On 08/09/2010 05:07 AM, Michal Simek wrote:
> CONSOLE_POLL support for uartlite enables
> KGDB debugging over serial line.
>
> Signed-off-by: Michal Simek <[email protected]>
> ---
> drivers/serial/uartlite.c | 26 +++++++++++++++++++++++++-
> 1 files changed, 25 insertions(+), 1 deletions(-)
>
> diff --git a/drivers/serial/uartlite.c b/drivers/serial/uartlite.c
> index caf085d..ce1f3e0 100644
> --- a/drivers/serial/uartlite.c
> +++ b/drivers/serial/uartlite.c
> @@ -322,6 +322,26 @@ static int ulite_verify_port(struct uart_port *port,
> struct serial_struct *ser)
> return -EINVAL;
> }
>
> +#ifdef CONFIG_CONSOLE_POLL
> +static int ulite_get_poll_char(struct uart_port *port)
> +{
> + while (!(ioread32be(port->membase + ULITE_STATUS)
> + & ULITE_STATUS_RXVALID))
> + cpu_relax();
> +
> + return ioread32be(port->membase + ULITE_RX);
> +}
After kdb was merged, the CONSOLE_POLL API had a subtle change where
all I/O drivers should return immediately with a return of
NO_POLL_CHAR in order to allow more than one I/O driver to operate in
the same poll loop.
The patch should look like the following for the get_poll_char.
+static int ulite_get_poll_char(struct uart_port *port)
+{
+ if (!(ioread32be(port->membase + ULITE_STATUS)
+ & ULITE_STATUS_RXVALID))
+ return NO_POLL_CHAR;
+
+ return ioread32be(port->membase + ULITE_RX);
+}
If you agree to the change you can add my SOB.
Signed-off-by: Jason Wessel <[email protected]>
Cheers,
Jason.
------------------------------------------------------------------------------
This SF.net email is sponsored by
Make an app they can't live without
Enter the BlackBerry Developer Challenge
http://p.sf.net/sfu/RIM-dev2dev
_______________________________________________
Kgdb-bugreport mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/kgdb-bugreport
