[U-Boot] [PATCH] dm: serial: Move a carriage return before a line feed
In general, a carriage return needs to execute before a line feed. The patch is to change serial DM driver based on this rule.
Signed-off-by: Alison Wang alison.wang@nxp.com --- drivers/serial/serial-uclass.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/serial/serial-uclass.c b/drivers/serial/serial-uclass.c index 1c447ff..f154eb1 100644 --- a/drivers/serial/serial-uclass.c +++ b/drivers/serial/serial-uclass.c @@ -123,11 +123,12 @@ static void _serial_putc(struct udevice *dev, char ch) struct dm_serial_ops *ops = serial_get_ops(dev); int err;
+ if (ch == '\n') + _serial_putc(dev, '\r'); + do { err = ops->putc(dev, ch); } while (err == -EAGAIN); - if (ch == '\n') - _serial_putc(dev, '\r'); }
static void _serial_puts(struct udevice *dev, const char *str)
On 23 February 2016 at 20:56, Alison Wang b18965@freescale.com wrote:
In general, a carriage return needs to execute before a line feed. The patch is to change serial DM driver based on this rule.
Signed-off-by: Alison Wang alison.wang@nxp.com
drivers/serial/serial-uclass.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
Reviewed-by: Simon Glass sjg@chromium.org
Hi Alison,
On Wed, Feb 24, 2016 at 11:56 AM, Alison Wang b18965@freescale.com wrote:
In general, a carriage return needs to execute before a line feed. The patch is to change serial DM driver based on this rule.
Signed-off-by: Alison Wang alison.wang@nxp.com
drivers/serial/serial-uclass.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/serial/serial-uclass.c b/drivers/serial/serial-uclass.c index 1c447ff..f154eb1 100644 --- a/drivers/serial/serial-uclass.c +++ b/drivers/serial/serial-uclass.c @@ -123,11 +123,12 @@ static void _serial_putc(struct udevice *dev, char ch) struct dm_serial_ops *ops = serial_get_ops(dev); int err;
if (ch == '\n')_serial_putc(dev, '\r');do { err = ops->putc(dev, ch); } while (err == -EAGAIN);
if (ch == '\n')_serial_putc(dev, '\r');}
We should also clean up all DM serial driver to remove this handling in their driver. eg: serial_lpuart.c/serial_mxc.c
Regards, Bin
Hi, Bin,
On Wed, Feb 24, 2016 at 11:56 AM, Alison Wang b18965@freescale.com wrote:
In general, a carriage return needs to execute before a line feed. The patch is to change serial DM driver based on this rule.
Signed-off-by: Alison Wang alison.wang@nxp.com
drivers/serial/serial-uclass.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/serial/serial-uclass.c b/drivers/serial/serial-uclass.c index 1c447ff..f154eb1 100644 --- a/drivers/serial/serial-uclass.c +++ b/drivers/serial/serial-uclass.c @@ -123,11 +123,12 @@ static void _serial_putc(struct udevice *dev,
char ch)
struct dm_serial_ops *ops = serial_get_ops(dev); int err;
if (ch == '\n')_serial_putc(dev, '\r');do { err = ops->putc(dev, ch); } while (err == -EAGAIN);
if (ch == '\n')_serial_putc(dev, '\r');}
We should also clean up all DM serial driver to remove this handling in their driver. eg: serial_lpuart.c/serial_mxc.c
[Alison Wang] Thanks for your advice. Yes, I agree. This handling in particular driver is redundant.
It seems only serial_lpuart.c and serial_arc.c has this handling. For serial_mxc.c, this handling is only in the non-DM .putc function, but not in the DM .putc function. Do you think it is necessary to change the non-DM .putc function?
Best Regards, Alison Wang
Hi Alison,
On Wed, Feb 24, 2016 at 3:35 PM, Huan Wang alison.wang@nxp.com wrote:
Hi, Bin,
On Wed, Feb 24, 2016 at 11:56 AM, Alison Wang b18965@freescale.com wrote:
In general, a carriage return needs to execute before a line feed. The patch is to change serial DM driver based on this rule.
Signed-off-by: Alison Wang alison.wang@nxp.com
drivers/serial/serial-uclass.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/serial/serial-uclass.c b/drivers/serial/serial-uclass.c index 1c447ff..f154eb1 100644 --- a/drivers/serial/serial-uclass.c +++ b/drivers/serial/serial-uclass.c @@ -123,11 +123,12 @@ static void _serial_putc(struct udevice *dev,
char ch)
struct dm_serial_ops *ops = serial_get_ops(dev); int err;
if (ch == '\n')_serial_putc(dev, '\r');do { err = ops->putc(dev, ch); } while (err == -EAGAIN);
if (ch == '\n')_serial_putc(dev, '\r');}
We should also clean up all DM serial driver to remove this handling in their driver. eg: serial_lpuart.c/serial_mxc.c
[Alison Wang] Thanks for your advice. Yes, I agree. This handling in particular driver is redundant.
It seems only serial_lpuart.c and serial_arc.c has this handling. For serial_mxc.c, this handling is only in the non-DM .putc function, but not in the DM .putc function. Do you think it is necessary to change the non-DM .putc function?
No, we should keep the non-DM putc function as it is now. But this serial_mxc.c, its non-DM putc version has the same issue as the serial-uclass.c, so it should be fixed. The legacy serial_pxa.c, serial_s3c24x0.c and usbtty.c also have the same problem.
Regards, Bin
Hi, Bin,
On Wed, Feb 24, 2016 at 3:35 PM, Huan Wang alison.wang@nxp.com wrote:
Hi, Bin,
On Wed, Feb 24, 2016 at 11:56 AM, Alison Wang b18965@freescale.com wrote:
In general, a carriage return needs to execute before a line feed. The patch is to change serial DM driver based on this rule.
Signed-off-by: Alison Wang alison.wang@nxp.com
drivers/serial/serial-uclass.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
diff --git a/drivers/serial/serial-uclass.c b/drivers/serial/serial-uclass.c index 1c447ff..f154eb1 100644 --- a/drivers/serial/serial-uclass.c +++ b/drivers/serial/serial-uclass.c @@ -123,11 +123,12 @@ static void _serial_putc(struct udevice *dev,
char ch)
struct dm_serial_ops *ops = serial_get_ops(dev); int err;
if (ch == '\n')_serial_putc(dev, '\r');do { err = ops->putc(dev, ch); } while (err == -EAGAIN);
if (ch == '\n')_serial_putc(dev, '\r');}
We should also clean up all DM serial driver to remove this handling in their driver. eg: serial_lpuart.c/serial_mxc.c
[Alison Wang] Thanks for your advice. Yes, I agree. This handling in particular driver is redundant.
It seems only serial_lpuart.c and serial_arc.c has this handling. For serial_mxc.c, this handling is only in the non-DM .putc function, but not in the DM .putc function. Do you think it is necessary to change the non-DM .putc function?
No, we should keep the non-DM putc function as it is now. But this serial_mxc.c, its non-DM putc version has the same issue as the serial- uclass.c, so it should be fixed. The legacy serial_pxa.c, serial_s3c24x0.c and usbtty.c also have the same problem.
[Alison Wang] Ok, I get your idea. I will send another patch to fix them.
Thanks.
Best Regards, Alison Wang
On Wed, Feb 24, 2016 at 11:56 AM, Alison Wang b18965@freescale.com wrote:
In general, a carriage return needs to execute before a line feed. The patch is to change serial DM driver based on this rule.
Signed-off-by: Alison Wang alison.wang@nxp.com
drivers/serial/serial-uclass.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-)
Reviewed-by: Bin Meng bmeng.cn@gmail.com
participants (4)
-
Alison Wang -
Bin Meng -
Huan Wang -
Simon Glass