Index: linux-2.6-working/drivers/serial/Kconfig
===================================================================
--- linux-2.6-working.orig/drivers/serial/Kconfig 2006-02-03 14:22:05.000000000 +0000
+++ linux-2.6-working/drivers/serial/Kconfig 2006-02-06 10:17:52.000000000 +0000
@@ -98,6 +98,7 @@
config SERIAL_8250_RUNTIME_UARTS
int "Number of 8250/16550 serial ports to register at runtime"
depends on SERIAL_8250
+ range 0 SERIAL_8250_NR_UARTS
default "4"
help
Set this to the maximum number of serial ports you want
@@ -105,6 +106,9 @@
with the module parameter "nr_uarts", or boot-time parameter
8250.nr_uarts
+ This must be less than or equal to the maximum number of 8250/16550
+ serial ports supported (SERIAL_8250_NR_UARTS).
+
config SERIAL_8250_EXTENDED
bool "Extended 8250/16550 serial driver options"
depends on SERIAL_8250
David Vrabel wrote:
>If SERIAL_8250_RUNTIME_UARTS is > SERIAL_8250_NR_UARTS then more serial
>ports are registered than we've allocated memory for. Prevent this by
>limiting SERIAL_8250_RUNTIME_UARTS in the serial Kconfig.
>
>Signed-off-by: David Vrabel <[email protected]>
>
>
Is there any real use case for having *less* registered serial ports and
having some spare?
- Yuki.
David Vrabel wrote:
> If SERIAL_8250_RUNTIME_UARTS is > SERIAL_8250_NR_UARTS then more serial
> ports are registered than we've allocated memory for. Prevent this by
> limiting SERIAL_8250_RUNTIME_UARTS in the serial Kconfig.
Nevermind. I see Russell has already applied an equivilent patch.
David Vrabel
--
David Vrabel, Design Engineer
Arcom, Clifton Road Tel: +44 (0)1223 411200 ext. 3233
Cambridge CB1 7EA, UK Web: http://www.arcom.com/
On Mon, Feb 06, 2006 at 09:42:19PM +1100, Yuki Cuss wrote:
> David Vrabel wrote:
>
> >If SERIAL_8250_RUNTIME_UARTS is > SERIAL_8250_NR_UARTS then more serial
> >ports are registered than we've allocated memory for. Prevent this by
> >limiting SERIAL_8250_RUNTIME_UARTS in the serial Kconfig.
> >
> >Signed-off-by: David Vrabel <[email protected]>
> >
> >
>
> Is there any real use case for having *less* registered serial ports and
> having some spare?
Having the ability to build a kernel image which supports many serial ports,
whilst at the same same time when booted on the common-case systems with
two serial ports, not creating so many /dev/ttyS* nodes or sysfs objects
wasting ram that'll never be used or reclaimed.
Dave