[PATCH] xfs: add tracepoints for each of the externally visible allocators

[PATCH] xfs: add tracepoints for each of the externally visible allocators

[Darrick J. Wong]

From: Darrick J. Wong <djwong@kernel.org>

There are now five separate space allocator interfaces exposed to the
rest of XFS for five different strategies to find space.  Add
tracepoints for each of them so that I can tell from a trace dump
exactly which ones got called and what happened underneath them.  Add a
sixth so it's more obvious if an allocation actually happened.

Signed-off-by: Darrick J. Wong <djwong@kernel.org>
---
 fs/xfs/libxfs/xfs_alloc.c |   17 +++++++++++++++++
 fs/xfs/xfs_trace.h        |    7 +++++++
 2 files changed, 24 insertions(+)

diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index bd7112d430b6..55ae08a6144c 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -3255,6 +3255,8 @@ xfs_alloc_vextent_finish(
 	XFS_STATS_INC(mp, xs_allocx);
 	XFS_STATS_ADD(mp, xs_allocb, args->len);
 
+	trace_xfs_alloc_vextent_finish(args);
+
 out_drop_perag:
 	if (drop_perag && args->pag) {
 		xfs_perag_rele(args->pag);
@@ -3284,6 +3286,9 @@ xfs_alloc_vextent_this_ag(
 
 	args->agno = agno;
 	args->agbno = 0;
+
+	trace_xfs_alloc_vextent_this_ag(args);
+
 	error = xfs_alloc_vextent_check_args(args, XFS_AGB_TO_FSB(mp, agno, 0),
 			&minimum_agno);
 	if (error) {
@@ -3405,6 +3410,9 @@ xfs_alloc_vextent_start_ag(
 
 	args->agno = NULLAGNUMBER;
 	args->agbno = NULLAGBLOCK;
+
+	trace_xfs_alloc_vextent_first_ag(args);
+
 	error = xfs_alloc_vextent_check_args(args, target, &minimum_agno);
 	if (error) {
 		if (error == -ENOSPC)
@@ -3455,6 +3463,9 @@ xfs_alloc_vextent_first_ag(
 
 	args->agno = NULLAGNUMBER;
 	args->agbno = NULLAGBLOCK;
+
+	trace_xfs_alloc_vextent_start_ag(args);
+
 	error = xfs_alloc_vextent_check_args(args, target, &minimum_agno);
 	if (error) {
 		if (error == -ENOSPC)
@@ -3486,6 +3497,9 @@ xfs_alloc_vextent_exact_bno(
 
 	args->agno = XFS_FSB_TO_AGNO(mp, target);
 	args->agbno = XFS_FSB_TO_AGBNO(mp, target);
+
+	trace_xfs_alloc_vextent_near_bno(args);
+
 	error = xfs_alloc_vextent_check_args(args, target, &minimum_agno);
 	if (error) {
 		if (error == -ENOSPC)
@@ -3521,6 +3535,9 @@ xfs_alloc_vextent_near_bno(
 
 	args->agno = XFS_FSB_TO_AGNO(mp, target);
 	args->agbno = XFS_FSB_TO_AGBNO(mp, target);
+
+	trace_xfs_alloc_vextent_exact_bno(args);
+
 	error = xfs_alloc_vextent_check_args(args, target, &minimum_agno);
 	if (error) {
 		if (error == -ENOSPC)
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index 7dc0fd6a6504..9c0006c55fec 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -1883,6 +1883,13 @@ DEFINE_ALLOC_EVENT(xfs_alloc_vextent_noagbp);
 DEFINE_ALLOC_EVENT(xfs_alloc_vextent_loopfailed);
 DEFINE_ALLOC_EVENT(xfs_alloc_vextent_allfailed);
 
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_this_ag);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_start_ag);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_first_ag);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_exact_bno);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_near_bno);
+DEFINE_ALLOC_EVENT(xfs_alloc_vextent_finish);
+
 TRACE_EVENT(xfs_alloc_cur_check,
 	TP_PROTO(struct xfs_mount *mp, xfs_btnum_t btnum, xfs_agblock_t bno,
 		 xfs_extlen_t len, xfs_extlen_t diff, bool new),

Re: [PATCH] xfs: add tracepoints for each of the externally visible allocators

[Dave Chinner]

On Thu, Mar 16, 2023 at 09:47:43AM -0700, Darrick J. Wong wrote:
> From: Darrick J. Wong <djwong@kernel.org>
> 
> There are now five separate space allocator interfaces exposed to the
> rest of XFS for five different strategies to find space.  Add
> tracepoints for each of them so that I can tell from a trace dump
> exactly which ones got called and what happened underneath them.  Add a
> sixth so it's more obvious if an allocation actually happened.
> 
> Signed-off-by: Darrick J. Wong <djwong@kernel.org>
> ---
>  fs/xfs/libxfs/xfs_alloc.c |   17 +++++++++++++++++
>  fs/xfs/xfs_trace.h        |    7 +++++++
>  2 files changed, 24 insertions(+)

Makes sense.

Reviewed-by: Dave Chinner <dchinner@redhat.com>

-- 
Dave Chinner
david@fromorbit.com