[parisc-linux] kernel update bug?

[parisc-linux] kernel update bug?

[b.gunreben]

Hi,

I recently got a C360 to play around with it. After installing an old beta
of woody (thats the distribution, I got on my private net), everything seemd
fine. The kernel --
Linux version 2.4.9-32 (root@paer) (gcc version 3.0.2 (Debian)) #1 Fri Nov
30 19:36:30 MST 2001 -- worked just fine. 

After this, I installed a newer kernel (well, I just copied the kernel, and
omitted the modules):  -- 
Linux version 2.4.20-pa18 (bame@dsl2) (gcc version 3.0.3) #1 Fri Dec 27
06:12:20 MST 2002 --
and with this kernel, the machine crashes reproduceable when starting up. 

Now the reason: I found, that with kernel 2.4.9, there were 3 serial devices
found

/dev/ttyS0 uart 16550A port 0x0000 irq 90 baud_base 454545 spd_normal
/dev/ttyS1 uart 16550A port 0x0000 irq 106 baud_base 454545 spd_normal
/dev/ttyS2 uart 16550A port 0x0000 irq 170 baud_base 454545 spd_normal

the later kernel finds only the first two of these interfaces, and restarts
when restoring the settings of /dev/ttyS2. Later it crashes completely of
course. I could fix this behaviour by commenting out the last line in
/etc/serial.conf.

Now my question: should kernel 2.4.20 find two or three serial devices on a
C360, and why does the kernel crash if setserial tries to configure a
(nonexistant?) device?


Berthold

Re: [parisc-linux] kernel update bug?

[Matthew Wilcox]

On Thu, Jan 23, 2003 at 10:19:00PM +0100, b.gunreben wrote:
> Now the reason: I found, that with kernel 2.4.9, there were 3 serial devices
> found
> 
> /dev/ttyS0 uart 16550A port 0x0000 irq 90 baud_base 454545 spd_normal
> /dev/ttyS1 uart 16550A port 0x0000 irq 106 baud_base 454545 spd_normal
> /dev/ttyS2 uart 16550A port 0x0000 irq 170 baud_base 454545 spd_normal
> 
> the later kernel finds only the first two of these interfaces, and restarts
> when restoring the settings of /dev/ttyS2. Later it crashes completely of
> course. I could fix this behaviour by commenting out the last line in
> /etc/serial.conf.
> 
> Now my question: should kernel 2.4.20 find two or three serial devices on a
> C360, and why does the kernel crash if setserial tries to configure a
> (nonexistant?) device?

Well, how many serial ports do you have?  ;-)

It was a bug up till (i think) a 2.4.17 kernel that we found a ghost
serial port.  And setserial is a broken piece of crap that will crash
the machine if it tries to set serial ports that don't exist.

-- 
"It's not Hollywood.  War is real, war is primarily not about defeat or
victory, it is about death.  I've seen thousands and thousands of dead bodies.
Do you think I want to have an academic debate on this subject?" -- Robert Fisk