ChangeLog

   1 2015-07-31  Lucas Forschler  <lforschler@apple.com>
   2
   3         Merge r187579
   4
   5     2015-07-29  Filip Pizlo  <fpizlo@apple.com>
   6
   7             DFG::ArgumentsEliminationPhase should emit a PutStack for all of the GetStacks that the ByteCodeParser emitted
   8             https://bugs.webkit.org/show_bug.cgi?id=147433
   9             rdar://problem/21668986
  10
  11             Reviewed by Mark Lam.
  12
  13             Ideally, the ByteCodeParser would only emit SetArgument nodes for named arguments.  But
  14             currently that's not what it does - it emits a SetArgument for every argument that a varargs
  15             call may pass.  Each SetArgument gets turned into a GetStack.  This means that if
  16             ArgumentsEliminationPhase optimizes away PutStacks for those varargs arguments that didn't
  17             get passed or used, we get degenerate IR where we have a GetStack of something that didn't
  18             have a PutStack.
  19
  20             This fixes the bug by removing the code to optimize away PutStacks in
  21             ArgumentsEliminationPhase.
  22
  23             * dfg/DFGArgumentsEliminationPhase.cpp:
  24             * tests/stress/varargs-inlining-underflow.js: Added.
  25             (baz):
  26             (bar):
  27             (foo):
  28
  29 2015-07-24  Matthew Hanson  <matthew_hanson@apple.com>
  30
  31         Merge r187139. rdar://problem/21847618
  32
  33     2015-07-21  Filip Pizlo  <fpizlo@apple.com>
  34
  35             Unreviewed, fix a lot of tests. Need to initialize WTF threading sooner.
  36
  37             * jsc.cpp:
  38             (main):
  39
  40 2015-07-23  Lucas Forschler  <lforschler@apple.com>
  41
  42         Merge r187125
  43
  44     2015-07-21  Filip Pizlo  <fpizlo@apple.com>
  45
  46             Fixed VM pool allocation should have a reserve for allocations that cannot fail
  47             https://bugs.webkit.org/show_bug.cgi?id=147154
  48             rdar://problem/21847618
  49
  50             Reviewed by Geoffrey Garen.
  51
  52             This adds the notion of a JIT pool reserve fraction. Some fraction, currently 1/4, of
  53             the JIT pool is reserved for allocations that cannot fail. It makes sense to make this
  54             a fraction rather than a constant because each allocation that can fail may cause some
  55             number of allocations that cannot fail (for example, the OSR exit thunks that we
  56             compile when we exit from some CodeBlock cannot fail).
  57
  58             I've tested this by adding a test mode where we artificially limit the JIT pool size.
  59             Prior to the fix, we had >20 failures. Now we have none.
  60
  61             * heap/GCLogging.cpp:
  62             (WTF::printInternal): I needed a dump method on Options members when debugging this.
  63             * heap/GCLogging.h:
  64             * jit/ExecutableAllocator.h: Raise the ARM64 limit to 32MB because 16MB is cutting it too close.
  65             * jit/ExecutableAllocatorFixedVMPool.cpp:
  66             (JSC::FixedVMPoolExecutableAllocator::FixedVMPoolExecutableAllocator): Add the ability to artificially limit JIT pool size for testing.
  67             (JSC::ExecutableAllocator::memoryPressureMultiplier): Implement the reserve when computing memory pressure for JIT tier-up heuristics.
  68             (JSC::ExecutableAllocator::allocate): Implement the reserve when allocating can-fail things.
  69             * jsc.cpp: Rewire some options parsing so that CommandLine happens before we create the JIT pool.
  70             (main):
  71             (CommandLine::parseArguments):
  72             (jscmain):
  73             * runtime/Options.cpp:
  74             (JSC::OptionRange::dump): I needed a dump method on Options members when debugging this.
  75             (JSC::Options::initialize): This can now be called more than once.
  76             * runtime/Options.h:
  77
  78 == Rolled over to ChangeLog-2015-07-23 ==