summaryrefslogtreecommitdiffstats
path: root/arch/parisc/kernel/inventory.c
diff options
context:
space:
mode:
authorLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
committerLinus Torvalds <torvalds@ppc970.osdl.org>2005-04-16 15:20:36 -0700
commit1da177e4c3f41524e886b7f1b8a0c1fc7321cac2 (patch)
tree0bba044c4ce775e45a88a51686b5d9f90697ea9d /arch/parisc/kernel/inventory.c
downloadop-kernel-dev-1da177e4c3f41524e886b7f1b8a0c1fc7321cac2.zip
op-kernel-dev-1da177e4c3f41524e886b7f1b8a0c1fc7321cac2.tar.gz
Linux-2.6.12-rc2v2.6.12-rc2
Initial git repository build. I'm not bothering with the full history, even though we have it. We can create a separate "historical" git archive of that later if we want to, and in the meantime it's about 3.2GB when imported into git - space that would just make the early git days unnecessarily complicated, when we don't have a lot of good infrastructure for it. Let it rip!
Diffstat (limited to 'arch/parisc/kernel/inventory.c')
-rw-r--r--arch/parisc/kernel/inventory.c612
1 files changed, 612 insertions, 0 deletions
diff --git a/arch/parisc/kernel/inventory.c b/arch/parisc/kernel/inventory.c
new file mode 100644
index 0000000..1a1c664
--- /dev/null
+++ b/arch/parisc/kernel/inventory.c
@@ -0,0 +1,612 @@
+/*
+ * inventory.c
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version
+ * 2 of the License, or (at your option) any later version.
+ *
+ * Copyright (c) 1999 The Puffin Group (David Kennedy and Alex deVries)
+ * Copyright (c) 2001 Matthew Wilcox for Hewlett-Packard
+ *
+ * These are the routines to discover what hardware exists in this box.
+ * This task is complicated by there being 3 different ways of
+ * performing an inventory, depending largely on the age of the box.
+ * The recommended way to do this is to check to see whether the machine
+ * is a `Snake' first, then try System Map, then try PAT. We try System
+ * Map before checking for a Snake -- this probably doesn't cause any
+ * problems, but...
+ */
+
+#include <linux/types.h>
+#include <linux/kernel.h>
+#include <linux/init.h>
+#include <linux/slab.h>
+#include <linux/mm.h>
+#include <asm/hardware.h>
+#include <asm/io.h>
+#include <asm/mmzone.h>
+#include <asm/pdc.h>
+#include <asm/pdcpat.h>
+#include <asm/processor.h>
+#include <asm/page.h>
+#include <asm/parisc-device.h>
+
+/*
+** Debug options
+** DEBUG_PAT Dump details which PDC PAT provides about ranges/devices.
+*/
+#undef DEBUG_PAT
+
+int pdc_type = PDC_TYPE_ILLEGAL;
+
+void __init setup_pdc(void)
+{
+ long status;
+ unsigned int bus_id;
+ struct pdc_system_map_mod_info module_result;
+ struct pdc_module_path module_path;
+ struct pdc_model model;
+#ifdef __LP64__
+ struct pdc_pat_cell_num cell_info;
+#endif
+
+ /* Determine the pdc "type" used on this machine */
+
+ printk(KERN_INFO "Determining PDC firmware type: ");
+
+ status = pdc_system_map_find_mods(&module_result, &module_path, 0);
+ if (status == PDC_OK) {
+ pdc_type = PDC_TYPE_SYSTEM_MAP;
+ printk("System Map.\n");
+ return;
+ }
+
+ /*
+ * If the machine doesn't support PDC_SYSTEM_MAP then either it
+ * is a pdc pat box, or it is an older box. All 64 bit capable
+ * machines are either pdc pat boxes or they support PDC_SYSTEM_MAP.
+ */
+
+ /*
+ * TODO: We should test for 64 bit capability and give a
+ * clearer message.
+ */
+
+#ifdef __LP64__
+ status = pdc_pat_cell_get_number(&cell_info);
+ if (status == PDC_OK) {
+ pdc_type = PDC_TYPE_PAT;
+ printk("64 bit PAT.\n");
+ return;
+ }
+#endif
+
+ /* Check the CPU's bus ID. There's probably a better test. */
+
+ status = pdc_model_info(&model);
+
+ bus_id = (model.hversion >> (4 + 7)) & 0x1f;
+
+ switch (bus_id) {
+ case 0x4: /* 720, 730, 750, 735, 755 */
+ case 0x6: /* 705, 710 */
+ case 0x7: /* 715, 725 */
+ case 0x8: /* 745, 747, 742 */
+ case 0xA: /* 712 and similiar */
+ case 0xC: /* 715/64, at least */
+
+ pdc_type = PDC_TYPE_SNAKE;
+ printk("Snake.\n");
+ return;
+
+ default: /* Everything else */
+
+ printk("Unsupported.\n");
+ panic("If this is a 64-bit machine, please try a 64-bit kernel.\n");
+ }
+}
+
+#define PDC_PAGE_ADJ_SHIFT (PAGE_SHIFT - 12) /* pdc pages are always 4k */
+
+static void __init
+set_pmem_entry(physmem_range_t *pmem_ptr, unsigned long start,
+ unsigned long pages4k)
+{
+ /* Rather than aligning and potentially throwing away
+ * memory, we'll assume that any ranges are already
+ * nicely aligned with any reasonable page size, and
+ * panic if they are not (it's more likely that the
+ * pdc info is bad in this case).
+ */
+
+ if ( ((start & (PAGE_SIZE - 1)) != 0)
+ || ((pages4k & ((1UL << PDC_PAGE_ADJ_SHIFT) - 1)) != 0) ) {
+
+ panic("Memory range doesn't align with page size!\n");
+ }
+
+ pmem_ptr->start_pfn = (start >> PAGE_SHIFT);
+ pmem_ptr->pages = (pages4k >> PDC_PAGE_ADJ_SHIFT);
+}
+
+static void __init pagezero_memconfig(void)
+{
+ unsigned long npages;
+
+ /* Use the 32 bit information from page zero to create a single
+ * entry in the pmem_ranges[] table.
+ *
+ * We currently don't support machines with contiguous memory
+ * >= 4 Gb, who report that memory using 64 bit only fields
+ * on page zero. It's not worth doing until it can be tested,
+ * and it is not clear we can support those machines for other
+ * reasons.
+ *
+ * If that support is done in the future, this is where it
+ * should be done.
+ */
+
+ npages = (PAGE_ALIGN(PAGE0->imm_max_mem) >> PAGE_SHIFT);
+ set_pmem_entry(pmem_ranges,0UL,npages);
+ npmem_ranges = 1;
+}
+
+#ifdef __LP64__
+
+/* All of the PDC PAT specific code is 64-bit only */
+
+/*
+** The module object is filled via PDC_PAT_CELL[Return Cell Module].
+** If a module is found, register module will get the IODC bytes via
+** pdc_iodc_read() using the PA view of conf_base_addr for the hpa parameter.
+**
+** The IO view can be used by PDC_PAT_CELL[Return Cell Module]
+** only for SBAs and LBAs. This view will cause an invalid
+** argument error for all other cell module types.
+**
+*/
+
+static int __init
+pat_query_module(ulong pcell_loc, ulong mod_index)
+{
+ pdc_pat_cell_mod_maddr_block_t pa_pdc_cell;
+ unsigned long bytecnt;
+ unsigned long temp; /* 64-bit scratch value */
+ long status; /* PDC return value status */
+ struct parisc_device *dev;
+
+ /* return cell module (PA or Processor view) */
+ status = pdc_pat_cell_module(&bytecnt, pcell_loc, mod_index,
+ PA_VIEW, &pa_pdc_cell);
+
+ if (status != PDC_OK) {
+ /* no more cell modules or error */
+ return status;
+ }
+
+ temp = pa_pdc_cell.cba;
+ dev = alloc_pa_dev(PAT_GET_CBA(temp), &pa_pdc_cell.mod_path);
+ if (!dev) {
+ return PDC_NE_MOD;
+ }
+
+ /* alloc_pa_dev sets dev->hpa */
+
+ /*
+ ** save parameters in the parisc_device
+ ** (The idea being the device driver will call pdc_pat_cell_module()
+ ** and store the results in its own data structure.)
+ */
+ dev->pcell_loc = pcell_loc;
+ dev->mod_index = mod_index;
+
+ /* save generic info returned from the call */
+ /* REVISIT: who is the consumer of this? not sure yet... */
+ dev->mod_info = pa_pdc_cell.mod_info; /* pass to PAT_GET_ENTITY() */
+ dev->pmod_loc = pa_pdc_cell.mod_location;
+
+ register_parisc_device(dev); /* advertise device */
+
+#ifdef DEBUG_PAT
+ pdc_pat_cell_mod_maddr_block_t io_pdc_cell;
+ /* dump what we see so far... */
+ switch (PAT_GET_ENTITY(dev->mod_info)) {
+ unsigned long i;
+
+ case PAT_ENTITY_PROC:
+ printk(KERN_DEBUG "PAT_ENTITY_PROC: id_eid 0x%lx\n",
+ pa_pdc_cell.mod[0]);
+ break;
+
+ case PAT_ENTITY_MEM:
+ printk(KERN_DEBUG
+ "PAT_ENTITY_MEM: amount 0x%lx min_gni_base 0x%lx min_gni_len 0x%lx\n",
+ pa_pdc_cell.mod[0], pa_pdc_cell.mod[1],
+ pa_pdc_cell.mod[2]);
+ break;
+ case PAT_ENTITY_CA:
+ printk(KERN_DEBUG "PAT_ENTITY_CA: %ld\n", pcell_loc);
+ break;
+
+ case PAT_ENTITY_PBC:
+ printk(KERN_DEBUG "PAT_ENTITY_PBC: ");
+ goto print_ranges;
+
+ case PAT_ENTITY_SBA:
+ printk(KERN_DEBUG "PAT_ENTITY_SBA: ");
+ goto print_ranges;
+
+ case PAT_ENTITY_LBA:
+ printk(KERN_DEBUG "PAT_ENTITY_LBA: ");
+
+ print_ranges:
+ pdc_pat_cell_module(&bytecnt, pcell_loc, mod_index,
+ IO_VIEW, &io_pdc_cell);
+ printk(KERN_DEBUG "ranges %ld\n", pa_pdc_cell.mod[1]);
+ for (i = 0; i < pa_pdc_cell.mod[1]; i++) {
+ printk(KERN_DEBUG
+ " PA_VIEW %ld: 0x%016lx 0x%016lx 0x%016lx\n",
+ i, pa_pdc_cell.mod[2 + i * 3], /* type */
+ pa_pdc_cell.mod[3 + i * 3], /* start */
+ pa_pdc_cell.mod[4 + i * 3]); /* finish (ie end) */
+ printk(KERN_DEBUG
+ " IO_VIEW %ld: 0x%016lx 0x%016lx 0x%016lx\n",
+ i, io_pdc_cell.mod[2 + i * 3], /* type */
+ io_pdc_cell.mod[3 + i * 3], /* start */
+ io_pdc_cell.mod[4 + i * 3]); /* finish (ie end) */
+ }
+ printk(KERN_DEBUG "\n");
+ break;
+ }
+#endif /* DEBUG_PAT */
+ return PDC_OK;
+}
+
+
+/* pat pdc can return information about a variety of different
+ * types of memory (e.g. firmware,i/o, etc) but we only care about
+ * the usable physical ram right now. Since the firmware specific
+ * information is allocated on the stack, we'll be generous, in
+ * case there is a lot of other information we don't care about.
+ */
+
+#define PAT_MAX_RANGES (4 * MAX_PHYSMEM_RANGES)
+
+static void __init pat_memconfig(void)
+{
+ unsigned long actual_len;
+ struct pdc_pat_pd_addr_map_entry mem_table[PAT_MAX_RANGES+1];
+ struct pdc_pat_pd_addr_map_entry *mtbl_ptr;
+ physmem_range_t *pmem_ptr;
+ long status;
+ int entries;
+ unsigned long length;
+ int i;
+
+ length = (PAT_MAX_RANGES + 1) * sizeof(struct pdc_pat_pd_addr_map_entry);
+
+ status = pdc_pat_pd_get_addr_map(&actual_len, mem_table, length, 0L);
+
+ if ((status != PDC_OK)
+ || ((actual_len % sizeof(struct pdc_pat_pd_addr_map_entry)) != 0)) {
+
+ /* The above pdc call shouldn't fail, but, just in
+ * case, just use the PAGE0 info.
+ */
+
+ printk("\n\n\n");
+ printk(KERN_WARNING "WARNING! Could not get full memory configuration. "
+ "All memory may not be used!\n\n\n");
+ pagezero_memconfig();
+ return;
+ }
+
+ entries = actual_len / sizeof(struct pdc_pat_pd_addr_map_entry);
+
+ if (entries > PAT_MAX_RANGES) {
+ printk(KERN_WARNING "This Machine has more memory ranges than we support!\n");
+ printk(KERN_WARNING "Some memory may not be used!\n");
+ }
+
+ /* Copy information into the firmware independent pmem_ranges
+ * array, skipping types we don't care about. Notice we said
+ * "may" above. We'll use all the entries that were returned.
+ */
+
+ npmem_ranges = 0;
+ mtbl_ptr = mem_table;
+ pmem_ptr = pmem_ranges; /* Global firmware independent table */
+ for (i = 0; i < entries; i++,mtbl_ptr++) {
+ if ( (mtbl_ptr->entry_type != PAT_MEMORY_DESCRIPTOR)
+ || (mtbl_ptr->memory_type != PAT_MEMTYPE_MEMORY)
+ || (mtbl_ptr->pages == 0)
+ || ( (mtbl_ptr->memory_usage != PAT_MEMUSE_GENERAL)
+ && (mtbl_ptr->memory_usage != PAT_MEMUSE_GI)
+ && (mtbl_ptr->memory_usage != PAT_MEMUSE_GNI) ) ) {
+
+ continue;
+ }
+
+ if (npmem_ranges == MAX_PHYSMEM_RANGES) {
+ printk(KERN_WARNING "This Machine has more memory ranges than we support!\n");
+ printk(KERN_WARNING "Some memory will not be used!\n");
+ break;
+ }
+
+ set_pmem_entry(pmem_ptr++,mtbl_ptr->paddr,mtbl_ptr->pages);
+ npmem_ranges++;
+ }
+}
+
+static int __init pat_inventory(void)
+{
+ int status;
+ ulong mod_index = 0;
+ struct pdc_pat_cell_num cell_info;
+
+ /*
+ ** Note: Prelude (and it's successors: Lclass, A400/500) only
+ ** implement PDC_PAT_CELL sub-options 0 and 2.
+ */
+ status = pdc_pat_cell_get_number(&cell_info);
+ if (status != PDC_OK) {
+ return 0;
+ }
+
+#ifdef DEBUG_PAT
+ printk(KERN_DEBUG "CELL_GET_NUMBER: 0x%lx 0x%lx\n", cell_info.cell_num,
+ cell_info.cell_loc);
+#endif
+
+ while (PDC_OK == pat_query_module(cell_info.cell_loc, mod_index)) {
+ mod_index++;
+ }
+
+ return mod_index;
+}
+
+/* We only look for extended memory ranges on a 64 bit capable box */
+static void __init sprockets_memconfig(void)
+{
+ struct pdc_memory_table_raddr r_addr;
+ struct pdc_memory_table mem_table[MAX_PHYSMEM_RANGES];
+ struct pdc_memory_table *mtbl_ptr;
+ physmem_range_t *pmem_ptr;
+ long status;
+ int entries;
+ int i;
+
+ status = pdc_mem_mem_table(&r_addr,mem_table,
+ (unsigned long)MAX_PHYSMEM_RANGES);
+
+ if (status != PDC_OK) {
+
+ /* The above pdc call only works on boxes with sprockets
+ * firmware (newer B,C,J class). Other non PAT PDC machines
+ * do support more than 3.75 Gb of memory, but we don't
+ * support them yet.
+ */
+
+ pagezero_memconfig();
+ return;
+ }
+
+ if (r_addr.entries_total > MAX_PHYSMEM_RANGES) {
+ printk(KERN_WARNING "This Machine has more memory ranges than we support!\n");
+ printk(KERN_WARNING "Some memory will not be used!\n");
+ }
+
+ entries = (int)r_addr.entries_returned;
+
+ npmem_ranges = 0;
+ mtbl_ptr = mem_table;
+ pmem_ptr = pmem_ranges; /* Global firmware independent table */
+ for (i = 0; i < entries; i++,mtbl_ptr++) {
+ set_pmem_entry(pmem_ptr++,mtbl_ptr->paddr,mtbl_ptr->pages);
+ npmem_ranges++;
+ }
+}
+
+#else /* !__LP64__ */
+
+#define pat_inventory() do { } while (0)
+#define pat_memconfig() do { } while (0)
+#define sprockets_memconfig() pagezero_memconfig()
+
+#endif /* !__LP64__ */
+
+
+#ifndef CONFIG_PA20
+
+/* Code to support Snake machines (7[2350], 7[235]5, 715/Scorpio) */
+
+static struct parisc_device * __init
+legacy_create_device(struct pdc_memory_map *r_addr,
+ struct pdc_module_path *module_path)
+{
+ struct parisc_device *dev;
+ int status = pdc_mem_map_hpa(r_addr, module_path);
+ if (status != PDC_OK)
+ return NULL;
+
+ dev = alloc_pa_dev(r_addr->hpa, &module_path->path);
+ if (dev == NULL)
+ return NULL;
+
+ register_parisc_device(dev);
+ return dev;
+}
+
+/**
+ * snake_inventory
+ *
+ * Before PDC_SYSTEM_MAP was invented, the PDC_MEM_MAP call was used.
+ * To use it, we initialise the mod_path.bc to 0xff and try all values of
+ * mod to get the HPA for the top-level devices. Bus adapters may have
+ * sub-devices which are discovered by setting bc[5] to 0 and bc[4] to the
+ * module, then trying all possible functions.
+ */
+static void __init snake_inventory(void)
+{
+ int mod;
+ for (mod = 0; mod < 16; mod++) {
+ struct parisc_device *dev;
+ struct pdc_module_path module_path;
+ struct pdc_memory_map r_addr;
+ unsigned int func;
+
+ memset(module_path.path.bc, 0xff, 6);
+ module_path.path.mod = mod;
+ dev = legacy_create_device(&r_addr, &module_path);
+ if ((!dev) || (dev->id.hw_type != HPHW_BA))
+ continue;
+
+ memset(module_path.path.bc, 0xff, 4);
+ module_path.path.bc[4] = mod;
+
+ for (func = 0; func < 16; func++) {
+ module_path.path.bc[5] = 0;
+ module_path.path.mod = func;
+ legacy_create_device(&r_addr, &module_path);
+ }
+ }
+}
+
+#else /* CONFIG_PA20 */
+#define snake_inventory() do { } while (0)
+#endif /* CONFIG_PA20 */
+
+/* Common 32/64 bit based code goes here */
+
+/**
+ * add_system_map_addresses - Add additional addresses to the parisc device.
+ * @dev: The parisc device.
+ * @num_addrs: Then number of addresses to add;
+ * @module_instance: The system_map module instance.
+ *
+ * This function adds any additional addresses reported by the system_map
+ * firmware to the parisc device.
+ */
+static void __init
+add_system_map_addresses(struct parisc_device *dev, int num_addrs,
+ int module_instance)
+{
+ int i;
+ long status;
+ struct pdc_system_map_addr_info addr_result;
+
+ dev->addr = kmalloc(num_addrs * sizeof(unsigned long), GFP_KERNEL);
+ if(!dev->addr) {
+ printk(KERN_ERR "%s %s(): memory allocation failure\n",
+ __FILE__, __FUNCTION__);
+ return;
+ }
+
+ for(i = 1; i <= num_addrs; ++i) {
+ status = pdc_system_map_find_addrs(&addr_result,
+ module_instance, i);
+ if(PDC_OK == status) {
+ dev->addr[dev->num_addrs] = (unsigned long)addr_result.mod_addr;
+ dev->num_addrs++;
+ } else {
+ printk(KERN_WARNING
+ "Bad PDC_FIND_ADDRESS status return (%ld) for index %d\n",
+ status, i);
+ }
+ }
+}
+
+/**
+ * system_map_inventory - Retrieve firmware devices via SYSTEM_MAP.
+ *
+ * This function attempts to retrieve and register all the devices firmware
+ * knows about via the SYSTEM_MAP PDC call.
+ */
+static void __init system_map_inventory(void)
+{
+ int i;
+ long status = PDC_OK;
+
+ for (i = 0; i < 256; i++) {
+ struct parisc_device *dev;
+ struct pdc_system_map_mod_info module_result;
+ struct pdc_module_path module_path;
+
+ status = pdc_system_map_find_mods(&module_result,
+ &module_path, i);
+ if ((status == PDC_BAD_PROC) || (status == PDC_NE_MOD))
+ break;
+ if (status != PDC_OK)
+ continue;
+
+ dev = alloc_pa_dev(module_result.mod_addr, &module_path.path);
+ if (!dev)
+ continue;
+
+ register_parisc_device(dev);
+
+ /* if available, get the additional addresses for a module */
+ if (!module_result.add_addrs)
+ continue;
+
+ add_system_map_addresses(dev, module_result.add_addrs, i);
+ }
+
+ walk_central_bus();
+ return;
+}
+
+void __init do_memory_inventory(void)
+{
+ switch (pdc_type) {
+
+ case PDC_TYPE_PAT:
+ pat_memconfig();
+ break;
+
+ case PDC_TYPE_SYSTEM_MAP:
+ sprockets_memconfig();
+ break;
+
+ case PDC_TYPE_SNAKE:
+ pagezero_memconfig();
+ return;
+
+ default:
+ panic("Unknown PDC type!\n");
+ }
+
+ if (npmem_ranges == 0 || pmem_ranges[0].start_pfn != 0) {
+ printk(KERN_WARNING "Bad memory configuration returned!\n");
+ printk(KERN_WARNING "Some memory may not be used!\n");
+ pagezero_memconfig();
+ }
+}
+
+void __init do_device_inventory(void)
+{
+ printk(KERN_INFO "Searching for devices...\n");
+
+ init_parisc_bus();
+
+ switch (pdc_type) {
+
+ case PDC_TYPE_PAT:
+ pat_inventory();
+ break;
+
+ case PDC_TYPE_SYSTEM_MAP:
+ system_map_inventory();
+ break;
+
+ case PDC_TYPE_SNAKE:
+ snake_inventory();
+ break;
+
+ default:
+ panic("Unknown PDC type!\n");
+ }
+ printk(KERN_INFO "Found devices:\n");
+ print_parisc_devices();
+}
OpenPOWER on IntegriCloud