LLVM with the emscripten fastcomp javascript backend https://git.amat.us/emscripten-fastcomp/atom/lib/Bitcode?h=1.12.1 2013-11-21T18:34:00Z 2013-11-21T18:34:00Z Karl Schimpf kschimpf@google.com 2013-11-21T18:34:00Z urn:sha1:7026af7138fccfb256456b04b375d39b025bdb7c Write out all of PNaCl bitcode headers using a single interface function. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3720 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/61753017 2013-11-21T17:55:24Z Karl Schimpf kschimpf@google.com 2013-11-21T17:55:24Z urn:sha1:86ff10150d76fa0db431a7c1c6db280f75977f35 The code previously was very inconsisted in modeling the number of bits needed for type ids within abbreviations. Frequently, the number of bits used was based on the total number of types. This is a bad choice in that this includes all function types, even though most cases do not use these types. This patch makes the selection more consistent. Also removing TYPE_CODE_ARRAY, since it is not allowed. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3720 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/68503018 2013-11-20T21:21:04Z Jan Voung jvoung@chromium.org 2013-11-20T21:21:04Z urn:sha1:32885ab95b44df7f2a3ace9ac7d227df9d648d46 This was kept in upstream LLVM for backwards compatibiilty with version 0 bitcode, but PNaCl only accepts version 1 bitcode, so UseRelativeID is always true. This reduces the number of branches taken while llvm-dis'ing the Ogre SampleBrowser from 2,680,184,988 branches to 2,670,798,955 branches. A 2.5% difference in wall-clock time... BUG=none trybots: http://chromegw.corp.google.com/i/tryserver.nacl/builders/nacl-toolchain-linux-pnacl-x86_32/builds/939 http://chromegw.corp.google.com/i/tryserver.nacl/builders/nacl-toolchain-linux-pnacl-x86_64/builds/991 http://chromegw.corp.google.com/i/tryserver.nacl/builders/nacl-toolchain-mac-pnacl-x86_32/builds/933 R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/77023007 2013-11-07T16:54:33Z Karl Schimpf kschimpf@google.com 2013-11-07T16:54:33Z urn:sha1:2796c5b8a31ea88cc24bf3449a95319928c93519 Factors out bitcode parser from pnacl-bcanalyzer, so that it can also be used for PNaCl bitcode to bitcode rewriters. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3720 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/48623003 2013-10-03T16:43:21Z Mark Seaborn mseaborn@chromium.org 2013-10-03T16:43:21Z urn:sha1:ed6b15283ff5f9d13cb48f88ba6e189be11291e8 The PNaCl bitcode writer had some complex logic for deciding when to omit casts in the output. This was left over from when the writer was trying to leave in the casts in some but not all cases (as part of incrementally removing casts). This is no longer needed now that the writer just omits all inttoptrs, all ptrtoints, and all pointer bitcasts. This cleanup also fixes the writer so that it elides an inttoptr of a ptrtoint. This sequence is allowed by the PNaCl ABI verifier, but never occurred in practice because ReplacePtrsWithInts' SimplifyCasts() function converts this sequence to an equivalent bitcast. Before this change, the writer would give this error for an inttoptr-of-ptrtoint: LLVM ERROR: Illegal (PNaCl ABI) pointer cast : %1 = ptrtoint [4 x i8]* @bytes to i32 BUG=https://code.google.com/p/nativeclient/issues/detail?id=3590 TEST=toolchain trybots Review URL: https://codereview.chromium.org/25817002 2013-10-03T15:08:25Z Mark Seaborn mseaborn@chromium.org 2013-10-03T15:08:25Z urn:sha1:8ed96f4d0ca9b8114875997f1f0196fc89a41a04 Make the reader and writer stricter so that we can be sure we're not accidentally generating ptrtoint or inttoptr instructions in pexe files. BUG=https://code.google.com/p/nativeclient/issues/detail?id=3590 TEST=toolchain trybots Review URL: https://codereview.chromium.org/25607006 2013-10-03T15:06:41Z Mark Seaborn mseaborn@chromium.org 2013-10-03T15:06:41Z urn:sha1:97e49363c618e0984b16bfaca3c5f60a97e442b8 This is a very Mac OS X-specific hack which isn't relevant to PNaCl. PNaCl doesn't use Xcode's ranlib for processing bitcode libraries. BUG=https://code.google.com/p/nativeclient/issues/detail?id=3590 TEST=toolchain trybots Review URL: https://codereview.chromium.org/25635004 2013-10-03T15:05:38Z Mark Seaborn mseaborn@chromium.org 2013-10-03T15:05:38Z urn:sha1:3b0e78e84065b446ec8a77f02bdbd5ac213af000 BUG=https://code.google.com/p/nativeclient/issues/detail?id=3590 TEST=build Review URL: https://codereview.chromium.org/25802002 2013-10-02T21:30:59Z Mark Seaborn mseaborn@chromium.org 2013-10-02T21:30:59Z urn:sha1:a85ac8f6999c04c5306efb5aa4fe77fc013c90e2 Declare NaClBitWriter's dependency on NaClBitReader. Without this change, pnacl-freeze fails at run time with: pnacl-freeze: symbol lookup error: .../native_client/pnacl/build/llvm_x86_64_ninja/lib/libLLVMNaClBitWriter.so: undefined symbol: _ZN4llvm17NaClBitcodeHeaderC1Ev BUG=none TEST=llvm-lit tests Review URL: https://codereview.chromium.org/24524003 2013-09-18T21:24:56Z Karl Schimpf kschimpf@google.com 2013-09-18T21:24:56Z urn:sha1:197cadba6026114ce62ed57dafe2125e70500a6b BUG= https://code.google.com/p/nativeclient/issues/detail?id=3590 R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/24232002