Skip to contentSkip to conversation

usb/serial/io_ti.c: inconsequent NULL checking

PATCH In review Adrian Bunk 4 participants 6 messages Feb 19, 2008 – Feb 22, 2008

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

The Coverity checker spotted the following inconsequent NULL checking introduced by commit d5f5bcd425b771c0b7ff5a650b2ce061ac8bbb87:

<-- snip -->

static int edge_open (struct usb_serial_port *port, struct file * filp) {
...

        if (port->tty)   <---------------------------------------------
                port->tty->low_latency = low_latency;

        port_number = port->number - port->serial->minor;
        switch (port_number) {
                case 0:
                        edge_port->uart_base = UMPMEM_BASE_UART1;
                        edge_port->dma_address = UMPD_OEDB1_ADDRESS;
                        break;
                case 1:
                        edge_port->uart_base = UMPMEM_BASE_UART2;
                        edge_port->dma_address = UMPD_OEDB2_ADDRESS;
                        break;
                default:
                        dev_err (&port->dev, "Unknown port number!!!\n");
                        return -ENODEV;
        }

        dbg ("%s - port_number = %d, uart_base = %04x, dma_address = %04x",
             __FUNCTION__, port_number, edge_port->uart_base, edge_port->dma_address);

        dev = port->serial->dev;

        memset (&(edge_port->icount), 0x00, sizeof(edge_port->icount));
        init_waitqueue_head (&edge_port->delta_msr_wait);

        /* turn off loopback */
        status = TIClearLoopBack (edge_port);
        if (status) {
                dev_err(&port->dev,"%s - cannot send clear loopback command, %d\n",
                        __FUNCTION__, status);
                return status;
        }
        
        /* set up the port settings */
        edge_set_termios (port, port->tty->termios);

... ^^^^^^^^^^^^^^^^^^

<-- snip -->

cu
Adrian

Signature · 4 lines
       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed
#1 lore raw
Quoted 4 lines · Adrian Bunk

The Coverity checker spotted the following inconsequent NULL checking introduced by commit d5f5bcd425b771c0b7ff5a650b2ce061ac8bbb87:

<-- snip -->

It's not a real problem, that pointer can't go away.

thanks,

greg k-h

#2 lore raw
Quoted 7 lines · Greg KH
Quoted 4 lines · Adrian Bunk

The Coverity checker spotted the following inconsequent NULL checking introduced by commit d5f5bcd425b771c0b7ff5a650b2ce061ac8bbb87:

<-- snip -->

It's not a real problem, that pointer can't go away.

I think he meant inconsistent not "inconsequent."

Either the test of port->tty here is unneeded:

if (port->tty)

               port->tty->low_latency = low_latency;

...or the lack of test of port->tty here is a mistake:

	 edge_set_termios (port, port->tty->termios);
#3 lore raw
Quoted 8 lines

Either the test of port->tty here is unneeded:

if (port->tty)

               port->tty->low_latency = low_latency;

...or the lack of test of port->tty here is a mistake:

	 edge_set_termios (port, port->tty->termios);

Interesting - so coverity doesn't understand the BKL. It's producing the right answer, for the right reason but the assumption it makes isn't 100% safe.

The check can go.

Alan

#4 lore raw

PATCH 1/1 usb/serial/io_ti.c: remove unneeded NULL check

Quoted 5 lines · Alan Cox
Quoted 5 lines

Either the test of port->tty here is unneeded:

if (port->tty)

               port->tty->low_latency = low_latency;
Quoted 7 lines
Quoted 3 lines

...or the lack of test of port->tty here is a mistake:

	 edge_set_termios (port, port->tty->termios);

Interesting - so coverity doesn't understand the BKL. It's producing the right answer, for the right reason but the assumption it makes isn't 100% safe.

Coverity doesn't perform magic.
It simply notices that it's once checked and once dereferenced unconditionally.

Quoted 1 line

The check can go.

Thanks, patch below.

Quoted 1 line

Alan

cu
Adrian

<-- snip -->

There's no reason for checking pdev->bus for being NULL here (and we'd anyway Oops below if it was).

---
afefd6935c788b3e4071092ebc04c4d356dd7496 diff --git a/drivers/usb/serial/io_ti.c b/drivers/usb/serial/io_ti.c index cd34059..700683c 100644

drivers/usb/serial/io_ti.c +1 -2
@@ -1944,8 +1944,7 @@ static int edge_open (struct usb_serial_port *port, struct file * filp)
 	if (edge_port == NULL)
 		return -ENODEV;
 
-	if (port->tty)
-		port->tty->low_latency = low_latency;
+	port->tty->low_latency = low_latency;
 
 	port_number = port->number - port->serial->minor;
 	switch (port_number) {
#5 lore raw

Re: [2.6 patch] usb/serial/io_ti.c: remove unneeded NULL check

Already patched in my tree