path: root/fs/exofs/ore.c

/*
 * Copyright (C) 2005, 2006
 * Avishay Traeger (avishay@gmail.com)
 * Copyright (C) 2008, 2009
 * Boaz Harrosh <bharrosh@panasas.com>
 *
 * This file is part of exofs.
 *
 * exofs 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.  Since it is based on ext2, and the only
 * valid version of GPL for the Linux kernel is version 2, the only valid
 * version of GPL for exofs is version 2.
 *
 * exofs is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with exofs; if not, write to the Free Software
 * Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
 */

#include <linux/slab.h>
#include <linux/module.h>
#include <asm/div64.h>
#include <linux/lcm.h>

#include "ore_raid.h"

MODULE_AUTHOR("Boaz Harrosh <bharrosh@panasas.com>");
MODULE_DESCRIPTION("Objects Raid Engine ore.ko");
MODULE_LICENSE("GPL");

/* ore_verify_layout does a couple of things:
 * 1. Given a minimum number of needed parameters fixes up the rest of the
 *    members to be operatonals for the ore. The needed parameters are those
 *    that are defined by the pnfs-objects layout STD.
 * 2. Check to see if the current ore code actually supports these parameters
 *    for example stripe_unit must be a multple of the system PAGE_SIZE,
 *    and etc...
 * 3. Cache some havily used calculations that will be needed by users.
 */

enum { BIO_MAX_PAGES_KMALLOC =
		(PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec),};

int ore_verify_layout(unsigned total_comps, struct ore_layout *layout)
{
	u64 stripe_length;

	switch (layout->raid_algorithm) {
	case PNFS_OSD_RAID_0:
		layout->parity = 0;
		break;
	case PNFS_OSD_RAID_5:
		layout->parity = 1;
		break;
	case PNFS_OSD_RAID_PQ:
		layout->parity = 2;
		break;
	case PNFS_OSD_RAID_4:
	default:
		ORE_ERR("Only RAID_0/5/6 for now received-enum=%d\n",
			layout->raid_algorithm);
		return -EINVAL;
	}
	if (0 != (layout->stripe_unit & ~PAGE_MASK)) {
		ORE_ERR("Stripe Unit(0x%llx)"
			  " must be Multples of PAGE_SIZE(0x%lx)\n",
			  _LLU(layout->stripe_unit), PAGE_SIZE);
		return -EINVAL;
	}
	if (layout->group_width) {
		if (!layout->group_depth) {
			ORE_ERR("group_depth == 0 && group_width != 0\n");
			return -EINVAL;
		}
		if (total_comps < (layout->group_width * layout->mirrors_p1)) {
			ORE_ERR("Data Map wrong, "
				"numdevs=%d < group_width=%d * mirrors=%d\n",
				total_comps, layout->group_width,
				layout->mirrors_p1);
			return -EINVAL;
		}
		layout->group_count = total_comps / layout->mirrors_p1 /
						layout->group_width;
	} else {
		if (layout->group_depth) {
			printk(KERN_NOTICE "Warning: group_depth ignored "
				"group_width == 0 && group_depth == %lld\n",
				_LLU(layout->group_depth));
		}
		layout->group_width = total_comps / layout->mirrors_p1;
		layout->group_depth = -1;
		layout->group_count = 1;
	}

	stripe_length = (u64)layout->group_width * layout->stripe_unit;
	if (stripe_length >= (1ULL << 32)) {
		ORE_ERR("Stripe_length(0x%llx) >= 32bit is not supported\n",
			_LLU(stripe_length));
		return -EINVAL;
	}

	layout->max_io_length =
		(BIO_MAX_PAGES_KMALLOC * PAGE_SIZE - layout->stripe_unit) *
					(layout->group_width - layout->parity);
	if (layout->parity) {
		unsigned stripe_length =
				(layout->group_width - layout->parity) *
				layout->stripe_unit;

		layout->max_io_length /= stripe_length;
		layout->max_io_length *= stripe_length;