[PATCH v2] Bluetooth: btintel_pcie: Avoid redundant buffer allocation

[PATCH v2] Bluetooth: btintel_pcie: Avoid redundant buffer allocation

[Kiran K]

Reuse the skb buffer provided by the PCIe driver to pass it onto the
stack, instead of copying it to a new skb.

Fixes: c2b636b3f788 ("Bluetooth: btintel_pcie: Add support for PCIe transport")
Signed-off-by: Kiran K <kiran.k@intel.com>
---
 drivers/bluetooth/btintel_pcie.c | 33 ++++++++++++--------------------
 1 file changed, 12 insertions(+), 21 deletions(-)

diff --git a/drivers/bluetooth/btintel_pcie.c b/drivers/bluetooth/btintel_pcie.c
index e0b50513403f..b73928e38d34 100644
--- a/drivers/bluetooth/btintel_pcie.c
+++ b/drivers/bluetooth/btintel_pcie.c
@@ -947,8 +947,10 @@ static int btintel_pcie_recv_event(struct hci_dev *hdev, struct sk_buff *skb)
 		/* This is a debug event that comes from IML and OP image when it
 		 * starts execution. There is no need pass this event to stack.
 		 */
-		if (skb->data[2] == 0x97)
+		if (skb->data[2] == 0x97) {
+			hci_recv_diag(hdev, skb);
 			return 0;
+		}
 	}
 
 	return hci_recv_frame(hdev, skb);
@@ -964,7 +966,6 @@ static int btintel_pcie_recv_frame(struct btintel_pcie_data *data,
 	u8 pkt_type;
 	u16 plen;
 	u32 pcie_pkt_type;
-	struct sk_buff *new_skb;
 	void *pdata;
 	struct hci_dev *hdev = data->hdev;
 
@@ -1041,24 +1042,20 @@ static int btintel_pcie_recv_frame(struct btintel_pcie_data *data,
 
 	bt_dev_dbg(hdev, "pkt_type: 0x%2.2x len: %u", pkt_type, plen);
 
-	new_skb = bt_skb_alloc(plen, GFP_ATOMIC);
-	if (!new_skb) {
-		bt_dev_err(hdev, "Failed to allocate memory for skb of len: %u",
-			   skb->len);
-		ret = -ENOMEM;
-		goto exit_error;
-	}
-
-	hci_skb_pkt_type(new_skb) = pkt_type;
-	skb_put_data(new_skb, skb->data, plen);
+	hci_skb_pkt_type(skb) = pkt_type;
 	hdev->stat.byte_rx += plen;
+	skb_trim(skb, plen);
 
 	if (pcie_pkt_type == BTINTEL_PCIE_HCI_EVT_PKT)
-		ret = btintel_pcie_recv_event(hdev, new_skb);
+		ret = btintel_pcie_recv_event(hdev, skb);
 	else
-		ret = hci_recv_frame(hdev, new_skb);
+		ret = hci_recv_frame(hdev, skb);
+	skb = NULL; /* skb is freed in the callee  */
 
 exit_error:
+	if (skb)
+		kfree_skb(skb);
+
 	if (ret)
 		hdev->stat.err_rx++;
 
@@ -1192,8 +1189,6 @@ static void btintel_pcie_rx_work(struct work_struct *work)
 	struct btintel_pcie_data *data = container_of(work,
 					struct btintel_pcie_data, rx_work);
 	struct sk_buff *skb;
-	int err;
-	struct hci_dev *hdev = data->hdev;
 
 	if (test_bit(BTINTEL_PCIE_HWEXP_INPROGRESS, &data->flags)) {
 		/* Unlike usb products, controller will not send hardware
@@ -1214,11 +1209,7 @@ static void btintel_pcie_rx_work(struct work_struct *work)
 
 	/* Process the sk_buf in queue and send to the HCI layer */
 	while ((skb = skb_dequeue(&data->rx_skb_q))) {
-		err = btintel_pcie_recv_frame(data, skb);
-		if (err)
-			bt_dev_err(hdev, "Failed to send received frame: %d",
-				   err);
-		kfree_skb(skb);
+		btintel_pcie_recv_frame(data, skb);
 	}
 }
 
-- 
2.43.0

RE: [v2] Bluetooth: btintel_pcie: Avoid redundant buffer allocation

[bluez.test.bot]

[-- Attachment #1: Type: text/plain, Size: 1949 bytes --]

This is automated email and please do not reply to this email!

Dear submitter,

Thank you for submitting the patches to the linux bluetooth mailing list.
This is a CI test results with your patch series:
PW Link:https://patchwork.kernel.org/project/bluetooth/list/?series=954259

---Test result---

Test Summary:
CheckPatch                    PENDING   0.31 seconds
GitLint                       PENDING   0.23 seconds
SubjectPrefix                 PASS      0.13 seconds
BuildKernel                   PASS      24.66 seconds
CheckAllWarning               PASS      26.92 seconds
CheckSparse                   PASS      37.45 seconds
BuildKernel32                 PASS      26.14 seconds
TestRunnerSetup               PASS      457.93 seconds
TestRunner_l2cap-tester       PASS      25.54 seconds
TestRunner_iso-tester         PASS      34.66 seconds
TestRunner_bnep-tester        PASS      6.88 seconds
TestRunner_mgmt-tester        PASS      136.85 seconds
TestRunner_rfcomm-tester      PASS      16.27 seconds
TestRunner_sco-tester         PASS      17.95 seconds
TestRunner_ioctl-tester       PASS      17.29 seconds
TestRunner_mesh-tester        FAIL      6.28 seconds
TestRunner_smp-tester         PASS      10.42 seconds
TestRunner_userchan-tester    PASS      9.44 seconds
IncrementalBuild              PENDING   0.72 seconds

Details
##############################
Test: CheckPatch - PENDING
Desc: Run checkpatch.pl script
Output:

##############################
Test: GitLint - PENDING
Desc: Run gitlint
Output:

##############################
Test: TestRunner_mesh-tester - FAIL
Desc: Run mesh-tester with test-runner
Output:
Total: 10, Passed: 9 (90.0%), Failed: 1, Not Run: 0

Failed Test Cases
Mesh - Send cancel - 2                               Failed       0.119 seconds
##############################
Test: IncrementalBuild - PENDING
Desc: Incremental build with the patches in the series
Output:



---
Regards,
Linux Bluetooth

Re: [PATCH v2] Bluetooth: btintel_pcie: Avoid redundant buffer allocation

[Paul Menzel]

Dear Kiran,


Thank you for your patch.


Am 17.04.25 um 05:48 schrieb Kiran K:
> Reuse the skb buffer provided by the PCIe driver to pass it onto the
> stack, instead of copying it to a new skb.

Does `scripts/bloat-o-meter` verify, that the optimization worked?

> Fixes: c2b636b3f788 ("Bluetooth: btintel_pcie: Add support for PCIe transport")
> Signed-off-by: Kiran K <kiran.k@intel.com>
> ---
>   drivers/bluetooth/btintel_pcie.c | 33 ++++++++++++--------------------
>   1 file changed, 12 insertions(+), 21 deletions(-)
> 
> diff --git a/drivers/bluetooth/btintel_pcie.c b/drivers/bluetooth/btintel_pcie.c
> index e0b50513403f..b73928e38d34 100644
> --- a/drivers/bluetooth/btintel_pcie.c
> +++ b/drivers/bluetooth/btintel_pcie.c
> @@ -947,8 +947,10 @@ static int btintel_pcie_recv_event(struct hci_dev *hdev, struct sk_buff *skb)
>   		/* This is a debug event that comes from IML and OP image when it
>   		 * starts execution. There is no need pass this event to stack.
>   		 */
> -		if (skb->data[2] == 0x97)
> +		if (skb->data[2] == 0x97) {
> +			hci_recv_diag(hdev, skb);
>   			return 0;
> +		}
>   	}
>   
>   	return hci_recv_frame(hdev, skb);
> @@ -964,7 +966,6 @@ static int btintel_pcie_recv_frame(struct btintel_pcie_data *data,
>   	u8 pkt_type;
>   	u16 plen;
>   	u32 pcie_pkt_type;
> -	struct sk_buff *new_skb;
>   	void *pdata;
>   	struct hci_dev *hdev = data->hdev;
>   
> @@ -1041,24 +1042,20 @@ static int btintel_pcie_recv_frame(struct btintel_pcie_data *data,
>   
>   	bt_dev_dbg(hdev, "pkt_type: 0x%2.2x len: %u", pkt_type, plen);
>   
> -	new_skb = bt_skb_alloc(plen, GFP_ATOMIC);
> -	if (!new_skb) {
> -		bt_dev_err(hdev, "Failed to allocate memory for skb of len: %u",
> -			   skb->len);
> -		ret = -ENOMEM;
> -		goto exit_error;
> -	}
> -
> -	hci_skb_pkt_type(new_skb) = pkt_type;
> -	skb_put_data(new_skb, skb->data, plen);
> +	hci_skb_pkt_type(skb) = pkt_type;
>   	hdev->stat.byte_rx += plen;
> +	skb_trim(skb, plen);
>   
>   	if (pcie_pkt_type == BTINTEL_PCIE_HCI_EVT_PKT)
> -		ret = btintel_pcie_recv_event(hdev, new_skb);
> +		ret = btintel_pcie_recv_event(hdev, skb);
>   	else
> -		ret = hci_recv_frame(hdev, new_skb);
> +		ret = hci_recv_frame(hdev, skb);
> +	skb = NULL; /* skb is freed in the callee  */
>   
>   exit_error:
> +	if (skb)
> +		kfree_skb(skb);
> +
>   	if (ret)
>   		hdev->stat.err_rx++;
>   
> @@ -1192,8 +1189,6 @@ static void btintel_pcie_rx_work(struct work_struct *work)
>   	struct btintel_pcie_data *data = container_of(work,
>   					struct btintel_pcie_data, rx_work);
>   	struct sk_buff *skb;
> -	int err;
> -	struct hci_dev *hdev = data->hdev;
>   
>   	if (test_bit(BTINTEL_PCIE_HWEXP_INPROGRESS, &data->flags)) {
>   		/* Unlike usb products, controller will not send hardware
> @@ -1214,11 +1209,7 @@ static void btintel_pcie_rx_work(struct work_struct *work)
>   
>   	/* Process the sk_buf in queue and send to the HCI layer */
>   	while ((skb = skb_dequeue(&data->rx_skb_q))) {
> -		err = btintel_pcie_recv_frame(data, skb);
> -		if (err)
> -			bt_dev_err(hdev, "Failed to send received frame: %d",
> -				   err);
> -		kfree_skb(skb);
> +		btintel_pcie_recv_frame(data, skb);
>   	}
>   }

Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>


Kind regards,

Paul

Re: [PATCH v2] Bluetooth: btintel_pcie: Avoid redundant buffer allocation

[patchwork-bot+bluetooth]

Hello:

This patch was applied to bluetooth/bluetooth-next.git (master)
by Luiz Augusto von Dentz <luiz.von.dentz@intel.com>:

On Thu, 17 Apr 2025 09:18:42 +0530 you wrote:
> Reuse the skb buffer provided by the PCIe driver to pass it onto the
> stack, instead of copying it to a new skb.
> 
> Fixes: c2b636b3f788 ("Bluetooth: btintel_pcie: Add support for PCIe transport")
> Signed-off-by: Kiran K <kiran.k@intel.com>
> ---
>  drivers/bluetooth/btintel_pcie.c | 33 ++++++++++++--------------------
>  1 file changed, 12 insertions(+), 21 deletions(-)

Here is the summary with links:
  - [v2] Bluetooth: btintel_pcie: Avoid redundant buffer allocation
    https://git.kernel.org/bluetooth/bluetooth-next/c/16b4f97defef

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html