Re: [RFC/CFT] cmd640 irqlocking fixes

Alan Cox (alan@lxorguk.ukuu.org.uk)
25 Jul 2002 13:52:35 +0100


On Thu, 2002-07-25 at 11:51, Andre Hedrick wrote:
> I am sorry Alan, but I fixed all of the locking code in 2.4 a long time
> ago, and you adopted it somewhere around this patch.

Andre - the PCI probe fixes which this is about is something I wrote.
The other locking stuff which is unrelated to this discussion is your
code but thats not Im talking about - OK.

> Again I have to call this patch and fix and take credit and full ownership
> of removing all the save/cli/sti/restore which littered the driver and
> were spread like cow patties through a chopper gun.

That patch is wrong by the way because I made a mistake in 2.4. Your PCI
config locking should be using pci_config lock not io_request. I'll fix
the 2.4 tree in a bit now I've tidied up the 2.5 version.

> @@ -748,10 +737,8 @@
> bus_type = "VLB";
> } else {
> cmd640_vlb = 0;
> - /*
> - * Don't leak I/O cycles on the PCI bus by blindly attempting
> - * a configuration cycle type that is not supported by the hardware.
> - */
> + /* Find out what kind of PCI probing is supported otherwise
> + Justin Gibbs will sulk.. */

Just ask Justin Gibbs. He didn't like my comment 8) Im sure he remembers
who wrote it 8)

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/