Here is the story:
Somewhere back near 2.1.56 I built my very first dev kernel while I was away
on vacation. When I got back, my pcmcia controller didn't work. I figured that
this was a hardware problem due to some rough treatment the computer had
received on the trip. I traced the problem through the kernel and could see
very clearly that the problem was that the i82365 chip was probed it returned
that there were 0 slots.
The interesting thing was that when I plugged the computer into a docking
station which also has two slots in it, the probe returned that two slots were
available. However, the two slots that worked were the ones internal to my
laptop rather than those in the docking station. i.e. The only thing that was
wrong was the counting of the slots not the physical functioning of the slots.
The hardware tech's swore up and down that it was a OS issue and I swore that
it was a hardware issue. To finally settle things they plugged a different HD
with Win95 on it into my laptop. Amazingly, all the pcmcia slots worked the way
they were supposed to. The thing that surprised me even more was that after
we put the linux drive back into my computer, the pcmcia slots continued to
work. I figured that it had something to do with the plug and play
section of the kernel writing something to some device's nvram and getting it
into a state where the i82365 didn't respond properly. So I disabled, the
plug and play section of the kernel, recompiled and lived happily ever after
all the way up to 2.1.103.
When 2.1.103 came out, I decided to take a quick tour through the options and
see if anything had changed. Along the way, I enabled plug and play (I had
completely forgotten about the problem I had previously) and then booted to
the new kernel. It seems like the same thing happened. My pcmcia controller
now reports "device or resource busy" when the module tries to load. Falling
back to earlier versions of the kernel is of no use. They all exhibit the same
problem. I also went into the CMOS settings and disabled all the devices so I
could to see if it might be a bit of a resource problem. No luck.
The questions:
Is there anyway I can figure out what got screwed up in my computer and fix it
without having to boot a M$ OS?
Is this a bug or am I being clueless?
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu