[PATCH net-next 0/4] ptp: vmclock: bugfixes and cleanups for error handling

[PATCH net-next 0/4] ptp: vmclock: bugfixes and cleanups for error handling

[Thomas Weißschuh]

Some error handling issues I noticed while looking at the code.

Only compile-tested.

Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
---
Thomas Weißschuh (4):
      ptp: vmclock: Set driver data before its usage
      ptp: vmclock: Don't unregister misc device if it was not registered
      ptp: vmclock: Clean up miscdev and ptp clock through devres
      ptp: vmclock: Remove goto-based cleanup logic

 drivers/ptp/ptp_vmclock.c | 46 ++++++++++++++++++++--------------------------
 1 file changed, 20 insertions(+), 26 deletions(-)
---
base-commit: 2014c95afecee3e76ca4a56956a936e23283f05b
change-id: 20250206-vmclock-probe-57cbcb770925

Best regards,
-- 
Thomas Weißschuh <thomas.weissschuh@linutronix.de>

[PATCH net-next 1/4] ptp: vmclock: Set driver data before its usage

[Thomas Weißschuh]

If vmlock_ptp_register() fails during probing, vmclock_remove() is
called to clean up the ptp clock and misc device.
It uses dev_get_drvdata() to access the vmclock state.
However the driver data is not yet set at this point.

Assign the driver data earlier.

Fixes: 205032724226 ("ptp: Add support for the AMZNC10C 'vmclock' device")
Cc: stable@vger.kernel.org
Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
---
 drivers/ptp/ptp_vmclock.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/ptp/ptp_vmclock.c b/drivers/ptp/ptp_vmclock.c
index 0a2cfc8ad3c5408a87fd8fedeff274ab895de3dd..1920698ae6eba6abfff5b61afae1b047910026fd 100644
--- a/drivers/ptp/ptp_vmclock.c
+++ b/drivers/ptp/ptp_vmclock.c
@@ -524,6 +524,8 @@ static int vmclock_probe(struct platform_device *pdev)
 		goto out;
 	}
 
+	dev_set_drvdata(dev, st);
+
 	if (le32_to_cpu(st->clk->magic) != VMCLOCK_MAGIC ||
 	    le32_to_cpu(st->clk->size) > resource_size(&st->res) ||
 	    le16_to_cpu(st->clk->version) != 1) {
@@ -587,8 +589,6 @@ static int vmclock_probe(struct platform_device *pdev)
 		 (st->miscdev.minor && st->ptp_clock) ? ", " : "",
 		 st->ptp_clock ? "PTP" : "");
 
-	dev_set_drvdata(dev, st);
-
  out:
 	return ret;
 }

-- 
2.48.1

[PATCH net-next 2/4] ptp: vmclock: Don't unregister misc device if it was not registered

[Thomas Weißschuh]

vmclock_remove() tries to detect the successful registration of the misc
device based on the value of its minor value.
However that check is incorrect if the misc device registration was not
attempted in the first place.

Always initialize the minor number, so the check works properly.

Fixes: 205032724226 ("ptp: Add support for the AMZNC10C 'vmclock' device")
Cc: stable@vger.kernel.org
Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
---
 drivers/ptp/ptp_vmclock.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/ptp/ptp_vmclock.c b/drivers/ptp/ptp_vmclock.c
index 1920698ae6eba6abfff5b61afae1b047910026fd..82e6bef72b1b6edef7d891964c3f9c4546f6ddba 100644
--- a/drivers/ptp/ptp_vmclock.c
+++ b/drivers/ptp/ptp_vmclock.c
@@ -549,6 +549,8 @@ static int vmclock_probe(struct platform_device *pdev)
 		goto out;
 	}
 
+	st->miscdev.minor = MISC_DYNAMIC_MINOR;
+
 	/*
 	 * If the structure is big enough, it can be mapped to userspace.
 	 * Theoretically a guest OS even using larger pages could still
@@ -556,7 +558,6 @@ static int vmclock_probe(struct platform_device *pdev)
 	 * cross that bridge if/when we come to it.
 	 */
 	if (le32_to_cpu(st->clk->size) >= PAGE_SIZE) {
-		st->miscdev.minor = MISC_DYNAMIC_MINOR;
 		st->miscdev.fops = &vmclock_miscdev_fops;
 		st->miscdev.name = st->name;
 

-- 
2.48.1

[PATCH net-next 3/4] ptp: vmclock: Clean up miscdev and ptp clock through devres

[Thomas Weißschuh]

Most resources owned by the vmclock device are managed through devres.
Only the miscdev and ptp clock are managed manually.
This makes the code slightly harder to understand than necessary.

Switch them over to devres and remove the now unnecessary drvdata.

Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
---
 drivers/ptp/ptp_vmclock.c | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

diff --git a/drivers/ptp/ptp_vmclock.c b/drivers/ptp/ptp_vmclock.c
index 82e6bef72b1b6edef7d891964c3f9c4546f6ddba..b7fa196975229cec6d27b34f7bb235c53b5b5732 100644
--- a/drivers/ptp/ptp_vmclock.c
+++ b/drivers/ptp/ptp_vmclock.c
@@ -420,10 +420,9 @@ static const struct file_operations vmclock_miscdev_fops = {
 
 /* module operations */
 
-static void vmclock_remove(struct platform_device *pdev)
+static void vmclock_remove(void *data)
 {
-	struct device *dev = &pdev->dev;
-	struct vmclock_state *st = dev_get_drvdata(dev);
+	struct vmclock_state *st = data;
 
 	if (st->ptp_clock)
 		ptp_clock_unregister(st->ptp_clock);
@@ -524,8 +523,6 @@ static int vmclock_probe(struct platform_device *pdev)
 		goto out;
 	}
 
-	dev_set_drvdata(dev, st);
-
 	if (le32_to_cpu(st->clk->magic) != VMCLOCK_MAGIC ||
 	    le32_to_cpu(st->clk->size) > resource_size(&st->res) ||
 	    le16_to_cpu(st->clk->version) != 1) {
@@ -551,6 +548,10 @@ static int vmclock_probe(struct platform_device *pdev)
 
 	st->miscdev.minor = MISC_DYNAMIC_MINOR;
 
+	ret = devm_add_action_or_reset(&pdev->dev, vmclock_remove, st);
+	if (ret)
+		goto out;
+
 	/*
 	 * If the structure is big enough, it can be mapped to userspace.
 	 * Theoretically a guest OS even using larger pages could still
@@ -573,7 +574,6 @@ static int vmclock_probe(struct platform_device *pdev)
 		if (IS_ERR(st->ptp_clock)) {
 			ret = PTR_ERR(st->ptp_clock);
 			st->ptp_clock = NULL;
-			vmclock_remove(pdev);
 			goto out;
 		}
 	}
@@ -602,7 +602,6 @@ MODULE_DEVICE_TABLE(acpi, vmclock_acpi_ids);
 
 static struct platform_driver vmclock_platform_driver = {
 	.probe		= vmclock_probe,
-	.remove		= vmclock_remove,
 	.driver	= {
 		.name	= "vmclock",
 		.acpi_match_table = vmclock_acpi_ids,

-- 
2.48.1

[PATCH net-next 4/4] ptp: vmclock: Remove goto-based cleanup logic

[Thomas Weißschuh]

vmclock_probe() uses an "out:" label to return from the function on
error. This indicates that some cleanup operation is necessary.
However the label does not do anything as all resources are managed
through devres, making the code slightly harder to read.

Remove the label and just return directly.

Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
---
 drivers/ptp/ptp_vmclock.c | 32 +++++++++++++-------------------
 1 file changed, 13 insertions(+), 19 deletions(-)

diff --git a/drivers/ptp/ptp_vmclock.c b/drivers/ptp/ptp_vmclock.c
index b7fa196975229cec6d27b34f7bb235c53b5b5732..26fcc1185ee72999a43842a59164498657799b1f 100644
--- a/drivers/ptp/ptp_vmclock.c
+++ b/drivers/ptp/ptp_vmclock.c
@@ -505,14 +505,13 @@ static int vmclock_probe(struct platform_device *pdev)
 
 	if (ret) {
 		dev_info(dev, "Failed to obtain physical address: %d\n", ret);
-		goto out;
+		return ret;
 	}
 
 	if (resource_size(&st->res) < VMCLOCK_MIN_SIZE) {
 		dev_info(dev, "Region too small (0x%llx)\n",
 			 resource_size(&st->res));
-		ret = -EINVAL;
-		goto out;
+		return -EINVAL;
 	}
 	st->clk = devm_memremap(dev, st->res.start, resource_size(&st->res),
 				MEMREMAP_WB | MEMREMAP_DEC);
@@ -520,37 +519,34 @@ static int vmclock_probe(struct platform_device *pdev)
 		ret = PTR_ERR(st->clk);
 		dev_info(dev, "failed to map shared memory\n");
 		st->clk = NULL;
-		goto out;
+		return ret;
 	}
 
 	if (le32_to_cpu(st->clk->magic) != VMCLOCK_MAGIC ||
 	    le32_to_cpu(st->clk->size) > resource_size(&st->res) ||
 	    le16_to_cpu(st->clk->version) != 1) {
 		dev_info(dev, "vmclock magic fields invalid\n");
-		ret = -EINVAL;
-		goto out;
+		return -EINVAL;
 	}
 
 	ret = ida_alloc(&vmclock_ida, GFP_KERNEL);
 	if (ret < 0)
-		goto out;
+		return ret;
 
 	st->index = ret;
 	ret = devm_add_action_or_reset(&pdev->dev, vmclock_put_idx, st);
 	if (ret)
-		goto out;
+		return ret;
 
 	st->name = devm_kasprintf(&pdev->dev, GFP_KERNEL, "vmclock%d", st->index);
-	if (!st->name) {
-		ret = -ENOMEM;
-		goto out;
-	}
+	if (!st->name)
+		return -ENOMEM;
 
 	st->miscdev.minor = MISC_DYNAMIC_MINOR;
 
 	ret = devm_add_action_or_reset(&pdev->dev, vmclock_remove, st);
 	if (ret)
-		goto out;
+		return ret;
 
 	/*
 	 * If the structure is big enough, it can be mapped to userspace.
@@ -564,7 +560,7 @@ static int vmclock_probe(struct platform_device *pdev)
 
 		ret = misc_register(&st->miscdev);
 		if (ret)
-			goto out;
+			return ret;
 	}
 
 	/* If there is valid clock information, register a PTP clock */
@@ -574,15 +570,14 @@ static int vmclock_probe(struct platform_device *pdev)
 		if (IS_ERR(st->ptp_clock)) {
 			ret = PTR_ERR(st->ptp_clock);
 			st->ptp_clock = NULL;
-			goto out;
+			return ret;
 		}
 	}
 
 	if (!st->miscdev.minor && !st->ptp_clock) {
 		/* Neither miscdev nor PTP registered */
 		dev_info(dev, "vmclock: Neither miscdev nor PTP available; not registering\n");
-		ret = -ENODEV;
-		goto out;
+		return -ENODEV;
 	}
 
 	dev_info(dev, "%s: registered %s%s%s\n", st->name,
@@ -590,8 +585,7 @@ static int vmclock_probe(struct platform_device *pdev)
 		 (st->miscdev.minor && st->ptp_clock) ? ", " : "",
 		 st->ptp_clock ? "PTP" : "");
 
- out:
-	return ret;
+	return 0;
 }
 
 static const struct acpi_device_id vmclock_acpi_ids[] = {

-- 
2.48.1

Re: [PATCH net-next 0/4] ptp: vmclock: bugfixes and cleanups for error handling

[Richard Cochran]

On Thu, Feb 06, 2025 at 06:45:00PM +0100, Thomas Weißschuh wrote:
> Some error handling issues I noticed while looking at the code.
> 
> Only compile-tested.
> 
> Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>

For the series:

Acked-by: Richard Cochran <richardcochran@gmail.com>

Re: [PATCH net-next 1/4] ptp: vmclock: Set driver data before its usage

[Mateusz Polchlopek]

On 2/6/2025 6:45 PM, Thomas Weißschuh wrote:
> If vmlock_ptp_register() fails during probing, vmclock_remove() is

Typo: you missed 'c' in function name - vmclock_ptp_register

> called to clean up the ptp clock and misc device.
> It uses dev_get_drvdata() to access the vmclock state.
> However the driver data is not yet set at this point.
> 
> Assign the driver data earlier.
> 
> Fixes: 205032724226 ("ptp: Add support for the AMZNC10C 'vmclock' device")
> Cc: stable@vger.kernel.org
> Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
> ---
>   drivers/ptp/ptp_vmclock.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/ptp/ptp_vmclock.c b/drivers/ptp/ptp_vmclock.c
> index 0a2cfc8ad3c5408a87fd8fedeff274ab895de3dd..1920698ae6eba6abfff5b61afae1b047910026fd 100644
> --- a/drivers/ptp/ptp_vmclock.c
> +++ b/drivers/ptp/ptp_vmclock.c
> @@ -524,6 +524,8 @@ static int vmclock_probe(struct platform_device *pdev)
>   		goto out;
>   	}
>   
> +	dev_set_drvdata(dev, st);
> +
>   	if (le32_to_cpu(st->clk->magic) != VMCLOCK_MAGIC ||
>   	    le32_to_cpu(st->clk->size) > resource_size(&st->res) ||
>   	    le16_to_cpu(st->clk->version) != 1) {
> @@ -587,8 +589,6 @@ static int vmclock_probe(struct platform_device *pdev)
>   		 (st->miscdev.minor && st->ptp_clock) ? ", " : "",
>   		 st->ptp_clock ? "PTP" : "");
>   
> -	dev_set_drvdata(dev, st);
> -
>    out:
>   	return ret;
>   }
> 

Nice catch! The code looks good but please fix the typo in the
commit msg, as this may be misleading. When applied please add my:

Reviewed-by: Mateusz Polchlopek <mateusz.polchlopek@intel.com>

Re: [PATCH net-next 0/4] ptp: vmclock: bugfixes and cleanups for error handling

[Mateusz Polchlopek]

On 2/6/2025 6:45 PM, Thomas Weißschuh wrote:
> Some error handling issues I noticed while looking at the code.
> 
> Only compile-tested.
> 
> Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
> ---
> Thomas Weißschuh (4):
>        ptp: vmclock: Set driver data before its usage
>        ptp: vmclock: Don't unregister misc device if it was not registered
>        ptp: vmclock: Clean up miscdev and ptp clock through devres
>        ptp: vmclock: Remove goto-based cleanup logic
> 
>   drivers/ptp/ptp_vmclock.c | 46 ++++++++++++++++++++--------------------------
>   1 file changed, 20 insertions(+), 26 deletions(-)
> ---
> base-commit: 2014c95afecee3e76ca4a56956a936e23283f05b
> change-id: 20250206-vmclock-probe-57cbcb770925
> 
> Best regards,

As those all are fixes and cleanups then I think it should be tagged to
net not net-next.

thanks

Re: [PATCH net-next 0/4] ptp: vmclock: bugfixes and cleanups for error handling

[David Woodhouse]

[-- Attachment #1: Type: text/plain, Size: 2209 bytes --]

On Fri, 2025-02-07 at 08:13 +0100, Mateusz Polchlopek wrote:
> 
> 
> On 2/6/2025 6:45 PM, Thomas Weißschuh wrote:
> > Some error handling issues I noticed while looking at the code.
> > 
> > Only compile-tested.
> > 
> > Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
> > ---
> > Thomas Weißschuh (4):
> >        ptp: vmclock: Set driver data before its usage
> >        ptp: vmclock: Don't unregister misc device if it was not registered
> >        ptp: vmclock: Clean up miscdev and ptp clock through devres
> >        ptp: vmclock: Remove goto-based cleanup logic
> > 
> >   drivers/ptp/ptp_vmclock.c | 46 ++++++++++++++++++++--------------------------
> >   1 file changed, 20 insertions(+), 26 deletions(-)
> > ---
> > base-commit: 2014c95afecee3e76ca4a56956a936e23283f05b
> > change-id: 20250206-vmclock-probe-57cbcb770925
> > 
> > Best regards,
> 
> As those all are fixes and cleanups then I think it should be tagged to
> net not net-next.

Agreed. Thanks, Thomas. For all four:

Reviewed-by: David Woodhouse <dwmw@amazon.co.uk>

I'm about to post a fifth which adds a .owner to vmclock_miscdev_fops.

Tested with the existing '-device vmclock' support in QEMU, plus this
hack to actually expose a PTP clock to the guest (which we haven't
worked out how to do *properly* from the timekeeping subsystem of a
Linux host yet; qv).

--- a/hw/acpi/vmclock.c
+++ b/hw/acpi/vmclock.c
@@ -151,6 +151,18 @@ static void vmclock_realize(DeviceState *dev,
Error **errp)
 
     qemu_register_reset(vmclock_handle_reset, vms);
 
+	vms->clk->time_type = VMCLOCK_TIME_TAI;
+    vms->clk->flags = VMCLOCK_FLAG_TAI_OFFSET_VALID;
+    vms->clk->tai_offset_sec = -3600;
+    vms->clk->clock_status = VMCLOCK_STATUS_SYNCHRONIZED;
+    vms->clk->counter_value = 0;
+    vms->clk->counter_id = VMCLOCK_COUNTER_X86_TSC;
+    vms->clk->time_sec = 1704067200;
+    vms->clk->time_frac_sec = 0x8000000000000000ULL;
+    vms->clk->counter_period_frac_sec = 0x1a6e39b3e0ULL;
+    vms->clk->counter_period_shift = 4;
+    //vms->clk->counter_period_frac_sec = 0x1934c67f9b2ce6ULL;
+
     vmclock_update_guest(vms);
 }
 


[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5069 bytes --]

[PATCH net-next 5/4] ptp: vmclock: Add .owner to vmclock_miscdev_fops

[David Woodhouse]

[-- Attachment #1: Type: text/plain, Size: 831 bytes --]

From: David Woodhouse <dwmw@amazon.co.uk>

Without the .owner field, the module can be unloaded while /dev/vmclock0
is open, leading to an oops.

Fixes: 205032724226 ("ptp: Add support for the AMZNC10C 'vmclock' device")
Cc: stable@vger.kernel.org
Signed-off-by: David Woodhouse <dwmw@amazon.co.uk>
---
 drivers/ptp/ptp_vmclock.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/ptp/ptp_vmclock.c b/drivers/ptp/ptp_vmclock.c
index 26fcc1185ee7..b3a83b03d9c1 100644
--- a/drivers/ptp/ptp_vmclock.c
+++ b/drivers/ptp/ptp_vmclock.c
@@ -414,6 +414,7 @@ static ssize_t vmclock_miscdev_read(struct file *fp, char __user *buf,
 }
 
 static const struct file_operations vmclock_miscdev_fops = {
+	.owner = THIS_MODULE,
 	.mmap = vmclock_miscdev_mmap,
 	.read = vmclock_miscdev_read,
 };
-- 
2.48.1



[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5069 bytes --]

Re: [PATCH net-next 0/4] ptp: vmclock: bugfixes and cleanups for error handling

[Thomas Weißschuh]

On Fri, Feb 07, 2025 at 09:10:42AM +0000, David Woodhouse wrote:
> On Fri, 2025-02-07 at 08:13 +0100, Mateusz Polchlopek wrote:
> > 
> > 
> > On 2/6/2025 6:45 PM, Thomas Weißschuh wrote:
> > > Some error handling issues I noticed while looking at the code.
> > > 
> > > Only compile-tested.
> > > 
> > > Signed-off-by: Thomas Weißschuh <thomas.weissschuh@linutronix.de>
> > > ---
> > > Thomas Weißschuh (4):
> > >        ptp: vmclock: Set driver data before its usage
> > >        ptp: vmclock: Don't unregister misc device if it was not registered
> > >        ptp: vmclock: Clean up miscdev and ptp clock through devres
> > >        ptp: vmclock: Remove goto-based cleanup logic
> > > 
> > >   drivers/ptp/ptp_vmclock.c | 46 ++++++++++++++++++++--------------------------
> > >   1 file changed, 20 insertions(+), 26 deletions(-)
> > > ---
> > > base-commit: 2014c95afecee3e76ca4a56956a936e23283f05b
> > > change-id: 20250206-vmclock-probe-57cbcb770925
> > > 
> > > Best regards,
> > 
> > As those all are fixes and cleanups then I think it should be tagged to
> > net not net-next.
> 
> Agreed. Thanks, Thomas. For all four:

Ack.


> Reviewed-by: David Woodhouse <dwmw@amazon.co.uk>

Thanks.

> I'm about to post a fifth which adds a .owner to vmclock_miscdev_fops.

I assume you want me to include this in my series.

> Tested with the existing '-device vmclock' support in QEMU, plus this
> hack to actually expose a PTP clock to the guest (which we haven't
> worked out how to do *properly* from the timekeeping subsystem of a
> Linux host yet; qv).
> 
> --- a/hw/acpi/vmclock.c
> +++ b/hw/acpi/vmclock.c
> @@ -151,6 +151,18 @@ static void vmclock_realize(DeviceState *dev,
> Error **errp)
>  
>      qemu_register_reset(vmclock_handle_reset, vms);
>  
> +	vms->clk->time_type = VMCLOCK_TIME_TAI;
> +    vms->clk->flags = VMCLOCK_FLAG_TAI_OFFSET_VALID;
> +    vms->clk->tai_offset_sec = -3600;
> +    vms->clk->clock_status = VMCLOCK_STATUS_SYNCHRONIZED;
> +    vms->clk->counter_value = 0;
> +    vms->clk->counter_id = VMCLOCK_COUNTER_X86_TSC;
> +    vms->clk->time_sec = 1704067200;
> +    vms->clk->time_frac_sec = 0x8000000000000000ULL;
> +    vms->clk->counter_period_frac_sec = 0x1a6e39b3e0ULL;
> +    vms->clk->counter_period_shift = 4;
> +    //vms->clk->counter_period_frac_sec = 0x1934c67f9b2ce6ULL;
> +
>      vmclock_update_guest(vms);
>  }
>  
> 

Re: [PATCH net-next 0/4] ptp: vmclock: bugfixes and cleanups for error handling

[David Woodhouse]

[-- Attachment #1: Type: text/plain, Size: 346 bytes --]

On Fri, 2025-02-07 at 10:25 +0100, Thomas Weißschuh wrote:
> > I'm about to post a fifth which adds a .owner to vmclock_miscdev_fops.
> 
> I assume you want me to include this in my series.

If you do need to repost the series, yes please. Otherwise I've posted
it as [PATCH 5/4] so hopefully it'll get swept up with the original
set.


[-- Attachment #2: smime.p7s --]
[-- Type: application/pkcs7-signature, Size: 5069 bytes --]