Date: Thu, 16 Mar 2000 13:02:57 +0100
From: Martin Mares <mj@suse.cz>
Yes, of course, but first of all I'd like to know what exactly
the problems are. I dislike creating yet another architecture
specific function -- I'd prefer to keep the function generic, but
allow the architecture to set the right constants (either by macros
in pci.h [which can be defined to call functions if needed, of
course] or by adding a new entry in struct pci_dev).
In my case for 2.4.x I'm going to need to play "BIOS" on sparc64 and
init all the latency timer (based upon min_gnt, max_latency, and 66MHZ
settings) and cache line size values because the boot firmware doesn't
do it. I want to remove all of the ifdef __sparc__ garbage in PCI
drivers which exists now to accomplish this.
Also for some controllers I know cases where if you set the latency
timer such that PIO's from the host bridge are blocked out for more
than a certain amount of clocks, it can hang the host bridge. I'd
like to be able to control this.
Later,
David S. Miller
davem@redhat.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Thu Mar 23 2000 - 21:00:17 EST