lloop.c

lustre 1.6.5 source code

/* -*- mode: c; c-basic-offset: 8; indent-tabs-mode: nil; -*-
 * vim:expandtab:shiftwidth=8:tabstop=8:
 *
 * Lustre virtual block device emulator.
 *
 *  Copyright (c) 2001-2003 Cluster File Systems, Inc.
 *
 *   This file is part of Lustre, http://www.lustre.org.
 *
 *   Lustre is free software; you can redistribute it and/or
 *   modify it under the terms of version 2 of the GNU General Public
 *   License as published by the Free Software Foundation.
 *
 *   Lustre 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 Lustre; if not, write to the Free Software
 *   Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA.
 */

/*
 *  linux/drivers/block/loop.c
 *
 *  Written by Theodore Ts'o, 3/29/93
 *
 * Copyright 1993 by Theodore Ts'o.  Redistribution of this file is
 * permitted under the GNU General Public License.
 *
 * DES encryption plus some minor changes by Werner Almesberger, 30-MAY-1993
 * more DES encryption plus IDEA encryption by Nicholas J. Leon, June 20, 1996
 *
 * Modularized and updated for 1.1.16 kernel - Mitch Dsouza 28th May 1994
 * Adapted for 1.3.59 kernel - Andries Brouwer, 1 Feb 1996
 *
 * Fixed do_loop_request() re-entrancy - [e-mail] Mar 20, 1997
 *
 * Added devfs support - Richard Gooch <[e-mail]> 16-Jan-1998
 *
 * Handle sparse backing files correctly - Kenn Humborg, Jun 28, 1998
 *
 * Loadable modules and other fixes by AK, 1998
 *
 * Make real block number available to downstream transfer functions, enables
 * CBC (and relatives) mode encryption requiring unique IVs per data block.
 * Reed H. Petty, [e-mail]
 *
 * Maximum number of loop devices now dynamic via max_loop module parameter.
 * Russell Kroll <[e-mail]> 19990701
 *
 * Maximum number of loop devices when compiled-in now selectable by passing
 * max_loop=<1-255> to the kernel on boot.
 * Erik I. Bols?, <[e-mail]>, Oct 31, 1999
 *
 * Completely rewrite request handling to be make_request_fn style and
 * non blocking, pushing work to a helper thread. Lots of fixes from
 * Al Viro too.
 * Jens Axboe <[e-mail]>, Nov 2000
 *
 * Support up to 256 loop devices
 * Heinz Mauelshagen <[e-mail]>, Feb 2002
 *
 * Support for falling back on the write file operation when the address space
 * operations prepare_write and/or commit_write are not available on the
 * backing filesystem.
 * Anton Altaparmakov, 16 Feb 2005
 *
 * Still To Fix:
 * - Advisory locking is ignored here.
 * - Should use an own CAP_* category instead of CAP_SYS_ADMIN
 *
 */

#ifndef AUTOCONF_INCLUDED
#include <linux/config.h>
#endif
#include <linux/module.h>

#include <linux/sched.h>
#include <linux/fs.h>
#include <linux/file.h>
#include <linux/stat.h>
#include <linux/errno.h>
#include <linux/major.h>
#include <linux/wait.h>
#include <linux/blkdev.h>
#include <linux/blkpg.h>
#include <linux/init.h>
#include <linux/smp_lock.h>
#include <linux/swap.h>
#include <linux/slab.h>
#include <linux/suspend.h>
#include <linux/writeback.h>
#include <linux/buffer_head.h>                /* for invalidate_bdev() */
#include <linux/completion.h>
#include <linux/highmem.h>
#include <linux/gfp.h>
#include <linux/swap.h>
#include <linux/pagevec.h>

#include <asm/uaccess.h>

#include <lustre_lib.h>
#include <lustre_lite.h>
#include "llite_internal.h"

#define LLOOP_MAX_SEGMENTS        PTLRPC_MAX_BRW_PAGES

/* Possible states of device */
enum {
        LLOOP_UNBOUND,
        LLOOP_BOUND,
        LLOOP_RUNDOWN,
};

struct lloop_device {
        int                lo_number;
        int                lo_refcnt;
        loff_t             lo_offset;
        loff_t             lo_sizelimit;
        int                lo_flags;
        int                (*ioctl)(struct lloop_device *, int cmd, 
                                 unsigned long arg); 

        struct file *      lo_backing_file;
        struct block_device *lo_device;
        unsigned           lo_blocksize;

        int                old_gfp_mask;

        spinlock_t         lo_lock;
        struct bio         *lo_bio;
        struct bio         *lo_biotail;
        int                lo_state;
        struct semaphore   lo_sem;
        struct semaphore   lo_ctl_mutex;
        struct semaphore   lo_bh_mutex;
        atomic_t           lo_pending;

        request_queue_t    *lo_queue;

        /* data to handle bio for lustre. */
        struct lo_request_data {
                struct brw_page    lrd_pages[LLOOP_MAX_SEGMENTS];
                struct obdo        lrd_oa;
        } lo_requests[1];

};

/*
 * Loop flags
 */
enum {
        LO_FLAGS_READ_ONLY       = 1,
};

static int lloop_major;
static int max_loop = 8;
static struct lloop_device *loop_dev;
static struct gendisk **disks;
static struct semaphore lloop_mutex;
static void *ll_iocontrol_magic = NULL;

static loff_t get_loop_size(struct lloop_device *lo, struct file *file)
{
        loff_t size, offset, loopsize;

        /* Compute loopsize in bytes */
        size = i_size_read(file->f_mapping->host);
        offset = lo->lo_offset;
        loopsize = size - offset;
        if (lo->lo_sizelimit > 0 && lo->lo_sizelimit < loopsize)
                loopsize = lo->lo_sizelimit;

        /*
         * Unfortunately, if we want to do I/O on the device,
         * the number of 512-byte sectors has to fit into a sector_t.
         */
        return loopsize >> 9;
}

static int do_bio_filebacked(struct lloop_device *lo, struct bio *bio)
{
        struct inode *inode = lo->lo_backing_file->f_dentry->d_inode;
        struct ll_inode_info *lli = ll_i2info(inode);
        struct lov_stripe_md *lsm = lli->lli_smd;
        struct obd_info oinfo = {{{0}}};
        struct brw_page *pg = lo->lo_requests[0].lrd_pages;
        struct obdo *oa = &lo->lo_requests[0].lrd_oa;
        pgoff_t offset;
        int ret, cmd, i;
        struct bio_vec *bvec;

        BUG_ON(bio->bi_hw_segments > LLOOP_MAX_SEGMENTS);

        offset = (pgoff_t)(bio->bi_sector << 9) + lo->lo_offset;
        bio_for_each_segment(bvec, bio, i) {
                BUG_ON(bvec->bv_offset != 0);
                BUG_ON(bvec->bv_len != CFS_PAGE_SIZE);

                pg->pg = bvec->bv_page;
                pg->off = offset;
                pg->count = bvec->bv_len;
                pg->flag = OBD_BRW_SRVLOCK;

                pg++;
                offset += bvec->bv_len;
        }

        oa->o_mode = inode->i_mode;
        oa->o_id = lsm->lsm_object_id;
        oa->o_valid = OBD_MD_FLID | OBD_MD_FLMODE | OBD_MD_FLTYPE;
        obdo_from_inode(oa, inode, OBD_MD_FLFID | OBD_MD_FLGENER);

        cmd = OBD_BRW_READ;
        if (bio_rw(bio) == WRITE)
                cmd = OBD_BRW_WRITE;

        if (cmd == OBD_BRW_WRITE)
                ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_BRW_WRITE, bio->bi_size);
        else
                ll_stats_ops_tally(ll_i2sbi(inode), LPROC_LL_BRW_READ, bio->bi_size);
        oinfo.oi_oa = oa;
        oinfo.oi_md = lsm;
        ret = obd_brw(cmd, ll_i2obdexp(inode), &oinfo, 
                      (obd_count)(i - bio->bi_idx), 
                      lo->lo_requests[0].lrd_pages, NULL);
        if (ret == 0)
                obdo_to_inode(inode, oa, OBD_MD_FLBLOCKS);
        return ret;
}


/*
 * Add bio to back of pending list
 */
static void loop_add_bio(struct lloop_device *lo, struct bio *bio)
{
        unsigned long flags;

        spin_lock_irqsave(&lo->lo_lock, flags);
        if (lo->lo_biotail) {
                lo->lo_biotail->bi_next = bio;
                lo->lo_biotail = bio;
        } else
                lo->lo_bio = lo->lo_biotail = bio;
        spin_unlock_irqrestore(&lo->lo_lock, flags);

        up(&lo->lo_bh_mutex);
}

/*
 * Grab first pending buffer
 */
static struct bio *loop_get_bio(struct lloop_device *lo)
{
        struct bio *bio;

        spin_lock_irq(&lo->lo_lock);
        if ((bio = lo->lo_bio)) {
                if (bio == lo->lo_biotail)
                        lo->lo_biotail = NULL;
                lo->lo_bio = bio->bi_next;
                bio->bi_next = NULL;
        }
        spin_unlock_irq(&lo->lo_lock);

        return bio;
}

static int loop_make_request(request_queue_t *q, struct bio *old_bio)
{
        struct lloop_device *lo = q->queuedata;
        int rw = bio_rw(old_bio);

        if (!lo)
                goto out;

        spin_lock_irq(&lo->lo_lock);
        if (lo->lo_state != LLOOP_BOUND)
                goto inactive;
        atomic_inc(&lo->lo_pending);
        spin_unlock_irq(&lo->lo_lock);

        if (rw == WRITE) {
                if (lo->lo_flags & LO_FLAGS_READ_ONLY)
                        goto err;
        } else if (rw == READA) {
                rw = READ;
        } else if (rw != READ) {
                CERROR("lloop: unknown command (%x)\n", rw);
                goto err;
        }
        loop_add_bio(lo, old_bio);
        return 0;
err:
        if (atomic_dec_and_test(&lo->lo_pending))
                up(&lo->lo_bh_mutex);
out:
        bio_io_error(old_bio, old_bio->bi_size);
        return 0;
inactive:
        spin_unlock_irq(&lo->lo_lock);
        goto out;
}

/*
 * kick off io on the underlying address space
 */
static void loop_unplug(request_queue_t *q)
{
        struct lloop_device *lo = q->queuedata;

        clear_bit(QUEUE_FLAG_PLUGGED, &q->queue_flags);
        blk_run_address_space(lo->lo_backing_file->f_mapping);
}

static inline void loop_handle_bio(struct lloop_device *lo, struct bio *bio)
{
        int ret;
        ret = do_bio_filebacked(lo, bio);
        bio_endio(bio, bio->bi_size, ret);
}

/*
 * worker thread that handles reads/writes to file backed loop devices,
 * to avoid blocking in our make_request_fn. it also does loop decrypting
 * on reads for block backed loop, as that is too heavy to do from
 * b_end_io context where irqs may be disabled.
 */
static int loop_thread(void *data)
{
        struct lloop_device *lo = data;
        struct bio *bio;

        daemonize("lloop%d", lo->lo_number);

        set_user_nice(current, -20);

        lo->lo_state = LLOOP_BOUND;
        atomic_inc(&lo->lo_pending);

        /*
         * up sem, we are running
         */
        up(&lo->lo_sem);

        for (;;) {
                down_interruptible(&lo->lo_bh_mutex);
                /*
                 * could be upped because of tear-down, not because of
                 * pending work
                 */
                if (!atomic_read(&lo->lo_pending))
                        break;

                bio = loop_get_bio(lo);
                if (!bio) {
                        CWARN("lloop(minor: %d): missing bio\n", lo->lo_number);
                        continue;
                }
                loop_handle_bio(lo, bio);

                /*
                 * upped both for pending work and tear-down, lo_pending
                 * will hit zero then
                 */
                if (atomic_dec_and_test(&lo->lo_pending))
                        break;
        }

        up(&lo->lo_sem);
        return 0;
}

static int loop_set_fd(struct lloop_device *lo, struct file *unused,
                       struct block_device *bdev, struct file *file)
{
        struct inode        *inode;
        struct address_space *mapping;
        int                lo_flags = 0;