[PATCH 0/4] mfd: qnap-mcu: Some fixes / improvements

[PATCH 0/4] mfd: qnap-mcu: Some fixes / improvements

[Heiko Stuebner]

While digging through the mcu functions, I came across some deficits I
introduced with the initial driver submission, so trying to make that
a bit nicer with this series.


I struggled a bit with the ordering of qnap_mcu_write error-check
and conversion to guard(mutex). Converting to guard before the
error check would need dropping the ret variable, just to re-add it
one patch later - to not cause unused variable warning.

Additionally the error handling could be considered a fix, so makes
sense to have in front.


Heiko Stuebner (4):
  mfd: qnap-mcu: include linux/types.h in qnap-mcu.h shared header
  mfd: qnap-mcu: handle errors returned from qnap_mcu_write
  mfd: qnap-mcu: convert to guard(mutex) in qnap_mcu_exec
  mfd: qnap-mcu: improve structure in qnap_mcu_exec

 drivers/mfd/qnap-mcu.c       | 30 +++++++++++++++---------------
 include/linux/mfd/qnap-mcu.h |  2 ++
 2 files changed, 17 insertions(+), 15 deletions(-)

-- 
2.47.2

[PATCH 1/4] mfd: qnap-mcu: include linux/types.h in qnap-mcu.h shared header

[Heiko Stuebner]

Relying on other components to include those basic types is unrelyable
and may cause compile errors like:

../include/linux/mfd/qnap-mcu.h:13:9: error: unknown type name ‘u32’
   13 |         u32 baud_rate;
      |         ^~~
../include/linux/mfd/qnap-mcu.h:17:9: error: unknown type name ‘bool’
   17 |         bool usb_led;
      |         ^~~~

So make sure, the types used in the header are available.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
---
 include/linux/mfd/qnap-mcu.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/linux/mfd/qnap-mcu.h b/include/linux/mfd/qnap-mcu.h
index 8d48c212fd44..42bf523f9a5b 100644
--- a/include/linux/mfd/qnap-mcu.h
+++ b/include/linux/mfd/qnap-mcu.h
@@ -7,6 +7,8 @@
 #ifndef _LINUX_QNAP_MCU_H_
 #define _LINUX_QNAP_MCU_H_
 
+#include <linux/types.h>
+
 struct qnap_mcu;
 
 struct qnap_mcu_variant {
-- 
2.47.2

[PATCH 2/4] mfd: qnap-mcu: handle errors returned from qnap_mcu_write

[Heiko Stuebner]

qnap_mcu_write can return errors and those were not checked before.
So do that now.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
---
 drivers/mfd/qnap-mcu.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/mfd/qnap-mcu.c b/drivers/mfd/qnap-mcu.c
index e340e9337b98..6448237c4c35 100644
--- a/drivers/mfd/qnap-mcu.c
+++ b/drivers/mfd/qnap-mcu.c
@@ -163,7 +163,11 @@ int qnap_mcu_exec(struct qnap_mcu *mcu,
 	reply->received = 0;
 	reinit_completion(&reply->done);
 
-	qnap_mcu_write(mcu, cmd_data, cmd_data_size);
+	ret = qnap_mcu_write(mcu, cmd_data, cmd_data_size);
+	if (ret < 0) {
+		mutex_unlock(&mcu->bus_lock);
+		return ret;
+	}
 
 	serdev_device_wait_until_sent(mcu->serdev, msecs_to_jiffies(QNAP_MCU_TIMEOUT_MS));
 
-- 
2.47.2

[PATCH 3/4] mfd: qnap-mcu: convert to guard(mutex) in qnap_mcu_exec

[Heiko Stuebner]

guard() makes sure that the mutex gets unlocked when the function returns
and thus removes the need for unlock gotos or similar mechanisms and
therefore allows for a simpler function structure.

So convert the qnap_mcu_exec function to use it.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
---
 drivers/mfd/qnap-mcu.c | 13 +++++--------
 1 file changed, 5 insertions(+), 8 deletions(-)

diff --git a/drivers/mfd/qnap-mcu.c b/drivers/mfd/qnap-mcu.c
index 6448237c4c35..7bc44c64b7fd 100644
--- a/drivers/mfd/qnap-mcu.c
+++ b/drivers/mfd/qnap-mcu.c
@@ -156,7 +156,7 @@ int qnap_mcu_exec(struct qnap_mcu *mcu,
 		return -EINVAL;
 	}
 
-	mutex_lock(&mcu->bus_lock);
+	guard(mutex)(&mcu->bus_lock);
 
 	reply->data = rx;
 	reply->length = length;
@@ -164,30 +164,27 @@ int qnap_mcu_exec(struct qnap_mcu *mcu,
 	reinit_completion(&reply->done);
 
 	ret = qnap_mcu_write(mcu, cmd_data, cmd_data_size);
-	if (ret < 0) {
-		mutex_unlock(&mcu->bus_lock);
+	if (ret < 0)
 		return ret;
-	}
 
 	serdev_device_wait_until_sent(mcu->serdev, msecs_to_jiffies(QNAP_MCU_TIMEOUT_MS));
 
 	if (!wait_for_completion_timeout(&reply->done, msecs_to_jiffies(QNAP_MCU_TIMEOUT_MS))) {
 		dev_err(&mcu->serdev->dev, "Command timeout\n");
-		ret = -ETIMEDOUT;
+		return -ETIMEDOUT;
 	} else {
 		u8 crc = qnap_mcu_csum(rx, reply_data_size);
 
 		if (crc != rx[reply_data_size]) {
 			dev_err(&mcu->serdev->dev,
 				"Invalid Checksum received\n");
-			ret = -EIO;
+			return -EIO;
 		} else {
 			memcpy(reply_data, rx, reply_data_size);
 		}
 	}
 
-	mutex_unlock(&mcu->bus_lock);
-	return ret;
+	return 0;
 }
 EXPORT_SYMBOL_GPL(qnap_mcu_exec);
 
-- 
2.47.2

[PATCH 4/4] mfd: qnap-mcu: improve structure in qnap_mcu_exec

[Heiko Stuebner]

Now with guard(mutex) in place, we can make the function's structure
a bit easier to read, by removing the nested if-else-clauses.

Signed-off-by: Heiko Stuebner <heiko@sntech.de>
---
 drivers/mfd/qnap-mcu.c | 19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)

diff --git a/drivers/mfd/qnap-mcu.c b/drivers/mfd/qnap-mcu.c
index 7bc44c64b7fd..2be429a50611 100644
--- a/drivers/mfd/qnap-mcu.c
+++ b/drivers/mfd/qnap-mcu.c
@@ -150,6 +150,7 @@ int qnap_mcu_exec(struct qnap_mcu *mcu,
 	size_t length = reply_data_size + QNAP_MCU_CHECKSUM_SIZE;
 	struct qnap_mcu_reply *reply = &mcu->reply;
 	int ret = 0;
+	u8 crc;
 
 	if (length > sizeof(rx)) {
 		dev_err(&mcu->serdev->dev, "expected data too big for receive buffer");
@@ -172,18 +173,16 @@ int qnap_mcu_exec(struct qnap_mcu *mcu,
 	if (!wait_for_completion_timeout(&reply->done, msecs_to_jiffies(QNAP_MCU_TIMEOUT_MS))) {
 		dev_err(&mcu->serdev->dev, "Command timeout\n");
 		return -ETIMEDOUT;
-	} else {
-		u8 crc = qnap_mcu_csum(rx, reply_data_size);
-
-		if (crc != rx[reply_data_size]) {
-			dev_err(&mcu->serdev->dev,
-				"Invalid Checksum received\n");
-			return -EIO;
-		} else {
-			memcpy(reply_data, rx, reply_data_size);
-		}
 	}
 
+	crc = qnap_mcu_csum(rx, reply_data_size);
+	if (crc != rx[reply_data_size]) {
+		dev_err(&mcu->serdev->dev, "Invalid Checksum received\n");
+		return -EIO;
+	}
+
+	memcpy(reply_data, rx, reply_data_size);
+
 	return 0;
 }
 EXPORT_SYMBOL_GPL(qnap_mcu_exec);
-- 
2.47.2

Re: [PATCH 0/4] mfd: qnap-mcu: Some fixes / improvements

[Lee Jones]

On Mon, 04 Aug 2025 15:07:22 +0200, Heiko Stuebner wrote:
> While digging through the mcu functions, I came across some deficits I
> introduced with the initial driver submission, so trying to make that
> a bit nicer with this series.
> 
> 
> I struggled a bit with the ordering of qnap_mcu_write error-check
> and conversion to guard(mutex). Converting to guard before the
> error check would need dropping the ret variable, just to re-add it
> one patch later - to not cause unused variable warning.
> 
> [...]

Applied, thanks!

[1/4] mfd: qnap-mcu: include linux/types.h in qnap-mcu.h shared header
      commit: f7ef7c03d8599a0d86b2a05929da73358cd56dcf
[2/4] mfd: qnap-mcu: handle errors returned from qnap_mcu_write
      commit: 3d02c538ec5337b66750d83ce6f861aef263fbff
[3/4] mfd: qnap-mcu: convert to guard(mutex) in qnap_mcu_exec
      commit: 5fd101541c6d0f0ad3b14d86dfcf9347c3f3bffd
[4/4] mfd: qnap-mcu: improve structure in qnap_mcu_exec
      commit: 6cdb0fbe090be966432db041d5650907a4dceac4

--
Lee Jones [李琼斯]

Re: [PATCH 0/4] mfd: qnap-mcu: Some fixes / improvements

[Lee Jones]

On Tue, 02 Sep 2025, Lee Jones wrote:

> On Mon, 04 Aug 2025 15:07:22 +0200, Heiko Stuebner wrote:
> > While digging through the mcu functions, I came across some deficits I
> > introduced with the initial driver submission, so trying to make that
> > a bit nicer with this series.
> > 
> > 
> > I struggled a bit with the ordering of qnap_mcu_write error-check
> > and conversion to guard(mutex). Converting to guard before the
> > error check would need dropping the ret variable, just to re-add it
> > one patch later - to not cause unused variable warning.
> > 
> > [...]
> 
> Applied, thanks!
> 
> [1/4] mfd: qnap-mcu: include linux/types.h in qnap-mcu.h shared header
>       commit: f7ef7c03d8599a0d86b2a05929da73358cd56dcf
> [2/4] mfd: qnap-mcu: handle errors returned from qnap_mcu_write
>       commit: 3d02c538ec5337b66750d83ce6f861aef263fbff
> [3/4] mfd: qnap-mcu: convert to guard(mutex) in qnap_mcu_exec
>       commit: 5fd101541c6d0f0ad3b14d86dfcf9347c3f3bffd
> [4/4] mfd: qnap-mcu: improve structure in qnap_mcu_exec
>       commit: 6cdb0fbe090be966432db041d5650907a4dceac4

I fixed some issues in the commit messages for you pertaining to
spelling mistakes and expected format.  Please not that in future, MFD
and LED (the submissions I've seen from you recently) take a capitalised
word as the start of the subject description.

-- 
Lee Jones [李琼斯]