some armv7 decode helpers are trivial functions but didn't inline

both from_u8 and the build function here compiled to truly trivial code:
four instructions (mov rdi, rax; cmp 0xlim, rax; jae panic; ret) in the
hot path, and constrained register choice on the caller side. inlining
these makes for a *smaller* armv7 decoder, on the order of 5kb down from
5.5kb. in the process it also gets about 45% faster (400mb/s to 560mb/s)

inlining decode_into, then, really just helps the standalone decoder
benchmark case. this moves decode throughput from 560mb/s to 724mb/s.

1 files changed, 4 insertions, 0 deletions

diff --git a/src/armv7.rs b/src/armv7.rs
index e025267..b8f33d3 100644
--- a/src/armv7.rs
+++ b/src/armv7.rs
@@ -1291,6 +1291,7 @@ impl Reg {
     /// create a new `Reg` with the specified number.
     ///
     /// panics if `bits` is out of range (16 or above).
+    #[inline]
     pub fn from_u8(bits: u8) -> Reg {
         if bits > 0b1111 {
             panic!("register number out of range");
@@ -1322,6 +1323,7 @@ impl CReg {
     /// create a new `CReg` with the specified number.
     ///
     /// panics if `bits` is out of range (16 or above).
+    #[inline]
     pub fn from_u8(bits: u8) -> CReg {
         if bits > 0b1111 {
             panic!("register number out of range");
@@ -1936,6 +1938,7 @@ impl Display for ConditionCode {
 }
 
 impl ConditionCode {
+    #[inline]
     fn build(value: u8) -> ConditionCode {
         match value {
             0b0000 => ConditionCode::EQ,
@@ -2204,6 +2207,7 @@ impl InstDecoder {
 
 #[allow(non_snake_case)]
 impl Decoder<ARMv7> for InstDecoder {
+    #[inline]
     fn decode_into<T: Reader<<ARMv7 as Arch>::Address, <ARMv7 as Arch>::Word>>(&self, inst: &mut Instruction, words: &mut T) -> Result<(), <ARMv7 as Arch>::DecodeError> {
         inst.set_w(false);
         inst.set_wide(false);