xfs: make xfs btree stats less huge

Embedding a switch statement in every btree stats inc/add adds a lot
of code overhead to the core btree infrastructure paths. Stats are
supposed to be small and lightweight, but the btree stats have
become big and bloated as we've added more btrees. It needs fixing
because the reflink code will just add more overhead again.

Convert the v2 btree stats to arrays instead of independent
variables, and instead use the type to index the specific btree
array via an enum. This allows us to use array based indexing
to update the stats, rather than having to derefence variables
specific to the btree type.

If we then wrap the xfsstats structure in a union and place uint32_t
array beside it, and calculate the correct btree stats array base
array index when creating a btree cursor,  we can easily access
entries in the stats structure without having to switch names based
on the btree type.

We then replace with the switch statement with a simple set of stats
wrapper macros, resulting in a significant simplification of the
btree stats code, and:

   text	   data	    bss	    dec	    hex	filename
  48905	    144	      8	  49057	   bfa1	fs/xfs/libxfs/xfs_btree.o.old
  36793	    144	      8	  36945	   9051	fs/xfs/libxfs/xfs_btree.o

it reduces the core btree infrastructure code size by close to 25%!

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Dave Chinner <david@fromorbit.com>

fs/xfs/libxfs/xfs_alloc_btree.c

@@ -428,6 +428,10 @@ xfs_allocbt_init_cursor(
 	cur->bc_btnum = btnum;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 	cur->bc_ops = &xfs_allocbt_ops;
+	if (btnum == XFS_BTNUM_BNO)
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtb_2);
+	else
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtc_2);
 
 	if (btnum == XFS_BTNUM_CNT) {
 		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);

fs/xfs/libxfs/xfs_bmap_btree.c

@@ -803,6 +803,7 @@ xfs_bmbt_init_cursor(
 	cur->bc_nlevels = be16_to_cpu(ifp->if_broot->bb_level) + 1;
 	cur->bc_btnum = XFS_BTNUM_BMAP;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_bmbt_2);
 
 	cur->bc_ops = &xfs_bmbt_ops;
 	cur->bc_flags = XFS_BTREE_LONG_PTRS | XFS_BTREE_ROOT_IN_INODE;

fs/xfs/libxfs/xfs_btree.h

@@ -96,46 +96,10 @@ union xfs_btree_rec {
 /*
  * Generic stats interface
  */
-#define __XFS_BTREE_STATS_INC(mp, type, stat) \
-	XFS_STATS_INC(mp, xs_ ## type ## _2_ ## stat)
 #define XFS_BTREE_STATS_INC(cur, stat)	\
-do {    \
-	struct xfs_mount *__mp = cur->bc_mp; \
-	switch (cur->bc_btnum) {  \
-	case XFS_BTNUM_BNO: __XFS_BTREE_STATS_INC(__mp, abtb, stat); break; \
-	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_INC(__mp, abtc, stat); break; \
-	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(__mp, bmbt, stat); break; \
-	case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(__mp, ibt, stat); break; \
-	case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(__mp, fibt, stat); break; \
-	case XFS_BTNUM_RMAP: __XFS_BTREE_STATS_INC(__mp, rmap, stat); break; \
-	case XFS_BTNUM_REFC: __XFS_BTREE_STATS_INC(__mp, refcbt, stat); break; \
-	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
-	}       \
-} while (0)
-
-#define __XFS_BTREE_STATS_ADD(mp, type, stat, val) \
-	XFS_STATS_ADD(mp, xs_ ## type ## _2_ ## stat, val)
-#define XFS_BTREE_STATS_ADD(cur, stat, val)  \
-do {    \
-	struct xfs_mount *__mp = cur->bc_mp; \
-	switch (cur->bc_btnum) {  \
-	case XFS_BTNUM_BNO:	\
-		__XFS_BTREE_STATS_ADD(__mp, abtb, stat, val); break; \
-	case XFS_BTNUM_CNT:	\
-		__XFS_BTREE_STATS_ADD(__mp, abtc, stat, val); break; \
-	case XFS_BTNUM_BMAP:	\
-		__XFS_BTREE_STATS_ADD(__mp, bmbt, stat, val); break; \
-	case XFS_BTNUM_INO:	\
-		__XFS_BTREE_STATS_ADD(__mp, ibt, stat, val); break; \
-	case XFS_BTNUM_FINO:	\
-		__XFS_BTREE_STATS_ADD(__mp, fibt, stat, val); break; \
-	case XFS_BTNUM_RMAP:	\
-		__XFS_BTREE_STATS_ADD(__mp, rmap, stat, val); break; \
-	case XFS_BTNUM_REFC:	\
-		__XFS_BTREE_STATS_ADD(__mp, refcbt, stat, val); break; \
-	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break; \
-	}       \
-} while (0)
+	XFS_STATS_INC_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat)
+#define XFS_BTREE_STATS_ADD(cur, stat, val)	\
+	XFS_STATS_ADD_OFF((cur)->bc_mp, (cur)->bc_statoff + __XBTS_ ## stat, val)
 
 #define	XFS_BTREE_MAXLEVELS	9	/* max of all btrees */
 
@@ -253,6 +217,7 @@ typedef struct xfs_btree_cur
 	__uint8_t	bc_nlevels;	/* number of levels in the tree */
 	__uint8_t	bc_blocklog;	/* log2(blocksize) of btree blocks */
 	xfs_btnum_t	bc_btnum;	/* identifies which btree type */
+	int		bc_statoff;	/* offset of btre stats array */
 	union {
 		struct {			/* needed for BNO, CNT, INO */
 			struct xfs_buf	*agbp;	/* agf/agi buffer pointer */

fs/xfs/libxfs/xfs_ialloc_btree.c

@@ -365,9 +365,11 @@ xfs_inobt_init_cursor(
 	if (btnum == XFS_BTNUM_INO) {
 		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
 		cur->bc_ops = &xfs_inobt_ops;
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2);
 	} else {
 		cur->bc_nlevels = be32_to_cpu(agi->agi_free_level);
 		cur->bc_ops = &xfs_finobt_ops;
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2);
 	}
 
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;

fs/xfs/libxfs/xfs_refcount_btree.c

@@ -354,6 +354,7 @@ xfs_refcountbt_init_cursor(
 	cur->bc_btnum = XFS_BTNUM_REFC;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 	cur->bc_ops = &xfs_refcountbt_ops;
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_refcbt_2);
 
 	cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
 

fs/xfs/libxfs/xfs_rmap_btree.c

@@ -484,6 +484,7 @@ xfs_rmapbt_init_cursor(
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 	cur->bc_ops = &xfs_rmapbt_ops;
 	cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]);
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2);
 
 	cur->bc_private.a.agbp = agbp;
 	cur->bc_private.a.agno = agno;

fs/xfs/xfs_stats.c

@@ -80,9 +80,9 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 	}
 	/* extra precision counters */
 	for_each_possible_cpu(i) {
-		xs_xstrat_bytes += per_cpu_ptr(stats, i)->xs_xstrat_bytes;
-		xs_write_bytes += per_cpu_ptr(stats, i)->xs_write_bytes;
-		xs_read_bytes += per_cpu_ptr(stats, i)->xs_read_bytes;
+		xs_xstrat_bytes += per_cpu_ptr(stats, i)->s.xs_xstrat_bytes;
+		xs_write_bytes += per_cpu_ptr(stats, i)->s.xs_write_bytes;
+		xs_read_bytes += per_cpu_ptr(stats, i)->s.xs_read_bytes;
 	}
 
 	len += snprintf(buf + len, PATH_MAX-len, "xpc %Lu %Lu %Lu\n",
@@ -106,9 +106,9 @@ void xfs_stats_clearall(struct xfsstats __percpu *stats)
 	for_each_possible_cpu(c) {
 		preempt_disable();
 		/* save vn_active, it's a universal truth! */
-		vn_active = per_cpu_ptr(stats, c)->vn_active;
+		vn_active = per_cpu_ptr(stats, c)->s.vn_active;
 		memset(per_cpu_ptr(stats, c), 0, sizeof(*stats));
-		per_cpu_ptr(stats, c)->vn_active = vn_active;
+		per_cpu_ptr(stats, c)->s.vn_active = vn_active;
 		preempt_enable();
 	}
 }

fs/xfs/xfs_stats.h

@@ -21,10 +21,38 @@
 
 #include <linux/percpu.h>
 
+/*
+ * The btree stats arrays have fixed offsets for the different stats. We
+ * store the base index in the btree cursor via XFS_STATS_CALC_INDEX() and
+ * that allows us to use fixed offsets into the stats array for each btree
+ * stat. These index offsets are defined in the order they will be emitted
+ * in the stats files, so it is possible to add new btree stat types by
+ * appending to the enum list below.
+ */
+enum {
+	__XBTS_lookup = 0,
+	__XBTS_compare = 1,
+	__XBTS_insrec = 2,
+	__XBTS_delrec = 3,
+	__XBTS_newroot = 4,
+	__XBTS_killroot = 5,
+	__XBTS_increment = 6,
+	__XBTS_decrement = 7,
+	__XBTS_lshift = 8,
+	__XBTS_rshift = 9,
+	__XBTS_split = 10,
+	__XBTS_join = 11,
+	__XBTS_alloc = 12,
+	__XBTS_free = 13,
+	__XBTS_moves = 14,
+
+	__XBTS_MAX = 15,
+};
+
 /*
  * XFS global statistics
  */
-struct xfsstats {
+struct __xfsstats {
 # define XFSSTAT_END_EXTENT_ALLOC	4
 	__uint32_t		xs_allocx;
 	__uint32_t		xs_allocb;
@@ -117,118 +145,20 @@ struct xfsstats {
 	__uint32_t		xb_page_found;
 	__uint32_t		xb_get_read;
 /* Version 2 btree counters */
-#define XFSSTAT_END_ABTB_V2		(XFSSTAT_END_BUF+15)
-	__uint32_t		xs_abtb_2_lookup;
-	__uint32_t		xs_abtb_2_compare;
-	__uint32_t		xs_abtb_2_insrec;
-	__uint32_t		xs_abtb_2_delrec;
-	__uint32_t		xs_abtb_2_newroot;
-	__uint32_t		xs_abtb_2_killroot;
-	__uint32_t		xs_abtb_2_increment;
-	__uint32_t		xs_abtb_2_decrement;
-	__uint32_t		xs_abtb_2_lshift;
-	__uint32_t		xs_abtb_2_rshift;
-	__uint32_t		xs_abtb_2_split;
-	__uint32_t		xs_abtb_2_join;
-	__uint32_t		xs_abtb_2_alloc;
-	__uint32_t		xs_abtb_2_free;
-	__uint32_t		xs_abtb_2_moves;
-#define XFSSTAT_END_ABTC_V2		(XFSSTAT_END_ABTB_V2+15)
-	__uint32_t		xs_abtc_2_lookup;
-	__uint32_t		xs_abtc_2_compare;
-	__uint32_t		xs_abtc_2_insrec;
-	__uint32_t		xs_abtc_2_delrec;
-	__uint32_t		xs_abtc_2_newroot;
-	__uint32_t		xs_abtc_2_killroot;
-	__uint32_t		xs_abtc_2_increment;
-	__uint32_t		xs_abtc_2_decrement;
-	__uint32_t		xs_abtc_2_lshift;
-	__uint32_t		xs_abtc_2_rshift;
-	__uint32_t		xs_abtc_2_split;
-	__uint32_t		xs_abtc_2_join;
-	__uint32_t		xs_abtc_2_alloc;
-	__uint32_t		xs_abtc_2_free;
-	__uint32_t		xs_abtc_2_moves;
-#define XFSSTAT_END_BMBT_V2		(XFSSTAT_END_ABTC_V2+15)
-	__uint32_t		xs_bmbt_2_lookup;
-	__uint32_t		xs_bmbt_2_compare;
-	__uint32_t		xs_bmbt_2_insrec;
-	__uint32_t		xs_bmbt_2_delrec;
-	__uint32_t		xs_bmbt_2_newroot;
-	__uint32_t		xs_bmbt_2_killroot;
-	__uint32_t		xs_bmbt_2_increment;
-	__uint32_t		xs_bmbt_2_decrement;
-	__uint32_t		xs_bmbt_2_lshift;
-	__uint32_t		xs_bmbt_2_rshift;
-	__uint32_t		xs_bmbt_2_split;
-	__uint32_t		xs_bmbt_2_join;
-	__uint32_t		xs_bmbt_2_alloc;
-	__uint32_t		xs_bmbt_2_free;
-	__uint32_t		xs_bmbt_2_moves;
-#define XFSSTAT_END_IBT_V2		(XFSSTAT_END_BMBT_V2+15)
-	__uint32_t		xs_ibt_2_lookup;
-	__uint32_t		xs_ibt_2_compare;
-	__uint32_t		xs_ibt_2_insrec;
-	__uint32_t		xs_ibt_2_delrec;
-	__uint32_t		xs_ibt_2_newroot;
-	__uint32_t		xs_ibt_2_killroot;
-	__uint32_t		xs_ibt_2_increment;
-	__uint32_t		xs_ibt_2_decrement;
-	__uint32_t		xs_ibt_2_lshift;
-	__uint32_t		xs_ibt_2_rshift;
-	__uint32_t		xs_ibt_2_split;
-	__uint32_t		xs_ibt_2_join;
-	__uint32_t		xs_ibt_2_alloc;
-	__uint32_t		xs_ibt_2_free;
-	__uint32_t		xs_ibt_2_moves;
-#define XFSSTAT_END_FIBT_V2		(XFSSTAT_END_IBT_V2+15)
-	__uint32_t		xs_fibt_2_lookup;
-	__uint32_t		xs_fibt_2_compare;
-	__uint32_t		xs_fibt_2_insrec;
-	__uint32_t		xs_fibt_2_delrec;
-	__uint32_t		xs_fibt_2_newroot;
-	__uint32_t		xs_fibt_2_killroot;
-	__uint32_t		xs_fibt_2_increment;
-	__uint32_t		xs_fibt_2_decrement;
-	__uint32_t		xs_fibt_2_lshift;
-	__uint32_t		xs_fibt_2_rshift;
-	__uint32_t		xs_fibt_2_split;
-	__uint32_t		xs_fibt_2_join;
-	__uint32_t		xs_fibt_2_alloc;
-	__uint32_t		xs_fibt_2_free;
-	__uint32_t		xs_fibt_2_moves;
-#define XFSSTAT_END_RMAP_V2		(XFSSTAT_END_FIBT_V2+15)
-	__uint32_t		xs_rmap_2_lookup;
-	__uint32_t		xs_rmap_2_compare;
-	__uint32_t		xs_rmap_2_insrec;
-	__uint32_t		xs_rmap_2_delrec;
-	__uint32_t		xs_rmap_2_newroot;
-	__uint32_t		xs_rmap_2_killroot;
-	__uint32_t		xs_rmap_2_increment;
-	__uint32_t		xs_rmap_2_decrement;
-	__uint32_t		xs_rmap_2_lshift;
-	__uint32_t		xs_rmap_2_rshift;
-	__uint32_t		xs_rmap_2_split;
-	__uint32_t		xs_rmap_2_join;
-	__uint32_t		xs_rmap_2_alloc;
-	__uint32_t		xs_rmap_2_free;
-	__uint32_t		xs_rmap_2_moves;
-#define XFSSTAT_END_REFCOUNT		(XFSSTAT_END_RMAP_V2 + 15)
-	__uint32_t		xs_refcbt_2_lookup;
-	__uint32_t		xs_refcbt_2_compare;
-	__uint32_t		xs_refcbt_2_insrec;
-	__uint32_t		xs_refcbt_2_delrec;
-	__uint32_t		xs_refcbt_2_newroot;
-	__uint32_t		xs_refcbt_2_killroot;
-	__uint32_t		xs_refcbt_2_increment;
-	__uint32_t		xs_refcbt_2_decrement;
-	__uint32_t		xs_refcbt_2_lshift;
-	__uint32_t		xs_refcbt_2_rshift;
-	__uint32_t		xs_refcbt_2_split;
-	__uint32_t		xs_refcbt_2_join;
-	__uint32_t		xs_refcbt_2_alloc;
-	__uint32_t		xs_refcbt_2_free;
-	__uint32_t		xs_refcbt_2_moves;
+#define XFSSTAT_END_ABTB_V2		(XFSSTAT_END_BUF + __XBTS_MAX)
+	__uint32_t		xs_abtb_2[__XBTS_MAX];
+#define XFSSTAT_END_ABTC_V2		(XFSSTAT_END_ABTB_V2 + __XBTS_MAX)
+	__uint32_t		xs_abtc_2[__XBTS_MAX];
+#define XFSSTAT_END_BMBT_V2		(XFSSTAT_END_ABTC_V2 + __XBTS_MAX)
+	__uint32_t		xs_bmbt_2[__XBTS_MAX];
+#define XFSSTAT_END_IBT_V2		(XFSSTAT_END_BMBT_V2 + __XBTS_MAX)
+	__uint32_t		xs_ibt_2[__XBTS_MAX];
+#define XFSSTAT_END_FIBT_V2		(XFSSTAT_END_IBT_V2 + __XBTS_MAX)
+	__uint32_t		xs_fibt_2[__XBTS_MAX];
+#define XFSSTAT_END_RMAP_V2		(XFSSTAT_END_FIBT_V2 + __XBTS_MAX)
+	__uint32_t		xs_rmap_2[__XBTS_MAX];
+#define XFSSTAT_END_REFCOUNT		(XFSSTAT_END_RMAP_V2 + __XBTS_MAX)
+	__uint32_t		xs_refcbt_2[__XBTS_MAX];
 #define XFSSTAT_END_XQMSTAT		(XFSSTAT_END_REFCOUNT + 6)
 	__uint32_t		xs_qm_dqreclaims;
 	__uint32_t		xs_qm_dqreclaim_misses;
@@ -245,26 +175,58 @@ struct xfsstats {
 	__uint64_t		xs_read_bytes;
 };
 
+struct xfsstats {
+	union {
+		struct __xfsstats	s;
+		uint32_t		a[XFSSTAT_END_XQMSTAT];
+	};
+};
+
+/*
+ * simple wrapper for getting the array index of s struct member offset
+ */
+#define XFS_STATS_CALC_INDEX(member)	\
+	(offsetof(struct __xfsstats, member) / (int)sizeof(__uint32_t))
+
+
 int xfs_stats_format(struct xfsstats __percpu *stats, char *buf);
 void xfs_stats_clearall(struct xfsstats __percpu *stats);
 extern struct xstats xfsstats;
 
 #define XFS_STATS_INC(mp, v)					\
 do {								\
-	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v++;	\
-	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v++;	\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v++;	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v++;	\
 } while (0)
 
 #define XFS_STATS_DEC(mp, v)					\
 do {								\
-	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v--;	\
-	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v--;	\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v--;	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v--;	\
 } while (0)
 
 #define XFS_STATS_ADD(mp, v, inc)					\
 do {									\
-	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->v += (inc);	\
-	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->v += (inc);	\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->s.v += (inc);	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->s.v += (inc);	\
+} while (0)
+
+#define XFS_STATS_INC_OFF(mp, off)				\
+do {								\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off]++;	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off]++;	\
+} while (0)
+
+#define XFS_STATS_DEC_OFF(mp, off)					\
+do {								\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off];	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off];	\
+} while (0)
+
+#define XFS_STATS_ADD_OFF(mp, off, inc)					\
+do {									\
+	per_cpu_ptr(xfsstats.xs_stats, current_cpu())->a[off] += (inc);	\
+	per_cpu_ptr(mp->m_stats.xs_stats, current_cpu())->a[off] += (inc);	\
 } while (0)
 
 #if defined(CONFIG_PROC_FS)