Hi Rob,

On 09/16/2016 02:20 PM, Rob Herring wrote:
On Thu, Sep 08, 2016 at 11:12:19AM -0500, ttha...@opensource.altera.com wrote:
From: Thor Thayer <ttha...@opensource.altera.com>

Initialize the tx_loadsz parameter if it is defined in the
device tree.

Signed-off-by: Thor Thayer <ttha...@opensource.altera.com>
---
 drivers/tty/serial/8250/8250_of.c |    6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/drivers/tty/serial/8250/8250_of.c 
b/drivers/tty/serial/8250/8250_of.c
index 38963d7..3e4be2f 100644
--- a/drivers/tty/serial/8250/8250_of.c
+++ b/drivers/tty/serial/8250/8250_of.c
@@ -195,6 +195,7 @@ static int of_platform_serial_probe(struct platform_device 
*ofdev)
        switch (port_type) {
        case PORT_8250 ... PORT_MAX_8250:
        {
+               u32 prop;
                struct uart_8250_port port8250;
                memset(&port8250, 0, sizeof(port8250));
                port8250.port = port;
@@ -202,6 +203,11 @@ static int of_platform_serial_probe(struct platform_device 
*ofdev)
                if (port.fifosize)
                        port8250.capabilities = UART_CAP_FIFO;

+               /* Check for TX fifo load size */
+               if (of_property_read_u32(ofdev->dev.of_node,
+                                        "tx-loadsz", &prop) == 0)
+                       port8250.tx_loadsz = prop;

This can be simplified to:

of_property_read_u32(ofdev->dev.of_node, "tx-loadsz", &port8250.tx_loadsz);


Yes, I originally had it implemented using that function but then I realized the port8250.tx_loadsz is an unsigned int and therefore can change size on different platforms.

The assignment handles that. In retrospect, I should probably change this to a of_property_read_u8() anyway so that it won't cause compile warnings in the smaller architectures.

Thank you for reviewing and the comments.

Thor

+
                if (of_property_read_bool(ofdev->dev.of_node,
                                          "auto-flow-control"))
                        port8250.capabilities |= UART_CAP_AFE;
--
1.7.9.5

Reply via email to