* [RFC v2 3/5] Bluetooth: Add buffer count to tx scheduler parameters
@ 2011-08-18 0:37 Peter Hurley
2011-08-18 7:57 ` Luiz Augusto von Dentz
0 siblings, 1 reply; 2+ messages in thread
From: Peter Hurley @ 2011-08-18 0:37 UTC (permalink / raw)
To: linux-bluetooth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^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [RFC v2 3/5] Bluetooth: Add buffer count to tx scheduler parameters
2011-08-18 0:37 [RFC v2 3/5] Bluetooth: Add buffer count to tx scheduler parameters Peter Hurley
@ 2011-08-18 7:57 ` Luiz Augusto von Dentz
0 siblings, 0 replies; 2+ messages in thread
From: Luiz Augusto von Dentz @ 2011-08-18 7:57 UTC (permalink / raw)
To: Peter Hurley; +Cc: linux-bluetooth
Hi Peter,
On Thu, Aug 18, 2011 at 3:37 AM, Peter Hurley <peter@hurleysoftware.com> wrote:
> Utilize already known tx buffer count values as parameters to
> the tx scheduler, rather than determining the values on each
> iteration of the tx scheduler.
>
> This is preparatory for merging transmission types (such as SCO
> and ESCO) and also for handling tx buffer counters atomically.
>
> Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
> ---
> net/bluetooth/hci_core.c | 51 ++++++++++++++++-----------------------------
> 1 files changed, 18 insertions(+), 33 deletions(-)
>
> diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c
> index 0defa83..b2a1b9a 100644
> --- a/net/bluetooth/hci_core.c
> +++ b/net/bluetooth/hci_core.c
> @@ -1798,7 +1798,8 @@ EXPORT_SYMBOL(hci_send_sco);
> /* ---- HCI TX task (outgoing data) ---- */
>
> /* HCI Connection scheduler */
> -static inline struct hci_conn *hci_low_sent(struct hci_dev *hdev, __u8 type, int *quote)
> +static inline struct hci_conn *hci_low_sent(struct hci_dev *hdev, __u8 type,
> + int cnt, int *quote)
> {
> struct hci_conn_hash *h = &hdev->conn_hash;
> struct hci_conn *conn = NULL;
> @@ -1825,28 +1826,8 @@ static inline struct hci_conn *hci_low_sent(struct hci_dev *hdev, __u8 type, int
> }
> }
>
> - if (conn) {
> - int cnt;
> -
> - switch (conn->type) {
> - case ACL_LINK:
> - cnt = hdev->acl_cnt;
> - break;
> - case SCO_LINK:
> - case ESCO_LINK:
> - cnt = hdev->sco_cnt;
> - break;
> - case LE_LINK:
> - cnt = hdev->le_mtu ? hdev->le_cnt : hdev->acl_cnt;
> - break;
> - default:
> - cnt = 0;
> - BT_ERR("Unknown link type");
> - }
> -
> + if (conn)
> *quote = (cnt + (num - 1)) / num;
> - } else
> - *quote = 0;
>
> BT_DBG("conn %p quote %d", conn, *quote);
> return conn;
> @@ -1875,18 +1856,19 @@ static inline void hci_sched_acl(struct hci_dev *hdev)
> {
> struct hci_conn *conn;
> struct sk_buff *skb;
> - int quote;
> + int quote = 0;
> + int cnt = hdev->acl_cnt;
>
> BT_DBG("%s", hdev->name);
>
> if (!test_bit(HCI_RAW, &hdev->flags)) {
> /* ACL tx timeout must be longer than maximum
> * link supervision timeout (40.9 seconds) */
> - if (!hdev->acl_cnt && time_after(jiffies, hdev->acl_last_tx + HZ * 45))
> + if (!cnt && time_after(jiffies, hdev->acl_last_tx + HZ * 45))
> hci_link_tx_to(hdev, ACL_LINK);
> }
>
> - while (hdev->acl_cnt && (conn = hci_low_sent(hdev, ACL_LINK, "e))) {
> + while (cnt && (conn = hci_low_sent(hdev, ACL_LINK, cnt, "e))) {
> while (quote-- && (skb = skb_dequeue(&conn->data_q))) {
> BT_DBG("skb %p len %d", skb, skb->len);
>
> @@ -1895,10 +1877,11 @@ static inline void hci_sched_acl(struct hci_dev *hdev)
> hci_send_frame(skb);
> hdev->acl_last_tx = jiffies;
>
> - hdev->acl_cnt--;
> + cnt--;
> conn->sent++;
> }
> }
> + hdev->acl_cnt = cnt;
> }
>
> /* Schedule SCO */
> @@ -1906,11 +1889,12 @@ static inline void hci_sched_sco(struct hci_dev *hdev)
> {
> struct hci_conn *conn;
> struct sk_buff *skb;
> - int quote;
> + int quote = 0;
> + int cnt = hdev->sco_cnt;
>
> BT_DBG("%s", hdev->name);
>
> - while (hdev->sco_cnt && (conn = hci_low_sent(hdev, SCO_LINK, "e))) {
> + while (cnt && (conn = hci_low_sent(hdev, SCO_LINK, cnt, "e))) {
> while (quote-- && (skb = skb_dequeue(&conn->data_q))) {
> BT_DBG("skb %p len %d", skb, skb->len);
> hci_send_frame(skb);
> @@ -1926,11 +1910,12 @@ static inline void hci_sched_esco(struct hci_dev *hdev)
> {
> struct hci_conn *conn;
> struct sk_buff *skb;
> - int quote;
> + int quote = 0;
> + int cnt = hdev->sco_cnt;
>
> BT_DBG("%s", hdev->name);
>
> - while (hdev->sco_cnt && (conn = hci_low_sent(hdev, ESCO_LINK, "e))) {
> + while (cnt && (conn = hci_low_sent(hdev, ESCO_LINK, cnt, "e))) {
> while (quote-- && (skb = skb_dequeue(&conn->data_q))) {
> BT_DBG("skb %p len %d", skb, skb->len);
> hci_send_frame(skb);
> @@ -1946,7 +1931,8 @@ static inline void hci_sched_le(struct hci_dev *hdev)
> {
> struct hci_conn *conn;
> struct sk_buff *skb;
> - int quote, cnt;
> + int quote = 0;
> + int cnt = hdev->le_pkts ? hdev->le_cnt : hdev->acl_cnt;
>
> BT_DBG("%s", hdev->name);
>
> @@ -1958,8 +1944,7 @@ static inline void hci_sched_le(struct hci_dev *hdev)
> hci_link_tx_to(hdev, LE_LINK);
> }
>
> - cnt = hdev->le_pkts ? hdev->le_cnt : hdev->acl_cnt;
> - while (cnt && (conn = hci_low_sent(hdev, LE_LINK, "e))) {
> + while (cnt && (conn = hci_low_sent(hdev, LE_LINK, cnt, "e))) {
> while (quote-- && (skb = skb_dequeue(&conn->data_q))) {
> BT_DBG("skb %p len %d", skb, skb->len);
>
> --
> 1.7.4.1
>
>
Nice work, it probably conflict with one of my patches to check if
there is any connection, but I guess it is easy to rebase, btw it is
probably a good idea to no initialize the variables in the declaration
since with hci_conn_num check we may bail out without using them.
--
Luiz Augusto von Dentz
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2011-08-18 7:57 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2011-08-18 0:37 [RFC v2 3/5] Bluetooth: Add buffer count to tx scheduler parameters Peter Hurley
2011-08-18 7:57 ` Luiz Augusto von Dentz
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.