summaryrefslogtreecommitdiff
path: root/arch/arm/vfp
diff options
context:
space:
mode:
authorRussell King <rmk+kernel@arm.linux.org.uk>2014-06-30 19:29:12 +0400
committerRussell King <rmk+kernel@arm.linux.org.uk>2014-07-18 15:29:04 +0400
commit6ebbf2ce437b33022d30badd49dc94d33ecfa498 (patch)
treebc015e35b456a28bb0e501803a454dc0c0d3291a /arch/arm/vfp
parentaf040ffc9ba1e079ee4c0748aff64fa3d4716fa5 (diff)
downloadlinux-6ebbf2ce437b33022d30badd49dc94d33ecfa498.tar.xz
ARM: convert all "mov.* pc, reg" to "bx reg" for ARMv6+
ARMv6 and greater introduced a new instruction ("bx") which can be used to return from function calls. Recent CPUs perform better when the "bx lr" instruction is used rather than the "mov pc, lr" instruction, and this sequence is strongly recommended to be used by the ARM architecture manual (section A.4.1.1). We provide a new macro "ret" with all its variants for the condition code which will resolve to the appropriate instruction. Rather than doing this piecemeal, and miss some instances, change all the "mov pc" instances to use the new macro, with the exception of the "movs" instruction and the kprobes code. This allows us to detect the "mov pc, lr" case and fix it up - and also gives us the possibility of deploying this for other registers depending on the CPU selection. Reported-by: Will Deacon <will.deacon@arm.com> Tested-by: Stephen Warren <swarren@nvidia.com> # Tegra Jetson TK1 Tested-by: Robert Jarzmik <robert.jarzmik@free.fr> # mioa701_bootresume.S Tested-by: Andrew Lunn <andrew@lunn.ch> # Kirkwood Tested-by: Shawn Guo <shawn.guo@freescale.com> Tested-by: Tony Lindgren <tony@atomide.com> # OMAPs Tested-by: Gregory CLEMENT <gregory.clement@free-electrons.com> # Armada XP, 375, 385 Acked-by: Sekhar Nori <nsekhar@ti.com> # DaVinci Acked-by: Christoffer Dall <christoffer.dall@linaro.org> # kvm/hyp Acked-by: Haojian Zhuang <haojian.zhuang@gmail.com> # PXA3xx Acked-by: Stefano Stabellini <stefano.stabellini@eu.citrix.com> # Xen Tested-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de> # ARMv7M Tested-by: Simon Horman <horms+renesas@verge.net.au> # Shmobile Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Diffstat (limited to 'arch/arm/vfp')
-rw-r--r--arch/arm/vfp/entry.S4
-rw-r--r--arch/arm/vfp/vfphw.S26
2 files changed, 15 insertions, 15 deletions
diff --git a/arch/arm/vfp/entry.S b/arch/arm/vfp/entry.S
index fe6ca574d093..2e78760f3495 100644
--- a/arch/arm/vfp/entry.S
+++ b/arch/arm/vfp/entry.S
@@ -34,7 +34,7 @@ ENDPROC(do_vfp)
ENTRY(vfp_null_entry)
dec_preempt_count_ti r10, r4
- mov pc, lr
+ ret lr
ENDPROC(vfp_null_entry)
.align 2
@@ -49,7 +49,7 @@ ENTRY(vfp_testing_entry)
dec_preempt_count_ti r10, r4
ldr r0, VFP_arch_address
str r0, [r0] @ set to non-zero value
- mov pc, r9 @ we have handled the fault
+ ret r9 @ we have handled the fault
ENDPROC(vfp_testing_entry)
.align 2
diff --git a/arch/arm/vfp/vfphw.S b/arch/arm/vfp/vfphw.S
index be807625ed8c..cda654cbf2c2 100644
--- a/arch/arm/vfp/vfphw.S
+++ b/arch/arm/vfp/vfphw.S
@@ -183,7 +183,7 @@ vfp_hw_state_valid:
@ always subtract 4 from the following
@ instruction address.
dec_preempt_count_ti r10, r4
- mov pc, r9 @ we think we have handled things
+ ret r9 @ we think we have handled things
look_for_VFP_exceptions:
@@ -202,7 +202,7 @@ look_for_VFP_exceptions:
DBGSTR "not VFP"
dec_preempt_count_ti r10, r4
- mov pc, lr
+ ret lr
process_exception:
DBGSTR "bounce"
@@ -234,7 +234,7 @@ ENTRY(vfp_save_state)
VFPFMRX r12, FPINST2 @ FPINST2 if needed (and present)
1:
stmia r0, {r1, r2, r3, r12} @ save FPEXC, FPSCR, FPINST, FPINST2
- mov pc, lr
+ ret lr
ENDPROC(vfp_save_state)
.align
@@ -245,7 +245,7 @@ vfp_current_hw_state_address:
#ifdef CONFIG_THUMB2_KERNEL
adr \tmp, 1f
add \tmp, \tmp, \base, lsl \shift
- mov pc, \tmp
+ ret \tmp
#else
add pc, pc, \base, lsl \shift
mov r0, r0
@@ -257,10 +257,10 @@ ENTRY(vfp_get_float)
tbl_branch r0, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mrc p10, 0, r0, c\dr, c0, 0 @ fmrs r0, s0
- mov pc, lr
+ ret lr
.org 1b + 8
1: mrc p10, 0, r0, c\dr, c0, 4 @ fmrs r0, s1
- mov pc, lr
+ ret lr
.org 1b + 8
.endr
ENDPROC(vfp_get_float)
@@ -269,10 +269,10 @@ ENTRY(vfp_put_float)
tbl_branch r1, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mcr p10, 0, r0, c\dr, c0, 0 @ fmsr r0, s0
- mov pc, lr
+ ret lr
.org 1b + 8
1: mcr p10, 0, r0, c\dr, c0, 4 @ fmsr r0, s1
- mov pc, lr
+ ret lr
.org 1b + 8
.endr
ENDPROC(vfp_put_float)
@@ -281,14 +281,14 @@ ENTRY(vfp_get_double)
tbl_branch r0, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: fmrrd r0, r1, d\dr
- mov pc, lr
+ ret lr
.org 1b + 8
.endr
#ifdef CONFIG_VFPv3
@ d16 - d31 registers
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mrrc p11, 3, r0, r1, c\dr @ fmrrd r0, r1, d\dr
- mov pc, lr
+ ret lr
.org 1b + 8
.endr
#endif
@@ -296,21 +296,21 @@ ENTRY(vfp_get_double)
@ virtual register 16 (or 32 if VFPv3) for compare with zero
mov r0, #0
mov r1, #0
- mov pc, lr
+ ret lr
ENDPROC(vfp_get_double)
ENTRY(vfp_put_double)
tbl_branch r2, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: fmdrr d\dr, r0, r1
- mov pc, lr
+ ret lr
.org 1b + 8
.endr
#ifdef CONFIG_VFPv3
@ d16 - d31 registers
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mcrr p11, 3, r0, r1, c\dr @ fmdrr r0, r1, d\dr
- mov pc, lr
+ ret lr
.org 1b + 8
.endr
#endif