similar to: [GlobalISel] [X86] unable to legalize instruction

Thanks for your quick reply. Here its is: =========== ; ModuleID = 'simple_foo.c' source_filename = "simple_foo.c" target datalayout = "e-m:o-i64:64-f80:128-n8:16:32:64-S128" target triple = "x86_64-apple-macosx10.13.0" ; Function Attrs: noinline nounwind optnone ssp uwtable define i32 @foo(i32 %a, i32 %b, i32 %c) #0 { entry: %retval = alloca i32, align

I believe if you pass(iirc) -pass-remarks-missed=“gisel-*”, it’ll print the instruction it failed to legalize. Sent from my iPhone > On Oct 11, 2017, at 6:44 PM, S. Bharadwaj Yadavalli via llvm-dev <llvm-dev at lists.llvm.org> wrote: > > Thanks for your quick reply. > > Here its is: > > =========== > > ; ModuleID = 'simple_foo.c' > source_filename

Hi , Currently GobalIsel like FastIsel designed to use fallback to DAGIsel in case of failure. You can use -global-isel-abort=2 option. llc -global-isel -pass-remarks-missed="gisel-*" -global-isel-abort=2 simple_foo.ll Regards, Igor From: S. Bharadwaj Yadavalli [mailto:bharadwajy at gmail.com] Sent: Thursday, November 09, 2017 03:36 To: Aditya Nandakumar <proaditya at

Hi, Mips32 has 64 bit floating point instructions, while i64 instructions have to be emulated with i32 instructions. This means that G_LOAD should be custom legalized for s64 integer value, and be legal for s64 floating point value. There are also other generic instructions with the same problem: G_STORE, G_SELECT, G_EXTRACT, and G_INSERT. There are also other configurations where integer

I am trying to categorize the machine instructions based on associated static (i.e., as encoded in .td file) machine description and the corresponding APIs. I would like to perform appropriate actions based on the kind of instruction in a tool that I am working on. For example, I'd like to distinguish between memop instructions involving immediate vs register. While it appears that I would be

Thanks for your reply, Paul. That seems to be a plausible explanation. I ended up adding code in my project to massage the output of llvm-config to the right paths based of the values of CMAKE_GENERATOR and build-mode. Bharadwaj On Dec 15, 2017 9:46 AM, "Robinson, Paul" <paul.robinson at sony.com> wrote: It may be that llvm-config simply doesn't understand a

Hi, I had (naively?) expected that the instruction to move immediate to register or memory (such as MOV32mi, MOV32ri, MOV64mi32, MOV64ri, MOV64ri32) would be marked with the flag MCID::MovImm via the X86InstrInfo.td (and hence in the generated X86GenInstrInfo.inc). I do not see that to be the case. Can someone please tell me if my expectation is flawed? Is there a better/different way to

I think even if we did use it, MoveImmediate is intended for instructions that move an immediate into a register rather than into memory. It's supposed to indicate instructions that can be folded with the user of the register by changing the user to an immediate instruction. And it wouldn't be set on an instruction like "addl $0, %eax" or "addl $0, (%ecx)" either since

Hi Duncan, Thanks for your reply. I did try the method suggested by you. I got a foo.bc with -emit-llvm switch. llc on foo.bc gets me the same assertion failure, as you pointed out. Then I ran bugpoint to generate bugpoint-reduced-simplified.bc and bugpoint-reduced-function.bc. My previous experience is that if I compile either if them, I would get the same failure as I did with foo.bc. That is

Hi, Is there some place I can get instructions to load, build and debug LLVM and/or Clang using XCode on Mac OSX? Sorry if this is an FAQ - my (possibly quick) search did not yield an answer. Thanks, Bharadwaj

Hi, Sorry for the newbie question. I downloaded llvm and tried out the simple "Hello, World" program but got the following error. What am I missing? I am running RHAS 3 Update 4 with GCC 3.2.3. Thanks, Bharadwaj $ ./hello lli: /home/proj/skokomish/syadaval/ia32/Sandbox/llvm/lib/Target/X86/X86CodeEmitter.cpp:208: unsigned char ModRMByte(unsigned int, unsigned int, unsigned int):

Hi, I configured and built LLVM as follows on Mac OS 10.13.2: $ pwd <llvm-build-dir> $ cmake -G "Xcode" -DCMAKE_INSTALL_PREFIX=<llvm-install-dir> -DLLVM_ENABLE_PIC=0 <llvm-src-dir>/llvm $ xcodebuild -project LLVM.xcodeproj All binaries are in <llvm-build-dir>/Debug/bin. I am interested in using the output of llvm-config. Now, when I run $

I am trying to build a x86_64 to ARM cross compiler. I configured, built and installed LLVM as follows: $ ../../../src/llvm/configure --with-llvmgccdir=<llvm-root>/install/x86_64-arm/llvm-gcc-4.2 --enable-optimized --enable-jit --prefix=<llvm-root>/install/x86_64-arm/llvm --target=arm-unknown-linux-gnueabi $ make $ make install Then I configured gcc front end as follows: $

Hi, I built LLVM + Clang on Mac OS X and ran make check. I get the following result summary: Failing Tests (11): LLVM :: LLVMC/C++/dash-x.cpp LLVM :: LLVMC/C++/hello.cpp LLVM :: LLVMC/C++/just-compile.cpp LLVM :: LLVMC/C++/together.cpp LLVM :: LLVMC/C++/unknown_suffix.unk LLVM :: LLVMC/C/hello.c LLVM :: LLVMC/C/opt-test.c LLVM :: LLVMC/C/sink.c LLVM ::

I would like to generate the tags for LLVM sources to be used by emacs. I noticed that there is a tags target that can be specified to make to generate the tags file. It appears that the rules for this target are generated only with the specification of --with-tags option to configure script. However, I get the following errors respectively when I specify --with-tags or --with-tags=TAGS option to

Hi, I notice that the following recent addition X86::CondCode X86::getCondFromBranch(const MachineInstr &MI) { switch (MI.getOpcode()) { default: return X86::COND_INVALID; case X86::JCC_1: return static_cast<X86::CondCode>( MI.getOperand(MI.getDesc().getNumOperands() - 1).getImm()); } } returns an invalid condition for JCC_2 and JCC_4 conditional opcodes. What is

That flag is specifically used by the foldImmediate optimization in the Peephole pass. We don't implement the TLI foldImmediate hook the peephole pass uses on x86 so we have no reason to set the flag on any instructions What are you trying to do? ~Craig On Fri, Feb 9, 2018 at 11:45 AM, S. Bharadwaj Yadavalli via llvm-dev < llvm-dev at lists.llvm.org> wrote: > Hi, > > I had

Luke Dalessandro wrote: > S. Bharadwaj Yadavalli wrote: >> $ ../../../src/llvm-gcc-4.2/configure >> --prefix=<llvm-root>/install/x86_64-arm/llvm-gcc-4.2 >> --program-prefix=llvm-x86_64-arm >> --enable-llvm=<llvm-root>/install/x86_64-arm/llvm >> --enable-languages=c,c++ --disable-multilib >> --target=arm-unknown-linux-gnueabi > > I

On Mar 4, 2009, at 1:33 PM, S. Bharadwaj Yadavalli wrote: > Hi Duncan, > > Thanks for your reply. > > I did try the method suggested by you. I got a foo.bc with -emit- > llvm switch. llc on foo.bc gets me the same assertion failure, as > you pointed out. > > Then I ran bugpoint to generate bugpoint-reduced-simplified.bc and > bugpoint-reduced-function.bc. My

On Mon, Sep 5, 2011 at 6:16 PM, S. Bharadwaj Yadavalli <bharadwajy at gmail.com> wrote: > Hi, > > Is there some place I can get instructions to load, build and debug > LLVM and/or Clang using XCode on Mac OSX? > > Sorry if this is an FAQ - my (possibly quick) search did not yield an answer. You can generate an XCode project using CMake. See