[Bluez-devel] ttyS0 problem

[Bluez-devel] ttyS0 problem

[david]

Hi marcel,

=D4=DA 2005-10-10=D2=BB=B5=C4 09:13 +0800=A3=ACdavid=D0=B4=B5=C0=A3=BA
> Hi David,
> =20
>  > I want to write some commands to the uart(ttyS0), but it fails.
>  > The steps are as follows:
>  > =20
>  > 1. hciattach ttyS0 texas
>  > 2. Open ttyS0 device again and get a fd, init ttyS0 with speed,
>  > flow_ctl and so on,then write a command(0x30) to the fd but return
>  > 0; That is nothing is written to ttyS0.
>  > =20
>  > If I write the same command  in the  function texas(int fd, struct
>  > uart_t *u, struct termios *ti), it succeeds.
>  > From bluez-utiles I see the following code which maybe
>  > the reason why I can't write the command to ttyS0 after running
>  > hciattach ttyS0 texas.
>  > =20
>  > /* Set TTY to N_HCI line discipline */
>  >  i =3D N_HCI;
>  >  if (ioctl(fd, TIOCSETD, &i) < 0) {
>  >   perror("Can't set line discipline");
>  >   return -1;
>  >  }
>  > =20
>  > It seems the codes is very important ,because if I delete these
>  > codes, the following codes can't run any more.
>  > =20
>  > Is there any method that I can write the command to the ttyS0
>  > successfully after step 1?
> =20
>  after step 1, you  must use the HCI raw socket to send commands. What
>  commands do you wanna send?
> =20
>  Regards
> =20
>  Marcel
=20

> What commands do you wanna send?
The command is HCI_GO_TO_SLEEP_IND(0x30) which is only one byte. It is a
vendor specific command used to ask the BT device to go to sleep. If BT
device receives the command, maybe it would response
HCILL_GO_TO_SLEEP_ACK(0x31) command to the host.

> after step 1, you  must use the HCI raw socket to send commands

Is this means we cannot write command to ttyS0 directly through write()
function?


Regards

David


-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

Re: [Bluez-devel] ttyS0 problem

[Marcel Holtmann]

Hi David,

> > What commands do you wanna send?
> The command is HCI_GO_TO_SLEEP_IND(0x30) which is only one byte. It is a
> vendor specific command used to ask the BT device to go to sleep. If BT
> device receives the command, maybe it would response
> HCILL_GO_TO_SLEEP_ACK(0x31) command to the host.

do you have the specification of these vendor commands? Is this command
according to the HCI specification or is it an extension of the H:4
transport protocol?

> > after step 1, you  must use the HCI raw socket to send commands
> 
> Is this means we cannot write command to ttyS0 directly through write()
> function?

After calling hciattach the kernel takes over the serial port via the
line discipline and the reads and writes are useless.

Regards

Marcel




-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

[Bluez-devel] ttyS0 problem

[david]

Hi marcel,

=D4=DA 2005-10-11=B6=FE=B5=C4 14:03 +0800=A3=ACdavid=D0=B4=B5=C0=A3=BA
>  Hi David,

>  > I want to write some commands to the uart(ttyS0), but it fails.
 >  > The steps are as follows:
 >  > =20
 >  > 1. hciattach ttyS0 texas
 >  > 2. Open ttyS0 device again and get a fd, init ttyS0 with speed,
 >  > flow_ctl and so on,then write a command(0x30) to the fd but return
 >  > 0; That is nothing is written to ttyS0.
 >  > =20
> marcel write,=20
>  > > What commands do you wanna send?
>  > The command is HCI_GO_TO_SLEEP_IND(0x30) which is only one byte. It is=
 a
>  > vendor specific command used to ask the BT device to go to sleep. If B=
T
>  > device receives the command, maybe it would response
>  > HCILL_GO_TO_SLEEP_ACK(0x31) command to the host.
> =20
>  do you have the specification of these vendor commands? Is this command
>  according to the HCI specification or is it an extension of the H:4
>  transport protocol?
> =20
>  > > after step 1, you  must use the HCI raw socket to send commands
>  >=20
>  > Is this means we cannot write command to ttyS0 directly through write(=
)
>  > function?
> =20
>  After calling hciattach the kernel takes over the serial port via the
>  line discipline and the reads and writes are useless.
> =20
>  Regards
> =20
>  Marcel



Now, I can write command(0x30) to ttyS0 driectly after hciattach tty0
texas(assume step 1);
The function is as follows(assume step 2):

static int write_cmd_to_ttyS0()

{
                int fdd;
                int len;
                char cmd[5];
                int temp;
                int dd;
   =20
                fdd =3D open("/dev/ttyS0", O_RDWR | O_NOCTTY);
                if (fdd < 0) {
                        printf("Can't open serial port, fd =3D %d\n",fdd);
                        return -1;
       		  }
                else
                        printf("open serial ok, fd =3D %d\n", fdd);
             =20
		  temp =3D 0;
                if (ioctl(fdd, TIOCSETD, &temp) < 0) {
                        perror("Can't set line discipline");
                        return -1;
                  }

                cmd[0] =3D 0x30;//HCILL_GO_TO_SLEEP_IND

                len =3D write(fdd, cmd, 1);
                printf("write %d byte\n", len);
                len =3D read(fdd, cmd, 1);
                printf("read %d byte cmd[0] =3D %d\n", len, cmd[0]);

                temp =3D N_HCI;
                if (ioctl(fdd, TIOCSETD, &temp) < 0) {
                        perror("Can't set line discipline");
                        return -1;
                }
                if (ioctl(fdd, HCIUARTSETPROTO, 0) < 0) {
                        perror("Can't set device");
                        return -1;
                }
 }            =20

After step 2,  step3 and step 4 run successfully.=20

step 3: hciconfig hci0 up
step 4: hcitool scan

If I change the steps as follows, an error occures in step 4, the error
is "Device is not available". Do you know why? How can I step 4 can run
successfully after step 3?

step 1: hciattach ttyS0 texas
Step 2: hciconfig hci0 up
step 3: write_cmd_to_ttyS0()
step 4: hcitool scan

/*** cmd_scan() ****/
   if (dev_id < 0) {
                dev_id =3D hci_get_route(NULL);
                if (dev_id < 0) {
                        perror("Device is not available");
                        exit(1);
                }
        }
/**********************/




Regards,

David



-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel

[Bluez-devel] ttyS0 problem

[davidgchen]

[-- Attachment #1: Type: text/plain, Size: 940 bytes --]

Hi marcel,

I want to write some commands to the uart(ttyS0), but it fails. The steps are as follows:

1. hciattach ttyS0 texas
2. Open ttyS0 device again and get a fd, init ttyS0 with speed, flow_ctl and so on,then write a command(0x30) to the fd but return 0; That is nothing is written to ttyS0.

If I write the same command  in the  function texas(int fd, struct uart_t *u, struct termios *ti), it succeeds.
From bluez-utiles I see the following code which maybe the reason why I can't write the command to ttyS0 after running hciattach ttyS0 texas.

/* Set TTY to N_HCI line discipline */
 i = N_HCI;
 if (ioctl(fd, TIOCSETD, &i) < 0) {
  perror("Can't set line discipline");
  return -1;
 }

It seems the codes is very important ,because if I delete these codes, the following codes can't run any more.

Is there any method that I can write the command to the ttyS0 successfully after step 1?

regurds,

David

[-- Attachment #2: Type: text/html, Size: 2097 bytes --]

Re: [Bluez-devel] ttyS0 problem

[Marcel Holtmann]

Hi David,

> I want to write some commands to the uart(ttyS0), but it fails.
> The steps are as follows:
>  
> 1. hciattach ttyS0 texas
> 2. Open ttyS0 device again and get a fd, init ttyS0 with speed,
> flow_ctl and so on,then write a command(0x30) to the fd but return
> 0; That is nothing is written to ttyS0.
>  
> If I write the same command  in the  function texas(int fd, struct
> uart_t *u, struct termios *ti), it succeeds.
> From bluez-utiles I see the following code which maybe
> the reason why I can't write the command to ttyS0 after running
> hciattach ttyS0 texas.
>  
> /* Set TTY to N_HCI line discipline */
>  i = N_HCI;
>  if (ioctl(fd, TIOCSETD, &i) < 0) {
>   perror("Can't set line discipline");
>   return -1;
>  }
>  
> It seems the codes is very important ,because if I delete these
> codes, the following codes can't run any more.
>  
> Is there any method that I can write the command to the ttyS0
> successfully after step 1?

after step 1, you  must use the HCI raw socket to send commands. What
commands do you wanna send?

Regards

Marcel




-------------------------------------------------------
This SF.Net email is sponsored by:
Power Architecture Resource Center: Free content, downloads, discussions,
and more. http://solutions.newsforge.com/ibmarch.tmpl
_______________________________________________
Bluez-devel mailing list
Bluez-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bluez-devel