Hi Masahiro,
On 22 October 2014 03:13, Masahiro Yamada yamada.m@jp.panasonic.com wrote:
Without this commit, functions such as printf(), puts() stop working after the console is ready (= after GD_FLG_DEVINIT is set in console_init_r() function).
The function serial_putc() is called to print a character before the console is available, while serial_stub_putc() is used on the console.
The cause of the problem is that the error handling of ops->putc handler is missing from serial_stub_putc(); it should behave like serial_putc().
Signed-off-by: Masahiro Yamada yamada.m@jp.panasonic.com
drivers/serial/serial-uclass.c | 7 ++++++- 1 file changed, 6 insertions(+), 1 deletion(-)
diff --git a/drivers/serial/serial-uclass.c b/drivers/serial/serial-uclass.c index 6dde4ea..163308b 100644 --- a/drivers/serial/serial-uclass.c +++ b/drivers/serial/serial-uclass.c @@ -127,8 +127,13 @@ void serial_stub_putc(struct stdio_dev *sdev, const char ch) { struct udevice *dev = sdev->priv; struct dm_serial_ops *ops = serial_get_ops(dev);
int err;
ops->putc(dev, ch);
do {err = ops->putc(cur_dev, ch);} while (err == -EAGAIN);if (ch == '\n')serial_putc('\r');}
void serial_stub_puts(struct stdio_dev *sdev, const char *str)
1.9.1
Sorry for the bug, but I did fix it some weeks OK. Please see this patch:
http://patchwork.ozlabs.org/patch/395725/
I recommend basing on dm/working if you are sending new patches because code there is generally on its way upstream ahead of your patch.
I will see if I can start using dm/testing more, as that is really its intended purpose.
Regards, Simon