Re: [PATCH] Bluetooth: Fix HCI_RESET command syncronization
From: Szymon Janc <hidden>
Date: 2011-03-17 10:35:20
We can't send new commands before a cmd_complete for the HCI_RESET commnnd
Typo: comm*a*nd
quoted hunk ↗ jump to hunk
shows up. Reported-by: Mikko Vinni <redacted> Reported-by: Justin P. Mattock <redacted> Reported-by: Ed Tomlinson <redacted> Signed-off-by: Gustavo F. Padovan <redacted> --- include/net/bluetooth/hci.h | 2 ++ net/bluetooth/hci_core.c | 3 +++ net/bluetooth/hci_event.c | 4 +++- 3 files changed, 8 insertions(+), 1 deletions(-)diff --git a/include/net/bluetooth/hci.h b/include/net/bluetooth/hci.h index ec6acf2..2c0d309 100644 --- a/include/net/bluetooth/hci.h +++ b/include/net/bluetooth/hci.h@@ -84,6 +84,8 @@ enum { HCI_SERVICE_CACHE, HCI_LINK_KEYS, HCI_DEBUG_KEYS, + + HCI_RESET, }; /* HCI ioctl defines */diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c index b372fb8..32b82e2 100644 --- a/net/bluetooth/hci_core.c +++ b/net/bluetooth/hci_core.c@@ -534,6 +534,8 @@ int hci_dev_open(__u16 dev) set_bit(HCI_INIT, &hdev->flags); hdev->init_last_cmd = 0; + set_bit(HCI_RESET, &hdev->flags);
Does this work with HCI_QUIRK_NO_RESET enabled? Shouldn't this bit be set only right before sending HCI_OP_RESET (in hci_init_req and hci_reset_req) for doing what commit msg says? Now it looks more like a workaround for initialization issue..
quoted hunk ↗ jump to hunk
+ ret = __hci_request(hdev, hci_init_req, 0, msecs_to_jiffies(HCI_INIT_TIMEOUT));@@ -1074,6 +1076,7 @@ static void hci_cmd_timer(unsigned long arg) BT_ERR("%s command tx timeout", hdev->name); atomic_set(&hdev->cmd_cnt, 1); + clear_bit(HCI_RESET, &hdev->flags); tasklet_schedule(&hdev->cmd_task); }diff --git a/net/bluetooth/hci_event.c b/net/bluetooth/hci_event.c index 3fbfa50..cebe7588 100644 --- a/net/bluetooth/hci_event.c +++ b/net/bluetooth/hci_event.c@@ -183,6 +183,8 @@ static void hci_cc_reset(struct hci_dev *hdev, struct sk_buff *skb) BT_DBG("%s status 0x%x", hdev->name, status); + clear_bit(HCI_RESET, &hdev->flags); + hci_req_complete(hdev, HCI_OP_RESET, status); }@@ -1847,7 +1849,7 @@ static inline void hci_cmd_status_evt(struct hci_dev *hdev, struct sk_buff *skb) if (ev->opcode != HCI_OP_NOP) del_timer(&hdev->cmd_timer); - if (ev->ncmd) { + if (ev->ncmd && !test_bit(HCI_RESET, &hdev->flags)) { atomic_set(&hdev->cmd_cnt, 1); if (!skb_queue_empty(&hdev->cmd_q)) tasklet_schedule(&hdev->cmd_task);
Maybe use unlikely() here? -- BR Szymon Janc