Toggle navigation
Patchwork
GNU Compiler Collection
Patches
Bundles
About this project
Login
Register
Mail settings
Show patches with
: Submitter =
Roger Sayle
 State =
Action Required
 Archived =
No
 141 patches
Series
Submitter
State
any
Action Required
New
Under Review
Accepted
Rejected
RFC
Not applicable
Changes Requested
Failed CI
Superseded
Deferred
Committed
Dropped
DCO or assignment missing
Search
Archived
No
Yes
Both
Delegate

Nobody
siddhesh
siddhesh
siddhesh
codonell
azanella
gbenson
jankratochvil
ppluzhnikov
neleai
tromey
palves
keiths
sergiodj
StanShebs
teawater
simark
rth
roland
davem
schwab
Allan
aj
ldv
tschwinge
tschwinge
tschwinge
andreaskr
fweimer
ChrisMetcalf
jsm28
aurel32
will
vapier
vapier
tuliom
hjl
triegel
jwlemke
cltang
macro
nsz
pwbot
arjun
stli
ludo
rj
zack
mfabian
rluzynski
dj
sthibaul
mscastanho
lamm
girish946
maennich
dodji
zimmerma
rzinsly
lukma
mjw
goldsteinn
raoni
jason
jwakely
jwakely
maximk
trodgers
trodgers
ams
ams
ams
siddhesh_staff
lancesix
aburgess
Apply
«
1
2
»
Patch
Series
rb/tb
S/W/F
Date
Submitter
Delegate
State
Optimize (X<<C)+(Y<<C) as (X+Y)<<C for signed addition.
Optimize (X<<C)+(Y<<C) as (X+Y)<<C for signed addition.
 



20220913
Roger Sayle
New
PR treeoptimization/71343: Value number X<<2 as X*4.
PR treeoptimization/71343: Value number X<<2 as X*4.
 



20220913
Roger Sayle
New
PR rtloptimization/106594: Preserve zero_extend when cheap.
PR rtloptimization/106594: Preserve zero_extend when cheap.
 



20220911
Roger Sayle
New
[x86_64] Support shifts and rotates by integer constants in TImode STV.
[x86_64] Support shifts and rotates by integer constants in TImode STV.
 



20220815
Roger Sayle
New
[x86,take,#2] Move V1TI shift/rotate lowering from expand to prereload split.
[x86,take,#2] Move V1TI shift/rotate lowering from expand to prereload split.
 



20220812
Roger Sayle
New
[x86] PR target/106577: force_reg may clobber operands during split.
[x86] PR target/106577: force_reg may clobber operands during split.
 



20220812
Roger Sayle
New
[x86_64] Use PTEST to perform AND in TImode STV of (A & B) != 0.
[x86_64] Use PTEST to perform AND in TImode STV of (A & B) != 0.
 



20220809
Roger Sayle
New
PR treeoptimization/64992: (B << 2) != 0 is B when B is Boolean.
PR treeoptimization/64992: (B << 2) != 0 is B when B is Boolean.
 



20220808
Roger Sayle
New
PR treeoptimization/71343: Optimize (X<<C)&(Y<<C) as (X&Y)<<C.
PR treeoptimization/71343: Optimize (X<<C)&(Y<<C) as (X&Y)<<C.
 



20220808
Roger Sayle
New
middleend: Optimize ((X >> C1) & C2) != C3 for more cases.
middleend: Optimize ((X >> C1) & C2) != C3 for more cases.
 



20220807
Roger Sayle
New
[x86,take,#2] Add peephole2 to reduce double word register shuffling
[x86,take,#2] Add peephole2 to reduce double word register shuffling
 



20220807
Roger Sayle
New
[x86] Move V1TI shift/rotate lowering from expand to prereload split.
[x86] Move V1TI shift/rotate lowering from expand to prereload split.
 



20220805
Roger Sayle
New
middleend: Allow backend to expand/split double word compare to 0/1.
middleend: Allow backend to expand/split double word compare to 0/1.
 



20220803
Roger Sayle
New
[x86] PR target/47949: Use xchg to move from/to AX_REG with Oz.
[x86] PR target/47949: Use xchg to move from/to AX_REG with Oz.
 



20220802
Roger Sayle
New
[take,#2] Some additional zeroextension related optimizations in simplifyrtx.
[take,#2] Some additional zeroextension related optimizations in simplifyrtx.
 



20220802
Roger Sayle
New
[x86_64] PR target/106481: Handle CONST_WIDE_INT in REG_EQUAL during STV.
[x86_64] PR target/106481: Handle CONST_WIDE_INT in REG_EQUAL during STV.
 



20220801
Roger Sayle
New
[Ada] Update configure to check for a recent gnat Ada compiler.
[Ada] Update configure to check for a recent gnat Ada compiler.
 



20220730
Roger Sayle
New
PR bootstrap/106472: Add libgo depends on libbacktrace to Makefile.def
PR bootstrap/106472: Add libgo depends on libbacktrace to Makefile.def
 



20220730
Roger Sayle
New
[x86_64,take,#2] PR target/106450: Tweak timode_remove_non_convertible_regs.
[x86_64,take,#2] PR target/106450: Tweak timode_remove_non_convertible_regs.
 



20220730
Roger Sayle
New
[x86_64] PR target/106450: Tweak timode_remove_non_convertible_regs.
[x86_64] PR target/106450: Tweak timode_remove_non_convertible_regs.
 



20220728
Roger Sayle
New
Some additional zeroextension related optimizations in simplifyrtx.
Some additional zeroextension related optimizations in simplifyrtx.
 



20220727
Roger Sayle
New
middleend: More support for ABIs that pass FP values as wider ints.
middleend: More support for ABIs that pass FP values as wider ints.
 



20220726
Roger Sayle
New
Add new target hook: simplify_modecc_const.
Add new target hook: simplify_modecc_const.
 



20220726
Roger Sayle
New
[Documentation] Correct RTL documentation: (use (mem ...)) is allowed.
[Documentation] Correct RTL documentation: (use (mem ...)) is allowed.
 



20220723
Roger Sayle
New
[x86,take,#3] PR target/91681: zero_extendditi2 pattern for more optimizations.
[x86,take,#3] PR target/91681: zero_extendditi2 pattern for more optimizations.
 



20220723
Roger Sayle
New
[x86] PR target/106303: Fix TImode STV related failures.
[x86] PR target/106303: Fix TImode STV related failures.
 



20220723
Roger Sayle
New
[middleend] PR c/106264: Silence warnings from __builtin_modf et al.
[middleend] PR c/106264: Silence warnings from __builtin_modf et al.
 



20220716
Roger Sayle
New
[x86] PR target/106273: Add earlyclobber to *andn<dwi>3_doubleword_bmi
[x86] PR target/106273: Add earlyclobber to *andn<dwi>3_doubleword_bmi
 



20220715
Roger Sayle
New
PR target/106278: Keep REG_EQUAL notes consistent during TImode STV.
PR target/106278: Keep REG_EQUAL notes consistent during TImode STV.
 



20220714
Roger Sayle
New
Move reload_completed and other rtl.h globals to crtl structure.
Move reload_completed and other rtl.h globals to crtl structure.
 



20220710
Roger Sayle
New
[x86_64] Improved ScalarToVector (STV) support for TImode to V1TImode.
[x86_64] Improved ScalarToVector (STV) support for TImode to V1TImode.
 



20220709
Roger Sayle
New
[gcc12,backport] PR target/105930: Split *xordi3_doubleword after reload on x86.
[gcc12,backport] PR target/105930: Split *xordi3_doubleword after reload on x86.
 



20220709
Roger Sayle
New
[x86] Fun with flags: Adding stc/clc instructions to i386.md.
[x86] Fun with flags: Adding stc/clc instructions to i386.md.
 



20220708
Roger Sayle
New
Be careful with MODE_CC in simplify_const_relational_operation.
Be careful with MODE_CC in simplify_const_relational_operation.
 



20220707
Roger Sayle
New
[PATCH/RFC] combine_completed global variable.
[PATCH/RFC] combine_completed global variable.
 



20220707
Roger Sayle
New
[x86] Support *testdi_not_doubleword during STV pass.
[x86] Support *testdi_not_doubleword during STV pass.
 



20220707
Roger Sayle
New
[x86] UNSPEC_PALIGNR optimizations and cleanups.
[x86] UNSPEC_PALIGNR optimizations and cleanups.
 



20220630
Roger Sayle
New
[x86,take,#2] Double word logical operation cleanups in i386.md.
[x86,take,#2] Double word logical operation cleanups in i386.md.
 



20220630
Roger Sayle
New
[x86] Double word logical operation cleanups in i386.md.
[x86] Double word logical operation cleanups in i386.md.
 



20220628
Roger Sayle
New
[rs6000] Improve constant integer multiply using rldimi.
[rs6000] Improve constant integer multiply using rldimi.
 



20220626
Roger Sayle
New
[x86] Use xchg for DImode double word rotate by 32 bits with m32.
[x86] Use xchg for DImode double word rotate by 32 bits with m32.
 



20220626
Roger Sayle
New
[take,2] middleend: Support ABIs that pass FP values as wider integers.
[take,2] middleend: Support ABIs that pass FP values as wider integers.
 



20220626
Roger Sayle
New
[x86] PR rtloptimization/96692: ((AB)^C)^A using andn with mbmi.
[x86] PR rtloptimization/96692: ((AB)^C)^A using andn with mbmi.
 



20220626
Roger Sayle
New
[x86] PR target/105930: Split *xordi3_doubleword after reload.
[x86] PR target/105930: Split *xordi3_doubleword after reload.
 



20220622
Roger Sayle
New
[rs6000] PR target/105991: Recognize PLUS and XOR forms of rldimi.
[rs6000] PR target/105991: Recognize PLUS and XOR forms of rldimi.
 



20220617
Roger Sayle
New
[x86] Double word implementation of and; cmp to not; test optimization.
[x86] Double word implementation of and; cmp to not; test optimization.
 



20220605
Roger Sayle
New
[take,#2] Fold truncations of left shifts in match.pd
[take,#2] Fold truncations of left shifts in match.pd
 



20220605
Roger Sayle
New
[x86] Recognize vpcmov in combine with mxop.
[x86] Recognize vpcmov in combine with mxop.
 



20220604
Roger Sayle
New
[x86] PR target/91681: zero_extendditi2 pattern for more optimizations.
[x86] PR target/91681: zero_extendditi2 pattern for more optimizations.
 



20220603
Roger Sayle
New
[PATCH/RFC] cprop_hardreg... Third time's a charm.
[PATCH/RFC] cprop_hardreg... Third time's a charm.
 



20220602
Roger Sayle
New
[x86] Add peephole2 to reduce double word register shuffling.
[x86] Add peephole2 to reduce double word register shuffling.
 



20220602
Roger Sayle
New
Fold truncations of left shifts in match.pd
Fold truncations of left shifts in match.pd
 



20220530
Roger Sayle
New
[x86,PING] PR target/70321: Split double word equality/inequality after STV.
[x86,PING] PR target/70321: Split double word equality/inequality after STV.
 



20220530
Roger Sayle
New
Canonicalize X&Y as X*Y in match.pd when Y is [0,1].
Canonicalize X&Y as X*Y in match.pd when Y is [0,1].
 



20220524
Roger Sayle
New
[PATCH/RFC] PR treeoptimization/96912: Recognize VEC_COND_EXPR in match.pd
[PATCH/RFC] PR treeoptimization/96912: Recognize VEC_COND_EXPR in match.pd
 



20220523
Roger Sayle
New
[x86,PING] Peephole pand;pxor into pandn
[x86,PING] Peephole pand;pxor into pandn
 



20220523
Roger Sayle
New
PR treeoptimization/105668: Provide RTL expansion for VEC_COND_EXPR.
PR treeoptimization/105668: Provide RTL expansion for VEC_COND_EXPR.
 



20220523
Roger Sayle
New
Simplify vec_unpack of uniform_vector_p constructors in match.pd.
Simplify vec_unpack of uniform_vector_p constructors in match.pd.
 



20220521
Roger Sayle
New
[PING] PR middleend/95126: Expand small const structs as immediate constants
[PING] PR middleend/95126: Expand small const structs as immediate constants
 



20220521
Roger Sayle
New
[x86] Some additional ix86_rtx_costs cleanups: NEG, AND and pandn.
[x86] Some additional ix86_rtx_costs cleanups: NEG, AND and pandn.
 



20220518
Roger Sayle
New
Simplify logic in treescalarevolution's expensive_expression_p.
Simplify logic in treescalarevolution's expensive_expression_p.
 



20220517
Roger Sayle
New
[x86,take,2] Improved V1TI (and V2DI) mode equality/inequality.
[x86,take,2] Improved V1TI (and V2DI) mode equality/inequality.
 



20220513
Roger Sayle
New
[x86] PR target/92578: Peephole2s to tweak cmove register allocation.
[x86] PR target/92578: Peephole2s to tweak cmove register allocation.
 



20220425
Roger Sayle
New
PR middleend/98865: Optimize (a>>63)*b as (a>>63)&b in match.pd.
PR middleend/98865: Optimize (a>>63)*b as (a>>63)&b in match.pd.
 



20220420
Roger Sayle
New
[x86_64] PR middleend/105135: Catch more cmov idioms in combine.
[x86_64] PR middleend/105135: Catch more cmov idioms in combine.
 



20220419
Roger Sayle
New
[x86] PR target/70321: Split double word equality/inequality after STV.
[x86] PR target/70321: Split double word equality/inequality after STV.
 



20220413
Roger Sayle
New
[x86] Avoid andn and generate shorter not;and with Oz.
[x86] Avoid andn and generate shorter not;and with Oz.
 



20220413
Roger Sayle
New
[x86] Peephole pand;pxor into pandn.
[x86] Peephole pand;pxor into pandn.
 



20220412
Roger Sayle
New
[x86_64] Avoid andb %dil when optimizing for size.
[x86_64] Avoid andb %dil when optimizing for size.
 



20220412
Roger Sayle
New
[x86_64] PR 90356: Use xor to load const_double 0.0 on SSE (always)
[x86_64] PR 90356: Use xor to load const_double 0.0 on SSE (always)
 



20220317
Roger Sayle
New
Avoid generating unused labels in genmatch.
Avoid generating unused labels in genmatch.
 



20220316
Roger Sayle
New
Generate an if instead of a switch with one case in genmatch.
Generate an if instead of a switch with one case in genmatch.
 



20220316
Roger Sayle
New
[x86] PR target/94680: Clear upper bits of V2DF using movq (like V2DI).
[x86] PR target/94680: Clear upper bits of V2DF using movq (like V2DI).
 



20220315
Roger Sayle
New
Performance/size improvement to single_use when matching GIMPLE.
Performance/size improvement to single_use when matching GIMPLE.
 



20220315
Roger Sayle
New
Ignore (possible) signed zeros in operands of FP comparisons.
Ignore (possible) signed zeros in operands of FP comparisons.
 



20220314
Roger Sayle
New
[x86] Fix libitm.c/memset1.c test fails with new peephole2s.
[x86] Fix libitm.c/memset1.c test fails with new peephole2s.
 



20220312
Roger Sayle
New
PR treeoptimization/101895: Fold VEC_PERM to help recognize FMA.
PR treeoptimization/101895: Fold VEC_PERM to help recognize FMA.
 



20220311
Roger Sayle
New
PR middleend/98420: Don't fold x  x to 0.0 with froundingmath
PR middleend/98420: Don't fold x  x to 0.0 with froundingmath
 



20220310
Roger Sayle
New
[v2] PR treeoptimization/98335: Improvements to DSE's compute_trims.
[v2] PR treeoptimization/98335: Improvements to DSE's compute_trims.
 



20220309
Roger Sayle
New
[C++] PR c++/96440: ICEoninvalidcode error recovery.
[C++] PR c++/96440: ICEoninvalidcode error recovery.
 



20220307
Roger Sayle
New
[C++] PR c++/96437: ICEoninvalidcode error recovery.
[C++] PR c++/96437: ICEoninvalidcode error recovery.
 



20220307
Roger Sayle
New
[C++] PR c++/96329: ICEoninvalidcode error recovery.
[C++] PR c++/96329: ICEoninvalidcode error recovery.
 



20220307
Roger Sayle
New
[x86] PR treeoptimization/98335: New peephole2 xorl; movb > movzbl
[x86] PR treeoptimization/98335: New peephole2 xorl; movb > movzbl
 



20220307
Roger Sayle
New
PR treeoptimization/98335: Improvements to DSE's compute_trims.
PR treeoptimization/98335: Improvements to DSE's compute_trims.
 



20220307
Roger Sayle
New
[i386] PR 104732: Simplify/fix DI mode logic expansion/splitting on m32.
[i386] PR 104732: Simplify/fix DI mode logic expansion/splitting on m32.
 



20220304
Roger Sayle
New
PR c++/84964: Middleend patch to expand_call for ICE after sorry.
PR c++/84964: Middleend patch to expand_call for ICE after sorry.
 



20220228
Roger Sayle
New
PR middleend/80270: ICE in extract_bit_field_1
PR middleend/80270: ICE in extract_bit_field_1
 



20220227
Roger Sayle
New
PR c++/39751: ICEoninvalid parsing regression.
PR c++/39751: ICEoninvalid parsing regression.
 



20220226
Roger Sayle
New
PR middleend/95126: Expand small const structs as immediate constants.
PR middleend/95126: Expand small const structs as immediate constants.
 



20220226
Roger Sayle
New
[C++] PR c++/96442: Another improved error recovery in enumerations.
[C++] PR c++/96442: Another improved error recovery in enumerations.
 



20220222
Roger Sayle
New
c++: PR c++/95999: Improved error recovery in enumeration lists.
c++: PR c++/95999: Improved error recovery in enumeration lists.
 



20220221
Roger Sayle
New
Implement constantfolding simplifications of reductions.
Implement constantfolding simplifications of reductions.
 



20220221
Roger Sayle
New
PR treeoptimization/83907: Improved memset handling in strlen pass.
PR treeoptimization/83907: Improved memset handling in strlen pass.
 



20220220
Roger Sayle
New
Improved constant folding for scalar evolution.
Improved constant folding for scalar evolution.
 



20220220
Roger Sayle
New
PR middleend/65855: Scalar evolution for quadratic chrecs
PR middleend/65855: Scalar evolution for quadratic chrecs
 



20220218
Roger Sayle
New
PR c/104506: Tolerate error_mark_node in useless_type_conversion_p.
PR c/104506: Tolerate error_mark_node in useless_type_conversion_p.
 



20220214
Roger Sayle
New
gfortran: Respect target's NO_DOT_IN_LABEL in transcommon.cc
gfortran: Respect target's NO_DOT_IN_LABEL in transcommon.cc
 



20220210
Roger Sayle
New
middleend: Support ABIs that pass FP values as wider integers.
middleend: Support ABIs that pass FP values as wider integers.
 



20220209
Roger Sayle
New
PR rtloptimization/101885: Prevent combine from clobbering flags
PR rtloptimization/101885: Prevent combine from clobbering flags
 



20220203
Roger Sayle
New
PR middleend/104140: bootstrap ICE on riscv.
PR middleend/104140: bootstrap ICE on riscv.
 



20220121
Roger Sayle
New
«
1
2
»