Return-Path: <cyrus@holtmann.org>
Date: Wed, 7 Mar 2012 11:19:31 +0200
From: Andrei Emeltchenko <Andrei.Emeltchenko.news@gmail.com>
To: Szymon Janc <szymon.janc@tieto.com>
Cc: "linux-bluetooth@vger.kernel.org" <linux-bluetooth@vger.kernel.org>
Subject: Re: [RFC 1/2] Bluetooth: General HCI callback implementation
Message-ID: <20120307091929.GC3647@aemeltch-MOBL1>
References: <1331039789-31519-1-git-send-email-Andrei.Emeltchenko.news@gmail.com>
 <201203061602.17443.szymon.janc@tieto.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
In-Reply-To: <201203061602.17443.szymon.janc@tieto.com>
Sender: linux-bluetooth-owner@vger.kernel.org
List-ID: <linux-bluetooth.vger.kernel.org>

Hi Janc,

On Tue, Mar 06, 2012 at 04:02:17PM +0100, Szymon Janc wrote:
> > +
> > +struct cb_cmd {
> 
> This name seems a bit too more generic to me, maybe prefix it with hci_?

OK.

> 
> > +	struct list_head list;
> > +	u16 opcode;
> > +	u8 status;
> > +	void *opt;
> > +	void (*cb)(struct hci_dev *hdev, struct cb_cmd *cmd);
> > +	void (*destructor)(struct cb_cmd *cmd);
> > +};
> > +
> >  #define NUM_REASSEMBLY 4
> >  struct hci_dev {
> >  	struct list_head list;
> > @@ -237,6 +248,7 @@ struct hci_dev {
> >  	__u16			init_last_cmd;
> >  
> >  	struct list_head	mgmt_pending;
> > +	struct list_head	cb_list;
> >  
> >  	struct discovery_state	discovery;
> >  	struct hci_conn_hash	conn_hash;
> > @@ -1086,4 +1098,10 @@ int hci_cancel_inquiry(struct hci_dev *hdev);
> >  int hci_le_scan(struct hci_dev *hdev, u8 type, u16 interval, u16 window,
> >  								int timeout);
> >  
> > +struct cb_cmd *hci_find_cb(struct hci_dev *hdev, __u16 opcode);
> > +int hci_cmd_cb(struct hci_dev *hdev, __u16 opcode, __u32 plen, void *param,
> > +	void (*cb)(struct hci_dev *hdev, struct cb_cmd *cmd), void *opt,
> > +				void (*destructor)(struct cb_cmd *cmd));
> > +void hci_remove_cb(struct cb_cmd *cmd);
> > +
> >  #endif /* __HCI_CORE_H */
> > diff --git a/net/bluetooth/hci_core.c b/net/bluetooth/hci_core.c
> > index 5b1ddab..cdc0220 100644
> > --- a/net/bluetooth/hci_core.c
> > +++ b/net/bluetooth/hci_core.c
> > @@ -1826,6 +1826,8 @@ int hci_register_dev(struct hci_dev *hdev)
> >  
> >  	INIT_LIST_HEAD(&hdev->mgmt_pending);
> >  
> > +	INIT_LIST_HEAD(&hdev->cb_list);
> > +
> >  	INIT_LIST_HEAD(&hdev->blacklist);
> >  
> >  	INIT_LIST_HEAD(&hdev->uuids);
> > @@ -2239,6 +2241,61 @@ int hci_send_cmd(struct hci_dev *hdev, __u16 opcode, __u32 plen, void *param)
> >  	return 0;
> >  }
> >  
> 
> hci_add_cb could return error on failure and that could be used in hci_cmd_cb().
> 
> > +void hci_add_cb(struct hci_dev *hdev, __u16 opcode,
> > +			void (*cb)(struct hci_dev *hdev, struct cb_cmd *cmd),
> > +			void *opt, void (*destructor)(struct cb_cmd *cmd))
> > +{
> > +	struct cb_cmd *cmd;
> > +
> > +	cmd = kmalloc(sizeof(*cmd), GFP_ATOMIC);
> 
> Why atomic? Maybe allow to pass custom gfp mask?

I think I will use GFP_KERNEL as Ulisses suggested. I have used the code
when we had tasklets.

> 
> > +	if (!cmd)
> > +		return;
> > +
> > +	cmd->cb = cb;
> > +	cmd->opcode = opcode;
> > +	cmd->opt = opt;
> > +	cmd->status = 0;
> > +
> > +	if (destructor)
> > +		cmd->destructor = destructor;
> 
> That could leave cmd->destructor uninitialized (cmd is allocated with kmalloc).
> I would assign destructor unconditionally.

This makes sense.

> > +
> > +	list_add(&cmd->list, &hdev->cb_list);
> > +}
> > +
> > +struct cb_cmd *hci_find_cb(struct hci_dev *hdev, __u16 opcode)
> > +{
> > +	struct cb_cmd *cmd;
> > +
> > +	list_for_each_entry(cmd, &hdev->cb_list, list)
> > +		if (cmd->opcode == opcode)
> > +			return cmd;
> > +
> > +	return NULL;
> > +}
> > +
> > +void hci_remove_cb(struct cb_cmd *cmd)
> > +{
> > +	list_del(&cmd->list);
> > +
> > +	if (cmd->destructor) {
> > +		cmd->destructor(cmd);
> > +	} else {
> > +		kfree(cmd->opt);
> > +		kfree(cmd);
> > +	}
> > +}
> > +
> > +/* Send HCI command with callback */
> > +int hci_cmd_cb(struct hci_dev *hdev, __u16 opcode, __u32 plen, void *param,
> > +			void (*cb)(struct hci_dev *hdev, struct cb_cmd *cmd),
> > +			void *opt, void (*destructor)(struct cb_cmd *cmd))
> > +{
> > +	if (cb)
> > +		hci_add_cb(hdev, opcode, cb, opt, destructor);
> 
> Is this if really needed? If one would like to send cmd without cb he can just
> call hci_send_cmd directly. And I would also check if hci_add_cb failed before
> sending command.

OK I will change the code this way.

Best regards 
Andrei Emeltchenko