[PATCH v2] drm: renesas: rz-du: mipi_dsi: Fix return path on error

[PATCH v2] drm: renesas: rz-du: mipi_dsi: Fix return path on error

[Chris Brandt]

In case of error, we should unwind correctly.
Switching to using dmam_ instead of dma_ and moving the code earlier
fixes the issue.

Fixes: 6f392f371650 ("drm: renesas: rz-du: Implement MIPI DSI host transfers")
Suggested-by: Pavel Machek <pavel@nabladev.com>
Signed-off-by: Chris Brandt <chris.brandt@renesas.com>
Acked-by: Hugo Villeneuve <hvilleneuve@dimonoff.com>

---
V1->V2:
 - Need to take into account mipi_dsi_host_register was already called (Geert)
 - Switch to using dmam_alloc_coherent to automate unwinding (Chris)
 - Added Acked-by (Hugo)
---
 drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c | 12 +++++-------
 1 file changed, 5 insertions(+), 7 deletions(-)

diff --git a/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c b/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c
index 29f2b7d24fe5..b908e702b607 100644
--- a/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c
+++ b/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c
@@ -1442,6 +1442,11 @@ static int rzg2l_mipi_dsi_probe(struct platform_device *pdev)
 		return dev_err_probe(dsi->dev, PTR_ERR(dsi->prstc),
 				     "failed to get prst\n");

+	dsi->dcs_buf_virt = dmam_alloc_coherent(dsi->dev, RZG2L_DCS_BUF_SIZE,
+						&dsi->dcs_buf_phys, GFP_KERNEL);
+	if (!dsi->dcs_buf_virt)
+		return -ENOMEM;
+
 	platform_set_drvdata(pdev, dsi);

 	pm_runtime_enable(dsi->dev);
@@ -1474,11 +1479,6 @@ static int rzg2l_mipi_dsi_probe(struct platform_device *pdev)
 	if (ret < 0)
 		goto err_pm_disable;

-	dsi->dcs_buf_virt = dma_alloc_coherent(dsi->host.dev, RZG2L_DCS_BUF_SIZE,
-					       &dsi->dcs_buf_phys, GFP_KERNEL);
-	if (!dsi->dcs_buf_virt)
-		return -ENOMEM;
-
 	return 0;

 err_phy:
@@ -1493,8 +1493,6 @@ static void rzg2l_mipi_dsi_remove(struct platform_device *pdev)
 {
 	struct rzg2l_mipi_dsi *dsi = platform_get_drvdata(pdev);

-	dma_free_coherent(dsi->host.dev, RZG2L_DCS_BUF_SIZE, dsi->dcs_buf_virt,
-			  dsi->dcs_buf_phys);
 	mipi_dsi_host_unregister(&dsi->host);
 	pm_runtime_disable(&pdev->dev);
 }
--
2.52.0

Re: [PATCH v2] drm: renesas: rz-du: mipi_dsi: Fix return path on error

[Hugo Villeneuve]

Hi Chris,

On Wed, 22 Apr 2026 10:51:46 -0400
Chris Brandt <chris.brandt@renesas.com> wrote:

> In case of error, we should unwind correctly.
> Switching to using dmam_ instead of dma_ and moving the code earlier
> fixes the issue.
> 
> Fixes: 6f392f371650 ("drm: renesas: rz-du: Implement MIPI DSI host transfers")
> Suggested-by: Pavel Machek <pavel@nabladev.com>
> Signed-off-by: Chris Brandt <chris.brandt@renesas.com>
> Acked-by: Hugo Villeneuve <hvilleneuve@dimonoff.com>
> 
> ---
> V1->V2:
>  - Need to take into account mipi_dsi_host_register was already called (Geert)
>  - Switch to using dmam_alloc_coherent to automate unwinding (Chris)
>  - Added Acked-by (Hugo)

You should not add "Acked-by" tags if you make
modifications _after_ someone has given its Ack (unless it is just a
minor cosmetic or coding style change)...

In other words, my ack was for V1, not V2 which is quite different from
V1 :)

Don't resend a V3 just for that, I will have a look later at V2...


> ---
>  drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c | 12 +++++-------
>  1 file changed, 5 insertions(+), 7 deletions(-)
> 
> diff --git a/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c b/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c
> index 29f2b7d24fe5..b908e702b607 100644
> --- a/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c
> +++ b/drivers/gpu/drm/renesas/rz-du/rzg2l_mipi_dsi.c
> @@ -1442,6 +1442,11 @@ static int rzg2l_mipi_dsi_probe(struct platform_device *pdev)
>  		return dev_err_probe(dsi->dev, PTR_ERR(dsi->prstc),
>  				     "failed to get prst\n");
> 
> +	dsi->dcs_buf_virt = dmam_alloc_coherent(dsi->dev, RZG2L_DCS_BUF_SIZE,
> +						&dsi->dcs_buf_phys, GFP_KERNEL);
> +	if (!dsi->dcs_buf_virt)
> +		return -ENOMEM;
> +
>  	platform_set_drvdata(pdev, dsi);
> 
>  	pm_runtime_enable(dsi->dev);
> @@ -1474,11 +1479,6 @@ static int rzg2l_mipi_dsi_probe(struct platform_device *pdev)
>  	if (ret < 0)
>  		goto err_pm_disable;
> 
> -	dsi->dcs_buf_virt = dma_alloc_coherent(dsi->host.dev, RZG2L_DCS_BUF_SIZE,
> -					       &dsi->dcs_buf_phys, GFP_KERNEL);
> -	if (!dsi->dcs_buf_virt)
> -		return -ENOMEM;
> -
>  	return 0;
> 
>  err_phy:
> @@ -1493,8 +1493,6 @@ static void rzg2l_mipi_dsi_remove(struct platform_device *pdev)
>  {
>  	struct rzg2l_mipi_dsi *dsi = platform_get_drvdata(pdev);
> 
> -	dma_free_coherent(dsi->host.dev, RZG2L_DCS_BUF_SIZE, dsi->dcs_buf_virt,
> -			  dsi->dcs_buf_phys);
>  	mipi_dsi_host_unregister(&dsi->host);
>  	pm_runtime_disable(&pdev->dev);
>  }
> --
> 2.52.0
> 
> 


-- 
Hugo Villeneuve

Claude review: drm: renesas: rz-du: mipi_dsi: Fix return path on error

[Claude Code Review Bot]

Patch Review

**The bug (confirmed in tree):** In the current code at line 1476-1479 of `rzg2l_mipi_dsi.c`:

```c
dsi->dcs_buf_virt = dma_alloc_coherent(dsi->host.dev, RZG2L_DCS_BUF_SIZE,
                                       &dsi->dcs_buf_phys, GFP_KERNEL);
if (!dsi->dcs_buf_virt)
    return -ENOMEM;
```

This bare `return -ENOMEM` leaks two resources:
1. The `mipi_dsi_host_register` at line 1472 is never unregistered.
2. The `pm_runtime_enable` at line 1446 is never disabled.

**The fix is correct and clean:**

1. **`dma_alloc_coherent` → `dmam_alloc_coherent`**: The managed variant ties the DMA buffer lifetime to the device via devres, so no explicit free is needed. This eliminates the `dma_free_coherent` call in `rzg2l_mipi_dsi_remove`, which is correct.

2. **Moving allocation before `platform_set_drvdata` / `pm_runtime_enable`**: At this point in probe, only `devm_`-managed resources (clocks, resets, ioremap) have been acquired. A plain `return -ENOMEM` is safe here — all prior allocations are device-managed and will be cleaned up automatically. The DMA buffer isn't used until actual DSI host transfers, so early allocation is fine.

3. **Device parameter change**: The original used `dsi->host.dev`, the new code uses `dsi->dev`. This is correct because `dsi->host.dev` hasn't been assigned yet at the new call site, and they are set to the same value anyway (line 1470: `dsi->host.dev = dsi->dev`).

**Minor observations (not blocking):**

- The commit message could be slightly more explicit about *what* was leaking (host registration + pm_runtime), but "unwind correctly" is adequate for a Fixes patch.

- The ordering in `remove` now has `mipi_dsi_host_unregister` first and `pm_runtime_disable` second, with the DMA free handled by devres (which runs after `remove` returns). This ordering is correct — devres cleanup happens during device unbind after the driver's remove callback.

**No issues found. Patch is ready.**

Reviewed-by: Dave Airlie <airlied@gmail.com>

---
Generated by Claude Code Patch Reviewer

Claude review: drm: renesas: rz-du: mipi_dsi: Fix return path on error

[Claude Code Review Bot]

Overall Series Review

Subject: drm: renesas: rz-du: mipi_dsi: Fix return path on error
Author: Chris Brandt <chris.brandt@renesas.com>
Patches: 2
Reviewed: 2026-04-23T07:31:41.505756

---

This is a single-patch fix for a resource leak bug in the RZ/G2L MIPI DSI driver's probe function. The original code calls `dma_alloc_coherent` after `mipi_dsi_host_register` and `pm_runtime_enable`, but on failure returns `-ENOMEM` directly — skipping the cleanup of the host registration and pm_runtime. The fix takes a clean approach: switch to the device-managed `dmam_alloc_coherent` and move the allocation earlier in probe, before any resources that would need unwinding. This is a well-structured v2 addressing reviewer feedback.

**Verdict: Looks good to merge.**

---
Generated by Claude Code Patch Reviewer