* [PATCH 01/17] xfs: remove done roadmap item from xfs-delayed-logging-design.txt
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 4:33 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 02/17] xfs: skip writeback from reclaim context Christoph Hellwig
` (17 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-remove-roadmap-item --]
[-- Type: text/plain, Size: 878 bytes --]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/Documentation/filesystems/xfs-delayed-logging-design.txt
===================================================================
--- xfs.orig/Documentation/filesystems/xfs-delayed-logging-design.txt 2010-05-30 12:23:13.223254240 +0200
+++ xfs/Documentation/filesystems/xfs-delayed-logging-design.txt 2010-05-30 12:23:57.363005673 +0200
@@ -794,11 +794,6 @@ designed.
Roadmap:
-2.6.35 Inclusion in mainline as an experimental mount option
- => approximately 2-3 months to merge window
- => needs to be in xfs-dev tree in 4-6 weeks
- => code is nearing readiness for review
-
2.6.37 Remove experimental tag from mount option
=> should be roughly 6 months after initial merge
=> enough time to:
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 01/17] xfs: remove done roadmap item from xfs-delayed-logging-design.txt
2010-05-31 16:07 ` [PATCH 01/17] xfs: remove done roadmap item from xfs-delayed-logging-design.txt Christoph Hellwig
@ 2010-06-02 4:33 ` Dave Chinner
0 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 4:33 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:28PM -0400, Christoph Hellwig wrote:
> Signed-off-by: Christoph Hellwig <hch@lst.de>
>
> Index: xfs/Documentation/filesystems/xfs-delayed-logging-design.txt
> ===================================================================
> --- xfs.orig/Documentation/filesystems/xfs-delayed-logging-design.txt 2010-05-30 12:23:13.223254240 +0200
> +++ xfs/Documentation/filesystems/xfs-delayed-logging-design.txt 2010-05-30 12:23:57.363005673 +0200
> @@ -794,11 +794,6 @@ designed.
>
> Roadmap:
>
> -2.6.35 Inclusion in mainline as an experimental mount option
> - => approximately 2-3 months to merge window
> - => needs to be in xfs-dev tree in 4-6 weeks
> - => code is nearing readiness for review
> -
> 2.6.37 Remove experimental tag from mount option
> => should be roughly 6 months after initial merge
> => enough time to:
Reviewed-by: Dave Chinner <dchinner@redhat.com>
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* [PATCH 02/17] xfs: skip writeback from reclaim context
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
2010-05-31 16:07 ` [PATCH 01/17] xfs: remove done roadmap item from xfs-delayed-logging-design.txt Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 4:39 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 03/17] xfs: improve xfs_isilocked Christoph Hellwig
` (16 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-disable-allocations-from-reclaim --]
[-- Type: text/plain, Size: 1772 bytes --]
Allowing writeback from reclaim context causes massive problems with stack
overflows as we can call into the writeback code which tends to be a heavy
stack user both in the generic code and XFS from random contexts that
perform memory allocations.
Follow the example of btrfs (and in slightly different form ext4) and refuse
to write out data from reclaim context. This issue should really be handled
by the VM so that we can tune better for this case, but until we get it
sorted out there we have to hack around this in each filesystem with a
complex writeback path.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c 2010-05-29 18:09:01.713253891 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c 2010-05-29 18:11:46.847261713 +0200
@@ -1333,6 +1333,21 @@ xfs_vm_writepage(
trace_xfs_writepage(inode, page, 0);
/*
+ * Refuse to write the page out if we are called from reclaim context.
+ *
+ * This is primarily to avoid stack overflows when called from deep
+ * used stacks in random callers for direct reclaim, but disabling
+ * reclaim for kswap is a nice side-effect as kswapd causes rather
+ * suboptimal I/O patters, too.
+ *
+ * This should really be done by the core VM, but until that happens
+ * filesystems like XFS, btrfs and ext4 have to take care of this
+ * by themselves.
+ */
+ if (current->flags & PF_MEMALLOC)
+ goto out_fail;
+
+ /*
* We need a transaction if:
* 1. There are delalloc buffers on the page
* 2. The page is uptodate and we have unmapped buffers
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 02/17] xfs: skip writeback from reclaim context
2010-05-31 16:07 ` [PATCH 02/17] xfs: skip writeback from reclaim context Christoph Hellwig
@ 2010-06-02 4:39 ` Dave Chinner
2010-06-02 10:08 ` Christoph Hellwig
0 siblings, 1 reply; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 4:39 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:29PM -0400, Christoph Hellwig wrote:
> Allowing writeback from reclaim context causes massive problems with stack
> overflows as we can call into the writeback code which tends to be a heavy
> stack user both in the generic code and XFS from random contexts that
> perform memory allocations.
>
> Follow the example of btrfs (and in slightly different form ext4) and refuse
> to write out data from reclaim context. This issue should really be handled
> by the VM so that we can tune better for this case, but until we get it
> sorted out there we have to hack around this in each filesystem with a
> complex writeback path.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Yup, I agree that we have to do this for the moment. It hasn't
caused any problems in my usual OOM testing, so it seems safe
to do.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Also worth thinking about is if should be checked in
xfs_vm_releasepage() as well to avoid the same stack issues if it
triggers allocation...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* Re: [PATCH 02/17] xfs: skip writeback from reclaim context
2010-06-02 4:39 ` Dave Chinner
@ 2010-06-02 10:08 ` Christoph Hellwig
2010-06-02 23:02 ` Dave Chinner
0 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-06-02 10:08 UTC (permalink / raw)
To: Dave Chinner; +Cc: Christoph Hellwig, xfs
On Wed, Jun 02, 2010 at 02:39:57PM +1000, Dave Chinner wrote:
> Also worth thinking about is if should be checked in
> xfs_vm_releasepage() as well to avoid the same stack issues if it
> triggers allocation...
I agree this is a potential problem as well.
I did some QA runs with this thrown in, and it causes massive OOM killer
wreckage in xfstests. I've also cross-checked btrfs and ext4 and
neither one skips ->releasepage from reclaim context.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* Re: [PATCH 02/17] xfs: skip writeback from reclaim context
2010-06-02 10:08 ` Christoph Hellwig
@ 2010-06-02 23:02 ` Dave Chinner
2010-06-03 6:52 ` Christoph Hellwig
0 siblings, 1 reply; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 23:02 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Wed, Jun 02, 2010 at 06:08:12AM -0400, Christoph Hellwig wrote:
> On Wed, Jun 02, 2010 at 02:39:57PM +1000, Dave Chinner wrote:
> > Also worth thinking about is if should be checked in
> > xfs_vm_releasepage() as well to avoid the same stack issues if it
> > triggers allocation...
>
> I agree this is a potential problem as well.
>
> I did some QA runs with this thrown in, and it causes massive OOM killer
> wreckage in xfstests. I've also cross-checked btrfs and ext4 and
> neither one skips ->releasepage from reclaim context.
Did you skip it unconditionally, or only when a transaction was
required?
The scary part is that I've seen stack traces (i.e. most stack used)
through this reclaim path for delalloc conversion even for
allocations that are GFP_NOFS and the only thing saving us from
deadlocks is th PF_FSTRANS check. Even worse is that
shrinker_page_list() will call try_to_release_pages() without
checking whether it's allowed to enter the filesystem or not, so we
can be doing block allocation in places we've specifically told the
memory allocation subsystem not to....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* Re: [PATCH 02/17] xfs: skip writeback from reclaim context
2010-06-02 23:02 ` Dave Chinner
@ 2010-06-03 6:52 ` Christoph Hellwig
0 siblings, 0 replies; 43+ messages in thread
From: Christoph Hellwig @ 2010-06-03 6:52 UTC (permalink / raw)
To: Dave Chinner; +Cc: linux-mm, xfs
On Thu, Jun 03, 2010 at 09:02:09AM +1000, Dave Chinner wrote:
> Did you skip it unconditionally, or only when a transaction was
> required?
xfs_vm_releasepage is mostly a no-op if no transaction is required.
If we have neither delalloc nor unwritten buffer we do not actually
enter xfs_page_state_convert, and ->releasepage also doesn't touch
unampped buffers at all.
> The scary part is that I've seen stack traces (i.e. most stack used)
> through this reclaim path for delalloc conversion even for
> allocations that are GFP_NOFS and the only thing saving us from
> deadlocks is th PF_FSTRANS check. Even worse is that
> shrinker_page_list() will call try_to_release_pages() without
> checking whether it's allowed to enter the filesystem or not, so we
> can be doing block allocation in places we've specifically told the
> memory allocation subsystem not to....
s/shrinker_page_list/shrink_page_list/ and
s/try_to_release_pages/try_to_release_page/ above.
shrink_page_list takes the gfp_mask for try_to_release_page from
the scan_control structure passed to it from all the top of the long
callchain. I can't find anobvious bug, but this could cause a lot
more harm.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* Re: [PATCH 02/17] xfs: skip writeback from reclaim context
@ 2010-06-03 6:52 ` Christoph Hellwig
0 siblings, 0 replies; 43+ messages in thread
From: Christoph Hellwig @ 2010-06-03 6:52 UTC (permalink / raw)
To: Dave Chinner; +Cc: xfs, linux-mm
On Thu, Jun 03, 2010 at 09:02:09AM +1000, Dave Chinner wrote:
> Did you skip it unconditionally, or only when a transaction was
> required?
xfs_vm_releasepage is mostly a no-op if no transaction is required.
If we have neither delalloc nor unwritten buffer we do not actually
enter xfs_page_state_convert, and ->releasepage also doesn't touch
unampped buffers at all.
> The scary part is that I've seen stack traces (i.e. most stack used)
> through this reclaim path for delalloc conversion even for
> allocations that are GFP_NOFS and the only thing saving us from
> deadlocks is th PF_FSTRANS check. Even worse is that
> shrinker_page_list() will call try_to_release_pages() without
> checking whether it's allowed to enter the filesystem or not, so we
> can be doing block allocation in places we've specifically told the
> memory allocation subsystem not to....
s/shrinker_page_list/shrink_page_list/ and
s/try_to_release_pages/try_to_release_page/ above.
shrink_page_list takes the gfp_mask for try_to_release_page from
the scan_control structure passed to it from all the top of the long
callchain. I can't find anobvious bug, but this could cause a lot
more harm.
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>
^ permalink raw reply [flat|nested] 43+ messages in thread
* [PATCH 03/17] xfs: improve xfs_isilocked
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
2010-05-31 16:07 ` [PATCH 01/17] xfs: remove done roadmap item from xfs-delayed-logging-design.txt Christoph Hellwig
2010-05-31 16:07 ` [PATCH 02/17] xfs: skip writeback from reclaim context Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 4:41 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 04/17] xfs: drop dmapi hooks Christoph Hellwig
` (15 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-improve-mrlock-debug --]
[-- Type: text/plain, Size: 1580 bytes --]
Use rwsem_is_locked to make the assertations for shared locks work.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/xfs_iget.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iget.c 2010-05-29 18:09:02.083254031 +0200
+++ xfs/fs/xfs/xfs_iget.c 2010-05-29 18:11:58.194003577 +0200
@@ -741,30 +741,24 @@ xfs_ilock_demote(
}
#ifdef DEBUG
-/*
- * Debug-only routine, without additional rw_semaphore APIs, we can
- * now only answer requests regarding whether we hold the lock for write
- * (reader state is outside our visibility, we only track writer state).
- *
- * Note: this means !xfs_isilocked would give false positives, so don't do that.
- */
int
xfs_isilocked(
xfs_inode_t *ip,
uint lock_flags)
{
- if ((lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) ==
- XFS_ILOCK_EXCL) {
- if (!ip->i_lock.mr_writer)
- return 0;
+ if (lock_flags & (XFS_ILOCK_EXCL|XFS_ILOCK_SHARED)) {
+ if (!(lock_flags & XFS_ILOCK_SHARED))
+ return !!ip->i_lock.mr_writer;
+ return rwsem_is_locked(&ip->i_lock.mr_lock);
}
- if ((lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) ==
- XFS_IOLOCK_EXCL) {
- if (!ip->i_iolock.mr_writer)
- return 0;
+ if (lock_flags & (XFS_IOLOCK_EXCL|XFS_IOLOCK_SHARED)) {
+ if (!(lock_flags & XFS_IOLOCK_SHARED))
+ return !!ip->i_iolock.mr_writer;
+ return rwsem_is_locked(&ip->i_iolock.mr_lock);
}
- return 1;
+ ASSERT(0);
+ return 0;
}
#endif
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 04/17] xfs: drop dmapi hooks
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (2 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 03/17] xfs: improve xfs_isilocked Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 4:45 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 05/17] xfs: remove unneeded #include statements Christoph Hellwig
` (14 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-drop-dmapi-hooks --]
[-- Type: text/plain, Size: 62954 bytes --]
Dmapi support was never merged upstream, but we still have a lot of hooks
bloating XFS for it, all over the fast pathes of the filesystem.
This patch drops over 700 lines of dmapi overhead. If we'll ever get HSM
support in mainline at least the namespace events can be done much saner
in the VFS instead of the individual filesystem, so it's not like this
is much help for future work.
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
V1 -> V2
- updated to account for the delayed logging branch merge
- remove a few more left overs pointed out by Eric
Index: xfs/fs/xfs/linux-2.6/xfs_file.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_file.c 2010-05-22 00:40:37.890004594 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_file.c 2010-05-25 11:14:59.236274548 +0200
@@ -24,7 +24,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_trans.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
@@ -275,20 +274,6 @@ xfs_file_aio_read(
mutex_lock(&inode->i_mutex);
xfs_ilock(ip, XFS_IOLOCK_SHARED);
- if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
- int dmflags = FILP_DELAY_FLAG(file) | DM_SEM_FLAG_RD(ioflags);
- int iolock = XFS_IOLOCK_SHARED;
-
- ret = -XFS_SEND_DATA(mp, DM_EVENT_READ, ip, iocb->ki_pos, size,
- dmflags, &iolock);
- if (ret) {
- xfs_iunlock(ip, XFS_IOLOCK_SHARED);
- if (unlikely(ioflags & IO_ISDIRECT))
- mutex_unlock(&inode->i_mutex);
- return ret;
- }
- }
-
if (unlikely(ioflags & IO_ISDIRECT)) {
if (inode->i_mapping->nrpages) {
ret = -xfs_flushinval_pages(ip,
@@ -321,7 +306,6 @@ xfs_file_splice_read(
unsigned int flags)
{
struct xfs_inode *ip = XFS_I(infilp->f_mapping->host);
- struct xfs_mount *mp = ip->i_mount;
int ioflags = 0;
ssize_t ret;
@@ -335,18 +319,6 @@ xfs_file_splice_read(
xfs_ilock(ip, XFS_IOLOCK_SHARED);
- if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) && !(ioflags & IO_INVIS)) {
- int iolock = XFS_IOLOCK_SHARED;
- int error;
-
- error = XFS_SEND_DATA(mp, DM_EVENT_READ, ip, *ppos, count,
- FILP_DELAY_FLAG(infilp), &iolock);
- if (error) {
- xfs_iunlock(ip, XFS_IOLOCK_SHARED);
- return -error;
- }
- }
-
trace_xfs_file_splice_read(ip, count, *ppos, ioflags);
ret = generic_file_splice_read(infilp, ppos, pipe, count, flags);
@@ -367,7 +339,6 @@ xfs_file_splice_write(
{
struct inode *inode = outfilp->f_mapping->host;
struct xfs_inode *ip = XFS_I(inode);
- struct xfs_mount *mp = ip->i_mount;
xfs_fsize_t isize, new_size;
int ioflags = 0;
ssize_t ret;
@@ -382,18 +353,6 @@ xfs_file_splice_write(
xfs_ilock(ip, XFS_IOLOCK_EXCL);
- if (DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) && !(ioflags & IO_INVIS)) {
- int iolock = XFS_IOLOCK_EXCL;
- int error;
-
- error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, ip, *ppos, count,
- FILP_DELAY_FLAG(outfilp), &iolock);
- if (error) {
- xfs_iunlock(ip, XFS_IOLOCK_EXCL);
- return -error;
- }
- }
-
new_size = *ppos + count;
xfs_ilock(ip, XFS_ILOCK_EXCL);
@@ -627,7 +586,6 @@ xfs_file_aio_write(
int ioflags = 0;
xfs_fsize_t isize, new_size;
int iolock;
- int eventsent = 0;
size_t ocount = 0, count;
int need_i_mutex;
@@ -673,33 +631,6 @@ start:
goto out_unlock_mutex;
}
- if ((DM_EVENT_ENABLED(ip, DM_EVENT_WRITE) &&
- !(ioflags & IO_INVIS) && !eventsent)) {
- int dmflags = FILP_DELAY_FLAG(file);
-
- if (need_i_mutex)
- dmflags |= DM_FLAGS_IMUX;
-
- xfs_iunlock(ip, XFS_ILOCK_EXCL);
- error = XFS_SEND_DATA(ip->i_mount, DM_EVENT_WRITE, ip,
- pos, count, dmflags, &iolock);
- if (error) {
- goto out_unlock_internal;
- }
- xfs_ilock(ip, XFS_ILOCK_EXCL);
- eventsent = 1;
-
- /*
- * The iolock was dropped and reacquired in XFS_SEND_DATA
- * so we have to recheck the size when appending.
- * We will only "goto start;" once, since having sent the
- * event prevents another call to XFS_SEND_DATA, which is
- * what allows the size to change in the first place.
- */
- if ((file->f_flags & O_APPEND) && pos != ip->i_size)
- goto start;
- }
-
if (ioflags & IO_ISDIRECT) {
xfs_buftarg_t *target =
XFS_IS_REALTIME_INODE(ip) ?
@@ -830,22 +761,6 @@ write_retry:
xfs_iunlock(ip, XFS_ILOCK_EXCL);
}
- if (ret == -ENOSPC &&
- DM_EVENT_ENABLED(ip, DM_EVENT_NOSPACE) && !(ioflags & IO_INVIS)) {
- xfs_iunlock(ip, iolock);
- if (need_i_mutex)
- mutex_unlock(&inode->i_mutex);
- error = XFS_SEND_NAMESP(ip->i_mount, DM_EVENT_NOSPACE, ip,
- DM_RIGHT_NULL, ip, DM_RIGHT_NULL, NULL, NULL,
- 0, 0, 0); /* Delay flag intentionally unused */
- if (need_i_mutex)
- mutex_lock(&inode->i_mutex);
- xfs_ilock(ip, iolock);
- if (error)
- goto out_unlock_internal;
- goto start;
- }
-
error = -ret;
if (ret <= 0)
goto out_unlock_internal;
Index: xfs/fs/xfs/linux-2.6/xfs_ioctl.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_ioctl.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_ioctl.c 2010-05-25 11:14:59.237275526 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
@@ -1117,16 +1116,7 @@ xfs_ioctl_setattr(
xfs_qm_dqrele(udqp);
xfs_qm_dqrele(gdqp);
- if (code)
- return code;
-
- if (DM_EVENT_ENABLED(ip, DM_EVENT_ATTRIBUTE)) {
- XFS_SEND_NAMESP(mp, DM_EVENT_ATTRIBUTE, ip, DM_RIGHT_NULL,
- NULL, DM_RIGHT_NULL, NULL, NULL, 0, 0,
- (mask & FSX_NONBLOCK) ? DM_FLAGS_NDELAY : 0);
- }
-
- return 0;
+ return code;
error_return:
xfs_qm_dqrele(udqp);
Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c 2010-05-24 23:11:41.018253918 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.c 2010-05-25 11:15:23.496005936 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
@@ -116,9 +115,6 @@ mempool_t *xfs_ioend_pool;
#define MNTOPT_GQUOTANOENF "gqnoenforce"/* group quota limit enforcement */
#define MNTOPT_PQUOTANOENF "pqnoenforce"/* project quota limit enforcement */
#define MNTOPT_QUOTANOENF "qnoenforce" /* same as uqnoenforce */
-#define MNTOPT_DMAPI "dmapi" /* DMI enabled (DMAPI / XDSM) */
-#define MNTOPT_XDSM "xdsm" /* DMI enabled (DMAPI / XDSM) */
-#define MNTOPT_DMI "dmi" /* DMI enabled (DMAPI / XDSM) */
#define MNTOPT_DELAYLOG "delaylog" /* Delayed loging enabled */
#define MNTOPT_NODELAYLOG "nodelaylog" /* Delayed loging disabled */
@@ -172,15 +168,13 @@ suffix_strtoul(char *s, char **endp, uns
STATIC int
xfs_parseargs(
struct xfs_mount *mp,
- char *options,
- char **mtpt)
+ char *options)
{
struct super_block *sb = mp->m_super;
char *this_char, *value, *eov;
int dsunit = 0;
int dswidth = 0;
int iosize = 0;
- int dmapi_implies_ikeep = 1;
__uint8_t iosizelog = 0;
/*
@@ -243,15 +237,10 @@ xfs_parseargs(
if (!mp->m_logname)
return ENOMEM;
} else if (!strcmp(this_char, MNTOPT_MTPT)) {
- if (!value || !*value) {
- cmn_err(CE_WARN,
- "XFS: %s option requires an argument",
- this_char);
- return EINVAL;
- }
- *mtpt = kstrndup(value, MAXNAMELEN, GFP_KERNEL);
- if (!*mtpt)
- return ENOMEM;
+ cmn_err(CE_WARN,
+ "XFS: %s option not allowed on this system",
+ this_char);
+ return EINVAL;
} else if (!strcmp(this_char, MNTOPT_RTDEV)) {
if (!value || !*value) {
cmn_err(CE_WARN,
@@ -329,7 +318,6 @@ xfs_parseargs(
} else if (!strcmp(this_char, MNTOPT_IKEEP)) {
mp->m_flags |= XFS_MOUNT_IKEEP;
} else if (!strcmp(this_char, MNTOPT_NOIKEEP)) {
- dmapi_implies_ikeep = 0;
mp->m_flags &= ~XFS_MOUNT_IKEEP;
} else if (!strcmp(this_char, MNTOPT_LARGEIO)) {
mp->m_flags &= ~XFS_MOUNT_COMPAT_IOSIZE;
@@ -370,12 +358,6 @@ xfs_parseargs(
} else if (!strcmp(this_char, MNTOPT_GQUOTANOENF)) {
mp->m_qflags |= (XFS_GQUOTA_ACCT | XFS_GQUOTA_ACTIVE);
mp->m_qflags &= ~XFS_OQUOTA_ENFD;
- } else if (!strcmp(this_char, MNTOPT_DMAPI)) {
- mp->m_flags |= XFS_MOUNT_DMAPI;
- } else if (!strcmp(this_char, MNTOPT_XDSM)) {
- mp->m_flags |= XFS_MOUNT_DMAPI;
- } else if (!strcmp(this_char, MNTOPT_DMI)) {
- mp->m_flags |= XFS_MOUNT_DMAPI;
} else if (!strcmp(this_char, MNTOPT_DELAYLOG)) {
mp->m_flags |= XFS_MOUNT_DELAYLOG;
cmn_err(CE_WARN,
@@ -430,12 +412,6 @@ xfs_parseargs(
return EINVAL;
}
- if ((mp->m_flags & XFS_MOUNT_DMAPI) && (!*mtpt || *mtpt[0] == '\0')) {
- printk("XFS: %s option needs the mount point option as well\n",
- MNTOPT_DMAPI);
- return EINVAL;
- }
-
if ((dsunit && !dswidth) || (!dsunit && dswidth)) {
cmn_err(CE_WARN,
"XFS: sunit and swidth must be specified together");
@@ -449,18 +425,6 @@ xfs_parseargs(
return EINVAL;
}
- /*
- * Applications using DMI filesystems often expect the
- * inode generation number to be monotonically increasing.
- * If we delete inode chunks we break this assumption, so
- * keep unused inode chunks on disk for DMI filesystems
- * until we come up with a better solution.
- * Note that if "ikeep" or "noikeep" mount options are
- * supplied, then they are honored.
- */
- if ((mp->m_flags & XFS_MOUNT_DMAPI) && dmapi_implies_ikeep)
- mp->m_flags |= XFS_MOUNT_IKEEP;
-
done:
if (!(mp->m_flags & XFS_MOUNT_NOALIGN)) {
/*
@@ -542,7 +506,6 @@ xfs_showargs(
{ XFS_MOUNT_OSYNCISOSYNC, "," MNTOPT_OSYNCISOSYNC },
{ XFS_MOUNT_ATTR2, "," MNTOPT_ATTR2 },
{ XFS_MOUNT_FILESTREAMS, "," MNTOPT_FILESTREAM },
- { XFS_MOUNT_DMAPI, "," MNTOPT_DMAPI },
{ XFS_MOUNT_GRPID, "," MNTOPT_GRPID },
{ XFS_MOUNT_DELAYLOG, "," MNTOPT_DELAYLOG },
{ 0, NULL }
@@ -1206,8 +1169,6 @@ xfs_fs_put_super(
xfs_sync_attr(mp, 0);
}
- XFS_SEND_PREUNMOUNT(mp);
-
/*
* Blow away any referenced inode in the filestreams cache.
* This can and will cause log traffic as inodes go inactive
@@ -1217,14 +1178,11 @@ xfs_fs_put_super(
XFS_bflush(mp->m_ddev_targp);
- XFS_SEND_UNMOUNT(mp);
-
xfs_unmountfs(mp);
xfs_freesb(mp);
xfs_inode_shrinker_unregister(mp);
xfs_icsb_destroy_counters(mp);
xfs_close_devices(mp);
- xfs_dmops_put(mp);
xfs_free_fsname(mp);
kfree(mp);
}
@@ -1542,7 +1500,6 @@ xfs_fs_fill_super(
struct inode *root;
struct xfs_mount *mp = NULL;
int flags = 0, error = ENOMEM;
- char *mtpt = NULL;
mp = kzalloc(sizeof(struct xfs_mount), GFP_KERNEL);
if (!mp)
@@ -1558,7 +1515,7 @@ xfs_fs_fill_super(
mp->m_super = sb;
sb->s_fs_info = mp;
- error = xfs_parseargs(mp, (char *)data, &mtpt);
+ error = xfs_parseargs(mp, (char *)data);
if (error)
goto out_free_fsname;
@@ -1570,16 +1527,12 @@ xfs_fs_fill_super(
#endif
sb->s_op = &xfs_super_operations;
- error = xfs_dmops_get(mp);
- if (error)
- goto out_free_fsname;
-
if (silent)
flags |= XFS_MFSI_QUIET;
error = xfs_open_devices(mp);
if (error)
- goto out_put_dmops;
+ goto out_free_fsname;
if (xfs_icsb_init_counters(mp))
mp->m_flags |= XFS_MOUNT_NO_PERCPU_SB;
@@ -1607,8 +1560,6 @@ xfs_fs_fill_super(
if (error)
goto out_filestream_unmount;
- XFS_SEND_MOUNT(mp, DM_RIGHT_NULL, mtpt, mp->m_fsname);
-
sb->s_magic = XFS_SB_MAGIC;
sb->s_blocksize = mp->m_sb.sb_blocksize;
sb->s_blocksize_bits = ffs(sb->s_blocksize) - 1;
@@ -1637,7 +1588,6 @@ xfs_fs_fill_super(
xfs_inode_shrinker_register(mp);
- kfree(mtpt);
return 0;
out_filestream_unmount:
@@ -1647,11 +1597,8 @@ xfs_fs_fill_super(
out_destroy_counters:
xfs_icsb_destroy_counters(mp);
xfs_close_devices(mp);
- out_put_dmops:
- xfs_dmops_put(mp);
out_free_fsname:
xfs_free_fsname(mp);
- kfree(mtpt);
kfree(mp);
out:
return -error;
Index: xfs/fs/xfs/xfs_bmap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_bmap.c 2010-05-25 11:14:59.261004260 +0200
@@ -34,7 +34,6 @@
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_ialloc.h"
#include "xfs_itable.h"
@@ -5605,28 +5604,6 @@ xfs_getbmap(
prealloced = 0;
fixlen = 1LL << 32;
} else {
- /*
- * If the BMV_IF_NO_DMAPI_READ interface bit specified, do
- * not generate a DMAPI read event. Otherwise, if the
- * DM_EVENT_READ bit is set for the file, generate a read
- * event in order that the DMAPI application may do its thing
- * before we return the extents. Usually this means restoring
- * user file data to regions of the file that look like holes.
- *
- * The "old behavior" (from XFS_IOC_GETBMAP) is to not specify
- * BMV_IF_NO_DMAPI_READ so that read events are generated.
- * If this were not true, callers of ioctl(XFS_IOC_GETBMAP)
- * could misinterpret holes in a DMAPI file as true holes,
- * when in fact they may represent offline user data.
- */
- if (DM_EVENT_ENABLED(ip, DM_EVENT_READ) &&
- !(iflags & BMV_IF_NO_DMAPI_READ)) {
- error = XFS_SEND_DATA(mp, DM_EVENT_READ, ip,
- 0, 0, 0, NULL);
- if (error)
- return XFS_ERROR(error);
- }
-
if (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS &&
ip->i_d.di_format != XFS_DINODE_FMT_BTREE &&
ip->i_d.di_format != XFS_DINODE_FMT_LOCAL)
Index: xfs/fs/xfs/xfs_rename.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rename.c 2010-05-22 00:39:04.104004455 +0200
+++ xfs/fs/xfs/xfs_rename.c 2010-05-25 11:14:59.266265051 +0200
@@ -24,7 +24,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
@@ -119,18 +118,6 @@ xfs_rename(
xfs_itrace_entry(src_dp);
xfs_itrace_entry(target_dp);
- if (DM_EVENT_ENABLED(src_dp, DM_EVENT_RENAME) ||
- DM_EVENT_ENABLED(target_dp, DM_EVENT_RENAME)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_RENAME,
- src_dp, DM_RIGHT_NULL,
- target_dp, DM_RIGHT_NULL,
- src_name->name, target_name->name,
- 0, 0, 0);
- if (error)
- return error;
- }
- /* Return through std_return after this point. */
-
new_parent = (src_dp != target_dp);
src_is_directory = ((src_ip->i_d.di_mode & S_IFMT) == S_IFDIR);
@@ -369,26 +356,13 @@ xfs_rename(
* trans_commit will unlock src_ip, target_ip & decrement
* the vnode references.
*/
- error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
-
- /* Fall through to std_return with error = 0 or errno from
- * xfs_trans_commit */
-std_return:
- if (DM_EVENT_ENABLED(src_dp, DM_EVENT_POSTRENAME) ||
- DM_EVENT_ENABLED(target_dp, DM_EVENT_POSTRENAME)) {
- (void) XFS_SEND_NAMESP (mp, DM_EVENT_POSTRENAME,
- src_dp, DM_RIGHT_NULL,
- target_dp, DM_RIGHT_NULL,
- src_name->name, target_name->name,
- 0, error, 0);
- }
- return error;
+ return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
abort_return:
cancel_flags |= XFS_TRANS_ABORT;
- /* FALLTHROUGH */
error_return:
xfs_bmap_cancel(&free_list);
xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
+ std_return:
+ return error;
}
Index: xfs/fs/xfs/xfs_vnodeops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_vnodeops.c 2010-05-22 00:40:37.991004874 +0200
+++ xfs/fs/xfs/xfs_vnodeops.c 2010-05-25 11:14:59.274016343 +0200
@@ -26,7 +26,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
@@ -143,16 +142,6 @@ xfs_setattr(
goto error_return;
}
} else {
- if (DM_EVENT_ENABLED(ip, DM_EVENT_TRUNCATE) &&
- !(flags & XFS_ATTR_DMI)) {
- int dmflags = AT_DELAY_FLAG(flags) | DM_SEM_FLAG_WR;
- code = XFS_SEND_DATA(mp, DM_EVENT_TRUNCATE, ip,
- iattr->ia_size, 0, dmflags, NULL);
- if (code) {
- lock_flags = 0;
- goto error_return;
- }
- }
if (need_iolock)
lock_flags |= XFS_IOLOCK_EXCL;
}
@@ -470,17 +459,10 @@ xfs_setattr(
return XFS_ERROR(code);
}
- if (DM_EVENT_ENABLED(ip, DM_EVENT_ATTRIBUTE) &&
- !(flags & XFS_ATTR_DMI)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_ATTRIBUTE, ip, DM_RIGHT_NULL,
- NULL, DM_RIGHT_NULL, NULL, NULL,
- 0, 0, AT_DELAY_FLAG(flags));
- }
return 0;
abort_return:
commit_flags |= XFS_TRANS_ABORT;
- /* FALLTHROUGH */
error_return:
xfs_qm_dqrele(udqp);
xfs_qm_dqrele(gdqp);
@@ -1060,9 +1042,6 @@ xfs_inactive(
mp = ip->i_mount;
- if (ip->i_d.di_nlink == 0 && DM_EVENT_ENABLED(ip, DM_EVENT_DESTROY))
- XFS_SEND_DESTROY(mp, ip, DM_RIGHT_NULL);
-
error = 0;
/* If this is a read-only mount, don't do this (would generate I/O) */
@@ -1314,16 +1293,6 @@ xfs_create(
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
- if (DM_EVENT_ENABLED(dp, DM_EVENT_CREATE)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_CREATE,
- dp, DM_RIGHT_NULL, NULL,
- DM_RIGHT_NULL, name->name, NULL,
- mode, 0, 0);
-
- if (error)
- return error;
- }
-
if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
prid = dp->i_d.di_projid;
else
@@ -1487,16 +1456,7 @@ xfs_create(
xfs_qm_dqrele(gdqp);
*ipp = ip;
-
- /* Fallthrough to std_return with error = 0 */
- std_return:
- if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTCREATE)) {
- XFS_SEND_NAMESP(mp, DM_EVENT_POSTCREATE, dp, DM_RIGHT_NULL,
- ip, DM_RIGHT_NULL, name->name, NULL, mode,
- error, 0);
- }
-
- return error;
+ return 0;
out_bmap_cancel:
xfs_bmap_cancel(&free_list);
@@ -1510,8 +1470,8 @@ xfs_create(
if (unlock_dp_on_error)
xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
- goto std_return;
+ std_return:
+ return error;
out_abort_rele:
/*
@@ -1732,14 +1692,6 @@ xfs_remove(
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
- if (DM_EVENT_ENABLED(dp, DM_EVENT_REMOVE)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_REMOVE, dp, DM_RIGHT_NULL,
- NULL, DM_RIGHT_NULL, name->name, NULL,
- ip->i_d.di_mode, 0, 0);
- if (error)
- return error;
- }
-
error = xfs_qm_dqattach(dp, 0);
if (error)
goto std_return;
@@ -1877,21 +1829,15 @@ xfs_remove(
if (!is_dir && link_zero && xfs_inode_is_filestream(ip))
xfs_filestream_deassociate(ip);
- std_return:
- if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTREMOVE)) {
- XFS_SEND_NAMESP(mp, DM_EVENT_POSTREMOVE, dp, DM_RIGHT_NULL,
- NULL, DM_RIGHT_NULL, name->name, NULL,
- ip->i_d.di_mode, error, 0);
- }
-
- return error;
+ return 0;
out_bmap_cancel:
xfs_bmap_cancel(&free_list);
cancel_flags |= XFS_TRANS_ABORT;
out_trans_cancel:
xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
+ std_return:
+ return error;
}
int
@@ -1917,17 +1863,6 @@ xfs_link(
if (XFS_FORCED_SHUTDOWN(mp))
return XFS_ERROR(EIO);
- if (DM_EVENT_ENABLED(tdp, DM_EVENT_LINK)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_LINK,
- tdp, DM_RIGHT_NULL,
- sip, DM_RIGHT_NULL,
- target_name->name, NULL, 0, 0, 0);
- if (error)
- return error;
- }
-
- /* Return through std_return after this point. */
-
error = xfs_qm_dqattach(sip, 0);
if (error)
goto std_return;
@@ -2014,27 +1949,14 @@ xfs_link(
goto abort_return;
}
- error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
- if (error)
- goto std_return;
-
- /* Fall through to std_return with error = 0. */
-std_return:
- if (DM_EVENT_ENABLED(sip, DM_EVENT_POSTLINK)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTLINK,
- tdp, DM_RIGHT_NULL,
- sip, DM_RIGHT_NULL,
- target_name->name, NULL, 0, error, 0);
- }
- return error;
+ return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
abort_return:
cancel_flags |= XFS_TRANS_ABORT;
- /* FALLTHROUGH */
-
error_return:
xfs_trans_cancel(tp, cancel_flags);
- goto std_return;
+ std_return:
+ return error;
}
int
@@ -2086,17 +2008,6 @@ xfs_symlink(
if (pathlen >= MAXPATHLEN) /* total string too long */
return XFS_ERROR(ENAMETOOLONG);
- if (DM_EVENT_ENABLED(dp, DM_EVENT_SYMLINK)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_SYMLINK, dp,
- DM_RIGHT_NULL, NULL, DM_RIGHT_NULL,
- link_name->name,
- (unsigned char *)target_path, 0, 0, 0);
- if (error)
- return error;
- }
-
- /* Return through std_return after this point. */
-
udqp = gdqp = NULL;
if (dp->i_d.di_flags & XFS_DIFLAG_PROJINHERIT)
prid = dp->i_d.di_projid;
@@ -2278,21 +2189,8 @@ xfs_symlink(
xfs_qm_dqrele(udqp);
xfs_qm_dqrele(gdqp);
- /* Fall through to std_return with error = 0 or errno from
- * xfs_trans_commit */
-std_return:
- if (DM_EVENT_ENABLED(dp, DM_EVENT_POSTSYMLINK)) {
- (void) XFS_SEND_NAMESP(mp, DM_EVENT_POSTSYMLINK,
- dp, DM_RIGHT_NULL,
- error ? NULL : ip,
- DM_RIGHT_NULL, link_name->name,
- (unsigned char *)target_path,
- 0, error, 0);
- }
-
- if (!error)
- *ipp = ip;
- return error;
+ *ipp = ip;
+ return 0;
error2:
IRELE(ip);
@@ -2306,8 +2204,8 @@ std_return:
if (unlock_dp_on_error)
xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
- goto std_return;
+ std_return:
+ return error;
}
int
@@ -2412,25 +2310,9 @@ xfs_alloc_file_space(
startoffset_fsb = XFS_B_TO_FSBT(mp, offset);
allocatesize_fsb = XFS_B_TO_FSB(mp, count);
- /* Generate a DMAPI event if needed. */
- if (alloc_type != 0 && offset < ip->i_size &&
- (attr_flags & XFS_ATTR_DMI) == 0 &&
- DM_EVENT_ENABLED(ip, DM_EVENT_WRITE)) {
- xfs_off_t end_dmi_offset;
-
- end_dmi_offset = offset+len;
- if (end_dmi_offset > ip->i_size)
- end_dmi_offset = ip->i_size;
- error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, ip, offset,
- end_dmi_offset - offset, 0, NULL);
- if (error)
- return error;
- }
-
/*
* Allocate file space until done or until there is an error
*/
-retry:
while (allocatesize_fsb && !error) {
xfs_fileoff_t s, e;
@@ -2527,17 +2409,6 @@ retry:
startoffset_fsb += allocated_fsb;
allocatesize_fsb -= allocated_fsb;
}
-dmapi_enospc_check:
- if (error == ENOSPC && (attr_flags & XFS_ATTR_DMI) == 0 &&
- DM_EVENT_ENABLED(ip, DM_EVENT_NOSPACE)) {
- error = XFS_SEND_NAMESP(mp, DM_EVENT_NOSPACE,
- ip, DM_RIGHT_NULL,
- ip, DM_RIGHT_NULL,
- NULL, NULL, 0, 0, 0); /* Delay flag intentionally unused */
- if (error == 0)
- goto retry; /* Maybe DMAPI app. has made space */
- /* else fall through with error from XFS_SEND_DATA */
- }
return error;
@@ -2548,7 +2419,7 @@ error0: /* Cancel bmap, unlock inode, un
error1: /* Just cancel transaction */
xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES | XFS_TRANS_ABORT);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
- goto dmapi_enospc_check;
+ return error;
}
/*
@@ -2661,7 +2532,6 @@ xfs_free_file_space(
{
int committed;
int done;
- xfs_off_t end_dmi_offset;
xfs_fileoff_t endoffset_fsb;
int error;
xfs_fsblock_t firstfsb;
@@ -2691,19 +2561,7 @@ xfs_free_file_space(
return error;
rt = XFS_IS_REALTIME_INODE(ip);
startoffset_fsb = XFS_B_TO_FSB(mp, offset);
- end_dmi_offset = offset + len;
- endoffset_fsb = XFS_B_TO_FSBT(mp, end_dmi_offset);
-
- if (offset < ip->i_size && (attr_flags & XFS_ATTR_DMI) == 0 &&
- DM_EVENT_ENABLED(ip, DM_EVENT_WRITE)) {
- if (end_dmi_offset > ip->i_size)
- end_dmi_offset = ip->i_size;
- error = XFS_SEND_DATA(mp, DM_EVENT_WRITE, ip,
- offset, end_dmi_offset - offset,
- AT_DELAY_FLAG(attr_flags), NULL);
- if (error)
- return error;
- }
+ endoffset_fsb = XFS_B_TO_FSBT(mp, offset + len);
if (attr_flags & XFS_ATTR_NOLOCK)
need_iolock = 0;
Index: xfs/fs/xfs/Makefile
===================================================================
--- xfs.orig/fs/xfs/Makefile 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/Makefile 2010-05-25 11:14:59.282031778 +0200
@@ -90,8 +90,7 @@ xfs-y += xfs_alloc.o \
xfs_trans_item.o \
xfs_utils.o \
xfs_vnodeops.o \
- xfs_rw.o \
- xfs_dmops.o
+ xfs_rw.o
xfs-$(CONFIG_XFS_TRACE) += xfs_btree_trace.o
Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c 2010-05-25 11:14:59.286031150 +0200
@@ -23,7 +23,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_trans.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/linux-2.6/xfs_buf.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_buf.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_buf.c 2010-05-25 11:14:59.297023956 +0200
@@ -39,7 +39,6 @@
#include "xfs_inum.h"
#include "xfs_log.h"
#include "xfs_ag.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_trace.h"
Index: xfs/fs/xfs/linux-2.6/xfs_export.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_export.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_export.c 2010-05-25 11:14:59.303023816 +0200
@@ -23,7 +23,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_export.h"
#include "xfs_vnodeops.h"
Index: xfs/fs/xfs/linux-2.6/xfs_fs_subr.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_fs_subr.c 2010-05-22 00:40:37.890004594 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_fs_subr.c 2010-05-25 11:14:59.310005937 +0200
@@ -21,10 +21,6 @@
#include "xfs_inode.h"
#include "xfs_trace.h"
-int fs_noerr(void) { return 0; }
-int fs_nosys(void) { return ENOSYS; }
-void fs_noval(void) { return; }
-
/*
* note: all filemap functions return negative error codes. These
* need to be inverted before returning to the xfs core functions.
Index: xfs/fs/xfs/linux-2.6/xfs_fs_subr.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_fs_subr.h 2010-05-22 00:39:04.000000000 +0200
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,25 +0,0 @@
-/*
- * Copyright (c) 2000,2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * 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.
- *
- * This program is distributed in the hope that it would 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 this program; if not, write the Free Software Foundation,
- * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- */
-#ifndef __XFS_FS_SUBR_H__
-#define __XFS_FS_SUBR_H__
-
-extern int fs_noerr(void);
-extern int fs_nosys(void);
-extern void fs_noval(void);
-
-#endif /* __XFS_FS_SUBR_H__ */
Index: xfs/fs/xfs/linux-2.6/xfs_ioctl32.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_ioctl32.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_ioctl32.c 2010-05-25 11:14:59.317003422 +0200
@@ -29,7 +29,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_attr_sf.h"
Index: xfs/fs/xfs/linux-2.6/xfs_iops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_iops.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_iops.c 2010-05-25 11:14:59.319042045 +0200
@@ -26,7 +26,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/linux-2.6/xfs_linux.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_linux.h 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_linux.h 2010-05-25 11:14:59.324043582 +0200
@@ -87,7 +87,6 @@
#include <xfs_aops.h>
#include <xfs_super.h>
#include <xfs_globals.h>
-#include <xfs_fs_subr.h>
#include <xfs_buf.h>
/*
Index: xfs/fs/xfs/linux-2.6/xfs_quotaops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_quotaops.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_quotaops.c 2010-05-25 11:14:59.329274410 +0200
@@ -16,7 +16,6 @@
* Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "xfs.h"
-#include "xfs_dmapi.h"
#include "xfs_sb.h"
#include "xfs_inum.h"
#include "xfs_log.h"
Index: xfs/fs/xfs/linux-2.6/xfs_sync.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_sync.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_sync.c 2010-05-25 11:14:59.335006286 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/linux-2.6/xfs_trace.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_trace.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_trace.c 2010-05-25 11:14:59.338005727 +0200
@@ -34,7 +34,6 @@
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_ialloc.h"
#include "xfs_itable.h"
Index: xfs/fs/xfs/quota/xfs_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_dquot.c 2010-05-25 11:14:59.344006076 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/quota/xfs_dquot_item.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot_item.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_dquot_item.c 2010-05-25 11:14:59.350033035 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/quota/xfs_qm.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_qm.c 2010-05-25 11:14:59.357004121 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/quota/xfs_qm_bhv.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm_bhv.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_qm_bhv.c 2010-05-25 11:14:59.361013061 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/quota/xfs_qm_stats.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm_stats.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_qm_stats.c 2010-05-25 11:14:59.363010267 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/quota/xfs_qm_syscalls.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm_syscalls.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_qm_syscalls.c 2010-05-25 11:14:59.368004121 +0200
@@ -28,7 +28,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/quota/xfs_trans_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_trans_dquot.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_trans_dquot.c 2010-05-25 11:14:59.372003632 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/support/debug.c
===================================================================
--- xfs.orig/fs/xfs/support/debug.c 2010-05-22 00:39:04.000000000 +0200
+++ xfs/fs/xfs/support/debug.c 2010-05-25 11:14:59.376020813 +0200
@@ -22,7 +22,6 @@
#include "xfs_sb.h"
#include "xfs_inum.h"
#include "xfs_ag.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_error.h"
Index: xfs/fs/xfs/xfs_alloc.c
===================================================================
--- xfs.orig/fs/xfs/xfs_alloc.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_alloc.c 2010-05-25 11:14:59.381273222 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_alloc_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_alloc_btree.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_alloc_btree.c 2010-05-25 11:14:59.386036737 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_attr.c
===================================================================
--- xfs.orig/fs/xfs/xfs_attr.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_attr.c 2010-05-25 11:14:59.390255762 +0200
@@ -26,7 +26,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_attr_leaf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_attr_leaf.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_attr_leaf.c 2010-05-25 11:14:59.398255692 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_bmap_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap_btree.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_bmap_btree.c 2010-05-25 11:14:59.404255552 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_btree.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_btree.c 2010-05-25 11:14:59.410003911 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_buf_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_buf_item.c 2010-05-25 11:14:59.418034083 +0200
@@ -24,7 +24,6 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_buf_item.h"
#include "xfs_trans_priv.h"
Index: xfs/fs/xfs/xfs_da_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_da_btree.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_da_btree.c 2010-05-25 11:14:59.424006216 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_dfrag.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dfrag.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_dfrag.c 2010-05-25 11:14:59.430031918 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_dir2.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_dir2.c 2010-05-25 11:14:59.435037366 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_dir2_block.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_block.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_dir2_block.c 2010-05-25 11:14:59.440014527 +0200
@@ -24,7 +24,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_dir2_data.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_data.c 2010-05-22 00:39:04.000000000 +0200
+++ xfs/fs/xfs/xfs_dir2_data.c 2010-05-25 11:14:59.442011873 +0200
@@ -24,7 +24,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_dir2_leaf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_leaf.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_dir2_leaf.c 2010-05-25 11:14:59.454006216 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_dir2_node.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_node.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_dir2_node.c 2010-05-25 11:14:59.671255345 +0200
@@ -24,7 +24,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_dir2_sf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_sf.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_dir2_sf.c 2010-05-25 11:14:59.678033455 +0200
@@ -24,7 +24,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_dmapi.h
===================================================================
--- xfs.orig/fs/xfs/xfs_dmapi.h 2010-05-22 00:39:04.000000000 +0200
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,170 +0,0 @@
-/*
- * Copyright (c) 2000-2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * 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.
- *
- * This program is distributed in the hope that it would 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 this program; if not, write the Free Software Foundation,
- * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- */
-#ifndef __XFS_DMAPI_H__
-#define __XFS_DMAPI_H__
-
-/* Values used to define the on-disk version of dm_attrname_t. All
- * on-disk attribute names start with the 8-byte string "SGI_DMI_".
- *
- * In the on-disk inode, DMAPI attribute names consist of the user-provided
- * name with the DMATTR_PREFIXSTRING pre-pended. This string must NEVER be
- * changed.
- */
-
-#define DMATTR_PREFIXLEN 8
-#define DMATTR_PREFIXSTRING "SGI_DMI_"
-
-typedef enum {
- DM_EVENT_INVALID = -1,
- DM_EVENT_CANCEL = 0, /* not supported */
- DM_EVENT_MOUNT = 1,
- DM_EVENT_PREUNMOUNT = 2,
- DM_EVENT_UNMOUNT = 3,
- DM_EVENT_DEBUT = 4, /* not supported */
- DM_EVENT_CREATE = 5,
- DM_EVENT_CLOSE = 6, /* not supported */
- DM_EVENT_POSTCREATE = 7,
- DM_EVENT_REMOVE = 8,
- DM_EVENT_POSTREMOVE = 9,
- DM_EVENT_RENAME = 10,
- DM_EVENT_POSTRENAME = 11,
- DM_EVENT_LINK = 12,
- DM_EVENT_POSTLINK = 13,
- DM_EVENT_SYMLINK = 14,
- DM_EVENT_POSTSYMLINK = 15,
- DM_EVENT_READ = 16,
- DM_EVENT_WRITE = 17,
- DM_EVENT_TRUNCATE = 18,
- DM_EVENT_ATTRIBUTE = 19,
- DM_EVENT_DESTROY = 20,
- DM_EVENT_NOSPACE = 21,
- DM_EVENT_USER = 22,
- DM_EVENT_MAX = 23
-} dm_eventtype_t;
-#define HAVE_DM_EVENTTYPE_T
-
-typedef enum {
- DM_RIGHT_NULL,
- DM_RIGHT_SHARED,
- DM_RIGHT_EXCL
-} dm_right_t;
-#define HAVE_DM_RIGHT_T
-
-/* Defines for determining if an event message should be sent. */
-#ifdef HAVE_DMAPI
-#define DM_EVENT_ENABLED(ip, event) ( \
- unlikely ((ip)->i_mount->m_flags & XFS_MOUNT_DMAPI) && \
- ( ((ip)->i_d.di_dmevmask & (1 << event)) || \
- ((ip)->i_mount->m_dmevmask & (1 << event)) ) \
- )
-#else
-#define DM_EVENT_ENABLED(ip, event) (0)
-#endif
-
-#define DM_XFS_VALID_FS_EVENTS ( \
- (1 << DM_EVENT_PREUNMOUNT) | \
- (1 << DM_EVENT_UNMOUNT) | \
- (1 << DM_EVENT_NOSPACE) | \
- (1 << DM_EVENT_DEBUT) | \
- (1 << DM_EVENT_CREATE) | \
- (1 << DM_EVENT_POSTCREATE) | \
- (1 << DM_EVENT_REMOVE) | \
- (1 << DM_EVENT_POSTREMOVE) | \
- (1 << DM_EVENT_RENAME) | \
- (1 << DM_EVENT_POSTRENAME) | \
- (1 << DM_EVENT_LINK) | \
- (1 << DM_EVENT_POSTLINK) | \
- (1 << DM_EVENT_SYMLINK) | \
- (1 << DM_EVENT_POSTSYMLINK) | \
- (1 << DM_EVENT_ATTRIBUTE) | \
- (1 << DM_EVENT_DESTROY) )
-
-/* Events valid in dm_set_eventlist() when called with a file handle for
- a regular file or a symlink. These events are persistent.
-*/
-
-#define DM_XFS_VALID_FILE_EVENTS ( \
- (1 << DM_EVENT_ATTRIBUTE) | \
- (1 << DM_EVENT_DESTROY) )
-
-/* Events valid in dm_set_eventlist() when called with a file handle for
- a directory. These events are persistent.
-*/
-
-#define DM_XFS_VALID_DIRECTORY_EVENTS ( \
- (1 << DM_EVENT_CREATE) | \
- (1 << DM_EVENT_POSTCREATE) | \
- (1 << DM_EVENT_REMOVE) | \
- (1 << DM_EVENT_POSTREMOVE) | \
- (1 << DM_EVENT_RENAME) | \
- (1 << DM_EVENT_POSTRENAME) | \
- (1 << DM_EVENT_LINK) | \
- (1 << DM_EVENT_POSTLINK) | \
- (1 << DM_EVENT_SYMLINK) | \
- (1 << DM_EVENT_POSTSYMLINK) | \
- (1 << DM_EVENT_ATTRIBUTE) | \
- (1 << DM_EVENT_DESTROY) )
-
-/* Events supported by the XFS filesystem. */
-#define DM_XFS_SUPPORTED_EVENTS ( \
- (1 << DM_EVENT_MOUNT) | \
- (1 << DM_EVENT_PREUNMOUNT) | \
- (1 << DM_EVENT_UNMOUNT) | \
- (1 << DM_EVENT_NOSPACE) | \
- (1 << DM_EVENT_CREATE) | \
- (1 << DM_EVENT_POSTCREATE) | \
- (1 << DM_EVENT_REMOVE) | \
- (1 << DM_EVENT_POSTREMOVE) | \
- (1 << DM_EVENT_RENAME) | \
- (1 << DM_EVENT_POSTRENAME) | \
- (1 << DM_EVENT_LINK) | \
- (1 << DM_EVENT_POSTLINK) | \
- (1 << DM_EVENT_SYMLINK) | \
- (1 << DM_EVENT_POSTSYMLINK) | \
- (1 << DM_EVENT_READ) | \
- (1 << DM_EVENT_WRITE) | \
- (1 << DM_EVENT_TRUNCATE) | \
- (1 << DM_EVENT_ATTRIBUTE) | \
- (1 << DM_EVENT_DESTROY) )
-
-
-/*
- * Definitions used for the flags field on dm_send_*_event().
- */
-
-#define DM_FLAGS_NDELAY 0x001 /* return EAGAIN after dm_pending() */
-#define DM_FLAGS_UNWANTED 0x002 /* event not in fsys dm_eventset_t */
-#define DM_FLAGS_IMUX 0x004 /* thread holds i_mutex */
-#define DM_FLAGS_IALLOCSEM_RD 0x010 /* thread holds i_alloc_sem rd */
-#define DM_FLAGS_IALLOCSEM_WR 0x020 /* thread holds i_alloc_sem wr */
-
-/*
- * Pull in platform specific event flags defines
- */
-#include "xfs_dmapi_priv.h"
-
-/*
- * Macros to turn caller specified delay/block flags into
- * dm_send_xxxx_event flag DM_FLAGS_NDELAY.
- */
-
-#define FILP_DELAY_FLAG(filp) ((filp->f_flags&(O_NDELAY|O_NONBLOCK)) ? \
- DM_FLAGS_NDELAY : 0)
-#define AT_DELAY_FLAG(f) ((f & XFS_ATTR_NONBLOCK) ? DM_FLAGS_NDELAY : 0)
-
-#endif /* __XFS_DMAPI_H__ */
Index: xfs/fs/xfs/xfs_dmops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dmops.c 2010-05-22 00:39:04.000000000 +0200
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,55 +0,0 @@
-/*
- * Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * 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.
- *
- * This program is distributed in the hope that it would 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 this program; if not, write the Free Software Foundation,
- * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_types.h"
-#include "xfs_log.h"
-#include "xfs_trans.h"
-#include "xfs_sb.h"
-#include "xfs_dmapi.h"
-#include "xfs_inum.h"
-#include "xfs_ag.h"
-#include "xfs_mount.h"
-
-
-static struct xfs_dmops xfs_dmcore_stub = {
- .xfs_send_data = (xfs_send_data_t)fs_nosys,
- .xfs_send_mmap = (xfs_send_mmap_t)fs_noerr,
- .xfs_send_destroy = (xfs_send_destroy_t)fs_nosys,
- .xfs_send_namesp = (xfs_send_namesp_t)fs_nosys,
- .xfs_send_mount = (xfs_send_mount_t)fs_nosys,
- .xfs_send_unmount = (xfs_send_unmount_t)fs_noerr,
-};
-
-int
-xfs_dmops_get(struct xfs_mount *mp)
-{
- if (mp->m_flags & XFS_MOUNT_DMAPI) {
- cmn_err(CE_WARN,
- "XFS: dmapi support not available in this kernel.");
- return EINVAL;
- }
-
- mp->m_dm_ops = &xfs_dmcore_stub;
- return 0;
-}
-
-void
-xfs_dmops_put(struct xfs_mount *mp)
-{
-}
Index: xfs/fs/xfs/xfs_error.c
===================================================================
--- xfs.orig/fs/xfs/xfs_error.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_error.c 2010-05-25 11:14:59.684255901 +0200
@@ -24,7 +24,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_dir2_sf.h"
Index: xfs/fs/xfs/xfs_extfree_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_extfree_item.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_extfree_item.c 2010-05-25 11:14:59.687282511 +0200
@@ -24,7 +24,6 @@
#include "xfs_buf_item.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_trans_priv.h"
#include "xfs_extfree_item.h"
Index: xfs/fs/xfs/xfs_filestream.c
===================================================================
--- xfs.orig/fs/xfs/xfs_filestream.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_filestream.c 2010-05-25 11:14:59.694256111 +0200
@@ -24,7 +24,6 @@
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_ag.h"
-#include "xfs_dmapi.h"
#include "xfs_log.h"
#include "xfs_trans.h"
#include "xfs_sb.h"
Index: xfs/fs/xfs/xfs_fsops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_fsops.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_fsops.c 2010-05-25 11:14:59.700275318 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_ialloc.c
===================================================================
--- xfs.orig/fs/xfs/xfs_ialloc.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_ialloc.c 2010-05-25 11:14:59.705254086 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_ialloc_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_ialloc_btree.c 2010-05-22 00:39:04.000000000 +0200
+++ xfs/fs/xfs/xfs_ialloc_btree.c 2010-05-25 11:14:59.709254155 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_iget.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iget.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_iget.c 2010-05-25 11:14:59.713253736 +0200
@@ -26,7 +26,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_inode.c 2010-05-25 11:14:59.716253946 +0200
@@ -28,7 +28,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_inode_item.c 2010-05-25 11:14:59.722275178 +0200
@@ -26,7 +26,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_trans_priv.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_iomap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_iomap.c 2010-05-25 11:14:59.733005937 +0200
@@ -25,7 +25,6 @@
#include "xfs_ag.h"
#include "xfs_dir2.h"
#include "xfs_alloc.h"
-#include "xfs_dmapi.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_itable.c
===================================================================
--- xfs.orig/fs/xfs/xfs_itable.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_itable.c 2010-05-25 11:14:59.740006635 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_log.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_log.c 2010-05-25 11:14:59.753006007 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_log_priv.h"
Index: xfs/fs/xfs/xfs_log_recover.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log_recover.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_log_recover.c 2010-05-25 11:14:59.762003492 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_bmap_btree.h"
Index: xfs/fs/xfs/xfs_mount.c
===================================================================
--- xfs.orig/fs/xfs/xfs_mount.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_mount.c 2010-05-25 11:14:59.764011664 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_mount.h
===================================================================
--- xfs.orig/fs/xfs/xfs_mount.h 2010-05-24 23:11:41.025253988 +0200
+++ xfs/fs/xfs/xfs_mount.h 2010-05-25 11:15:46.687030871 +0200
@@ -66,65 +66,6 @@ struct xfs_nameops;
struct xfs_ail;
struct xfs_quotainfo;
-
-/*
- * Prototypes and functions for the Data Migration subsystem.
- */
-
-typedef int (*xfs_send_data_t)(int, struct xfs_inode *,
- xfs_off_t, size_t, int, int *);
-typedef int (*xfs_send_mmap_t)(struct vm_area_struct *, uint);
-typedef int (*xfs_send_destroy_t)(struct xfs_inode *, dm_right_t);
-typedef int (*xfs_send_namesp_t)(dm_eventtype_t, struct xfs_mount *,
- struct xfs_inode *, dm_right_t,
- struct xfs_inode *, dm_right_t,
- const unsigned char *, const unsigned char *,
- mode_t, int, int);
-typedef int (*xfs_send_mount_t)(struct xfs_mount *, dm_right_t,
- char *, char *);
-typedef void (*xfs_send_unmount_t)(struct xfs_mount *, struct xfs_inode *,
- dm_right_t, mode_t, int, int);
-
-typedef struct xfs_dmops {
- xfs_send_data_t xfs_send_data;
- xfs_send_mmap_t xfs_send_mmap;
- xfs_send_destroy_t xfs_send_destroy;
- xfs_send_namesp_t xfs_send_namesp;
- xfs_send_mount_t xfs_send_mount;
- xfs_send_unmount_t xfs_send_unmount;
-} xfs_dmops_t;
-
-#define XFS_DMAPI_UNMOUNT_FLAGS(mp) \
- (((mp)->m_dmevmask & (1 << DM_EVENT_UNMOUNT)) ? 0 : DM_FLAGS_UNWANTED)
-
-#define XFS_SEND_DATA(mp, ev,ip,off,len,fl,lock) \
- (*(mp)->m_dm_ops->xfs_send_data)(ev,ip,off,len,fl,lock)
-#define XFS_SEND_MMAP(mp, vma,fl) \
- (*(mp)->m_dm_ops->xfs_send_mmap)(vma,fl)
-#define XFS_SEND_DESTROY(mp, ip,right) \
- (*(mp)->m_dm_ops->xfs_send_destroy)(ip,right)
-#define XFS_SEND_NAMESP(mp, ev,b1,r1,b2,r2,n1,n2,mode,rval,fl) \
- (*(mp)->m_dm_ops->xfs_send_namesp)(ev,NULL,b1,r1,b2,r2,n1,n2,mode,rval,fl)
-#define XFS_SEND_MOUNT(mp,right,path,name) \
- (*(mp)->m_dm_ops->xfs_send_mount)(mp,right,path,name)
-#define XFS_SEND_PREUNMOUNT(mp) \
-do { \
- if (mp->m_flags & XFS_MOUNT_DMAPI) { \
- (*(mp)->m_dm_ops->xfs_send_namesp)(DM_EVENT_PREUNMOUNT, mp, \
- (mp)->m_rootip, DM_RIGHT_NULL, \
- (mp)->m_rootip, DM_RIGHT_NULL, \
- NULL, NULL, 0, 0, XFS_DMAPI_UNMOUNT_FLAGS(mp)); \
- } \
-} while (0)
-#define XFS_SEND_UNMOUNT(mp) \
-do { \
- if (mp->m_flags & XFS_MOUNT_DMAPI) { \
- (*(mp)->m_dm_ops->xfs_send_unmount)(mp, (mp)->m_rootip, \
- DM_RIGHT_NULL, 0, 0, XFS_DMAPI_UNMOUNT_FLAGS(mp)); \
- } \
-} while (0)
-
-
#ifdef HAVE_PERCPU_SB
/*
@@ -241,8 +182,6 @@ typedef struct xfs_mount {
uint m_chsize; /* size of next field */
struct xfs_chash *m_chash; /* fs private inode per-cluster
* hash table */
- struct xfs_dmops *m_dm_ops; /* vector of DMI ops */
- struct xfs_qmops *m_qm_ops; /* vector of XQM ops */
atomic_t m_active_trans; /* number trans frozen */
#ifdef HAVE_PERCPU_SB
xfs_icsb_cnts_t __percpu *m_sb_cnts; /* per-cpu superblock counters */
@@ -269,7 +208,6 @@ typedef struct xfs_mount {
must be synchronous except
for space allocations */
#define XFS_MOUNT_DELAYLOG (1ULL << 1) /* delayed logging is enabled */
-#define XFS_MOUNT_DMAPI (1ULL << 2) /* dmapi is enabled */
#define XFS_MOUNT_WAS_CLEAN (1ULL << 3)
#define XFS_MOUNT_FS_SHUTDOWN (1ULL << 4) /* atomic stop of all filesystem
operations, typically for
@@ -440,11 +378,6 @@ extern int xfs_sb_validate_fsb_count(str
extern int xfs_dev_is_read_only(struct xfs_mount *, char *);
-extern int xfs_dmops_get(struct xfs_mount *);
-extern void xfs_dmops_put(struct xfs_mount *);
-
-extern struct xfs_dmops xfs_dmcore_xfs;
-
#endif /* __KERNEL__ */
extern void xfs_mod_sb(struct xfs_trans *, __int64_t);
Index: xfs/fs/xfs/xfs_rtalloc.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rtalloc.c 2010-05-22 00:39:04.000000000 +0200
+++ xfs/fs/xfs/xfs_rtalloc.c 2010-05-25 11:14:59.778003841 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_rw.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rw.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_rw.c 2010-05-25 11:14:59.784026331 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_trans.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_trans.c 2010-05-25 11:14:59.797257578 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_da_btree.h"
Index: xfs/fs/xfs/xfs_trans_ail.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_ail.c 2010-05-22 00:40:37.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_ail.c 2010-05-25 11:14:59.802255482 +0200
@@ -24,7 +24,6 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_trans_priv.h"
#include "xfs_error.h"
Index: xfs/fs/xfs/xfs_trans_buf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_buf.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_buf.c 2010-05-25 11:14:59.808006007 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_trans_extfree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_extfree.c 2010-05-22 00:39:04.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_extfree.c 2010-05-25 11:14:59.812006356 +0200
@@ -23,7 +23,6 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_trans_priv.h"
#include "xfs_extfree_item.h"
Index: xfs/fs/xfs/xfs_trans_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_inode.c 2010-05-22 00:39:04.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_inode.c 2010-05-25 11:14:59.816006845 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
Index: xfs/fs/xfs/xfs_utils.c
===================================================================
--- xfs.orig/fs/xfs/xfs_utils.c 2010-05-22 00:39:04.000000000 +0200
+++ xfs/fs/xfs/xfs_utils.c 2010-05-25 11:14:59.820006076 +0200
@@ -25,7 +25,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_dir2_sf.h"
Index: xfs/fs/xfs/xfs_trans_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_item.c 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_item.c 2010-05-25 11:14:59.823006076 +0200
@@ -28,7 +28,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
STATIC int xfs_trans_unlock_chunk(xfs_log_item_chunk_t *,
Index: xfs/fs/xfs/linux-2.6/xfs_dmapi_priv.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_dmapi_priv.h 2010-05-22 00:39:04.000000000 +0200
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,28 +0,0 @@
-/*
- * Copyright (c) 2000-2006 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * 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.
- *
- * This program is distributed in the hope that it would 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 this program; if not, write the Free Software Foundation,
- * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- */
-#ifndef __XFS_DMAPI_PRIV_H__
-#define __XFS_DMAPI_PRIV_H__
-
-/*
- * Based on IO_ISDIRECT, decide which i_ flag is set.
- */
-#define DM_SEM_FLAG_RD(ioflags) (((ioflags) & IO_ISDIRECT) ? \
- DM_FLAGS_IMUX : 0)
-#define DM_SEM_FLAG_WR (DM_FLAGS_IALLOCSEM_WR | DM_FLAGS_IMUX)
-
-#endif /*__XFS_DMAPI_PRIV_H__*/
Index: xfs/fs/xfs/linux-2.6/xfs_super.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.h 2010-05-22 00:39:04.089004246 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.h 2010-05-25 11:14:59.826005657 +0200
@@ -56,12 +56,6 @@ extern void xfs_qm_exit(void);
# define XFS_BIGFS_STRING
#endif
-#ifdef CONFIG_XFS_DMAPI
-# define XFS_DMAPI_STRING "dmapi support, "
-#else
-# define XFS_DMAPI_STRING
-#endif
-
#ifdef DEBUG
# define XFS_DBG_STRING "debug"
#else
@@ -72,7 +66,6 @@ extern void xfs_qm_exit(void);
XFS_SECURITY_STRING \
XFS_REALTIME_STRING \
XFS_BIGFS_STRING \
- XFS_DMAPI_STRING \
XFS_DBG_STRING /* DBG must be last */
struct xfs_inode;
Index: xfs/fs/xfs/xfs_log_cil.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log_cil.c 2010-05-25 11:16:08.428003561 +0200
+++ xfs/fs/xfs/xfs_log_cil.c 2010-05-25 11:16:11.151344325 +0200
@@ -27,7 +27,6 @@
#include "xfs_sb.h"
#include "xfs_ag.h"
#include "xfs_dir2.h"
-#include "xfs_dmapi.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_alloc.h"
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 04/17] xfs: drop dmapi hooks
2010-05-31 16:07 ` [PATCH 04/17] xfs: drop dmapi hooks Christoph Hellwig
@ 2010-06-02 4:45 ` Dave Chinner
0 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 4:45 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:31PM -0400, Christoph Hellwig wrote:
> Dmapi support was never merged upstream, but we still have a lot of hooks
> bloating XFS for it, all over the fast pathes of the filesystem.
>
> This patch drops over 700 lines of dmapi overhead. If we'll ever get HSM
> support in mainline at least the namespace events can be done much saner
> in the VFS instead of the individual filesystem, so it's not like this
> is much help for future work.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Looks pretty good to me, and it runs through QA just fine.
There are still a couple of references to dmapi in comments,
but I don't think that's a big issue.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* [PATCH 05/17] xfs: remove unneeded #include statements
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (3 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 04/17] xfs: drop dmapi hooks Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 4:45 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 06/17] xfs: simplify log item descriptor tracking Christoph Hellwig
` (13 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-reduce-includes --]
[-- Type: text/plain, Size: 34482 bytes --]
Signed-off-by: Christoph Hellwig <hch@lst.de>
---
V1 -> V2
- updated to account for the delayed logging branch merge
Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c 2010-05-25 11:14:59.286031150 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c 2010-05-25 11:38:00.497004330 +0200
@@ -21,18 +21,12 @@
#include "xfs_inum.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_trans.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_alloc.h"
-#include "xfs_btree.h"
#include "xfs_error.h"
#include "xfs_rw.h"
#include "xfs_iomap.h"
Index: xfs/fs/xfs/linux-2.6/xfs_file.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_file.c 2010-05-25 11:14:59.236274548 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_file.c 2010-05-25 11:38:00.498004261 +0200
@@ -22,22 +22,15 @@
#include "xfs_inum.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_trans.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
#include "xfs_alloc.h"
-#include "xfs_btree.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dir2_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_bmap.h"
#include "xfs_error.h"
-#include "xfs_rw.h"
#include "xfs_vnodeops.h"
#include "xfs_da_btree.h"
#include "xfs_ioctl.h"
Index: xfs/fs/xfs/linux-2.6/xfs_ioctl.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_ioctl.c 2010-05-25 11:14:59.237275526 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_ioctl.c 2010-05-25 11:38:00.505033035 +0200
@@ -23,23 +23,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dir2_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_ioctl.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_rtalloc.h"
#include "xfs_itable.h"
#include "xfs_error.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_bmap.h"
#include "xfs_buf_item.h"
Index: xfs/fs/xfs/linux-2.6/xfs_ioctl32.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_ioctl32.c 2010-05-25 11:14:59.317003422 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_ioctl32.c 2010-05-25 11:38:00.512033733 +0200
@@ -28,11 +28,8 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dir2_sf.h"
#include "xfs_vnode.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
Index: xfs/fs/xfs/linux-2.6/xfs_iops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_iops.c 2010-05-25 11:14:59.319042045 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_iops.c 2010-05-25 11:38:00.519033454 +0200
@@ -24,20 +24,13 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c 2010-05-25 11:15:23.496005936 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.c 2010-05-25 11:38:00.525005726 +0200
@@ -30,8 +30,6 @@
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
@@ -42,7 +40,6 @@
#include "xfs_error.h"
#include "xfs_itable.h"
#include "xfs_fsops.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
Index: xfs/fs/xfs/linux-2.6/xfs_sync.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_sync.c 2010-05-25 11:14:59.335006286 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_sync.c 2010-05-25 11:38:00.531005657 +0200
@@ -24,24 +24,14 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_inode.h"
#include "xfs_dinode.h"
#include "xfs_error.h"
-#include "xfs_mru_cache.h"
#include "xfs_filestream.h"
#include "xfs_vnodeops.h"
-#include "xfs_utils.h"
-#include "xfs_buf_item.h"
#include "xfs_inode_item.h"
-#include "xfs_rw.h"
#include "xfs_quota.h"
#include "xfs_trace.h"
Index: xfs/fs/xfs/linux-2.6/xfs_trace.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_trace.c 2010-05-25 11:14:59.338005727 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_trace.c 2010-05-25 11:38:00.534004958 +0200
@@ -24,13 +24,10 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
Index: xfs/fs/xfs/quota/xfs_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot.c 2010-05-25 11:14:59.344006076 +0200
+++ xfs/fs/xfs/quota/xfs_dquot.c 2010-05-25 11:38:00.540006006 +0200
@@ -23,24 +23,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_space.h"
Index: xfs/fs/xfs/quota/xfs_dquot_item.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot_item.c 2010-05-25 11:14:59.350033035 +0200
+++ xfs/fs/xfs/quota/xfs_dquot_item.c 2010-05-25 11:38:00.546006565 +0200
@@ -23,24 +23,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_itable.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_priv.h"
Index: xfs/fs/xfs/quota/xfs_qm.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm.c 2010-05-25 11:14:59.357004121 +0200
+++ xfs/fs/xfs/quota/xfs_qm.c 2010-05-25 11:38:00.552267005 +0200
@@ -23,24 +23,18 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
#include "xfs_bmap.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_space.h"
Index: xfs/fs/xfs/quota/xfs_qm_bhv.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm_bhv.c 2010-05-25 11:14:59.361013061 +0200
+++ xfs/fs/xfs/quota/xfs_qm_bhv.c 2010-05-25 11:38:00.557255481 +0200
@@ -23,24 +23,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_ialloc.h"
#include "xfs_itable.h"
-#include "xfs_btree.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_qm.h"
Index: xfs/fs/xfs/quota/xfs_qm_stats.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm_stats.c 2010-05-25 11:14:59.363010267 +0200
+++ xfs/fs/xfs/quota/xfs_qm_stats.c 2010-05-25 11:38:00.560284396 +0200
@@ -23,24 +23,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_bmap.h"
-#include "xfs_btree.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_qm.h"
Index: xfs/fs/xfs/quota/xfs_qm_syscalls.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm_syscalls.c 2010-05-25 11:14:59.368004121 +0200
+++ xfs/fs/xfs/quota/xfs_qm_syscalls.c 2010-05-25 11:38:00.565285095 +0200
@@ -26,24 +26,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_bmap.h"
-#include "xfs_btree.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_utils.h"
Index: xfs/fs/xfs/quota/xfs_trans_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_trans_dquot.c 2010-05-25 11:14:59.372003632 +0200
+++ xfs/fs/xfs/quota/xfs_trans_dquot.c 2010-05-25 11:38:00.570255481 +0200
@@ -23,24 +23,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_attr_sf.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_ialloc.h"
#include "xfs_itable.h"
-#include "xfs_btree.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
#include "xfs_error.h"
-#include "xfs_rw.h"
#include "xfs_attr.h"
#include "xfs_buf_item.h"
#include "xfs_trans_priv.h"
Index: xfs/fs/xfs/xfs_alloc.c
===================================================================
--- xfs.orig/fs/xfs/xfs_alloc.c 2010-05-25 11:14:59.381273222 +0200
+++ xfs/fs/xfs/xfs_alloc.c 2010-05-25 11:38:00.578255621 +0200
@@ -24,17 +24,13 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_alloc.h"
#include "xfs_error.h"
#include "xfs_trace.h"
Index: xfs/fs/xfs/xfs_alloc_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_alloc_btree.c 2010-05-25 11:14:59.386036737 +0200
+++ xfs/fs/xfs/xfs_alloc_btree.c 2010-05-25 11:38:00.584254923 +0200
@@ -24,18 +24,14 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
#include "xfs_btree_trace.h"
-#include "xfs_ialloc.h"
#include "xfs_alloc.h"
#include "xfs_error.h"
#include "xfs_trace.h"
Index: xfs/fs/xfs/xfs_attr.c
===================================================================
--- xfs.orig/fs/xfs/xfs_attr.c 2010-05-25 11:14:59.390255762 +0200
+++ xfs/fs/xfs/xfs_attr.c 2010-05-25 11:38:00.592255901 +0200
@@ -25,18 +25,13 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_alloc.h"
-#include "xfs_btree.h"
#include "xfs_inode_item.h"
#include "xfs_bmap.h"
#include "xfs_attr.h"
Index: xfs/fs/xfs/xfs_attr_leaf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_attr_leaf.c 2010-05-25 11:14:59.398255692 +0200
+++ xfs/fs/xfs/xfs_attr_leaf.c 2010-05-25 11:38:00.599285444 +0200
@@ -24,7 +24,6 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
@@ -32,7 +31,6 @@
#include "xfs_ialloc_btree.h"
#include "xfs_alloc.h"
#include "xfs_btree.h"
-#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
Index: xfs/fs/xfs/xfs_bmap_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap_btree.c 2010-05-25 11:14:59.404255552 +0200
+++ xfs/fs/xfs/xfs_bmap_btree.c 2010-05-25 11:38:00.605255621 +0200
@@ -24,20 +24,16 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_alloc.h"
#include "xfs_btree.h"
#include "xfs_btree_trace.h"
-#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_bmap.h"
#include "xfs_error.h"
Index: xfs/fs/xfs/xfs_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_btree.c 2010-05-25 11:14:59.410003911 +0200
+++ xfs/fs/xfs/xfs_btree.c 2010-05-25 11:38:00.611255412 +0200
@@ -24,19 +24,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_btree.h"
#include "xfs_btree_trace.h"
-#include "xfs_ialloc.h"
#include "xfs_error.h"
#include "xfs_trace.h"
Index: xfs/fs/xfs/xfs_dfrag.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dfrag.c 2010-05-25 11:14:59.430031918 +0200
+++ xfs/fs/xfs/xfs_dfrag.c 2010-05-25 11:38:00.613284745 +0200
@@ -24,23 +24,15 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_bmap.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_dfrag.h"
#include "xfs_error.h"
-#include "xfs_rw.h"
#include "xfs_vnodeops.h"
#include "xfs_trace.h"
Index: xfs/fs/xfs/xfs_error.c
===================================================================
--- xfs.orig/fs/xfs/xfs_error.c 2010-05-25 11:14:59.684255901 +0200
+++ xfs/fs/xfs/xfs_error.c 2010-05-25 11:38:00.620282861 +0200
@@ -23,11 +23,8 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_utils.h"
Index: xfs/fs/xfs/xfs_filestream.c
===================================================================
--- xfs.orig/fs/xfs/xfs_filestream.c 2010-05-25 11:14:59.694256111 +0200
+++ xfs/fs/xfs/xfs_filestream.c 2010-05-25 11:38:00.627283908 +0200
@@ -18,9 +18,6 @@
#include "xfs.h"
#include "xfs_bmap_btree.h"
#include "xfs_inum.h"
-#include "xfs_dir2.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_ag.h"
Index: xfs/fs/xfs/xfs_fsops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_fsops.c 2010-05-25 11:14:59.700275318 +0200
+++ xfs/fs/xfs/xfs_fsops.c 2010-05-25 11:38:00.634282791 +0200
@@ -24,13 +24,10 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
Index: xfs/fs/xfs/xfs_ialloc.c
===================================================================
--- xfs.orig/fs/xfs/xfs_ialloc.c 2010-05-25 11:14:59.705254086 +0200
+++ xfs/fs/xfs/xfs_ialloc.c 2010-05-25 11:38:00.640284956 +0200
@@ -24,13 +24,10 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
Index: xfs/fs/xfs/xfs_ialloc_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_ialloc_btree.c 2010-05-25 11:14:59.709254155 +0200
+++ xfs/fs/xfs/xfs_ialloc_btree.c 2010-05-25 11:38:00.646284886 +0200
@@ -24,13 +24,10 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
Index: xfs/fs/xfs/xfs_iget.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iget.c 2010-05-25 11:14:59.713253736 +0200
+++ xfs/fs/xfs/xfs_iget.c 2010-05-25 11:38:00.653255482 +0200
@@ -25,13 +25,10 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
Index: xfs/fs/xfs/xfs_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.c 2010-05-25 11:14:59.716253946 +0200
+++ xfs/fs/xfs/xfs_inode.c 2010-05-25 11:38:00.661255413 +0200
@@ -27,12 +27,10 @@
#include "xfs_trans_priv.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
@@ -43,7 +41,6 @@
#include "xfs_alloc.h"
#include "xfs_ialloc.h"
#include "xfs_bmap.h"
-#include "xfs_rw.h"
#include "xfs_error.h"
#include "xfs_utils.h"
#include "xfs_quota.h"
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c 2010-05-25 11:14:59.722275178 +0200
+++ xfs/fs/xfs/xfs_inode_item.c 2010-05-25 11:38:00.664256321 +0200
@@ -22,23 +22,14 @@
#include "xfs_log.h"
#include "xfs_inum.h"
#include "xfs_trans.h"
-#include "xfs_buf_item.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_trans_priv.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
-#include "xfs_rw.h"
#include "xfs_error.h"
#include "xfs_trace.h"
Index: xfs/fs/xfs/xfs_iomap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.c 2010-05-25 11:14:59.733005937 +0200
+++ xfs/fs/xfs/xfs_iomap.c 2010-05-25 11:38:00.669284676 +0200
@@ -23,18 +23,14 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_alloc.h"
#include "xfs_quota.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_ialloc.h"
#include "xfs_btree.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
Index: xfs/fs/xfs/xfs_itable.c
===================================================================
--- xfs.orig/fs/xfs/xfs_itable.c 2010-05-25 11:14:59.740006635 +0200
+++ xfs/fs/xfs/xfs_itable.c 2010-05-25 11:38:00.674285794 +0200
@@ -24,13 +24,10 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_ialloc.h"
Index: xfs/fs/xfs/xfs_log.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log.c 2010-05-25 11:14:59.753006007 +0200
+++ xfs/fs/xfs/xfs_log.c 2010-05-25 11:38:00.683255622 +0200
@@ -24,7 +24,6 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_log_priv.h"
@@ -34,8 +33,6 @@
#include "xfs_ialloc_btree.h"
#include "xfs_log_recover.h"
#include "xfs_trans_priv.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_rw.h"
Index: xfs/fs/xfs/xfs_log_recover.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log_recover.c 2010-05-25 11:14:59.762003492 +0200
+++ xfs/fs/xfs/xfs_log_recover.c 2010-05-25 11:38:00.689033734 +0200
@@ -24,14 +24,11 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
Index: xfs/fs/xfs/xfs_mount.c
===================================================================
--- xfs.orig/fs/xfs/xfs_mount.c 2010-05-25 11:14:59.764011664 +0200
+++ xfs/fs/xfs/xfs_mount.c 2010-05-25 11:38:00.697283978 +0200
@@ -29,8 +29,6 @@
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
Index: xfs/fs/xfs/xfs_rename.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rename.c 2010-05-25 11:14:59.266265051 +0200
+++ xfs/fs/xfs/xfs_rename.c 2010-05-25 11:38:00.703005448 +0200
@@ -27,8 +27,6 @@
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
Index: xfs/fs/xfs/xfs_rtalloc.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rtalloc.c 2010-05-25 11:14:59.778003841 +0200
+++ xfs/fs/xfs/xfs_rtalloc.c 2010-05-25 11:38:00.706025842 +0200
@@ -27,14 +27,8 @@
#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_alloc.h"
#include "xfs_bmap.h"
#include "xfs_rtalloc.h"
Index: xfs/fs/xfs/xfs_rw.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rw.c 2010-05-25 11:14:59.784026331 +0200
+++ xfs/fs/xfs/xfs_rw.c 2010-05-25 11:38:00.711005378 +0200
@@ -24,26 +24,12 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
-#include "xfs_inode_item.h"
-#include "xfs_itable.h"
-#include "xfs_btree.h"
-#include "xfs_alloc.h"
-#include "xfs_ialloc.h"
-#include "xfs_attr.h"
-#include "xfs_bmap.h"
#include "xfs_error.h"
-#include "xfs_buf_item.h"
#include "xfs_rw.h"
-#include "xfs_trace.h"
/*
* Force a shutdown of the filesystem instantly while keeping
Index: xfs/fs/xfs/xfs_trans.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.c 2010-05-25 11:14:59.797257578 +0200
+++ xfs/fs/xfs/xfs_trans.c 2010-05-25 11:38:00.715067607 +0200
@@ -24,15 +24,12 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
Index: xfs/fs/xfs/xfs_trans_buf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_buf.c 2010-05-25 11:14:59.808006007 +0200
+++ xfs/fs/xfs/xfs_trans_buf.c 2010-05-25 11:38:00.720005588 +0200
@@ -24,13 +24,10 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_buf_item.h"
Index: xfs/fs/xfs/xfs_trans_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_inode.c 2010-05-25 11:14:59.816006845 +0200
+++ xfs/fs/xfs/xfs_trans_inode.c 2010-05-25 11:38:00.729033873 +0200
@@ -24,17 +24,13 @@
#include "xfs_trans.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
-#include "xfs_ialloc.h"
#include "xfs_trans_priv.h"
#include "xfs_inode_item.h"
Index: xfs/fs/xfs/xfs_utils.c
===================================================================
--- xfs.orig/fs/xfs/xfs_utils.c 2010-05-25 11:14:59.820006076 +0200
+++ xfs/fs/xfs/xfs_utils.c 2010-05-25 11:38:00.736033663 +0200
@@ -27,15 +27,12 @@
#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_bmap_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_bmap.h"
#include "xfs_error.h"
#include "xfs_quota.h"
-#include "xfs_rw.h"
#include "xfs_itable.h"
#include "xfs_utils.h"
Index: xfs/fs/xfs/xfs_vnodeops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_vnodeops.c 2010-05-25 11:14:59.274016343 +0200
+++ xfs/fs/xfs/xfs_vnodeops.c 2010-05-25 11:38:00.742005377 +0200
@@ -29,15 +29,11 @@
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
-#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_itable.h"
-#include "xfs_btree.h"
#include "xfs_ialloc.h"
#include "xfs_alloc.h"
#include "xfs_bmap.h"
Index: xfs/fs/xfs/xfs_bmap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.c 2010-05-25 11:14:59.261004260 +0200
+++ xfs/fs/xfs/xfs_bmap.c 2010-05-25 11:38:00.753005587 +0200
@@ -30,12 +30,10 @@
#include "xfs_alloc_btree.h"
#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_btree.h"
#include "xfs_mount.h"
-#include "xfs_ialloc.h"
#include "xfs_itable.h"
#include "xfs_dir2_data.h"
#include "xfs_dir2_leaf.h"
Index: xfs/fs/xfs/xfs_da_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_da_btree.c 2010-05-25 11:14:59.424006216 +0200
+++ xfs/fs/xfs/xfs_da_btree.c 2010-05-25 11:38:00.760005726 +0200
@@ -28,15 +28,11 @@
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
-#include "xfs_alloc_btree.h"
-#include "xfs_ialloc_btree.h"
#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
#include "xfs_alloc.h"
-#include "xfs_btree.h"
#include "xfs_bmap.h"
#include "xfs_attr.h"
#include "xfs_attr_leaf.h"
Index: xfs/fs/xfs/xfs_dir2.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2.c 2010-05-25 11:14:59.435037366 +0200
+++ xfs/fs/xfs/xfs_dir2.c 2010-05-25 11:38:00.762068235 +0200
@@ -30,7 +30,6 @@
#include "xfs_bmap_btree.h"
#include "xfs_alloc_btree.h"
#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
Index: xfs/fs/xfs/xfs_dir2_block.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_block.c 2010-05-25 11:14:59.440014527 +0200
+++ xfs/fs/xfs/xfs_dir2_block.c 2010-05-25 11:38:00.767067886 +0200
@@ -28,7 +28,6 @@
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
Index: xfs/fs/xfs/xfs_dir2_data.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_data.c 2010-05-25 11:14:59.442011873 +0200
+++ xfs/fs/xfs/xfs_dir2_data.c 2010-05-25 11:38:00.775006634 +0200
@@ -28,7 +28,6 @@
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_dir2_data.h"
Index: xfs/fs/xfs/xfs_dir2_leaf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_leaf.c 2010-05-25 11:14:59.454006216 +0200
+++ xfs/fs/xfs/xfs_dir2_leaf.c 2010-05-25 11:38:00.780034013 +0200
@@ -28,7 +28,6 @@
#include "xfs_mount.h"
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
-#include "xfs_attr_sf.h"
#include "xfs_dir2_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
Index: xfs/fs/xfs/xfs_dir2_node.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_node.c 2010-05-25 11:14:59.671255345 +0200
+++ xfs/fs/xfs/xfs_dir2_node.c 2010-05-25 11:38:00.788005377 +0200
@@ -28,7 +28,6 @@
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_bmap.h"
Index: xfs/fs/xfs/xfs_dir2_sf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_sf.c 2010-05-25 11:14:59.678033455 +0200
+++ xfs/fs/xfs/xfs_dir2_sf.c 2010-05-25 11:38:00.794039181 +0200
@@ -28,7 +28,6 @@
#include "xfs_da_btree.h"
#include "xfs_bmap_btree.h"
#include "xfs_dir2_sf.h"
-#include "xfs_attr_sf.h"
#include "xfs_dinode.h"
#include "xfs_inode.h"
#include "xfs_inode_item.h"
Index: xfs/fs/xfs/xfs_log_cil.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log_cil.c 2010-05-25 11:38:09.226004051 +0200
+++ xfs/fs/xfs/xfs_log_cil.c 2010-05-25 11:40:49.242004818 +0200
@@ -26,7 +26,6 @@
#include "xfs_log_priv.h"
#include "xfs_sb.h"
#include "xfs_ag.h"
-#include "xfs_dir2.h"
#include "xfs_mount.h"
#include "xfs_error.h"
#include "xfs_alloc.h"
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 06/17] xfs: simplify log item descriptor tracking
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (4 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 05/17] xfs: remove unneeded #include statements Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 5:11 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 07/17] xfs: merge iop_unpin_remove into iop_unpin Christoph Hellwig
` (12 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-log_item_desc-list --]
[-- Type: text/plain, Size: 42277 bytes --]
Currently we track log item descriptor belonging to a transaction using a
complex opencoded chunk allocator. This code has been there since day one
and seems to work around the lack of an efficient slab allocator.
This patch replaces it with dynamically allocated log item descriptors
from a dedicated slab pool, linked to the transaction by a linked list.
This allows to greatly simplify the log item descriptor tracking to the
point where it's just a couple hundred lines in xfs_trans.c instead of
a separate file. The external API has also been simplified while we're
at it - the xfs_trans_add_item and xfs_trans_del_item functions to add/
delete items from a transaction have been simplified to the bare minium,
and the xfs_trans_find_item function is replaced with a direct dereference
of the li_desc field. All debug code walking the list of log items in
a transaction is down to a simple list_for_each_entry.
Note that we could easily use a singly linked list here instead of the
double linked list from list.h as the fastpath only does deletion from
sequential traversal. But given that we don't have one available as
a library function yet I use the list.h functions for simplicity.
[the patch applies ontop of Dave's delayed-logging branch]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/quota/xfs_trans_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_trans_dquot.c 2010-05-25 11:40:59.000000000 +0200
+++ xfs/fs/xfs/quota/xfs_trans_dquot.c 2010-05-25 12:43:47.925004680 +0200
@@ -49,16 +49,14 @@ xfs_trans_dqjoin(
xfs_trans_t *tp,
xfs_dquot_t *dqp)
{
- xfs_dq_logitem_t *lp = &dqp->q_logitem;
-
ASSERT(dqp->q_transp != tp);
ASSERT(XFS_DQ_IS_LOCKED(dqp));
- ASSERT(lp->qli_dquot == dqp);
+ ASSERT(dqp->q_logitem.qli_dquot == dqp);
/*
* Get a log_item_desc to point at the new item.
*/
- (void) xfs_trans_add_item(tp, (xfs_log_item_t*)(lp));
+ xfs_trans_add_item(tp, &dqp->q_logitem.qli_item);
/*
* Initialize i_transp so we can later determine if this dquot is
@@ -83,16 +81,11 @@ xfs_trans_log_dquot(
xfs_trans_t *tp,
xfs_dquot_t *dqp)
{
- xfs_log_item_desc_t *lidp;
-
ASSERT(dqp->q_transp == tp);
ASSERT(XFS_DQ_IS_LOCKED(dqp));
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)(&dqp->q_logitem));
- ASSERT(lidp != NULL);
-
tp->t_flags |= XFS_TRANS_DIRTY;
- lidp->lid_flags |= XFS_LID_DIRTY;
+ dqp->q_logitem.qli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
}
/*
@@ -864,9 +857,8 @@ xfs_trans_get_qoff_item(
/*
* Get a log_item_desc to point at the new item.
*/
- (void) xfs_trans_add_item(tp, (xfs_log_item_t*)q);
-
- return (q);
+ xfs_trans_add_item(tp, &q->qql_item);
+ return q;
}
@@ -880,13 +872,8 @@ xfs_trans_log_quotaoff_item(
xfs_trans_t *tp,
xfs_qoff_logitem_t *qlp)
{
- xfs_log_item_desc_t *lidp;
-
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t *)qlp);
- ASSERT(lidp != NULL);
-
tp->t_flags |= XFS_TRANS_DIRTY;
- lidp->lid_flags |= XFS_LID_DIRTY;
+ qlp->qql_item.li_desc->lid_flags |= XFS_LID_DIRTY;
}
STATIC void
Index: xfs/fs/xfs/xfs_buf_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.c 2010-05-25 11:40:58.000000000 +0200
+++ xfs/fs/xfs/xfs_buf_item.c 2010-05-25 12:43:47.926004470 +0200
@@ -460,13 +460,10 @@ xfs_buf_item_unpin_remove(
* occurs later in the xfs_trans_uncommit() will try to
* reference the buffer which we no longer have a hold on.
*/
- struct xfs_log_item_desc *lidp;
-
ASSERT(XFS_BUF_VALUSEMA(bip->bli_buf) <= 0);
trace_xfs_buf_item_unpin_stale(bip);
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t *)bip);
- xfs_trans_free_item(tp, lidp);
+ xfs_trans_del_item(&bip->bli_item);
/*
* Since the transaction no longer refers to the buffer, the
Index: xfs/fs/xfs/xfs_extfree_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_extfree_item.c 2010-05-25 11:40:58.000000000 +0200
+++ xfs/fs/xfs/xfs_extfree_item.c 2010-05-25 12:43:47.938004470 +0200
@@ -131,18 +131,18 @@ STATIC void
xfs_efi_item_unpin_remove(xfs_efi_log_item_t *efip, xfs_trans_t *tp)
{
struct xfs_ail *ailp = efip->efi_item.li_ailp;
- xfs_log_item_desc_t *lidp;
spin_lock(&ailp->xa_lock);
if (efip->efi_flags & XFS_EFI_CANCELED) {
+ struct xfs_log_item *lip = &efip->efi_item;
+
/*
* free the xaction descriptor pointing to this item
*/
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t *) efip);
- xfs_trans_free_item(tp, lidp);
+ xfs_trans_del_item(lip);
/* xfs_trans_ail_delete() drops the AIL lock. */
- xfs_trans_ail_delete(ailp, (xfs_log_item_t *)efip);
+ xfs_trans_ail_delete(ailp, lip);
xfs_efi_item_free(efip);
} else {
efip->efi_flags |= XFS_EFI_COMMITTED;
Index: xfs/fs/xfs/xfs_trans_buf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_buf.c 2010-05-25 11:40:59.292005935 +0200
+++ xfs/fs/xfs/xfs_trans_buf.c 2010-05-25 12:43:47.945004470 +0200
@@ -47,36 +47,17 @@ xfs_trans_buf_item_match(
xfs_daddr_t blkno,
int len)
{
- xfs_log_item_chunk_t *licp;
- xfs_log_item_desc_t *lidp;
- xfs_buf_log_item_t *blip;
- int i;
+ struct xfs_log_item_desc *lidp;
+ struct xfs_buf_log_item *blip;
len = BBTOB(len);
- for (licp = &tp->t_items; licp != NULL; licp = licp->lic_next) {
- if (xfs_lic_are_all_free(licp)) {
- ASSERT(licp == &tp->t_items);
- ASSERT(licp->lic_next == NULL);
- return NULL;
- }
-
- for (i = 0; i < licp->lic_unused; i++) {
- /*
- * Skip unoccupied slots.
- */
- if (xfs_lic_isfree(licp, i))
- continue;
-
- lidp = xfs_lic_slot(licp, i);
- blip = (xfs_buf_log_item_t *)lidp->lid_item;
- if (blip->bli_item.li_type != XFS_LI_BUF)
- continue;
-
- if (XFS_BUF_TARGET(blip->bli_buf) == target &&
- XFS_BUF_ADDR(blip->bli_buf) == blkno &&
- XFS_BUF_COUNT(blip->bli_buf) == len)
- return blip->bli_buf;
- }
+ list_for_each_entry(lidp, &tp->t_items, lid_trans) {
+ blip = (struct xfs_buf_log_item *)lidp->lid_item;
+ if (blip->bli_item.li_type == XFS_LI_BUF &&
+ XFS_BUF_TARGET(blip->bli_buf) == target &&
+ XFS_BUF_ADDR(blip->bli_buf) == blkno &&
+ XFS_BUF_COUNT(blip->bli_buf) == len)
+ return blip->bli_buf;
}
return NULL;
@@ -123,7 +104,7 @@ _xfs_trans_bjoin(
/*
* Get a log_item_desc to point at the new item.
*/
- (void) xfs_trans_add_item(tp, (xfs_log_item_t *)bip);
+ xfs_trans_add_item(tp, &bip->bli_item);
/*
* Initialize b_fsprivate2 so we can find it with incore_match()
@@ -479,7 +460,6 @@ xfs_trans_brelse(xfs_trans_t *tp,
{
xfs_buf_log_item_t *bip;
xfs_log_item_t *lip;
- xfs_log_item_desc_t *lidp;
/*
* Default to a normal brelse() call if the tp is NULL.
@@ -510,13 +490,6 @@ xfs_trans_brelse(xfs_trans_t *tp,
ASSERT(!(bip->bli_format.blf_flags & XFS_BLF_CANCEL));
ASSERT(atomic_read(&bip->bli_refcount) > 0);
- /*
- * Find the item descriptor pointing to this buffer's
- * log item. It must be there.
- */
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)bip);
- ASSERT(lidp != NULL);
-
trace_xfs_trans_brelse(bip);
/*
@@ -532,7 +505,7 @@ xfs_trans_brelse(xfs_trans_t *tp,
* If the buffer is dirty within this transaction, we can't
* release it until we commit.
*/
- if (lidp->lid_flags & XFS_LID_DIRTY)
+ if (bip->bli_item.li_desc->lid_flags & XFS_LID_DIRTY)
return;
/*
@@ -549,7 +522,7 @@ xfs_trans_brelse(xfs_trans_t *tp,
/*
* Free up the log item descriptor tracking the released item.
*/
- xfs_trans_free_item(tp, lidp);
+ xfs_trans_del_item(&bip->bli_item);
/*
* Clear the hold flag in the buf log item if it is set.
@@ -661,7 +634,6 @@ xfs_trans_log_buf(xfs_trans_t *tp,
uint last)
{
xfs_buf_log_item_t *bip;
- xfs_log_item_desc_t *lidp;
ASSERT(XFS_BUF_ISBUSY(bp));
ASSERT(XFS_BUF_FSPRIVATE2(bp, xfs_trans_t *) == tp);
@@ -703,11 +675,8 @@ xfs_trans_log_buf(xfs_trans_t *tp,
bip->bli_format.blf_flags &= ~XFS_BLF_CANCEL;
}
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)bip);
- ASSERT(lidp != NULL);
-
tp->t_flags |= XFS_TRANS_DIRTY;
- lidp->lid_flags |= XFS_LID_DIRTY;
+ bip->bli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
bip->bli_flags |= XFS_BLI_LOGGED;
xfs_buf_item_log(bip, first, last);
}
@@ -736,7 +705,6 @@ xfs_trans_binval(
xfs_trans_t *tp,
xfs_buf_t *bp)
{
- xfs_log_item_desc_t *lidp;
xfs_buf_log_item_t *bip;
ASSERT(XFS_BUF_ISBUSY(bp));
@@ -744,8 +712,6 @@ xfs_trans_binval(
ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);
bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)bip);
- ASSERT(lidp != NULL);
ASSERT(atomic_read(&bip->bli_refcount) > 0);
trace_xfs_trans_binval(bip);
@@ -760,7 +726,7 @@ xfs_trans_binval(
ASSERT(!(bip->bli_flags & (XFS_BLI_LOGGED | XFS_BLI_DIRTY)));
ASSERT(!(bip->bli_format.blf_flags & XFS_BLF_INODE_BUF));
ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
- ASSERT(lidp->lid_flags & XFS_LID_DIRTY);
+ ASSERT(bip->bli_item.li_desc->lid_flags & XFS_LID_DIRTY);
ASSERT(tp->t_flags & XFS_TRANS_DIRTY);
return;
}
@@ -793,7 +759,7 @@ xfs_trans_binval(
bip->bli_format.blf_flags |= XFS_BLF_CANCEL;
memset((char *)(bip->bli_format.blf_data_map), 0,
(bip->bli_format.blf_map_size * sizeof(uint)));
- lidp->lid_flags |= XFS_LID_DIRTY;
+ bip->bli_item.li_desc->lid_flags |= XFS_LID_DIRTY;
tp->t_flags |= XFS_TRANS_DIRTY;
}
Index: xfs/fs/xfs/xfs_trans_extfree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_extfree.c 2010-05-25 11:40:58.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_extfree.c 2010-05-25 12:43:47.950004191 +0200
@@ -48,9 +48,8 @@ xfs_trans_get_efi(xfs_trans_t *tp,
/*
* Get a log_item_desc to point at the new item.
*/
- (void) xfs_trans_add_item(tp, (xfs_log_item_t*)efip);
-
- return (efip);
+ xfs_trans_add_item(tp, &efip->efi_item);
+ return efip;
}
/*
@@ -64,15 +63,11 @@ xfs_trans_log_efi_extent(xfs_trans_t *t
xfs_fsblock_t start_block,
xfs_extlen_t ext_len)
{
- xfs_log_item_desc_t *lidp;
uint next_extent;
xfs_extent_t *extp;
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)efip);
- ASSERT(lidp != NULL);
-
tp->t_flags |= XFS_TRANS_DIRTY;
- lidp->lid_flags |= XFS_LID_DIRTY;
+ efip->efi_item.li_desc->lid_flags |= XFS_LID_DIRTY;
next_extent = efip->efi_next_extent;
ASSERT(next_extent < efip->efi_format.efi_nextents);
@@ -105,9 +100,8 @@ xfs_trans_get_efd(xfs_trans_t *tp,
/*
* Get a log_item_desc to point at the new item.
*/
- (void) xfs_trans_add_item(tp, (xfs_log_item_t*)efdp);
-
- return (efdp);
+ xfs_trans_add_item(tp, &efdp->efd_item);
+ return efdp;
}
/*
@@ -121,15 +115,11 @@ xfs_trans_log_efd_extent(xfs_trans_t *t
xfs_fsblock_t start_block,
xfs_extlen_t ext_len)
{
- xfs_log_item_desc_t *lidp;
uint next_extent;
xfs_extent_t *extp;
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)efdp);
- ASSERT(lidp != NULL);
-
tp->t_flags |= XFS_TRANS_DIRTY;
- lidp->lid_flags |= XFS_LID_DIRTY;
+ efdp->efd_item.li_desc->lid_flags |= XFS_LID_DIRTY;
next_extent = efdp->efd_next_extent;
ASSERT(next_extent < efdp->efd_format.efd_nextents);
Index: xfs/fs/xfs/xfs_trans_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_inode.c 2010-05-25 11:40:59.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_inode.c 2010-05-25 12:43:47.956003422 +0200
@@ -88,7 +88,7 @@ xfs_trans_ijoin(
/*
* Get a log_item_desc to point at the new item.
*/
- (void) xfs_trans_add_item(tp, (xfs_log_item_t*)(iip));
+ xfs_trans_add_item(tp, &iip->ili_item);
xfs_trans_inode_broot_debug(ip);
@@ -144,17 +144,12 @@ xfs_trans_log_inode(
xfs_inode_t *ip,
uint flags)
{
- xfs_log_item_desc_t *lidp;
-
ASSERT(ip->i_transp == tp);
ASSERT(ip->i_itemp != NULL);
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
- lidp = xfs_trans_find_item(tp, (xfs_log_item_t*)(ip->i_itemp));
- ASSERT(lidp != NULL);
-
tp->t_flags |= XFS_TRANS_DIRTY;
- lidp->lid_flags |= XFS_LID_DIRTY;
+ ip->i_itemp->ili_item.li_desc->lid_flags |= XFS_LID_DIRTY;
/*
* Always OR in the bits from the ili_last_fields field.
Index: xfs/fs/xfs/xfs_trans_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_item.c 2010-05-25 12:43:55.924003980 +0200
+++ /dev/null 1970-01-01 00:00:00.000000000 +0000
@@ -1,440 +0,0 @@
-/*
- * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
- * All Rights Reserved.
- *
- * 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.
- *
- * This program is distributed in the hope that it would 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 this program; if not, write the Free Software Foundation,
- * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
- */
-#include "xfs.h"
-#include "xfs_fs.h"
-#include "xfs_types.h"
-#include "xfs_log.h"
-#include "xfs_inum.h"
-#include "xfs_trans.h"
-#include "xfs_trans_priv.h"
-/* XXX: from here down needed until struct xfs_trans has its own ailp */
-#include "xfs_bit.h"
-#include "xfs_buf_item.h"
-#include "xfs_sb.h"
-#include "xfs_ag.h"
-#include "xfs_dir2.h"
-#include "xfs_mount.h"
-
-STATIC int xfs_trans_unlock_chunk(xfs_log_item_chunk_t *,
- int, int, xfs_lsn_t);
-
-/*
- * This is called to add the given log item to the transaction's
- * list of log items. It must find a free log item descriptor
- * or allocate a new one and add the item to that descriptor.
- * The function returns a pointer to item descriptor used to point
- * to the new item. The log item will now point to its new descriptor
- * with its li_desc field.
- */
-xfs_log_item_desc_t *
-xfs_trans_add_item(xfs_trans_t *tp, xfs_log_item_t *lip)
-{
- xfs_log_item_desc_t *lidp;
- xfs_log_item_chunk_t *licp;
- int i=0;
-
- /*
- * If there are no free descriptors, allocate a new chunk
- * of them and put it at the front of the chunk list.
- */
- if (tp->t_items_free == 0) {
- licp = (xfs_log_item_chunk_t*)
- kmem_alloc(sizeof(xfs_log_item_chunk_t), KM_SLEEP);
- ASSERT(licp != NULL);
- /*
- * Initialize the chunk, and then
- * claim the first slot in the newly allocated chunk.
- */
- xfs_lic_init(licp);
- xfs_lic_claim(licp, 0);
- licp->lic_unused = 1;
- xfs_lic_init_slot(licp, 0);
- lidp = xfs_lic_slot(licp, 0);
-
- /*
- * Link in the new chunk and update the free count.
- */
- licp->lic_next = tp->t_items.lic_next;
- tp->t_items.lic_next = licp;
- tp->t_items_free = XFS_LIC_NUM_SLOTS - 1;
-
- /*
- * Initialize the descriptor and the generic portion
- * of the log item.
- *
- * Point the new slot at this item and return it.
- * Also point the log item at its currently active
- * descriptor and set the item's mount pointer.
- */
- lidp->lid_item = lip;
- lidp->lid_flags = 0;
- lidp->lid_size = 0;
- lip->li_desc = lidp;
- lip->li_mountp = tp->t_mountp;
- lip->li_ailp = tp->t_mountp->m_ail;
- return lidp;
- }
-
- /*
- * Find the free descriptor. It is somewhere in the chunklist
- * of descriptors.
- */
- licp = &tp->t_items;
- while (licp != NULL) {
- if (xfs_lic_vacancy(licp)) {
- if (licp->lic_unused <= XFS_LIC_MAX_SLOT) {
- i = licp->lic_unused;
- ASSERT(xfs_lic_isfree(licp, i));
- break;
- }
- for (i = 0; i <= XFS_LIC_MAX_SLOT; i++) {
- if (xfs_lic_isfree(licp, i))
- break;
- }
- ASSERT(i <= XFS_LIC_MAX_SLOT);
- break;
- }
- licp = licp->lic_next;
- }
- ASSERT(licp != NULL);
- /*
- * If we find a free descriptor, claim it,
- * initialize it, and return it.
- */
- xfs_lic_claim(licp, i);
- if (licp->lic_unused <= i) {
- licp->lic_unused = i + 1;
- xfs_lic_init_slot(licp, i);
- }
- lidp = xfs_lic_slot(licp, i);
- tp->t_items_free--;
- lidp->lid_item = lip;
- lidp->lid_flags = 0;
- lidp->lid_size = 0;
- lip->li_desc = lidp;
- lip->li_mountp = tp->t_mountp;
- lip->li_ailp = tp->t_mountp->m_ail;
- return lidp;
-}
-
-/*
- * Free the given descriptor.
- *
- * This requires setting the bit in the chunk's free mask corresponding
- * to the given slot.
- */
-void
-xfs_trans_free_item(xfs_trans_t *tp, xfs_log_item_desc_t *lidp)
-{
- uint slot;
- xfs_log_item_chunk_t *licp;
- xfs_log_item_chunk_t **licpp;
-
- slot = xfs_lic_desc_to_slot(lidp);
- licp = xfs_lic_desc_to_chunk(lidp);
- xfs_lic_relse(licp, slot);
- lidp->lid_item->li_desc = NULL;
- tp->t_items_free++;
-
- /*
- * If there are no more used items in the chunk and this is not
- * the chunk embedded in the transaction structure, then free
- * the chunk. First pull it from the chunk list and then
- * free it back to the heap. We didn't bother with a doubly
- * linked list here because the lists should be very short
- * and this is not a performance path. It's better to save
- * the memory of the extra pointer.
- *
- * Also decrement the transaction structure's count of free items
- * by the number in a chunk since we are freeing an empty chunk.
- */
- if (xfs_lic_are_all_free(licp) && (licp != &(tp->t_items))) {
- licpp = &(tp->t_items.lic_next);
- while (*licpp != licp) {
- ASSERT(*licpp != NULL);
- licpp = &((*licpp)->lic_next);
- }
- *licpp = licp->lic_next;
- kmem_free(licp);
- tp->t_items_free -= XFS_LIC_NUM_SLOTS;
- }
-}
-
-/*
- * This is called to find the descriptor corresponding to the given
- * log item. It returns a pointer to the descriptor.
- * The log item MUST have a corresponding descriptor in the given
- * transaction. This routine does not return NULL, it panics.
- *
- * The descriptor pointer is kept in the log item's li_desc field.
- * Just return it.
- */
-/*ARGSUSED*/
-xfs_log_item_desc_t *
-xfs_trans_find_item(xfs_trans_t *tp, xfs_log_item_t *lip)
-{
- ASSERT(lip->li_desc != NULL);
-
- return lip->li_desc;
-}
-
-
-/*
- * Return a pointer to the first descriptor in the chunk list.
- * This does not return NULL if there are none, it panics.
- *
- * The first descriptor must be in either the first or second chunk.
- * This is because the only chunk allowed to be empty is the first.
- * All others are freed when they become empty.
- *
- * At some point this and xfs_trans_next_item() should be optimized
- * to quickly look at the mask to determine if there is anything to
- * look at.
- */
-xfs_log_item_desc_t *
-xfs_trans_first_item(xfs_trans_t *tp)
-{
- xfs_log_item_chunk_t *licp;
- int i;
-
- licp = &tp->t_items;
- /*
- * If it's not in the first chunk, skip to the second.
- */
- if (xfs_lic_are_all_free(licp)) {
- licp = licp->lic_next;
- }
-
- /*
- * Return the first non-free descriptor in the chunk.
- */
- ASSERT(!xfs_lic_are_all_free(licp));
- for (i = 0; i < licp->lic_unused; i++) {
- if (xfs_lic_isfree(licp, i)) {
- continue;
- }
-
- return xfs_lic_slot(licp, i);
- }
- cmn_err(CE_WARN, "xfs_trans_first_item() -- no first item");
- return NULL;
-}
-
-
-/*
- * Given a descriptor, return the next descriptor in the chunk list.
- * This returns NULL if there are no more used descriptors in the list.
- *
- * We do this by first locating the chunk in which the descriptor resides,
- * and then scanning forward in the chunk and the list for the next
- * used descriptor.
- */
-/*ARGSUSED*/
-xfs_log_item_desc_t *
-xfs_trans_next_item(xfs_trans_t *tp, xfs_log_item_desc_t *lidp)
-{
- xfs_log_item_chunk_t *licp;
- int i;
-
- licp = xfs_lic_desc_to_chunk(lidp);
-
- /*
- * First search the rest of the chunk. The for loop keeps us
- * from referencing things beyond the end of the chunk.
- */
- for (i = (int)xfs_lic_desc_to_slot(lidp) + 1; i < licp->lic_unused; i++) {
- if (xfs_lic_isfree(licp, i)) {
- continue;
- }
-
- return xfs_lic_slot(licp, i);
- }
-
- /*
- * Now search the next chunk. It must be there, because the
- * next chunk would have been freed if it were empty.
- * If there is no next chunk, return NULL.
- */
- if (licp->lic_next == NULL) {
- return NULL;
- }
-
- licp = licp->lic_next;
- ASSERT(!xfs_lic_are_all_free(licp));
- for (i = 0; i < licp->lic_unused; i++) {
- if (xfs_lic_isfree(licp, i)) {
- continue;
- }
-
- return xfs_lic_slot(licp, i);
- }
- ASSERT(0);
- /* NOTREACHED */
- return NULL; /* keep gcc quite */
-}
-
-/*
- * This is called to unlock all of the items of a transaction and to free
- * all the descriptors of that transaction.
- *
- * It walks the list of descriptors and unlocks each item. It frees
- * each chunk except that embedded in the transaction as it goes along.
- */
-void
-xfs_trans_free_items(
- xfs_trans_t *tp,
- xfs_lsn_t commit_lsn,
- int flags)
-{
- xfs_log_item_chunk_t *licp;
- xfs_log_item_chunk_t *next_licp;
- int abort;
-
- abort = flags & XFS_TRANS_ABORT;
- licp = &tp->t_items;
- /*
- * Special case the embedded chunk so we don't free it below.
- */
- if (!xfs_lic_are_all_free(licp)) {
- (void) xfs_trans_unlock_chunk(licp, 1, abort, commit_lsn);
- xfs_lic_all_free(licp);
- licp->lic_unused = 0;
- }
- licp = licp->lic_next;
-
- /*
- * Unlock each item in each chunk and free the chunks.
- */
- while (licp != NULL) {
- ASSERT(!xfs_lic_are_all_free(licp));
- (void) xfs_trans_unlock_chunk(licp, 1, abort, commit_lsn);
- next_licp = licp->lic_next;
- kmem_free(licp);
- licp = next_licp;
- }
-
- /*
- * Reset the transaction structure's free item count.
- */
- tp->t_items_free = XFS_LIC_NUM_SLOTS;
- tp->t_items.lic_next = NULL;
-}
-
-
-
-/*
- * This is called to unlock the items associated with a transaction.
- * Items which were not logged should be freed.
- * Those which were logged must still be tracked so they can be unpinned
- * when the transaction commits.
- */
-void
-xfs_trans_unlock_items(xfs_trans_t *tp, xfs_lsn_t commit_lsn)
-{
- xfs_log_item_chunk_t *licp;
- xfs_log_item_chunk_t *next_licp;
- xfs_log_item_chunk_t **licpp;
- int freed;
-
- freed = 0;
- licp = &tp->t_items;
-
- /*
- * Special case the embedded chunk so we don't free.
- */
- if (!xfs_lic_are_all_free(licp)) {
- freed = xfs_trans_unlock_chunk(licp, 0, 0, commit_lsn);
- }
- licpp = &(tp->t_items.lic_next);
- licp = licp->lic_next;
-
- /*
- * Unlock each item in each chunk, free non-dirty descriptors,
- * and free empty chunks.
- */
- while (licp != NULL) {
- ASSERT(!xfs_lic_are_all_free(licp));
- freed += xfs_trans_unlock_chunk(licp, 0, 0, commit_lsn);
- next_licp = licp->lic_next;
- if (xfs_lic_are_all_free(licp)) {
- *licpp = next_licp;
- kmem_free(licp);
- freed -= XFS_LIC_NUM_SLOTS;
- } else {
- licpp = &(licp->lic_next);
- }
- ASSERT(*licpp == next_licp);
- licp = next_licp;
- }
-
- /*
- * Fix the free descriptor count in the transaction.
- */
- tp->t_items_free += freed;
-}
-
-/*
- * Unlock each item pointed to by a descriptor in the given chunk.
- * Stamp the commit lsn into each item if necessary.
- * Free descriptors pointing to items which are not dirty if freeing_chunk
- * is zero. If freeing_chunk is non-zero, then we need to unlock all
- * items in the chunk.
- *
- * Return the number of descriptors freed.
- */
-STATIC int
-xfs_trans_unlock_chunk(
- xfs_log_item_chunk_t *licp,
- int freeing_chunk,
- int abort,
- xfs_lsn_t commit_lsn)
-{
- xfs_log_item_desc_t *lidp;
- xfs_log_item_t *lip;
- int i;
- int freed;
-
- freed = 0;
- lidp = licp->lic_descs;
- for (i = 0; i < licp->lic_unused; i++, lidp++) {
- if (xfs_lic_isfree(licp, i)) {
- continue;
- }
- lip = lidp->lid_item;
- lip->li_desc = NULL;
-
- if (commit_lsn != NULLCOMMITLSN)
- IOP_COMMITTING(lip, commit_lsn);
- if (abort)
- lip->li_flags |= XFS_LI_ABORTED;
- IOP_UNLOCK(lip);
-
- /*
- * Free the descriptor if the item is not dirty
- * within this transaction and the caller is not
- * going to just free the entire thing regardless.
- */
- if (!(freeing_chunk) &&
- (!(lidp->lid_flags & XFS_LID_DIRTY) || abort)) {
- xfs_lic_relse(licp, i);
- freed++;
- }
- }
-
- return freed;
-}
Index: xfs/fs/xfs/xfs_trans_priv.h
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_priv.h 2010-05-24 23:11:41.000000000 +0200
+++ xfs/fs/xfs/xfs_trans_priv.h 2010-05-25 12:43:47.958004191 +0200
@@ -23,22 +23,8 @@ struct xfs_log_item_desc;
struct xfs_mount;
struct xfs_trans;
-/*
- * From xfs_trans_item.c
- */
-struct xfs_log_item_desc *xfs_trans_add_item(struct xfs_trans *,
- struct xfs_log_item *);
-void xfs_trans_free_item(struct xfs_trans *,
- struct xfs_log_item_desc *);
-struct xfs_log_item_desc *xfs_trans_find_item(struct xfs_trans *,
- struct xfs_log_item *);
-struct xfs_log_item_desc *xfs_trans_first_item(struct xfs_trans *);
-struct xfs_log_item_desc *xfs_trans_next_item(struct xfs_trans *,
- struct xfs_log_item_desc *);
-
-void xfs_trans_unlock_items(struct xfs_trans *tp, xfs_lsn_t commit_lsn);
-void xfs_trans_free_items(struct xfs_trans *tp, xfs_lsn_t commit_lsn,
- int flags);
+void xfs_trans_add_item(struct xfs_trans *, struct xfs_log_item *);
+void xfs_trans_del_item(struct xfs_log_item *);
void xfs_trans_item_committed(struct xfs_log_item *lip,
xfs_lsn_t commit_lsn, int aborted);
Index: xfs/fs/xfs/Makefile
===================================================================
--- xfs.orig/fs/xfs/Makefile 2010-05-25 11:40:58.000000000 +0200
+++ xfs/fs/xfs/Makefile 2010-05-25 12:43:47.961004819 +0200
@@ -87,7 +87,6 @@ xfs-y += xfs_alloc.o \
xfs_trans_buf.o \
xfs_trans_extfree.o \
xfs_trans_inode.o \
- xfs_trans_item.o \
xfs_utils.o \
xfs_vnodeops.o \
xfs_rw.o
Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c 2010-05-25 11:40:59.000000000 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.c 2010-05-25 12:43:47.966004540 +0200
@@ -1702,6 +1702,12 @@ xfs_init_zones(void)
if (!xfs_trans_zone)
goto out_destroy_ifork_zone;
+ xfs_log_item_desc_zone =
+ kmem_zone_init(sizeof(struct xfs_log_item_desc),
+ "xfs_log_item_desc");
+ if (!xfs_log_item_desc_zone)
+ goto out_destroy_trans_zone;
+
/*
* The size of the zone allocated buf log item is the maximum
* size possible under XFS. This wastes a little bit of memory,
@@ -1711,7 +1717,7 @@ xfs_init_zones(void)
(((XFS_MAX_BLOCKSIZE / XFS_BLF_CHUNK) /
NBWORD) * sizeof(int))), "xfs_buf_item");
if (!xfs_buf_item_zone)
- goto out_destroy_trans_zone;
+ goto out_destroy_log_item_desc_zone;
xfs_efd_zone = kmem_zone_init((sizeof(xfs_efd_log_item_t) +
((XFS_EFD_MAX_FAST_EXTENTS - 1) *
@@ -1748,6 +1754,8 @@ xfs_init_zones(void)
kmem_zone_destroy(xfs_efd_zone);
out_destroy_buf_item_zone:
kmem_zone_destroy(xfs_buf_item_zone);
+ out_destroy_log_item_desc_zone:
+ kmem_zone_destroy(xfs_log_item_desc_zone);
out_destroy_trans_zone:
kmem_zone_destroy(xfs_trans_zone);
out_destroy_ifork_zone:
@@ -1778,6 +1786,7 @@ xfs_destroy_zones(void)
kmem_zone_destroy(xfs_efi_zone);
kmem_zone_destroy(xfs_efd_zone);
kmem_zone_destroy(xfs_buf_item_zone);
+ kmem_zone_destroy(xfs_log_item_desc_zone);
kmem_zone_destroy(xfs_trans_zone);
kmem_zone_destroy(xfs_ifork_zone);
kmem_zone_destroy(xfs_dabuf_zone);
Index: xfs/fs/xfs/xfs_bmap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.c 2010-05-25 11:40:59.000000000 +0200
+++ xfs/fs/xfs/xfs_bmap.c 2010-05-25 12:43:47.970004889 +0200
@@ -5857,43 +5857,18 @@ xfs_bmap_get_bp(
bp = NULL;
if (!bp) { /* Chase down all the log items to see if the bp is there */
- xfs_log_item_chunk_t *licp;
- xfs_trans_t *tp;
+ struct xfs_log_item_desc *lidp;
+ struct xfs_buf_log_item *bip;
- tp = cur->bc_tp;
- licp = &tp->t_items;
- while (!bp && licp != NULL) {
- if (xfs_lic_are_all_free(licp)) {
- licp = licp->lic_next;
- continue;
- }
- for (i = 0; i < licp->lic_unused; i++) {
- xfs_log_item_desc_t *lidp;
- xfs_log_item_t *lip;
- xfs_buf_log_item_t *bip;
- xfs_buf_t *lbp;
-
- if (xfs_lic_isfree(licp, i)) {
- continue;
- }
-
- lidp = xfs_lic_slot(licp, i);
- lip = lidp->lid_item;
- if (lip->li_type != XFS_LI_BUF)
- continue;
-
- bip = (xfs_buf_log_item_t *)lip;
- lbp = bip->bli_buf;
-
- if (XFS_BUF_ADDR(lbp) == bno) {
- bp = lbp;
- break; /* Found it */
- }
- }
- licp = licp->lic_next;
+ list_for_each_entry(lidp, &cur->bc_tp->t_items, lid_trans) {
+ bip = (struct xfs_buf_log_item *)lidp->lid_item;
+ if (bip->bli_item.li_type == XFS_LI_BUF &&
+ XFS_BUF_ADDR(bip->bli_buf) == bno)
+ return bip->bli_buf;
}
}
- return(bp);
+
+ return bp;
}
STATIC void
Index: xfs/fs/xfs/xfs_trans.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.c 2010-05-25 12:43:44.000000000 +0200
+++ xfs/fs/xfs/xfs_trans.c 2010-05-25 12:43:47.977004330 +0200
@@ -1,5 +1,6 @@
/*
* Copyright (c) 2000-2003,2005 Silicon Graphics, Inc.
+ * Copyright (C) 2010 Red Hat, Inc.
* All Rights Reserved.
*
* This program is free software; you can redistribute it and/or
@@ -43,6 +44,7 @@
#include "xfs_trace.h"
kmem_zone_t *xfs_trans_zone;
+kmem_zone_t *xfs_log_item_desc_zone;
/*
@@ -593,8 +595,7 @@ _xfs_trans_alloc(
tp->t_magic = XFS_TRANS_MAGIC;
tp->t_type = type;
tp->t_mountp = mp;
- tp->t_items_free = XFS_LIC_NUM_SLOTS;
- xfs_lic_init(&(tp->t_items));
+ INIT_LIST_HEAD(&tp->t_items);
INIT_LIST_HEAD(&tp->t_busy);
return tp;
}
@@ -639,8 +640,7 @@ xfs_trans_dup(
ntp->t_magic = XFS_TRANS_MAGIC;
ntp->t_type = tp->t_type;
ntp->t_mountp = tp->t_mountp;
- ntp->t_items_free = XFS_LIC_NUM_SLOTS;
- xfs_lic_init(&(ntp->t_items));
+ INIT_LIST_HEAD(&ntp->t_items);
INIT_LIST_HEAD(&ntp->t_busy);
ASSERT(tp->t_flags & XFS_TRANS_PERM_LOG_RES);
@@ -1120,6 +1120,108 @@ xfs_trans_unreserve_and_mod_sb(
}
/*
+ * Add the given log item to the transaction's list of log items.
+ *
+ * The log item will now point to its new descriptor with its li_desc field.
+ */
+void
+xfs_trans_add_item(
+ struct xfs_trans *tp,
+ struct xfs_log_item *lip)
+{
+ struct xfs_log_item_desc *lidp;
+
+ ASSERT(lip->li_mountp = tp->t_mountp);
+ ASSERT(lip->li_ailp = tp->t_mountp->m_ail);
+
+ lidp = kmem_zone_zalloc(xfs_log_item_desc_zone, KM_SLEEP);
+
+ lidp->lid_item = lip;
+ lidp->lid_flags = 0;
+ lidp->lid_size = 0;
+ list_add_tail(&lidp->lid_trans, &tp->t_items);
+
+ lip->li_desc = lidp;
+}
+
+STATIC void
+xfs_trans_free_item_desc(
+ struct xfs_log_item_desc *lidp)
+{
+ list_del_init(&lidp->lid_trans);
+ kmem_zone_free(xfs_log_item_desc_zone, lidp);
+}
+
+/*
+ * Unlink and free the given descriptor.
+ */
+void
+xfs_trans_del_item(
+ struct xfs_log_item *lip)
+{
+ xfs_trans_free_item_desc(lip->li_desc);
+ lip->li_desc = NULL;
+}
+
+/*
+ * Unlock all of the items of a transaction and free all the descriptors
+ * of that transaction.
+ */
+STATIC void
+xfs_trans_free_items(
+ struct xfs_trans *tp,
+ xfs_lsn_t commit_lsn,
+ int flags)
+{
+ struct xfs_log_item_desc *lidp, *next;
+
+ list_for_each_entry_safe(lidp, next, &tp->t_items, lid_trans) {
+ struct xfs_log_item *lip = lidp->lid_item;
+
+ lip->li_desc = NULL;
+
+ if (commit_lsn != NULLCOMMITLSN)
+ IOP_COMMITTING(lip, commit_lsn);
+ if (flags & XFS_TRANS_ABORT)
+ lip->li_flags |= XFS_LI_ABORTED;
+ IOP_UNLOCK(lip);
+
+ xfs_trans_free_item_desc(lidp);
+ }
+}
+
+/*
+ * Unlock the items associated with a transaction.
+ *
+ * Items which were not logged should be freed. Those which were logged must
+ * still be tracked so they can be unpinned when the transaction commits.
+ */
+STATIC void
+xfs_trans_unlock_items(
+ struct xfs_trans *tp,
+ xfs_lsn_t commit_lsn)
+{
+ struct xfs_log_item_desc *lidp, *next;
+
+ list_for_each_entry_safe(lidp, next, &tp->t_items, lid_trans) {
+ struct xfs_log_item *lip = lidp->lid_item;
+
+ lip->li_desc = NULL;
+
+ if (commit_lsn != NULLCOMMITLSN)
+ IOP_COMMITTING(lip, commit_lsn);
+ IOP_UNLOCK(lip);
+
+ /*
+ * Free the descriptor if the item is not dirty
+ * within this transaction.
+ */
+ if (!(lidp->lid_flags & XFS_LID_DIRTY))
+ xfs_trans_free_item_desc(lidp);
+ }
+}
+
+/*
* Total up the number of log iovecs needed to commit this
* transaction. The transaction itself needs one for the
* transaction header. Ask each dirty item in turn how many
@@ -1130,30 +1232,27 @@ xfs_trans_count_vecs(
struct xfs_trans *tp)
{
int nvecs;
- xfs_log_item_desc_t *lidp;
+ struct xfs_log_item_desc *lidp;
nvecs = 1;
- lidp = xfs_trans_first_item(tp);
- ASSERT(lidp != NULL);
/* In the non-debug case we need to start bailing out if we
* didn't find a log_item here, return zero and let trans_commit
* deal with it.
*/
- if (lidp == NULL)
+ if (list_empty(&tp->t_items)) {
+ ASSERT(0);
return 0;
+ }
- while (lidp != NULL) {
+ list_for_each_entry(lidp, &tp->t_items, lid_trans) {
/*
* Skip items which aren't dirty in this transaction.
*/
- if (!(lidp->lid_flags & XFS_LID_DIRTY)) {
- lidp = xfs_trans_next_item(tp, lidp);
+ if (!(lidp->lid_flags & XFS_LID_DIRTY))
continue;
- }
lidp->lid_size = IOP_SIZE(lidp->lid_item);
nvecs += lidp->lid_size;
- lidp = xfs_trans_next_item(tp, lidp);
}
return nvecs;
@@ -1173,7 +1272,7 @@ xfs_trans_fill_vecs(
struct xfs_trans *tp,
struct xfs_log_iovec *log_vector)
{
- xfs_log_item_desc_t *lidp;
+ struct xfs_log_item_desc *lidp;
struct xfs_log_iovec *vecp;
uint nitems;
@@ -1184,14 +1283,11 @@ xfs_trans_fill_vecs(
vecp = log_vector + 1;
nitems = 0;
- lidp = xfs_trans_first_item(tp);
- ASSERT(lidp);
- while (lidp) {
+ ASSERT(!list_empty(&tp->t_items));
+ list_for_each_entry(lidp, &tp->t_items, lid_trans) {
/* Skip items which aren't dirty in this transaction. */
- if (!(lidp->lid_flags & XFS_LID_DIRTY)) {
- lidp = xfs_trans_next_item(tp, lidp);
+ if (!(lidp->lid_flags & XFS_LID_DIRTY))
continue;
- }
/*
* The item may be marked dirty but not log anything. This can
@@ -1202,7 +1298,6 @@ xfs_trans_fill_vecs(
IOP_FORMAT(lidp->lid_item, vecp);
vecp += lidp->lid_size;
IOP_PIN(lidp->lid_item);
- lidp = xfs_trans_next_item(tp, lidp);
}
/*
@@ -1297,24 +1392,15 @@ xfs_trans_committed(
struct xfs_trans *tp,
int abortflag)
{
- xfs_log_item_desc_t *lidp;
- xfs_log_item_chunk_t *licp;
- xfs_log_item_chunk_t *next_licp;
+ struct xfs_log_item_desc *lidp, *next;
/* Call the transaction's completion callback if there is one. */
if (tp->t_callback != NULL)
tp->t_callback(tp, tp->t_callarg);
- for (lidp = xfs_trans_first_item(tp);
- lidp != NULL;
- lidp = xfs_trans_next_item(tp, lidp)) {
+ list_for_each_entry_safe(lidp, next, &tp->t_items, lid_trans) {
xfs_trans_item_committed(lidp->lid_item, tp->t_lsn, abortflag);
- }
-
- /* free the item chunks, ignoring the embedded chunk */
- for (licp = tp->t_items.lic_next; licp != NULL; licp = next_licp) {
- next_licp = licp->lic_next;
- kmem_free(licp);
+ xfs_trans_free_item_desc(lidp);
}
xfs_trans_free(tp);
@@ -1329,11 +1415,9 @@ xfs_trans_uncommit(
struct xfs_trans *tp,
uint flags)
{
- xfs_log_item_desc_t *lidp;
+ struct xfs_log_item_desc *lidp;
- for (lidp = xfs_trans_first_item(tp);
- lidp != NULL;
- lidp = xfs_trans_next_item(tp, lidp)) {
+ list_for_each_entry(lidp, &tp->t_items, lid_trans) {
/*
* Unpin all but those that aren't dirty.
*/
@@ -1504,33 +1588,28 @@ STATIC struct xfs_log_vec *
xfs_trans_alloc_log_vecs(
xfs_trans_t *tp)
{
- xfs_log_item_desc_t *lidp;
+ struct xfs_log_item_desc *lidp;
struct xfs_log_vec *lv = NULL;
struct xfs_log_vec *ret_lv = NULL;
- lidp = xfs_trans_first_item(tp);
/* Bail out if we didn't find a log item. */
- if (!lidp) {
+ if (list_empty(&tp->t_items)) {
ASSERT(0);
return NULL;
}
- while (lidp != NULL) {
+ list_for_each_entry(lidp, &tp->t_items, lid_trans) {
struct xfs_log_vec *new_lv;
/* Skip items which aren't dirty in this transaction. */
- if (!(lidp->lid_flags & XFS_LID_DIRTY)) {
- lidp = xfs_trans_next_item(tp, lidp);
+ if (!(lidp->lid_flags & XFS_LID_DIRTY))
continue;
- }
/* Skip items that do not have any vectors for writing */
lidp->lid_size = IOP_SIZE(lidp->lid_item);
- if (!lidp->lid_size) {
- lidp = xfs_trans_next_item(tp, lidp);
+ if (!lidp->lid_size)
continue;
- }
new_lv = kmem_zalloc(sizeof(*new_lv) +
lidp->lid_size * sizeof(struct xfs_log_iovec),
@@ -1545,7 +1624,6 @@ xfs_trans_alloc_log_vecs(
else
lv->lv_next = new_lv;
lv = new_lv;
- lidp = xfs_trans_next_item(tp, lidp);
}
return ret_lv;
@@ -1704,12 +1782,6 @@ xfs_trans_cancel(
int flags)
{
int log_flags;
-#ifdef DEBUG
- xfs_log_item_chunk_t *licp;
- xfs_log_item_desc_t *lidp;
- xfs_log_item_t *lip;
- int i;
-#endif
xfs_mount_t *mp = tp->t_mountp;
/*
@@ -1728,21 +1800,11 @@ xfs_trans_cancel(
xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
}
#ifdef DEBUG
- if (!(flags & XFS_TRANS_ABORT)) {
- licp = &(tp->t_items);
- while (licp != NULL) {
- lidp = licp->lic_descs;
- for (i = 0; i < licp->lic_unused; i++, lidp++) {
- if (xfs_lic_isfree(licp, i)) {
- continue;
- }
-
- lip = lidp->lid_item;
- if (!XFS_FORCED_SHUTDOWN(mp))
- ASSERT(!(lip->li_type == XFS_LI_EFD));
- }
- licp = licp->lic_next;
- }
+ if (!(flags & XFS_TRANS_ABORT) && !XFS_FORCED_SHUTDOWN(mp)) {
+ struct xfs_log_item_desc *lidp;
+
+ list_for_each_entry(lidp, &tp->t_items, lid_trans)
+ ASSERT(!(lidp->lid_item->li_type == XFS_LI_EFD));
}
#endif
xfs_trans_unreserve_and_mod_sb(tp);
Index: xfs/fs/xfs/xfs_trans.h
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.h 2010-05-25 12:43:44.000000000 +0200
+++ xfs/fs/xfs/xfs_trans.h 2010-05-25 12:43:47.986004051 +0200
@@ -161,105 +161,14 @@ typedef struct xfs_trans_header {
* the amount of space needed to log the item it describes
* once we get to commit processing (see xfs_trans_commit()).
*/
-typedef struct xfs_log_item_desc {
+struct xfs_log_item_desc {
struct xfs_log_item *lid_item;
- ushort lid_size;
- unsigned char lid_flags;
- unsigned char lid_index;
-} xfs_log_item_desc_t;
+ ushort lid_size;
+ unsigned char lid_flags;
+ struct list_head lid_trans;
+};
#define XFS_LID_DIRTY 0x1
-#define XFS_LID_PINNED 0x2
-
-/*
- * This structure is used to maintain a chunk list of log_item_desc
- * structures. The free field is a bitmask indicating which descriptors
- * in this chunk's array are free. The unused field is the first value
- * not used since this chunk was allocated.
- */
-#define XFS_LIC_NUM_SLOTS 15
-typedef struct xfs_log_item_chunk {
- struct xfs_log_item_chunk *lic_next;
- ushort lic_free;
- ushort lic_unused;
- xfs_log_item_desc_t lic_descs[XFS_LIC_NUM_SLOTS];
-} xfs_log_item_chunk_t;
-
-#define XFS_LIC_MAX_SLOT (XFS_LIC_NUM_SLOTS - 1)
-#define XFS_LIC_FREEMASK ((1 << XFS_LIC_NUM_SLOTS) - 1)
-
-
-/*
- * Initialize the given chunk. Set the chunk's free descriptor mask
- * to indicate that all descriptors are free. The caller gets to set
- * lic_unused to the right value (0 matches all free). The
- * lic_descs.lid_index values are set up as each desc is allocated.
- */
-static inline void xfs_lic_init(xfs_log_item_chunk_t *cp)
-{
- cp->lic_free = XFS_LIC_FREEMASK;
-}
-
-static inline void xfs_lic_init_slot(xfs_log_item_chunk_t *cp, int slot)
-{
- cp->lic_descs[slot].lid_index = (unsigned char)(slot);
-}
-
-static inline int xfs_lic_vacancy(xfs_log_item_chunk_t *cp)
-{
- return cp->lic_free & XFS_LIC_FREEMASK;
-}
-
-static inline void xfs_lic_all_free(xfs_log_item_chunk_t *cp)
-{
- cp->lic_free = XFS_LIC_FREEMASK;
-}
-
-static inline int xfs_lic_are_all_free(xfs_log_item_chunk_t *cp)
-{
- return ((cp->lic_free & XFS_LIC_FREEMASK) == XFS_LIC_FREEMASK);
-}
-
-static inline int xfs_lic_isfree(xfs_log_item_chunk_t *cp, int slot)
-{
- return (cp->lic_free & (1 << slot));
-}
-
-static inline void xfs_lic_claim(xfs_log_item_chunk_t *cp, int slot)
-{
- cp->lic_free &= ~(1 << slot);
-}
-
-static inline void xfs_lic_relse(xfs_log_item_chunk_t *cp, int slot)
-{
- cp->lic_free |= 1 << slot;
-}
-
-static inline xfs_log_item_desc_t *
-xfs_lic_slot(xfs_log_item_chunk_t *cp, int slot)
-{
- return &(cp->lic_descs[slot]);
-}
-
-static inline int xfs_lic_desc_to_slot(xfs_log_item_desc_t *dp)
-{
- return (uint)dp->lid_index;
-}
-
-/*
- * Calculate the address of a chunk given a descriptor pointer:
- * dp - dp->lid_index give the address of the start of the lic_descs array.
- * From this we subtract the offset of the lic_descs field in a chunk.
- * All of this yields the address of the chunk, which is
- * cast to a chunk pointer.
- */
-static inline xfs_log_item_chunk_t *
-xfs_lic_desc_to_chunk(xfs_log_item_desc_t *dp)
-{
- return (xfs_log_item_chunk_t*) \
- (((xfs_caddr_t)((dp) - (dp)->lid_index)) - \
- (xfs_caddr_t)(((xfs_log_item_chunk_t*)0)->lic_descs));
-}
#define XFS_TRANS_MAGIC 0x5452414E /* 'TRAN' */
/*
@@ -516,8 +425,7 @@ typedef struct xfs_trans {
int64_t t_rblocks_delta;/* superblock rblocks change */
int64_t t_rextents_delta;/* superblocks rextents chg */
int64_t t_rextslog_delta;/* superblocks rextslog chg */
- unsigned int t_items_free; /* log item descs free */
- xfs_log_item_chunk_t t_items; /* first log item desc chunk */
+ struct list_head t_items; /* log item descriptors */
xfs_trans_header_t t_header; /* header for in-log trans */
struct list_head t_busy; /* list of busy extents */
unsigned long t_pflags; /* saved process flags state */
@@ -595,6 +503,7 @@ int xfs_trans_ail_init(struct xfs_mount
void xfs_trans_ail_destroy(struct xfs_mount *);
extern kmem_zone_t *xfs_trans_zone;
+extern kmem_zone_t *xfs_log_item_desc_zone;
#endif /* __KERNEL__ */
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 06/17] xfs: simplify log item descriptor tracking
2010-05-31 16:07 ` [PATCH 06/17] xfs: simplify log item descriptor tracking Christoph Hellwig
@ 2010-06-02 5:11 ` Dave Chinner
0 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 5:11 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:33PM -0400, Christoph Hellwig wrote:
> Currently we track log item descriptor belonging to a transaction using a
> complex opencoded chunk allocator. This code has been there since day one
> and seems to work around the lack of an efficient slab allocator.
>
> This patch replaces it with dynamically allocated log item descriptors
> from a dedicated slab pool, linked to the transaction by a linked list.
>
> This allows to greatly simplify the log item descriptor tracking to the
> point where it's just a couple hundred lines in xfs_trans.c instead of
> a separate file. The external API has also been simplified while we're
> at it - the xfs_trans_add_item and xfs_trans_del_item functions to add/
> delete items from a transaction have been simplified to the bare minium,
> and the xfs_trans_find_item function is replaced with a direct dereference
> of the li_desc field. All debug code walking the list of log items in
> a transaction is down to a simple list_for_each_entry.
>
> Note that we could easily use a singly linked list here instead of the
> double linked list from list.h as the fastpath only does deletion from
> sequential traversal. But given that we don't have one available as
> a library function yet I use the list.h functions for simplicity.
>
> [the patch applies ontop of Dave's delayed-logging branch]
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Looks OK - nothing jumps out at me as being incorrect. I'm still
running QA on it, but it looks good so far.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* [PATCH 07/17] xfs: merge iop_unpin_remove into iop_unpin
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (5 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 06/17] xfs: simplify log item descriptor tracking Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 5:14 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 08/17] xfs: give xfs_item_ops methods the correct prototypes Christoph Hellwig
` (11 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-merge-IOP-remove --]
[-- Type: text/plain, Size: 14794 bytes --]
The unpin_remove item operation instances always share most of the
implementation with the respective unpin implementation. So instead
of keeping two different entry points add a remove flag to the unpin
operation and share the code more easily.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/quota/xfs_dquot_item.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot_item.c 2010-05-25 11:40:59.115285445 +0200
+++ xfs/fs/xfs/quota/xfs_dquot_item.c 2010-05-25 12:45:02.895004400 +0200
@@ -97,7 +97,8 @@ xfs_qm_dquot_logitem_pin(
/* ARGSUSED */
STATIC void
xfs_qm_dquot_logitem_unpin(
- xfs_dq_logitem_t *logitem)
+ xfs_dq_logitem_t *logitem,
+ int remove)
{
xfs_dquot_t *dqp = logitem->qli_dquot;
@@ -106,15 +107,6 @@ xfs_qm_dquot_logitem_unpin(
wake_up(&dqp->q_pinwait);
}
-/* ARGSUSED */
-STATIC void
-xfs_qm_dquot_logitem_unpin_remove(
- xfs_dq_logitem_t *logitem,
- xfs_trans_t *tp)
-{
- xfs_qm_dquot_logitem_unpin(logitem);
-}
-
/*
* Given the logitem, this writes the corresponding dquot entry to disk
* asynchronously. This is called with the dquot entry securely locked;
@@ -318,9 +310,7 @@ static struct xfs_item_ops xfs_dquot_ite
.iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
xfs_qm_dquot_logitem_format,
.iop_pin = (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_unpin,
- .iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t*))
- xfs_qm_dquot_logitem_unpin_remove,
+ .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_qm_dquot_logitem_unpin,
.iop_trylock = (uint(*)(xfs_log_item_t*))
xfs_qm_dquot_logitem_trylock,
.iop_unlock = (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_unlock,
@@ -413,14 +403,7 @@ xfs_qm_qoff_logitem_pin(xfs_qoff_logitem
*/
/*ARGSUSED*/
STATIC void
-xfs_qm_qoff_logitem_unpin(xfs_qoff_logitem_t *qf)
-{
- return;
-}
-
-/*ARGSUSED*/
-STATIC void
-xfs_qm_qoff_logitem_unpin_remove(xfs_qoff_logitem_t *qf, xfs_trans_t *tp)
+xfs_qm_qoff_logitem_unpin(xfs_qoff_logitem_t *qf, int remove)
{
return;
}
@@ -524,9 +507,7 @@ static struct xfs_item_ops xfs_qm_qoffen
.iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
xfs_qm_qoff_logitem_format,
.iop_pin = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unpin,
- .iop_unpin_remove = (void(*)(xfs_log_item_t*,xfs_trans_t*))
- xfs_qm_qoff_logitem_unpin_remove,
+ .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_qm_qoff_logitem_unpin,
.iop_trylock = (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
.iop_unlock = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
.iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
@@ -545,9 +526,7 @@ static struct xfs_item_ops xfs_qm_qoff_l
.iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
xfs_qm_qoff_logitem_format,
.iop_pin = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unpin,
- .iop_unpin_remove = (void(*)(xfs_log_item_t*,xfs_trans_t*))
- xfs_qm_qoff_logitem_unpin_remove,
+ .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_qm_qoff_logitem_unpin,
.iop_trylock = (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
.iop_unlock = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
.iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
Index: xfs/fs/xfs/xfs_buf_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.c 2010-05-25 12:43:47.926004470 +0200
+++ xfs/fs/xfs/xfs_buf_item.c 2010-05-25 12:45:02.901022489 +0200
@@ -388,20 +388,25 @@ xfs_buf_item_pin(
* Also drop the reference to the buf item for the current transaction.
* If the XFS_BLI_STALE flag is set and we are the last reference,
* then free up the buf log item and unlock the buffer.
+ *
+ * If the remove flag is set we are called from uncommit in the
+ * forced-shutdown path. If that is true and the reference count on
+ * the log item is going to drop to zero we need to free the item's
+ * descriptor in the transaction.
*/
STATIC void
xfs_buf_item_unpin(
- xfs_buf_log_item_t *bip)
+ xfs_buf_log_item_t *bip,
+ int remove)
{
struct xfs_ail *ailp;
- xfs_buf_t *bp;
+ xfs_buf_t *bp = bip->bli_buf;
int freed;
int stale = bip->bli_flags & XFS_BLI_STALE;
- bp = bip->bli_buf;
- ASSERT(bp != NULL);
ASSERT(XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *) == bip);
ASSERT(atomic_read(&bip->bli_refcount) > 0);
+
trace_xfs_buf_item_unpin(bip);
freed = atomic_dec_and_test(&bip->bli_refcount);
@@ -413,8 +418,26 @@ xfs_buf_item_unpin(
ASSERT(!(XFS_BUF_ISDELAYWRITE(bp)));
ASSERT(XFS_BUF_ISSTALE(bp));
ASSERT(bip->bli_format.blf_flags & XFS_BLF_CANCEL);
+
trace_xfs_buf_item_unpin_stale(bip);
+ if (remove) {
+ /*
+ * We have to remove the log item from the transaction
+ * as we are about to release our reference to the
+ * buffer. If we don't, the unlock that occurs later
+ * in xfs_trans_uncommit() will ry to reference the
+ * buffer which we no longer have a hold on.
+ */
+ xfs_trans_del_item(&bip->bli_item);
+
+ /*
+ * Since the transaction no longer refers to the buffer,
+ * the buffer should no longer refer to the transaction.
+ */
+ XFS_BUF_SET_FSPRIVATE2(bp, NULL);
+ }
+
/*
* If we get called here because of an IO error, we may
* or may not have the item on the AIL. xfs_trans_ail_delete()
@@ -436,45 +459,6 @@ xfs_buf_item_unpin(
}
/*
- * this is called from uncommit in the forced-shutdown path.
- * we need to check to see if the reference count on the log item
- * is going to drop to zero. If so, unpin will free the log item
- * so we need to free the item's descriptor (that points to the item)
- * in the transaction.
- */
-STATIC void
-xfs_buf_item_unpin_remove(
- xfs_buf_log_item_t *bip,
- xfs_trans_t *tp)
-{
- /* will xfs_buf_item_unpin() call xfs_buf_item_relse()? */
- if ((atomic_read(&bip->bli_refcount) == 1) &&
- (bip->bli_flags & XFS_BLI_STALE)) {
- /*
- * yes -- We can safely do some work here and then call
- * buf_item_unpin to do the rest because we are
- * are holding the buffer locked so no one else will be
- * able to bump up the refcount. We have to remove the
- * log item from the transaction as we are about to release
- * our reference to the buffer. If we don't, the unlock that
- * occurs later in the xfs_trans_uncommit() will try to
- * reference the buffer which we no longer have a hold on.
- */
- ASSERT(XFS_BUF_VALUSEMA(bip->bli_buf) <= 0);
- trace_xfs_buf_item_unpin_stale(bip);
-
- xfs_trans_del_item(&bip->bli_item);
-
- /*
- * Since the transaction no longer refers to the buffer, the
- * buffer should no longer refer to the transaction.
- */
- XFS_BUF_SET_FSPRIVATE2(bip->bli_buf, NULL);
- }
- xfs_buf_item_unpin(bip);
-}
-
-/*
* This is called to attempt to lock the buffer associated with this
* buf log item. Don't sleep on the buffer lock. If we can't get
* the lock right away, return 0. If we can get the lock, take a
@@ -669,9 +653,7 @@ static struct xfs_item_ops xfs_buf_item_
.iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
xfs_buf_item_format,
.iop_pin = (void(*)(xfs_log_item_t*))xfs_buf_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*))xfs_buf_item_unpin,
- .iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t *))
- xfs_buf_item_unpin_remove,
+ .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_buf_item_unpin,
.iop_trylock = (uint(*)(xfs_log_item_t*))xfs_buf_item_trylock,
.iop_unlock = (void(*)(xfs_log_item_t*))xfs_buf_item_unlock,
.iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
Index: xfs/fs/xfs/xfs_extfree_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_extfree_item.c 2010-05-25 12:43:47.938004470 +0200
+++ xfs/fs/xfs/xfs_extfree_item.c 2010-05-25 12:45:02.903023258 +0200
@@ -103,32 +103,8 @@ xfs_efi_item_pin(xfs_efi_log_item_t *efi
* Here we coordinate with xfs_efi_cancel() to determine who gets to
* free the EFI.
*/
-/*ARGSUSED*/
-STATIC void
-xfs_efi_item_unpin(xfs_efi_log_item_t *efip)
-{
- struct xfs_ail *ailp = efip->efi_item.li_ailp;
-
- spin_lock(&ailp->xa_lock);
- if (efip->efi_flags & XFS_EFI_CANCELED) {
- /* xfs_trans_ail_delete() drops the AIL lock. */
- xfs_trans_ail_delete(ailp, (xfs_log_item_t *)efip);
- xfs_efi_item_free(efip);
- } else {
- efip->efi_flags |= XFS_EFI_COMMITTED;
- spin_unlock(&ailp->xa_lock);
- }
-}
-
-/*
- * like unpin only we have to also clear the xaction descriptor
- * pointing the log item if we free the item. This routine duplicates
- * unpin because efi_flags is protected by the AIL lock. Freeing
- * the descriptor and then calling unpin would force us to drop the AIL
- * lock which would open up a race condition.
- */
STATIC void
-xfs_efi_item_unpin_remove(xfs_efi_log_item_t *efip, xfs_trans_t *tp)
+xfs_efi_item_unpin(xfs_efi_log_item_t *efip, int remove)
{
struct xfs_ail *ailp = efip->efi_item.li_ailp;
@@ -136,10 +112,8 @@ xfs_efi_item_unpin_remove(xfs_efi_log_it
if (efip->efi_flags & XFS_EFI_CANCELED) {
struct xfs_log_item *lip = &efip->efi_item;
- /*
- * free the xaction descriptor pointing to this item
- */
- xfs_trans_del_item(lip);
+ if (remove)
+ xfs_trans_del_item(lip);
/* xfs_trans_ail_delete() drops the AIL lock. */
xfs_trans_ail_delete(ailp, lip);
@@ -223,9 +197,7 @@ static struct xfs_item_ops xfs_efi_item_
.iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
xfs_efi_item_format,
.iop_pin = (void(*)(xfs_log_item_t*))xfs_efi_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*))xfs_efi_item_unpin,
- .iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t *))
- xfs_efi_item_unpin_remove,
+ .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_efi_item_unpin,
.iop_trylock = (uint(*)(xfs_log_item_t*))xfs_efi_item_trylock,
.iop_unlock = (void(*)(xfs_log_item_t*))xfs_efi_item_unlock,
.iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
@@ -424,14 +396,7 @@ xfs_efd_item_pin(xfs_efd_log_item_t *efd
*/
/*ARGSUSED*/
STATIC void
-xfs_efd_item_unpin(xfs_efd_log_item_t *efdp)
-{
- return;
-}
-
-/*ARGSUSED*/
-STATIC void
-xfs_efd_item_unpin_remove(xfs_efd_log_item_t *efdp, xfs_trans_t *tp)
+xfs_efd_item_unpin(xfs_efd_log_item_t *efdp, int remove)
{
return;
}
@@ -514,9 +479,7 @@ static struct xfs_item_ops xfs_efd_item_
.iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
xfs_efd_item_format,
.iop_pin = (void(*)(xfs_log_item_t*))xfs_efd_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*))xfs_efd_item_unpin,
- .iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t*))
- xfs_efd_item_unpin_remove,
+ .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_efd_item_unpin,
.iop_trylock = (uint(*)(xfs_log_item_t*))xfs_efd_item_trylock,
.iop_unlock = (void(*)(xfs_log_item_t*))xfs_efd_item_unlock,
.iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c 2010-05-25 11:40:59.225034362 +0200
+++ xfs/fs/xfs/xfs_inode_item.c 2010-05-25 12:45:02.915017599 +0200
@@ -544,10 +544,10 @@ xfs_inode_item_pin(
*
* Also wake up anyone in xfs_iunpin_wait() if the count goes to 0.
*/
-/* ARGSUSED */
STATIC void
xfs_inode_item_unpin(
- xfs_inode_log_item_t *iip)
+ xfs_inode_log_item_t *iip,
+ int remove)
{
struct xfs_inode *ip = iip->ili_inode;
@@ -557,15 +557,6 @@ xfs_inode_item_unpin(
wake_up(&ip->i_ipin_wait);
}
-/* ARGSUSED */
-STATIC void
-xfs_inode_item_unpin_remove(
- xfs_inode_log_item_t *iip,
- xfs_trans_t *tp)
-{
- xfs_inode_item_unpin(iip);
-}
-
/*
* This is called to attempt to lock the inode associated with this
* inode log item, in preparation for the push routine which does the actual
@@ -829,9 +820,7 @@ static struct xfs_item_ops xfs_inode_ite
.iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
xfs_inode_item_format,
.iop_pin = (void(*)(xfs_log_item_t*))xfs_inode_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*))xfs_inode_item_unpin,
- .iop_unpin_remove = (void(*)(xfs_log_item_t*, xfs_trans_t*))
- xfs_inode_item_unpin_remove,
+ .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_inode_item_unpin,
.iop_trylock = (uint(*)(xfs_log_item_t*))xfs_inode_item_trylock,
.iop_unlock = (void(*)(xfs_log_item_t*))xfs_inode_item_unlock,
.iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
Index: xfs/fs/xfs/xfs_trans.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.c 2010-05-25 12:43:47.977004330 +0200
+++ xfs/fs/xfs/xfs_trans.c 2010-05-25 12:45:02.918006215 +0200
@@ -1375,7 +1375,7 @@ xfs_trans_item_committed(
* log item flags, if anyone else stales the buffer we do not want to
* pay any attention to it.
*/
- IOP_UNPIN(lip);
+ IOP_UNPIN(lip, 0);
}
/*
@@ -1422,7 +1422,7 @@ xfs_trans_uncommit(
* Unpin all but those that aren't dirty.
*/
if (lidp->lid_flags & XFS_LID_DIRTY)
- IOP_UNPIN_REMOVE(lidp->lid_item, tp);
+ IOP_UNPIN(lidp->lid_item, 1);
}
xfs_trans_unreserve_and_mod_sb(tp);
Index: xfs/fs/xfs/xfs_trans.h
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.h 2010-05-25 12:43:47.986004051 +0200
+++ xfs/fs/xfs/xfs_trans.h 2010-05-25 12:45:02.942046166 +0200
@@ -347,8 +347,7 @@ typedef struct xfs_item_ops {
uint (*iop_size)(xfs_log_item_t *);
void (*iop_format)(xfs_log_item_t *, struct xfs_log_iovec *);
void (*iop_pin)(xfs_log_item_t *);
- void (*iop_unpin)(xfs_log_item_t *);
- void (*iop_unpin_remove)(xfs_log_item_t *, struct xfs_trans *);
+ void (*iop_unpin)(xfs_log_item_t *, int remove);
uint (*iop_trylock)(xfs_log_item_t *);
void (*iop_unlock)(xfs_log_item_t *);
xfs_lsn_t (*iop_committed)(xfs_log_item_t *, xfs_lsn_t);
@@ -360,8 +359,7 @@ typedef struct xfs_item_ops {
#define IOP_SIZE(ip) (*(ip)->li_ops->iop_size)(ip)
#define IOP_FORMAT(ip,vp) (*(ip)->li_ops->iop_format)(ip, vp)
#define IOP_PIN(ip) (*(ip)->li_ops->iop_pin)(ip)
-#define IOP_UNPIN(ip) (*(ip)->li_ops->iop_unpin)(ip)
-#define IOP_UNPIN_REMOVE(ip,tp) (*(ip)->li_ops->iop_unpin_remove)(ip, tp)
+#define IOP_UNPIN(ip, remove) (*(ip)->li_ops->iop_unpin)(ip, remove)
#define IOP_TRYLOCK(ip) (*(ip)->li_ops->iop_trylock)(ip)
#define IOP_UNLOCK(ip) (*(ip)->li_ops->iop_unlock)(ip)
#define IOP_COMMITTED(ip, lsn) (*(ip)->li_ops->iop_committed)(ip, lsn)
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 08/17] xfs: give xfs_item_ops methods the correct prototypes
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (6 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 07/17] xfs: merge iop_unpin_remove into iop_unpin Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 5:30 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 09/17] xfs: give li_cb callbacks the correct prototype Christoph Hellwig
` (10 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-iop-prototype-cleanup --]
[-- Type: text/plain, Size: 45727 bytes --]
Stop the function pointer casting madness and give all the xfs_item_ops the
correct prototypes.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/quota/xfs_dquot_item.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot_item.c 2010-05-29 18:12:37.757004346 +0200
+++ xfs/fs/xfs/quota/xfs_dquot_item.c 2010-05-29 18:14:21.307004136 +0200
@@ -37,18 +37,22 @@
#include "xfs_trans_priv.h"
#include "xfs_qm.h"
+static inline struct xfs_dq_logitem *DQUOT_ITEM(struct xfs_log_item *lip)
+{
+ return container_of(lip, struct xfs_dq_logitem, qli_item);
+}
+
/*
* returns the number of iovecs needed to log the given dquot item.
*/
-/* ARGSUSED */
STATIC uint
xfs_qm_dquot_logitem_size(
- xfs_dq_logitem_t *logitem)
+ struct xfs_log_item *lip)
{
/*
* we need only two iovecs, one for the format, one for the real thing
*/
- return (2);
+ return 2;
}
/*
@@ -56,22 +60,21 @@ xfs_qm_dquot_logitem_size(
*/
STATIC void
xfs_qm_dquot_logitem_format(
- xfs_dq_logitem_t *logitem,
- xfs_log_iovec_t *logvec)
+ struct xfs_log_item *lip,
+ struct xfs_log_iovec *logvec)
{
- ASSERT(logitem);
- ASSERT(logitem->qli_dquot);
+ struct xfs_dq_logitem *qlip = DQUOT_ITEM(lip);
- logvec->i_addr = (xfs_caddr_t)&logitem->qli_format;
+ logvec->i_addr = (xfs_caddr_t)&qlip->qli_format;
logvec->i_len = sizeof(xfs_dq_logformat_t);
logvec->i_type = XLOG_REG_TYPE_QFORMAT;
logvec++;
- logvec->i_addr = (xfs_caddr_t)&logitem->qli_dquot->q_core;
+ logvec->i_addr = (xfs_caddr_t)&qlip->qli_dquot->q_core;
logvec->i_len = sizeof(xfs_disk_dquot_t);
logvec->i_type = XLOG_REG_TYPE_DQUOT;
- ASSERT(2 == logitem->qli_item.li_desc->lid_size);
- logitem->qli_format.qlf_size = 2;
+ ASSERT(2 == lip->li_desc->lid_size);
+ qlip->qli_format.qlf_size = 2;
}
@@ -80,9 +83,9 @@ xfs_qm_dquot_logitem_format(
*/
STATIC void
xfs_qm_dquot_logitem_pin(
- xfs_dq_logitem_t *logitem)
+ struct xfs_log_item *lip)
{
- xfs_dquot_t *dqp = logitem->qli_dquot;
+ struct xfs_dquot *dqp = DQUOT_ITEM(lip)->qli_dquot;
ASSERT(XFS_DQ_IS_LOCKED(dqp));
atomic_inc(&dqp->q_pincount);
@@ -94,13 +97,12 @@ xfs_qm_dquot_logitem_pin(
* dquot must have been previously pinned with a call to
* xfs_qm_dquot_logitem_pin().
*/
-/* ARGSUSED */
STATIC void
xfs_qm_dquot_logitem_unpin(
- xfs_dq_logitem_t *logitem,
+ struct xfs_log_item *lip,
int remove)
{
- xfs_dquot_t *dqp = logitem->qli_dquot;
+ struct xfs_dquot *dqp = DQUOT_ITEM(lip)->qli_dquot;
ASSERT(atomic_read(&dqp->q_pincount) > 0);
if (atomic_dec_and_test(&dqp->q_pincount))
@@ -115,12 +117,10 @@ xfs_qm_dquot_logitem_unpin(
*/
STATIC void
xfs_qm_dquot_logitem_push(
- xfs_dq_logitem_t *logitem)
+ struct xfs_log_item *lip)
{
- xfs_dquot_t *dqp;
- int error;
-
- dqp = logitem->qli_dquot;
+ struct xfs_dquot *dqp = DQUOT_ITEM(lip)->qli_dquot;
+ int error;
ASSERT(XFS_DQ_IS_LOCKED(dqp));
ASSERT(!completion_done(&dqp->q_flush));
@@ -142,27 +142,25 @@ xfs_qm_dquot_logitem_push(
xfs_dqunlock(dqp);
}
-/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_qm_dquot_logitem_committed(
- xfs_dq_logitem_t *l,
+ struct xfs_log_item *lip,
xfs_lsn_t lsn)
{
/*
* We always re-log the entire dquot when it becomes dirty,
* so, the latest copy _is_ the only one that matters.
*/
- return (lsn);
+ return lsn;
}
-
/*
* This is called to wait for the given dquot to be unpinned.
* Most of these pin/unpin routines are plagiarized from inode code.
*/
void
xfs_qm_dqunpin_wait(
- xfs_dquot_t *dqp)
+ struct xfs_dquot *dqp)
{
ASSERT(XFS_DQ_IS_LOCKED(dqp));
if (atomic_read(&dqp->q_pincount) == 0)
@@ -188,13 +186,12 @@ xfs_qm_dqunpin_wait(
*/
STATIC void
xfs_qm_dquot_logitem_pushbuf(
- xfs_dq_logitem_t *qip)
+ struct xfs_log_item *lip)
{
- xfs_dquot_t *dqp;
- xfs_mount_t *mp;
- xfs_buf_t *bp;
+ struct xfs_dq_logitem *qlip = DQUOT_ITEM(lip);
+ struct xfs_dquot *dqp = qlip->qli_dquot;
+ struct xfs_buf *bp;
- dqp = qip->qli_dquot;
ASSERT(XFS_DQ_IS_LOCKED(dqp));
/*
@@ -202,22 +199,20 @@ xfs_qm_dquot_logitem_pushbuf(
* inode flush completed and the inode was taken off the AIL.
* So, just get out.
*/
- if (completion_done(&dqp->q_flush) ||
- ((qip->qli_item.li_flags & XFS_LI_IN_AIL) == 0)) {
+ if (completion_done(&dqp->q_flush) ||
+ !(lip->li_flags & XFS_LI_IN_AIL)) {
xfs_dqunlock(dqp);
return;
}
- mp = dqp->q_mount;
- bp = xfs_incore(mp->m_ddev_targp, qip->qli_format.qlf_blkno,
- mp->m_quotainfo->qi_dqchunklen, XBF_TRYLOCK);
+
+ bp = xfs_incore(dqp->q_mount->m_ddev_targp, qlip->qli_format.qlf_blkno,
+ dqp->q_mount->m_quotainfo->qi_dqchunklen, XBF_TRYLOCK);
xfs_dqunlock(dqp);
if (!bp)
return;
if (XFS_BUF_ISDELAYWRITE(bp))
xfs_buf_delwri_promote(bp);
xfs_buf_relse(bp);
- return;
-
}
/*
@@ -232,15 +227,14 @@ xfs_qm_dquot_logitem_pushbuf(
*/
STATIC uint
xfs_qm_dquot_logitem_trylock(
- xfs_dq_logitem_t *qip)
+ struct xfs_log_item *lip)
{
- xfs_dquot_t *dqp;
+ struct xfs_dquot *dqp = DQUOT_ITEM(lip)->qli_dquot;
- dqp = qip->qli_dquot;
if (atomic_read(&dqp->q_pincount) > 0)
return XFS_ITEM_PINNED;
- if (! xfs_qm_dqlock_nowait(dqp))
+ if (!xfs_qm_dqlock_nowait(dqp))
return XFS_ITEM_LOCKED;
if (!xfs_dqflock_nowait(dqp)) {
@@ -251,11 +245,10 @@ xfs_qm_dquot_logitem_trylock(
return XFS_ITEM_PUSHBUF;
}
- ASSERT(qip->qli_item.li_flags & XFS_LI_IN_AIL);
+ ASSERT(lip->li_flags & XFS_LI_IN_AIL);
return XFS_ITEM_SUCCESS;
}
-
/*
* Unlock the dquot associated with the log item.
* Clear the fields of the dquot and dquot log item that
@@ -264,12 +257,10 @@ xfs_qm_dquot_logitem_trylock(
*/
STATIC void
xfs_qm_dquot_logitem_unlock(
- xfs_dq_logitem_t *ql)
+ struct xfs_log_item *lip)
{
- xfs_dquot_t *dqp;
+ struct xfs_dquot *dqp = DQUOT_ITEM(lip)->qli_dquot;
- ASSERT(ql != NULL);
- dqp = ql->qli_dquot;
ASSERT(XFS_DQ_IS_LOCKED(dqp));
/*
@@ -286,41 +277,32 @@ xfs_qm_dquot_logitem_unlock(
xfs_dqunlock(dqp);
}
-
/*
* this needs to stamp an lsn into the dquot, I think.
* rpc's that look at user dquot's would then have to
* push on the dependency recorded in the dquot
*/
-/* ARGSUSED */
STATIC void
xfs_qm_dquot_logitem_committing(
- xfs_dq_logitem_t *l,
+ struct xfs_log_item *lip,
xfs_lsn_t lsn)
{
- return;
}
-
/*
* This is the ops vector for dquots
*/
static struct xfs_item_ops xfs_dquot_item_ops = {
- .iop_size = (uint(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_size,
- .iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
- xfs_qm_dquot_logitem_format,
- .iop_pin = (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_qm_dquot_logitem_unpin,
- .iop_trylock = (uint(*)(xfs_log_item_t*))
- xfs_qm_dquot_logitem_trylock,
- .iop_unlock = (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_unlock,
- .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_qm_dquot_logitem_committed,
- .iop_push = (void(*)(xfs_log_item_t*))xfs_qm_dquot_logitem_push,
- .iop_pushbuf = (void(*)(xfs_log_item_t*))
- xfs_qm_dquot_logitem_pushbuf,
- .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_qm_dquot_logitem_committing
+ .iop_size = xfs_qm_dquot_logitem_size,
+ .iop_format = xfs_qm_dquot_logitem_format,
+ .iop_pin = xfs_qm_dquot_logitem_pin,
+ .iop_unpin = xfs_qm_dquot_logitem_unpin,
+ .iop_trylock = xfs_qm_dquot_logitem_trylock,
+ .iop_unlock = xfs_qm_dquot_logitem_unlock,
+ .iop_committed = xfs_qm_dquot_logitem_committed,
+ .iop_push = xfs_qm_dquot_logitem_push,
+ .iop_pushbuf = xfs_qm_dquot_logitem_pushbuf,
+ .iop_committing = xfs_qm_dquot_logitem_committing
};
/*
@@ -330,10 +312,9 @@ static struct xfs_item_ops xfs_dquot_ite
*/
void
xfs_qm_dquot_logitem_init(
- struct xfs_dquot *dqp)
+ struct xfs_dquot *dqp)
{
- xfs_dq_logitem_t *lp;
- lp = &dqp->q_logitem;
+ struct xfs_dq_logitem *lp = &dqp->q_logitem;
xfs_log_item_init(dqp->q_mount, &lp->qli_item, XFS_LI_DQUOT,
&xfs_dquot_item_ops);
@@ -354,16 +335,22 @@ xfs_qm_dquot_logitem_init(
/*------------------ QUOTAOFF LOG ITEMS -------------------*/
+static inline struct xfs_qoff_logitem *QOFF_ITEM(struct xfs_log_item *lip)
+{
+ return container_of(lip, struct xfs_qoff_logitem, qql_item);
+}
+
+
/*
* This returns the number of iovecs needed to log the given quotaoff item.
* We only need 1 iovec for an quotaoff item. It just logs the
* quotaoff_log_format structure.
*/
-/*ARGSUSED*/
STATIC uint
-xfs_qm_qoff_logitem_size(xfs_qoff_logitem_t *qf)
+xfs_qm_qoff_logitem_size(
+ struct xfs_log_item *lip)
{
- return (1);
+ return 1;
}
/*
@@ -374,46 +361,46 @@ xfs_qm_qoff_logitem_size(xfs_qoff_logite
* slots in the quotaoff item have been filled.
*/
STATIC void
-xfs_qm_qoff_logitem_format(xfs_qoff_logitem_t *qf,
- xfs_log_iovec_t *log_vector)
+xfs_qm_qoff_logitem_format(
+ struct xfs_log_item *lip,
+ struct xfs_log_iovec *log_vector)
{
- ASSERT(qf->qql_format.qf_type == XFS_LI_QUOTAOFF);
+ struct xfs_qoff_logitem *qflip = QOFF_ITEM(lip);
- log_vector->i_addr = (xfs_caddr_t)&(qf->qql_format);
+ ASSERT(qflip->qql_format.qf_type == XFS_LI_QUOTAOFF);
+
+ log_vector->i_addr = (xfs_caddr_t)&(qflip->qql_format);
log_vector->i_len = sizeof(xfs_qoff_logitem_t);
log_vector->i_type = XLOG_REG_TYPE_QUOTAOFF;
- qf->qql_format.qf_size = 1;
+ qflip->qql_format.qf_size = 1;
}
-
/*
* Pinning has no meaning for an quotaoff item, so just return.
*/
-/*ARGSUSED*/
STATIC void
-xfs_qm_qoff_logitem_pin(xfs_qoff_logitem_t *qf)
+xfs_qm_qoff_logitem_pin(
+ struct xfs_log_item *lip)
{
- return;
}
-
/*
* Since pinning has no meaning for an quotaoff item, unpinning does
* not either.
*/
-/*ARGSUSED*/
STATIC void
-xfs_qm_qoff_logitem_unpin(xfs_qoff_logitem_t *qf, int remove)
+xfs_qm_qoff_logitem_unpin(
+ struct xfs_log_item *lip,
+ int remove)
{
- return;
}
/*
* Quotaoff items have no locking, so just return success.
*/
-/*ARGSUSED*/
STATIC uint
-xfs_qm_qoff_logitem_trylock(xfs_qoff_logitem_t *qf)
+xfs_qm_qoff_logitem_trylock(
+ struct xfs_log_item *lip)
{
return XFS_ITEM_LOCKED;
}
@@ -422,53 +409,51 @@ xfs_qm_qoff_logitem_trylock(xfs_qoff_log
* Quotaoff items have no locking or pushing, so return failure
* so that the caller doesn't bother with us.
*/
-/*ARGSUSED*/
STATIC void
-xfs_qm_qoff_logitem_unlock(xfs_qoff_logitem_t *qf)
+xfs_qm_qoff_logitem_unlock(
+ struct xfs_log_item *lip)
{
- return;
}
/*
* The quotaoff-start-item is logged only once and cannot be moved in the log,
* so simply return the lsn at which it's been logged.
*/
-/*ARGSUSED*/
STATIC xfs_lsn_t
-xfs_qm_qoff_logitem_committed(xfs_qoff_logitem_t *qf, xfs_lsn_t lsn)
+xfs_qm_qoff_logitem_committed(
+ struct xfs_log_item *lip,
+ xfs_lsn_t lsn)
{
- return (lsn);
+ return lsn;
}
/*
* There isn't much you can do to push on an quotaoff item. It is simply
* stuck waiting for the log to be flushed to disk.
*/
-/*ARGSUSED*/
STATIC void
-xfs_qm_qoff_logitem_push(xfs_qoff_logitem_t *qf)
+xfs_qm_qoff_logitem_push(
+ struct xfs_log_item *lip)
{
- return;
}
-/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_qm_qoffend_logitem_committed(
- xfs_qoff_logitem_t *qfe,
- xfs_lsn_t lsn)
+ struct xfs_log_item *lip,
+ xfs_lsn_t lsn)
{
- xfs_qoff_logitem_t *qfs;
- struct xfs_ail *ailp;
+ struct xfs_qoff_logitem *qfe = QOFF_ITEM(lip);
+ struct xfs_qoff_logitem *qfs = qfe->qql_start_lip;
+ struct xfs_ail *ailp = qfs->qql_item.li_ailp;
- qfs = qfe->qql_start_lip;
- ailp = qfs->qql_item.li_ailp;
- spin_lock(&ailp->xa_lock);
/*
* Delete the qoff-start logitem from the AIL.
* xfs_trans_ail_delete() drops the AIL lock.
*/
+ spin_lock(&ailp->xa_lock);
xfs_trans_ail_delete(ailp, (xfs_log_item_t *)qfs);
+
kmem_free(qfs);
kmem_free(qfe);
return (xfs_lsn_t)-1;
@@ -488,67 +473,52 @@ xfs_qm_qoffend_logitem_committed(
* (truly makes the quotaoff irrevocable). If we do something else,
* then maybe we don't need two.
*/
-/* ARGSUSED */
-STATIC void
-xfs_qm_qoff_logitem_committing(xfs_qoff_logitem_t *qip, xfs_lsn_t commit_lsn)
-{
- return;
-}
-
-/* ARGSUSED */
STATIC void
-xfs_qm_qoffend_logitem_committing(xfs_qoff_logitem_t *qip, xfs_lsn_t commit_lsn)
+xfs_qm_qoff_logitem_committing(
+ struct xfs_log_item *lip,
+ xfs_lsn_t commit_lsn)
{
- return;
}
static struct xfs_item_ops xfs_qm_qoffend_logitem_ops = {
- .iop_size = (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_size,
- .iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
- xfs_qm_qoff_logitem_format,
- .iop_pin = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_qm_qoff_logitem_unpin,
- .iop_trylock = (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
- .iop_unlock = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
- .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_qm_qoffend_logitem_committed,
- .iop_push = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_push,
- .iop_pushbuf = NULL,
- .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_qm_qoffend_logitem_committing
+ .iop_size = xfs_qm_qoff_logitem_size,
+ .iop_format = xfs_qm_qoff_logitem_format,
+ .iop_pin = xfs_qm_qoff_logitem_pin,
+ .iop_unpin = xfs_qm_qoff_logitem_unpin,
+ .iop_trylock = xfs_qm_qoff_logitem_trylock,
+ .iop_unlock = xfs_qm_qoff_logitem_unlock,
+ .iop_committed = xfs_qm_qoffend_logitem_committed,
+ .iop_push = xfs_qm_qoff_logitem_push,
+ .iop_committing = xfs_qm_qoff_logitem_committing
};
/*
* This is the ops vector shared by all quotaoff-start log items.
*/
static struct xfs_item_ops xfs_qm_qoff_logitem_ops = {
- .iop_size = (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_size,
- .iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
- xfs_qm_qoff_logitem_format,
- .iop_pin = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_qm_qoff_logitem_unpin,
- .iop_trylock = (uint(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_trylock,
- .iop_unlock = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_unlock,
- .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_qm_qoff_logitem_committed,
- .iop_push = (void(*)(xfs_log_item_t*))xfs_qm_qoff_logitem_push,
- .iop_pushbuf = NULL,
- .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_qm_qoff_logitem_committing
+ .iop_size = xfs_qm_qoff_logitem_size,
+ .iop_format = xfs_qm_qoff_logitem_format,
+ .iop_pin = xfs_qm_qoff_logitem_pin,
+ .iop_unpin = xfs_qm_qoff_logitem_unpin,
+ .iop_trylock = xfs_qm_qoff_logitem_trylock,
+ .iop_unlock = xfs_qm_qoff_logitem_unlock,
+ .iop_committed = xfs_qm_qoff_logitem_committed,
+ .iop_push = xfs_qm_qoff_logitem_push,
+ .iop_committing = xfs_qm_qoff_logitem_committing
};
/*
* Allocate and initialize an quotaoff item of the correct quota type(s).
*/
-xfs_qoff_logitem_t *
+struct xfs_qoff_logitem *
xfs_qm_qoff_logitem_init(
- struct xfs_mount *mp,
- xfs_qoff_logitem_t *start,
- uint flags)
+ struct xfs_mount *mp,
+ struct xfs_qoff_logitem *start,
+ uint flags)
{
- xfs_qoff_logitem_t *qf;
+ struct xfs_qoff_logitem *qf;
- qf = (xfs_qoff_logitem_t*) kmem_zalloc(sizeof(xfs_qoff_logitem_t), KM_SLEEP);
+ qf = kmem_zalloc(sizeof(struct xfs_qoff_logitem), KM_SLEEP);
xfs_log_item_init(mp, &qf->qql_item, XFS_LI_QUOTAOFF, start ?
&xfs_qm_qoffend_logitem_ops : &xfs_qm_qoff_logitem_ops);
@@ -556,5 +526,5 @@ xfs_qm_qoff_logitem_init(
qf->qql_format.qf_type = XFS_LI_QUOTAOFF;
qf->qql_format.qf_flags = flags;
qf->qql_start_lip = start;
- return (qf);
+ return qf;
}
Index: xfs/fs/xfs/xfs_extfree_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_extfree_item.c 2010-05-29 18:12:37.767005742 +0200
+++ xfs/fs/xfs/xfs_extfree_item.c 2010-05-29 18:14:21.308004066 +0200
@@ -32,18 +32,19 @@
kmem_zone_t *xfs_efi_zone;
kmem_zone_t *xfs_efd_zone;
-STATIC void xfs_efi_item_unlock(xfs_efi_log_item_t *);
+static inline struct xfs_efi_log_item *EFI_ITEM(struct xfs_log_item *lip)
+{
+ return container_of(lip, struct xfs_efi_log_item, efi_item);
+}
void
-xfs_efi_item_free(xfs_efi_log_item_t *efip)
+xfs_efi_item_free(
+ struct xfs_efi_log_item *efip)
{
- int nexts = efip->efi_format.efi_nextents;
-
- if (nexts > XFS_EFI_MAX_FAST_EXTENTS) {
+ if (efip->efi_format.efi_nextents > XFS_EFI_MAX_FAST_EXTENTS)
kmem_free(efip);
- } else {
+ else
kmem_zone_free(xfs_efi_zone, efip);
- }
}
/*
@@ -51,9 +52,9 @@ xfs_efi_item_free(xfs_efi_log_item_t *ef
* We only need 1 iovec for an efi item. It just logs the efi_log_format
* structure.
*/
-/*ARGSUSED*/
STATIC uint
-xfs_efi_item_size(xfs_efi_log_item_t *efip)
+xfs_efi_item_size(
+ struct xfs_log_item *lip)
{
return 1;
}
@@ -66,10 +67,12 @@ xfs_efi_item_size(xfs_efi_log_item_t *ef
* slots in the efi item have been filled.
*/
STATIC void
-xfs_efi_item_format(xfs_efi_log_item_t *efip,
- xfs_log_iovec_t *log_vector)
+xfs_efi_item_format(
+ struct xfs_log_item *lip,
+ struct xfs_log_iovec *log_vector)
{
- uint size;
+ struct xfs_efi_log_item *efip = EFI_ITEM(lip);
+ uint size;
ASSERT(efip->efi_next_extent == efip->efi_format.efi_nextents);
@@ -79,7 +82,7 @@ xfs_efi_item_format(xfs_efi_log_item_t *
size += (efip->efi_format.efi_nextents - 1) * sizeof(xfs_extent_t);
efip->efi_format.efi_size = 1;
- log_vector->i_addr = (xfs_caddr_t)&(efip->efi_format);
+ log_vector->i_addr = (xfs_caddr_t)&efip->efi_format;
log_vector->i_len = size;
log_vector->i_type = XLOG_REG_TYPE_EFI_FORMAT;
ASSERT(size >= sizeof(xfs_efi_log_format_t));
@@ -89,14 +92,12 @@ xfs_efi_item_format(xfs_efi_log_item_t *
/*
* Pinning has no meaning for an efi item, so just return.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efi_item_pin(xfs_efi_log_item_t *efip)
+xfs_efi_item_pin(
+ struct xfs_log_item *lip)
{
- return;
}
-
/*
* While EFIs cannot really be pinned, the unpin operation is the
* last place at which the EFI is manipulated during a transaction.
@@ -104,14 +105,15 @@ xfs_efi_item_pin(xfs_efi_log_item_t *efi
* free the EFI.
*/
STATIC void
-xfs_efi_item_unpin(xfs_efi_log_item_t *efip, int remove)
+xfs_efi_item_unpin(
+ struct xfs_log_item *lip,
+ int remove)
{
- struct xfs_ail *ailp = efip->efi_item.li_ailp;
+ struct xfs_efi_log_item *efip = EFI_ITEM(lip);
+ struct xfs_ail *ailp = lip->li_ailp;
spin_lock(&ailp->xa_lock);
if (efip->efi_flags & XFS_EFI_CANCELED) {
- struct xfs_log_item *lip = &efip->efi_item;
-
if (remove)
xfs_trans_del_item(lip);
@@ -131,9 +133,9 @@ xfs_efi_item_unpin(xfs_efi_log_item_t *e
* XFS_ITEM_PINNED so that the caller will eventually flush the log.
* This should help in getting the EFI out of the AIL.
*/
-/*ARGSUSED*/
STATIC uint
-xfs_efi_item_trylock(xfs_efi_log_item_t *efip)
+xfs_efi_item_trylock(
+ struct xfs_log_item *lip)
{
return XFS_ITEM_PINNED;
}
@@ -141,13 +143,12 @@ xfs_efi_item_trylock(xfs_efi_log_item_t
/*
* Efi items have no locking, so just return.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efi_item_unlock(xfs_efi_log_item_t *efip)
+xfs_efi_item_unlock(
+ struct xfs_log_item *lip)
{
- if (efip->efi_item.li_flags & XFS_LI_ABORTED)
- xfs_efi_item_free(efip);
- return;
+ if (lip->li_flags & XFS_LI_ABORTED)
+ xfs_efi_item_free(EFI_ITEM(lip));
}
/*
@@ -156,9 +157,10 @@ xfs_efi_item_unlock(xfs_efi_log_item_t *
* flag is not paid any attention here. Checking for that is delayed
* until the EFI is unpinned.
*/
-/*ARGSUSED*/
STATIC xfs_lsn_t
-xfs_efi_item_committed(xfs_efi_log_item_t *efip, xfs_lsn_t lsn)
+xfs_efi_item_committed(
+ struct xfs_log_item *lip,
+ xfs_lsn_t lsn)
{
return lsn;
}
@@ -168,11 +170,10 @@ xfs_efi_item_committed(xfs_efi_log_item_
* stuck waiting for all of its corresponding efd items to be
* committed to disk.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efi_item_push(xfs_efi_log_item_t *efip)
+xfs_efi_item_push(
+ struct xfs_log_item *lip)
{
- return;
}
/*
@@ -182,59 +183,55 @@ xfs_efi_item_push(xfs_efi_log_item_t *ef
* example, for inodes, the inode is locked throughout the extent freeing
* so the dependency should be recorded there.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efi_item_committing(xfs_efi_log_item_t *efip, xfs_lsn_t lsn)
+xfs_efi_item_committing(
+ struct xfs_log_item *lip,
+ xfs_lsn_t lsn)
{
- return;
}
/*
* This is the ops vector shared by all efi log items.
*/
static struct xfs_item_ops xfs_efi_item_ops = {
- .iop_size = (uint(*)(xfs_log_item_t*))xfs_efi_item_size,
- .iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
- xfs_efi_item_format,
- .iop_pin = (void(*)(xfs_log_item_t*))xfs_efi_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_efi_item_unpin,
- .iop_trylock = (uint(*)(xfs_log_item_t*))xfs_efi_item_trylock,
- .iop_unlock = (void(*)(xfs_log_item_t*))xfs_efi_item_unlock,
- .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_efi_item_committed,
- .iop_push = (void(*)(xfs_log_item_t*))xfs_efi_item_push,
- .iop_pushbuf = NULL,
- .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_efi_item_committing
+ .iop_size = xfs_efi_item_size,
+ .iop_format = xfs_efi_item_format,
+ .iop_pin = xfs_efi_item_pin,
+ .iop_unpin = xfs_efi_item_unpin,
+ .iop_trylock = xfs_efi_item_trylock,
+ .iop_unlock = xfs_efi_item_unlock,
+ .iop_committed = xfs_efi_item_committed,
+ .iop_push = xfs_efi_item_push,
+ .iop_committing = xfs_efi_item_committing
};
/*
* Allocate and initialize an efi item with the given number of extents.
*/
-xfs_efi_log_item_t *
-xfs_efi_init(xfs_mount_t *mp,
- uint nextents)
+struct xfs_efi_log_item *
+xfs_efi_init(
+ struct xfs_mount *mp,
+ uint nextents)
{
- xfs_efi_log_item_t *efip;
+ struct xfs_efi_log_item *efip;
uint size;
ASSERT(nextents > 0);
if (nextents > XFS_EFI_MAX_FAST_EXTENTS) {
size = (uint)(sizeof(xfs_efi_log_item_t) +
((nextents - 1) * sizeof(xfs_extent_t)));
- efip = (xfs_efi_log_item_t*)kmem_zalloc(size, KM_SLEEP);
+ efip = kmem_zalloc(size, KM_SLEEP);
} else {
- efip = (xfs_efi_log_item_t*)kmem_zone_zalloc(xfs_efi_zone,
- KM_SLEEP);
+ efip = kmem_zone_zalloc(xfs_efi_zone, KM_SLEEP);
}
xfs_log_item_init(mp, &efip->efi_item, XFS_LI_EFI, &xfs_efi_item_ops);
efip->efi_format.efi_nextents = nextents;
efip->efi_format.efi_id = (__psint_t)(void*)efip;
- return (efip);
+ return efip;
}
/*
@@ -327,16 +324,18 @@ xfs_efi_release(xfs_efi_log_item_t *efip
}
}
-STATIC void
-xfs_efd_item_free(xfs_efd_log_item_t *efdp)
+static inline struct xfs_efd_log_item *EFD_ITEM(struct xfs_log_item *lip)
{
- int nexts = efdp->efd_format.efd_nextents;
+ return container_of(lip, struct xfs_efd_log_item, efd_item);
+}
- if (nexts > XFS_EFD_MAX_FAST_EXTENTS) {
+STATIC void
+xfs_efd_item_free(struct xfs_efd_log_item *efdp)
+{
+ if (efdp->efd_format.efd_nextents > XFS_EFD_MAX_FAST_EXTENTS)
kmem_free(efdp);
- } else {
+ else
kmem_zone_free(xfs_efd_zone, efdp);
- }
}
/*
@@ -344,9 +343,9 @@ xfs_efd_item_free(xfs_efd_log_item_t *ef
* We only need 1 iovec for an efd item. It just logs the efd_log_format
* structure.
*/
-/*ARGSUSED*/
STATIC uint
-xfs_efd_item_size(xfs_efd_log_item_t *efdp)
+xfs_efd_item_size(
+ struct xfs_log_item *lip)
{
return 1;
}
@@ -359,10 +358,12 @@ xfs_efd_item_size(xfs_efd_log_item_t *ef
* slots in the efd item have been filled.
*/
STATIC void
-xfs_efd_item_format(xfs_efd_log_item_t *efdp,
- xfs_log_iovec_t *log_vector)
+xfs_efd_item_format(
+ struct xfs_log_item *lip,
+ struct xfs_log_iovec *log_vector)
{
- uint size;
+ struct xfs_efd_log_item *efdp = EFD_ITEM(lip);
+ uint size;
ASSERT(efdp->efd_next_extent == efdp->efd_format.efd_nextents);
@@ -372,41 +373,38 @@ xfs_efd_item_format(xfs_efd_log_item_t *
size += (efdp->efd_format.efd_nextents - 1) * sizeof(xfs_extent_t);
efdp->efd_format.efd_size = 1;
- log_vector->i_addr = (xfs_caddr_t)&(efdp->efd_format);
+ log_vector->i_addr = (xfs_caddr_t)&efdp->efd_format;
log_vector->i_len = size;
log_vector->i_type = XLOG_REG_TYPE_EFD_FORMAT;
ASSERT(size >= sizeof(xfs_efd_log_format_t));
}
-
/*
* Pinning has no meaning for an efd item, so just return.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efd_item_pin(xfs_efd_log_item_t *efdp)
+xfs_efd_item_pin(
+ struct xfs_log_item *lip)
{
- return;
}
-
/*
* Since pinning has no meaning for an efd item, unpinning does
* not either.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efd_item_unpin(xfs_efd_log_item_t *efdp, int remove)
+xfs_efd_item_unpin(
+ struct xfs_log_item *lip,
+ int remove)
{
- return;
}
/*
* Efd items have no locking, so just return success.
*/
-/*ARGSUSED*/
STATIC uint
-xfs_efd_item_trylock(xfs_efd_log_item_t *efdp)
+xfs_efd_item_trylock(
+ struct xfs_log_item *lip)
{
return XFS_ITEM_LOCKED;
}
@@ -415,13 +413,12 @@ xfs_efd_item_trylock(xfs_efd_log_item_t
* Efd items have no locking or pushing, so return failure
* so that the caller doesn't bother with us.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efd_item_unlock(xfs_efd_log_item_t *efdp)
+xfs_efd_item_unlock(
+ struct xfs_log_item *lip)
{
- if (efdp->efd_item.li_flags & XFS_LI_ABORTED)
- xfs_efd_item_free(efdp);
- return;
+ if (lip->li_flags & XFS_LI_ABORTED)
+ xfs_efd_item_free(EFD_ITEM(lip));
}
/*
@@ -431,15 +428,18 @@ xfs_efd_item_unlock(xfs_efd_log_item_t *
* return -1 to keep the transaction code from further referencing
* this item.
*/
-/*ARGSUSED*/
STATIC xfs_lsn_t
-xfs_efd_item_committed(xfs_efd_log_item_t *efdp, xfs_lsn_t lsn)
+xfs_efd_item_committed(
+ struct xfs_log_item *lip,
+ xfs_lsn_t lsn)
{
+ struct xfs_efd_log_item *efdp = EFD_ITEM(lip);
+
/*
* If we got a log I/O error, it's always the case that the LR with the
* EFI got unpinned and freed before the EFD got aborted.
*/
- if ((efdp->efd_item.li_flags & XFS_LI_ABORTED) == 0)
+ if (!(lip->li_flags & XFS_LI_ABORTED))
xfs_efi_release(efdp->efd_efip, efdp->efd_format.efd_nextents);
xfs_efd_item_free(efdp);
@@ -450,11 +450,10 @@ xfs_efd_item_committed(xfs_efd_log_item_
* There isn't much you can do to push on an efd item. It is simply
* stuck waiting for the log to be flushed to disk.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efd_item_push(xfs_efd_log_item_t *efdp)
+xfs_efd_item_push(
+ struct xfs_log_item *lip)
{
- return;
}
/*
@@ -464,53 +463,48 @@ xfs_efd_item_push(xfs_efd_log_item_t *ef
* example, for inodes, the inode is locked throughout the extent freeing
* so the dependency should be recorded there.
*/
-/*ARGSUSED*/
STATIC void
-xfs_efd_item_committing(xfs_efd_log_item_t *efip, xfs_lsn_t lsn)
+xfs_efd_item_committing(
+ struct xfs_log_item *lip,
+ xfs_lsn_t lsn)
{
- return;
}
/*
* This is the ops vector shared by all efd log items.
*/
static struct xfs_item_ops xfs_efd_item_ops = {
- .iop_size = (uint(*)(xfs_log_item_t*))xfs_efd_item_size,
- .iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
- xfs_efd_item_format,
- .iop_pin = (void(*)(xfs_log_item_t*))xfs_efd_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_efd_item_unpin,
- .iop_trylock = (uint(*)(xfs_log_item_t*))xfs_efd_item_trylock,
- .iop_unlock = (void(*)(xfs_log_item_t*))xfs_efd_item_unlock,
- .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_efd_item_committed,
- .iop_push = (void(*)(xfs_log_item_t*))xfs_efd_item_push,
- .iop_pushbuf = NULL,
- .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_efd_item_committing
+ .iop_size = xfs_efd_item_size,
+ .iop_format = xfs_efd_item_format,
+ .iop_pin = xfs_efd_item_pin,
+ .iop_unpin = xfs_efd_item_unpin,
+ .iop_trylock = xfs_efd_item_trylock,
+ .iop_unlock = xfs_efd_item_unlock,
+ .iop_committed = xfs_efd_item_committed,
+ .iop_push = xfs_efd_item_push,
+ .iop_committing = xfs_efd_item_committing
};
-
/*
* Allocate and initialize an efd item with the given number of extents.
*/
-xfs_efd_log_item_t *
-xfs_efd_init(xfs_mount_t *mp,
- xfs_efi_log_item_t *efip,
- uint nextents)
+struct xfs_efd_log_item *
+xfs_efd_init(
+ struct xfs_mount *mp,
+ struct xfs_efi_log_item *efip,
+ uint nextents)
{
- xfs_efd_log_item_t *efdp;
+ struct xfs_efd_log_item *efdp;
uint size;
ASSERT(nextents > 0);
if (nextents > XFS_EFD_MAX_FAST_EXTENTS) {
size = (uint)(sizeof(xfs_efd_log_item_t) +
((nextents - 1) * sizeof(xfs_extent_t)));
- efdp = (xfs_efd_log_item_t*)kmem_zalloc(size, KM_SLEEP);
+ efdp = kmem_zalloc(size, KM_SLEEP);
} else {
- efdp = (xfs_efd_log_item_t*)kmem_zone_zalloc(xfs_efd_zone,
- KM_SLEEP);
+ efdp = kmem_zone_zalloc(xfs_efd_zone, KM_SLEEP);
}
xfs_log_item_init(mp, &efdp->efd_item, XFS_LI_EFD, &xfs_efd_item_ops);
@@ -518,5 +512,5 @@ xfs_efd_init(xfs_mount_t *mp,
efdp->efd_format.efd_nextents = nextents;
efdp->efd_format.efd_efi_id = efip->efi_format.efi_id;
- return (efdp);
+ return efdp;
}
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c 2010-05-29 18:12:37.772005254 +0200
+++ xfs/fs/xfs/xfs_inode_item.c 2010-05-29 18:14:21.313003857 +0200
@@ -36,6 +36,12 @@
kmem_zone_t *xfs_ili_zone; /* inode log item zone */
+static inline struct xfs_inode_log_item *INODE_ITEM(struct xfs_log_item *lip)
+{
+ return container_of(lip, struct xfs_inode_log_item, ili_item);
+}
+
+
/*
* This returns the number of iovecs needed to log the given inode item.
*
@@ -45,13 +51,11 @@ kmem_zone_t *xfs_ili_zone; /* inode log
*/
STATIC uint
xfs_inode_item_size(
- xfs_inode_log_item_t *iip)
+ struct xfs_log_item *lip)
{
- uint nvecs;
- xfs_inode_t *ip;
-
- ip = iip->ili_inode;
- nvecs = 2;
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
+ uint nvecs = 2;
/*
* Only log the data/extents/b-tree root if there is something
@@ -202,20 +206,17 @@ xfs_inode_item_size(
*/
STATIC void
xfs_inode_item_format(
- xfs_inode_log_item_t *iip,
- xfs_log_iovec_t *log_vector)
+ struct xfs_log_item *lip,
+ struct xfs_log_iovec *vecp)
{
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
uint nvecs;
- xfs_log_iovec_t *vecp;
- xfs_inode_t *ip;
size_t data_bytes;
xfs_bmbt_rec_t *ext_buffer;
int nrecs;
xfs_mount_t *mp;
- ip = iip->ili_inode;
- vecp = log_vector;
-
vecp->i_addr = (xfs_caddr_t)&iip->ili_format;
vecp->i_len = sizeof(xfs_inode_log_format_t);
vecp->i_type = XLOG_REG_TYPE_IFORMAT;
@@ -427,7 +428,7 @@ xfs_inode_item_format(
* Assert that no attribute-related log flags are set.
*/
if (!XFS_IFORK_Q(ip)) {
- ASSERT(nvecs == iip->ili_item.li_desc->lid_size);
+ ASSERT(nvecs == lip->li_desc->lid_size);
iip->ili_format.ilf_size = nvecs;
ASSERT(!(iip->ili_format.ilf_fields &
(XFS_ILOG_ADATA | XFS_ILOG_ABROOT | XFS_ILOG_AEXT)));
@@ -518,7 +519,7 @@ xfs_inode_item_format(
break;
}
- ASSERT(nvecs == iip->ili_item.li_desc->lid_size);
+ ASSERT(nvecs == lip->li_desc->lid_size);
iip->ili_format.ilf_size = nvecs;
}
@@ -529,12 +530,14 @@ xfs_inode_item_format(
*/
STATIC void
xfs_inode_item_pin(
- xfs_inode_log_item_t *iip)
+ struct xfs_log_item *lip)
{
- ASSERT(xfs_isilocked(iip->ili_inode, XFS_ILOCK_EXCL));
+ struct xfs_inode *ip = INODE_ITEM(lip)->ili_inode;
- trace_xfs_inode_pin(iip->ili_inode, _RET_IP_);
- atomic_inc(&iip->ili_inode->i_pincount);
+ ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+
+ trace_xfs_inode_pin(ip, _RET_IP_);
+ atomic_inc(&ip->i_pincount);
}
@@ -546,10 +549,10 @@ xfs_inode_item_pin(
*/
STATIC void
xfs_inode_item_unpin(
- xfs_inode_log_item_t *iip,
+ struct xfs_log_item *lip,
int remove)
{
- struct xfs_inode *ip = iip->ili_inode;
+ struct xfs_inode *ip = INODE_ITEM(lip)->ili_inode;
trace_xfs_inode_unpin(ip, _RET_IP_);
ASSERT(atomic_read(&ip->i_pincount) > 0);
@@ -572,19 +575,16 @@ xfs_inode_item_unpin(
*/
STATIC uint
xfs_inode_item_trylock(
- xfs_inode_log_item_t *iip)
+ struct xfs_log_item *lip)
{
- register xfs_inode_t *ip;
-
- ip = iip->ili_inode;
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
- if (xfs_ipincount(ip) > 0) {
+ if (xfs_ipincount(ip) > 0)
return XFS_ITEM_PINNED;
- }
- if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) {
+ if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED))
return XFS_ITEM_LOCKED;
- }
if (!xfs_iflock_nowait(ip)) {
/*
@@ -610,7 +610,7 @@ xfs_inode_item_trylock(
if (!XFS_FORCED_SHUTDOWN(ip->i_mount)) {
ASSERT(iip->ili_format.ilf_fields != 0);
ASSERT(iip->ili_logged == 0);
- ASSERT(iip->ili_item.li_flags & XFS_LI_IN_AIL);
+ ASSERT(lip->li_flags & XFS_LI_IN_AIL);
}
#endif
return XFS_ITEM_SUCCESS;
@@ -624,12 +624,13 @@ xfs_inode_item_trylock(
*/
STATIC void
xfs_inode_item_unlock(
- xfs_inode_log_item_t *iip)
+ struct xfs_log_item *lip)
{
- uint hold;
- uint iolocked;
- uint lock_flags;
- xfs_inode_t *ip;
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
+ uint hold;
+ uint iolocked;
+ uint lock_flags;
ASSERT(iip != NULL);
ASSERT(iip->ili_inode->i_itemp != NULL);
@@ -640,10 +641,10 @@ xfs_inode_item_unlock(
ASSERT((!(iip->ili_inode->i_itemp->ili_flags &
XFS_ILI_IOLOCKED_SHARED)) ||
xfs_isilocked(iip->ili_inode, XFS_IOLOCK_SHARED));
+
/*
* Clear the transaction pointer in the inode.
*/
- ip = iip->ili_inode;
ip->i_transp = NULL;
/*
@@ -706,13 +707,12 @@ xfs_inode_item_unlock(
* is the only one that matters. Therefore, simply return the
* given lsn.
*/
-/*ARGSUSED*/
STATIC xfs_lsn_t
xfs_inode_item_committed(
- xfs_inode_log_item_t *iip,
+ struct xfs_log_item *lip,
xfs_lsn_t lsn)
{
- return (lsn);
+ return lsn;
}
/*
@@ -724,13 +724,12 @@ xfs_inode_item_committed(
*/
STATIC void
xfs_inode_item_pushbuf(
- xfs_inode_log_item_t *iip)
+ struct xfs_log_item *lip)
{
- xfs_inode_t *ip;
- xfs_mount_t *mp;
- xfs_buf_t *bp;
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
+ struct xfs_buf *bp;
- ip = iip->ili_inode;
ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED));
/*
@@ -738,14 +737,13 @@ xfs_inode_item_pushbuf(
* inode was taken off the AIL. So, just get out.
*/
if (completion_done(&ip->i_flush) ||
- ((iip->ili_item.li_flags & XFS_LI_IN_AIL) == 0)) {
+ !(lip->li_flags & XFS_LI_IN_AIL)) {
xfs_iunlock(ip, XFS_ILOCK_SHARED);
return;
}
- mp = ip->i_mount;
- bp = xfs_incore(mp->m_ddev_targp, iip->ili_format.ilf_blkno,
- iip->ili_format.ilf_len, XBF_TRYLOCK);
+ bp = xfs_incore(ip->i_mount->m_ddev_targp, iip->ili_format.ilf_blkno,
+ iip->ili_format.ilf_len, XBF_TRYLOCK);
xfs_iunlock(ip, XFS_ILOCK_SHARED);
if (!bp)
@@ -753,10 +751,8 @@ xfs_inode_item_pushbuf(
if (XFS_BUF_ISDELAYWRITE(bp))
xfs_buf_delwri_promote(bp);
xfs_buf_relse(bp);
- return;
}
-
/*
* This is called to asynchronously write the inode associated with this
* inode log item out to disk. The inode will already have been locked by
@@ -764,14 +760,14 @@ xfs_inode_item_pushbuf(
*/
STATIC void
xfs_inode_item_push(
- xfs_inode_log_item_t *iip)
+ struct xfs_log_item *lip)
{
- xfs_inode_t *ip;
-
- ip = iip->ili_inode;
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+ struct xfs_inode *ip = iip->ili_inode;
ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED));
ASSERT(!completion_done(&ip->i_flush));
+
/*
* Since we were able to lock the inode's flush lock and
* we found it on the AIL, the inode must be dirty. This
@@ -794,41 +790,34 @@ xfs_inode_item_push(
*/
(void) xfs_iflush(ip, 0);
xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
- return;
}
/*
* XXX rcc - this one really has to do something. Probably needs
* to stamp in a new field in the incore inode.
*/
-/* ARGSUSED */
STATIC void
xfs_inode_item_committing(
- xfs_inode_log_item_t *iip,
+ struct xfs_log_item *lip,
xfs_lsn_t lsn)
{
- iip->ili_last_lsn = lsn;
- return;
+ INODE_ITEM(lip)->ili_last_lsn = lsn;
}
/*
* This is the ops vector shared by all buf log items.
*/
static struct xfs_item_ops xfs_inode_item_ops = {
- .iop_size = (uint(*)(xfs_log_item_t*))xfs_inode_item_size,
- .iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
- xfs_inode_item_format,
- .iop_pin = (void(*)(xfs_log_item_t*))xfs_inode_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_inode_item_unpin,
- .iop_trylock = (uint(*)(xfs_log_item_t*))xfs_inode_item_trylock,
- .iop_unlock = (void(*)(xfs_log_item_t*))xfs_inode_item_unlock,
- .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_inode_item_committed,
- .iop_push = (void(*)(xfs_log_item_t*))xfs_inode_item_push,
- .iop_pushbuf = (void(*)(xfs_log_item_t*))xfs_inode_item_pushbuf,
- .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_inode_item_committing
+ .iop_size = xfs_inode_item_size,
+ .iop_format = xfs_inode_item_format,
+ .iop_pin = xfs_inode_item_pin,
+ .iop_unpin = xfs_inode_item_unpin,
+ .iop_trylock = xfs_inode_item_trylock,
+ .iop_unlock = xfs_inode_item_unlock,
+ .iop_committed = xfs_inode_item_committed,
+ .iop_push = xfs_inode_item_push,
+ .iop_pushbuf = xfs_inode_item_pushbuf,
+ .iop_committing = xfs_inode_item_committing
};
@@ -837,10 +826,10 @@ static struct xfs_item_ops xfs_inode_ite
*/
void
xfs_inode_item_init(
- xfs_inode_t *ip,
- xfs_mount_t *mp)
+ struct xfs_inode *ip,
+ struct xfs_mount *mp)
{
- xfs_inode_log_item_t *iip;
+ struct xfs_inode_log_item *iip;
ASSERT(ip->i_itemp == NULL);
iip = ip->i_itemp = kmem_zone_zalloc(xfs_ili_zone, KM_SLEEP);
Index: xfs/fs/xfs/xfs_buf_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.c 2010-05-29 18:12:37.760004276 +0200
+++ xfs/fs/xfs/xfs_buf_item.c 2010-05-29 18:14:21.317046949 +0200
@@ -33,6 +33,12 @@
kmem_zone_t *xfs_buf_item_zone;
+static inline struct xfs_buf_log_item *BUF_ITEM(struct xfs_log_item *lip)
+{
+ return container_of(lip, struct xfs_buf_log_item, bli_item);
+}
+
+
#ifdef XFS_TRANS_DEBUG
/*
* This function uses an alternate strategy for tracking the bytes
@@ -150,12 +156,13 @@ STATIC void xfs_buf_do_callbacks(xfs_buf
*/
STATIC uint
xfs_buf_item_size(
- xfs_buf_log_item_t *bip)
+ struct xfs_log_item *lip)
{
- uint nvecs;
- int next_bit;
- int last_bit;
- xfs_buf_t *bp;
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
+ struct xfs_buf *bp = bip->bli_buf;
+ uint nvecs;
+ int next_bit;
+ int last_bit;
ASSERT(atomic_read(&bip->bli_refcount) > 0);
if (bip->bli_flags & XFS_BLI_STALE) {
@@ -169,7 +176,6 @@ xfs_buf_item_size(
return 1;
}
- bp = bip->bli_buf;
ASSERT(bip->bli_flags & XFS_BLI_LOGGED);
nvecs = 1;
last_bit = xfs_next_bit(bip->bli_format.blf_data_map,
@@ -218,13 +224,13 @@ xfs_buf_item_size(
*/
STATIC void
xfs_buf_item_format(
- xfs_buf_log_item_t *bip,
- xfs_log_iovec_t *log_vector)
+ struct xfs_log_item *lip,
+ struct xfs_log_iovec *vecp)
{
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
+ struct xfs_buf *bp = bip->bli_buf;
uint base_size;
uint nvecs;
- xfs_log_iovec_t *vecp;
- xfs_buf_t *bp;
int first_bit;
int last_bit;
int next_bit;
@@ -234,8 +240,6 @@ xfs_buf_item_format(
ASSERT(atomic_read(&bip->bli_refcount) > 0);
ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
(bip->bli_flags & XFS_BLI_STALE));
- bp = bip->bli_buf;
- vecp = log_vector;
/*
* The size of the base structure is the size of the
@@ -262,7 +266,7 @@ xfs_buf_item_format(
*/
if (bip->bli_flags & XFS_BLI_INODE_BUF) {
if (!((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) &&
- xfs_log_item_in_current_chkpt(&bip->bli_item)))
+ xfs_log_item_in_current_chkpt(lip)))
bip->bli_format.blf_flags |= XFS_BLF_INODE_BUF;
bip->bli_flags &= ~XFS_BLI_INODE_BUF;
}
@@ -365,21 +369,20 @@ xfs_buf_item_format(
STATIC void
xfs_buf_item_pin(
- xfs_buf_log_item_t *bip)
+ struct xfs_log_item *lip)
{
- xfs_buf_t *bp;
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
- bp = bip->bli_buf;
- ASSERT(XFS_BUF_ISBUSY(bp));
+ ASSERT(XFS_BUF_ISBUSY(bip->bli_buf));
ASSERT(atomic_read(&bip->bli_refcount) > 0);
ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
(bip->bli_flags & XFS_BLI_STALE));
+
atomic_inc(&bip->bli_refcount);
trace_xfs_buf_item_pin(bip);
- xfs_bpin(bp);
+ xfs_bpin(bip->bli_buf);
}
-
/*
* This is called to unpin the buffer associated with the buf log
* item which was previously pinned with a call to xfs_buf_item_pin().
@@ -396,13 +399,14 @@ xfs_buf_item_pin(
*/
STATIC void
xfs_buf_item_unpin(
- xfs_buf_log_item_t *bip,
+ struct xfs_log_item *lip,
int remove)
{
- struct xfs_ail *ailp;
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
xfs_buf_t *bp = bip->bli_buf;
- int freed;
+ struct xfs_ail *ailp = lip->li_ailp;
int stale = bip->bli_flags & XFS_BLI_STALE;
+ int freed;
ASSERT(XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *) == bip);
ASSERT(atomic_read(&bip->bli_refcount) > 0);
@@ -410,8 +414,8 @@ xfs_buf_item_unpin(
trace_xfs_buf_item_unpin(bip);
freed = atomic_dec_and_test(&bip->bli_refcount);
- ailp = bip->bli_item.li_ailp;
xfs_bunpin(bp);
+
if (freed && stale) {
ASSERT(bip->bli_flags & XFS_BLI_STALE);
ASSERT(XFS_BUF_VALUSEMA(bp) <= 0);
@@ -429,7 +433,7 @@ xfs_buf_item_unpin(
* in xfs_trans_uncommit() will ry to reference the
* buffer which we no longer have a hold on.
*/
- xfs_trans_del_item(&bip->bli_item);
+ xfs_trans_del_item(lip);
/*
* Since the transaction no longer refers to the buffer,
@@ -468,11 +472,11 @@ xfs_buf_item_unpin(
*/
STATIC uint
xfs_buf_item_trylock(
- xfs_buf_log_item_t *bip)
+ struct xfs_log_item *lip)
{
- xfs_buf_t *bp;
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
+ struct xfs_buf *bp = bip->bli_buf;
- bp = bip->bli_buf;
if (XFS_BUF_ISPINNED(bp))
return XFS_ITEM_PINNED;
if (!XFS_BUF_CPSEMA(bp))
@@ -509,13 +513,12 @@ xfs_buf_item_trylock(
*/
STATIC void
xfs_buf_item_unlock(
- xfs_buf_log_item_t *bip)
+ struct xfs_log_item *lip)
{
- int aborted;
- xfs_buf_t *bp;
- uint hold;
-
- bp = bip->bli_buf;
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
+ struct xfs_buf *bp = bip->bli_buf;
+ int aborted;
+ uint hold;
/* Clear the buffer's association with this transaction. */
XFS_BUF_SET_FSPRIVATE2(bp, NULL);
@@ -526,7 +529,7 @@ xfs_buf_item_unlock(
* (cancelled) buffers at unpin time, but we'll never go through the
* pin/unpin cycle if we abort inside commit.
*/
- aborted = (bip->bli_item.li_flags & XFS_LI_ABORTED) != 0;
+ aborted = (lip->li_flags & XFS_LI_ABORTED) != 0;
/*
* Before possibly freeing the buf item, determine if we should
@@ -587,16 +590,16 @@ xfs_buf_item_unlock(
*/
STATIC xfs_lsn_t
xfs_buf_item_committed(
- xfs_buf_log_item_t *bip,
+ struct xfs_log_item *lip,
xfs_lsn_t lsn)
{
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
+
trace_xfs_buf_item_committed(bip);
- if ((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) &&
- (bip->bli_item.li_lsn != 0)) {
- return bip->bli_item.li_lsn;
- }
- return (lsn);
+ if ((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) && lip->li_lsn != 0)
+ return lip->li_lsn;
+ return lsn;
}
/*
@@ -606,15 +609,16 @@ xfs_buf_item_committed(
*/
STATIC void
xfs_buf_item_push(
- xfs_buf_log_item_t *bip)
+ struct xfs_log_item *lip)
{
- xfs_buf_t *bp;
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
+ struct xfs_buf *bp = bip->bli_buf;
ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
+ ASSERT(!XFS_BUF_ISDELAYWRITE(bp));
+
trace_xfs_buf_item_push(bip);
- bp = bip->bli_buf;
- ASSERT(!XFS_BUF_ISDELAYWRITE(bp));
xfs_buf_relse(bp);
}
@@ -626,22 +630,24 @@ xfs_buf_item_push(
*/
STATIC void
xfs_buf_item_pushbuf(
- xfs_buf_log_item_t *bip)
+ struct xfs_log_item *lip)
{
- xfs_buf_t *bp;
+ struct xfs_buf_log_item *bip = BUF_ITEM(lip);
+ struct xfs_buf *bp = bip->bli_buf;
ASSERT(!(bip->bli_flags & XFS_BLI_STALE));
+ ASSERT(XFS_BUF_ISDELAYWRITE(bp));
+
trace_xfs_buf_item_pushbuf(bip);
- bp = bip->bli_buf;
- ASSERT(XFS_BUF_ISDELAYWRITE(bp));
xfs_buf_delwri_promote(bp);
xfs_buf_relse(bp);
}
-/* ARGSUSED */
STATIC void
-xfs_buf_item_committing(xfs_buf_log_item_t *bip, xfs_lsn_t commit_lsn)
+xfs_buf_item_committing(
+ struct xfs_log_item *lip,
+ xfs_lsn_t commit_lsn)
{
}
@@ -649,19 +655,16 @@ xfs_buf_item_committing(xfs_buf_log_item
* This is the ops vector shared by all buf log items.
*/
static struct xfs_item_ops xfs_buf_item_ops = {
- .iop_size = (uint(*)(xfs_log_item_t*))xfs_buf_item_size,
- .iop_format = (void(*)(xfs_log_item_t*, xfs_log_iovec_t*))
- xfs_buf_item_format,
- .iop_pin = (void(*)(xfs_log_item_t*))xfs_buf_item_pin,
- .iop_unpin = (void(*)(xfs_log_item_t*, int))xfs_buf_item_unpin,
- .iop_trylock = (uint(*)(xfs_log_item_t*))xfs_buf_item_trylock,
- .iop_unlock = (void(*)(xfs_log_item_t*))xfs_buf_item_unlock,
- .iop_committed = (xfs_lsn_t(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_buf_item_committed,
- .iop_push = (void(*)(xfs_log_item_t*))xfs_buf_item_push,
- .iop_pushbuf = (void(*)(xfs_log_item_t*))xfs_buf_item_pushbuf,
- .iop_committing = (void(*)(xfs_log_item_t*, xfs_lsn_t))
- xfs_buf_item_committing
+ .iop_size = xfs_buf_item_size,
+ .iop_format = xfs_buf_item_format,
+ .iop_pin = xfs_buf_item_pin,
+ .iop_unpin = xfs_buf_item_unpin,
+ .iop_trylock = xfs_buf_item_trylock,
+ .iop_unlock = xfs_buf_item_unlock,
+ .iop_committed = xfs_buf_item_committed,
+ .iop_push = xfs_buf_item_push,
+ .iop_pushbuf = xfs_buf_item_pushbuf,
+ .iop_committing = xfs_buf_item_committing
};
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 09/17] xfs: give li_cb callbacks the correct prototype
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (7 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 08/17] xfs: give xfs_item_ops methods the correct prototypes Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 5:45 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 10/17] xfs: simplify buffer pinning Christoph Hellwig
` (9 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-fix-li_cb-types --]
[-- Type: text/plain, Size: 10753 bytes --]
Stop the function pointer casting madness and give all the li_cb instances
correct prototype.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/quota/xfs_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot.c 2010-05-29 18:12:09.768255567 +0200
+++ xfs/fs/xfs/quota/xfs_dquot.c 2010-05-29 18:16:52.004003788 +0200
@@ -54,8 +54,6 @@
flush lock - ditto.
*/
-STATIC void xfs_qm_dqflush_done(xfs_buf_t *, xfs_dq_logitem_t *);
-
#ifdef DEBUG
xfs_buftarg_t *xfs_dqerror_target;
int xfs_do_dqerror;
@@ -1131,6 +1129,46 @@ xfs_qm_dqrele(
xfs_qm_dqput(dqp);
}
+/*
+ * This is the dquot flushing I/O completion routine. It is called
+ * from interrupt level when the buffer containing the dquot is
+ * flushed to disk. It is responsible for removing the dquot logitem
+ * from the AIL if it has not been re-logged, and unlocking the dquot's
+ * flush lock. This behavior is very similar to that of inodes..
+ */
+STATIC void
+xfs_qm_dqflush_done(
+ struct xfs_buf *bp,
+ struct xfs_log_item *lip)
+{
+ xfs_dq_logitem_t *qip = (struct xfs_dq_logitem *)lip;
+ xfs_dquot_t *dqp = qip->qli_dquot;
+ struct xfs_ail *ailp = lip->li_ailp;
+
+ /*
+ * We only want to pull the item from the AIL if its
+ * location in the log has not changed since we started the flush.
+ * Thus, we only bother if the dquot's lsn has
+ * not changed. First we check the lsn outside the lock
+ * since it's cheaper, and then we recheck while
+ * holding the lock before removing the dquot from the AIL.
+ */
+ if ((lip->li_flags & XFS_LI_IN_AIL) &&
+ lip->li_lsn == qip->qli_flush_lsn) {
+
+ /* xfs_trans_ail_delete() drops the AIL lock. */
+ spin_lock(&ailp->xa_lock);
+ if (lip->li_lsn == qip->qli_flush_lsn)
+ xfs_trans_ail_delete(ailp, lip);
+ else
+ spin_unlock(&ailp->xa_lock);
+ }
+
+ /*
+ * Release the dq's flush lock since we're done with it.
+ */
+ xfs_dqfunlock(dqp);
+}
/*
* Write a modified dquot to disk.
@@ -1212,8 +1250,9 @@ xfs_qm_dqflush(
* Attach an iodone routine so that we can remove this dquot from the
* AIL and release the flush lock once the dquot is synced to disk.
*/
- xfs_buf_attach_iodone(bp, (void(*)(xfs_buf_t *, xfs_log_item_t *))
- xfs_qm_dqflush_done, &(dqp->q_logitem.qli_item));
+ xfs_buf_attach_iodone(bp, xfs_qm_dqflush_done,
+ &dqp->q_logitem.qli_item);
+
/*
* If the buffer is pinned then push on the log so we won't
* get stuck waiting in the write for too long.
@@ -1237,50 +1276,6 @@ xfs_qm_dqflush(
}
-/*
- * This is the dquot flushing I/O completion routine. It is called
- * from interrupt level when the buffer containing the dquot is
- * flushed to disk. It is responsible for removing the dquot logitem
- * from the AIL if it has not been re-logged, and unlocking the dquot's
- * flush lock. This behavior is very similar to that of inodes..
- */
-/*ARGSUSED*/
-STATIC void
-xfs_qm_dqflush_done(
- xfs_buf_t *bp,
- xfs_dq_logitem_t *qip)
-{
- xfs_dquot_t *dqp;
- struct xfs_ail *ailp;
-
- dqp = qip->qli_dquot;
- ailp = qip->qli_item.li_ailp;
-
- /*
- * We only want to pull the item from the AIL if its
- * location in the log has not changed since we started the flush.
- * Thus, we only bother if the dquot's lsn has
- * not changed. First we check the lsn outside the lock
- * since it's cheaper, and then we recheck while
- * holding the lock before removing the dquot from the AIL.
- */
- if ((qip->qli_item.li_flags & XFS_LI_IN_AIL) &&
- qip->qli_item.li_lsn == qip->qli_flush_lsn) {
-
- /* xfs_trans_ail_delete() drops the AIL lock. */
- spin_lock(&ailp->xa_lock);
- if (qip->qli_item.li_lsn == qip->qli_flush_lsn)
- xfs_trans_ail_delete(ailp, (xfs_log_item_t*)qip);
- else
- spin_unlock(&ailp->xa_lock);
- }
-
- /*
- * Release the dq's flush lock since we're done with it.
- */
- xfs_dqfunlock(dqp);
-}
-
int
xfs_qm_dqlock_nowait(
xfs_dquot_t *dqp)
Index: xfs/fs/xfs/xfs_buf_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.c 2010-05-29 18:14:21.317046949 +0200
+++ xfs/fs/xfs/xfs_buf_item.c 2010-05-29 18:17:16.277004555 +0200
@@ -1079,15 +1079,14 @@ xfs_buf_error_relse(
* It is called by xfs_buf_iodone_callbacks() above which will take
* care of cleaning up the buffer itself.
*/
-/* ARGSUSED */
void
xfs_buf_iodone(
- xfs_buf_t *bp,
- xfs_buf_log_item_t *bip)
+ struct xfs_buf *bp,
+ struct xfs_log_item *lip)
{
- struct xfs_ail *ailp = bip->bli_item.li_ailp;
+ struct xfs_ail *ailp = lip->li_ailp;
- ASSERT(bip->bli_buf == bp);
+ ASSERT(BUF_ITEM(lip)->bli_buf == bp);
xfs_buf_rele(bp);
@@ -1101,6 +1100,6 @@ xfs_buf_iodone(
* Either way, AIL is useless if we're forcing a shutdown.
*/
spin_lock(&ailp->xa_lock);
- xfs_trans_ail_delete(ailp, (xfs_log_item_t *)bip);
- xfs_buf_item_free(bip);
+ xfs_trans_ail_delete(ailp, lip);
+ xfs_buf_item_free(BUF_ITEM(lip));
}
Index: xfs/fs/xfs/xfs_buf_item.h
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.h 2010-05-25 13:01:10.104254015 +0200
+++ xfs/fs/xfs/xfs_buf_item.h 2010-05-29 18:17:16.279048625 +0200
@@ -124,7 +124,7 @@ void xfs_buf_attach_iodone(struct xfs_bu
void(*)(struct xfs_buf *, xfs_log_item_t *),
xfs_log_item_t *);
void xfs_buf_iodone_callbacks(struct xfs_buf *);
-void xfs_buf_iodone(struct xfs_buf *, xfs_buf_log_item_t *);
+void xfs_buf_iodone(struct xfs_buf *, struct xfs_log_item *);
#ifdef XFS_TRANS_DEBUG
void
Index: xfs/fs/xfs/xfs_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.c 2010-05-29 18:17:08.230003507 +0200
+++ xfs/fs/xfs/xfs_inode.c 2010-05-29 18:17:15.205004416 +0200
@@ -1984,7 +1984,7 @@ xfs_ifree_cluster(
if (lip->li_type == XFS_LI_INODE) {
iip = (xfs_inode_log_item_t *)lip;
ASSERT(iip->ili_logged == 1);
- lip->li_cb = (void(*)(xfs_buf_t*,xfs_log_item_t*)) xfs_istale_done;
+ lip->li_cb = xfs_istale_done;
xfs_trans_ail_copy_lsn(mp->m_ail,
&iip->ili_flush_lsn,
&iip->ili_item.li_lsn);
@@ -2051,9 +2051,8 @@ xfs_ifree_cluster(
xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
&iip->ili_item.li_lsn);
- xfs_buf_attach_iodone(bp,
- (void(*)(xfs_buf_t*,xfs_log_item_t*))
- xfs_istale_done, (xfs_log_item_t *)iip);
+ xfs_buf_attach_iodone(bp, xfs_istale_done,
+ &iip->ili_item);
if (ip != free_ip)
xfs_iunlock(ip, XFS_ILOCK_EXCL);
@@ -3065,8 +3064,7 @@ xfs_iflush_int(
* and unlock the inode's flush lock when the inode is
* completely written to disk.
*/
- xfs_buf_attach_iodone(bp, (void(*)(xfs_buf_t*,xfs_log_item_t*))
- xfs_iflush_done, (xfs_log_item_t *)iip);
+ xfs_buf_attach_iodone(bp, xfs_iflush_done, &iip->ili_item);
ASSERT(XFS_BUF_FSPRIVATE(bp, void *) != NULL);
ASSERT(XFS_BUF_IODONE_FUNC(bp) != NULL);
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c 2010-05-29 18:17:08.236004764 +0200
+++ xfs/fs/xfs/xfs_inode_item.c 2010-05-29 18:17:15.206003508 +0200
@@ -867,14 +867,14 @@ xfs_inode_item_destroy(
* from the AIL if it has not been re-logged, and unlocking the inode's
* flush lock.
*/
-/*ARGSUSED*/
void
xfs_iflush_done(
- xfs_buf_t *bp,
- xfs_inode_log_item_t *iip)
+ struct xfs_buf *bp,
+ struct xfs_log_item *lip)
{
+ struct xfs_inode_log_item *iip = INODE_ITEM(lip);
xfs_inode_t *ip = iip->ili_inode;
- struct xfs_ail *ailp = iip->ili_item.li_ailp;
+ struct xfs_ail *ailp = lip->li_ailp;
/*
* We only want to pull the item from the AIL if it is
@@ -885,12 +885,11 @@ xfs_iflush_done(
* the lock since it's cheaper, and then we recheck while
* holding the lock before removing the inode from the AIL.
*/
- if (iip->ili_logged &&
- (iip->ili_item.li_lsn == iip->ili_flush_lsn)) {
+ if (iip->ili_logged && lip->li_lsn == iip->ili_flush_lsn) {
spin_lock(&ailp->xa_lock);
- if (iip->ili_item.li_lsn == iip->ili_flush_lsn) {
+ if (lip->li_lsn == iip->ili_flush_lsn) {
/* xfs_trans_ail_delete() drops the AIL lock. */
- xfs_trans_ail_delete(ailp, (xfs_log_item_t*)iip);
+ xfs_trans_ail_delete(ailp, lip);
} else {
spin_unlock(&ailp->xa_lock);
}
@@ -908,8 +907,6 @@ xfs_iflush_done(
* Release the inode's flush lock since we're done with it.
*/
xfs_ifunlock(ip);
-
- return;
}
/*
@@ -959,10 +956,10 @@ xfs_iflush_abort(
void
xfs_istale_done(
- xfs_buf_t *bp,
- xfs_inode_log_item_t *iip)
+ struct xfs_buf *bp,
+ struct xfs_log_item *lip)
{
- xfs_iflush_abort(iip->ili_inode);
+ xfs_iflush_abort(INODE_ITEM(lip)->ili_inode);
}
/*
Index: xfs/fs/xfs/xfs_inode_item.h
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.h 2010-05-29 18:17:08.256254380 +0200
+++ xfs/fs/xfs/xfs_inode_item.h 2010-05-29 18:17:15.211004626 +0200
@@ -161,8 +161,8 @@ static inline int xfs_inode_clean(xfs_in
extern void xfs_inode_item_init(struct xfs_inode *, struct xfs_mount *);
extern void xfs_inode_item_destroy(struct xfs_inode *);
-extern void xfs_iflush_done(struct xfs_buf *, xfs_inode_log_item_t *);
-extern void xfs_istale_done(struct xfs_buf *, xfs_inode_log_item_t *);
+extern void xfs_iflush_done(struct xfs_buf *, struct xfs_log_item *);
+extern void xfs_istale_done(struct xfs_buf *, struct xfs_log_item *);
extern void xfs_iflush_abort(struct xfs_inode *);
extern int xfs_inode_item_format_convert(xfs_log_iovec_t *,
xfs_inode_log_format_t *);
Index: xfs/fs/xfs/xfs_trans_buf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_buf.c 2010-05-29 18:12:27.256254730 +0200
+++ xfs/fs/xfs/xfs_trans_buf.c 2010-05-29 18:17:16.285032981 +0200
@@ -658,7 +658,7 @@ xfs_trans_log_buf(xfs_trans_t *tp,
bip = XFS_BUF_FSPRIVATE(bp, xfs_buf_log_item_t *);
ASSERT(atomic_read(&bip->bli_refcount) > 0);
XFS_BUF_SET_IODONE_FUNC(bp, xfs_buf_iodone_callbacks);
- bip->bli_item.li_cb = (void(*)(xfs_buf_t*,xfs_log_item_t*))xfs_buf_iodone;
+ bip->bli_item.li_cb = xfs_buf_iodone;
trace_xfs_trans_log_buf(bip);
@@ -815,12 +815,9 @@ xfs_trans_stale_inode_buf(
ASSERT(atomic_read(&bip->bli_refcount) > 0);
bip->bli_flags |= XFS_BLI_STALE_INODE;
- bip->bli_item.li_cb = (void(*)(xfs_buf_t*,xfs_log_item_t*))
- xfs_buf_iodone;
+ bip->bli_item.li_cb = xfs_buf_iodone;
}
-
-
/*
* Mark the buffer as being one which contains newly allocated
* inodes. We need to make sure that even if this buffer is
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 10/17] xfs: simplify buffer pinning
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (8 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 09/17] xfs: give li_cb callbacks the correct prototype Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 5:47 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 11/17] xfs: simplify inode to transaction joining Christoph Hellwig
` (8 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-simplify-buffer-pinning --]
[-- Type: text/plain, Size: 5340 bytes --]
Get rid of the xfs_buf_pin/xfs_buf_unpin/xfs_buf_ispin helpers and opencode
them in their only callers, just like we did for the inode pinning a while
ago. Also remove duplicate trace points - the bufitem tracepoints cover
all the information that is present in a buffer tracepoint.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/linux-2.6/xfs_buf.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_buf.c 2010-05-29 18:12:07.996183820 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_buf.c 2010-05-29 18:19:18.653023552 +0200
@@ -896,36 +896,6 @@ xfs_buf_unlock(
trace_xfs_buf_unlock(bp, _RET_IP_);
}
-
-/*
- * Pinning Buffer Storage in Memory
- * Ensure that no attempt to force a buffer to disk will succeed.
- */
-void
-xfs_buf_pin(
- xfs_buf_t *bp)
-{
- trace_xfs_buf_pin(bp, _RET_IP_);
- atomic_inc(&bp->b_pin_count);
-}
-
-void
-xfs_buf_unpin(
- xfs_buf_t *bp)
-{
- trace_xfs_buf_unpin(bp, _RET_IP_);
-
- if (atomic_dec_and_test(&bp->b_pin_count))
- wake_up_all(&bp->b_waiters);
-}
-
-int
-xfs_buf_ispin(
- xfs_buf_t *bp)
-{
- return atomic_read(&bp->b_pin_count);
-}
-
STATIC void
xfs_buf_wait_unpin(
xfs_buf_t *bp)
@@ -1802,7 +1772,7 @@ xfs_buf_delwri_split(
trace_xfs_buf_delwri_split(bp, _RET_IP_);
ASSERT(bp->b_flags & XBF_DELWRI);
- if (!xfs_buf_ispin(bp) && !xfs_buf_cond_lock(bp)) {
+ if (!XFS_BUF_ISPINNED(bp) && !xfs_buf_cond_lock(bp)) {
if (!force &&
time_before(jiffies, bp->b_queuetime + age)) {
xfs_buf_unlock(bp);
Index: xfs/fs/xfs/linux-2.6/xfs_buf.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_buf.h 2010-05-22 00:40:37.889004315 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_buf.h 2010-05-29 18:19:18.654010422 +0200
@@ -258,11 +258,6 @@ static inline int xfs_buf_geterror(xfs_b
/* Buffer Utility Routines */
extern xfs_caddr_t xfs_buf_offset(xfs_buf_t *, size_t);
-/* Pinning Buffer Storage in Memory */
-extern void xfs_buf_pin(xfs_buf_t *);
-extern void xfs_buf_unpin(xfs_buf_t *);
-extern int xfs_buf_ispin(xfs_buf_t *);
-
/* Delayed Write Buffer Routines */
extern void xfs_buf_delwri_dequeue(xfs_buf_t *);
extern void xfs_buf_delwri_promote(xfs_buf_t *);
@@ -351,7 +346,7 @@ extern void xfs_buf_terminate(void);
#define XFS_BUF_SET_VTYPE(bp, type) do { } while (0)
#define XFS_BUF_SET_REF(bp, ref) do { } while (0)
-#define XFS_BUF_ISPINNED(bp) xfs_buf_ispin(bp)
+#define XFS_BUF_ISPINNED(bp) atomic_read(&((bp)->b_pin_count))
#define XFS_BUF_VALUSEMA(bp) xfs_buf_lock_value(bp)
#define XFS_BUF_CPSEMA(bp) (xfs_buf_cond_lock(bp) == 0)
@@ -370,8 +365,6 @@ static inline void xfs_buf_relse(xfs_buf
xfs_buf_rele(bp);
}
-#define xfs_bpin(bp) xfs_buf_pin(bp)
-#define xfs_bunpin(bp) xfs_buf_unpin(bp)
#define xfs_biodone(bp) xfs_buf_ioend(bp, 0)
#define xfs_biomove(bp, off, len, data, rw) \
Index: xfs/fs/xfs/linux-2.6/xfs_trace.h
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_trace.h 2010-05-29 12:30:33.890033540 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_trace.h 2010-05-29 18:19:18.660003717 +0200
@@ -314,8 +314,6 @@ DEFINE_BUF_EVENT(xfs_buf_init);
DEFINE_BUF_EVENT(xfs_buf_free);
DEFINE_BUF_EVENT(xfs_buf_hold);
DEFINE_BUF_EVENT(xfs_buf_rele);
-DEFINE_BUF_EVENT(xfs_buf_pin);
-DEFINE_BUF_EVENT(xfs_buf_unpin);
DEFINE_BUF_EVENT(xfs_buf_iodone);
DEFINE_BUF_EVENT(xfs_buf_iorequest);
DEFINE_BUF_EVENT(xfs_buf_bawrite);
Index: xfs/fs/xfs/xfs_buf_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.c 2010-05-29 18:17:16.277004555 +0200
+++ xfs/fs/xfs/xfs_buf_item.c 2010-05-29 18:19:18.662028022 +0200
@@ -359,14 +359,13 @@ xfs_buf_item_format(
/*
* This is called to pin the buffer associated with the buf log item in memory
- * so it cannot be written out. Simply call bpin() on the buffer to do this.
+ * so it cannot be written out.
*
* We also always take a reference to the buffer log item here so that the bli
* is held while the item is pinned in memory. This means that we can
* unconditionally drop the reference count a transaction holds when the
* transaction is completed.
*/
-
STATIC void
xfs_buf_item_pin(
struct xfs_log_item *lip)
@@ -378,15 +377,15 @@ xfs_buf_item_pin(
ASSERT((bip->bli_flags & XFS_BLI_LOGGED) ||
(bip->bli_flags & XFS_BLI_STALE));
- atomic_inc(&bip->bli_refcount);
trace_xfs_buf_item_pin(bip);
- xfs_bpin(bip->bli_buf);
+
+ atomic_inc(&bip->bli_refcount);
+ atomic_inc(&bip->bli_buf->b_pin_count);
}
/*
* This is called to unpin the buffer associated with the buf log
* item which was previously pinned with a call to xfs_buf_item_pin().
- * Just call bunpin() on the buffer to do this.
*
* Also drop the reference to the buf item for the current transaction.
* If the XFS_BLI_STALE flag is set and we are the last reference,
@@ -414,7 +413,9 @@ xfs_buf_item_unpin(
trace_xfs_buf_item_unpin(bip);
freed = atomic_dec_and_test(&bip->bli_refcount);
- xfs_bunpin(bp);
+
+ if (atomic_dec_and_test(&bp->b_pin_count))
+ wake_up_all(&bp->b_waiters);
if (freed && stale) {
ASSERT(bip->bli_flags & XFS_BLI_STALE);
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 10/17] xfs: simplify buffer pinning
2010-05-31 16:07 ` [PATCH 10/17] xfs: simplify buffer pinning Christoph Hellwig
@ 2010-06-02 5:47 ` Dave Chinner
0 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 5:47 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:37PM -0400, Christoph Hellwig wrote:
> Get rid of the xfs_buf_pin/xfs_buf_unpin/xfs_buf_ispin helpers and opencode
> them in their only callers, just like we did for the inode pinning a while
> ago. Also remove duplicate trace points - the bufitem tracepoints cover
> all the information that is present in a buffer tracepoint.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Looks good.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* [PATCH 11/17] xfs: simplify inode to transaction joining
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (9 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 10/17] xfs: simplify buffer pinning Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 5:57 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 12/17] xfs: fix the xfs_log_iovec i_addr type Christoph Hellwig
` (7 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-simplify-xfs_trans_ijoin --]
[-- Type: text/plain, Size: 28310 bytes --]
Currently we need to either call IHOLD or xfs_trans_ihold on an inode when
joining it to a transaction via xfs_trans_ijoin.
This patches instead makes xfs_trans_ijoin usable on it's own by doing
an implicity xfs_trans_ihold, which also allows us to drop the third
argument. For the case where we want to hold a reference on the inode
a xfs_trans_ijoin_ref wrapper is added which does the IHOLD and marks
the inode for needing an xfs_iput. In addition to the cleaner interface
to the caller this also simplifies the implementation.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/quota/xfs_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot.c 2010-05-29 18:44:47.554255638 +0200
+++ xfs/fs/xfs/quota/xfs_dquot.c 2010-05-29 18:49:10.749003438 +0200
@@ -378,14 +378,7 @@ xfs_qm_dqalloc(
return (ESRCH);
}
- /*
- * xfs_trans_commit normally decrements the vnode ref count
- * when it unlocks the inode. Since we want to keep the quota
- * inode around, we bump the vnode ref count now.
- */
- IHOLD(quotip);
-
- xfs_trans_ijoin(tp, quotip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, quotip, XFS_ILOCK_EXCL);
nmaps = 1;
if ((error = xfs_bmapi(tp, quotip,
offset_fsb, XFS_DQUOT_CLUSTER_SIZE_FSB,
Index: xfs/fs/xfs/xfs_bmap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.c 2010-05-29 18:44:44.719005883 +0200
+++ xfs/fs/xfs/xfs_bmap.c 2010-05-29 18:49:11.611005462 +0200
@@ -3739,9 +3739,10 @@ xfs_bmap_add_attrfork(
ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
}
ASSERT(ip->i_d.di_anextents == 0);
- IHOLD(ip);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+
+ xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+
switch (ip->i_d.di_format) {
case XFS_DINODE_FMT_DEV:
ip->i_d.di_forkoff = roundup(sizeof(xfs_dev_t), 8) >> 3;
Index: xfs/fs/xfs/xfs_rename.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rename.c 2010-05-29 18:44:44.198005742 +0200
+++ xfs/fs/xfs/xfs_rename.c 2010-05-29 18:44:47.678039336 +0200
@@ -169,26 +169,14 @@ xfs_rename(
/*
* Join all the inodes to the transaction. From this point on,
* we can rely on either trans_commit or trans_cancel to unlock
- * them. Note that we need to add a vnode reference to the
- * directories since trans_commit & trans_cancel will decrement
- * them when they unlock the inodes. Also, we need to be careful
- * not to add an inode to the transaction more than once.
+ * them.
*/
- IHOLD(src_dp);
- xfs_trans_ijoin(tp, src_dp, XFS_ILOCK_EXCL);
-
- if (new_parent) {
- IHOLD(target_dp);
- xfs_trans_ijoin(tp, target_dp, XFS_ILOCK_EXCL);
- }
-
- IHOLD(src_ip);
- xfs_trans_ijoin(tp, src_ip, XFS_ILOCK_EXCL);
-
- if (target_ip) {
- IHOLD(target_ip);
- xfs_trans_ijoin(tp, target_ip, XFS_ILOCK_EXCL);
- }
+ xfs_trans_ijoin_ref(tp, src_dp, XFS_ILOCK_EXCL);
+ if (new_parent)
+ xfs_trans_ijoin_ref(tp, target_dp, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, src_ip, XFS_ILOCK_EXCL);
+ if (target_ip)
+ xfs_trans_ijoin_ref(tp, target_ip, XFS_ILOCK_EXCL);
/*
* If we are using project inheritance, we only allow renames
Index: xfs/fs/xfs/xfs_trans.h
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.h 2010-05-29 18:44:44.783005813 +0200
+++ xfs/fs/xfs/xfs_trans.h 2010-05-29 18:49:12.008004765 +0200
@@ -475,8 +475,8 @@ void xfs_trans_dquot_buf(xfs_trans_t *,
void xfs_trans_inode_alloc_buf(xfs_trans_t *, struct xfs_buf *);
int xfs_trans_iget(struct xfs_mount *, xfs_trans_t *,
xfs_ino_t , uint, uint, struct xfs_inode **);
-void xfs_trans_ijoin(xfs_trans_t *, struct xfs_inode *, uint);
-void xfs_trans_ihold(xfs_trans_t *, struct xfs_inode *);
+void xfs_trans_ijoin_ref(struct xfs_trans *, struct xfs_inode *, uint);
+void xfs_trans_ijoin(struct xfs_trans *, struct xfs_inode *);
void xfs_trans_log_buf(xfs_trans_t *, struct xfs_buf *, uint, uint);
void xfs_trans_log_inode(xfs_trans_t *, struct xfs_inode *, uint);
struct xfs_efi_log_item *xfs_trans_get_efi(xfs_trans_t *, uint);
Index: xfs/fs/xfs/xfs_trans_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans_inode.c 2010-05-29 18:44:44.695207098 +0200
+++ xfs/fs/xfs/xfs_trans_inode.c 2010-05-29 18:44:47.682003507 +0200
@@ -33,6 +33,7 @@
#include "xfs_btree.h"
#include "xfs_trans_priv.h"
#include "xfs_inode_item.h"
+#include "xfs_trace.h"
#ifdef XFS_TRANS_DEBUG
STATIC void
@@ -42,7 +43,6 @@ xfs_trans_inode_broot_debug(
#define xfs_trans_inode_broot_debug(ip)
#endif
-
/*
* Get an inode and join it to the transaction.
*/
@@ -58,32 +58,31 @@ xfs_trans_iget(
int error;
error = xfs_iget(mp, tp, ino, flags, lock_flags, ipp, 0);
- if (!error && tp)
- xfs_trans_ijoin(tp, *ipp, lock_flags);
+ if (!error && tp) {
+ xfs_trans_ijoin(tp, *ipp);
+ (*ipp)->i_itemp->ili_lock_flags = lock_flags;
+ }
return error;
}
/*
- * Add the locked inode to the transaction.
- * The inode must be locked, and it cannot be associated with any
- * transaction. The caller must specify the locks already held
- * on the inode.
+ * Add a locked inode to the transaction.
+ *
+ * The inode must be locked, and it cannot be associated with any transaction.
*/
void
xfs_trans_ijoin(
- xfs_trans_t *tp,
- xfs_inode_t *ip,
- uint lock_flags)
+ struct xfs_trans *tp,
+ struct xfs_inode *ip)
{
xfs_inode_log_item_t *iip;
ASSERT(ip->i_transp == NULL);
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
- ASSERT(lock_flags & XFS_ILOCK_EXCL);
if (ip->i_itemp == NULL)
xfs_inode_item_init(ip, ip->i_mount);
iip = ip->i_itemp;
- ASSERT(iip->ili_flags == 0);
+ ASSERT(iip->ili_lock_flags == 0);
/*
* Get a log_item_desc to point at the new item.
@@ -93,42 +92,31 @@ xfs_trans_ijoin(
xfs_trans_inode_broot_debug(ip);
/*
- * If the IO lock is already held, mark that in the inode log item.
- */
- if (lock_flags & XFS_IOLOCK_EXCL) {
- iip->ili_flags |= XFS_ILI_IOLOCKED_EXCL;
- } else if (lock_flags & XFS_IOLOCK_SHARED) {
- iip->ili_flags |= XFS_ILI_IOLOCKED_SHARED;
- }
-
- /*
* Initialize i_transp so we can find it with xfs_inode_incore()
* in xfs_trans_iget() above.
*/
ip->i_transp = tp;
}
-
-
/*
- * Mark the inode as not needing to be unlocked when the inode item's
- * IOP_UNLOCK() routine is called. The inode must already be locked
- * and associated with the given transaction.
+ * Add a locked inode to the transaction.
+ *
+ *
+ * Grabs a reference to the inode which will be dropped when the transaction
+ * is commited. The inode will also be unlocked at that point. The inode
+ * must be locked, and it cannot be associated with any transaction.
*/
-/*ARGSUSED*/
void
-xfs_trans_ihold(
- xfs_trans_t *tp,
- xfs_inode_t *ip)
+xfs_trans_ijoin_ref(
+ struct xfs_trans *tp,
+ struct xfs_inode *ip,
+ uint lock_flags)
{
- ASSERT(ip->i_transp == tp);
- ASSERT(ip->i_itemp != NULL);
- ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-
- ip->i_itemp->ili_flags |= XFS_ILI_HOLD;
+ xfs_trans_ijoin(tp, ip);
+ IHOLD(ip);
+ ip->i_itemp->ili_lock_flags = lock_flags;
}
-
/*
* This is called to mark the fields indicated in fieldmask as needing
* to be logged when the transaction is committed. The inode must
Index: xfs/fs/xfs/xfs_vnodeops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_vnodeops.c 2010-05-29 18:44:44.233005323 +0200
+++ xfs/fs/xfs/xfs_vnodeops.c 2010-05-29 18:49:10.853004415 +0200
@@ -268,8 +268,7 @@ xfs_setattr(
commit_flags = XFS_TRANS_RELEASE_LOG_RES;
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, lock_flags);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
/*
* Only change the c/mtime if we are changing the size
@@ -319,8 +318,7 @@ xfs_setattr(
xfs_iflags_set(ip, XFS_ITRUNCATED);
}
} else if (tp) {
- xfs_trans_ijoin(tp, ip, lock_flags);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
}
/*
@@ -653,10 +651,7 @@ xfs_free_eofblocks(
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip,
- XFS_IOLOCK_EXCL |
- XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
error = xfs_itruncate_finish(&tp, ip,
ip->i_size,
@@ -728,8 +723,7 @@ xfs_inactive_symlink_rmt(
xfs_ilock(ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
size = (int)ip->i_d.di_size;
ip->i_d.di_size = 0;
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
/*
* Find the block(s) so we can inval and unmap them.
@@ -773,8 +767,7 @@ xfs_inactive_symlink_rmt(
* Mark it dirty so it will be logged and moved forward in the log as
* part of every commit.
*/
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
/*
* Get a new, empty transaction to return to our caller.
@@ -907,8 +900,7 @@ xfs_inactive_attrs(
goto error_cancel;
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_idestroy_fork(ip, XFS_ATTR_FORK);
ASSERT(ip->i_d.di_anextents == 0);
@@ -1095,8 +1087,7 @@ xfs_inactive(
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
/*
* normally, we have to run xfs_itruncate_finish sync.
@@ -1129,8 +1120,7 @@ xfs_inactive(
return VN_INACTIVE_CACHE;
}
- xfs_trans_ijoin(tp, ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
} else {
error = xfs_trans_reserve(tp, 0,
XFS_IFREE_LOG_RES(mp),
@@ -1143,8 +1133,7 @@ xfs_inactive(
}
xfs_ilock(ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_IOLOCK_EXCL | XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
}
/*
@@ -1392,8 +1381,7 @@ xfs_create(
* the transaction cancel unlocking dp so don't do it explicitly in the
* error path.
*/
- IHOLD(dp);
- xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, dp, XFS_ILOCK_EXCL);
unlock_dp_on_error = B_FALSE;
error = xfs_dir_createname(tp, dp, name, ip->i_ino,
@@ -1730,15 +1718,8 @@ xfs_remove(
xfs_lock_two_inodes(dp, ip, XFS_ILOCK_EXCL);
- /*
- * At this point, we've gotten both the directory and the entry
- * inodes locked.
- */
- IHOLD(ip);
- xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
-
- IHOLD(dp);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, dp, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL);
/*
* If we're removing a directory perform some additional validation.
@@ -1884,15 +1865,8 @@ xfs_link(
xfs_lock_two_inodes(sip, tdp, XFS_ILOCK_EXCL);
- /*
- * Increment vnode ref counts since xfs_trans_commit &
- * xfs_trans_cancel will both unlock the inodes and
- * decrement the associated ref counts.
- */
- IHOLD(sip);
- IHOLD(tdp);
- xfs_trans_ijoin(tp, sip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, tdp, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, sip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, tdp, XFS_ILOCK_EXCL);
/*
* If the source has too many links, we can't make any more to it.
@@ -2087,8 +2061,7 @@ xfs_symlink(
* transaction cancel unlocking dp so don't do it explicitly in the
* error path.
*/
- IHOLD(dp);
- xfs_trans_ijoin(tp, dp, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, dp, XFS_ILOCK_EXCL);
unlock_dp_on_error = B_FALSE;
/*
@@ -2227,13 +2200,12 @@ xfs_set_dmattrs(
return error;
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL);
ip->i_d.di_dmevmask = evmask;
ip->i_d.di_dmstate = state;
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
- IHOLD(ip);
error = xfs_trans_commit(tp, 0);
return error;
@@ -2366,8 +2338,7 @@ xfs_alloc_file_space(
if (error)
goto error1;
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
/*
* Issue the xfs_bmapi() call to allocate the blocks
@@ -2668,8 +2639,7 @@ xfs_free_file_space(
if (error)
goto error1;
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
/*
* issue the bunmapi() call to free the blocks
@@ -2839,8 +2809,7 @@ xfs_change_file_space(
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
if ((attr_flags & XFS_ATTR_DMI) == 0) {
ip->i_d.di_mode &= ~S_ISUID;
Index: xfs/fs/xfs/linux-2.6/xfs_ioctl.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_ioctl.c 2010-05-29 18:44:44.024034028 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_ioctl.c 2010-05-29 18:44:47.689026206 +0200
@@ -1035,8 +1035,7 @@ xfs_ioctl_setattr(
}
}
- xfs_trans_ijoin(tp, ip, lock_flags);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
/*
* Change file ownership. Must be the owner or privileged.
Index: xfs/fs/xfs/linux-2.6/xfs_sync.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_sync.c 2010-05-29 18:44:44.048005742 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_sync.c 2010-05-29 18:44:47.695005533 +0200
@@ -328,8 +328,7 @@ xfs_commit_dummy_trans(
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
error = xfs_trans_commit(tp, 0);
xfs_iunlock(ip, XFS_ILOCK_EXCL);
Index: xfs/fs/xfs/xfs_attr.c
===================================================================
--- xfs.orig/fs/xfs/xfs_attr.c 2010-05-29 18:44:44.103255638 +0200
+++ xfs/fs/xfs/xfs_attr.c 2010-05-29 18:49:10.764003787 +0200
@@ -319,8 +319,7 @@ xfs_attr_set_int(
return (error);
}
- xfs_trans_ijoin(args.trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args.trans, dp);
+ xfs_trans_ijoin(args.trans, dp);
/*
* If the attribute list is non-existent or a shortform list,
@@ -390,10 +389,8 @@ xfs_attr_set_int(
* bmap_finish() may have committed the last trans and started
* a new one. We need the inode to be in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args.trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args.trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args.trans, dp);
/*
* Commit the leaf transformation. We'll need another (linked)
@@ -538,8 +535,7 @@ xfs_attr_remove_int(xfs_inode_t *dp, str
* No need to make quota reservations here. We expect to release some
* blocks not allocate in the common case.
*/
- xfs_trans_ijoin(args.trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args.trans, dp);
+ xfs_trans_ijoin(args.trans, dp);
/*
* Decide on what work routines to call based on the inode size.
@@ -815,8 +811,7 @@ xfs_attr_inactive(xfs_inode_t *dp)
* No need to make quota reservations here. We expect to release some
* blocks, not allocate, in the common case.
*/
- xfs_trans_ijoin(trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(trans, dp);
+ xfs_trans_ijoin(trans, dp);
/*
* Decide on what work routines to call based on the inode size.
@@ -975,10 +970,8 @@ xfs_attr_leaf_addname(xfs_da_args_t *arg
* bmap_finish() may have committed the last trans and started
* a new one. We need the inode to be in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
/*
* Commit the current trans (including the inode) and start
@@ -1079,10 +1072,8 @@ xfs_attr_leaf_addname(xfs_da_args_t *arg
* and started a new one. We need the inode to be
* in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
} else
xfs_da_buf_done(bp);
@@ -1155,10 +1146,8 @@ xfs_attr_leaf_removename(xfs_da_args_t *
* bmap_finish() may have committed the last trans and started
* a new one. We need the inode to be in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
} else
xfs_da_buf_done(bp);
return(0);
@@ -1311,10 +1300,8 @@ restart:
* and started a new one. We need the inode to be
* in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
/*
* Commit the node conversion and start the next
@@ -1350,10 +1337,8 @@ restart:
* bmap_finish() may have committed the last trans and started
* a new one. We need the inode to be in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
} else {
/*
* Addition succeeded, update Btree hashvals.
@@ -1464,10 +1449,8 @@ restart:
* and started a new one. We need the inode to be
* in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
}
/*
@@ -1598,10 +1581,8 @@ xfs_attr_node_removename(xfs_da_args_t *
* bmap_finish() may have committed the last trans and started
* a new one. We need the inode to be in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
/*
* Commit the Btree join operation and start a new trans.
@@ -1652,10 +1633,8 @@ xfs_attr_node_removename(xfs_da_args_t *
* and started a new one. We need the inode to be
* in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
} else
xfs_da_brelse(args->trans, bp);
}
@@ -2093,10 +2072,8 @@ xfs_attr_rmtval_set(xfs_da_args_t *args)
* bmap_finish() may have committed the last trans and started
* a new one. We need the inode to be in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, dp);
ASSERT(nmap == 1);
ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
@@ -2249,10 +2226,8 @@ xfs_attr_rmtval_remove(xfs_da_args_t *ar
* bmap_finish() may have committed the last trans and started
* a new one. We need the inode to be in all transactions.
*/
- if (committed) {
- xfs_trans_ijoin(args->trans, args->dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(args->trans, args->dp);
- }
+ if (committed)
+ xfs_trans_ijoin(args->trans, args->dp);
/*
* Close out trans and start the next one in the chain.
Index: xfs/fs/xfs/xfs_dfrag.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dfrag.c 2010-05-29 18:44:44.127260247 +0200
+++ xfs/fs/xfs/xfs_dfrag.c 2010-05-29 18:44:47.708005952 +0200
@@ -413,11 +413,8 @@ xfs_swap_extents(
}
- IHOLD(ip);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
-
- IHOLD(tip);
- xfs_trans_ijoin(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
+ xfs_trans_ijoin_ref(tp, tip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
xfs_trans_log_inode(tp, ip, ilf_fields);
xfs_trans_log_inode(tp, tip, tilf_fields);
Index: xfs/fs/xfs/xfs_fsops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_fsops.c 2010-05-29 18:44:44.139031165 +0200
+++ xfs/fs/xfs/xfs_fsops.c 2010-05-29 18:44:47.713005812 +0200
@@ -633,8 +633,7 @@ xfs_fs_log_dummy(
ip = mp->m_rootip;
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0);
Index: xfs/fs/xfs/xfs_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.c 2010-05-29 18:44:47.572011120 +0200
+++ xfs/fs/xfs/xfs_inode.c 2010-05-29 18:49:11.627003926 +0200
@@ -1459,7 +1459,7 @@ xfs_itruncate_finish(
ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES);
ASSERT(ip->i_transp == *tp);
ASSERT(ip->i_itemp != NULL);
- ASSERT(ip->i_itemp->ili_flags & XFS_ILI_HOLD);
+ ASSERT(ip->i_itemp->ili_lock_flags == 0);
ntp = *tp;
@@ -1611,12 +1611,8 @@ xfs_itruncate_finish(
*/
error = xfs_bmap_finish(tp, &free_list, &committed);
ntp = *tp;
- if (committed) {
- /* link the inode into the next xact in the chain */
- xfs_trans_ijoin(ntp, ip,
- XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
- xfs_trans_ihold(ntp, ip);
- }
+ if (committed)
+ xfs_trans_ijoin(ntp, ip);
if (error) {
/*
@@ -1645,9 +1641,7 @@ xfs_itruncate_finish(
error = xfs_trans_commit(*tp, 0);
*tp = ntp;
- /* link the inode into the next transaction in the chain */
- xfs_trans_ijoin(ntp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
- xfs_trans_ihold(ntp, ip);
+ xfs_trans_ijoin(ntp, ip);
if (error)
return error;
Index: xfs/fs/xfs/xfs_iomap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.c 2010-05-29 18:44:44.173005812 +0200
+++ xfs/fs/xfs/xfs_iomap.c 2010-05-29 18:49:10.832274285 +0200
@@ -329,8 +329,7 @@ xfs_iomap_write_direct(
if (error)
goto error1;
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
bmapi_flag = XFS_BMAPI_WRITE;
if ((flags & BMAPI_DIRECT) && (offset < ip->i_size || extsz))
@@ -597,8 +596,7 @@ xfs_iomap_write_allocate(
return XFS_ERROR(error);
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_bmap_init(&free_list, &first_block);
@@ -761,8 +759,7 @@ xfs_iomap_write_unwritten(
}
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
/*
* Modify the unwritten extent state of the buffer.
Index: xfs/fs/xfs/xfs_trans.c
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.c 2010-05-29 18:44:44.776255080 +0200
+++ xfs/fs/xfs/xfs_trans.c 2010-05-29 18:44:47.726070626 +0200
@@ -1892,7 +1892,6 @@ xfs_trans_roll(
if (error)
return error;
- xfs_trans_ijoin(trans, dp, XFS_ILOCK_EXCL);
- xfs_trans_ihold(trans, dp);
+ xfs_trans_ijoin(trans, dp);
return 0;
}
Index: xfs/fs/xfs/xfs_utils.c
===================================================================
--- xfs.orig/fs/xfs/xfs_utils.c 2010-05-29 18:44:44.226005812 +0200
+++ xfs/fs/xfs/xfs_utils.c 2010-05-29 18:44:47.732005603 +0200
@@ -374,8 +374,8 @@ xfs_truncate_file(
* of references will stay constant.
*/
xfs_ilock(ip, XFS_ILOCK_EXCL);
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | XFS_IOLOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
+
/*
* Signal a sync xaction. The only case where that isn't
* the case is if we're truncating an already unlinked file
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c 2010-05-29 18:44:47.580032003 +0200
+++ xfs/fs/xfs/xfs_inode_item.c 2010-05-29 18:49:14.203004136 +0200
@@ -628,19 +628,10 @@ xfs_inode_item_unlock(
{
struct xfs_inode_log_item *iip = INODE_ITEM(lip);
struct xfs_inode *ip = iip->ili_inode;
- uint hold;
- uint iolocked;
- uint lock_flags;
+ unsigned short lock_flags;
- ASSERT(iip != NULL);
ASSERT(iip->ili_inode->i_itemp != NULL);
ASSERT(xfs_isilocked(iip->ili_inode, XFS_ILOCK_EXCL));
- ASSERT((!(iip->ili_inode->i_itemp->ili_flags &
- XFS_ILI_IOLOCKED_EXCL)) ||
- xfs_isilocked(iip->ili_inode, XFS_IOLOCK_EXCL));
- ASSERT((!(iip->ili_inode->i_itemp->ili_flags &
- XFS_ILI_IOLOCKED_SHARED)) ||
- xfs_isilocked(iip->ili_inode, XFS_IOLOCK_SHARED));
/*
* Clear the transaction pointer in the inode.
@@ -668,35 +659,10 @@ xfs_inode_item_unlock(
iip->ili_aextents_buf = NULL;
}
- /*
- * Figure out if we should unlock the inode or not.
- */
- hold = iip->ili_flags & XFS_ILI_HOLD;
-
- /*
- * Before clearing out the flags, remember whether we
- * are holding the inode's IO lock.
- */
- iolocked = iip->ili_flags & XFS_ILI_IOLOCKED_ANY;
-
- /*
- * Clear out the fields of the inode log item particular
- * to the current transaction.
- */
- iip->ili_flags = 0;
-
- /*
- * Unlock the inode if XFS_ILI_HOLD was not set.
- */
- if (!hold) {
- lock_flags = XFS_ILOCK_EXCL;
- if (iolocked & XFS_ILI_IOLOCKED_EXCL) {
- lock_flags |= XFS_IOLOCK_EXCL;
- } else if (iolocked & XFS_ILI_IOLOCKED_SHARED) {
- lock_flags |= XFS_IOLOCK_SHARED;
- }
+ lock_flags = iip->ili_lock_flags;
+ iip->ili_lock_flags = 0;
+ if (lock_flags)
xfs_iput(iip->ili_inode, lock_flags);
- }
}
/*
Index: xfs/fs/xfs/xfs_inode_item.h
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.h 2010-05-29 18:44:47.582032213 +0200
+++ xfs/fs/xfs/xfs_inode_item.h 2010-05-29 18:49:35.192005952 +0200
@@ -103,12 +103,6 @@ typedef struct xfs_inode_log_format_64 {
XFS_ILOG_ADATA | XFS_ILOG_AEXT | \
XFS_ILOG_ABROOT)
-#define XFS_ILI_HOLD 0x1
-#define XFS_ILI_IOLOCKED_EXCL 0x2
-#define XFS_ILI_IOLOCKED_SHARED 0x4
-
-#define XFS_ILI_IOLOCKED_ANY (XFS_ILI_IOLOCKED_EXCL | XFS_ILI_IOLOCKED_SHARED)
-
static inline int xfs_ilog_fbroot(int w)
{
return (w == XFS_DATA_FORK ? XFS_ILOG_DBROOT : XFS_ILOG_ABROOT);
@@ -137,7 +131,7 @@ typedef struct xfs_inode_log_item {
struct xfs_inode *ili_inode; /* inode ptr */
xfs_lsn_t ili_flush_lsn; /* lsn at last flush */
xfs_lsn_t ili_last_lsn; /* lsn at last transaction */
- unsigned short ili_flags; /* misc flags */
+ unsigned short ili_lock_flags; /* lock flags */
unsigned short ili_logged; /* flushed logged data */
unsigned int ili_last_fields; /* fields when flushed */
struct xfs_bmbt_rec *ili_extents_buf; /* array of logged
Index: xfs/fs/xfs/linux-2.6/xfs_file.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_file.c 2010-05-29 18:44:44.020005812 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_file.c 2010-05-29 18:49:10.742004625 +0200
@@ -158,8 +158,7 @@ xfs_file_fsync(
* transaction. So we play it safe and fire off the
* transaction anyway.
*/
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = _xfs_trans_commit(tp, 0, &log_flushed);
Index: xfs/fs/xfs/linux-2.6/xfs_super.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_super.c 2010-05-29 18:44:44.709255988 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_super.c 2010-05-29 18:44:47.751005673 +0200
@@ -1023,8 +1023,7 @@ xfs_log_inode(
* an inode in another recent transaction. So we play it safe and
* fire off the transaction anyway.
*/
- xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
- xfs_trans_ihold(tp, ip);
+ xfs_trans_ijoin(tp, ip);
xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
xfs_trans_set_sync(tp);
error = xfs_trans_commit(tp, 0);
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 11/17] xfs: simplify inode to transaction joining
2010-05-31 16:07 ` [PATCH 11/17] xfs: simplify inode to transaction joining Christoph Hellwig
@ 2010-06-02 5:57 ` Dave Chinner
0 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 5:57 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:38PM -0400, Christoph Hellwig wrote:
> Currently we need to either call IHOLD or xfs_trans_ihold on an inode when
> joining it to a transaction via xfs_trans_ijoin.
>
> This patches instead makes xfs_trans_ijoin usable on it's own by doing
> an implicity xfs_trans_ihold, which also allows us to drop the third
> argument. For the case where we want to hold a reference on the inode
> a xfs_trans_ijoin_ref wrapper is added which does the IHOLD and marks
> the inode for needing an xfs_iput. In addition to the cleaner interface
> to the caller this also simplifies the implementation.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Can't say I'm fond of the xfs_trans_iget() setting of the lock
flags, but that's a small price to pay forthe rest of the cleanup.
So it looks good to me.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* [PATCH 12/17] xfs: fix the xfs_log_iovec i_addr type
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (10 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 11/17] xfs: simplify inode to transaction joining Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 6:01 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 13/17] xfs: kill the unused xlog_debug variable Christoph Hellwig
` (6 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-fix-i_addr-type --]
[-- Type: text/plain, Size: 14537 bytes --]
By making this member a void pointer we can get rid of a lot of pointless
casts.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/quota/xfs_dquot_item.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot_item.c 2010-05-29 18:44:47.514010352 +0200
+++ xfs/fs/xfs/quota/xfs_dquot_item.c 2010-05-29 18:44:55.990253681 +0200
@@ -65,11 +65,11 @@ xfs_qm_dquot_logitem_format(
{
struct xfs_dq_logitem *qlip = DQUOT_ITEM(lip);
- logvec->i_addr = (xfs_caddr_t)&qlip->qli_format;
+ logvec->i_addr = &qlip->qli_format;
logvec->i_len = sizeof(xfs_dq_logformat_t);
logvec->i_type = XLOG_REG_TYPE_QFORMAT;
logvec++;
- logvec->i_addr = (xfs_caddr_t)&qlip->qli_dquot->q_core;
+ logvec->i_addr = &qlip->qli_dquot->q_core;
logvec->i_len = sizeof(xfs_disk_dquot_t);
logvec->i_type = XLOG_REG_TYPE_DQUOT;
@@ -369,7 +369,7 @@ xfs_qm_qoff_logitem_format(
ASSERT(qflip->qql_format.qf_type == XFS_LI_QUOTAOFF);
- log_vector->i_addr = (xfs_caddr_t)&(qflip->qql_format);
+ log_vector->i_addr = &qflip->qql_format;
log_vector->i_len = sizeof(xfs_qoff_logitem_t);
log_vector->i_type = XLOG_REG_TYPE_QUOTAOFF;
qflip->qql_format.qf_size = 1;
Index: xfs/fs/xfs/xfs_buf_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_buf_item.c 2010-05-29 18:44:47.641007209 +0200
+++ xfs/fs/xfs/xfs_buf_item.c 2010-05-29 18:44:55.991253611 +0200
@@ -251,7 +251,7 @@ xfs_buf_item_format(
base_size =
(uint)(sizeof(xfs_buf_log_format_t) +
((bip->bli_format.blf_map_size - 1) * sizeof(uint)));
- vecp->i_addr = (xfs_caddr_t)&bip->bli_format;
+ vecp->i_addr = &bip->bli_format;
vecp->i_len = base_size;
vecp->i_type = XLOG_REG_TYPE_BFORMAT;
vecp++;
Index: xfs/fs/xfs/xfs_extfree_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_extfree_item.c 2010-05-29 18:44:47.519254590 +0200
+++ xfs/fs/xfs/xfs_extfree_item.c 2010-05-29 18:44:55.996299498 +0200
@@ -82,7 +82,7 @@ xfs_efi_item_format(
size += (efip->efi_format.efi_nextents - 1) * sizeof(xfs_extent_t);
efip->efi_format.efi_size = 1;
- log_vector->i_addr = (xfs_caddr_t)&efip->efi_format;
+ log_vector->i_addr = &efip->efi_format;
log_vector->i_len = size;
log_vector->i_type = XLOG_REG_TYPE_EFI_FORMAT;
ASSERT(size >= sizeof(xfs_efi_log_format_t));
@@ -244,7 +244,7 @@ xfs_efi_init(
int
xfs_efi_copy_format(xfs_log_iovec_t *buf, xfs_efi_log_format_t *dst_efi_fmt)
{
- xfs_efi_log_format_t *src_efi_fmt = (xfs_efi_log_format_t *)buf->i_addr;
+ xfs_efi_log_format_t *src_efi_fmt = buf->i_addr;
uint i;
uint len = sizeof(xfs_efi_log_format_t) +
(src_efi_fmt->efi_nextents - 1) * sizeof(xfs_extent_t);
@@ -257,8 +257,7 @@ xfs_efi_copy_format(xfs_log_iovec_t *buf
memcpy((char *)dst_efi_fmt, (char*)src_efi_fmt, len);
return 0;
} else if (buf->i_len == len32) {
- xfs_efi_log_format_32_t *src_efi_fmt_32 =
- (xfs_efi_log_format_32_t *)buf->i_addr;
+ xfs_efi_log_format_32_t *src_efi_fmt_32 = buf->i_addr;
dst_efi_fmt->efi_type = src_efi_fmt_32->efi_type;
dst_efi_fmt->efi_size = src_efi_fmt_32->efi_size;
@@ -272,8 +271,7 @@ xfs_efi_copy_format(xfs_log_iovec_t *buf
}
return 0;
} else if (buf->i_len == len64) {
- xfs_efi_log_format_64_t *src_efi_fmt_64 =
- (xfs_efi_log_format_64_t *)buf->i_addr;
+ xfs_efi_log_format_64_t *src_efi_fmt_64 = buf->i_addr;
dst_efi_fmt->efi_type = src_efi_fmt_64->efi_type;
dst_efi_fmt->efi_size = src_efi_fmt_64->efi_size;
@@ -373,7 +371,7 @@ xfs_efd_item_format(
size += (efdp->efd_format.efd_nextents - 1) * sizeof(xfs_extent_t);
efdp->efd_format.efd_size = 1;
- log_vector->i_addr = (xfs_caddr_t)&efdp->efd_format;
+ log_vector->i_addr = &efdp->efd_format;
log_vector->i_len = size;
log_vector->i_type = XLOG_REG_TYPE_EFD_FORMAT;
ASSERT(size >= sizeof(xfs_efd_log_format_t));
Index: xfs/fs/xfs/xfs_inode_item.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode_item.c 2010-05-29 18:44:47.737033819 +0200
+++ xfs/fs/xfs/xfs_inode_item.c 2010-05-29 18:44:55.999022224 +0200
@@ -217,7 +217,7 @@ xfs_inode_item_format(
int nrecs;
xfs_mount_t *mp;
- vecp->i_addr = (xfs_caddr_t)&iip->ili_format;
+ vecp->i_addr = &iip->ili_format;
vecp->i_len = sizeof(xfs_inode_log_format_t);
vecp->i_type = XLOG_REG_TYPE_IFORMAT;
vecp++;
@@ -268,7 +268,7 @@ xfs_inode_item_format(
*/
xfs_synchronize_times(ip);
- vecp->i_addr = (xfs_caddr_t)&ip->i_d;
+ vecp->i_addr = &ip->i_d;
vecp->i_len = sizeof(struct xfs_icdinode);
vecp->i_type = XLOG_REG_TYPE_ICORE;
vecp++;
@@ -324,8 +324,7 @@ xfs_inode_item_format(
* extents, so just point to the
* real extents array.
*/
- vecp->i_addr =
- (char *)(ip->i_df.if_u1.if_extents);
+ vecp->i_addr = ip->i_df.if_u1.if_extents;
vecp->i_len = ip->i_df.if_bytes;
vecp->i_type = XLOG_REG_TYPE_IEXT;
} else
@@ -343,7 +342,7 @@ xfs_inode_item_format(
ext_buffer = kmem_alloc(ip->i_df.if_bytes,
KM_SLEEP);
iip->ili_extents_buf = ext_buffer;
- vecp->i_addr = (xfs_caddr_t)ext_buffer;
+ vecp->i_addr = ext_buffer;
vecp->i_len = xfs_iextents_copy(ip, ext_buffer,
XFS_DATA_FORK);
vecp->i_type = XLOG_REG_TYPE_IEXT;
@@ -362,7 +361,7 @@ xfs_inode_item_format(
if (iip->ili_format.ilf_fields & XFS_ILOG_DBROOT) {
ASSERT(ip->i_df.if_broot_bytes > 0);
ASSERT(ip->i_df.if_broot != NULL);
- vecp->i_addr = (xfs_caddr_t)ip->i_df.if_broot;
+ vecp->i_addr = ip->i_df.if_broot;
vecp->i_len = ip->i_df.if_broot_bytes;
vecp->i_type = XLOG_REG_TYPE_IBROOT;
vecp++;
@@ -380,7 +379,7 @@ xfs_inode_item_format(
ASSERT(ip->i_df.if_u1.if_data != NULL);
ASSERT(ip->i_d.di_size > 0);
- vecp->i_addr = (xfs_caddr_t)ip->i_df.if_u1.if_data;
+ vecp->i_addr = ip->i_df.if_u1.if_data;
/*
* Round i_bytes up to a word boundary.
* The underlying memory is guaranteed to
@@ -454,7 +453,7 @@ xfs_inode_item_format(
* There are not delayed allocation extents
* for attributes, so just point at the array.
*/
- vecp->i_addr = (char *)(ip->i_afp->if_u1.if_extents);
+ vecp->i_addr = ip->i_afp->if_u1.if_extents;
vecp->i_len = ip->i_afp->if_bytes;
#else
ASSERT(iip->ili_aextents_buf == NULL);
@@ -464,7 +463,7 @@ xfs_inode_item_format(
ext_buffer = kmem_alloc(ip->i_afp->if_bytes,
KM_SLEEP);
iip->ili_aextents_buf = ext_buffer;
- vecp->i_addr = (xfs_caddr_t)ext_buffer;
+ vecp->i_addr = ext_buffer;
vecp->i_len = xfs_iextents_copy(ip, ext_buffer,
XFS_ATTR_FORK);
#endif
@@ -481,7 +480,7 @@ xfs_inode_item_format(
if (iip->ili_format.ilf_fields & XFS_ILOG_ABROOT) {
ASSERT(ip->i_afp->if_broot_bytes > 0);
ASSERT(ip->i_afp->if_broot != NULL);
- vecp->i_addr = (xfs_caddr_t)ip->i_afp->if_broot;
+ vecp->i_addr = ip->i_afp->if_broot;
vecp->i_len = ip->i_afp->if_broot_bytes;
vecp->i_type = XLOG_REG_TYPE_IATTR_BROOT;
vecp++;
@@ -497,7 +496,7 @@ xfs_inode_item_format(
ASSERT(ip->i_afp->if_bytes > 0);
ASSERT(ip->i_afp->if_u1.if_data != NULL);
- vecp->i_addr = (xfs_caddr_t)ip->i_afp->if_u1.if_data;
+ vecp->i_addr = ip->i_afp->if_u1.if_data;
/*
* Round i_bytes up to a word boundary.
* The underlying memory is guaranteed to
@@ -938,9 +937,8 @@ xfs_inode_item_format_convert(
xfs_inode_log_format_t *in_f)
{
if (buf->i_len == sizeof(xfs_inode_log_format_32_t)) {
- xfs_inode_log_format_32_t *in_f32;
+ xfs_inode_log_format_32_t *in_f32 = buf->i_addr;
- in_f32 = (xfs_inode_log_format_32_t *)buf->i_addr;
in_f->ilf_type = in_f32->ilf_type;
in_f->ilf_size = in_f32->ilf_size;
in_f->ilf_fields = in_f32->ilf_fields;
@@ -956,9 +954,8 @@ xfs_inode_item_format_convert(
in_f->ilf_boffset = in_f32->ilf_boffset;
return 0;
} else if (buf->i_len == sizeof(xfs_inode_log_format_64_t)){
- xfs_inode_log_format_64_t *in_f64;
+ xfs_inode_log_format_64_t *in_f64 = buf->i_addr;
- in_f64 = (xfs_inode_log_format_64_t *)buf->i_addr;
in_f->ilf_type = in_f64->ilf_type;
in_f->ilf_size = in_f64->ilf_size;
in_f->ilf_fields = in_f64->ilf_fields;
Index: xfs/fs/xfs/xfs_log.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log.c 2010-05-29 18:44:44.180005323 +0200
+++ xfs/fs/xfs/xfs_log.c 2010-05-29 18:44:56.006006719 +0200
@@ -548,7 +548,7 @@ xfs_log_unmount_write(xfs_mount_t *mp)
.magic = XLOG_UNMOUNT_TYPE,
};
struct xfs_log_iovec reg = {
- .i_addr = (void *)&magic,
+ .i_addr = &magic,
.i_len = sizeof(magic),
.i_type = XLOG_REG_TYPE_UNMOUNT,
};
Index: xfs/fs/xfs/xfs_log.h
===================================================================
--- xfs.orig/fs/xfs/xfs_log.h 2010-05-29 18:34:05.795254730 +0200
+++ xfs/fs/xfs/xfs_log.h 2010-05-29 18:44:56.012049463 +0200
@@ -104,7 +104,7 @@ static inline xfs_lsn_t _lsn_cmp(xfs_lsn
#define XLOG_REG_TYPE_MAX 19
typedef struct xfs_log_iovec {
- xfs_caddr_t i_addr; /* beginning address of region */
+ void *i_addr; /* beginning address of region */
int i_len; /* length in bytes of region */
uint i_type; /* type of region */
} xfs_log_iovec_t;
Index: xfs/fs/xfs/xfs_log_cil.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log_cil.c 2010-05-29 18:44:44.299005460 +0200
+++ xfs/fs/xfs/xfs_log_cil.c 2010-05-29 18:44:56.016261643 +0200
@@ -552,7 +552,7 @@ xlog_cil_push(
thdr.th_type = XFS_TRANS_CHECKPOINT;
thdr.th_tid = tic->t_tid;
thdr.th_num_items = num_iovecs;
- lhdr.i_addr = (xfs_caddr_t)&thdr;
+ lhdr.i_addr = &thdr;
lhdr.i_len = sizeof(xfs_trans_header_t);
lhdr.i_type = XLOG_REG_TYPE_TRANSHDR;
tic->t_curr_res -= lhdr.i_len + sizeof(xlog_op_header_t);
Index: xfs/fs/xfs/xfs_log_recover.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log_recover.c 2010-05-29 18:44:44.185264996 +0200
+++ xfs/fs/xfs/xfs_log_recover.c 2010-05-29 18:44:56.019268488 +0200
@@ -1561,9 +1561,7 @@ xlog_recover_reorder_trans(
list_splice_init(&trans->r_itemq, &sort_list);
list_for_each_entry_safe(item, n, &sort_list, ri_list) {
- xfs_buf_log_format_t *buf_f;
-
- buf_f = (xfs_buf_log_format_t *)item->ri_buf[0].i_addr;
+ xfs_buf_log_format_t *buf_f = item->ri_buf[0].i_addr;
switch (ITEM_TYPE(item)) {
case XFS_LI_BUF:
@@ -1888,9 +1886,8 @@ xlog_recover_do_inode_buffer(
* current di_next_unlinked field. Extract its value
* and copy it to the buffer copy.
*/
- logged_nextp = (xfs_agino_t *)
- ((char *)(item->ri_buf[item_index].i_addr) +
- (next_unlinked_offset - reg_buf_offset));
+ logged_nextp = item->ri_buf[item_index].i_addr +
+ next_unlinked_offset - reg_buf_offset;
if (unlikely(*logged_nextp == 0)) {
xfs_fs_cmn_err(CE_ALERT, mp,
"bad inode buffer log record (ptr = 0x%p, bp = 0x%p). XFS trying to replay bad (0) inode di_next_unlinked field",
@@ -1969,8 +1966,7 @@ xlog_recover_do_reg_buffer(
item->ri_buf[i].i_len, __func__);
goto next;
}
- error = xfs_qm_dqcheck((xfs_disk_dquot_t *)
- item->ri_buf[i].i_addr,
+ error = xfs_qm_dqcheck(item->ri_buf[i].i_addr,
-1, 0, XFS_QMOPT_DOWARN,
"dquot_buf_recover");
if (error)
@@ -2183,7 +2179,7 @@ xlog_recover_do_buffer_trans(
xlog_recover_item_t *item,
int pass)
{
- xfs_buf_log_format_t *buf_f;
+ xfs_buf_log_format_t *buf_f = item->ri_buf[0].i_addr;
xfs_mount_t *mp;
xfs_buf_t *bp;
int error;
@@ -2193,8 +2189,6 @@ xlog_recover_do_buffer_trans(
ushort flags;
uint buf_flags;
- buf_f = (xfs_buf_log_format_t *)item->ri_buf[0].i_addr;
-
if (pass == XLOG_RECOVER_PASS1) {
/*
* In this pass we're only looking for buf items
@@ -2315,10 +2309,9 @@ xlog_recover_do_inode_trans(
}
if (item->ri_buf[0].i_len == sizeof(xfs_inode_log_format_t)) {
- in_f = (xfs_inode_log_format_t *)item->ri_buf[0].i_addr;
+ in_f = item->ri_buf[0].i_addr;
} else {
- in_f = (xfs_inode_log_format_t *)kmem_alloc(
- sizeof(xfs_inode_log_format_t), KM_SLEEP);
+ in_f = kmem_alloc(sizeof(xfs_inode_log_format_t), KM_SLEEP);
need_free = 1;
error = xfs_inode_item_format_convert(&item->ri_buf[0], in_f);
if (error)
@@ -2366,7 +2359,7 @@ xlog_recover_do_inode_trans(
error = EFSCORRUPTED;
goto error;
}
- dicp = (xfs_icdinode_t *)(item->ri_buf[1].i_addr);
+ dicp = item->ri_buf[1].i_addr;
if (unlikely(dicp->di_magic != XFS_DINODE_MAGIC)) {
xfs_buf_relse(bp);
xfs_fs_cmn_err(CE_ALERT, mp,
@@ -2457,7 +2450,7 @@ xlog_recover_do_inode_trans(
}
/* The core is in in-core format */
- xfs_dinode_to_disk(dip, (xfs_icdinode_t *)item->ri_buf[1].i_addr);
+ xfs_dinode_to_disk(dip, item->ri_buf[1].i_addr);
/* the rest is in on-disk format */
if (item->ri_buf[1].i_len > sizeof(struct xfs_icdinode)) {
@@ -2574,7 +2567,7 @@ xlog_recover_do_quotaoff_trans(
return (0);
}
- qoff_f = (xfs_qoff_logformat_t *)item->ri_buf[0].i_addr;
+ qoff_f = item->ri_buf[0].i_addr;
ASSERT(qoff_f);
/*
@@ -2618,9 +2611,8 @@ xlog_recover_do_dquot_trans(
if (mp->m_qflags == 0)
return (0);
- recddq = (xfs_disk_dquot_t *)item->ri_buf[1].i_addr;
-
- if (item->ri_buf[1].i_addr == NULL) {
+ recddq = item->ri_buf[1].i_addr;
+ if (recddq == NULL) {
cmn_err(CE_ALERT,
"XFS: NULL dquot in %s.", __func__);
return XFS_ERROR(EIO);
@@ -2650,7 +2642,7 @@ xlog_recover_do_dquot_trans(
* The other possibility, of course, is that the quota subsystem was
* removed since the last mount - ENOSYS.
*/
- dq_f = (xfs_dq_logformat_t *)item->ri_buf[0].i_addr;
+ dq_f = item->ri_buf[0].i_addr;
ASSERT(dq_f);
if ((error = xfs_qm_dqcheck(recddq,
dq_f->qlf_id,
@@ -2717,7 +2709,7 @@ xlog_recover_do_efi_trans(
return 0;
}
- efi_formatp = (xfs_efi_log_format_t *)item->ri_buf[0].i_addr;
+ efi_formatp = item->ri_buf[0].i_addr;
mp = log->l_mp;
efip = xfs_efi_init(mp, efi_formatp->efi_nextents);
@@ -2763,7 +2755,7 @@ xlog_recover_do_efd_trans(
return;
}
- efd_formatp = (xfs_efd_log_format_t *)item->ri_buf[0].i_addr;
+ efd_formatp = item->ri_buf[0].i_addr;
ASSERT((item->ri_buf[0].i_len == (sizeof(xfs_efd_log_format_32_t) +
((efd_formatp->efd_nextents - 1) * sizeof(xfs_extent_32_t)))) ||
(item->ri_buf[0].i_len == (sizeof(xfs_efd_log_format_64_t) +
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 13/17] xfs: kill the unused xlog_debug variable
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (11 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 12/17] xfs: fix the xfs_log_iovec i_addr type Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 6:02 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 14/17] xfs: remove the unused XFS_LOG_SLEEP and XFS_LOG_NOSLEEP flags Christoph Hellwig
` (5 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-kill-xlog_debug --]
[-- Type: text/plain, Size: 620 bytes --]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/xfs_log.h
===================================================================
--- xfs.orig/fs/xfs/xfs_log.h 2010-05-29 18:44:56.012049463 +0200
+++ xfs/fs/xfs/xfs_log.h 2010-05-29 18:44:57.709003926 +0200
@@ -201,9 +201,4 @@ int xfs_log_commit_cil(struct xfs_mount
bool xfs_log_item_in_current_chkpt(struct xfs_log_item *lip);
#endif
-
-
-extern int xlog_debug; /* set to 1 to enable real log */
-
-
#endif /* __XFS_LOG_H__ */
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 14/17] xfs: remove the unused XFS_LOG_SLEEP and XFS_LOG_NOSLEEP flags
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (12 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 13/17] xfs: kill the unused xlog_debug variable Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 6:02 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 15/17] xfs: remove the unused XFS_TRANS_NOSLEEP/XFS_TRANS_WAIT flags Christoph Hellwig
` (4 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-kill-dead-log-flags --]
[-- Type: text/plain, Size: 1358 bytes --]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/xfs_log.c
===================================================================
--- xfs.orig/fs/xfs/xfs_log.c 2010-05-29 18:21:05.762255846 +0200
+++ xfs/fs/xfs/xfs_log.c 2010-05-29 18:22:52.309004835 +0200
@@ -333,7 +333,6 @@ xfs_log_reserve(
int retval = 0;
ASSERT(client == XFS_TRANSACTION || client == XFS_LOG);
- ASSERT((flags & XFS_LOG_NOSLEEP) == 0);
if (XLOG_FORCED_SHUTDOWN(log))
return XFS_ERROR(EIO);
Index: xfs/fs/xfs/xfs_log.h
===================================================================
--- xfs.orig/fs/xfs/xfs_log.h 2010-05-29 18:22:26.023003369 +0200
+++ xfs/fs/xfs/xfs_log.h 2010-05-29 18:22:52.309004835 +0200
@@ -55,14 +55,10 @@ static inline xfs_lsn_t _lsn_cmp(xfs_lsn
/*
* Flags to xfs_log_reserve()
*
- * XFS_LOG_SLEEP: If space is not available, sleep (default)
- * XFS_LOG_NOSLEEP: If space is not available, return error
* XFS_LOG_PERM_RESERV: Permanent reservation. When writes are
* performed against this type of reservation, the reservation
* is not decreased. Long running transactions should use this.
*/
-#define XFS_LOG_SLEEP 0x0
-#define XFS_LOG_NOSLEEP 0x1
#define XFS_LOG_PERM_RESERV 0x2
/*
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 15/17] xfs: remove the unused XFS_TRANS_NOSLEEP/XFS_TRANS_WAIT flags
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (13 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 14/17] xfs: remove the unused XFS_LOG_SLEEP and XFS_LOG_NOSLEEP flags Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 6:03 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 16/17] xfs: remove unused XFS_BMAPI_ flags Christoph Hellwig
` (3 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-kill-dead-trans-flags --]
[-- Type: text/plain, Size: 625 bytes --]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/xfs_trans.h
===================================================================
--- xfs.orig/fs/xfs/xfs_trans.h 2010-05-29 18:12:37.780017895 +0200
+++ xfs/fs/xfs/xfs_trans.h 2010-05-29 18:23:22.618003507 +0200
@@ -184,8 +184,6 @@ struct xfs_log_item_desc {
/*
* Values for call flags parameter.
*/
-#define XFS_TRANS_NOSLEEP 0x1
-#define XFS_TRANS_WAIT 0x2
#define XFS_TRANS_RELEASE_LOG_RES 0x4
#define XFS_TRANS_ABORT 0x8
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 16/17] xfs: remove unused XFS_BMAPI_ flags
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (14 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 15/17] xfs: remove the unused XFS_TRANS_NOSLEEP/XFS_TRANS_WAIT flags Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 6:04 ` Dave Chinner
2010-05-31 16:07 ` [PATCH 17/17] xfs: remove unused delta tracking code in xfs_bmapi Christoph Hellwig
` (2 subsequent siblings)
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-bmapi-cleanup-flags --]
[-- Type: text/plain, Size: 3603 bytes --]
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/xfs_bmap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.c 2010-05-29 18:31:14.248017615 +0200
+++ xfs/fs/xfs/xfs_bmap.c 2010-05-29 18:31:15.146003157 +0200
@@ -4605,19 +4605,13 @@ xfs_bmapi(
* allocate the stuff asked for in this bmap call
* but that wouldn't be as good.
*/
- if (wasdelay && !(flags & XFS_BMAPI_EXACT)) {
+ if (wasdelay) {
alen = (xfs_extlen_t)got.br_blockcount;
aoff = got.br_startoff;
if (lastx != NULLEXTNUM && lastx) {
ep = xfs_iext_get_ext(ifp, lastx - 1);
xfs_bmbt_get_all(ep, &prev);
}
- } else if (wasdelay) {
- alen = (xfs_extlen_t)
- XFS_FILBLKS_MIN(len,
- (got.br_startoff +
- got.br_blockcount) - bno);
- aoff = bno;
} else {
alen = (xfs_extlen_t)
XFS_FILBLKS_MIN(len, MAXEXTLEN);
Index: xfs/fs/xfs/xfs_bmap.h
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.h 2010-05-29 18:09:00.436004974 +0200
+++ xfs/fs/xfs/xfs_bmap.h 2010-05-29 18:31:15.149023970 +0200
@@ -82,16 +82,13 @@ typedef struct xfs_bmap_free
#define XFS_BMAPI_DELAY 0x002 /* delayed write operation */
#define XFS_BMAPI_ENTIRE 0x004 /* return entire extent, not trimmed */
#define XFS_BMAPI_METADATA 0x008 /* mapping metadata not user data */
-#define XFS_BMAPI_EXACT 0x010 /* allocate only to spec'd bounds */
-#define XFS_BMAPI_ATTRFORK 0x020 /* use attribute fork not data */
-#define XFS_BMAPI_ASYNC 0x040 /* bunmapi xactions can be async */
-#define XFS_BMAPI_RSVBLOCKS 0x080 /* OK to alloc. reserved data blocks */
-#define XFS_BMAPI_PREALLOC 0x100 /* preallocation op: unwritten space */
-#define XFS_BMAPI_IGSTATE 0x200 /* Ignore state - */
+#define XFS_BMAPI_ATTRFORK 0x010 /* use attribute fork not data */
+#define XFS_BMAPI_RSVBLOCKS 0x020 /* OK to alloc. reserved data blocks */
+#define XFS_BMAPI_PREALLOC 0x040 /* preallocation op: unwritten space */
+#define XFS_BMAPI_IGSTATE 0x080 /* Ignore state - */
/* combine contig. space */
-#define XFS_BMAPI_CONTIG 0x400 /* must allocate only one extent */
-/* XFS_BMAPI_DIRECT_IO 0x800 */
-#define XFS_BMAPI_CONVERT 0x1000 /* unwritten extent conversion - */
+#define XFS_BMAPI_CONTIG 0x100 /* must allocate only one extent */
+#define XFS_BMAPI_CONVERT 0x200 /* unwritten extent conversion - */
/* need write cache flushing and no */
/* additional allocation alignments */
@@ -100,9 +97,7 @@ typedef struct xfs_bmap_free
{ XFS_BMAPI_DELAY, "DELAY" }, \
{ XFS_BMAPI_ENTIRE, "ENTIRE" }, \
{ XFS_BMAPI_METADATA, "METADATA" }, \
- { XFS_BMAPI_EXACT, "EXACT" }, \
{ XFS_BMAPI_ATTRFORK, "ATTRFORK" }, \
- { XFS_BMAPI_ASYNC, "ASYNC" }, \
{ XFS_BMAPI_RSVBLOCKS, "RSVBLOCKS" }, \
{ XFS_BMAPI_PREALLOC, "PREALLOC" }, \
{ XFS_BMAPI_IGSTATE, "IGSTATE" }, \
Index: xfs/fs/xfs/xfs_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.c 2010-05-29 18:17:15.205004416 +0200
+++ xfs/fs/xfs/xfs_inode.c 2010-05-29 18:31:15.156026205 +0200
@@ -1588,8 +1588,7 @@ xfs_itruncate_finish(
xfs_bmap_init(&free_list, &first_block);
error = xfs_bunmapi(ntp, ip,
first_unmap_block, unmap_len,
- xfs_bmapi_aflag(fork) |
- (sync ? 0 : XFS_BMAPI_ASYNC),
+ xfs_bmapi_aflag(fork),
XFS_ITRUNC_MAX_EXTENTS,
&first_block, &free_list,
NULL, &done);
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* [PATCH 17/17] xfs: remove unused delta tracking code in xfs_bmapi
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (15 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 16/17] xfs: remove unused XFS_BMAPI_ flags Christoph Hellwig
@ 2010-05-31 16:07 ` Christoph Hellwig
2010-06-02 6:11 ` Dave Chinner
2010-06-02 6:13 ` [PATCH 00/17] pending patches Dave Chinner
2010-06-03 17:01 ` Alex Elder
18 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-05-31 16:07 UTC (permalink / raw)
To: xfs
[-- Attachment #1: xfs-bmapi-delta-cleanup --]
[-- Type: text/plain, Size: 38742 bytes --]
This code was introduced four years ago in commit
3e57ecf640428c01ba1ed8c8fc538447ada1715b without any review and has
been unused since. Remove it just as the rest of the code introduced
in that commit to reduce that stack usage and complexity in this central
piece of code.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Index: xfs/fs/xfs/linux-2.6/xfs_aops.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_aops.c 2010-05-29 18:44:44.017005812 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_aops.c 2010-05-29 18:45:03.756004695 +0200
@@ -974,7 +974,7 @@ xfs_aops_discard_page(
*/
error = xfs_bmapi(NULL, ip, offset_fsb, 1,
XFS_BMAPI_ENTIRE, NULL, 0, &imap,
- &nimaps, NULL, NULL);
+ &nimaps, NULL);
if (error) {
/* something screwed, just bail */
@@ -1002,7 +1002,7 @@ xfs_aops_discard_page(
*/
xfs_bmap_init(&flist, &firstblock);
error = xfs_bunmapi(NULL, ip, offset_fsb, 1, 0, 1, &firstblock,
- &flist, NULL, &done);
+ &flist, &done);
ASSERT(!flist.xbf_count && !flist.xbf_first);
if (error) {
Index: xfs/fs/xfs/linux-2.6/xfs_file.c
===================================================================
--- xfs.orig/fs/xfs/linux-2.6/xfs_file.c 2010-05-29 18:44:47.747005393 +0200
+++ xfs/fs/xfs/linux-2.6/xfs_file.c 2010-05-29 18:45:03.757024530 +0200
@@ -414,7 +414,7 @@ xfs_zero_last_block(
last_fsb = XFS_B_TO_FSBT(mp, isize);
nimaps = 1;
error = xfs_bmapi(NULL, ip, last_fsb, 1, 0, NULL, 0, &imap,
- &nimaps, NULL, NULL);
+ &nimaps, NULL);
if (error) {
return error;
}
@@ -509,7 +509,7 @@ xfs_zero_eof(
nimaps = 1;
zero_count_fsb = end_zero_fsb - start_zero_fsb + 1;
error = xfs_bmapi(NULL, ip, start_zero_fsb, zero_count_fsb,
- 0, NULL, 0, &imap, &nimaps, NULL, NULL);
+ 0, NULL, 0, &imap, &nimaps, NULL);
if (error) {
ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_IOLOCK_EXCL));
return error;
Index: xfs/fs/xfs/quota/xfs_dquot.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_dquot.c 2010-05-29 18:44:47.671005673 +0200
+++ xfs/fs/xfs/quota/xfs_dquot.c 2010-05-29 18:45:03.762254380 +0200
@@ -385,7 +385,7 @@ xfs_qm_dqalloc(
XFS_BMAPI_METADATA | XFS_BMAPI_WRITE,
&firstblock,
XFS_QM_DQALLOC_SPACE_RES(mp),
- &map, &nmaps, &flist, NULL))) {
+ &map, &nmaps, &flist))) {
goto error0;
}
ASSERT(map.br_blockcount == XFS_DQUOT_CLUSTER_SIZE_FSB);
@@ -501,7 +501,7 @@ xfs_qm_dqtobp(
error = xfs_bmapi(NULL, quotip, dqp->q_fileoffset,
XFS_DQUOT_CLUSTER_SIZE_FSB,
XFS_BMAPI_METADATA,
- NULL, 0, &map, &nmaps, NULL, NULL);
+ NULL, 0, &map, &nmaps, NULL);
xfs_iunlock(quotip, XFS_ILOCK_SHARED);
if (error)
Index: xfs/fs/xfs/quota/xfs_qm.c
===================================================================
--- xfs.orig/fs/xfs/quota/xfs_qm.c 2010-05-29 18:44:44.061032213 +0200
+++ xfs/fs/xfs/quota/xfs_qm.c 2010-05-29 18:45:03.769254450 +0200
@@ -1490,7 +1490,7 @@ xfs_qm_dqiterate(
maxlblkcnt - lblkno,
XFS_BMAPI_METADATA,
NULL,
- 0, map, &nmaps, NULL, NULL);
+ 0, map, &nmaps, NULL);
xfs_iunlock(qip, XFS_ILOCK_SHARED);
if (error)
break;
Index: xfs/fs/xfs/xfs_attr.c
===================================================================
--- xfs.orig/fs/xfs/xfs_attr.c 2010-05-29 18:44:47.702005742 +0200
+++ xfs/fs/xfs/xfs_attr.c 2010-05-29 18:45:03.775276311 +0200
@@ -1977,7 +1977,7 @@ xfs_attr_rmtval_get(xfs_da_args_t *args)
error = xfs_bmapi(args->trans, args->dp, (xfs_fileoff_t)lblkno,
args->rmtblkcnt,
XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
- NULL, 0, map, &nmap, NULL, NULL);
+ NULL, 0, map, &nmap, NULL);
if (error)
return(error);
ASSERT(nmap >= 1);
@@ -2056,7 +2056,7 @@ xfs_attr_rmtval_set(xfs_da_args_t *args)
XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA |
XFS_BMAPI_WRITE,
args->firstblock, args->total, &map, &nmap,
- args->flist, NULL);
+ args->flist);
if (!error) {
error = xfs_bmap_finish(&args->trans, args->flist,
&committed);
@@ -2107,7 +2107,7 @@ xfs_attr_rmtval_set(xfs_da_args_t *args)
args->rmtblkcnt,
XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
args->firstblock, 0, &map, &nmap,
- NULL, NULL);
+ NULL);
if (error) {
return(error);
}
@@ -2172,7 +2172,7 @@ xfs_attr_rmtval_remove(xfs_da_args_t *ar
args->rmtblkcnt,
XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
args->firstblock, 0, &map, &nmap,
- args->flist, NULL);
+ args->flist);
if (error) {
return(error);
}
@@ -2210,7 +2210,7 @@ xfs_attr_rmtval_remove(xfs_da_args_t *ar
error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
1, args->firstblock, args->flist,
- NULL, &done);
+ &done);
if (!error) {
error = xfs_bmap_finish(&args->trans, args->flist,
&committed);
Index: xfs/fs/xfs/xfs_attr_leaf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_attr_leaf.c 2010-05-29 18:44:44.107255638 +0200
+++ xfs/fs/xfs/xfs_attr_leaf.c 2010-05-29 18:45:03.779011469 +0200
@@ -2928,7 +2928,7 @@ xfs_attr_leaf_freextent(xfs_trans_t **tr
nmap = 1;
error = xfs_bmapi(*trans, dp, (xfs_fileoff_t)tblkno, tblkcnt,
XFS_BMAPI_ATTRFORK | XFS_BMAPI_METADATA,
- NULL, 0, &map, &nmap, NULL, NULL);
+ NULL, 0, &map, &nmap, NULL);
if (error) {
return(error);
}
Index: xfs/fs/xfs/xfs_bmap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.c 2010-05-29 18:45:02.175004206 +0200
+++ xfs/fs/xfs/xfs_bmap.c 2010-05-29 18:45:03.785016777 +0200
@@ -101,7 +101,6 @@ xfs_bmap_add_extent(
xfs_fsblock_t *first, /* pointer to firstblock variable */
xfs_bmap_free_t *flist, /* list of extents to be freed */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int whichfork, /* data or attr fork */
int rsvd); /* OK to allocate reserved blocks */
@@ -119,7 +118,6 @@ xfs_bmap_add_extent_delay_real(
xfs_fsblock_t *first, /* pointer to firstblock variable */
xfs_bmap_free_t *flist, /* list of extents to be freed */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int rsvd); /* OK to allocate reserved blocks */
/*
@@ -132,7 +130,6 @@ xfs_bmap_add_extent_hole_delay(
xfs_extnum_t idx, /* extent number to update/insert */
xfs_bmbt_irec_t *new, /* new data to add to file extents */
int *logflagsp,/* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int rsvd); /* OK to allocate reserved blocks */
/*
@@ -146,7 +143,6 @@ xfs_bmap_add_extent_hole_real(
xfs_btree_cur_t *cur, /* if null, not a btree */
xfs_bmbt_irec_t *new, /* new data to add to file extents */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int whichfork); /* data or attr fork */
/*
@@ -159,8 +155,7 @@ xfs_bmap_add_extent_unwritten_real(
xfs_extnum_t idx, /* extent number to update/insert */
xfs_btree_cur_t **curp, /* if *curp is null, not a btree */
xfs_bmbt_irec_t *new, /* new data to add to file extents */
- int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta); /* Change made to incore extents */
+ int *logflagsp); /* inode logging flags */
/*
* xfs_bmap_alloc is called by xfs_bmapi to allocate an extent for a file.
@@ -197,7 +192,6 @@ xfs_bmap_del_extent(
xfs_btree_cur_t *cur, /* if null, not a btree */
xfs_bmbt_irec_t *new, /* new data to add to file extents */
int *logflagsp,/* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int whichfork, /* data or attr fork */
int rsvd); /* OK to allocate reserved blocks */
@@ -486,7 +480,6 @@ xfs_bmap_add_extent(
xfs_fsblock_t *first, /* pointer to firstblock variable */
xfs_bmap_free_t *flist, /* list of extents to be freed */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int whichfork, /* data or attr fork */
int rsvd) /* OK to use reserved data blocks */
{
@@ -521,15 +514,6 @@ xfs_bmap_add_extent(
logflags = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
} else
logflags = 0;
- /* DELTA: single new extent */
- if (delta) {
- if (delta->xed_startoff > new->br_startoff)
- delta->xed_startoff = new->br_startoff;
- if (delta->xed_blockcount <
- new->br_startoff + new->br_blockcount)
- delta->xed_blockcount = new->br_startoff +
- new->br_blockcount;
- }
}
/*
* Any kind of new delayed allocation goes here.
@@ -539,7 +523,7 @@ xfs_bmap_add_extent(
ASSERT((cur->bc_private.b.flags &
XFS_BTCUR_BPRV_WASDEL) == 0);
if ((error = xfs_bmap_add_extent_hole_delay(ip, idx, new,
- &logflags, delta, rsvd)))
+ &logflags, rsvd)))
goto done;
}
/*
@@ -550,7 +534,7 @@ xfs_bmap_add_extent(
ASSERT((cur->bc_private.b.flags &
XFS_BTCUR_BPRV_WASDEL) == 0);
if ((error = xfs_bmap_add_extent_hole_real(ip, idx, cur, new,
- &logflags, delta, whichfork)))
+ &logflags, whichfork)))
goto done;
} else {
xfs_bmbt_irec_t prev; /* old extent at offset idx */
@@ -575,17 +559,17 @@ xfs_bmap_add_extent(
XFS_BTCUR_BPRV_WASDEL);
if ((error = xfs_bmap_add_extent_delay_real(ip,
idx, &cur, new, &da_new, first, flist,
- &logflags, delta, rsvd)))
+ &logflags, rsvd)))
goto done;
} else if (new->br_state == XFS_EXT_NORM) {
ASSERT(new->br_state == XFS_EXT_NORM);
if ((error = xfs_bmap_add_extent_unwritten_real(
- ip, idx, &cur, new, &logflags, delta)))
+ ip, idx, &cur, new, &logflags)))
goto done;
} else {
ASSERT(new->br_state == XFS_EXT_UNWRITTEN);
if ((error = xfs_bmap_add_extent_unwritten_real(
- ip, idx, &cur, new, &logflags, delta)))
+ ip, idx, &cur, new, &logflags)))
goto done;
}
ASSERT(*curp == cur || *curp == NULL);
@@ -598,7 +582,7 @@ xfs_bmap_add_extent(
ASSERT((cur->bc_private.b.flags &
XFS_BTCUR_BPRV_WASDEL) == 0);
if ((error = xfs_bmap_add_extent_hole_real(ip, idx, cur,
- new, &logflags, delta, whichfork)))
+ new, &logflags, whichfork)))
goto done;
}
}
@@ -663,7 +647,6 @@ xfs_bmap_add_extent_delay_real(
xfs_fsblock_t *first, /* pointer to firstblock variable */
xfs_bmap_free_t *flist, /* list of extents to be freed */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int rsvd) /* OK to use reserved data block allocation */
{
xfs_btree_cur_t *cur; /* btree cursor */
@@ -794,11 +777,6 @@ xfs_bmap_add_extent_delay_real(
goto done;
}
*dnew = 0;
- /* DELTA: Three in-core extents are replaced by one. */
- temp = LEFT.br_startoff;
- temp2 = LEFT.br_blockcount +
- PREV.br_blockcount +
- RIGHT.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
@@ -829,10 +807,6 @@ xfs_bmap_add_extent_delay_real(
goto done;
}
*dnew = 0;
- /* DELTA: Two in-core extents are replaced by one. */
- temp = LEFT.br_startoff;
- temp2 = LEFT.br_blockcount +
- PREV.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
@@ -864,10 +838,6 @@ xfs_bmap_add_extent_delay_real(
goto done;
}
*dnew = 0;
- /* DELTA: Two in-core extents are replaced by one. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount +
- RIGHT.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
@@ -897,9 +867,6 @@ xfs_bmap_add_extent_delay_real(
XFS_WANT_CORRUPTED_GOTO(i == 1, done);
}
*dnew = 0;
- /* DELTA: The in-core extent described by new changed type. */
- temp = new->br_startoff;
- temp2 = new->br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
@@ -939,10 +906,6 @@ xfs_bmap_add_extent_delay_real(
xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
*dnew = temp;
- /* DELTA: The boundary between two in-core extents moved. */
- temp = LEFT.br_startoff;
- temp2 = LEFT.br_blockcount +
- PREV.br_blockcount;
break;
case BMAP_LEFT_FILLING:
@@ -987,9 +950,6 @@ xfs_bmap_add_extent_delay_real(
xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
trace_xfs_bmap_post_update(ip, idx + 1, state, _THIS_IP_);
*dnew = temp;
- /* DELTA: One in-core extent is split in two. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount;
break;
case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
@@ -1028,10 +988,6 @@ xfs_bmap_add_extent_delay_real(
xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
*dnew = temp;
- /* DELTA: The boundary between two in-core extents moved. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount +
- RIGHT.br_blockcount;
break;
case BMAP_RIGHT_FILLING:
@@ -1075,9 +1031,6 @@ xfs_bmap_add_extent_delay_real(
xfs_bmbt_set_startblock(ep, nullstartblock((int)temp));
trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
*dnew = temp;
- /* DELTA: One in-core extent is split in two. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount;
break;
case 0:
@@ -1158,9 +1111,6 @@ xfs_bmap_add_extent_delay_real(
nullstartblock((int)temp2));
trace_xfs_bmap_post_update(ip, idx + 2, state, _THIS_IP_);
*dnew = temp + temp2;
- /* DELTA: One in-core extent is split in three. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
@@ -1176,13 +1126,6 @@ xfs_bmap_add_extent_delay_real(
ASSERT(0);
}
*curp = cur;
- if (delta) {
- temp2 += temp;
- if (delta->xed_startoff > temp)
- delta->xed_startoff = temp;
- if (delta->xed_blockcount < temp2)
- delta->xed_blockcount = temp2;
- }
done:
*logflagsp = rval;
return error;
@@ -1201,8 +1144,7 @@ xfs_bmap_add_extent_unwritten_real(
xfs_extnum_t idx, /* extent number to update/insert */
xfs_btree_cur_t **curp, /* if *curp is null, not a btree */
xfs_bmbt_irec_t *new, /* new data to add to file extents */
- int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta) /* Change made to incore extents */
+ int *logflagsp) /* inode logging flags */
{
xfs_btree_cur_t *cur; /* btree cursor */
xfs_bmbt_rec_host_t *ep; /* extent entry for idx */
@@ -1216,8 +1158,6 @@ xfs_bmap_add_extent_unwritten_real(
/* left is 0, right is 1, prev is 2 */
int rval=0; /* return value (logging flags) */
int state = 0;/* state bits, accessed thru macros */
- xfs_filblks_t temp=0;
- xfs_filblks_t temp2=0;
#define LEFT r[0]
#define RIGHT r[1]
@@ -1338,11 +1278,6 @@ xfs_bmap_add_extent_unwritten_real(
RIGHT.br_blockcount, LEFT.br_state)))
goto done;
}
- /* DELTA: Three in-core extents are replaced by one. */
- temp = LEFT.br_startoff;
- temp2 = LEFT.br_blockcount +
- PREV.br_blockcount +
- RIGHT.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_LEFT_CONTIG:
@@ -1379,10 +1314,6 @@ xfs_bmap_add_extent_unwritten_real(
LEFT.br_state)))
goto done;
}
- /* DELTA: Two in-core extents are replaced by one. */
- temp = LEFT.br_startoff;
- temp2 = LEFT.br_blockcount +
- PREV.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
@@ -1419,10 +1350,6 @@ xfs_bmap_add_extent_unwritten_real(
newext)))
goto done;
}
- /* DELTA: Two in-core extents are replaced by one. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount +
- RIGHT.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_RIGHT_FILLING:
@@ -1450,9 +1377,6 @@ xfs_bmap_add_extent_unwritten_real(
newext)))
goto done;
}
- /* DELTA: The in-core extent described by new changed type. */
- temp = new->br_startoff;
- temp2 = new->br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG:
@@ -1498,10 +1422,6 @@ xfs_bmap_add_extent_unwritten_real(
LEFT.br_state))
goto done;
}
- /* DELTA: The boundary between two in-core extents moved. */
- temp = LEFT.br_startoff;
- temp2 = LEFT.br_blockcount +
- PREV.br_blockcount;
break;
case BMAP_LEFT_FILLING:
@@ -1541,9 +1461,6 @@ xfs_bmap_add_extent_unwritten_real(
goto done;
XFS_WANT_CORRUPTED_GOTO(i == 1, done);
}
- /* DELTA: One in-core extent is split in two. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount;
break;
case BMAP_RIGHT_FILLING | BMAP_RIGHT_CONTIG:
@@ -1584,10 +1501,6 @@ xfs_bmap_add_extent_unwritten_real(
newext)))
goto done;
}
- /* DELTA: The boundary between two in-core extents moved. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount +
- RIGHT.br_blockcount;
break;
case BMAP_RIGHT_FILLING:
@@ -1627,9 +1540,6 @@ xfs_bmap_add_extent_unwritten_real(
goto done;
XFS_WANT_CORRUPTED_GOTO(i == 1, done);
}
- /* DELTA: One in-core extent is split in two. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount;
break;
case 0:
@@ -1689,9 +1599,6 @@ xfs_bmap_add_extent_unwritten_real(
goto done;
XFS_WANT_CORRUPTED_GOTO(i == 1, done);
}
- /* DELTA: One in-core extent is split in three. */
- temp = PREV.br_startoff;
- temp2 = PREV.br_blockcount;
break;
case BMAP_LEFT_FILLING | BMAP_LEFT_CONTIG | BMAP_RIGHT_CONTIG:
@@ -1707,13 +1614,6 @@ xfs_bmap_add_extent_unwritten_real(
ASSERT(0);
}
*curp = cur;
- if (delta) {
- temp2 += temp;
- if (delta->xed_startoff > temp)
- delta->xed_startoff = temp;
- if (delta->xed_blockcount < temp2)
- delta->xed_blockcount = temp2;
- }
done:
*logflagsp = rval;
return error;
@@ -1733,7 +1633,6 @@ xfs_bmap_add_extent_hole_delay(
xfs_extnum_t idx, /* extent number to update/insert */
xfs_bmbt_irec_t *new, /* new data to add to file extents */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int rsvd) /* OK to allocate reserved blocks */
{
xfs_bmbt_rec_host_t *ep; /* extent record for idx */
@@ -1744,7 +1643,6 @@ xfs_bmap_add_extent_hole_delay(
xfs_bmbt_irec_t right; /* right neighbor extent entry */
int state; /* state bits, accessed thru macros */
xfs_filblks_t temp=0; /* temp for indirect calculations */
- xfs_filblks_t temp2=0;
ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
ep = xfs_iext_get_ext(ifp, idx);
@@ -1816,9 +1714,6 @@ xfs_bmap_add_extent_hole_delay(
xfs_iext_remove(ip, idx, 1, state);
ip->i_df.if_lastex = idx - 1;
- /* DELTA: Two in-core extents were replaced by one. */
- temp2 = temp;
- temp = left.br_startoff;
break;
case BMAP_LEFT_CONTIG:
@@ -1838,9 +1733,6 @@ xfs_bmap_add_extent_hole_delay(
trace_xfs_bmap_post_update(ip, idx - 1, state, _THIS_IP_);
ip->i_df.if_lastex = idx - 1;
- /* DELTA: One in-core extent grew into a hole. */
- temp2 = temp;
- temp = left.br_startoff;
break;
case BMAP_RIGHT_CONTIG:
@@ -1859,9 +1751,6 @@ xfs_bmap_add_extent_hole_delay(
trace_xfs_bmap_post_update(ip, idx, state, _THIS_IP_);
ip->i_df.if_lastex = idx;
- /* DELTA: One in-core extent grew into a hole. */
- temp2 = temp;
- temp = new->br_startoff;
break;
case 0:
@@ -1873,9 +1762,6 @@ xfs_bmap_add_extent_hole_delay(
oldlen = newlen = 0;
xfs_iext_insert(ip, idx, 1, new, state);
ip->i_df.if_lastex = idx;
- /* DELTA: A new in-core extent was added in a hole. */
- temp2 = new->br_blockcount;
- temp = new->br_startoff;
break;
}
if (oldlen != newlen) {
@@ -1886,13 +1772,6 @@ xfs_bmap_add_extent_hole_delay(
* Nothing to do for disk quota accounting here.
*/
}
- if (delta) {
- temp2 += temp;
- if (delta->xed_startoff > temp)
- delta->xed_startoff = temp;
- if (delta->xed_blockcount < temp2)
- delta->xed_blockcount = temp2;
- }
*logflagsp = 0;
return 0;
}
@@ -1908,7 +1787,6 @@ xfs_bmap_add_extent_hole_real(
xfs_btree_cur_t *cur, /* if null, not a btree */
xfs_bmbt_irec_t *new, /* new data to add to file extents */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int whichfork) /* data or attr fork */
{
xfs_bmbt_rec_host_t *ep; /* pointer to extent entry ins. point */
@@ -1919,8 +1797,6 @@ xfs_bmap_add_extent_hole_real(
xfs_bmbt_irec_t right; /* right neighbor extent entry */
int rval=0; /* return value (logging flags) */
int state; /* state bits, accessed thru macros */
- xfs_filblks_t temp=0;
- xfs_filblks_t temp2=0;
ifp = XFS_IFORK_PTR(ip, whichfork);
ASSERT(idx <= ifp->if_bytes / (uint)sizeof(xfs_bmbt_rec_t));
@@ -2017,11 +1893,6 @@ xfs_bmap_add_extent_hole_real(
left.br_state)))
goto done;
}
- /* DELTA: Two in-core extents were replaced by one. */
- temp = left.br_startoff;
- temp2 = left.br_blockcount +
- new->br_blockcount +
- right.br_blockcount;
break;
case BMAP_LEFT_CONTIG:
@@ -2053,10 +1924,6 @@ xfs_bmap_add_extent_hole_real(
left.br_state)))
goto done;
}
- /* DELTA: One in-core extent grew. */
- temp = left.br_startoff;
- temp2 = left.br_blockcount +
- new->br_blockcount;
break;
case BMAP_RIGHT_CONTIG:
@@ -2089,10 +1956,6 @@ xfs_bmap_add_extent_hole_real(
right.br_state)))
goto done;
}
- /* DELTA: One in-core extent grew. */
- temp = new->br_startoff;
- temp2 = new->br_blockcount +
- right.br_blockcount;
break;
case 0:
@@ -2120,18 +1983,8 @@ xfs_bmap_add_extent_hole_real(
goto done;
XFS_WANT_CORRUPTED_GOTO(i == 1, done);
}
- /* DELTA: A new extent was added in a hole. */
- temp = new->br_startoff;
- temp2 = new->br_blockcount;
break;
}
- if (delta) {
- temp2 += temp;
- if (delta->xed_startoff > temp)
- delta->xed_startoff = temp;
- if (delta->xed_blockcount < temp2)
- delta->xed_blockcount = temp2;
- }
done:
*logflagsp = rval;
return error;
@@ -2944,7 +2797,6 @@ xfs_bmap_del_extent(
xfs_btree_cur_t *cur, /* if null, not a btree */
xfs_bmbt_irec_t *del, /* data to remove from extents */
int *logflagsp, /* inode logging flags */
- xfs_extdelta_t *delta, /* Change made to incore extents */
int whichfork, /* data or attr fork */
int rsvd) /* OK to allocate reserved blocks */
{
@@ -3250,14 +3102,6 @@ xfs_bmap_del_extent(
if (da_old > da_new)
xfs_mod_incore_sb(mp, XFS_SBS_FDBLOCKS, (int64_t)(da_old - da_new),
rsvd);
- if (delta) {
- /* DELTA: report the original extent. */
- if (delta->xed_startoff > got.br_startoff)
- delta->xed_startoff = got.br_startoff;
- if (delta->xed_blockcount < got.br_startoff+got.br_blockcount)
- delta->xed_blockcount = got.br_startoff +
- got.br_blockcount;
- }
done:
*logflagsp = flags;
return error;
@@ -4469,8 +4313,7 @@ xfs_bmapi(
xfs_extlen_t total, /* total blocks needed */
xfs_bmbt_irec_t *mval, /* output: map values */
int *nmap, /* i/o: mval size/count */
- xfs_bmap_free_t *flist, /* i/o: list extents to free */
- xfs_extdelta_t *delta) /* o: change made to incore extents */
+ xfs_bmap_free_t *flist) /* i/o: list extents to free */
{
xfs_fsblock_t abno; /* allocated block number */
xfs_extlen_t alen; /* allocated extent length */
@@ -4582,10 +4425,7 @@ xfs_bmapi(
end = bno + len;
obno = bno;
bma.ip = NULL;
- if (delta) {
- delta->xed_startoff = NULLFILEOFF;
- delta->xed_blockcount = 0;
- }
+
while (bno < end && n < *nmap) {
/*
* Reading past eof, act as though there's a hole
@@ -4811,7 +4651,7 @@ xfs_bmapi(
got.br_state = XFS_EXT_UNWRITTEN;
}
error = xfs_bmap_add_extent(ip, lastx, &cur, &got,
- firstblock, flist, &tmp_logflags, delta,
+ firstblock, flist, &tmp_logflags,
whichfork, (flags & XFS_BMAPI_RSVBLOCKS));
logflags |= tmp_logflags;
if (error)
@@ -4907,7 +4747,7 @@ xfs_bmapi(
}
mval->br_state = XFS_EXT_NORM;
error = xfs_bmap_add_extent(ip, lastx, &cur, mval,
- firstblock, flist, &tmp_logflags, delta,
+ firstblock, flist, &tmp_logflags,
whichfork, (flags & XFS_BMAPI_RSVBLOCKS));
logflags |= tmp_logflags;
if (error)
@@ -4997,14 +4837,6 @@ xfs_bmapi(
ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE ||
XFS_IFORK_NEXTENTS(ip, whichfork) > ifp->if_ext_max);
error = 0;
- if (delta && delta->xed_startoff != NULLFILEOFF) {
- /* A change was actually made.
- * Note that delta->xed_blockount is an offset at this
- * point and needs to be converted to a block count.
- */
- ASSERT(delta->xed_blockcount > delta->xed_startoff);
- delta->xed_blockcount -= delta->xed_startoff;
- }
error0:
/*
* Log everything. Do this after conversion, there's no point in
@@ -5116,8 +4948,6 @@ xfs_bunmapi(
xfs_fsblock_t *firstblock, /* first allocated block
controls a.g. for allocs */
xfs_bmap_free_t *flist, /* i/o: list extents to free */
- xfs_extdelta_t *delta, /* o: change made to incore
- extents */
int *done) /* set if not done yet */
{
xfs_btree_cur_t *cur; /* bmap btree cursor */
@@ -5176,10 +5006,7 @@ xfs_bunmapi(
bno = start + len - 1;
ep = xfs_bmap_search_extents(ip, bno, whichfork, &eof, &lastx, &got,
&prev);
- if (delta) {
- delta->xed_startoff = NULLFILEOFF;
- delta->xed_blockcount = 0;
- }
+
/*
* Check to see if the given block number is past the end of the
* file, back up to the last block if so...
@@ -5277,7 +5104,7 @@ xfs_bunmapi(
}
del.br_state = XFS_EXT_UNWRITTEN;
error = xfs_bmap_add_extent(ip, lastx, &cur, &del,
- firstblock, flist, &logflags, delta,
+ firstblock, flist, &logflags,
XFS_DATA_FORK, 0);
if (error)
goto error0;
@@ -5332,7 +5159,7 @@ xfs_bunmapi(
prev.br_state = XFS_EXT_UNWRITTEN;
error = xfs_bmap_add_extent(ip, lastx - 1, &cur,
&prev, firstblock, flist, &logflags,
- delta, XFS_DATA_FORK, 0);
+ XFS_DATA_FORK, 0);
if (error)
goto error0;
goto nodelete;
@@ -5341,7 +5168,7 @@ xfs_bunmapi(
del.br_state = XFS_EXT_UNWRITTEN;
error = xfs_bmap_add_extent(ip, lastx, &cur,
&del, firstblock, flist, &logflags,
- delta, XFS_DATA_FORK, 0);
+ XFS_DATA_FORK, 0);
if (error)
goto error0;
goto nodelete;
@@ -5394,7 +5221,7 @@ xfs_bunmapi(
goto error0;
}
error = xfs_bmap_del_extent(ip, tp, lastx, flist, cur, &del,
- &tmp_logflags, delta, whichfork, rsvd);
+ &tmp_logflags, whichfork, rsvd);
logflags |= tmp_logflags;
if (error)
goto error0;
@@ -5451,14 +5278,6 @@ nodelete:
ASSERT(ifp->if_ext_max ==
XFS_IFORK_SIZE(ip, whichfork) / (uint)sizeof(xfs_bmbt_rec_t));
error = 0;
- if (delta && delta->xed_startoff != NULLFILEOFF) {
- /* A change was actually made.
- * Note that delta->xed_blockount is an offset at this
- * point and needs to be converted to a block count.
- */
- ASSERT(delta->xed_blockcount > delta->xed_startoff);
- delta->xed_blockcount -= delta->xed_startoff;
- }
error0:
/*
* Log everything. Do this after conversion, there's no point in
@@ -5671,7 +5490,7 @@ xfs_getbmap(
error = xfs_bmapi(NULL, ip, XFS_BB_TO_FSBT(mp, bmv->bmv_offset),
XFS_BB_TO_FSB(mp, bmv->bmv_length),
bmapi_flags, NULL, 0, map, &nmap,
- NULL, NULL);
+ NULL);
if (error)
goto out_free_map;
ASSERT(nmap <= subnex);
Index: xfs/fs/xfs/xfs_bmap.h
===================================================================
--- xfs.orig/fs/xfs/xfs_bmap.h 2010-05-29 18:45:02.175004206 +0200
+++ xfs/fs/xfs/xfs_bmap.h 2010-05-29 18:45:03.792276730 +0200
@@ -28,20 +28,6 @@ struct xfs_trans;
extern kmem_zone_t *xfs_bmap_free_item_zone;
/*
- * DELTA: describe a change to the in-core extent list.
- *
- * Internally the use of xed_blockount is somewhat funky.
- * xed_blockcount contains an offset much of the time because this
- * makes merging changes easier. (xfs_fileoff_t and xfs_filblks_t are
- * the same underlying type).
- */
-typedef struct xfs_extdelta
-{
- xfs_fileoff_t xed_startoff; /* offset of range */
- xfs_filblks_t xed_blockcount; /* blocks in range */
-} xfs_extdelta_t;
-
-/*
* List of extents to be free "later".
* The list is kept sorted on xbf_startblock.
*/
@@ -305,9 +291,7 @@ xfs_bmapi(
xfs_extlen_t total, /* total blocks needed */
struct xfs_bmbt_irec *mval, /* output: map values */
int *nmap, /* i/o: mval size/count */
- xfs_bmap_free_t *flist, /* i/o: list extents to free */
- xfs_extdelta_t *delta); /* o: change made to incore
- extents */
+ xfs_bmap_free_t *flist); /* i/o: list extents to free */
/*
* Map file blocks to filesystem blocks, simple version.
@@ -341,8 +325,6 @@ xfs_bunmapi(
xfs_fsblock_t *firstblock, /* first allocated block
controls a.g. for allocs */
xfs_bmap_free_t *flist, /* i/o: list extents to free */
- xfs_extdelta_t *delta, /* o: change made to incore
- extents */
int *done); /* set if not done yet */
/*
Index: xfs/fs/xfs/xfs_da_btree.c
===================================================================
--- xfs.orig/fs/xfs/xfs_da_btree.c 2010-05-29 18:44:44.247005742 +0200
+++ xfs/fs/xfs/xfs_da_btree.c 2010-05-29 18:45:03.800254799 +0200
@@ -1596,7 +1596,7 @@ xfs_da_grow_inode(xfs_da_args_t *args, x
xfs_bmapi_aflag(w)|XFS_BMAPI_WRITE|XFS_BMAPI_METADATA|
XFS_BMAPI_CONTIG,
args->firstblock, args->total, &map, &nmap,
- args->flist, NULL))) {
+ args->flist))) {
return error;
}
ASSERT(nmap <= 1);
@@ -1617,8 +1617,7 @@ xfs_da_grow_inode(xfs_da_args_t *args, x
xfs_bmapi_aflag(w)|XFS_BMAPI_WRITE|
XFS_BMAPI_METADATA,
args->firstblock, args->total,
- &mapp[mapi], &nmap, args->flist,
- NULL))) {
+ &mapp[mapi], &nmap, args->flist))) {
kmem_free(mapp);
return error;
}
@@ -1879,7 +1878,7 @@ xfs_da_shrink_inode(xfs_da_args_t *args,
*/
if ((error = xfs_bunmapi(tp, dp, dead_blkno, count,
xfs_bmapi_aflag(w)|XFS_BMAPI_METADATA,
- 0, args->firstblock, args->flist, NULL,
+ 0, args->firstblock, args->flist,
&done)) == ENOSPC) {
if (w != XFS_DATA_FORK)
break;
@@ -1984,7 +1983,7 @@ xfs_da_do_buf(
nfsb,
XFS_BMAPI_METADATA |
xfs_bmapi_aflag(whichfork),
- NULL, 0, mapp, &nmap, NULL, NULL)))
+ NULL, 0, mapp, &nmap, NULL)))
goto exit0;
}
} else {
Index: xfs/fs/xfs/xfs_dir2.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2.c 2010-05-29 18:44:44.250006161 +0200
+++ xfs/fs/xfs/xfs_dir2.c 2010-05-29 18:45:03.805255777 +0200
@@ -547,7 +547,7 @@ xfs_dir2_grow_inode(
if ((error = xfs_bmapi(tp, dp, bno, count,
XFS_BMAPI_WRITE|XFS_BMAPI_METADATA|XFS_BMAPI_CONTIG,
args->firstblock, args->total, &map, &nmap,
- args->flist, NULL)))
+ args->flist)))
return error;
ASSERT(nmap <= 1);
if (nmap == 1) {
@@ -579,8 +579,7 @@ xfs_dir2_grow_inode(
if ((error = xfs_bmapi(tp, dp, b, c,
XFS_BMAPI_WRITE|XFS_BMAPI_METADATA,
args->firstblock, args->total,
- &mapp[mapi], &nmap, args->flist,
- NULL))) {
+ &mapp[mapi], &nmap, args->flist))) {
kmem_free(mapp);
return error;
}
@@ -713,7 +712,7 @@ xfs_dir2_shrink_inode(
*/
if ((error = xfs_bunmapi(tp, dp, da, mp->m_dirblkfsbs,
XFS_BMAPI_METADATA, 0, args->firstblock, args->flist,
- NULL, &done))) {
+ &done))) {
/*
* ENOSPC actually can happen if we're in a removename with
* no space reservation, and the resulting block removal
Index: xfs/fs/xfs/xfs_dir2_leaf.c
===================================================================
--- xfs.orig/fs/xfs/xfs_dir2_leaf.c 2010-05-29 18:44:44.260005393 +0200
+++ xfs/fs/xfs/xfs_dir2_leaf.c 2010-05-29 18:45:03.809255847 +0200
@@ -873,7 +873,7 @@ xfs_dir2_leaf_getdents(
xfs_dir2_byte_to_da(mp,
XFS_DIR2_LEAF_OFFSET) - map_off,
XFS_BMAPI_METADATA, NULL, 0,
- &map[map_valid], &nmap, NULL, NULL);
+ &map[map_valid], &nmap, NULL);
/*
* Don't know if we should ignore this or
* try to return an error.
Index: xfs/fs/xfs/xfs_inode.c
===================================================================
--- xfs.orig/fs/xfs/xfs_inode.c 2010-05-29 18:45:02.178253752 +0200
+++ xfs/fs/xfs/xfs_inode.c 2010-05-29 18:45:03.814005812 +0200
@@ -1225,7 +1225,7 @@ xfs_isize_check(
(xfs_ufsize_t)XFS_MAXIOFFSET(mp)) -
map_first),
XFS_BMAPI_ENTIRE, NULL, 0, imaps, &nimaps,
- NULL, NULL))
+ NULL))
return;
ASSERT(nimaps == 1);
ASSERT(imaps[0].br_startblock == HOLESTARTBLOCK);
@@ -1591,7 +1591,7 @@ xfs_itruncate_finish(
xfs_bmapi_aflag(fork),
XFS_ITRUNC_MAX_EXTENTS,
&first_block, &free_list,
- NULL, &done);
+ &done);
if (error) {
/*
* If the bunmapi call encounters an error,
Index: xfs/fs/xfs/xfs_iomap.c
===================================================================
--- xfs.orig/fs/xfs/xfs_iomap.c 2010-05-29 18:44:47.721006022 +0200
+++ xfs/fs/xfs/xfs_iomap.c 2010-05-29 18:45:03.820005533 +0200
@@ -118,7 +118,7 @@ xfs_iomap(
error = xfs_bmapi(NULL, ip, offset_fsb,
(xfs_filblks_t)(end_fsb - offset_fsb),
bmapi_flags, NULL, 0, imap,
- nimaps, NULL, NULL);
+ nimaps, NULL);
if (error)
goto out;
@@ -341,7 +341,7 @@ xfs_iomap_write_direct(
xfs_bmap_init(&free_list, &firstfsb);
nimaps = 1;
error = xfs_bmapi(tp, ip, offset_fsb, count_fsb, bmapi_flag,
- &firstfsb, 0, &imap, &nimaps, &free_list, NULL);
+ &firstfsb, 0, &imap, &nimaps, &free_list);
if (error)
goto error0;
@@ -419,7 +419,7 @@ xfs_iomap_eof_want_preallocate(
imaps = nimaps;
firstblock = NULLFSBLOCK;
error = xfs_bmapi(NULL, ip, start_fsb, count_fsb, 0,
- &firstblock, 0, imap, &imaps, NULL, NULL);
+ &firstblock, 0, imap, &imaps, NULL);
if (error)
return error;
for (n = 0; n < imaps; n++) {
@@ -494,7 +494,7 @@ retry:
(xfs_filblks_t)(last_fsb - offset_fsb),
XFS_BMAPI_DELAY | XFS_BMAPI_WRITE |
XFS_BMAPI_ENTIRE, &firstblock, 1, imap,
- &nimaps, NULL, NULL);
+ &nimaps, NULL);
if (error && (error != ENOSPC))
return XFS_ERROR(error);
@@ -650,7 +650,7 @@ xfs_iomap_write_allocate(
/* Go get the actual blocks */
error = xfs_bmapi(tp, ip, map_start_fsb, count_fsb,
XFS_BMAPI_WRITE, &first_block, 1,
- &imap, &nimaps, &free_list, NULL);
+ &imap, &nimaps, &free_list);
if (error)
goto trans_cancel;
@@ -768,7 +768,7 @@ xfs_iomap_write_unwritten(
nimaps = 1;
error = xfs_bmapi(tp, ip, offset_fsb, count_fsb,
XFS_BMAPI_WRITE|XFS_BMAPI_CONVERT, &firstfsb,
- 1, &imap, &nimaps, &free_list, NULL);
+ 1, &imap, &nimaps, &free_list);
if (error)
goto error_on_bmapi_transaction;
Index: xfs/fs/xfs/xfs_rtalloc.c
===================================================================
--- xfs.orig/fs/xfs/xfs_rtalloc.c 2010-05-29 18:44:44.204005882 +0200
+++ xfs/fs/xfs/xfs_rtalloc.c 2010-05-29 18:45:03.826256825 +0200
@@ -122,7 +122,7 @@ xfs_growfs_rt_alloc(
cancelflags |= XFS_TRANS_ABORT;
error = xfs_bmapi(tp, ip, oblocks, nblocks - oblocks,
XFS_BMAPI_WRITE | XFS_BMAPI_METADATA, &firstblock,
- resblks, &map, &nmap, &flist, NULL);
+ resblks, &map, &nmap, &flist);
if (!error && nmap < 1)
error = XFS_ERROR(ENOSPC);
if (error)
Index: xfs/fs/xfs/xfs_vnodeops.c
===================================================================
--- xfs.orig/fs/xfs/xfs_vnodeops.c 2010-05-29 18:44:47.686004066 +0200
+++ xfs/fs/xfs/xfs_vnodeops.c 2010-05-29 18:45:03.835005393 +0200
@@ -492,7 +492,7 @@ xfs_readlink_bmap(
int error = 0;
error = xfs_bmapi(NULL, ip, 0, XFS_B_TO_FSB(mp, pathlen), 0, NULL, 0,
- mval, &nmaps, NULL, NULL);
+ mval, &nmaps, NULL);
if (error)
goto out;
@@ -596,7 +596,7 @@ xfs_free_eofblocks(
nimaps = 1;
xfs_ilock(ip, XFS_ILOCK_SHARED);
error = xfs_bmapi(NULL, ip, end_fsb, map_len, 0,
- NULL, 0, &imap, &nimaps, NULL, NULL);
+ NULL, 0, &imap, &nimaps, NULL);
xfs_iunlock(ip, XFS_ILOCK_SHARED);
if (!error && (nimaps != 0) &&
@@ -733,7 +733,7 @@ xfs_inactive_symlink_rmt(
nmaps = ARRAY_SIZE(mval);
if ((error = xfs_bmapi(tp, ip, 0, XFS_B_TO_FSB(mp, size),
XFS_BMAPI_METADATA, &first_block, 0, mval, &nmaps,
- &free_list, NULL)))
+ &free_list)))
goto error0;
/*
* Invalidate the block(s).
@@ -748,7 +748,7 @@ xfs_inactive_symlink_rmt(
* Unmap the dead block(s) to the free_list.
*/
if ((error = xfs_bunmapi(tp, ip, 0, size, XFS_BMAPI_METADATA, nmaps,
- &first_block, &free_list, NULL, &done)))
+ &first_block, &free_list, &done)))
goto error1;
ASSERT(done);
/*
@@ -2095,7 +2095,7 @@ xfs_symlink(
error = xfs_bmapi(tp, ip, first_fsb, fs_blocks,
XFS_BMAPI_WRITE | XFS_BMAPI_METADATA,
&first_block, resblks, mval, &nmaps,
- &free_list, NULL);
+ &free_list);
if (error) {
goto error1;
}
@@ -2347,7 +2347,7 @@ xfs_alloc_file_space(
error = xfs_bmapi(tp, ip, startoffset_fsb,
allocatesize_fsb, bmapi_flag,
&firstfsb, 0, imapp, &nimaps,
- &free_list, NULL);
+ &free_list);
if (error) {
goto error0;
}
@@ -2436,7 +2436,7 @@ xfs_zero_remaining_bytes(
offset_fsb = XFS_B_TO_FSBT(mp, offset);
nimap = 1;
error = xfs_bmapi(NULL, ip, offset_fsb, 1, 0,
- NULL, 0, &imap, &nimap, NULL, NULL);
+ NULL, 0, &imap, &nimap, NULL);
if (error || nimap < 1)
break;
ASSERT(imap.br_blockcount >= 1);
@@ -2556,7 +2556,7 @@ xfs_free_file_space(
if (rt && !xfs_sb_version_hasextflgbit(&mp->m_sb)) {
nimap = 1;
error = xfs_bmapi(NULL, ip, startoffset_fsb,
- 1, 0, NULL, 0, &imap, &nimap, NULL, NULL);
+ 1, 0, NULL, 0, &imap, &nimap, NULL);
if (error)
goto out_unlock_iolock;
ASSERT(nimap == 0 || nimap == 1);
@@ -2571,7 +2571,7 @@ xfs_free_file_space(
}
nimap = 1;
error = xfs_bmapi(NULL, ip, endoffset_fsb - 1,
- 1, 0, NULL, 0, &imap, &nimap, NULL, NULL);
+ 1, 0, NULL, 0, &imap, &nimap, NULL);
if (error)
goto out_unlock_iolock;
ASSERT(nimap == 0 || nimap == 1);
@@ -2647,7 +2647,7 @@ xfs_free_file_space(
xfs_bmap_init(&free_list, &firstfsb);
error = xfs_bunmapi(tp, ip, startoffset_fsb,
endoffset_fsb - startoffset_fsb,
- 0, 2, &firstfsb, &free_list, NULL, &done);
+ 0, 2, &firstfsb, &free_list, &done);
if (error) {
goto error0;
}
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 17/17] xfs: remove unused delta tracking code in xfs_bmapi
2010-05-31 16:07 ` [PATCH 17/17] xfs: remove unused delta tracking code in xfs_bmapi Christoph Hellwig
@ 2010-06-02 6:11 ` Dave Chinner
0 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 6:11 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:44PM -0400, Christoph Hellwig wrote:
> This code was introduced four years ago in commit
> 3e57ecf640428c01ba1ed8c8fc538447ada1715b without any review and has
> been unused since. Remove it just as the rest of the code introduced
> in that commit to reduce that stack usage and complexity in this central
> piece of code.
>
> Signed-off-by: Christoph Hellwig <hch@lst.de>
Agreed. if we have need for it in the future, it's there in
revision history.
Reviewed-by: Dave Chinner <dchinner@redhat.com>
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* Re: [PATCH 00/17] pending patches
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (16 preceding siblings ...)
2010-05-31 16:07 ` [PATCH 17/17] xfs: remove unused delta tracking code in xfs_bmapi Christoph Hellwig
@ 2010-06-02 6:13 ` Dave Chinner
2010-06-03 17:01 ` Alex Elder
18 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-02 6:13 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, May 31, 2010 at 12:07:27PM -0400, Christoph Hellwig wrote:
> Here's my list of pending patches. Ordering them is getting a bit
> difficult as various patches that have been out of a while haven't
> been commited yet. Note that some patches require Dave's
>
> [PATCH] xfs: fix race in inode cluster freeing failing to stale inodes
>
> applies first. That patch is a clear candidate for something that
> should go to Linus ASAP IHMO.
That's a fine set of cleanups, Christoph:
79 files changed, 974 insertions(+), 3106 deletions(-)
I'm still running it through QA. If it survives another day or so
I'll push them all out into my git tree for Alex....
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* Re: [PATCH 00/17] pending patches
2010-05-31 16:07 [PATCH 00/17] pending patches Christoph Hellwig
` (17 preceding siblings ...)
2010-06-02 6:13 ` [PATCH 00/17] pending patches Dave Chinner
@ 2010-06-03 17:01 ` Alex Elder
2010-06-03 17:08 ` dropping dmapi support, was " Christoph Hellwig
18 siblings, 1 reply; 43+ messages in thread
From: Alex Elder @ 2010-06-03 17:01 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs
On Mon, 2010-05-31 at 12:07 -0400, Christoph Hellwig wrote:
> Here's my list of pending patches. Ordering them is getting a bit
> difficult as various patches that have been out of a while haven't
> been commited yet. Note that some patches require Dave's
>
> [PATCH] xfs: fix race in inode cluster freeing failing to stale inodes
>
> applies first. That patch is a clear candidate for something that
> should go to Linus ASAP IHMO.
I agree it should go into 2.6.35. Last I saw on that one, Dave
said he was going to update it, so I was waiting for that. I
now see it's available on his kernel.org tree so I can grab
that and get it in the pipeline to send to Linus soon.
I mentioned to you privately a few weeks ago that I had some concerns
about removing DMAPI support, and had promised to send a note out to
the list to that effect, but regretfully I never did that.
I would like to have a chance to submit an alternative to simply
removing that code. I recognize it sits in the first part of your
patch series, and I will gladly do the work to rearrange them to
put it at the end, in order to give me some time to develop my
proposed change.
Basically what I'd like to do is update the DMAPI support code
so that it is much better isolated. I would like to replace
the big ugly hunks that lie in common code paths with small
function calls, so that their footprint is minimal and not
distracting (along the lines of tracing calls).
I got a start on doing this, and had hoped to send the result
pretty soon after your initial posting of the patch, but that
work unfortunately got preempted by other more pressing stuff.
I wanted to provide actual code to help make the discussion
of the merits of removal versus cleanup more concrete. I
now think I'll be able to put something together within the
next week or so.
I'll wait to hear back from you and others. In the mean time
I'll make my way through the rest of the patch series.
-Alex
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread* dropping dmapi support, was Re: [PATCH 00/17] pending patches
2010-06-03 17:01 ` Alex Elder
@ 2010-06-03 17:08 ` Christoph Hellwig
2010-06-03 22:33 ` Dave Chinner
0 siblings, 1 reply; 43+ messages in thread
From: Christoph Hellwig @ 2010-06-03 17:08 UTC (permalink / raw)
To: Alex Elder; +Cc: xfs
On Thu, Jun 03, 2010 at 12:01:46PM -0500, Alex Elder wrote:
> I would like to have a chance to submit an alternative to simply
> removing that code. I recognize it sits in the first part of your
> patch series, and I will gladly do the work to rearrange them to
> put it at the end, in order to give me some time to develop my
> proposed change.
>
> Basically what I'd like to do is update the DMAPI support code
> so that it is much better isolated. I would like to replace
> the big ugly hunks that lie in common code paths with small
> function calls, so that their footprint is minimal and not
> distracting (along the lines of tracing calls).
>
> I got a start on doing this, and had hoped to send the result
> pretty soon after your initial posting of the patch, but that
> work unfortunately got preempted by other more pressing stuff.
> I wanted to provide actual code to help make the discussion
> of the merits of removal versus cleanup more concrete. I
> now think I'll be able to put something together within the
> next week or so.
I don't think it's a good idea. I'm happy to not burn all bridges
and leave certain code structured in a way that makes adding it easier,
but if the hooks are as easy as you say above they can easily live in
an out of tree patchset. The general Linux kernel policy is that we
don't keep hooks for out of tree code around, and I tend to agree to
it. We kept all that dmapi cruft in, and it's never served any
purpose for us. I think that HSM support is actually a very useful
feature, but the a kernel interface based on the DMAPI specification
much less so, and the horrible SGI implementation that used to be
in the XFS CVS tree even less so.
If you want to push a new one the metadata hooks really need to be
entirely outside the low-level filesystem, that is before calling
into the namespace inode operations, which is easily doable even
while keeping the current DMAPI core.
But what's much more difficult is the read/write path. The dmapi
code really gets in the way there, and I have additional simplification
of this code pending that require this cruft to go away. XFS currently
has a needlessly complicated write path, and getting closer to the
generic code will help us with lots of things like the upcoming multi
page write support.
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread
* Re: dropping dmapi support, was Re: [PATCH 00/17] pending patches
2010-06-03 17:08 ` dropping dmapi support, was " Christoph Hellwig
@ 2010-06-03 22:33 ` Dave Chinner
0 siblings, 0 replies; 43+ messages in thread
From: Dave Chinner @ 2010-06-03 22:33 UTC (permalink / raw)
To: Christoph Hellwig; +Cc: xfs, Alex Elder
On Thu, Jun 03, 2010 at 01:08:18PM -0400, Christoph Hellwig wrote:
> On Thu, Jun 03, 2010 at 12:01:46PM -0500, Alex Elder wrote:
> > I would like to have a chance to submit an alternative to simply
> > removing that code. I recognize it sits in the first part of your
> > patch series, and I will gladly do the work to rearrange them to
> > put it at the end, in order to give me some time to develop my
> > proposed change.
> >
> > Basically what I'd like to do is update the DMAPI support code
> > so that it is much better isolated. I would like to replace
> > the big ugly hunks that lie in common code paths with small
> > function calls, so that their footprint is minimal and not
> > distracting (along the lines of tracing calls).
> >
> > I got a start on doing this, and had hoped to send the result
> > pretty soon after your initial posting of the patch, but that
> > work unfortunately got preempted by other more pressing stuff.
> > I wanted to provide actual code to help make the discussion
> > of the merits of removal versus cleanup more concrete. I
> > now think I'll be able to put something together within the
> > next week or so.
>
> I don't think it's a good idea. I'm happy to not burn all bridges
> and leave certain code structured in a way that makes adding it easier,
> but if the hooks are as easy as you say above they can easily live in
> an out of tree patchset. The general Linux kernel policy is that we
> don't keep hooks for out of tree code around, and I tend to agree to
> it. We kept all that dmapi cruft in, and it's never served any
> purpose for us. I think that HSM support is actually a very useful
> feature, but the a kernel interface based on the DMAPI specification
> much less so, and the horrible SGI implementation that used to be
> in the XFS CVS tree even less so.
>
> If you want to push a new one the metadata hooks really need to be
> entirely outside the low-level filesystem, that is before calling
> into the namespace inode operations, which is easily doable even
> while keeping the current DMAPI core.
Regardless of the implementation cruftiness, I think this a much
better approach. The events and checks really aren't XFS specific,
and putting them at a higher level cleanly separates the filesystem
functionality from the event+blocking functionality of DMAPI.
> But what's much more difficult is the read/write path. The dmapi
> code really gets in the way there, and I have additional simplification
> of this code pending that require this cruft to go away. XFS currently
> has a needlessly complicated write path, and getting closer to the
> generic code will help us with lots of things like the upcoming multi
> page write support.
That is true, and also intervening higher up in the IO path for
DMAPI would avoid a lot of the locking complexity that XFS has to go
through now to be able to block on events in dmapi calls.
Further, with ext4 gaining a persitent handle interface, adding
DMAPI to the VFS would also enable HSMs to work on more than just
XFS...
Cheers,
Dave.
--
Dave Chinner
david@fromorbit.com
_______________________________________________
xfs mailing list
xfs@oss.sgi.com
http://oss.sgi.com/mailman/listinfo/xfs
^ permalink raw reply [flat|nested] 43+ messages in thread