[PATCH] ASoC: cs-amp-lib: Replace __free(kfree) with normal kfree() cleanup

[PATCH] ASoC: cs-amp-lib: Replace __free(kfree) with normal kfree() cleanup

[Richard Fitzgerald]

Replace the use of __free(kfree) in _cs_amp_set_efi_calibration_data()
with normal kfree() at the end of the function.

Krzysztof Kozlowski pointed out that __free() can be dangerous.
It can introduce new cleanup bugs. These are more subtle and difficult to
spot than a missing goto in traditional cleanup, because they are triggered
by writing regular idiomatic C code instead of using C++ conventions. As
it's regular C style it's more likely to be missed because the code is as
would be expected for C. The traditional goto also more obviously flags
to anyone changing the code in the future that they must be careful about
the cleanup.

Arguably the traditional approach is more readable, and it avoids the
manipulation of __free() pointers when the buffer is reallocated for
resizing.

Signed-off-by: Richard Fitzgerald <rf@opensource.cirrus.com>
---
 sound/soc/codecs/cs-amp-lib.c | 27 ++++++++++++++++-----------
 1 file changed, 16 insertions(+), 11 deletions(-)

diff --git a/sound/soc/codecs/cs-amp-lib.c b/sound/soc/codecs/cs-amp-lib.c
index d8f8b0259cd1..b4d183e7501d 100644
--- a/sound/soc/codecs/cs-amp-lib.c
+++ b/sound/soc/codecs/cs-amp-lib.c
@@ -452,7 +452,7 @@ static int _cs_amp_set_efi_calibration_data(struct device *dev, int amp_index, i
 {
 	u64 cal_target = cs_amp_cal_target_u64(in_data);
 	unsigned long num_entries;
-	struct cirrus_amp_efi_data *data __free(kfree) = NULL;
+	struct cirrus_amp_efi_data *data;
 	efi_char16_t *name = CIRRUS_LOGIC_CALIBRATION_EFI_NAME;
 	efi_guid_t *guid = &CIRRUS_LOGIC_CALIBRATION_EFI_GUID;
 	u32 attr = CS_AMP_CAL_DEFAULT_EFI_ATTR;
@@ -515,28 +515,33 @@ static int _cs_amp_set_efi_calibration_data(struct device *dev, int amp_index, i
 
 	num_entries = max(num_amps, amp_index + 1);
 	if (!data || (data->count < num_entries)) {
-		struct cirrus_amp_efi_data *old_data __free(kfree) = no_free_ptr(data);
+		struct cirrus_amp_efi_data *new_data;
 		unsigned int new_data_size = struct_size(data, data, num_entries);
 
-		data = kzalloc(new_data_size, GFP_KERNEL);
-		if (!data)
-			return -ENOMEM;
+		new_data = kzalloc(new_data_size, GFP_KERNEL);
+		if (!new_data) {
+			ret = -ENOMEM;
+			goto err;
+		}
 
-		if (old_data)
-			memcpy(data, old_data, struct_size(old_data, data, old_data->count));
+		if (data) {
+			memcpy(new_data, data, struct_size(data, data, data->count));
+			kfree(data);
+		}
 
+		data = new_data;
 		data->count = num_entries;
 		data->size = new_data_size;
 	}
 
 	data->data[amp_index] = *in_data;
 	ret = cs_amp_set_cal_efi_buffer(dev, name, guid, attr, data);
-	if (ret) {
+	if (ret)
 		dev_err(dev, "Failed writing calibration to EFI: %d\n", ret);
-		return ret;
-	}
+err:
+	kfree(data);
 
-	return 0;
+	return ret;
 }
 
 /**
-- 
2.47.3

Re: [PATCH] ASoC: cs-amp-lib: Replace __free(kfree) with normal kfree() cleanup

[Mark Brown]

On Mon, 01 Dec 2025 12:39:06 +0000, Richard Fitzgerald wrote:
> Replace the use of __free(kfree) in _cs_amp_set_efi_calibration_data()
> with normal kfree() at the end of the function.
> 
> Krzysztof Kozlowski pointed out that __free() can be dangerous.
> It can introduce new cleanup bugs. These are more subtle and difficult to
> spot than a missing goto in traditional cleanup, because they are triggered
> by writing regular idiomatic C code instead of using C++ conventions. As
> it's regular C style it's more likely to be missed because the code is as
> would be expected for C. The traditional goto also more obviously flags
> to anyone changing the code in the future that they must be careful about
> the cleanup.
> 
> [...]

Applied to

   https://git.kernel.org/pub/scm/linux/kernel/git/broonie/sound.git for-next

Thanks!

[1/1] ASoC: cs-amp-lib: Replace __free(kfree) with normal kfree() cleanup
      commit: 9a123f222e1889d020d873aa6e0799098d22cdb1

All being well this means that it will be integrated into the linux-next
tree (usually sometime in the next 24 hours) and sent to Linus during
the next merge window (or sooner if it is a bug fix), however if
problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing
and review of the tree, please engage with people reporting problems and
send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they
should be sent as incremental updates against current git, existing
patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying
to this mail.

Thanks,
Mark