Re: [hylafax-users] /dev/ttySx is not a terminal device.

[Aidan Van Dyk <aidan@xxxxxxxx>]

* Lee Howard <faxguy@xxxxxxxxxxxxxxxx> [051129 16:14]:

> >fax2:~# dmesg | grep tty
> >ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
> >ttyS1 at I/O 0x2f8 (irq = 3) is a 16550A
> >ttyS14 at I/O 0xb400 (irq = 193) is a 16550A
> >
> 
> This is something that I've noticed with recent (kernel 2.6) versions of 
> the serial driver.
> 
> It *should* be ttyS4 and probably not IRQ 193.

Note that on a 2.6 kernel and modern motherboard, you can *very well*
get IRQ numbers like that.  These modern machines use IOAPIC/LAPIC,
where each pci slot can have it's interrupt wires directly connected to
the interrupt controller.  


	[aidan@host ~]$ dmesg | grep tty
	ttyS0 at I/O 0x3f8 (irq = 4) is a 16550A
	ttyS4 at I/O 0xcc80 (irq = 233) is a 16550A
	[aidan@host ~]$ cat /proc/interrupts 
		   CPU0       CPU1       CPU2       CPU3       
	  0:  215372042  215381090  215374171  215427572    IO-APIC-edge  timer
[SNIP]
	209:     531909          0          0         12   IO-APIC-level  eth0
	233:      38725      48651      47778      38803   IO-APIC-level  serial
	NMI:  861554605  861554564  861554563  861554562 
	LOC:  861538695  861542006  861538726  861541991 
	ERR:          0
	MIS:          0


Has udev properly created /dev/ttyS14?

Note that you also *could* force (with setserial) it to be
ttyS4/ttyS5/ttyS6/whatever by setting the IO port and IRQs.

a.

Attachment: signature.asc
Description: Digital signature