| Age | Commit message (Collapse) | Author |
|---|
|
- don't panic if you can't setup interrupts, just quit attach
- cleanup printf's, show a better error message
|
|
|
|
|
|
Finally my "Cyclades Cyclom-8Y" works correctly.
Cleanups:
- no more stupid arrays, we have softc structs, use them.
- no more global poll timeout that walks an array, timeouts are now cheap,
one poll per softc.
- sc_dev.dv_xname is the correct thing to print in diagnostic printfs.
- don't have bus-specific code in the generic attach for @#%$! sake, it
took 20 seconds to fix (I suspect that the workaround took much more
time to write).
- spltty, not splhigh.
- use pci_mapreg_map, not pci_{io,mem}_find + bus_space_map.
- use defines, not magic constants in a few places.
- KNF in a few places (much more work needed).
Now the interesting parts:
- when handling the interrupt, don't just ignore it when the tty is not
open. Read and discard the data. If we don't read it, the card will
interrupt again and again and again and ... (inspired by NetBSD).
- actually enable interrupts on PLX9050
- From FreeBSD: "Work around a PLX9050 bug that causes system lockup in
certain systems"
deraadt@ ok
|
|
|
|
All callers actually took all arguments to pci_intr_map from pci_attach_args
structs, so this simplifies code.
This also allows more complicated interrupt assignment schemes like the one
on sparc64.
This makes sparc64 pci interrupts work.
Inspired by the same change in NetBSD.
|
|
Not finished yet, but cleaned up and moved around code for
better organization. Inspired by NetBSD and other PCI drivers.
From my dmesg:
cy1 at pci0 dev 10 function 0 "Cyclades Cyclom-8Y" rev 0x01
|
|
|
|
|
|
Reported by Berndt Josef Wulf <wulf@ping.net.au>
|
|
NetBSD 961107 state. Other has not, partly because sometimes the bus.h
changes was coupled with other changes.
|
|
possibly taken in small steps
|
|
|
