ static void ufs_mtk_init_host_caps(struct ufs_hba *hba)
if (of_property_read_bool(np, "mediatek,ufs-rtff-mtcmos"))
host->caps |= UFS_MTK_CAP_RTFF_MTCMOS;
- if (of_property_read_bool(np, "mediatek,ufs-broken-rtc"))
- host->caps |= UFS_MTK_CAP_MCQ_BROKEN_RTC;
-
dev_info(hba->dev, "caps: 0x%x", host->caps);
}
@@ -1185,8 +1182,6 @@ static int ufs_mtk_init(struct ufs_hba *hba)
hba->quirks |= UFSHCI_QUIRK_SKIP_MANUAL_WB_FLUSH_CTRL;
hba->quirks |= UFSHCD_QUIRK_MCQ_BROKEN_INTR;
- if (host->caps & UFS_MTK_CAP_MCQ_BROKEN_RTC)
- hba->quirks |= UFSHCD_QUIRK_MCQ_BROKEN_RTC;
hba->vps->wb_flush_threshold = UFS_WB_BUF_REMAIN_PERCENT(80);
diff --git a/drivers/ufs/host/ufs-mediatek.h b/drivers/ufs/host/ufs-mediatek.h
index 8b3cd17e6227..d73093900d49 100644
--- a/drivers/ufs/host/ufs-mediatek.h
+++ b/drivers/ufs/host/ufs-mediatek.h
@@ -138,8 +138,6 @@ enum ufs_mtk_host_caps {
UFS_MTK_CAP_DISABLE_MCQ = 1 << 8,
/* Control MTCMOS with RTFF */
UFS_MTK_CAP_RTFF_MTCMOS = 1 << 9,
-
- UFS_MTK_CAP_MCQ_BROKEN_RTC = 1 << 10,
};
struct ufs_mtk_crypt_cfg {
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 15/26] scsi: ufs: mediatek: Rework _ufs_mtk_clk_scale error paths
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (13 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 14/26] scsi: ufs: mediatek: Remove mediatek,ufs-broken-rtc property Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:34 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 16/26] scsi: ufs: mediatek: Clean up logging prints Louis-Alexis Eyraud
` (10 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Errors should be printed at the correct log level. Additionally, it
looks like some "goto out"'s were omitted in the scale up case, which
looks like a mistake, as the scale down branch of the code does use
them.
Rework the error messages to make them nicer and at the correct
verbosity, and add the missing gotos.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 41 +++++++++++++++++++----------------------
1 file changed, 19 insertions(+), 22 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index b4867c9c2743..cc4b5f4c4a07 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -1971,16 +1971,16 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
ret = clk_prepare_enable(clki->clk);
if (ret) {
- dev_info(hba->dev,
- "clk_prepare_enable() fail, ret: %d\n", ret);
+ dev_err(hba->dev, "%s: Failed to enable clock: %pe\n", __func__, ERR_PTR(ret));
return;
}
if (clk_fde_scale) {
ret = clk_prepare_enable(fde_clki->clk);
if (ret) {
- dev_info(hba->dev,
- "fde clk_prepare_enable() fail, ret: %d\n", ret);
+ dev_err(hba->dev, "%s: Failed to enable FDE clock: %pe\n",
+ __func__, ERR_PTR(ret));
+ clk_disable_unprepare(clki->clk);
return;
}
}
@@ -1989,51 +1989,48 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
if (clk_bind_vcore) {
ret = regulator_set_voltage(reg, volt, INT_MAX);
if (ret) {
- dev_info(hba->dev,
- "Failed to set vcore to %d\n", volt);
+ dev_err(hba->dev, "Failed to set vcore to %d\n", volt);
goto out;
}
}
ret = clk_set_parent(clki->clk, mclk->ufs_sel_max_clki->clk);
if (ret) {
- dev_info(hba->dev, "Failed to set clk mux, ret = %d\n",
- ret);
+ dev_err(hba->dev, "%s: Failed to set clock mux: %pe\n",
+ __func__, ERR_PTR(ret));
+ goto out;
}
if (clk_fde_scale) {
- ret = clk_set_parent(fde_clki->clk,
- mclk->ufs_fde_max_clki->clk);
+ ret = clk_set_parent(fde_clki->clk, mclk->ufs_fde_max_clki->clk);
if (ret) {
- dev_info(hba->dev,
- "Failed to set fde clk mux, ret = %d\n",
- ret);
+ dev_err(hba->dev, "%s: Failed to set fde clock mux: %pe\n",
+ __func__, ERR_PTR(ret));
+ goto out;
}
}
} else {
if (clk_fde_scale) {
- ret = clk_set_parent(fde_clki->clk,
- mclk->ufs_fde_min_clki->clk);
+ ret = clk_set_parent(fde_clki->clk, mclk->ufs_fde_min_clki->clk);
if (ret) {
- dev_info(hba->dev,
- "Failed to set fde clk mux, ret = %d\n",
- ret);
+ dev_err(hba->dev, "%s: Failed to set fde clock mux: %pe\n",
+ __func__, ERR_PTR(ret));
goto out;
}
}
ret = clk_set_parent(clki->clk, mclk->ufs_sel_min_clki->clk);
if (ret) {
- dev_info(hba->dev, "Failed to set clk mux, ret = %d\n",
- ret);
+ dev_err(hba->dev, "%s: Failed to set clock mux: %pe\n",
+ __func__, ERR_PTR(ret));
goto out;
}
if (clk_bind_vcore) {
ret = regulator_set_voltage(reg, 0, INT_MAX);
if (ret) {
- dev_info(hba->dev,
- "failed to set vcore to MIN\n");
+ dev_err(hba->dev, "%s: Failed to set vcore to minimum: %pe\n",
+ __func__, ERR_PTR(ret));
}
}
}
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 15/26] scsi: ufs: mediatek: Rework _ufs_mtk_clk_scale error paths
2026-07-30 9:49 ` [PATCH v10 15/26] scsi: ufs: mediatek: Rework _ufs_mtk_clk_scale error paths Louis-Alexis Eyraud
@ 2026-07-30 10:34 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:34 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: vkoul, linux-phy, robh, neil.armstrong, conor+dt, olteanv,
linux-scsi, devicetree
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] Failures during hardware operations in `_ufs_mtk_clk_scale` are not propagated to callers, causing software state tracking and the UFS core to desynchronize from the actual hardware state.
--
--- Patch [15]: [PATCH v10 15/26] scsi: ufs: mediatek: Rework _ufs_mtk_clk_scale error paths ---
commit fc8d96d1151c17bc40e741378caf17dd49c71248
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Rework _ufs_mtk_clk_scale error paths
This commit reworks the error paths in _ufs_mtk_clk_scale to print errors
at the correct log level and adds missing gotos in the scale up case.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index b4867c9c27434..cc4b5f4c4a079 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
> @@ -1989,51 +1989,48 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
> if (clk_bind_vcore) {
> ret = regulator_set_voltage(reg, volt, INT_MAX);
> if (ret) {
> - dev_info(hba->dev,
> - "Failed to set vcore to %d\n", volt);
> + dev_err(hba->dev, "Failed to set vcore to %d\n", volt);
> goto out;
[Severity: High]
This is a pre-existing issue, but since the error paths are being reworked
here, should _ufs_mtk_clk_scale() be changed to return an int so these
errors can be propagated?
Currently, _ufs_mtk_clk_scale() returns void. When hardware operations like
regulator_set_voltage() or clk_set_parent() fail, the function jumps out
but drops the error code.
The caller ufs_mtk_clk_scale() then unconditionally updates the software
tracking state assuming success:
ufs_mtk_clk_scale() {
...
if (scale_up)
_ufs_mtk_clk_scale(hba, true);
else
_ufs_mtk_clk_scale(hba, false);
host->clk_scale_up = scale_up;
/* Must always set before clk_set_rate() */
if (scale_up)
clki->curr_freq = clki->max_freq;
...
}
Could this desynchronize the driver and devfreq state from the hardware,
potentially causing the UFS core to overdrive the wrong clock parent or
fail to scale frequencies properly?
> }
> }
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=15
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 16/26] scsi: ufs: mediatek: Clean up logging prints
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (14 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 15/26] scsi: ufs: mediatek: Rework _ufs_mtk_clk_scale error paths Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 17/26] scsi: ufs: mediatek: Rework ufs_mtk_wait_idle_state Louis-Alexis Eyraud
` (9 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The Linux kernel's log buffer provides many levels of verbosity,
associated with different semantic meanings. Care should be taken to
only log useful information to the info level, and log errors to the
error level.
The MediaTek UFS driver does not do this. It freely logs verbose debug
information to the info level, errors to the info level, and sometimes
errors to the warning level.
Adjust all the wrapped kprintf invocations to rectify this situation.
Use user-friendly %pe format codes for printing errors where possible.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 91 ++++++++++++++++++-----------------------
1 file changed, 40 insertions(+), 51 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index cc4b5f4c4a07..33917c345b29 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -192,8 +192,8 @@ static void ufs_mtk_crypto_enable(struct ufs_hba *hba)
ufs_mtk_crypto_ctrl(res, 1);
if (res.a0) {
- dev_info(hba->dev, "%s: crypto enable failed, err: %lu\n",
- __func__, res.a0);
+ dev_err(hba->dev, "%s: crypto enable failed with error %lu, disabling\n",
+ __func__, res.a0);
hba->caps &= ~UFSHCD_CAP_CRYPTO;
}
}
@@ -542,40 +542,38 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
ret = clk_prepare_enable(cfg->clk_crypt_mux);
if (ret) {
- dev_info(hba->dev, "clk_prepare_enable(): %d\n",
- ret);
+ dev_err(hba->dev, "%s: Failed to enable clk_crypt_mux: %pe\n",
+ __func__, ERR_PTR(ret));
return;
}
if (boost) {
ret = regulator_set_voltage(reg, volt, INT_MAX);
if (ret) {
- dev_info(hba->dev,
- "failed to set vcore to %d\n", volt);
+ dev_err(hba->dev, "%s: Failed to set vcore to %d: %pe\n",
+ __func__, volt, ERR_PTR(ret));
goto out;
}
- ret = clk_set_parent(cfg->clk_crypt_mux,
- cfg->clk_crypt_perf);
+ ret = clk_set_parent(cfg->clk_crypt_mux, cfg->clk_crypt_perf);
if (ret) {
- dev_info(hba->dev,
- "failed to set clk_crypt_perf\n");
+ dev_err(hba->dev, "%s: Failed to reparent clk_crypt_perf: %pe\n",
+ __func__, ERR_PTR(ret));
regulator_set_voltage(reg, 0, INT_MAX);
goto out;
}
} else {
- ret = clk_set_parent(cfg->clk_crypt_mux,
- cfg->clk_crypt_lp);
+ ret = clk_set_parent(cfg->clk_crypt_mux, cfg->clk_crypt_lp);
if (ret) {
- dev_info(hba->dev,
- "failed to set clk_crypt_lp\n");
+ dev_err(hba->dev, "%s: Failed to reparent clk_crypt_lp: %pe\n",
+ __func__, ERR_PTR(ret));
goto out;
}
ret = regulator_set_voltage(reg, 0, INT_MAX);
if (ret) {
- dev_info(hba->dev,
- "failed to set vcore to MIN\n");
+ dev_err(hba->dev, "%s: Failed to set vcore to minimum: %pe\n",
+ __func__, ERR_PTR(ret));
}
}
out:
@@ -763,10 +761,8 @@ static int ufs_mtk_setup_clocks(struct ufs_hba *hba, bool on,
if (clk_pwr_off) {
ufs_mtk_pwr_ctrl(hba, false);
} else {
- dev_warn(hba->dev, "Clock is not turned off, hba->ahit = 0x%x, AHIT = 0x%x\n",
- hba->ahit,
- ufshcd_readl(hba,
- REG_AUTO_HIBERNATE_IDLE_TIMER));
+ dev_warn(hba->dev, "Clock isn't off, hba->ahit = 0x%x, AHIT = 0x%x\n",
+ hba->ahit, ufshcd_readl(hba, REG_AUTO_HIBERNATE_IDLE_TIMER));
}
ufs_mtk_mcq_disable_irq(hba);
} else if (on && status == POST_CHANGE) {
@@ -810,11 +806,11 @@ static void ufs_mtk_mcq_set_irq_affinity(struct ufs_hba *hba, unsigned int cpu)
_cpu = (cpu == 0) ? 3 : cpu;
ret = irq_set_affinity(irq, cpumask_of(_cpu));
if (ret) {
- dev_err(hba->dev, "set irq %d affinity to CPU %d failed\n",
+ dev_err(hba->dev, "setting irq %d affinity to CPU %d failed\n",
irq, _cpu);
return;
}
- dev_info(hba->dev, "set irq %d affinity to CPU: %d\n", irq, _cpu);
+ dev_dbg(hba->dev, "set irq %d affinity to CPU %d\n", irq, _cpu);
}
static bool ufs_mtk_is_legacy_chipset(struct ufs_hba *hba, u32 hw_ip_ver)
@@ -830,7 +826,8 @@ static bool ufs_mtk_is_legacy_chipset(struct ufs_hba *hba, u32 hw_ip_ver)
default:
break;
}
- dev_info(hba->dev, "legacy IP version - 0x%x, is legacy : %d", hw_ip_ver, is_legacy);
+ dev_dbg(hba->dev, "IP version 0x%x, legacy = %s", hw_ip_ver,
+ str_true_false(is_legacy));
return is_legacy;
}
@@ -935,15 +932,12 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
}
}
- list_for_each_entry(clki, head, list) {
- dev_info(hba->dev, "clk \"%s\" present", clki->name);
- }
+ list_for_each_entry(clki, head, list)
+ dev_dbg(hba->dev, "clk \"%s\" present", clki->name);
if (!ufs_mtk_is_clk_scale_ready(hba)) {
hba->caps &= ~UFSHCD_CAP_CLK_SCALING;
- dev_info(hba->dev,
- "%s: Clk-scaling not ready. Feature disabled.",
- __func__);
+ dev_info(hba->dev, "%s: Clock scaling unavailable", __func__);
return;
}
@@ -953,8 +947,8 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
*/
reg = devm_regulator_get_optional(dev, "dvfsrc-vcore");
if (IS_ERR(reg)) {
- dev_info(dev, "failed to get dvfsrc-vcore: %ld",
- PTR_ERR(reg));
+ if (PTR_ERR(reg) != -ENODEV)
+ dev_err(dev, "Failed to get dvfsrc-vcore: %pe\n", reg);
return;
}
@@ -968,12 +962,9 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
host->mclk.vcore_volt = volt;
/* If default boot is max gear, request vcore */
- if (reg && volt && host->clk_scale_up) {
- if (regulator_set_voltage(reg, volt, INT_MAX)) {
- dev_info(hba->dev,
- "Failed to set vcore to %d\n", volt);
- }
- }
+ if (reg && volt && host->clk_scale_up)
+ if (regulator_set_voltage(reg, volt, INT_MAX))
+ dev_err(hba->dev, "Failed to set vcore to %d\n", volt);
}
static void ufs_mtk_setup_clk_gating(struct ufs_hba *hba)
@@ -1060,7 +1051,7 @@ static void ufs_mtk_init_mcq_irq(struct ufs_hba *hba)
}
host->mcq_intr_info[i].hba = hba;
host->mcq_intr_info[i].irq = irq;
- dev_info(hba->dev, "get platform mcq irq: %d, %d\n", i, irq);
+ dev_dbg(hba->dev, "get platform mcq irq: %d, %d\n", i, irq);
}
return;
@@ -1357,10 +1348,9 @@ static int ufs_mtk_pre_pwr_change(struct ufs_hba *hba,
ret = ufshcd_uic_change_pwr_mode(hba,
FASTAUTO_MODE << 4 | FASTAUTO_MODE);
- if (ret) {
- dev_err(hba->dev, "%s: HSG1B FASTAUTO failed ret=%d\n",
- __func__, ret);
- }
+ if (ret)
+ dev_err(hba->dev, "%s: HSG1B FASTAUTO failed: %pe\n",
+ __func__, ERR_PTR(ret));
}
/* if already configured to the requested pwr_mode, skip adapt */
@@ -1410,7 +1400,7 @@ static int ufs_mtk_auto_hibern8_disable(struct ufs_hba *hba)
out:
if (ret) {
- dev_warn(hba->dev, "exit h8 state fail, ret=%d\n", ret);
+ dev_err(hba->dev, "Failed to exit h8 state: %pe\n", ERR_PTR(ret));
ufshcd_force_error_recovery(hba);
@@ -1606,12 +1596,12 @@ static int ufs_mtk_link_set_hpm(struct ufs_hba *hba)
/* Check link state to make sure exit h8 success */
err = ufs_mtk_wait_idle_state(hba, 5);
if (err) {
- dev_warn(hba->dev, "wait idle fail, err=%d\n", err);
+ dev_err(hba->dev, "Failed to wait for idle: %pe\n", ERR_PTR(err));
return err;
}
err = ufs_mtk_wait_link_state(hba, VS_LINK_UP, 100);
if (err) {
- dev_warn(hba->dev, "exit h8 state fail, err=%d\n", err);
+ dev_err(hba->dev, "Failed to wait for link to be up: %pe\n", ERR_PTR(err));
return err;
}
ufshcd_set_link_active(hba);
@@ -1915,20 +1905,19 @@ static void ufs_mtk_event_notify(struct ufs_hba *hba,
/* Print details of UIC Errors */
if (evt <= UFS_EVT_DME_ERR) {
- dev_info(hba->dev,
- "Host UIC Error Code (%s): %08x\n",
- ufs_uic_err_str[evt], val);
+ dev_err(hba->dev, "Host UIC Error Code (%s): %08x\n",
+ ufs_uic_err_str[evt], val);
reg = val;
}
if (evt == UFS_EVT_PA_ERR) {
for_each_set_bit(bit, ®, ARRAY_SIZE(ufs_uic_pa_err_str))
- dev_info(hba->dev, "%s\n", ufs_uic_pa_err_str[bit]);
+ dev_err(hba->dev, "%s\n", ufs_uic_pa_err_str[bit]);
}
if (evt == UFS_EVT_DL_ERR) {
for_each_set_bit(bit, ®, ARRAY_SIZE(ufs_uic_dl_err_str))
- dev_info(hba->dev, "%s\n", ufs_uic_dl_err_str[bit]);
+ dev_err(hba->dev, "%s\n", ufs_uic_dl_err_str[bit]);
}
}
@@ -2143,7 +2132,7 @@ static int ufs_mtk_mcq_config_resource(struct ufs_hba *hba)
/* fail mcq initialization if interrupt is not filled properly */
if (!host->mcq_nr_intr) {
- dev_info(hba->dev, "IRQs not ready. MCQ disabled.");
+ dev_err(hba->dev, "IRQs not ready. MCQ disabled.");
return -EINVAL;
}
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 17/26] scsi: ufs: mediatek: Rework ufs_mtk_wait_idle_state
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (15 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 16/26] scsi: ufs: mediatek: Clean up logging prints Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 18/26] scsi: ufs: mediatek: Don't acquire dvfsrc-vcore twice Louis-Alexis Eyraud
` (8 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
While ufs_mtk_wait_idle state has some code smells for me (the
VS_HCE_BASE early exit seems racey at best), it can still benefit from
some general cleanup to make the code flow less convoluted.
Use the iopoll helpers, for one, and specifically the one that sleeps
and does not busy delay, as it's being done for up to 5ms.
The register read is split out to a helper function that branches
between new and old style flow.
Every called uses the same 5ms timeout value, so there is no point in
making this a parameter. Just assume a 5ms timeout in the function.
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 71 +++++++++++++++++------------------------
1 file changed, 30 insertions(+), 41 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 33917c345b29..0821de5b8a8f 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -10,6 +10,7 @@
#include <linux/bitfield.h>
#include <linux/clk.h>
#include <linux/delay.h>
+#include <linux/iopoll.h>
#include <linux/module.h>
#include <linux/of.h>
#include <linux/of_address.h>
@@ -380,51 +381,39 @@ static void ufs_mtk_dbg_sel(struct ufs_hba *hba)
}
}
-static int ufs_mtk_wait_idle_state(struct ufs_hba *hba,
- unsigned long retry_ms)
+static u32 ufs_mtk_read_state(struct ufs_hba *hba, bool old_style)
{
- u64 timeout, time_checked;
- u32 val, sm;
- bool wait_idle;
- struct ufs_mtk_host *host = ufshcd_get_variant(hba);
-
- /* cannot use plain ktime_get() in suspend */
- timeout = ktime_get_mono_fast_ns() + retry_ms * 1000000UL;
-
- /* wait a specific time after check base */
- udelay(10);
- wait_idle = false;
+ u32 val;
- do {
- time_checked = ktime_get_mono_fast_ns();
- if (host->legacy_ip_ver || host->ip_ver < IP_VER_MT6899) {
- ufs_mtk_dbg_sel(hba);
- val = ufshcd_readl(hba, REG_UFS_PROBE);
- } else {
- val = ufshcd_readl(hba, REG_UFS_UFS_MMIO_OTSD_CTRL);
- val = val >> 16;
- }
+ if (old_style) {
+ ufs_mtk_dbg_sel(hba);
+ val = ufshcd_readl(hba, REG_UFS_PROBE);
+ } else {
+ val = ufshcd_readl(hba, REG_UFS_UFS_MMIO_OTSD_CTRL) >> 16;
+ }
- sm = val & 0x1f;
+ return FIELD_GET(0x1f, val);
+}
- /*
- * if state is in H8 enter and H8 enter confirm
- * wait until return to idle state.
- */
- if ((sm >= VS_HIB_ENTER) && (sm <= VS_HIB_EXIT)) {
- wait_idle = true;
- udelay(50);
- continue;
- } else if (!wait_idle)
- break;
+static int ufs_mtk_wait_idle_state(struct ufs_hba *hba)
+{
+ struct ufs_mtk_host *host = ufshcd_get_variant(hba);
+ bool old_style = (host->legacy_ip_ver || host->ip_ver < IP_VER_MT6899);
+ u32 val;
+ int ret;
- if (wait_idle && (sm == VS_HCE_BASE))
- break;
- } while (time_checked < timeout);
+ /* If the device isn't in a hibernate state after 10us, don't wait. */
+ udelay(10);
+ val = ufs_mtk_read_state(hba, old_style);
+ if (val < VS_HIB_ENTER || val > VS_HIB_EXIT)
+ return 0;
- if (wait_idle && sm != VS_HCE_BASE) {
- dev_info(hba->dev, "wait idle tmo: 0x%x\n", val);
- return -ETIMEDOUT;
+ /* Poll to wait for idle */
+ ret = read_poll_timeout(ufs_mtk_read_state, val, (val == VS_HCE_BASE),
+ 50, 5 * USEC_PER_MSEC, false, hba, old_style);
+ if (ret) {
+ dev_err(hba->dev, "Timed out waiting for idle state, val = 0x%x\n", val);
+ return ret;
}
return 0;
@@ -1392,7 +1381,7 @@ static int ufs_mtk_auto_hibern8_disable(struct ufs_hba *hba)
ufshcd_writel(hba, 0, REG_AUTO_HIBERNATE_IDLE_TIMER);
/* wait host return to idle state when auto-hibern8 off */
- ret = ufs_mtk_wait_idle_state(hba, 5);
+ ret = ufs_mtk_wait_idle_state(hba);
if (ret)
goto out;
@@ -1594,7 +1583,7 @@ static int ufs_mtk_link_set_hpm(struct ufs_hba *hba)
return err;
/* Check link state to make sure exit h8 success */
- err = ufs_mtk_wait_idle_state(hba, 5);
+ err = ufs_mtk_wait_idle_state(hba);
if (err) {
dev_err(hba->dev, "Failed to wait for idle: %pe\n", ERR_PTR(err));
return err;
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 18/26] scsi: ufs: mediatek: Don't acquire dvfsrc-vcore twice
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (16 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 17/26] scsi: ufs: mediatek: Rework ufs_mtk_wait_idle_state Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:41 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 19/26] scsi: ufs: mediatek: Rework hardware version reading Louis-Alexis Eyraud
` (7 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
As part of its featureset, the ufs-mediatek driver needs to play with an
optional dvfsrc-vcore regulator for some of them.
However, it currently does this by acquiring two different references to
it in two different places, needlessly duplicating logic.
Move reg_vcore to the host struct, acquire it in the same function as
avdd09 is acquired, and rework the users of reg_vcore.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 71 +++++++++++++++++++----------------------
drivers/ufs/host/ufs-mediatek.h | 3 +-
2 files changed, 33 insertions(+), 41 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 0821de5b8a8f..8214cf0b3a16 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -519,7 +519,6 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
{
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
struct ufs_mtk_crypt_cfg *cfg;
- struct regulator *reg;
int volt, ret;
if (!ufs_mtk_is_boost_crypt_enabled(hba))
@@ -527,7 +526,6 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
cfg = host->crypt;
volt = cfg->vcore_volt;
- reg = cfg->reg_vcore;
ret = clk_prepare_enable(cfg->clk_crypt_mux);
if (ret) {
@@ -537,7 +535,7 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
}
if (boost) {
- ret = regulator_set_voltage(reg, volt, INT_MAX);
+ ret = regulator_set_voltage(host->reg_vcore, volt, INT_MAX);
if (ret) {
dev_err(hba->dev, "%s: Failed to set vcore to %d: %pe\n",
__func__, volt, ERR_PTR(ret));
@@ -548,7 +546,7 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
if (ret) {
dev_err(hba->dev, "%s: Failed to reparent clk_crypt_perf: %pe\n",
__func__, ERR_PTR(ret));
- regulator_set_voltage(reg, 0, INT_MAX);
+ regulator_set_voltage(host->reg_vcore, 0, INT_MAX);
goto out;
}
} else {
@@ -559,7 +557,7 @@ static void ufs_mtk_boost_crypt(struct ufs_hba *hba, bool boost)
goto out;
}
- ret = regulator_set_voltage(reg, 0, INT_MAX);
+ ret = regulator_set_voltage(host->reg_vcore, 0, INT_MAX);
if (ret) {
dev_err(hba->dev, "%s: Failed to set vcore to minimum: %pe\n",
__func__, ERR_PTR(ret));
@@ -576,15 +574,12 @@ static void ufs_mtk_init_boost_crypt(struct ufs_hba *hba)
struct device *dev = hba->dev;
int ret;
- cfg = devm_kzalloc(dev, sizeof(*cfg), GFP_KERNEL);
- if (!cfg)
+ if (!host->reg_vcore)
return;
- cfg->reg_vcore = devm_regulator_get_optional(dev, "dvfsrc-vcore");
- if (IS_ERR(cfg->reg_vcore)) {
- dev_err(dev, "Failed to get dvfsrc-vcore: %pe", cfg->reg_vcore);
+ cfg = devm_kzalloc(dev, sizeof(*cfg), GFP_KERNEL);
+ if (!cfg)
return;
- }
ret = of_property_read_u32(dev->of_node, "mediatek,boost-crypt-vcore-min",
&cfg->vcore_volt);
@@ -889,7 +884,6 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
struct list_head *head = &hba->clk_list_head;
struct ufs_clk_info *clki, *clki_tmp;
struct device *dev = hba->dev;
- struct regulator *reg;
u32 volt;
/*
@@ -930,16 +924,8 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
return;
}
- /*
- * Default get vcore if dts have these settings.
- * No matter clock scaling support or not. (may disable by customer)
- */
- reg = devm_regulator_get_optional(dev, "dvfsrc-vcore");
- if (IS_ERR(reg)) {
- if (PTR_ERR(reg) != -ENODEV)
- dev_err(dev, "Failed to get dvfsrc-vcore: %pe\n", reg);
+ if (!host->reg_vcore)
return;
- }
if (of_property_read_u32(dev->of_node, "clk-scale-up-vcore-min",
&volt)) {
@@ -947,12 +933,11 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
return;
}
- host->mclk.reg_vcore = reg;
host->mclk.vcore_volt = volt;
/* If default boot is max gear, request vcore */
- if (reg && volt && host->clk_scale_up)
- if (regulator_set_voltage(reg, volt, INT_MAX))
+ if (volt && host->clk_scale_up)
+ if (regulator_set_voltage(host->reg_vcore, volt, INT_MAX))
dev_err(hba->dev, "Failed to set vcore to %d\n", volt);
}
@@ -1064,6 +1049,17 @@ static int ufs_mtk_get_supplies(struct ufs_mtk_host *host)
const struct ufs_mtk_soc_data *data = of_device_get_match_data(dev);
int ret;
+ host->reg_vcore = devm_regulator_get_optional(dev, "dvfsrc-vcore");
+ if (IS_ERR(host->reg_vcore)) {
+ if (PTR_ERR(host->reg_vcore) != -ENODEV) {
+ dev_err(dev, "Failed to get dvfsrc-vcore supply: %pe\n",
+ host->reg_vcore);
+ return PTR_ERR(host->reg_vcore);
+ }
+
+ host->reg_vcore = NULL;
+ }
+
if (!data)
return 0;
@@ -1081,14 +1077,13 @@ static int ufs_mtk_get_supplies(struct ufs_mtk_host *host)
host->reg_avdd09 = devm_regulator_get_optional(dev, "avdd09");
if (IS_ERR(host->reg_avdd09)) {
- if (PTR_ERR(host->reg_avdd09) == -ENODEV) {
- host->reg_avdd09 = NULL;
- return 0;
+ if (PTR_ERR(host->reg_avdd09) != -ENODEV) {
+ dev_err(dev, "Failed to get avdd09 regulator: %pe\n",
+ host->reg_avdd09);
+ return PTR_ERR(host->reg_avdd09);
}
- dev_err(dev, "Failed to get avdd09 regulator: %pe\n",
- host->reg_avdd09);
- return PTR_ERR(host->reg_avdd09);
+ host->reg_avdd09 = NULL;
}
return 0;
@@ -1119,6 +1114,10 @@ static int ufs_mtk_init(struct ufs_hba *hba)
host->hba = hba;
ufshcd_set_variant(hba, host);
+ err = ufs_mtk_get_supplies(host);
+ if (err)
+ goto out_variant_clear;
+
/* Initialize host capability */
ufs_mtk_init_host_caps(hba);
@@ -1173,10 +1172,6 @@ static int ufs_mtk_init(struct ufs_hba *hba)
ufs_mtk_init_clocks(hba);
- err = ufs_mtk_get_supplies(host);
- if (err)
- goto out_phy_exit;
-
/*
* ufshcd_vops_init() is invoked after
* ufshcd_setup_clock(true) in ufshcd_hba_init() thus
@@ -1928,7 +1923,6 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
struct ufs_mtk_clk *mclk = &host->mclk;
struct ufs_clk_info *clki = mclk->ufs_sel_clki;
struct ufs_clk_info *fde_clki = mclk->ufs_fde_clki;
- struct regulator *reg;
int volt, ret = 0;
bool clk_bind_vcore = false;
bool clk_fde_scale = false;
@@ -1939,9 +1933,8 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
if (!clki || !fde_clki)
return;
- reg = host->mclk.reg_vcore;
volt = host->mclk.vcore_volt;
- if (reg && volt != 0)
+ if (host->reg_vcore && volt)
clk_bind_vcore = true;
if (mclk->ufs_fde_max_clki && mclk->ufs_fde_min_clki)
@@ -1965,7 +1958,7 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
if (scale_up) {
if (clk_bind_vcore) {
- ret = regulator_set_voltage(reg, volt, INT_MAX);
+ ret = regulator_set_voltage(host->reg_vcore, volt, INT_MAX);
if (ret) {
dev_err(hba->dev, "Failed to set vcore to %d\n", volt);
goto out;
@@ -2005,7 +1998,7 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
}
if (clk_bind_vcore) {
- ret = regulator_set_voltage(reg, 0, INT_MAX);
+ ret = regulator_set_voltage(host->reg_vcore, 0, INT_MAX);
if (ret) {
dev_err(hba->dev, "%s: Failed to set vcore to minimum: %pe\n",
__func__, ERR_PTR(ret));
diff --git a/drivers/ufs/host/ufs-mediatek.h b/drivers/ufs/host/ufs-mediatek.h
index d73093900d49..634c76c903c0 100644
--- a/drivers/ufs/host/ufs-mediatek.h
+++ b/drivers/ufs/host/ufs-mediatek.h
@@ -141,7 +141,6 @@ enum ufs_mtk_host_caps {
};
struct ufs_mtk_crypt_cfg {
- struct regulator *reg_vcore;
struct clk *clk_crypt_perf;
struct clk *clk_crypt_mux;
struct clk *clk_crypt_lp;
@@ -155,7 +154,6 @@ struct ufs_mtk_clk {
struct ufs_clk_info *ufs_fde_clki; /* Mux */
struct ufs_clk_info *ufs_fde_max_clki; /* Max src */
struct ufs_clk_info *ufs_fde_min_clki; /* Min src */
- struct regulator *reg_vcore;
int vcore_volt;
};
@@ -174,6 +172,7 @@ struct ufs_mtk_mcq_intr_info {
struct ufs_mtk_host {
struct phy *mphy;
struct regulator *reg_avdd09;
+ struct regulator *reg_vcore;
struct reset_control_bulk_data resets[MTK_UFS_NUM_RESETS];
struct ufs_hba *hba;
struct ufs_mtk_crypt_cfg *crypt;
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 18/26] scsi: ufs: mediatek: Don't acquire dvfsrc-vcore twice
2026-07-30 9:49 ` [PATCH v10 18/26] scsi: ufs: mediatek: Don't acquire dvfsrc-vcore twice Louis-Alexis Eyraud
@ 2026-07-30 10:41 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:41 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: conor+dt, olteanv, neil.armstrong, devicetree, linux-scsi, vkoul,
robh, linux-phy
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Consolidating `dvfsrc-vcore` into a single shared regulator handle prevents the regulator core from aggregating voltage constraints, causing independent features to overwrite each other's voltage requests and leading to hardware undervoltage.
--
--- Patch [18]: [PATCH v10 18/26] scsi: ufs: mediatek: Don't acquire dvfsrc-vcore twice ---
commit f445ade145082cbe60e3dee2e8a70aa32aec9d9b
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Don't acquire dvfsrc-vcore twice
This patch consolidates the dvfsrc-vcore regulator into a single shared
handle on the host structure, rather than acquiring two separate references.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index 0821de5b8a8fc..8214cf0b3a169 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
> @@ -2005,7 +1998,7 @@ static void _ufs_mtk_clk_scale(struct ufs_hba *hba, bool scale_up)
> }
>
> if (clk_bind_vcore) {
> - ret = regulator_set_voltage(reg, 0, INT_MAX);
> + ret = regulator_set_voltage(host->reg_vcore, 0, INT_MAX);
[Severity: High]
Will consolidating to a single host->reg_vcore handle prevent the regulator
core from properly aggregating voltage constraints?
During a clock scale-down, ufs_mtk_clk_scale_notify() runs _ufs_mtk_clk_scale()
with scale_up = false during the pre-change notification phase. Since
host->reg_vcore is now a single shared consumer handle, calling
regulator_set_voltage() with 0V here will immediately drop the physical
voltage constraint.
However, the crypto boost clock isn't downclocked until the post-change
notification phase calls ufs_mtk_boost_crypt(). Can this violate DVFS
constraints and cause a hardware undervoltage by dropping the voltage while
the crypto clock is still running at high frequency?
It appears the regulator core relies on separate consumer handles to aggregate
maximum voltage requests across independent features like clock scaling and
crypto boost, which might explain why two separate references were originally
used.
> if (ret) {
> dev_err(hba->dev, "%s: Failed to set vcore to minimum: %pe\n",
> __func__, ERR_PTR(ret));
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=18
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 19/26] scsi: ufs: mediatek: Rework hardware version reading
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (17 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 18/26] scsi: ufs: mediatek: Don't acquire dvfsrc-vcore twice Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 20/26] scsi: ufs: mediatek: Back up idle timer in per-instance struct Louis-Alexis Eyraud
` (6 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Split assignment to the host struct out from the read function, and
utilise bitfield helpers to simplify the code. Also move the debug print
out of the legacy version helper, which means it no longer has to take a
struct ufs_hba as an input, and can be rewritten as a pure function.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 65 +++++++++++++++++++++--------------------
1 file changed, 33 insertions(+), 32 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 8214cf0b3a16..c784ecb30e74 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -797,50 +797,47 @@ static void ufs_mtk_mcq_set_irq_affinity(struct ufs_hba *hba, unsigned int cpu)
dev_dbg(hba->dev, "set irq %d affinity to CPU %d\n", irq, _cpu);
}
-static bool ufs_mtk_is_legacy_chipset(struct ufs_hba *hba, u32 hw_ip_ver)
+static bool __pure ufs_mtk_is_legacy_chipset(u32 hw_ip_ver)
{
- bool is_legacy = false;
-
switch (hw_ip_ver) {
case IP_LEGACY_VER_MT6893:
case IP_LEGACY_VER_MT6781:
/* can add other legacy chipset ID here accordingly */
- is_legacy = true;
- break;
- default:
- break;
+ return true;
}
- dev_dbg(hba->dev, "IP version 0x%x, legacy = %s", hw_ip_ver,
- str_true_false(is_legacy));
- return is_legacy;
+ return false;
}
-/*
- * HW version format has been changed from 01MMmmmm to 1MMMmmmm, since
- * project MT6878. In order to perform correct version comparison,
- * version number is changed by SW for the following projects.
- * IP_VER_MT6983 0x00360000 to 0x10360000
- * IP_VER_MT6897 0x01440000 to 0x10440000
- * IP_VER_MT6989 0x01450000 to 0x10450000
- * IP_VER_MT6991 0x01460000 to 0x10460000
+#define MTK_UFS_VER_PREFIX_M GENMASK(31, 24)
+
+/**
+ * ufs_mtk_get_hw_ip_version - read and return adjusted hardware version
+ * @hba: pointer to this device's &struct ufs_hba
+ *
+ * Reads, transforms and returns the hardware version.
+ *
+ * Since MT6878, the versioning scheme was changed from 01MMmmmm to 1MMMmmmm.
+ * In order to support version comparisons across these different versioning
+ * schemes, this function transforms the older style to the newer one.
+ *
+ * For example:
+ * MT6983 is transformed from 0x00360000 to 0x10360000
+ * MT6897 is transformed from 0x01440000 to 0x10440000
+ * MT6989 is transformed from 0x01450000 to 0x10450000
+ * MT6991 is transformed from 0x01460000 to 0x10460000
+ *
+ * Returns a u32 representing the hardware version.
*/
-static void ufs_mtk_get_hw_ip_version(struct ufs_hba *hba)
+static u32 ufs_mtk_get_hw_ip_version(struct ufs_hba *hba)
{
- struct ufs_mtk_host *host = ufshcd_get_variant(hba);
- u32 hw_ip_ver;
+ u32 version = ufshcd_readl(hba, REG_UFS_MTK_IP_VER);
+ u32 prefix = FIELD_GET(MTK_UFS_VER_PREFIX_M, version);
- hw_ip_ver = ufshcd_readl(hba, REG_UFS_MTK_IP_VER);
+ if (prefix <= 1)
+ FIELD_MODIFY(MTK_UFS_VER_PREFIX_M, &version, 0x10);
- if (((hw_ip_ver & (0xFF << 24)) == (0x1 << 24)) ||
- ((hw_ip_ver & (0xFF << 24)) == 0)) {
- hw_ip_ver &= ~(0xFF << 24);
- hw_ip_ver |= (0x1 << 28);
- }
-
- host->ip_ver = hw_ip_ver;
-
- host->legacy_ip_ver = ufs_mtk_is_legacy_chipset(hba, hw_ip_ver);
+ return version;
}
static void ufs_mtk_get_controller_version(struct ufs_hba *hba)
@@ -1191,7 +1188,11 @@ static int ufs_mtk_init(struct ufs_hba *hba)
ufs_mtk_setup_clocks(hba, true, POST_CHANGE);
- ufs_mtk_get_hw_ip_version(hba);
+ host->ip_ver = ufs_mtk_get_hw_ip_version(hba);
+ host->legacy_ip_ver = ufs_mtk_is_legacy_chipset(host->ip_ver);
+
+ dev_dbg(hba->dev, "IP version 0x%x, legacy = %s", host->ip_ver,
+ str_true_false(host->legacy_ip_ver));
return 0;
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 20/26] scsi: ufs: mediatek: Back up idle timer in per-instance struct
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (18 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 19/26] scsi: ufs: mediatek: Rework hardware version reading Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:53 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 21/26] scsi: ufs: mediatek: Remove ret local from link_startup_notify Louis-Alexis Eyraud
` (5 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The MediaTek UFS driver uses a function-scope static variable to back up
a hardware register across a power change in the
ufs_mtk_pwr_change_notify function. This is dangerous, as it's only
correct if only ever one instance of the driver is loaded, which isn't
true if there's more than one device on a SoC that needs it, or it
otherwise gets loaded a second time.
Back it up into a member of the host struct instead, as this struct is
per-instance. Rework the function to not use a pointless "ret" local as
well.
Fixes: f5ca8d0c7a63 ("scsi: ufs: host: mediatek: Disable auto-hibern8 during power mode changes")
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Reviewed-by: Chaotian Jing <chaotian.jing@mediatek.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 19 ++++++++-----------
drivers/ufs/host/ufs-mediatek.h | 1 +
2 files changed, 9 insertions(+), 11 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index c784ecb30e74..115056927223 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -1400,27 +1400,24 @@ static int ufs_mtk_pwr_change_notify(struct ufs_hba *hba,
enum ufs_notify_change_status stage,
struct ufs_pa_layer_attr *dev_req_params)
{
- int ret = 0;
- static u32 reg;
+ struct ufs_mtk_host *host = ufshcd_get_variant(hba);
switch (stage) {
case PRE_CHANGE:
if (ufshcd_is_auto_hibern8_supported(hba)) {
- reg = ufshcd_readl(hba, REG_AUTO_HIBERNATE_IDLE_TIMER);
+ host->ahit = ufshcd_readl(
+ hba, REG_AUTO_HIBERNATE_IDLE_TIMER);
ufs_mtk_auto_hibern8_disable(hba);
}
- ret = ufs_mtk_pre_pwr_change(hba, dev_req_params);
- break;
+ return ufs_mtk_pre_pwr_change(hba, dev_req_params);
case POST_CHANGE:
if (ufshcd_is_auto_hibern8_supported(hba))
- ufshcd_writel(hba, reg, REG_AUTO_HIBERNATE_IDLE_TIMER);
- break;
- default:
- ret = -EINVAL;
- break;
+ ufshcd_writel(hba, host->ahit,
+ REG_AUTO_HIBERNATE_IDLE_TIMER);
+ return 0;
}
- return ret;
+ return -EINVAL;
}
static int ufs_mtk_unipro_set_lpm(struct ufs_hba *hba, bool lpm)
diff --git a/drivers/ufs/host/ufs-mediatek.h b/drivers/ufs/host/ufs-mediatek.h
index 634c76c903c0..7f2cdfd632be 100644
--- a/drivers/ufs/host/ufs-mediatek.h
+++ b/drivers/ufs/host/ufs-mediatek.h
@@ -187,6 +187,7 @@ struct ufs_mtk_host {
u16 ref_clk_gating_wait_us;
u32 ip_ver;
bool legacy_ip_ver;
+ u32 ahit;
bool mcq_set_intr;
bool is_mcq_intr_enabled;
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 20/26] scsi: ufs: mediatek: Back up idle timer in per-instance struct
2026-07-30 9:49 ` [PATCH v10 20/26] scsi: ufs: mediatek: Back up idle timer in per-instance struct Louis-Alexis Eyraud
@ 2026-07-30 10:53 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:53 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: conor+dt, linux-phy, vkoul, neil.armstrong, olteanv, robh,
linux-scsi, devicetree
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
Pre-existing issues:
- [High] Skipped POST_CHANGE notification on power change failure leads to permanent loss of auto-hibernate configuration, disabling Auto-Hibern8.
--
--- Patch [20]: [PATCH v10 20/26] scsi: ufs: mediatek: Back up idle timer in per-instance struct ---
commit 31da80f38058b251310b08adf616cded137a1bcd
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Back up idle timer in per-instance struct
This commit moves the backup of the auto-hibernate idle timer from a
static local variable to a per-instance host struct member. This ensures
correct behavior when multiple instances of the driver are loaded.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index c784ecb30e74e..115056927223f 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
> @@ -1400,27 +1400,24 @@ static int ufs_mtk_pwr_change_notify(struct ufs_hba *hba,
> enum ufs_notify_change_status stage,
> struct ufs_pa_layer_attr *dev_req_params)
> {
> - int ret = 0;
> - static u32 reg;
> + struct ufs_mtk_host *host = ufshcd_get_variant(hba);
>
> switch (stage) {
> case PRE_CHANGE:
> if (ufshcd_is_auto_hibern8_supported(hba)) {
> - reg = ufshcd_readl(hba, REG_AUTO_HIBERNATE_IDLE_TIMER);
> + host->ahit = ufshcd_readl(
> + hba, REG_AUTO_HIBERNATE_IDLE_TIMER);
[Severity: High]
This isn't a bug introduced by this patch, but does this sequence permanently
disable Auto-Hibern8 if a power change fails?
If a DME power mode change fails, the UFS core in ufshcd_change_power_mode()
skips the POST_CHANGE notification:
drivers/ufs/core/ufshcd.c:ufshcd_change_power_mode() {
...
ret = ufshcd_dme_change_power_mode(hba, pwr_mode);
if (!ret)
ufshcd_vops_pwr_change_notify(hba, POST_CHANGE, pwr_mode);
...
}
Because POST_CHANGE is skipped, the hardware register remains at 0 (since
it was disabled right below by ufs_mtk_auto_hibern8_disable).
On the next power mode change attempt, PRE_CHANGE will run again and read the
current register value (0) into host->ahit, overwriting the valid backup.
When POST_CHANGE eventually succeeds, it will write 0 back to the register,
leaving Auto-Hibern8 permanently disabled until a reset or suspend/resume.
> ufs_mtk_auto_hibern8_disable(hba);
> }
[ ... ]
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=20
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 21/26] scsi: ufs: mediatek: Remove ret local from link_startup_notify
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (19 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 20/26] scsi: ufs: mediatek: Back up idle timer in per-instance struct Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 22/26] scsi: ufs: mediatek: Remove undocumented "clk-scale-up-vcore-min" Louis-Alexis Eyraud
` (4 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Remove the "ret" local variable from ufs_mtk_link_startup_notify, as
it's pointless; in all cases it is assigned, it is returned right after
without being read first.
Rework the code to just return directly, and get rid of the default
branch while at it.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 12 +++---------
1 file changed, 3 insertions(+), 9 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 115056927223..90c8606d85a9 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -1502,21 +1502,15 @@ static void ufs_mtk_post_link(struct ufs_hba *hba)
static int ufs_mtk_link_startup_notify(struct ufs_hba *hba,
enum ufs_notify_change_status stage)
{
- int ret = 0;
-
switch (stage) {
case PRE_CHANGE:
- ret = ufs_mtk_pre_link(hba);
- break;
+ return ufs_mtk_pre_link(hba);
case POST_CHANGE:
ufs_mtk_post_link(hba);
- break;
- default:
- ret = -EINVAL;
- break;
+ return 0;
}
- return ret;
+ return -EINVAL;
}
static int ufs_mtk_device_reset(struct ufs_hba *hba)
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 22/26] scsi: ufs: mediatek: Remove undocumented "clk-scale-up-vcore-min"
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (20 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 21/26] scsi: ufs: mediatek: Remove ret local from link_startup_notify Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 10:49 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 23/26] scsi: ufs: mediatek: Add MT8196 compatible, update copyright Louis-Alexis Eyraud
` (3 subsequent siblings)
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
The MediaTek UFS driver contains support for an undocumented,
non-vendor-prefixed u32 property named "clk-scale-up-vcore-min".
Since it is not part of any binding, and would not pass a bindings
review in its current form, remove it.
To return this functionality, it needs to be resubmitted in a series
that also introduces it to the binding, and justifies what it is used
for. Compatibility with downstream device trees is not a valid
justification for its existence.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 19 -------------------
1 file changed, 19 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 90c8606d85a9..5dd72b826181 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -880,8 +880,6 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
struct ufs_mtk_host *host = ufshcd_get_variant(hba);
struct list_head *head = &hba->clk_list_head;
struct ufs_clk_info *clki, *clki_tmp;
- struct device *dev = hba->dev;
- u32 volt;
/*
* Find private clocks and store them in struct ufs_mtk_clk.
@@ -918,24 +916,7 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
if (!ufs_mtk_is_clk_scale_ready(hba)) {
hba->caps &= ~UFSHCD_CAP_CLK_SCALING;
dev_info(hba->dev, "%s: Clock scaling unavailable", __func__);
- return;
- }
-
- if (!host->reg_vcore)
- return;
-
- if (of_property_read_u32(dev->of_node, "clk-scale-up-vcore-min",
- &volt)) {
- dev_info(dev, "failed to get clk-scale-up-vcore-min");
- return;
}
-
- host->mclk.vcore_volt = volt;
-
- /* If default boot is max gear, request vcore */
- if (volt && host->clk_scale_up)
- if (regulator_set_voltage(host->reg_vcore, volt, INT_MAX))
- dev_err(hba->dev, "Failed to set vcore to %d\n", volt);
}
static void ufs_mtk_setup_clk_gating(struct ufs_hba *hba)
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 22/26] scsi: ufs: mediatek: Remove undocumented "clk-scale-up-vcore-min"
2026-07-30 9:49 ` [PATCH v10 22/26] scsi: ufs: mediatek: Remove undocumented "clk-scale-up-vcore-min" Louis-Alexis Eyraud
@ 2026-07-30 10:49 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 10:49 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: linux-phy, devicetree, neil.armstrong, conor+dt, linux-scsi, robh,
vkoul, olteanv
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] Incomplete feature removal leaves behind unused struct field and dead code for UFS clock vcore scaling.
--
--- Patch [22]: [PATCH v10 22/26] scsi: ufs: mediatek: Remove undocumented "clk-scale-up-vcore-min" ---
commit 35c917bdf36a893b4dd563ccc122d8d25ec6d9ed
Author: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
scsi: ufs: mediatek: Remove undocumented "clk-scale-up-vcore-min"
This commit removes support for the undocumented "clk-scale-up-vcore-min"
property from the MediaTek UFS driver. It deletes the device tree parsing
logic and the initial vcore regulator setup.
> diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
> index 90c8606d85a9d..5dd72b8261815 100644
> --- a/drivers/ufs/host/ufs-mediatek.c
> +++ b/drivers/ufs/host/ufs-mediatek.c
[ ... ]
> @@ -918,24 +916,7 @@ static void ufs_mtk_init_clocks(struct ufs_hba *hba)
> if (!ufs_mtk_is_clk_scale_ready(hba)) {
> hba->caps &= ~UFSHCD_CAP_CLK_SCALING;
> dev_info(hba->dev, "%s: Clock scaling unavailable", __func__);
> - return;
> - }
> -
> - if (!host->reg_vcore)
> - return;
> -
> - if (of_property_read_u32(dev->of_node, "clk-scale-up-vcore-min",
> - &volt)) {
> - dev_info(dev, "failed to get clk-scale-up-vcore-min");
> - return;
> }
> -
> - host->mclk.vcore_volt = volt;
[Severity: Low]
Since this patch removes the code that populates host->mclk.vcore_volt, does
it leave behind an unused field and dead code elsewhere in the driver?
In drivers/ufs/host/ufs-mediatek.h:
struct ufs_mtk_clk {
...
int vcore_volt;
...
};
And in drivers/ufs/host/ufs-mediatek.c:_ufs_mtk_clk_scale():
volt = host->mclk.vcore_volt;
if (host->reg_vcore && volt)
clk_bind_vcore = true;
Because vcore_volt is zero-initialized and no longer updated, clk_bind_vcore
appears to always be false, which makes the subsequent regulator scaling
branches permanently unreachable. Should these leftovers be cleaned up as
well?
> -
> - /* If default boot is max gear, request vcore */
> - if (volt && host->clk_scale_up)
> - if (regulator_set_voltage(host->reg_vcore, volt, INT_MAX))
> - dev_err(hba->dev, "Failed to set vcore to %d\n", volt);
> }
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=22
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 23/26] scsi: ufs: mediatek: Add MT8196 compatible, update copyright
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (21 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 22/26] scsi: ufs: mediatek: Remove undocumented "clk-scale-up-vcore-min" Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 24/26] arm64: dts: mediatek: mt8195: Align ufshci node to dt-bindings changes Louis-Alexis Eyraud
` (2 subsequent siblings)
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
From: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
THe MT8196's UFS controller has a new compatible. Add the necessary
struct definitions to support it.
Also update the copyrights and authors, without tabs following spaces to
avoid checkpatch errors, to list myself as having contributed to this
driver after the preceding rework patches.
Reviewed-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Reviewed-by: Peter Wang <peter.wang@mediatek.com>
Signed-off-by: Nicolas Frattaroli <nicolas.frattaroli@collabora.com>
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
drivers/ufs/host/ufs-mediatek.c | 17 +++++++++++++++--
1 file changed, 15 insertions(+), 2 deletions(-)
diff --git a/drivers/ufs/host/ufs-mediatek.c b/drivers/ufs/host/ufs-mediatek.c
index 5dd72b826181..decf2f5814b4 100644
--- a/drivers/ufs/host/ufs-mediatek.c
+++ b/drivers/ufs/host/ufs-mediatek.c
@@ -1,9 +1,11 @@
// SPDX-License-Identifier: GPL-2.0
/*
* Copyright (C) 2019 MediaTek Inc.
+ * Copyright (C) 2025 Collabora Ltd.
* Authors:
- * Stanley Chu <stanley.chu@mediatek.com>
- * Peter Wang <peter.wang@mediatek.com>
+ * Stanley Chu <stanley.chu@mediatek.com>
+ * Peter Wang <peter.wang@mediatek.com>
+ * Nicolas Frattaroli <nicolas.frattaroli@collabora.com> (Major cleanups)
*/
#include <linux/arm-smccc.h>
@@ -2222,6 +2224,10 @@ static const char *const ufs_mtk_regs_avdd12_ckbuf_avdd18[] = {
"avdd12", "avdd12-ckbuf", "avdd18"
};
+static const char *const ufs_mtk_regs_avdd12_ckbuf[] = {
+ "avdd12", "avdd12-ckbuf"
+};
+
static const struct ufs_mtk_soc_data mt8183_data = {
.has_avdd09 = true,
.reg_names = ufs_mtk_regs_avdd12_avdd18,
@@ -2234,10 +2240,17 @@ static const struct ufs_mtk_soc_data mt8192_8195_data = {
.num_reg_names = ARRAY_SIZE(ufs_mtk_regs_avdd12_ckbuf_avdd18),
};
+static const struct ufs_mtk_soc_data mt8196_data = {
+ .has_avdd09 = true,
+ .reg_names = ufs_mtk_regs_avdd12_ckbuf,
+ .num_reg_names = ARRAY_SIZE(ufs_mtk_regs_avdd12_ckbuf),
+};
+
static const struct of_device_id ufs_mtk_of_match[] = {
{ .compatible = "mediatek,mt8183-ufshci", .data = &mt8183_data },
{ .compatible = "mediatek,mt8192-ufshci", .data = &mt8192_8195_data },
{ .compatible = "mediatek,mt8195-ufshci", .data = &mt8192_8195_data },
+ { .compatible = "mediatek,mt8196-ufshci", .data = &mt8196_data },
{},
};
MODULE_DEVICE_TABLE(of, ufs_mtk_of_match);
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 24/26] arm64: dts: mediatek: mt8195: Align ufshci node to dt-bindings changes
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (22 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 23/26] scsi: ufs: mediatek: Add MT8196 compatible, update copyright Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 11:13 ` sashiko-bot
2026-07-30 9:49 ` [PATCH v10 25/26] arm64: dts: mediatek: mt8395-genio-1200-evk-ufs: Complete UFS power supplies Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 26/26] arm64: dts: mediatek: mt8395-radxa-nio-12l: " Louis-Alexis Eyraud
25 siblings, 1 reply; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
The MT8195 UFS host controller dt bindings now describe several extra
required clocks that were previously missing from clock description
and deprecates the freq-table-hz property use in favour of
operating-points-v2 instead.
Currently, the ufshci node declares the freq-table-hz property but all
values contained in the array are zero, thus converting it to an OPP
table is not needed.
So, add in ufshci node the rx_symbol clocks in clock list and remove
freq-table-hz property.
Also reorder properties in node.
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
arch/arm64/boot/dts/mediatek/mt8195.dtsi | 21 ++++++++++-----------
1 file changed, 10 insertions(+), 11 deletions(-)
diff --git a/arch/arm64/boot/dts/mediatek/mt8195.dtsi b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
index 9aaf8b6edcb5..5ff7b7e543b7 100644
--- a/arch/arm64/boot/dts/mediatek/mt8195.dtsi
+++ b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
@@ -1472,8 +1472,7 @@ mmc2: mmc@11250000 {
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>,
@@ -1481,15 +1480,15 @@ ufshci: ufshci@11270000 {
<&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>;
-
+ <&infracfg_ao CLK_INFRA_AO_PERI_UFS_MEM_SUB>,
+ <&topckgen CLK_TOP_UFS_RX_SYMBOL>,
+ <&topckgen CLK_TOP_UFS_RX_SYMBOL1>;
+ clock-names = "main", "aes", "tick", "unipro_sysclk",
+ "unipro_tick", "unipro_mp_bclk",
+ "tx_symbol", "mem_sub", "rx_symbol0",
+ "rx_symbol1";
+ interrupts = <GIC_SPI 137 IRQ_TYPE_LEVEL_HIGH 0>;
+ phys = <&ufsphy>;
mediatek,ufs-disable-mcq;
status = "disabled";
};
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* Re: [PATCH v10 24/26] arm64: dts: mediatek: mt8195: Align ufshci node to dt-bindings changes
2026-07-30 9:49 ` [PATCH v10 24/26] arm64: dts: mediatek: mt8195: Align ufshci node to dt-bindings changes Louis-Alexis Eyraud
@ 2026-07-30 11:13 ` sashiko-bot
0 siblings, 0 replies; 42+ messages in thread
From: sashiko-bot @ 2026-07-30 11:13 UTC (permalink / raw)
To: Louis-Alexis Eyraud
Cc: olteanv, robh, devicetree, conor+dt, linux-scsi, neil.armstrong,
linux-phy, vkoul
Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] Removing the `freq-table-hz` property without adding an `operating-points-v2` table causes the UFS core platform driver to skip parsing and enabling all hardware clocks, resulting in a controller initialization failure.
--
--- Patch [24]: [PATCH v10 24/26] arm64: dts: mediatek: mt8195: Align ufshci node to dt-bindings changes ---
commit eac3ed6e6791c53424e4e025e4dfd47b68309eb7
Author: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
arm64: dts: mediatek: mt8195: Align ufshci node to dt-bindings changes
This commit updates the MT8195 UFS host controller dt bindings by adding
new required clocks and removing the deprecated freq-table-hz property.
It states that converting to an OPP table is not needed since the values
were all zero.
> diff --git a/arch/arm64/boot/dts/mediatek/mt8195.dtsi b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> --- a/arch/arm64/boot/dts/mediatek/mt8195.dtsi
> +++ b/arch/arm64/boot/dts/mediatek/mt8195.dtsi
[ ... ]
> @@ -1481,15 +1480,15 @@ ufshci: ufshci@11270000 {
> <&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>;
> -
[Severity: High]
Does removing this property introduce a regression in the UFS controller
initialization?
Looking at ufshcd_parse_clock_info() in drivers/ufs/host/ufshcd-pltfrm.c,
the driver skips parsing the clock properties entirely if freq-table-hz is
missing:
drivers/ufs/host/ufshcd-pltfrm.c:ufshcd_parse_clock_info() {
...
sz = of_property_count_u32_elems(np, "freq-table-hz");
if (sz <= 0) {
dev_info(dev, "freq-table-hz property not specified\n");
goto out;
}
...
}
Since an operating-points-v2 node is also absent, the alternative parser
ufshcd_parse_operating_points() will also return early. This leaves
hba->clk_list_head empty.
If the list is empty, ufshcd_init_clocks() and ufshcd_setup_clocks()
will silently return without turning on any hardware clocks.
Could this cause a probe failure or a system crash (SError) when the driver
later attempts to access unclocked UFS MMIO registers?
> + <&infracfg_ao CLK_INFRA_AO_PERI_UFS_MEM_SUB>,
> + <&topckgen CLK_TOP_UFS_RX_SYMBOL>,
> + <&topckgen CLK_TOP_UFS_RX_SYMBOL1>;
> + clock-names = "main", "aes", "tick", "unipro_sysclk",
> + "unipro_tick", "unipro_mp_bclk",
> + "tx_symbol", "mem_sub", "rx_symbol0",
> + "rx_symbol1";
> + interrupts = <GIC_SPI 137 IRQ_TYPE_LEVEL_HIGH 0>;
> + phys = <&ufsphy>;
> mediatek,ufs-disable-mcq;
> status = "disabled";
> };
--
Sashiko AI review · https://sashiko.dev/#/patchset/20260730-mt8196-ufs-v10-0-bc5ecef202dc@collabora.com?part=24
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply [flat|nested] 42+ messages in thread
* [PATCH v10 25/26] arm64: dts: mediatek: mt8395-genio-1200-evk-ufs: Complete UFS power supplies
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (23 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 24/26] arm64: dts: mediatek: mt8195: Align ufshci node to dt-bindings changes Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
2026-07-30 9:49 ` [PATCH v10 26/26] arm64: dts: mediatek: mt8395-radxa-nio-12l: " Louis-Alexis Eyraud
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
The MT8195 UFS host controller dt binding now describes the power supplies
feeding the host controller, that were previously missing from supply
description.
So, add in the Mediatek Genio 1200-EVK board devicetree the regulators
used to power MT8395 UFS controller when the board is configured to use
its UFS storage as primary.
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
arch/arm64/boot/dts/mediatek/mt8395-genio-1200-evk-ufs.dts | 3 +++
1 file changed, 3 insertions(+)
diff --git a/arch/arm64/boot/dts/mediatek/mt8395-genio-1200-evk-ufs.dts b/arch/arm64/boot/dts/mediatek/mt8395-genio-1200-evk-ufs.dts
index e09a3ecd8773..f3d839f31918 100644
--- a/arch/arm64/boot/dts/mediatek/mt8395-genio-1200-evk-ufs.dts
+++ b/arch/arm64/boot/dts/mediatek/mt8395-genio-1200-evk-ufs.dts
@@ -16,6 +16,9 @@ / {
&ufshci {
status = "okay";
+ avdd12-supply = <&mt6359_vrf12_ldo_reg>;
+ avdd12-ckbuf-supply = <&mt6359_vbbck_ldo_reg>;
+ avdd18-supply = <&mt6359_vio18_ldo_reg>;
vcc-supply = <&mt6359_vemc_1_ldo_reg>;
vccq2-supply = <&mt6359_vufs_ldo_reg>;
};
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread* [PATCH v10 26/26] arm64: dts: mediatek: mt8395-radxa-nio-12l: Complete UFS power supplies
2026-07-30 9:48 [PATCH v10 00/26] MediaTek UFS Cleanup and MT8196 Enablement Louis-Alexis Eyraud
` (24 preceding siblings ...)
2026-07-30 9:49 ` [PATCH v10 25/26] arm64: dts: mediatek: mt8395-genio-1200-evk-ufs: Complete UFS power supplies Louis-Alexis Eyraud
@ 2026-07-30 9:49 ` Louis-Alexis Eyraud
25 siblings, 0 replies; 42+ messages in thread
From: Louis-Alexis Eyraud @ 2026-07-30 9:49 UTC (permalink / raw)
To: Chunfeng Yun, Vinod Koul, Neil Armstrong, Rob Herring,
Krzysztof Kozlowski, Conor Dooley, Matthias Brugger,
AngeloGioacchino Del Regno, Peter Wang, Chaotian Jing,
Alim Akhtar, Avri Altman, Bart Van Assche, Stanley Jhu,
James E.J. Bottomley, Martin K. Petersen, Philipp Zabel,
Liam Girdwood, Mark Brown
Cc: kernel, linux-arm-kernel, linux-mediatek, linux-phy, devicetree,
linux-kernel, linux-scsi, Nicolas Frattaroli, Louis-Alexis Eyraud
The MT8195 UFS host controller dt binding now describes the power supplies
feeding the host controller, that were previously missing from supply
description.
So, add in the Radxa NIO-12L board devicetree the regulators used to
power MT8395 UFS controller on the board.
Signed-off-by: Louis-Alexis Eyraud <louisalexis.eyraud@collabora.com>
---
arch/arm64/boot/dts/mediatek/mt8395-radxa-nio-12l.dts | 3 +++
1 file changed, 3 insertions(+)
diff --git a/arch/arm64/boot/dts/mediatek/mt8395-radxa-nio-12l.dts b/arch/arm64/boot/dts/mediatek/mt8395-radxa-nio-12l.dts
index 9cf0315da035..7e9cabdd2825 100644
--- a/arch/arm64/boot/dts/mediatek/mt8395-radxa-nio-12l.dts
+++ b/arch/arm64/boot/dts/mediatek/mt8395-radxa-nio-12l.dts
@@ -1184,6 +1184,9 @@ &uart1 {
};
&ufshci {
+ avdd12-supply = <&mt6359_vrf12_ldo_reg>;
+ avdd12-ckbuf-supply = <&mt6359_vbbck_ldo_reg>;
+ avdd18-supply = <&mt6359_vio18_ldo_reg>;
vcc-supply = <&mt6359_vemc_1_ldo_reg>;
vccq2-supply = <&mt6359_vufs_ldo_reg>;
status = "okay";
--
2.55.0
--
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy
^ permalink raw reply related [flat|nested] 42+ messages in thread