llvm/lib/CodeGen, branch release_28 http://llvm.org https://git.amat.us/llvm/atom/lib/CodeGen?h=release_28 2010-09-17T22:09:07Z Approved by Evan 2010-09-17T22:09:07Z Bill Wendling isanbard@gmail.com 2010-09-17T22:09:07Z urn:sha1:71f7f73dfa15e7f23630691778ee4d51686b552b $ svn merge -c 114220 https://llvm.org/svn/llvm-project/llvm/trunk --- Merging r114220 into '.': U lib/CodeGen/MachineCSE.cpp $ svn merge -c 114222 https://llvm.org/svn/llvm-project/llvm/trunk --- Merging r114222 into '.': G lib/CodeGen/MachineCSE.cpp git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_28@114224 91177308-0d34-0410-b5e6-96231b3b80d8 $ svn merge -c 113848 https://llvm.org/svn/llvm-project/llvm/trunk 2010-09-15T00:32:40Z Bill Wendling isanbard@gmail.com 2010-09-15T00:32:40Z urn:sha1:ed70cc3e1ba3b7b511c1bd6d3358ba58cc0eae63 --- Merging r113848 into '.': U lib/CodeGen/PeepholeOptimizer.cpp must not peephole away side effects git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_28@113913 91177308-0d34-0410-b5e6-96231b3b80d8 Approved by Evan: 2010-09-09T20:13:48Z Bill Wendling isanbard@gmail.com 2010-09-09T20:13:48Z urn:sha1:587f6ed6864af95997e16df8dcf930822c1416a2 $ svn merge -c 113299 https://llvm.org/svn/llvm-project/llvm/trunk --- Merging r113299 into '.': U lib/CodeGen/VirtRegRewriter.cpp git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_28@113531 91177308-0d34-0410-b5e6-96231b3b80d8 previous patch was a little too tricky for its own good. Don't try to 2010-09-03T21:45:15Z Jim Grosbach grosbach@apple.com 2010-09-03T21:45:15Z urn:sha1:ee72651df4b783c973bb682bef7eab2ff9a703e2 overload UserInInstr. Explicitly check Allocatable. The early exit in the condition will mean the performance impact of the extra test should be minimal. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@113016 91177308-0d34-0410-b5e6-96231b3b80d8 Add a missing check when legalizing a vector extending load. This doesn't 2010-09-03T19:20:37Z Bob Wilson bob.wilson@apple.com 2010-09-03T19:20:37Z urn:sha1:ef134d02e0036ef981983d4b5e40c678dd241583 solve the root problem, but it corrects the bug in the code I added to support legalizing in the case where the non-extended type is also legal. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112997 91177308-0d34-0410-b5e6-96231b3b80d8 VirtRegRewriter checks for early clobbers before it reuses an available stack 2010-09-03T18:36:56Z Jakob Stoklund Olesen stoklund@2pi.dk 2010-09-03T18:36:56Z urn:sha1:1e3aeefc088db2fe32958eb17f46e60011d34949 slot. Teach it to also check for early clobbered aliases, and early clobber operands following the current operand. This fixes the miscompilation in PR8044 where EC registers eax and ecx were being used for inputs. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112988 91177308-0d34-0410-b5e6-96231b3b80d8 Reapply commit 112702 which was speculatively reverted by echristo. 2010-09-03T08:31:48Z Duncan Sands baldrick@free.fr 2010-09-03T08:31:48Z urn:sha1:850fcd4705177d8630530ea6a18538782eb68a39 Original commit message: Use the SSAUpdator to turn calls to eh.exception that are not in a landing pad into uses of registers rather than loads from a stack slot. Doesn't touch the 'orrible hack code - Bill needs to persuade me harder :) git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112952 91177308-0d34-0410-b5e6-96231b3b80d8 There is no need to use .set here. 2010-09-02T23:01:10Z Devang Patel dpatel@apple.com 2010-09-02T23:01:10Z urn:sha1:a660be7587f52d269135a54e2d485bf7a7bc9c50 Thanks Chris! git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112900 91177308-0d34-0410-b5e6-96231b3b80d8 Detect undef value early and save unnecessary NodeMap query. 2010-09-02T21:29:42Z Devang Patel dpatel@apple.com 2010-09-02T21:29:42Z urn:sha1:3f74a116c80b483ffc7c2e6922ff62f0f73d83f5 git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112864 91177308-0d34-0410-b5e6-96231b3b80d8 Don't narrow the load and store in a load+twiddle+store sequence unless 2010-09-02T21:18:42Z Dan Gohman gohman@apple.com 2010-09-02T21:18:42Z urn:sha1:24bde5bce192119ee0fc4f94ef8757fd4031e5f6 there are clearly no stores between the load and the store. This fixes this miscompile reported as PR7833. This breaks the test/CodeGen/X86/narrow_op-2.ll optimization, which is safe, but awkward to prove safe. Move it to X86's README.txt. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@112861 91177308-0d34-0410-b5e6-96231b3b80d8