aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2013-07-29Fix formatting and remove TODO in PNaClLangRef.rstEli Bendersky
BUG=None R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/21011005
2013-07-29Updating the PNaCl reference manual to be in line with ABI verifierEli Bendersky
BUG=None R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/20862002
2013-07-29Remove handling of nongenerable instructions in PNaCl reader.Karl Schimpf
The PNaCl bitcode writer doesn't generate certain forms of instructions that the PNaCl bitcode writer accepts. Since these instructions do not get generated, remove them from the bitcode reader. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3590 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/20442002
2013-07-25Remove block address constants form PNaCl bitcode.Karl Schimpf
Removes block address constants because they are not part of the PNaClABI. Suggested in CL https://codereview.chromium.org/20172002 BUG= https://code.google.com/p/nativeclient/issues/detail?id=3590 R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/20402002
2013-07-25Remove dead instructions from the PNaCl bitcode reader/writer:Karl Schimpf
VAArg Invoke GetElementPtr LandingPad Resume IndirectBr ExtractElement InsertElement ShuffleVector ExtractValue InsertValue AtomicCmpXchg AtomicRMW BUG= https://code.google.com/p/nativeclient/issues/detail?id=3590 R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/20124002
2013-07-25Remove constant expressions and constant inline assembly from the PNaClKarl Schimpf
reader/writer, since they can't appear in PNaCl bitcode. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3590 R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/20172002
2013-07-25Clean some PNaCl-specific tests.Jim Stichnoth
BUG= https://code.google.com/p/nativeclient/issues/detail?id=3588 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/19606003
2013-07-25More pnacl-llc cleanups.Eli Bendersky
Focus on nacl_file and its interface with pnacl-llc.cpp; nacl_file as a name no longer makes sense, and neither do the comments describing it - so rename it. Also cleanup the interface a bit, and use more LLVM-y stuff instead of new data types. BUG=None R=dschuff@chromium.org, jfb@chromium.org Review URL: https://codereview.chromium.org/20186002
2013-07-24Hide the x86-64 sandbox base address.Jim Stichnoth
Prevent sandbox addresses from being written to the stack. This covers the following cases: 1. Function calls manually push a masked return address and jump to the target, rather than using the call instruction. 2. When the function prolog chooses to use a frame pointer (rbp), it saves a masked version of the old rbp. 3. Indirect branches (jumps, calls, and returns) uniformly use r11 to construct the 64-bit target address. 4. Register r11 is marked as reserved (similar to r15) so that the register allocator won't inadvertently spill a code address to the stack. These transformations can be disabled for performance testing with the flag "-sfi-hide-sandbox-base=false". BUG= https://code.google.com/p/nativeclient/issues/detail?id=1235 R=eliben@chromium.org, mseaborn@chromium.org Review URL: https://codereview.chromium.org/19505003
2013-07-24[MIPS] Fix LLVM merge issuesPetar Jovanovic
Several parts of the already in-place code have been omitted in the previous merge. These are: - missing lowering of Intrinsic::nacl_read_tp; - checks for forbidden instructiosn in branch-delay slots; - lowering operation for ISD::NACL_TP_TLS_OFFSET and ISD::NACL_TP_TDB_OFFSET. BUG= https://code.google.com/p/nativeclient/issues/detail?id=2275 TEST= run smoke tests R=eliben@chromium.org Review URL: https://codereview.chromium.org/19614006
2013-07-23More cleanups of pnacl-llc.Eli Bendersky
Simplifying code, style, removing unnecessary code & includes, etc. BUG=None R=jfb@chromium.org Review URL: https://codereview.chromium.org/19808004
2013-07-23Cherrypick test from r186226 - Fix ARM paired GPR COPY loweringJF Bastien
BUG= test the fix that was already cherrypicked TEST= self R=eliben@chromium.org Review URL: https://codereview.chromium.org/19704008
2013-07-22Unbreak the Windows/Cygwin toolchain build after the 3.3 merge.Eli Bendersky
Post 3.3-merge there seems to be a problem using mmap on cygwin. In the meantime, as a LOCALMOD, we disable usage of mmap in MemoryBuffer. The effect of this on the Windows/Cygwin toolchain can be a slightly slower developer-side linkage time. BUG=https://code.google.com/p/nativeclient/issues/detail?id=3578 R=dschuff@chromium.org Review URL: https://codereview.chromium.org/19923003
2013-07-22Cherrypick upstream ARM FastISel ext patchesJF Bastien
Specifically: r186489 - Fix ARMFastISel::ARMEmitIntExt shift emission r183794 - ARM FastISel fix sext/zext fold r183601 - Fix unused variable warning from my previous patch r183551 - ARM FastISel integer sext/zext improvements These should fix some failures that I had run into back then, as well as make ARM FastISel faster because it doesn't go to SelectionDAG. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3501 R=jvoung@chromium.org TEST= make check-all Review URL: https://codereview.chromium.org/19992002
2013-07-22Enable ability to have multiple PNaCl wire format versions.Karl Schimpf
Modifies PNaCl bitcode reader/writer to accept PNaClVersion=1 as supported, and all other versions are unsupported and unreadable. The PNaCl bitcode reader/writer will generate appropriate messages (including what version is unsupported if applicable). Also allows command-line option --pnacl-version for setting the PNaClVersion in the PNaCl bitcode writer. Also fixes some problems on PNaCl bitcode headers, using common support to determine when the read/written PNaCl bitcode file is valid. BUG=None R=jvoung@chromium.org Review URL: https://codereview.chromium.org/19400002
2013-07-22Fix default cases for printing instructions/constants in PNaCl wire format.Karl Schimpf
Fix printing of instructions/constants to explicitly check on cast opcodes and binary operators, so that we don't accidentally allow unknown opcodes. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3570 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/19242003
2013-07-22Remove module info records not allowed by PNaCl from the bitcode.Karl Schimpf
Remove dead code for: (1) Top-level inline assembly. (2) Remove code for named sections. (3) Remove code for GC names. (4) Global variable aliases. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3405 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/18770005
2013-07-22Disallow dev/private intrinsics by default (only llvm.nacl.target.arch remains)Jan Voung
BUG=https://code.google.com/p/nativeclient/issues/detail?id=3525 R=jfb@chromium.org Review URL: https://codereview.chromium.org/19668004
2013-07-20Remove prefetchJF Bastien
Following our discussion in the related bug, prefetch will not be part of our initial stable ABI. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3531 TEST= cd ./pnacl/build/llvm_x86_64; ninja check R=jvoung@chromium.org Review URL: https://codereview.chromium.org/19771015
2013-07-19Switch to the upstream way to measure IR parsing time, since we have 3.3 nowEli Bendersky
BUG=https://code.google.com/p/nativeclient/issues/detail?id=3349 R=stichnot@chromium.org Review URL: https://codereview.chromium.org/19500012
2013-07-19Clean up a debug printout left in by mistake during the 3.3 mergeEli Bendersky
BUG=None R=dschuff@chromium.org Review URL: https://codereview.chromium.org/19472003
2013-07-19Simplify the test by removing unnecessary function body.Eli Bendersky
Suggested in review: https://codereview.chromium.org/19705002/ BUG=None R=jvoung@chromium.org Review URL: https://codereview.chromium.org/19774012
2013-07-18Merge remote-tracking branch 'origin/master'Eli Bendersky
2013-07-18Some more OS-specific calls were added that are not supported in theEli Bendersky
sandboxed build. Condition them properly on __native_client__.
2013-07-18More safeguarding of Attr->getKindAsEnum by checking for isEnumAttributeEli Bendersky
first
2013-07-18Correctly adjust LOCALMOD for x64 frame register computation.Eli Bendersky
2013-07-18Adding checks for isEnumAttribute because getKindAsEnum now assertsEli Bendersky
2013-07-18Port new tests from origin/masterEli Bendersky
2013-07-18Increase coverage of NaCl-specific LLVM regression tests.Eli Bendersky
The issues these tests exercise came up in the recent merge of LLVM 3.3; Since it's considerably more difficult to debug problems in scons tests or sbtc tests, it makes sense to increase the coverate of LLVM regression tests as much as possible. These tests help resolve merging issues relatively quickly and should provide a first-line defense against problem arising in the NaCl specific LOCALMODs we have within LLVM. BUG=https://code.google.com/p/nativeclient/issues/detail?id=3578 R=jfb@chromium.org Review URL: https://codereview.chromium.org/19705002
2013-07-17Fixed constpool pattern matching problem that made the crtbeginS buildEli Bendersky
on ARM fail
2013-07-17Remove unnecessary debug printoutEli Bendersky
2013-07-17Disallow a global address in the x86-64 displacement field.Jim Stichnoth
This applies only to %r15 sandboxed memory references. The problem is that if the index register is negative, the sandboxing operation will cause the index to become a large positive 32-bit value, which combined with the displacement, will overflow and try to reference memory outside the sandbox. This situation may legitimately occur if the compiler happens to construct a (constant) interior pointer to the middle of the global struct/array, and then dereferences it with a variable offset. After this fix, pnacl/scripts/testsuite_known_failures_pnacl.txt can be updated to remove the "aha x86-64" known failure. BUG= https://code.google.com/p/nativeclient/issues/detail?id=3517 R=eliben@chromium.org Review URL: https://codereview.chromium.org/17987002
2013-07-16More tests passEli Bendersky
2013-07-16Fixing make check errors...Eli Bendersky
2013-07-16Remove useless field InstructionList from NaClBitcodeReader.Karl Schimpf
BUG=None R=mseaborn@chromium.org Review URL: https://codereview.chromium.org/19284013
2013-07-16Make it compileEli Bendersky
2013-07-15Trying to get the thing to copmile...Eli Bendersky
2013-07-15Merge commit '7dfcb84fc16b3bf6b2379713b53090757f0a45f9'Eli Bendersky
Conflicts: docs/LangRef.rst include/llvm/CodeGen/CallingConvLower.h include/llvm/IRReader/IRReader.h include/llvm/Target/TargetMachine.h lib/CodeGen/CallingConvLower.cpp lib/IRReader/IRReader.cpp lib/IRReader/LLVMBuild.txt lib/IRReader/Makefile lib/LLVMBuild.txt lib/Makefile lib/Support/MemoryBuffer.cpp lib/Support/Unix/PathV2.inc lib/Target/ARM/ARMBaseInstrInfo.cpp lib/Target/ARM/ARMISelLowering.cpp lib/Target/ARM/ARMInstrInfo.td lib/Target/ARM/ARMSubtarget.cpp lib/Target/ARM/ARMTargetMachine.cpp lib/Target/Mips/CMakeLists.txt lib/Target/Mips/MipsDelaySlotFiller.cpp lib/Target/Mips/MipsISelLowering.cpp lib/Target/Mips/MipsInstrInfo.td lib/Target/Mips/MipsSubtarget.cpp lib/Target/Mips/MipsSubtarget.h lib/Target/X86/X86FastISel.cpp lib/Target/X86/X86ISelDAGToDAG.cpp lib/Target/X86/X86ISelLowering.cpp lib/Target/X86/X86InstrControl.td lib/Target/X86/X86InstrFormats.td lib/Transforms/IPO/ExtractGV.cpp lib/Transforms/InstCombine/InstCombineCompares.cpp lib/Transforms/Utils/SimplifyLibCalls.cpp test/CodeGen/X86/fast-isel-divrem.ll test/MC/ARM/data-in-code.ll tools/Makefile tools/llvm-extract/llvm-extract.cpp tools/llvm-link/CMakeLists.txt tools/opt/CMakeLists.txt tools/opt/LLVMBuild.txt tools/opt/Makefile tools/opt/opt.cpp
2013-07-15Clean up some duplicated and/or unused code in pnacl-llc.Eli Bendersky
No functionality change. BUG=None R=jvoung@chromium.org Review URL: https://codereview.chromium.org/19235002
2013-07-15Run the LLVM IR verifier just once in pnacl-llc.Eli Bendersky
By default, using the path inherited from llc, the verifier is run twice. We only need it to run once right after reading the bitcode in and before the ABI verifier runs. BUG=https://code.google.com/p/nativeclient/issues/detail?id=3553 R=jvoung@chromium.org Review URL: https://codereview.chromium.org/18920004
2013-07-13Concurrency support for PNaCl ABIJF Bastien
Add portable support for concurrency in PNaCl's ABI: - Promote volatile to atomic. - Promote all memory ordering to sequential consistency. - Rewrite all atomic operations to frozen NaCl intrinsics for pexe. - Rewrite atomic intrinsics to LLVM instructions for translation. This change also adds documentation to the PNaCl language reference, as well as tests where it makes sense. A future CL could clean up more of our code which mentions atomics, volatiles, memory orderings. Multiple reviewers because this is a big patch: - eliben: LLVM-fu and ResolvePNaClIntrinsics. - dschuff: ABI stability. - mseaborn: ABI stability. - sehr: Tron-duty (fight for the user's programs to work). BUG= https://code.google.com/p/nativeclient/issues/detail?id=3475 R=dschuff@chromium.org, eliben@chromium.org, sehr@google.com TEST= (cd ./pnacl/build/llvm_x86_64; ninja check-all) && ./pnacl/test.sh test-x86-32 && ./pnacl/test.sh test-x86-64 && ./pnacl/test.sh test-arm && ./pnacl/test.sh test-x86-32-sbtc && ./pnacl/test.sh test-x86-64-sbtc && ./pnacl/test.sh test-arm-sbtc Review URL: https://codereview.chromium.org/17777004
2013-07-11Fix ARM paired GPR COPY loweringJF Bastien
ARM paired GPR COPY was being lowered to two MOVr without CC. This patch puts the CC back. I sent this patch upstream (with a test) but haven't received a review yet. This seems like a simple oversight in the code, and is holding my atomics patch so I'd like to get it into our repo. R=dschuff@chromium.org TEST= ./scons run_llvm_bitmanip_intrinsics_test platform=arm BUG= no CC on MOVr Review URL: https://codereview.chromium.org/18047006
2013-07-11Sandbox ARM load/store exclusive dualJF Bastien
PNaCl's LLVM sandboxing wasn't correct for ARM load/store exclusive dual. I encountered this while running our testsuite with my atomic changes: the tests which use volatile 64-bit values started failing validation. R=dschuff@chromium.org BUG= validation failure TEST= ./pnacl/test.sh test-arm Review URL: https://codereview.chromium.org/18978015
2013-07-10Fix ARM LDMIA MI.JF Bastien
The instruction that was generated for paired register stack slot load was wrong. This was fixed by Tim Northover in LLVM 3.3 commit 179977, but his patch does much more and doesn't apply as-is to our tree. I'll therefore punt applying the full patch to when we rebase to 3.3. I encountered the issue while working on atomics (64-bit atomics require paired registers on ARM), and saw the 3.3 fix when I tried upstreaming my fix. BUG= non TEST= ./pnacl/test.sh test-arm R=eliben@chromium.org Review URL: https://codereview.chromium.org/18699004
2013-07-10Reject the llvm.powi intrinsic in the PNaCl ABI verifier.Eli Bendersky
BUG=https://code.google.com/p/nativeclient/issues/detail?id=3524 R=jfb@chromium.org Review URL: https://codereview.chromium.org/18614013
2013-07-10Add option to pnacl-abicheck to handle pnacl bitcodeDerek Schuff
This allows running the utility on finalized/frozen pexes. R=kschimpf@google.com BUG=none Review URL: https://codereview.chromium.org/18884003
2013-07-09Make GlobalOpt's GV-by-alloca replacement work for PNaCl.Eli Bendersky
GlobalOpt currently assumes only an external "main" is the "real main". This is no longer the case for PNaCl, where we internalize "main". Make the test more strict and PNaCl specific by checking that "main" is just used once - in a call from "_start", but does not have to be external. Note that this also addresses a possible bug in the optimization for C code, since C does not guarantee that main is not recursive. This CL's purpose is to address a SPEC performance regression - 10% in 183.equake. The regression appeared after our ABI change that made 'main' internal, which disabled this particular optimization. The CL addresses this by re-enabling the optimization and also being more C-standard conforming. BUG=None Review URL: https://codereview.chromium.org/18615015
2013-07-09PNaCl: Fix negative relocation addends on x86-32Mark Seaborn
Disable an assertion. This assertion made the behaviour on x86-32 inconsistent with x86-64 and ARM. BUG=https://code.google.com/p/nativeclient/issues/detail?id=3548 TEST=*.ll tests + PNaCl toolchain trybots Review URL: https://codereview.chromium.org/18261008
2013-07-09Prune some arch-specific MC InstPrinter (.s printing), from sandboxed ↵Jan Voung
translator. Saves about 100KB per arch. As noted in the test, some of the functions like getRegisterName() are not pruned. Common infrastructure code may still be sitting around, only the ${ARCH}InstPrinter::* is touched... but this is just a one-liner. BUG=https://code.google.com/p/nativeclient/issues/detail?id=1222 TEST=https://codereview.chromium.org/18333008/ R=dschuff@chromium.org Review URL: https://codereview.chromium.org/18238011
2013-07-08Fix for a regression caused by the LoopVectorizer whenTom Stellard
vectorizing loops with memory accesses to non-zero address spaces. It simply dropped the AS info. Fixes PR16306. Merged from r184103 Author: Pekka Jaaskelainen <pekka.jaaskelainen@tut.fi> Date: Mon Jun 17 18:49:06 2013 +0000 git-svn-id: https://llvm.org/svn/llvm-project/llvm/branches/release_33@185869 91177308-0d34-0410-b5e6-96231b3b80d8
generated by cgit v1.2.3-18-g5258 (git 2.32.0) at 2025-05-04 04:23:26 +0000