aboutsummaryrefslogtreecommitdiffstats
path: root/arch/sparc/lib
diff options
context:
space:
mode:
authorBabu Moger <babu.moger@oracle.com>2017-03-17 14:52:21 -0600
committerDavid S. Miller <davem@davemloft.net>2017-03-27 21:51:40 -0700
commit0ae2d26ffe70c32d4a7fe77593f0a55ce416c09e (patch)
tree56fd33ac22298ef2ae162366727ada11c719323e /arch/sparc/lib
parentsparc64: kern_addr_valid regression (diff)
downloadlinux-dev-0ae2d26ffe70c32d4a7fe77593f0a55ce416c09e.tar.xz
linux-dev-0ae2d26ffe70c32d4a7fe77593f0a55ce416c09e.zip
arch/sparc: Avoid DCTI Couples
Avoid un-intended DCTI Couples. Use of DCTI couples is deprecated. Also address the "Programming Note" for optimal performance. Here is the complete text from Oracle SPARC Architecture Specs. 6.3.4.7 DCTI Couples "A delayed control transfer instruction (DCTI) in the delay slot of another DCTI is referred to as a “DCTI couple”. The use of DCTI couples is deprecated in the Oracle SPARC Architecture; no new software should place a DCTI in the delay slot of another DCTI, because on future Oracle SPARC Architecture implementations DCTI couples may execute either slowly or differently than the programmer assumes it will. SPARC V8 and SPARC V9 Compatibility Note The SPARC V8 architecture left behavior undefined for a DCTI couple. The SPARC V9 architecture defined behavior in that case, but as of UltraSPARC Architecture 2005, use of DCTI couples was deprecated. Software should not expect high performance from DCTI couples, and performance of DCTI couples should be expected to decline further in future processors. Programming Note As noted in TABLE 6-5 on page 115, an annulled branch-always (branch-always with a = 1) instruction is not architecturally a DCTI. However, since not all implementations make that distinction, for optimal performance, a DCTI should not be placed in the instruction word immediately following an annulled branch-always instruction (BA,A or BPA,A)." Signed-off-by: Babu Moger <babu.moger@oracle.com> Reviewed-by: Rob Gardner <rob.gardner@oracle.com> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'arch/sparc/lib')
-rw-r--r--arch/sparc/lib/NG2memcpy.S4
-rw-r--r--arch/sparc/lib/NG4memcpy.S1
-rw-r--r--arch/sparc/lib/NG4memset.S1
-rw-r--r--arch/sparc/lib/NGmemcpy.S1
4 files changed, 7 insertions, 0 deletions
diff --git a/arch/sparc/lib/NG2memcpy.S b/arch/sparc/lib/NG2memcpy.S
index c629dbd121b6..64dcd6cdb606 100644
--- a/arch/sparc/lib/NG2memcpy.S
+++ b/arch/sparc/lib/NG2memcpy.S
@@ -326,11 +326,13 @@ FUNC_NAME: /* %o0=dst, %o1=src, %o2=len */
blu 170f
nop
ba,a,pt %xcc, 180f
+ nop
4: /* 32 <= low bits < 48 */
blu 150f
nop
ba,a,pt %xcc, 160f
+ nop
5: /* 0 < low bits < 32 */
blu,a 6f
cmp %g2, 8
@@ -338,6 +340,7 @@ FUNC_NAME: /* %o0=dst, %o1=src, %o2=len */
blu 130f
nop
ba,a,pt %xcc, 140f
+ nop
6: /* 0 < low bits < 16 */
bgeu 120f
nop
@@ -475,6 +478,7 @@ FUNC_NAME: /* %o0=dst, %o1=src, %o2=len */
brz,pt %o2, 85f
sub %o0, %o1, GLOBAL_SPARE
ba,a,pt %XCC, 90f
+ nop
.align 64
75: /* 16 < len <= 64 */
diff --git a/arch/sparc/lib/NG4memcpy.S b/arch/sparc/lib/NG4memcpy.S
index 75bb93b1437f..78ea962edcbe 100644
--- a/arch/sparc/lib/NG4memcpy.S
+++ b/arch/sparc/lib/NG4memcpy.S
@@ -530,4 +530,5 @@ FUNC_NAME: /* %o0=dst, %o1=src, %o2=len */
bne,pt %icc, 1b
EX_ST(STORE(stb, %g1, %o0 - 0x01), NG4_retl_o2_plus_1)
ba,a,pt %icc, .Lexit
+ nop
.size FUNC_NAME, .-FUNC_NAME
diff --git a/arch/sparc/lib/NG4memset.S b/arch/sparc/lib/NG4memset.S
index 41da4bdd95cb..7c0c81f18837 100644
--- a/arch/sparc/lib/NG4memset.S
+++ b/arch/sparc/lib/NG4memset.S
@@ -102,4 +102,5 @@ NG4bzero:
bne,pt %icc, 1b
add %o0, 0x30, %o0
ba,a,pt %icc, .Lpostloop
+ nop
.size NG4bzero,.-NG4bzero
diff --git a/arch/sparc/lib/NGmemcpy.S b/arch/sparc/lib/NGmemcpy.S
index d88c4ed50a00..cd654a719b27 100644
--- a/arch/sparc/lib/NGmemcpy.S
+++ b/arch/sparc/lib/NGmemcpy.S
@@ -394,6 +394,7 @@ FUNC_NAME: /* %i0=dst, %i1=src, %i2=len */
brz,pt %i2, 85f
sub %o0, %i1, %i3
ba,a,pt %XCC, 90f
+ nop
.align 64
70: /* 16 < len <= 64 */