NOTE: This is a dirty brute force approach, I'm sure there is a proper and more elegant way of getting it running using FC's own tools.
It could work as a temporary solution, anyway, but I would look for a permanent fix.
Modules:
Code:
modprobe serial-8250
You can also try a brute-force approach listing all modules related to serial drivers:
Code:
find /lib/modules/$(uname -r) -name "*serial*"
and try to load them all with
modprobe.
Once you are down with the modules try if the device is there:
Code:
setserial -a /dev/ttyS0 --- tons of info
setserial -b /dev/ttyS0 --- summary of the configs
setserial -G /dev/ttyS0 --- formatted output
To set the device to autoconfig use:
Code:
setserial /dev/ttyS0 autoconfig
In order to get this running at startup you can add two lines to your /etc/rc.local, one to load the module and the second to setup the ttyS0 device:
[quote]
/etc/rc.local
...
modprobe WHICHEVER_MODULE_TO_LOAD
setserial /dev/ttyS0 autoconfig
[quote]
There is also a way to force things a bit more by manually creating device nodes, whichever their kind is: the
mknod command.
To take it short:
Code:
mknod --mode=0777 /dev/ttyS0 c 4 64
I would give the mknod manpage a read, it's a very useful tool.
The list where the mayor and minor numbers (the last two arguments to mknod) are listed in /usr/src/linux/Documentation/devices.txt, with this as a reference you will be able to create
device nodes when needed.
Try it and give us feedback. I'm anyway sure that this problem can be solved in a Fedora Core-Friendly way, specially if anaconda and the hardware related modules are all up and running,
watch for hotplug, coldplug and any FC4/Anaconda specific daemons. These daemons are started and configured from an anaconda GUI or using
ksysv, it's just drag and drop or
clicking check boxes.
Hope this helps
