Fix an obscure read-after-free bug that Duncan found.
[oota-llvm.git] / lib / Transforms /
2008-02-04 Owen AndersonBe more precise when eliminating pointers bue to memcpy...
2008-02-04 Owen AndersonAllow GVN to hack on memcpy's, making them open to...
2008-02-03 Nick LewyckyThere are some cases where icmp(add) can be folded...
2008-02-03 Nick LewyckyHack on vectors too.
2008-02-03 Nick LewyckyFold away one multiply in instcombine. This would norma...
2008-02-01 Duncan SandsDon't drop function/call return attributes like 'nounwind'.
2008-01-30 Nick LewyckyRemove a couple more cases of "getNumUses() == 0"....
2008-01-30 Nick LewyckyUse empty() instead of comparing size() with zero.
2008-01-30 Nick LewyckyDon't DCE FreeInst's. We were using those! Patch from...
2008-01-30 Owen AndersonMake DSE much more aggressive by performing DCE earlier...
2008-01-30 Chris LattnerFix a bug where scalarrepl would discard offset if...
2008-01-29 Chris LattnerDon't let globalopt hack on volatile loads or stores.
2008-01-29 Dan GohmanUse empty() instead of comparing size() with zero.
2008-01-29 Chris Lattnereliminate additions of 0.0 when they are obviously...
2008-01-29 Owen AndersonAdd support for eliminating memcpy's at the end of...
2008-01-29 Devang PatelFilter loops that subtract induction variables.
2008-01-28 Chris LattnerTransform calls to memcpy into llvm.memcpy calls, patch...
2008-01-28 Nick LewyckyHandle some more combinations of extend and icmp. Fixes...
2008-01-28 Chris LattnerFix PR1932 by disabling an xform invalid for fdiv.
2008-01-28 Chris LattnerFix PR1938 by forcing the code that uses an undefined...
2008-01-27 Nick LewyckyBe more careful modifying the use_list while also itera...
2008-01-27 Duncan SandsRevert r46393: readonly/readnone functions are no
2008-01-27 Bill WendlingThe CorrelatedExpressionElimination pass is known to...
2008-01-27 Chris LattnerFold fptrunc(add (fpextend x), (fpextend y)) -> add...
2008-01-26 Bill WendlingIf there are no machine instructions emitted for a...
2008-01-26 Duncan SandsCreate an explicit copy for byval parameters even
2008-01-26 Bill WendlingIf we have a function like this:
2008-01-25 Duncan SandsDo this more neatly.
2008-01-25 Owen AndersonDeadStoreElimination can treat byval parameters as...
2008-01-22 Nick LewyckyEnable the fix I just checked in, silly me.
2008-01-22 Nick LewyckyMultiply can be evaluated in a different type, so long...
2008-01-20 Duncan SandsMake sure the caller doesn't use freed memory.
2008-01-20 Duncan SandsInitializing an unsigned with ~0UL causes the compiler
2008-01-17 Evan ChengDAE bug fix. Don't lose parameter attributes on vararg...
2008-01-17 Chris LattnerFix arg promotion to propagate the correct attrs on...
2008-01-16 Chris LattnerHandle attribute(used) global variables that are i8.
2008-01-16 Devang PatelDo not strip llvm.used values.
2008-01-14 Duncan SandsI noticed that the trampoline straightening transformat...
2008-01-14 Chris LattnerFix the miscompilation of MiBench/consumer-lame that...
2008-01-14 Chris LattnerThe isNotSuitableForSRA property is now dead, don't...
2008-01-14 Chris LattnerChange SRAGlobal to not depend on isNotSuitableForSRA...
2008-01-14 Chris LattnerMake the 'shrink global to bool' optimization more...
2008-01-14 Chris LattnerTurn a memcpy from a double* into a load/store of doubl...
2008-01-13 Chris Lattnerfactor memcpy/memmove simplification out to its own...
2008-01-13 Chris Lattnersimplify some code. If we can infer alignment for...
2008-01-13 Chris Lattnersimplify some code by adding a InsertBitCastBefore...
2008-01-13 Chris LattnerFix PR1907, a nasty miscompilation because instcombine...
2008-01-13 Duncan SandsSmall simplification.
2008-01-13 Duncan SandsWhen turning a call to a bitcast function into a direct...
2008-01-12 Chris Lattnerwe don't have to make an explicit copy of a byval argum...
2008-01-12 Chris LattnerAllow clients to specify the inline threshold when...
2008-01-11 Duncan SandsWhen DAE drops the varargs part of a function, ensure any
2008-01-11 Chris LattnerTeach argpromote to ruthlessly hack small byval structs...
2008-01-11 Chris LattnerUse smallptrset instead of std::set for efficiency.
2008-01-11 Chris Lattnera byval argument is guaranteed to be valid to load.
2008-01-11 Chris LattnerUpdate this code to use eraseFromParent where possible...
2008-01-11 Chris Lattnerreplace a loop with a constant time check.
2008-01-11 Chris Lattneranother minor datastructure tweak.
2008-01-11 Chris Lattnerstart using smallvector to avoid vector heap thrashing.
2008-01-11 Chris LattnerWhen inlining a functino with a byval argument, make...
2008-01-08 Chris LattnerImplement PR1795, an instcombine hack for forming GEPs...
2008-01-07 Duncan SandsSmall cleanup for handling of type/parameter attribute
2008-01-07 Gordon HenriksenDeleting an empty file. Thanks, /usr/bin/patch!
2008-01-07 Gordon HenriksenWith this patch, the LowerGC transformation becomes the
2008-01-06 Duncan SandsThe transform that tries to turn calls to bitcast funct...
2008-01-06 Duncan SandsWhen transforming a call to a bitcast function into
2008-01-05 Chris Lattnerremove a couple more unsafe xforms in the face of overflow.
2008-01-05 Chris Lattnerremove the (x-y) < 0 comparison xform, it miscompiles
2008-01-04 Wojciech Matyjewiczfix typo
2008-01-04 Chris LattnerFix PR1896
2008-01-03 Chris Lattnerdon't hoist FP additions into unconditional adds +...
2008-01-02 Chris Lattneradd missing #include
2007-12-29 Chris LattnerRemove attribution from file headers, per discussion...
2007-12-29 Chris Lattnerremove attribution from lib Makefiles.
2007-12-29 Christopher LambDisable null pointer folding transforms for non-generic...
2007-12-29 Chris Lattnerdead calls to llvm.stacksave can be deleted, even thoug...
2007-12-28 Owen AndersonRepair a transform that Chris noticed a bug in. Thanks...
2007-12-28 Chris Lattnerdisable this instcombine xform, it miscompiles:
2007-12-25 Gordon HenriksenFixing several transforms which would drop the collecto...
2007-12-25 Chris LattnerDon't break critical edges for single-bb loops, this...
2007-12-25 Gordon HenriksenGC poses hazards to the inliner. Consider:
2007-12-24 Chris Lattneradd a -backedge-hack llc-beta option to codegenprepare.
2007-12-22 Chris Lattnerimplement InstCombine/shift-trunc-shift.ll. This allows
2007-12-22 Devang PatelIf succ has succ itself as one of the predecessors...
2007-12-21 Duncan SandsMake DAE not wipe out attributes on calls, and not...
2007-12-20 Christopher LambImplement review feedback, including additional transforms
2007-12-20 Evan ChengClean up previous patch: PHI uses should not prevent...
2007-12-20 Chris Lattnersimplify this code with the new m_Zero() pattern. ...
2007-12-19 Evan ChengAllow iv reuse if the user is a PHI node which is in...
2007-12-19 Duncan SandsWhen inlining through an 'nounwind' call, mark inlined
2007-12-18 Christopher LambFold subtracts into integer compares vs. zero. This...
2007-12-18 Christopher LambFix comments
2007-12-18 Christopher LambRemove an orthogonal transformation of the selection...
2007-12-18 Duncan SandsRename isNoReturn to doesNotReturn, and isNoUnwind to
2007-12-18 Christopher LambFix typos.
2007-12-18 Christopher LambFold certain additions through selects (and their compa...
2007-12-17 Duncan SandsMake invokes of inline asm legal. Teach codegen
2007-12-17 David GreeneGLIBCXX_DEBUG fix. std::vector<>::end() is invalidated...
2007-12-17 David GreeneGet rid of annoying spaces.
2007-12-17 David GreeneFix GLIBCXX_DEBUG errors. Erase invalidates std::vecto...
next