repo.or.cz
/
official-gcc.git
/
search
commit
grep
author
committer
pickaxe
?
search:
re
summary
|
log
|
graphiclog1
|
graphiclog2
|
commit
|
commitdiff
|
tree
|
refs
|
edit
|
fork
first
·
prev
·
next
[AArch64] Use gen_frame_mem for callee-saves
2017-08-07
wil
c
o
[A
A
rch64]
U
s
e gen_fr
a
me_mem for
callee-saves
commit
|
commitdiff
|
tree
2017-08-07
wilco
Fix unr
e
s
o
lved
in g
c
c
.
dg/pr46932
.
c
commit
|
commitdiff
|
tree
2017-07-26
wilco
Disabl
e
pr790
4
1-2
.
c with -
m
abi=i
l
p32
.
commit
|
commitdiff
|
tree
2017-07-26
w
ilc
o
Fix PR46932:
B
lock auto increment
on frame
poi
n
ter
commit
|
commitdiff
|
tree
2017-07-25
w
i
lco
Fi
x
PR7904
1
commit
|
commitdiff
|
tree
2017-07-24
wilc
o
Fix PR79
0
41
commit
|
commitdiff
|
tree
2017-06-28
wilco
T
h
is patch fi
x
es a f
a
ilure in
g
c
c
.
target/aarch64/reload
.
.
.
commit
|
commitdiff
|
tree
2017-06-28
wi
l
co
Im
p
rove Cortex-
A
53 shift bypass
commit
|
commitdiff
|
tree
2017-06-21
wilco
Em
i
t
S
IM
D
moves
a
s
mov
commit
|
commitdiff
|
tree
2017-06-21
wil
c
o
Imp
r
o
v
e dup pattern
commit
|
commitdiff
|
tree
2017-06-21
wilco
Mar
k
s
y
m
b
o
ls as constant
commit
|
commitdiff
|
tree
2017-06-14
wilco
Impr
o
ve Cortex-A53 FP scheduler
commit
|
commitdiff
|
tree
2017-05-30
wilco
Fix ARM bootstrap fail
u
re due to an odd
w
arning
:
commit
|
commitdiff
|
tree
2017-05-24
w
ilco
When lra-
r
emat remat
e
r
i
alizes an i
n
s
truction wi
t
h a
.
.
.
commit
|
commitdiff
|
tree
2017-05-18
w
i
lco
Add
a
n a
s
sert
to leaf_
f
unction_p to ensure it i
s
no
t
.
.
.
commit
|
commitdiff
|
tree
2017-05-10
wilco
Move a
n
use-af
t
er-free access
b
efore
the delete
.
commit
|
commitdiff
|
tree
2017-05-08
wilco
This f
i
xe
s
a
f
ew
failures on ARM and AArch64 due to
.
.
.
commit
|
commitdiff
|
tree
2017-05-05
w
ilco
Float to i
n
t mov
e
s currently generate
ineffi
c
ient
code
.
.
.
commit
|
commitdiff
|
tree
2017-05-05
wilco
Code scheduling
f
or Cortex-A53
isn't as good as it
.
.
.
commit
|
commitdiff
|
tree
2017-05-04
w
i
lco
Man
y
suppor
t
ed core
s
use
the AUTOPREFETCHER_
W
EAK
se
t
ting
.
.
.
commit
|
commitdiff
|
tree
2017-05-04
w
i
lco
Set j
u
mp alignment
t
o
4 for Cor
t
ex cores
as it redu
c
es
.
.
.
commit
|
commitdiff
|
tree
2017-05-04
wi
l
co
W
i
th -mcpu=
g
eneric the loop alignment is
c
urrently
.
.
.
commit
|
commitdiff
|
tree
2017-05-04
wilco
All cores
which add a cpu_addrcos
t
_tabl
e
use a non
.
.
.
commit
|
commitdiff
|
tree
2017-03-22
wi
l
co
Recently we'
v
e
p
ut a lot of
e
ffo
r
t in
t
o impr
o
ving ifcvt
.
.
.
commit
|
commitdiff
|
tree
2017-03-22
wilco
Many supported
c
ores implem
e
nt f
u
sion of AE
S
ins
t
ructio
n
s
.
.
.
commit
|
commitdiff
|
tree
2017-01-18
wilc
o
SHA1H
i
nstructions may
b
e scheduled after a SHA1C
i
nst
r
uction
commit
|
commitdiff
|
tree
2017-01-17
wilco
This
p
a
tch
simplifies th
e
handling
of
E
H return
.
We
.
.
.
commit
|
commitdiff
|
tree
2017-01-11
wi
l
co
M
y
previous change to t
h
e
Cortex
-
A53 sc
h
edu
l
er r
e
sul
t
e
d
.
.
.
commit
|
commitdiff
|
tree
2016-12-16
w
i
l
c
o
The negdi2
patterns for
A
RM and
T
humb-2 are duplicated
.
.
.
commit
|
commitdiff
|
tree
2016-12-16
wilco
Th
u
m
b
u
ses a
special register allocati
o
n order to increase
.
.
.
commit
|
commitdiff
|
tree
2016-12-15
wilco
Usi
n
g
l
ea
f
_function_p
i
n a backend is dangerou
s
as
.
.
.
commit
|
commitdiff
|
tree
2016-12-09
wilc
o
Add t
h
e
t
e
st this
t
ime
.
.
.
commit
|
commitdiff
|
tree
2016-12-08
w
ilco
This patch
fixes an issue
i
n
aarch64_classify_addre
s
s
.
.
.
commit
|
commitdiff
|
tree
2016-12-07
wi
l
co
GCC caches the
w
he
t
her
a
fun
c
tion is a leaf in cr
t
l
.
.
.
commit
|
commitdiff
|
tree
2016-12-07
wil
c
o
Improve T
I
mode address of
f
s
et
s
- these may eith
e
r
.
.
.
commit
|
commitdiff
|
tree
2016-11-16
wilco
L
o
o
king at PR773
0
8,
one of the issues is that the bs
w
ap
.
.
.
commit
|
commitdiff
|
tree
2016-11-14
wilco
The second patch updates
the C
o
rtex-A57 scheduler now
.
.
.
commit
|
commitdiff
|
tree
2016-11-14
wilco
Currently the SBFM, UBFM and BF
M
instructions all use
.
.
.
commit
|
commitdiff
|
tree
2016-11-14
wilco
The
e
xisti
n
g vector costs
s
t
o
p some
benefic
i
al ve
c
tori
z
ation
.
.
.
commit
|
commitdiff
|
tree
2016-11-02
wilco
Add LE/
B
E SHA1H patterns with a V2SI
i
npu
t
.
This avoids
.
.
.
commit
|
commitdiff
|
tree
2016-11-02
wilco
The add expander still
contains some expans
i
on code
.
.
.
commit
|
commitdiff
|
tree
2016-10-25
wilco
With -
f
pu=neon DI mode
s
hifts are expanded after rel
o
ad
.
.
.
commit
|
commitdiff
|
tree
2016-10-21
wilco
Really add the test this tim
e
.
.
.
commit
|
commitdiff
|
tree
2016-10-21
wilco
Imp
r
ove stack adju
s
tment
b
y reusin
g
a temporary move
.
.
.
commit
|
commitdiff
|
tree
2016-10-21
w
i
lco
If the number of integer ca
l
le
e
-saves
i
s odd, a
n
y FP
.
.
.
commit
|
commitdiff
|
tree
2016-10-05
wilco
Move all exi
s
t
ing strchr and st
r
rchr
f
ol
d
i
n
g fr
o
m builtins
.
.
.
commit
|
commitdiff
|
tree
2016-09-28
wi
l
co
If strchr can't
b
e
folde
d
in gimple-fold, we still
.
.
.
commit
|
commitdiff
|
tree
2016-09-28
wilco
Opti
m
ize strchr
(
s, 0)
to s + strl
e
n
(s)
.
strch
r
.
.
.
commit
|
commitdiff
|
tree
2016-09-07
wilco
Improve aarch64_legitimiz
e
_address - avoid sp
l
itti
n
g
.
.
.
commit
|
commitdiff
|
tree
2016-09-01
wilco
This p
a
tch adds legitimize_address_disp
l
acement ho
o
k
.
.
.
commit
|
commitdiff
|
tree
2016-08-01
w
i
lco
This pa
t
c
h opt
i
mizes the prolog and e
p
ilog code to
.
.
.
commit
|
commitdiff
|
tree
2016-07-28
wil
c
o
On AArch6
4
th
e
U
X
TB and UXTH ins
t
ructio
n
s ar
e
aliases
.
.
.
commit
|
commitdiff
|
tree
2016-07-28
w
i
lc
o
This patc
h
s
et im
p
roves
z
ero ext
e
nd
costs and
code genera
t
i
on
.
commit
|
commitdiff
|
tree
2016-07-28
wil
c
o
T
his patch impr
o
ves the readability of th
e
prolog and
.
.
.
commit
|
commitdiff
|
tree
2016-07-06
w
i
l
c
o
This patch impr
o
ves the accurac
y
of th
e
Cortex-A53
.
.
.
commit
|
commitdiff
|
tree
2016-07-06
wilco
F
i
x protot
y
pe in vst1Q_laneu64-1
.
c to
u
nsigned ch
a
r*
.
commit
|
commitdiff
|
tree
2016-06-30
wilco
This patch
s
ets th
e
branch cost
t
o the s
a
me most optim
a
l
.
.
.
commit
|
commitdiff
|
tree
2016-06-29
wilco
Inc
r
ease loop
a
lignment on Corte
x
cores to
8
and set
.
.
.
commit
|
commitdiff
|
tree
2016-06-28
wilco
This patch fix
e
s a bug in the bsw
a
p pass
.
I
n
big
-
endian
.
.
.
commit
|
commitdiff
|
tree
2016-06-21
w
i
lco
The recently added gcc
.
target/a
a
rch64/adv
s
i
m
d-intrinsics
.
.
.
commit
|
commitdiff
|
tree
2016-06-21
w
i
lc
o
Fix tree
-
ss
a
/attr-hot
c
old
-
2
.
c failures now that the
.
.
.
commit
|
commitdiff
|
tree
2016-06-21
wi
l
co
Due to recent
improveme
n
t
s to the vectorizer, the number
.
.
.
commit
|
commitdiff
|
tree
2016-06-20
wilco
This patch cleans
up the -mpc-relative-lo
a
ds option
.
.
.
commit
|
commitdiff
|
tree
2016-06-20
wilco
Improv
e
modes_tieable b
y
r
e
t
urning tru
e
in
m
ore cases
.
.
.
commit
|
commitdiff
|
tree
2016-06-20
wilco
Th
e
Cortex-A57 scheduler is missing
fcsel, so
add i
t
.
commit
|
commitdiff
|
tree
2016-05-27
wi
l
co
Remove aarch64_cannot_chan
g
e_mode_cla
s
s as
the underlying
.
.
.
commit
|
commitdiff
|
tree
2016-05-26
wi
l
co
GCC e
x
pands switch statements in a very si
m
plistic
.
.
.
commit
|
commitdiff
|
tree
2016-05-26
wilco
SIMD o
p
erations
like combine prefer to
h
a
ve
th
e
ir o
p
erands
.
.
.
commit
|
commitdiff
|
tree
2016-05-16
wilco
Update documentation of AArch64 o
p
tions
for GCC6 to
.
.
.
commit
|
commitdiff
|
tree
2016-05-16
wi
l
co
Some patterns are us
i
ng '%w2' for immediat
e
operands
.
.
.
commit
|
commitdiff
|
tree
2016-05-16
wilco
This patch
fixe
s
the attributes of i
n
teger im
m
ediate
.
.
.
commit
|
commitdiff
|
tree
2016-02-17
wilco
As part of
t
h
e
f
inal cleanup of aar
c
h
64_internal_mov_i
m
m
e
dia
.
.
.
2016-02-17
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-02-10
wilc
o
Enable instru
c
t
ion f
u
sion of dependent AESE; AESMC
.
.
.
commit
|
commitdiff
|
tree
2016-02-10
wilc
o
E
n
able instruction fusion of A
E
S in
s
tructions
o
n ARM
.
.
.
commit
|
commitdiff
|
tree
2016-02-04
wi
l
c
o
This patch fixes an ex
p
onential iss
u
e in
ccmp
.
c
.
When
.
.
.
2016-02-04
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-02-03
wilco
Fix the ccmp_1
.
c
t
est back to u
s
e '0
'
as r
e
gular e
x
p
r
e
ssions
.
.
.
2016-02-03
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-02-02
wilco
Improve TARGE
T
_IRA_CHANGE_PS
E
UDO_A
L
LOCNO_CLA
S
S target
.
.
.
2016-02-02
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-02-02
wil
c
o
This patch adds support for the T
A
RGE
T
_IRA_CHANGE_PSEUD
O
_AL
L
.
.
.
2016-02-02
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-01-28
wilco
A re
c
ent change added UNSPEC to
t
h
e
CCMP pattern
s
to
.
.
.
2016-01-28
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-01-28
w
i
lco
Add support for vector permu
t
e
c
ost since various permutes
.
.
.
2016-01-28
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-01-28
w
i
lco
Several instr
u
ctions di
s
assemble a zero im
m
e
diat
e
as
.
.
.
2016-01-28
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree
2016-01-27
w
ilc
o
20
1
5
-01-2
7
Wilco
Dijkstra <wdij
k
s
tr@ar
m
.
co
m
>
2015-01-27
Wilco
Dijkstra <wdijkstr@arm.com>
commit
|
commitdiff
|
tree