http://llvm.org https://git.amat.us/llvm/atom/lib?h=release_26 2009-10-12T16:46:55Z 2009-10-12T16:46:55Z Tanya Lattner tonic@nondot.org 2009-10-12T16:46:55Z urn:sha1:901cb73cc5640a6405e7d7f8f5d4b8627482a04f r83391 was completely broken since Twines keep references to their inputs, and some of the inputs were temporaries. Here's a real fix for the miscompilation. Thanks to sabre for pointing out the problem. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@83859 91177308-0d34-0410-b5e6-96231b3b80d8 2009-10-12T16:45:48Z Tanya Lattner tonic@nondot.org 2009-10-12T16:45:48Z urn:sha1:f97bad7a6f86eb2c281e536f6df98a677165c997 Fix PR5112, a miscompilation on gcc-4.0.3. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@83858 91177308-0d34-0410-b5e6-96231b3b80d8 2009-10-01T15:12:41Z Tanya Lattner tonic@nondot.org 2009-10-01T15:12:41Z urn:sha1:e07cc230d9f5997fa5faa23fd9203a8429175325 Don't constant propagate byval pointers, since they are not really pointers, but rather structs passed by value. This fixes PR5038. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@83200 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-21T03:01:01Z Tanya Lattner tonic@nondot.org 2009-09-21T03:01:01Z urn:sha1:f2641fb4d0090497b8a757581718c218d33f64f3 git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@82428 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-16T01:15:44Z Tanya Lattner tonic@nondot.org 2009-09-16T01:15:44Z urn:sha1:1d20c4e51ad24e6747da643afff71d2d0d3e9d91 Rework getPersonalityIndex slightly - 0 is now a valid and not-NULL personality function. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@81982 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-16T01:14:55Z Tanya Lattner tonic@nondot.org 2009-09-16T01:14:55Z urn:sha1:15a6377b2927930a13ddf7e359ec190ef6819b1e Change the marker byte for stubs from 0xcd to 0xce (another form of interrupt instruction, which shouldn't arise any other way). 0xcd is also used by JITMemoryManager to initialize the buffer to garbage, which means it could appear following a noreturn call even when that is not a stub, confusing X86CompilationCallback2. PR 4929. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@81981 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-16T01:13:59Z Tanya Lattner tonic@nondot.org 2009-09-16T01:13:59Z urn:sha1:b580bd1eb2cedb51fdc2944ff7a42d9ac6898f4f fix PR4963: folding insertvalue would sometimes turn a packed struct into an unpacked one. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@81980 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-16T01:13:06Z Tanya Lattner tonic@nondot.org 2009-09-16T01:13:06Z urn:sha1:565fb5a655d27ae3cb6d9297b33976c9eca7fa3e Don't pull a load through a callseq_start if the load's chain has multiple uses, as one of the other uses may be on a path to a different node above the callseq_start, because that leads to a cyclic graph. This problem is exposed when -combiner-global-alias-analysis is used. This fixes PR4880. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@81978 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-16T01:12:05Z Tanya Lattner tonic@nondot.org 2009-09-16T01:12:05Z urn:sha1:75a14269961ef14ac41fbdf10a062557b0800115 On x86-64, the 32-bit cmov doesn't actually clear the high 32-bit of its result if the condition is false. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@81977 91177308-0d34-0410-b5e6-96231b3b80d8 2009-09-16T01:10:57Z Tanya Lattner tonic@nondot.org 2009-09-16T01:10:57Z urn:sha1:dc10e65ed6ba47039b385aa11e267e5edb9a9065 When extending a memset range past the front, set the alignment of the memset region to the alignment of the new start address. git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_26@81976 91177308-0d34-0410-b5e6-96231b3b80d8