Reduce register pressure in substr()

Have substr() return the result string pointer. This means that we can delay allocating the register to hold the result until after the call to dt_substr() has been made. Signed-off-by: Kris Van Hees <kris.van.hees@oracle.com> Reviewed-by: Eugene Loh <eugene.loh@oracle.com>
oracle · Mar 18, 2022 · 6bd24a5 · 6bd24a5
1 parent 8d80e9a
commit 6bd24a5
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 14 deletions.
diff --git a/bpf/substr.S b/bpf/substr.S
@@ -6,7 +6,7 @@
 #define BPF_FUNC_probe_read_str	45
 
 /*
- * void dt_substr(char *dst, const char *src, int32_t idx, int32_t cnt,
+ * char *dt_substr(char *dst, const char *src, int32_t idx, int32_t cnt,
  *		  uint64_t argc)
  *
  * %r1 = dst, %r2 = src, %r3 = idx, %r4 = cnt, %r5 = argc
@@ -124,10 +124,12 @@ dt_substr :
 					 *		       &src[idx]);
 					 */
 
+	mov	%r0, %r9
 	exit
 
 .Lempty:
 	/* Store the empty string in the destination. */
 	stb	[%r9+0], 0
+	mov	%r0, %r9
 	exit
 	.size	dt_substr, .-dt_substr
diff --git a/libdtrace/dt_cg.c b/libdtrace/dt_cg.c
@@ -3997,26 +3997,20 @@ dt_cg_subr_substr(dt_node_t *dnp, dt_irlist_t *dlp, dt_regset_t *drp)
 		dt_cg_node(cnt, dlp, drp);
 
 	/*
-	 * Allocate the result register and associate it with a temporary
-	 * string slot.
+	 * Allocate a temporary string slot for the result.
 	 */
-	dnp->dn_reg = dt_regset_alloc(drp);
-	if (dnp->dn_reg == -1)
-		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
 	dt_cg_tstring_alloc(yypcb, dnp);
 
-        emit(dlp,  BPF_LOAD(BPF_DW, dnp->dn_reg, BPF_REG_FP, DT_STK_DCTX));
-        emit(dlp,  BPF_LOAD(BPF_DW, dnp->dn_reg, dnp->dn_reg, DCTX_MEM));
-        emit(dlp,  BPF_ALU64_IMM(BPF_ADD, dnp->dn_reg, dnp->dn_tstring->dn_value));
-
 	if (dt_regset_xalloc_args(drp) == -1)
 		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
 
-	emit(dlp, BPF_MOV_REG(BPF_REG_1, dnp->dn_reg));
-	emit(dlp, BPF_MOV_REG(BPF_REG_2, str->dn_reg));
+	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_1, BPF_REG_FP, DT_STK_DCTX));
+	emit(dlp,  BPF_LOAD(BPF_DW, BPF_REG_1, BPF_REG_1, DCTX_MEM));
+	emit(dlp,  BPF_ALU64_IMM(BPF_ADD, BPF_REG_1, dnp->dn_tstring->dn_value));
+	emit(dlp,  BPF_MOV_REG(BPF_REG_2, str->dn_reg));
 	dt_regset_free(drp, str->dn_reg);
 	dt_cg_tstring_free(yypcb, str);
-	emit(dlp, BPF_MOV_REG(BPF_REG_3, idx->dn_reg));
+	emit(dlp,  BPF_MOV_REG(BPF_REG_3, idx->dn_reg));
 	dt_regset_free(drp, idx->dn_reg);
 	if (cnt != NULL) {
 		emit(dlp, BPF_MOV_REG(BPF_REG_4, cnt->dn_reg));
@@ -4030,8 +4024,17 @@ dt_cg_subr_substr(dt_node_t *dnp, dt_irlist_t *dlp, dt_regset_t *drp)
 	idp = dt_dlib_get_func(yypcb->pcb_hdl, "dt_substr");
 	assert(idp != NULL);
 	emite(dlp,  BPF_CALL_FUNC(idp->di_id), idp);
-	dt_regset_free_args(drp);
+
+	/*
+	 * Allocate the result register, and assign the result to it..
+	 */
+	dnp->dn_reg = dt_regset_alloc(drp);
+	if (dnp->dn_reg == -1)
+		longjmp(yypcb->pcb_jmpbuf, EDT_NOREG);
+
+	emit(dlp, BPF_MOV_REG(dnp->dn_reg, BPF_REG_0));
 	dt_regset_free(drp, BPF_REG_0);
+	dt_regset_free_args(drp);
 
 	TRACE_REGSET("    subr-substr:End  ");
 }