chicago.c revision 5a61720e6468f16cdd018ced5bc5c81626103ee6
2N/A * The contents of this file are subject to the terms of the 2N/A * Common Development and Distribution License, Version 1.0 only 2N/A * (the "License"). You may not use this file except in compliance 2N/A * See the License for the specific language governing permissions 2N/A * and limitations under the License. 2N/A * When distributing Covered Code, include this CDDL HEADER in each 2N/A * If applicable, add the following below this CDDL HEADER, with the 2N/A * fields enclosed by brackets "[]" replaced with your own identifying 2N/A * information: Portions Copyright [yyyy] [name of copyright owner] 2N/A * Copyright 2005 Sun Microsystems, Inc. All rights reserved. 2N/A * Use is subject to license terms. 2N/A#
pragma ident "%Z%%M% %I% %E% SMI" 2N/A /* Set appropriate tod module for Chicago */ 2N/A * Definitions for accessing the pci config space of the isa node 2N/A * Install power driver which handles the power button. 2N/A * It is OK to return error because 'us' driver is not available 2N/A * in all clusters (e.g. missing in Core cluster). 2N/A * mc-us3i must stay loaded for plat_get_mem_unum() 2N/A * Figure out which mi2cv dip is shared with OBP for the nvram 2N/A * device, so the lock can be acquired. 2N/A * todds1337 needs early attach 2N/A return (
ENOTSUP);
/* not supported on this platform */ 2N/A return (
ENOTSUP);
/* not supported on this platform */ 2N/A * Fiesta support for lgroups. 2N/A * On fiesta platform, an lgroup platform handle == CPU id 2N/A * Macro for extracting the CPU number from the CPU id 2N/A * Return the platform handle for the lgroup containing the given CPU 2N/A * Platform specific lgroup initialization 2N/A * Count the number of CPUs installed to determine if 2N/A * NUMA optimization should be enabled or not. 2N/A * All CPU nodes reside in the root node and have a 2N/A * device type "cpu". 2N/A * Set tuneables for fiesta architecture 2N/A * lgrp_expand_proc_thresh is the minimum load on the lgroups 2N/A * this process is currently running on before considering 2N/A * expanding threads to another lgroup. 2N/A * lgrp_expand_proc_diff determines how much less the remote lgroup 2N/A * must be loaded before expanding to it. 2N/A * Optimize for memory bandwidth by spreading multi-threaded 2N/A * program to different lgroups. 2N/A * Return latency between "from" and "to" lgroups 2N/A * This latency number can only be used for relative comparison 2N/A * between lgroups on the running system, cannot be used across platforms, 2N/A * and may not reflect the actual latency. It is platform and implementation 2N/A * specific, so platform gets to decide its value. It would be nice if the 2N/A * number was at least proportional to make comparisons more meaningful though. 2N/A * NOTE: The numbers below are supposed to be load latencies for uncached 2N/A * memory divided by 10. 2N/A * Return remote latency when there are more than two lgroups 2N/A * (root and child) and getting latency between two different 2N/A * lgroups or root is involved 2N/A * Assign memnode to lgroups 2N/A * Chicago memory controller portid == global CPU id 2N/A * Boot install lists are arranged <addr, len>, <addr, len>, ... 2N/A * Common locking enter code 2N/A * Common locking exit code 2N/A * Called by mi2cv driver 2N/A * Called by mi2cv driver