Re: [RFC 4/5] support >32 bit ext3: sector_t convert in JBD
Mingming Cao <[email protected]>
| Newsgroups | gmane.comp.file-systems.ext2.devel |
|---|---|
| Organization | IBM LTC |
| Message-ID | <[email protected]> |
On Tue, 2006-05-09 at 22:05 +0100, Stephen C. Tweedie wrote:
> Hi,
>
> On Mon, 2006-05-08 at 12:22 -0700, Mingming Cao wrote:
>
> > --- linux-2.6.16/include/linux/jbd.h~sector_t-jbd 2006-05-08 11:53:51.915400370 -0700
> > +++ linux-2.6.16-ming/include/linux/jbd.h 2006-05-08 11:53:51.941397358 -0700
> > + sector_t t_log_start;
> > + sector_t j_head;
> > + sector_t j_tail;
> > + sector_t j_free;
> > + sector_t j_first;
> > + sector_t j_last;
> > + sector_t j_blk_offset;
> > + sector_t j_maxlen;
>
> Of all of these, only j_blk_offset seems useful to have as sector_t,
> unless we really want to go to a 64-bit journal size, which I doubt!
> The j_blk_offset is potentially useful when placing the journal on an
> external large device, but even then I doubt we'll be using external
> shared journal devices larger than 16TB!
>
> > -int journal_bmap(journal_t *journal, unsigned long blocknr,
> > - unsigned long *retp)
> > +int journal_bmap(journal_t *journal, sector_t blocknr,
> > + sector_t *retp)
>
> This, and a _lot_ of the other changes, assume we're going to a sector_t
> type for both the sectors the journal occupies on disk (which is
> reasonable), and for the sector offsets themselves (which is not.)
> There are a huge number of changes all over the place which fall into
> this category. We really don't want to be blindly peanut-buttering
> these changes everywhere: I just don't see the need for individual
> journal files larger than 16TB!
>
> > -static int do_readahead(journal_t *journal, unsigned int start)
> > +static int do_readahead(journal_t *journal, sector_t start)
> > {
> > int err;
> > - unsigned int max, nbufs, next;
> > - unsigned long blocknr;
> > + sector_t max, nbufs, next;
> > + sector_t blocknr;
>
> I don't think we need to cater for journals larger than 2^32 blocks;
> butnbufs is just an array index and *certainly* shouldn't be a sector_t.
>
> > /* Borrowed from buffer.c: this is a tried and tested block hash function */
> > -static inline int hash(journal_t *journal, unsigned long block)
> > +static inline int hash(journal_t *journal, sector_t block)
>
> This largely ignores the top 32 bits of the block in the hash function:
> perhaps use an "int hash = (int) block ^ ((int) block >> 32)" before
> computing it?
>
> --Stephen
>
Here is the updated sector_t-jbd patch.
Signed-Off-By: Mingming Cao <[email protected]>
---
linux-2.6.16-ming/fs/jbd/commit.c | 2 +-
linux-2.6.16-ming/fs/jbd/journal.c | 16 ++++++++--------
linux-2.6.16-ming/fs/jbd/recovery.c | 8 ++++----
linux-2.6.16-ming/fs/jbd/revoke.c | 24 +++++++++++++-----------
linux-2.6.16-ming/include/linux/ext3_jbd.h | 2 +-
linux-2.6.16-ming/include/linux/jbd.h | 17 ++++++++---------
6 files changed, 35 insertions(+), 34 deletions(-)
diff -puN fs/jbd/commit.c~sector_t-jbd fs/jbd/commit.c
--- linux-2.6.16/fs/jbd/commit.c~sector_t-jbd 2006-06-07 17:15:03.000000000 -0700
+++ linux-2.6.16-ming/fs/jbd/commit.c 2006-06-07 17:15:03.000000000 -0700
@@ -180,7 +180,7 @@ void journal_commit_transaction(journal_
int bufs;
int flags;
int err;
- unsigned long blocknr;
+ sector_t blocknr;
char *tagp = NULL;
journal_header_t *header;
journal_block_tag_t *tag = NULL;
diff -puN fs/jbd/recovery.c~sector_t-jbd fs/jbd/recovery.c
--- linux-2.6.16/fs/jbd/recovery.c~sector_t-jbd 2006-06-07 17:15:03.000000000 -0700
+++ linux-2.6.16-ming/fs/jbd/recovery.c 2006-06-07 17:15:03.000000000 -0700
@@ -70,7 +70,7 @@ static int do_readahead(journal_t *journ
{
int err;
unsigned int max, nbufs, next;
- unsigned long blocknr;
+ sector_t blocknr;
struct buffer_head *bh;
struct buffer_head * bufs[MAXBUF];
@@ -132,7 +132,7 @@ static int jread(struct buffer_head **bh
unsigned int offset)
{
int err;
- unsigned long blocknr;
+ sector_t blocknr;
struct buffer_head *bh;
*bhp = NULL;
@@ -451,7 +451,7 @@ static int do_one_pass(journal_t *journa
"block %ld in log\n",
err, io_block);
} else {
- unsigned long blocknr;
+ sector_t blocknr;
J_ASSERT(obh != NULL);
blocknr = read_split_be64(&tag->t_blocknr_high,
@@ -590,7 +590,7 @@ static int scan_revoke_records(journal_t
record_len = 8;
while (offset + record_len < max) {
- unsigned long blocknr;
+ sector_t blocknr;
int err;
if (record_len == 4)
diff -puN fs/jbd/revoke.c~sector_t-jbd fs/jbd/revoke.c
--- linux-2.6.16/fs/jbd/revoke.c~sector_t-jbd 2006-06-07 17:15:03.000000000 -0700
+++ linux-2.6.16-ming/fs/jbd/revoke.c 2006-06-08 13:46:44.664807799 -0700
@@ -81,7 +81,7 @@ struct jbd_revoke_record_s
{
struct list_head hash;
tid_t sequence; /* Used for recovery only */
- unsigned long blocknr;
+ sector_t blocknr;
};
@@ -106,17 +106,18 @@ static void flush_descriptor(journal_t *
/* Utility functions to maintain the revoke table */
/* Borrowed from buffer.c: this is a tried and tested block hash function */
-static inline int hash(journal_t *journal, unsigned long block)
+static inline int hash(journal_t *journal, sector_t block)
{
struct jbd_revoke_table_s *table = journal->j_revoke;
int hash_shift = table->hash_shift;
+ int hash = (int)block ^ (int)(block >> 32);
- return ((block << (hash_shift - 6)) ^
- (block >> 13) ^
- (block << (hash_shift - 12))) & (table->hash_size - 1);
+ return ((hash << (hash_shift - 6)) ^
+ (hash >> 13) ^
+ (hash << (hash_shift - 12))) & (table->hash_size - 1);
}
-static int insert_revoke_hash(journal_t *journal, unsigned long blocknr,
+static int insert_revoke_hash(journal_t *journal, sector_t blocknr,
tid_t seq)
{
struct list_head *hash_list;
@@ -146,7 +147,7 @@ oom:
/* Find a revoke record in the journal's hash table. */
static struct jbd_revoke_record_s *find_revoke_record(journal_t *journal,
- unsigned long blocknr)
+ sector_t blocknr)
{
struct list_head *hash_list;
struct jbd_revoke_record_s *record;
@@ -325,7 +326,7 @@ void journal_destroy_revoke(journal_t *j
* by one.
*/
-int journal_revoke(handle_t *handle, unsigned long blocknr,
+int journal_revoke(handle_t *handle, sector_t blocknr,
struct buffer_head *bh_in)
{
struct buffer_head *bh = NULL;
@@ -394,7 +395,8 @@ int journal_revoke(handle_t *handle, uns
}
}
- jbd_debug(2, "insert revoke for block %lu, bh_in=%p\n", blocknr, bh_in);
+ jbd_debug(2, "insert revoke for block %llu, bh_in=%p\n",
+ blocknr, bh_in);
err = insert_revoke_hash(journal, blocknr,
handle->h_transaction->t_tid);
BUFFER_TRACE(bh_in, "exit");
@@ -649,7 +651,7 @@ static void flush_descriptor(journal_t *
*/
int journal_set_revoke(journal_t *journal,
- unsigned long blocknr,
+ sector_t blocknr,
tid_t sequence)
{
struct jbd_revoke_record_s *record;
@@ -673,7 +675,7 @@ int journal_set_revoke(journal_t *journa
*/
int journal_test_revoke(journal_t *journal,
- unsigned long blocknr,
+ sector_t blocknr,
tid_t sequence)
{
struct jbd_revoke_record_s *record;
diff -puN include/linux/ext3_jbd.h~sector_t-jbd include/linux/ext3_jbd.h
--- linux-2.6.16/include/linux/ext3_jbd.h~sector_t-jbd 2006-06-07 17:15:03.000000000 -0700
+++ linux-2.6.16-ming/include/linux/ext3_jbd.h 2006-06-07 17:15:03.000000000 -0700
@@ -149,7 +149,7 @@ __ext3_journal_forget(const char *where,
static inline int
__ext3_journal_revoke(const char *where, handle_t *handle,
- unsigned long blocknr, struct buffer_head *bh)
+ ext3_fsblk_t blocknr, struct buffer_head *bh)
{
int err = journal_revoke(handle, blocknr, bh);
if (err)
diff -puN include/linux/jbd.h~sector_t-jbd include/linux/jbd.h
--- linux-2.6.16/include/linux/jbd.h~sector_t-jbd 2006-06-07 17:15:03.000000000 -0700
+++ linux-2.6.16-ming/include/linux/jbd.h 2006-06-07 17:15:03.000000000 -0700
@@ -729,7 +729,7 @@ struct journal_s
*/
struct block_device *j_dev;
int j_blocksize;
- unsigned int j_blk_offset;
+ sector_t j_blk_offset;
/*
* Device which holds the client fs. For internal journal this will be
@@ -848,7 +848,7 @@ extern void __journal_clean_data_list(tr
/* Log buffer allocation */
extern struct journal_head * journal_get_descriptor_buffer(journal_t *);
-int journal_next_log_block(journal_t *, unsigned long *);
+int journal_next_log_block(journal_t *, sector_t *);
/* Commit management */
extern void journal_commit_transaction(journal_t *);
@@ -863,7 +863,7 @@ extern int
journal_write_metadata_buffer(transaction_t *transaction,
struct journal_head *jh_in,
struct journal_head **jh_out,
- int blocknr);
+ sector_t blocknr);
/* Transaction locking */
extern void __wait_on_journal (journal_t *);
@@ -911,7 +911,7 @@ extern void journal_unlock_updates (jou
extern journal_t * journal_init_dev(struct block_device *bdev,
struct block_device *fs_dev,
- int start, int len, int bsize);
+ sector_t start, int len, int bsize);
extern journal_t * journal_init_inode (struct inode *);
extern int journal_update_format (journal_t *);
extern int journal_check_used_features
@@ -932,7 +932,7 @@ extern void journal_abort (journ
extern int journal_errno (journal_t *);
extern void journal_ack_err (journal_t *);
extern int journal_clear_err (journal_t *);
-extern int journal_bmap(journal_t *, unsigned long, unsigned long *);
+extern int journal_bmap(journal_t *, unsigned long, sector_t *);
extern int journal_force_commit(journal_t *);
/*
@@ -965,14 +965,13 @@ extern void journal_destroy_revoke_ca
extern int journal_init_revoke_caches(void);
extern void journal_destroy_revoke(journal_t *);
-extern int journal_revoke (handle_t *,
- unsigned long, struct buffer_head *);
+extern int journal_revoke (handle_t *, sector_t, struct buffer_head *);
extern int journal_cancel_revoke(handle_t *, struct journal_head *);
extern void journal_write_revoke_records(journal_t *, transaction_t *);
/* Recovery revoke support */
-extern int journal_set_revoke(journal_t *, unsigned long, tid_t);
-extern int journal_test_revoke(journal_t *, unsigned long, tid_t);
+extern int journal_set_revoke(journal_t *, sector_t, tid_t);
+extern int journal_test_revoke(journal_t *, sector_t, tid_t);
extern void journal_clear_revoke(journal_t *);
extern void journal_brelse_array(struct buffer_head *b[], int n);
extern void journal_switch_revoke_table(journal_t *journal);
diff -puN fs/jbd/journal.c~sector_t-jbd fs/jbd/journal.c
--- linux-2.6.16/fs/jbd/journal.c~sector_t-jbd 2006-06-07 17:15:03.000000000 -0700
+++ linux-2.6.16-ming/fs/jbd/journal.c 2006-06-07 17:15:03.000000000 -0700
@@ -269,7 +269,7 @@ static void journal_kill_thread(journal_
int journal_write_metadata_buffer(transaction_t *transaction,
struct journal_head *jh_in,
struct journal_head **jh_out,
- int blocknr)
+ sector_t blocknr)
{
int need_copy_out = 0;
int done_copy_out = 0;
@@ -553,7 +553,7 @@ int log_wait_commit(journal_t *journal,
* Log buffer allocation routines:
*/
-int journal_next_log_block(journal_t *journal, unsigned long *retp)
+int journal_next_log_block(journal_t *journal, sector_t *retp)
{
unsigned long blocknr;
@@ -577,10 +577,10 @@ int journal_next_log_block(journal_t *jo
* ready.
*/
int journal_bmap(journal_t *journal, unsigned long blocknr,
- unsigned long *retp)
+ sector_t *retp)
{
int err = 0;
- unsigned long ret;
+ sector_t ret;
if (journal->j_inode) {
ret = bmap(journal->j_inode, blocknr);
@@ -616,7 +616,7 @@ int journal_bmap(journal_t *journal, uns
struct journal_head *journal_get_descriptor_buffer(journal_t *journal)
{
struct buffer_head *bh;
- unsigned long blocknr;
+ sector_t blocknr;
int err;
err = journal_next_log_block(journal, &blocknr);
@@ -704,7 +704,7 @@ fail:
*/
journal_t * journal_init_dev(struct block_device *bdev,
struct block_device *fs_dev,
- int start, int len, int blocksize)
+ sector_t start, int len, int blocksize)
{
journal_t *journal = journal_init_common();
struct buffer_head *bh;
@@ -752,7 +752,7 @@ journal_t * journal_init_inode (struct i
journal_t *journal = journal_init_common();
int err;
int n;
- unsigned long blocknr;
+ sector_t blocknr;
if (!journal)
return NULL;
@@ -852,7 +852,7 @@ static int journal_reset(journal_t *jour
**/
int journal_create(journal_t *journal)
{
- unsigned long blocknr;
+ sector_t blocknr;
struct buffer_head *bh;
journal_superblock_t *sb;
int i, err;
_