/**
* mft.c - NTFS kernel mft record operations. Part of the Linux-NTFS project.
*
- * Copyright (c) 2001-2005 Anton Altaparmakov
+ * Copyright (c) 2001-2006 Anton Altaparmakov
* Copyright (c) 2002 Richard Russon
*
* This program/include file is free software; you can redistribute it and/or
ntfs_volume *vol = ni->vol;
struct inode *mft_vi = vol->mft_ino;
struct page *page;
- unsigned long index, ofs, end_index;
+ unsigned long index, end_index;
+ unsigned ofs;
BUG_ON(ni->page);
/*
* overflowing the unsigned long, but I don't think we would ever get
* here if the volume was that big...
*/
- index = ni->mft_no << vol->mft_record_size_bits >> PAGE_CACHE_SHIFT;
+ index = (u64)ni->mft_no << vol->mft_record_size_bits >>
+ PAGE_CACHE_SHIFT;
ofs = (ni->mft_no << vol->mft_record_size_bits) & ~PAGE_CACHE_MASK;
i_size = i_size_read(mft_vi);
"Run chkdsk.", ni->mft_no);
ntfs_unmap_page(page);
page = ERR_PTR(-EIO);
+ NVolSetErrors(vol);
}
err_out:
ni->page = NULL;
* map_mft_record - map, pin and lock an mft record
* @ni: ntfs inode whose MFT record to map
*
- * First, take the mrec_lock semaphore. We might now be sleeping, while waiting
- * for the semaphore if it was already locked by someone else.
+ * First, take the mrec_lock mutex. We might now be sleeping, while waiting
+ * for the mutex if it was already locked by someone else.
*
* The page of the record is mapped using map_mft_record_page() before being
* returned to the caller.
* So that code will end up having to own the mrec_lock of all mft
* records/inodes present in the page before I/O can proceed. In that case we
* wouldn't need to bother with PG_locked and PG_uptodate as nobody will be
- * accessing anything without owning the mrec_lock semaphore. But we do need
- * to use them because of the read_cache_page() invocation and the code becomes
- * so much simpler this way that it is well worth it.
+ * accessing anything without owning the mrec_lock mutex. But we do need to
+ * use them because of the read_cache_page() invocation and the code becomes so
+ * much simpler this way that it is well worth it.
*
* The mft record is now ours and we return a pointer to it. You need to check
* the returned pointer with IS_ERR() and if that is true, PTR_ERR() will return
atomic_inc(&ni->count);
/* Serialize access to this mft record. */
- down(&ni->mrec_lock);
+ mutex_lock(&ni->mrec_lock);
m = map_mft_record_page(ni);
if (likely(!IS_ERR(m)))
return m;
- up(&ni->mrec_lock);
+ mutex_unlock(&ni->mrec_lock);
atomic_dec(&ni->count);
ntfs_error(ni->vol->sb, "Failed with error code %lu.", -PTR_ERR(m));
return m;
ntfs_debug("Entering for mft_no 0x%lx.", ni->mft_no);
unmap_mft_record_page(ni);
- up(&ni->mrec_lock);
+ mutex_unlock(&ni->mrec_lock);
atomic_dec(&ni->count);
/*
* If pure ntfs_inode, i.e. no vfs inode attached, we leave it to
* in which case just return it. If not found, add it to the base
* inode before returning it.
*/
- down(&base_ni->extent_lock);
+ mutex_lock(&base_ni->extent_lock);
if (base_ni->nr_extents > 0) {
extent_nis = base_ni->ext.extent_ntfs_inos;
for (i = 0; i < base_ni->nr_extents; i++) {
}
}
if (likely(ni != NULL)) {
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
atomic_dec(&base_ni->count);
/* We found the record; just have to map and return it. */
m = map_mft_record(ni);
/* Record wasn't there. Get a new ntfs inode and initialize it. */
ni = ntfs_new_extent_inode(base_ni->vol->sb, mft_no);
if (unlikely(!ni)) {
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
atomic_dec(&base_ni->count);
return ERR_PTR(-ENOMEM);
}
/* Now map the record. */
m = map_mft_record(ni);
if (IS_ERR(m)) {
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
atomic_dec(&base_ni->count);
ntfs_clear_extent_inode(ni);
goto map_err_out;
ntfs_inode **tmp;
int new_size = (base_ni->nr_extents + 4) * sizeof(ntfs_inode *);
- tmp = (ntfs_inode **)kmalloc(new_size, GFP_NOFS);
+ tmp = kmalloc(new_size, GFP_NOFS);
if (unlikely(!tmp)) {
ntfs_error(base_ni->vol->sb, "Failed to allocate "
"internal buffer.");
base_ni->ext.extent_ntfs_inos = tmp;
}
base_ni->ext.extent_ntfs_inos[base_ni->nr_extents++] = ni;
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
atomic_dec(&base_ni->count);
ntfs_debug("Done 2.");
*ntfs_ino = ni;
return m;
unm_err_out:
unmap_mft_record(ni);
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
atomic_dec(&base_ni->count);
/*
* If the extent inode was not attached to the base inode we need to
BUG_ON(NInoAttr(ni));
mark_ntfs_record_dirty(ni->page, ni->page_ofs);
/* Determine the base vfs inode and mark it dirty, too. */
- down(&ni->extent_lock);
+ mutex_lock(&ni->extent_lock);
if (likely(ni->nr_extents >= 0))
base_ni = ni;
else
base_ni = ni->ext.base_ntfs_ino;
- up(&ni->extent_lock);
+ mutex_unlock(&ni->extent_lock);
__mark_inode_dirty(VFS_I(base_ni), I_DIRTY_SYNC | I_DIRTY_DATASYNC);
}
runlist_element *rl;
unsigned int block_start, block_end, m_start, m_end, page_ofs;
int i_bhs, nr_bhs, err = 0;
- unsigned char blocksize_bits = vol->mftmirr_ino->i_blkbits;
+ unsigned char blocksize_bits = vol->sb->s_blocksize_bits;
ntfs_debug("Entering for inode 0x%lx.", mft_no);
BUG_ON(!max_bhs);
} while (bh);
tail->b_this_page = head;
attach_page_buffers(page, head);
- BUG_ON(!page_has_buffers(page));
}
bh = head = page_buffers(page);
BUG_ON(!bh);
LCN lcn;
unsigned int vcn_ofs;
+ bh->b_bdev = vol->sb->s_bdev;
/* Obtain the vcn and offset of the current block. */
vcn = ((VCN)mft_no << vol->mft_record_size_bits) +
(block_start - m_start);
* fs/ntfs/aops.c::mark_ntfs_record_dirty().
*
* On success, clean the mft record and return 0. On error, leave the mft
- * record dirty and return -errno. The caller should call make_bad_inode() on
- * the base inode to ensure no more access happens to this inode. We do not do
- * it here as the caller may want to finish writing other extent mft records
- * first to minimize on-disk metadata inconsistencies.
+ * record dirty and return -errno.
*
* NOTE: We always perform synchronous i/o and ignore the @sync parameter.
* However, if the mft record has a counterpart in the mft mirror and @sync is
{
ntfs_volume *vol = ni->vol;
struct page *page = ni->page;
- unsigned char blocksize_bits = vol->mft_ino->i_blkbits;
- unsigned int blocksize = 1 << blocksize_bits;
+ unsigned int blocksize = vol->sb->s_blocksize;
+ unsigned char blocksize_bits = vol->sb->s_blocksize_bits;
int max_bhs = vol->mft_record_size / blocksize;
struct buffer_head *bhs[max_bhs];
struct buffer_head *bh, *head;
*/
if (!NInoTestClearDirty(ni))
goto done;
- BUG_ON(!page_has_buffers(page));
bh = head = page_buffers(page);
BUG_ON(!bh);
rl = NULL;
LCN lcn;
unsigned int vcn_ofs;
+ bh->b_bdev = vol->sb->s_bdev;
/* Obtain the vcn and offset of the current block. */
vcn = ((VCN)ni->mft_no << vol->mft_record_size_bits) +
(block_start - m_start);
na.name_len = 0;
na.type = AT_UNUSED;
/*
- * For inode 0, i.e. $MFT itself, we cannot use ilookup5() from here or
- * we deadlock because the inode is already locked by the kernel
- * (fs/fs-writeback.c::__sync_single_inode()) and ilookup5() waits
- * until the inode is unlocked before returning it and it never gets
- * unlocked because ntfs_should_write_mft_record() never returns. )-:
- * Fortunately, we have inode 0 pinned in icache for the duration of
- * the mount so we can access it directly.
+ * Optimize inode 0, i.e. $MFT itself, since we have it in memory and
+ * we get here for it rather often.
*/
if (!mft_no) {
/* Balance the below iput(). */
vi = igrab(mft_vi);
BUG_ON(vi != mft_vi);
- } else
- vi = ilookup5(sb, mft_no, (test_t)ntfs_test_inode, &na);
+ } else {
+ /*
+ * Have to use ilookup5_nowait() since ilookup5() waits for the
+ * inode lock which causes ntfs to deadlock when a concurrent
+ * inode write via the inode dirty code paths and the page
+ * dirty code path of the inode dirty code path when writing
+ * $MFT occurs.
+ */
+ vi = ilookup5_nowait(sb, mft_no, (test_t)ntfs_test_inode, &na);
+ }
if (vi) {
ntfs_debug("Base inode 0x%lx is in icache.", mft_no);
/* The inode is in icache. */
}
ntfs_debug("Inode 0x%lx is not dirty.", mft_no);
/* The inode is not dirty, try to take the mft record lock. */
- if (unlikely(down_trylock(&ni->mrec_lock))) {
+ if (unlikely(!mutex_trylock(&ni->mrec_lock))) {
ntfs_debug("Mft record 0x%lx is already locked, do "
"not write it.", mft_no);
atomic_dec(&ni->count);
na.mft_no = MREF_LE(m->base_mft_record);
ntfs_debug("Mft record 0x%lx is an extent record. Looking for base "
"inode 0x%lx in icache.", mft_no, na.mft_no);
- vi = ilookup5(sb, na.mft_no, (test_t)ntfs_test_inode, &na);
+ if (!na.mft_no) {
+ /* Balance the below iput(). */
+ vi = igrab(mft_vi);
+ BUG_ON(vi != mft_vi);
+ } else
+ vi = ilookup5_nowait(sb, na.mft_no, (test_t)ntfs_test_inode,
+ &na);
if (!vi) {
/*
* The base inode is not in icache, write this extent mft
* corresponding to this extent mft record attached.
*/
ni = NTFS_I(vi);
- down(&ni->extent_lock);
+ mutex_lock(&ni->extent_lock);
if (ni->nr_extents <= 0) {
/*
* The base inode has no attached extent inodes, write this
* extent mft record.
*/
- up(&ni->extent_lock);
+ mutex_unlock(&ni->extent_lock);
iput(vi);
ntfs_debug("Base inode 0x%lx has no attached extent inodes, "
"write the extent record.", na.mft_no);
* extent mft record.
*/
if (!eni) {
- up(&ni->extent_lock);
+ mutex_unlock(&ni->extent_lock);
iput(vi);
ntfs_debug("Extent inode 0x%lx is not attached to its base "
"inode 0x%lx, write the extent record.",
mft_no, na.mft_no);
/* Take a reference to the extent ntfs inode. */
atomic_inc(&eni->count);
- up(&ni->extent_lock);
+ mutex_unlock(&ni->extent_lock);
/*
* Found the extent inode coresponding to this extent mft record.
* Try to take the mft record lock.
*/
- if (unlikely(down_trylock(&eni->mrec_lock))) {
+ if (unlikely(!mutex_trylock(&eni->mrec_lock))) {
atomic_dec(&eni->count);
iput(vi);
ntfs_debug("Extent mft record 0x%lx is already locked, do "
read_lock_irqsave(&mftbmp_ni->size_lock, flags);
ll = mftbmp_ni->allocated_size;
read_unlock_irqrestore(&mftbmp_ni->size_lock, flags);
- rl = ntfs_find_vcn_nolock(mftbmp_ni,
- (ll - 1) >> vol->cluster_size_bits, TRUE);
+ rl = ntfs_attr_find_vcn_nolock(mftbmp_ni,
+ (ll - 1) >> vol->cluster_size_bits, NULL);
if (unlikely(IS_ERR(rl) || !rl->length || rl->lcn < 0)) {
up_write(&mftbmp_ni->runlist.lock);
ntfs_error(vol->sb, "Failed to determine last allocated "
up_write(&vol->lcnbmp_lock);
ntfs_unmap_page(page);
/* Allocate a cluster from the DATA_ZONE. */
- rl2 = ntfs_cluster_alloc(vol, rl[1].vcn, 1, lcn, DATA_ZONE);
+ rl2 = ntfs_cluster_alloc(vol, rl[1].vcn, 1, lcn, DATA_ZONE,
+ TRUE);
if (IS_ERR(rl2)) {
up_write(&mftbmp_ni->runlist.lock);
ntfs_error(vol->sb, "Failed to allocate a cluster for "
BUG_ON(ll < rl2->vcn);
BUG_ON(ll >= rl2->vcn + rl2->length);
/* Get the size for the new mapping pairs array for this extent. */
- mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll);
+ mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll, -1);
if (unlikely(mp_size <= 0)) {
ntfs_error(vol->sb, "Get size for mapping pairs failed for "
"mft bitmap attribute extent.");
/* Generate the mapping pairs array directly into the attr record. */
ret = ntfs_mapping_pairs_build(vol, (u8*)a +
le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
- mp_size, rl2, ll, NULL);
+ mp_size, rl2, ll, -1, NULL);
if (unlikely(ret)) {
ntfs_error(vol->sb, "Failed to build mapping pairs array for "
"mft bitmap attribute.");
a->data.non_resident.mapping_pairs_offset),
old_alen - le16_to_cpu(
a->data.non_resident.mapping_pairs_offset),
- rl2, ll, NULL)) {
+ rl2, ll, -1, NULL)) {
ntfs_error(vol->sb, "Failed to restore mapping pairs "
"array.%s", es);
NVolSetErrors(vol);
read_lock_irqsave(&mft_ni->size_lock, flags);
ll = mft_ni->allocated_size;
read_unlock_irqrestore(&mft_ni->size_lock, flags);
- rl = ntfs_find_vcn_nolock(mft_ni, (ll - 1) >> vol->cluster_size_bits,
- TRUE);
+ rl = ntfs_attr_find_vcn_nolock(mft_ni,
+ (ll - 1) >> vol->cluster_size_bits, NULL);
if (unlikely(IS_ERR(rl) || !rl->length || rl->lcn < 0)) {
up_write(&mft_ni->runlist.lock);
ntfs_error(vol->sb, "Failed to determine last allocated "
nr > min_nr ? "default" : "minimal", (long long)nr);
old_last_vcn = rl[1].vcn;
do {
- rl2 = ntfs_cluster_alloc(vol, old_last_vcn, nr, lcn, MFT_ZONE);
+ rl2 = ntfs_cluster_alloc(vol, old_last_vcn, nr, lcn, MFT_ZONE,
+ TRUE);
if (likely(!IS_ERR(rl2)))
break;
if (PTR_ERR(rl2) != -ENOSPC || nr == min_nr) {
BUG_ON(ll < rl2->vcn);
BUG_ON(ll >= rl2->vcn + rl2->length);
/* Get the size for the new mapping pairs array for this extent. */
- mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll);
+ mp_size = ntfs_get_size_for_mapping_pairs(vol, rl2, ll, -1);
if (unlikely(mp_size <= 0)) {
ntfs_error(vol->sb, "Get size for mapping pairs failed for "
"mft data attribute extent.");
/* Generate the mapping pairs array directly into the attr record. */
ret = ntfs_mapping_pairs_build(vol, (u8*)a +
le16_to_cpu(a->data.non_resident.mapping_pairs_offset),
- mp_size, rl2, ll, NULL);
+ mp_size, rl2, ll, -1, NULL);
if (unlikely(ret)) {
ntfs_error(vol->sb, "Failed to build mapping pairs array of "
"mft data attribute.");
NVolSetErrors(vol);
return ret;
}
- a = ctx->attr;
- a->data.non_resident.highest_vcn = cpu_to_sle64(old_last_vcn - 1);
+ ctx->attr->data.non_resident.highest_vcn =
+ cpu_to_sle64(old_last_vcn - 1);
undo_alloc:
- if (ntfs_cluster_free(vol->mft_ino, old_last_vcn, -1) < 0) {
+ if (ntfs_cluster_free(mft_ni, old_last_vcn, -1, ctx) < 0) {
ntfs_error(vol->sb, "Failed to free clusters from mft data "
"attribute.%s", es);
NVolSetErrors(vol);
}
+ a = ctx->attr;
if (ntfs_rl_truncate_nolock(vol, &mft_ni->runlist, old_last_vcn)) {
ntfs_error(vol->sb, "Failed to truncate mft data attribute "
"runlist.%s", es);
NVolSetErrors(vol);
}
- if (mp_rebuilt) {
+ if (mp_rebuilt && !IS_ERR(ctx->mrec)) {
if (ntfs_mapping_pairs_build(vol, (u8*)a + le16_to_cpu(
a->data.non_resident.mapping_pairs_offset),
old_alen - le16_to_cpu(
a->data.non_resident.mapping_pairs_offset),
- rl2, ll, NULL)) {
+ rl2, ll, -1, NULL)) {
ntfs_error(vol->sb, "Failed to restore mapping pairs "
"array.%s", es);
NVolSetErrors(vol);
}
flush_dcache_mft_record_page(ctx->ntfs_ino);
mark_mft_record_dirty(ctx->ntfs_ino);
+ } else if (IS_ERR(ctx->mrec)) {
+ ntfs_error(vol->sb, "Failed to restore attribute search "
+ "context.%s", es);
+ NVolSetErrors(vol);
}
if (ctx)
ntfs_attr_put_search_ctx(ctx);
ni->name_len = 4;
ni->itype.index.block_size = 4096;
- ni->itype.index.block_size_bits = generic_ffs(4096) - 1;
+ ni->itype.index.block_size_bits = ntfs_ffs(4096) - 1;
ni->itype.index.collation_rule = COLLATION_FILE_NAME;
if (vol->cluster_size <= ni->itype.index.block_size) {
ni->itype.index.vcn_size = vol->cluster_size;
* have its page mapped and it is very easy to do.
*/
atomic_inc(&ni->count);
- down(&ni->mrec_lock);
+ mutex_lock(&ni->mrec_lock);
ni->page = page;
ni->page_ofs = ofs;
/*
BUG_ON(NInoAttr(ni));
BUG_ON(ni->nr_extents != -1);
- down(&ni->extent_lock);
+ mutex_lock(&ni->extent_lock);
base_ni = ni->ext.base_ntfs_ino;
- up(&ni->extent_lock);
+ mutex_unlock(&ni->extent_lock);
BUG_ON(base_ni->nr_extents <= 0);
ntfs_debug("Entering for extent inode 0x%lx, base inode 0x%lx.\n",
mft_no, base_ni->mft_no);
- down(&base_ni->extent_lock);
+ mutex_lock(&base_ni->extent_lock);
/* Make sure we are holding the only reference to the extent inode. */
if (atomic_read(&ni->count) > 2) {
ntfs_error(vol->sb, "Tried to free busy extent inode 0x%lx, "
"not freeing.", base_ni->mft_no);
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
return -EBUSY;
}
break;
}
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
if (unlikely(err)) {
ntfs_error(vol->sb, "Extent inode 0x%lx is not attached to "
return 0;
rollback:
/* Rollback what we did... */
- down(&base_ni->extent_lock);
+ mutex_lock(&base_ni->extent_lock);
extent_nis = base_ni->ext.extent_ntfs_inos;
if (!(base_ni->nr_extents & 3)) {
int new_size = (base_ni->nr_extents + 4) * sizeof(ntfs_inode*);
- extent_nis = (ntfs_inode**)kmalloc(new_size, GFP_NOFS);
+ extent_nis = kmalloc(new_size, GFP_NOFS);
if (unlikely(!extent_nis)) {
ntfs_error(vol->sb, "Failed to allocate internal "
"buffer during rollback.%s", es);
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
NVolSetErrors(vol);
goto rollback_error;
}
m->flags |= MFT_RECORD_IN_USE;
m->sequence_number = old_seq_no;
extent_nis[base_ni->nr_extents++] = ni;
- up(&base_ni->extent_lock);
+ mutex_unlock(&base_ni->extent_lock);
mark_mft_record_dirty(ni);
return err;
}
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff