[PATCH v2 0/2] soc: renesas: mfis: add R-Car V4H/V4M support

[PATCH v2 0/2] soc: renesas: mfis: add R-Car V4H/V4M support

[Wolfram Sang]

Here are the additions to the MFIS driver in -next to add V4H support.
A full branch for testing with the SparrowHawk is here:

git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git renesas/v4h/mfis

To trigger the mailbox-test instances, I use:

$ for f in /sys/kernel/debug/e6300*/message; do echo "Hallo $f!" > $f; done

Looking forward to comments!

Happy hacking,

   Wolfram

Changes since (internal) v1:
* added tags from Geert (Thanks!)


Wolfram Sang (2):
  dt-bindings: soc: renesas: mfis: add R-Car V4H/V4M support
  soc: renesas: mfis: add R-Car V4H/V4M support

 .../soc/renesas/renesas,r8a78000-mfis.yaml    | 38 +++++++++++++++-
 drivers/soc/renesas/rcar-mfis.c               | 44 ++++++++++++++++++-
 2 files changed, 78 insertions(+), 4 deletions(-)

-- 
2.51.0

[PATCH v2 1/2] dt-bindings: soc: renesas: mfis: add R-Car V4H/V4M support

[Wolfram Sang]

The above SoCs have only 12 mailboxes and do not have an extra register
space for mailboxes. Everything is contained in the common register set.
In addition to adding these SoCs, the other entries get updated to
enforce 2 register spaces and their specific number of interrupts.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 .../soc/renesas/renesas,r8a78000-mfis.yaml    | 38 ++++++++++++++++++-
 1 file changed, 36 insertions(+), 2 deletions(-)

diff --git a/Documentation/devicetree/bindings/soc/renesas/renesas,r8a78000-mfis.yaml b/Documentation/devicetree/bindings/soc/renesas/renesas,r8a78000-mfis.yaml
index eef8c0a59e9c..72b9fa67d1f7 100644
--- a/Documentation/devicetree/bindings/soc/renesas/renesas,r8a78000-mfis.yaml
+++ b/Documentation/devicetree/bindings/soc/renesas/renesas,r8a78000-mfis.yaml
@@ -19,25 +19,29 @@ description:
 properties:
   compatible:
     enum:
+      - renesas,r8a779g0-mfis       # R-Car V4H
+      - renesas,r8a779h0-mfis       # R-Car V4M
       - renesas,r8a78000-mfis       # R-Car X5H (AP<->AP, with PRR)
       - renesas,r8a78000-mfis-scp   # R-Car X5H (AP<->SCP, without PRR)
 
   reg:
+    minItems: 1
     maxItems: 2
 
   reg-names:
+    minItems: 1
     items:
       - const: common
       - const: mboxes
 
   interrupts:
-    minItems: 32
+    minItems: 12
     maxItems: 128
     description:
       The interrupts raised by the remote doorbells.
 
   interrupt-names:
-    minItems: 32
+    minItems: 12
     maxItems: 128
     description:
       An interrupt name is constructed with the prefix 'ch'. Then, the
@@ -56,6 +60,26 @@ properties:
       <dt-bindings/soc/renesas,r8a78000-mfis.h>.
 
 allOf:
+  - if:
+      properties:
+        compatible:
+          contains:
+            enum:
+              - renesas,r8a779g0-mfis
+              - renesas,r8a779h0-mfis
+    then:
+      properties:
+        reg:
+          maxItems: 1
+        reg-names:
+          maxItems: 1
+        interrupts:
+          maxItems: 12
+        interrupt-names:
+          maxItems: 12
+          items:
+            pattern: "^ch[0-9]+e$"
+
   - if:
       properties:
         compatible:
@@ -63,6 +87,10 @@ allOf:
             const: renesas,r8a78000-mfis
     then:
       properties:
+        reg:
+          minItems: 2
+        reg-names:
+          minItems: 2
         interrupts:
           minItems: 128
         interrupt-names:
@@ -77,9 +105,15 @@ allOf:
             const: renesas,r8a78000-mfis-scp
     then:
       properties:
+        reg:
+          minItems: 2
+        reg-names:
+          minItems: 2
         interrupts:
+          minItems: 32
           maxItems: 32
         interrupt-names:
+          minItems: 32
           maxItems: 32
           items:
             pattern: "^ch[0-9]+i$"
-- 
2.51.0

[PATCH v2 2/2] soc: renesas: mfis: add R-Car V4H/V4M support

[Wolfram Sang]

The above SoCs have a weird register layout for the mailbox registers.
So, encapsulate register offset calculation in a per-SoC callback. Other
than that, only a separate config struct and compatibles are needed.

Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
---
 drivers/soc/renesas/rcar-mfis.c | 44 +++++++++++++++++++++++++++++++--
 1 file changed, 42 insertions(+), 2 deletions(-)

diff --git a/drivers/soc/renesas/rcar-mfis.c b/drivers/soc/renesas/rcar-mfis.c
index b334b35306ae..3435c3e16198 100644
--- a/drivers/soc/renesas/rcar-mfis.c
+++ b/drivers/soc/renesas/rcar-mfis.c
@@ -41,6 +41,7 @@ struct mfis_info {
 	unsigned int mb_reg_comes_from_dt:1;
 	unsigned int mb_tx_uses_eicr:1;
 	unsigned int mb_channels_are_unidir:1;
+	u32 (*mb_calc_reg)(u32 chan_num, bool tx_uses_eicr, bool is_only_rx);
 };
 
 struct mfis_chan_priv {
@@ -155,6 +156,35 @@ static const struct mbox_chan_ops mfis_iicr_ops = {
 	.last_tx_done = mfis_mb_iicr_last_tx_done,
 };
 
+static u32 mfis_mb_r8a779g0_calc_reg(u32 chan_num, bool tx_uses_eicr, bool is_only_rx)
+{
+	unsigned int i, k;
+	u32 reg;
+
+	i = chan_num & 3;
+	k = chan_num >> 2;
+
+	if (is_only_rx) {
+		if (k < 2)
+			reg = 0x9404 + 0x1020 * k + 0x08 * i;
+		else
+			reg = 0xb504 + 0x08 * i;
+	} else {
+		if (k < 2)
+			reg = 0x1400 + 0x1008 * i + 0x20 * k;
+		else
+			reg = 0x1500 + 0x1008 * i;
+	}
+
+	return reg;
+}
+
+static u32 mfis_mb_r8a78000_calc_reg(u32 chan_num, bool tx_uses_eicr, bool is_only_rx)
+{
+	return (tx_uses_eicr ^ is_only_rx) ? MFIS_X5H_EICR(chan_num) :
+					     MFIS_X5H_IICR(chan_num);
+}
+
 static struct mbox_chan *mfis_mb_of_xlate(struct mbox_controller *mbox,
 					  const struct of_phandle_args *sp)
 {
@@ -191,8 +221,7 @@ static struct mbox_chan *mfis_mb_of_xlate(struct mbox_controller *mbox,
 	}
 
 	chan_priv = chan->con_priv;
-	chan_priv->reg = (tx_uses_eicr ^ is_only_rx) ? MFIS_X5H_EICR(chan_num) :
-						       MFIS_X5H_IICR(chan_num);
+	chan_priv->reg = priv->info->mb_calc_reg(chan_num, tx_uses_eicr, is_only_rx);
 
 	if (!priv->info->mb_channels_are_unidir || is_only_rx) {
 		char irqname[8];
@@ -307,11 +336,19 @@ static int mfis_probe(struct platform_device *pdev)
 	return mfis_mb_probe(priv);
 }
 
+static const struct mfis_info mfis_info_r8a779g0 = {
+	.unprotect_mask	= 0x0000ffff,
+	.mb_num_channels = 12,
+	.mb_channels_are_unidir = true,
+	.mb_calc_reg = mfis_mb_r8a779g0_calc_reg,
+};
+
 static const struct mfis_info mfis_info_r8a78000 = {
 	.unprotect_mask	= 0x000fffff,
 	.mb_num_channels = 64,
 	.mb_reg_comes_from_dt = true,
 	.mb_channels_are_unidir = true,
+	.mb_calc_reg = mfis_mb_r8a78000_calc_reg,
 };
 
 static const struct mfis_info mfis_info_r8a78000_scp = {
@@ -319,9 +356,12 @@ static const struct mfis_info mfis_info_r8a78000_scp = {
 	.mb_num_channels = 32,
 	.mb_tx_uses_eicr = true,
 	.mb_channels_are_unidir = true,
+	.mb_calc_reg = mfis_mb_r8a78000_calc_reg,
 };
 
 static const struct of_device_id mfis_mfd_of_match[] = {
+	{ .compatible = "renesas,r8a779g0-mfis", .data = &mfis_info_r8a779g0, },
+	{ .compatible = "renesas,r8a779h0-mfis", .data = &mfis_info_r8a779g0, },
 	{ .compatible = "renesas,r8a78000-mfis", .data = &mfis_info_r8a78000, },
 	{ .compatible = "renesas,r8a78000-mfis-scp", .data = &mfis_info_r8a78000_scp, },
 	{}
-- 
2.51.0

Re: [PATCH v2 1/2] dt-bindings: soc: renesas: mfis: add R-Car V4H/V4M support

[Conor Dooley]

[-- Attachment #1: Type: text/plain, Size: 75 bytes --]

Acked-by: Conor Dooley <conor.dooley@microchip.com>
pw-bot: not-applicable

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 228 bytes --]

Re: [PATCH v2 0/2] soc: renesas: mfis: add R-Car V4H/V4M support

[Wolfram Sang]

[-- Attachment #1: Type: text/plain, Size: 613 bytes --]

Hi Geert,

On Tue, May 19, 2026 at 09:56:17AM +0200, Wolfram Sang wrote:
> Here are the additions to the MFIS driver in -next to add V4H support.
> A full branch for testing with the SparrowHawk is here:
> 
> git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux.git renesas/v4h/mfis
> 
> To trigger the mailbox-test instances, I use:
> 
> $ for f in /sys/kernel/debug/e6300*/message; do echo "Hallo $f!" > $f; done
> 
> Looking forward to comments!
> 
> Happy hacking,
> 
>    Wolfram

Any chance to get this into 7.2? All tags are there AFAICS.

Thanks and happy hacking,

   Wolfram


[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2 1/2] dt-bindings: soc: renesas: mfis: add R-Car V4H/V4M support

[Geert Uytterhoeven]

On Tue, 19 May 2026 at 09:56, Wolfram Sang
<wsa+renesas@sang-engineering.com> wrote:
> The above SoCs have only 12 mailboxes and do not have an extra register
> space for mailboxes. Everything is contained in the common register set.
> In addition to adding these SoCs, the other entries get updated to
> enforce 2 register spaces and their specific number of interrupts.
>
> Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
i.e. will queue in renesas-devel for v7.2.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH v2 2/2] soc: renesas: mfis: add R-Car V4H/V4M support

[Geert Uytterhoeven]

On Tue, 19 May 2026 at 09:56, Wolfram Sang
<wsa+renesas@sang-engineering.com> wrote:
> The above SoCs have a weird register layout for the mailbox registers.
> So, encapsulate register offset calculation in a per-SoC callback. Other
> than that, only a separate config struct and compatibles are needed.
>
> Signed-off-by: Wolfram Sang <wsa+renesas@sang-engineering.com>
> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
i.e. will queue in renesas-devel for v7.2.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds