On Thu, 12 Jul 2007 11:47:12 -0700 Andrew Morton
<akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> On Thu, 12 Jul 2007 18:39:00 +0100 (BST)
> "Maciej W. Rozycki" <macro@xxxxxxxxxxxxxx> wrote:
>
> > This is a driver for the SB1250 DUART, a dual serial port
> > implementation included in the Broadcom family of SOCs descending
> > from the SiByte SB1250 MIPS64 chip multiprocessor. It is a new
> > implementation replacing the old-fashioned driver currently present
> > in the linux-mips.org tree. It supports all the usual features one
> > would expect from a(n asynchronous) serial driver, including modem
> > line control (as far as hardware supports it -- there is edge
> > detection logic missing from the DCD and RI lines and the driver
> > does not implement polling of these lines at the moment), the
> > serial console, BREAK transmission and reception, including the
> > magic SysRq. The receive FIFO threshold is not maintained though.
> >
> > ...
> >
> > +
> > +#if defined(CONFIG_SIBYTE_BCM1x55) ||
> > defined(CONFIG_SIBYTE_BCM1x80) +#include <asm/sibyte/bcm1480_regs.h>
> > +#include <asm/sibyte/bcm1480_int.h>
> > +
> > +#define SBD_CHANREGS(line) A_BCM1480_DUART_CHANREG((line),
> > 0) +#define SBD_CTRLREGS(line)
> > A_BCM1480_DUART_CTRLREG((line), 0) +#define
> > SBD_INT(line) (K_BCM1480_INT_UART_0 + (line)) +
> > +#elif defined(CONFIG_SIBYTE_SB1250) ||
> > defined(CONFIG_SIBYTE_BCM112X) +#include <asm/sibyte/sb1250_regs.h>
> > +#include <asm/sibyte/sb1250_int.h>
> > +
> > +#define SBD_CHANREGS(line) A_DUART_CHANREG((line), 0)
> > +#define SBD_CTRLREGS(line) A_DUART_CTRLREG(0)
> > +#define SBD_INT(line) (K_INT_UART_0 + (line))
> > +
> > +#else
> > +#error invalid SB1250 UART configuration
> > +
> > +#endif
>
> If the #error can trigger, the Kconfig is broken, yes? (No action is
> required though - it's always good to have checks)
>
> > +#define to_sport(uport) container_of(uport, struct sbd_port, port)
>
> That didn't need to be implemented as a macro.
>
> > +#define __unused __attribute__((__unused__))
>
> Please use the generic implementations here. `grep unused
> include/linux/compiler*.h'.
>
> > +/*
> > + * In bug 1956, we get glitches that can mess up uart registers.
> > This
> > + * "read-mode-reg after any register access" is an accepted
> > workaround.
> > + */
>
> <looks in bugzilla.kernel.org>
>
> <wonders how a uart driver can fix an smbfs bug>
>
> Perhaps a reference to where that bug number came from?
It's a Sibyte errata/WAR number as I recall.
> > +static void __war_sbd1956(struct sbd_port *sport)
> > +{
> > + __read_sbdchn(sport, R_DUART_MODE_REG_1);
> > + __read_sbdchn(sport, R_DUART_MODE_REG_2);
> > +}
> >
> > ...
> >
> > +static struct uart_ops sbd_ops = {
>
> I suppose if we made this const, something would blow up.
>
>
>
>
> There is no power management support in this driver.
>
