[PATCH 0/3] fbdev: Request memory regions in platform drivers

[PATCH 0/3] fbdev: Request memory regions in platform drivers

[Amit Barzilai]

Several fbdev platform drivers call ioremap() on their hardware MMIO
without first claiming the range via request_mem_region(). This leaves
the kernel resource tree (/proc/iomem) with no record of the mapping,
allowing another driver to silently map the same registers.

This series fixes three platform_device drivers by switching to helpers
that combine resource claiming and ioremap in a single managed call.
cobalt_lcdfb and clps711x-fb are converted to
devm_platform_get_and_ioremap_resource(); goldfishfb is converted to
devm_ioremap_resource(), which also lets us drop the manual iounmap()
calls from the error path and remove function.

For clps711x-fb, resource 1 (the framebuffer range) already used
devm_platform_get_and_ioremap_resource() correctly; this series makes
resource 0 (the MMIO control registers) consistent with it.

This is part of the ongoing effort described in
Documentation/drm/todo.rst ("Request memory regions in all fbdev
drivers").

Amit Barzilai (3):
  fbdev: cobalt_lcdfb: Request memory region
  fbdev: clps711x-fb: Request memory region for MMIO
  fbdev: goldfishfb: Request memory region

 drivers/video/fbdev/clps711x-fb.c  |  9 +++------
 drivers/video/fbdev/cobalt_lcdfb.c | 12 +++---------
 drivers/video/fbdev/goldfishfb.c   |  8 +++-----
 3 files changed, 9 insertions(+), 20 deletions(-)


base-commit: 3f9357c30a44734d45e3093c521d52b2aefb09f5
-- 
2.53.0

[PATCH 1/3] fbdev: cobalt_lcdfb: Request memory region

[Amit Barzilai]

Use devm_platform_get_and_ioremap_resource() instead of open-coding
platform_get_resource() and devm_ioremap() separately. The helper
requests the memory region before mapping it, which registers the range
in /proc/iomem and prevents another driver from mapping the same
registers.

Assisted-by: Claude:claude-sonnet-4-6
---
 drivers/video/fbdev/cobalt_lcdfb.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/drivers/video/fbdev/cobalt_lcdfb.c b/drivers/video/fbdev/cobalt_lcdfb.c
index 308967b5096a..f7faa95fefd3 100644
--- a/drivers/video/fbdev/cobalt_lcdfb.c
+++ b/drivers/video/fbdev/cobalt_lcdfb.c
@@ -295,19 +295,13 @@ static int cobalt_lcdfb_probe(struct platform_device *dev)
 	if (!info)
 		return -ENOMEM;
 
-	res = platform_get_resource(dev, IORESOURCE_MEM, 0);
-	if (!res) {
+	info->screen_base = devm_platform_get_and_ioremap_resource(dev, 0, &res);
+	if (IS_ERR(info->screen_base)) {
 		framebuffer_release(info);
-		return -EBUSY;
+		return PTR_ERR(info->screen_base);
 	}
 
 	info->screen_size = resource_size(res);
-	info->screen_base = devm_ioremap(&dev->dev, res->start,
-					 info->screen_size);
-	if (!info->screen_base) {
-		framebuffer_release(info);
-		return -ENOMEM;
-	}
 
 	info->fbops = &cobalt_lcd_fbops;
 	info->fix = cobalt_lcdfb_fix;
-- 
2.53.0

Re: [PATCH 1/3] fbdev: cobalt_lcdfb: Request memory region

[Helge Deller]

Hi Amit,

On 4/20/26 15:44, Amit Barzilai wrote:
> Use devm_platform_get_and_ioremap_resource() instead of open-coding
> platform_get_resource() and devm_ioremap() separately. The helper
> requests the memory region before mapping it, which registers the range
> in /proc/iomem and prevents another driver from mapping the same
> registers.
> 
> Assisted-by: Claude:claude-sonnet-4-6

You missed your signed-off line, which I added while applying patches #1 and #2:
Signed-off-by: Amit Barzilai <amit.barzilai22@gmail.com>

Amit, please confirm that you wrote patches #2 and #3 and that you
are OK with the added Signed-off-by line, otherwise I need to drop the patches.

Helge

> ---
>   drivers/video/fbdev/cobalt_lcdfb.c | 12 +++---------
>   1 file changed, 3 insertions(+), 9 deletions(-)

Claude review: fbdev: cobalt_lcdfb: Request memory region

[Claude Code Review Bot]

Patch Review

**Status: Good**

Replaces separate `platform_get_resource()` + `devm_ioremap()` with a single `devm_platform_get_and_ioremap_resource()` call. The conversion is correct:

```c
info->screen_base = devm_platform_get_and_ioremap_resource(dev, 0, &res);
if (IS_ERR(info->screen_base)) {
    framebuffer_release(info);
    return PTR_ERR(info->screen_base);
}
```

The `res` output parameter is correctly used on the following line (`info->screen_size = resource_size(res)`), and the reordering (screen_base assigned before screen_size) is fine since `screen_size` is not needed for the mapping itself.

The original code returned `-EBUSY` when the resource wasn't found, which was arguably the wrong error code. The new code will return the actual error from the helper (likely `-EINVAL`), which is a minor improvement.

No issues.

---
Generated by Claude Code Patch Reviewer

[PATCH 2/3] fbdev: clps711x-fb: Request memory region for MMIO

[Amit Barzilai]

Use devm_platform_get_and_ioremap_resource() for resource 0 (the MMIO
control register range) instead of open-coding platform_get_resource()
and devm_ioremap() separately. The helper requests the memory region
before mapping it, which registers the range in /proc/iomem and prevents
another driver from mapping the same registers.

This makes resource 0 consistent with resource 1 (the framebuffer),
which already uses devm_platform_get_and_ioremap_resource().

Assisted-by: Claude:claude-sonnet-4-6
---
 drivers/video/fbdev/clps711x-fb.c | 9 +++------
 1 file changed, 3 insertions(+), 6 deletions(-)

diff --git a/drivers/video/fbdev/clps711x-fb.c b/drivers/video/fbdev/clps711x-fb.c
index 5e61a349a4ab..7a7db7100499 100644
--- a/drivers/video/fbdev/clps711x-fb.c
+++ b/drivers/video/fbdev/clps711x-fb.c
@@ -216,12 +216,9 @@ static int clps711x_fb_probe(struct platform_device *pdev)
 	cfb = info->par;
 	platform_set_drvdata(pdev, info);
 
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-	if (!res)
-		goto out_fb_release;
-	cfb->base = devm_ioremap(dev, res->start, resource_size(res));
-	if (!cfb->base) {
-		ret = -ENOMEM;
+	cfb->base = devm_platform_get_and_ioremap_resource(pdev, 0, &res);
+	if (IS_ERR(cfb->base)) {
+		ret = PTR_ERR(cfb->base);
 		goto out_fb_release;
 	}
 
-- 
2.53.0

Claude review: fbdev: clps711x-fb: Request memory region for MMIO

[Claude Code Review Bot]

Patch Review

**Status: Good**

Makes resource 0 (MMIO registers) consistent with resource 1 (framebuffer), which already uses `devm_platform_get_and_ioremap_resource()` at line 231 of the current source. Good consistency improvement.

```c
cfb->base = devm_platform_get_and_ioremap_resource(pdev, 0, &res);
if (IS_ERR(cfb->base)) {
    ret = PTR_ERR(cfb->base);
    goto out_fb_release;
}
```

The original code had a subtle deficiency: when `platform_get_resource()` returned NULL, it jumped to `out_fb_release` with `ret` still at its init value of `-ENOENT` (line 206 of current source). While not a bug per se, the new code is better — `ret` is explicitly set to the actual error from the helper.

`res` is still used afterward (lines 228-229: `info->fix.mmio_start = res->start; info->fix.mmio_len = resource_size(res)`), and it's correctly populated by the `&res` output parameter.

No issues.

---
Generated by Claude Code Patch Reviewer

[PATCH 3/3] fbdev: goldfishfb: Request memory region

[Amit Barzilai]

Use devm_ioremap_resource() instead of plain ioremap(). The helper
requests the memory region before mapping it, which registers the range
in /proc/iomem and prevents another driver from mapping the same
registers. As it is device-managed, remove the corresponding iounmap()
calls from the error unwind path and the remove function.

Assisted-by: Claude:claude-sonnet-4-6
---
 drivers/video/fbdev/goldfishfb.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/drivers/video/fbdev/goldfishfb.c b/drivers/video/fbdev/goldfishfb.c
index ffe33a36b944..14a7cb0a7a1c 100644
--- a/drivers/video/fbdev/goldfishfb.c
+++ b/drivers/video/fbdev/goldfishfb.c
@@ -194,9 +194,9 @@ static int goldfish_fb_probe(struct platform_device *pdev)
 		ret = -ENODEV;
 		goto err_no_io_base;
 	}
-	fb->reg_base = ioremap(r->start, PAGE_SIZE);
-	if (fb->reg_base == NULL) {
-		ret = -ENOMEM;
+	fb->reg_base = devm_ioremap_resource(&pdev->dev, r);
+	if (IS_ERR(fb->reg_base)) {
+		ret = PTR_ERR(fb->reg_base);
 		goto err_no_io_base;
 	}
 
@@ -273,7 +273,6 @@ static int goldfish_fb_probe(struct platform_device *pdev)
 				fb->fb.fix.smem_start);
 err_alloc_screen_base_failed:
 err_no_irq:
-	iounmap(fb->reg_base);
 err_no_io_base:
 	kfree(fb);
 err_fb_alloc_failed:
@@ -291,7 +290,6 @@ static void goldfish_fb_remove(struct platform_device *pdev)
 
 	dma_free_coherent(&pdev->dev, framesize, (void *)fb->fb.screen_base,
 						fb->fb.fix.smem_start);
-	iounmap(fb->reg_base);
 	kfree(fb);
 }
 
-- 
2.53.0

Re: [PATCH 3/3] fbdev: goldfishfb: Request memory region

[Helge Deller]

On 4/20/26 15:44, Amit Barzilai wrote:
> Use devm_ioremap_resource() instead of plain ioremap(). The helper
> requests the memory region before mapping it, which registers the range
> in /proc/iomem and prevents another driver from mapping the same
> registers. As it is device-managed, remove the corresponding iounmap()
> calls from the error unwind path and the remove function.
> 
> Assisted-by: Claude:claude-sonnet-4-6
> ---
>   drivers/video/fbdev/goldfishfb.c | 8 +++-----
>   1 file changed, 3 insertions(+), 5 deletions(-)

It was alreads fixed by another patch upstream...

Helge

Claude review: fbdev: goldfishfb: Request memory region

[Claude Code Review Bot]

Patch Review

**Status: Good, with one observation**

Converts from bare `ioremap()` to `devm_ioremap_resource()`, and correctly removes the manual `iounmap()` calls from both the error unwind path and `goldfish_fb_remove()`.

```c
fb->reg_base = devm_ioremap_resource(&pdev->dev, r);
if (IS_ERR(fb->reg_base)) {
    ret = PTR_ERR(fb->reg_base);
    goto err_no_io_base;
}
```

**Observation on mapping size:** The original code maps exactly `PAGE_SIZE` bytes:
```c
fb->reg_base = ioremap(r->start, PAGE_SIZE);
```
The new `devm_ioremap_resource()` maps `resource_size(r)` bytes. The actual register space only extends to offset `0x20` (36 bytes), so both PAGE_SIZE and the platform-defined resource size are more than sufficient. Mapping exactly what the resource describes is the right thing to do, so this is a correct change — but it's worth the reviewer being aware that the mapped size is now determined by the platform resource rather than hardcoded.

**Cleanup correctness:** The `iounmap()` removal is safe. The `kfree(fb)` calls in both the error path and `goldfish_fb_remove()` are retained, which is correct — freeing the struct that holds the `reg_base` pointer doesn't conflict with the devm-managed mapping lifetime (the mapping is tied to the device, not the struct).

No issues.

---
Generated by Claude Code Patch Reviewer

Claude review: fbdev: Request memory regions in platform drivers

[Claude Code Review Bot]

Overall Series Review

Subject: fbdev: Request memory regions in platform drivers
Author: Amit Barzilai <amit.barzilai22@gmail.com>
Patches: 6
Reviewed: 2026-04-23T09:34:40.480748

---

This is a clean, well-motivated 3-patch series converting three fbdev platform drivers to use managed ioremap helpers that also claim the memory region via `request_mem_region()`. The work addresses a documented TODO item in `Documentation/drm/todo.rst`. Each patch is small, self-contained, and mechanically correct. The commit messages are clear and explain the "why" well.

One process note: all three patches carry `Assisted-by: Claude:claude-sonnet-4-6`. The kernel community hasn't settled on a standard trailer format for AI assistance, and this format (`Name:model-id`) doesn't follow the standard trailer convention of `Tag: Name <email>`. The author should check current mailing list guidance on the preferred format.

No functional bugs found. One minor behavioral observation on patch 3 noted below.

---
Generated by Claude Code Patch Reviewer