similar to: [LLVMdev] Clang target specific test case question

Attached are the promised patches for the below proposed change. Cheers, Jack ________________________________________ From: Carter, Jack Sent: Tuesday, December 11, 2012 1:33 PM To: Jim Grosbach Cc: Rafael Espíndola; List Subject: RE: [LLVMdev] [MC] [llvm-mc] Getting target specific information to <target>ELFObjectWriter Jim, You are correct: MipsSubtarget. For llvm-mc we have a

On Dec 10, 2012, at 1:15 PM, "Carter, Jack" <jcarter at mips.com> wrote: > Here are some examples using the gnu assembler reacting to the same input file with different commandline options. > > These are using the GCC assembler on hello.c > // abi o32, arch mips32r2, relocation model pic+cpic > mips-linux-gnu-as -mips32r2 -EL -KPIC -o hello_gas.o hello_gas.s >

Jim, You are correct: MipsSubtarget. For llvm-mc we have a straight MCSubtargetInfo object. For llc we get a MipsSubtarget object which derives from MipsGenSubtargetInfo which derives from TargetSubtargetInfo which derives from MCSubtargetInfo. The patch I hope to send out for review will do this: Add a new data member to MCSubtargetInfo base class. It will be a set of integers that is used or

On 6 December 2012 17:49, Carter, Jack <jcarter at mips.com> wrote: > Older targets like Mips had/have assemblers and ABIs that carry a lot of > baggage. > > The small bit of baggage that is giving me fits is that MipsELFObjectWriter > needs to know the relocation model (static,pic,cpic), whether we are using > xgot (-mgot), which abi (old,new), which architecture

Hi Rafael, There are a lot of flags. Here are the ones you ask about: -KPIC, -call_shared generate SVR4 position independent code -call_nonpic generate non-PIC code that can operate with DSOs -mvxworks-pic generate VxWorks position independent code -non_shared do not generate code that can operate with DSOs -xgot assume a 32 bit GOT Just to make things fun, the SGI notion of cpic (call pic)

Here are some examples using the gnu assembler reacting to the same input file with different commandline options. These are using the GCC assembler on hello.c // abi o32, arch mips32r2, relocation model pic+cpic mips-linux-gnu-as -mips32r2 -EL -KPIC -o hello_gas.o hello_gas.s e_flags 0x70001007 EF_MIPS_NOREORDER EF_MIPS_PIC EF_MIPS_CPIC E_MIPS_ABI_O32 EF_MIPS_ARCH_32R2 // abi

On Tue, Dec 11, 2012 at 2:48 PM, Carter, Jack <jcarter at mips.com> wrote: > Attached are the promised patches for the below proposed change. > Just a quick question from an initial review: isn't the int->bool mapping of flags a bit limiting. Flag can have actual values and not only be there or not be there. Wouldn't a more generic mapping (string->string ?) be more

Eli, This is the kind of feedback I want. I believe I have to add to the base class so it should be generally useful. I can see string being better for the value. I still am enamoured with an enumeration for the tab though: int->string. How would that be a limitation? How about the rest of the patch? I appreciate the feedback, Jack ________________________________________ From: Eli

On 7 December 2012 18:57, Carter, Jack <jcarter at mips.com> wrote: > Hi Rafael, > > There are a lot of flags. Here are the ones you ask about: > > -KPIC, -call_shared generate SVR4 position independent code > -call_nonpic generate non-PIC code that can operate with DSOs > -mvxworks-pic generate VxWorks position independent code > -non_shared

Eli, Yes, SubtargetInfo is more of a container of convenience since it is available to all the assemblers. Working with the current framework it seemed the least disruptive. I'll describe the problem again. The Mips ABI for better or worse, uses the ELF headers e_flags extensively. The most pressing issue is the need to post the relocation model such as PIC, CPIC or non-shared. The object

Older targets like Mips had/have assemblers and ABIs that carry a lot of baggage. The small bit of baggage that is giving me fits is that MipsELFObjectWriter needs to know the relocation model (static,pic,cpic), whether we are using xgot (-mgot), which abi (old,new), which architecture (32r[123],64[123]), which if any coprocessor or extention instructions are used (mips16,micromips,etc.). I

On Fri, Dec 14, 2012 at 1:03 PM, Carter, Jack <jcarter at mips.com> wrote: > Eli, > > This is the kind of feedback I want. I believe I have to add to the base class so it should be generally useful. I can see string being better for the value. I still am enamoured with an enumeration for the tab though: int->string. How would that be a limitation? > I guess that's fine,

On Dec 6, 2011, at 7:43 PM, Eli Friedman wrote: > On Tue, Dec 6, 2011 at 7:29 PM, Carter, Jack <jcarter at mips.com> wrote: >> Is there an official llvm method of creating and submitting test cases that >> don't affect .s assembly files? >> >> When we check in changes that can affect the .s output we submit .ll files >> with the internalized semicolon

On Tue, Dec 6, 2011 at 7:29 PM, Carter, Jack <jcarter at mips.com> wrote: > Is there an official llvm method of creating and submitting test cases that > don't affect .s assembly files? > > When we check in changes that can affect the .s output we submit .ll files > with the internalized semicolon instructions on how to check the output .s > file. > > For direct

Let me see if I understand the response ;-) When you are saying integrated assembler do you mean llc --filetype=obj? If so, we currently have that for an option when running the test-suite. When you say that to test the llvm-mc assembler for your target you don't substitute the gcc assembler invocation for llvm-mc which would expect the resultant executable run to pass. Instead you have to

I would like to add some subdirectories to test-suite SingleSource/UnitTests/Mips MultiSource/UnitTests/Mips There are a lot of Mips specific tests. For Mips16/32 interoperability, some of these require multiple source files. The corresponding Makefiles would be changed to add these directories when $(ARCH)=MIPS We , at Mips, would take ownership of these subdirectories so we can add tests

Hi Jack, You're running into a fundamental problem with the current table generated asmmatcher. Specifically, wants to believe that assembly parsing is context insensitive, or at least close enough that operands can be parsed w/o knowing the context of the instruction. Its idea is to use the operand types to disambiguate which instruction should be selected. It sounds like MIPS 64vs.32 does

On Mon, Oct 15, 2012 at 2:22 PM, Carter, Jack <jcarter at mips.com> wrote: > Has anyone converted llvm/projects/test-suite to use the llvm assembler > instead of gcc? > > If so, what was needed to change and how? > > My assumption is that this would be a good way to test the llvm assembler. > Not quite sure what you mean, as far as I know there isn't any assembler

Hi Jack, On Jan 25, 2012, at 6:45 PM, "Carter, Jack" <jcarter at mips.com> wrote: > I'm trying to generate MipsGenAsmMatcher.inc for MipsAsmParser.cpp. > > What added restrictions for the .td file contents are there for tblgen -gen-asm-matcher? > Lots, as you're finding, almost all of them completely undocumented. :( > For the Mips platform we create

Yes, absolutely. There's two pieces of this that are handy. First, checking the normal integrated-assembler code path. That doesn't check the actual assembler, but rather the binary encoder and object file emitter. To test that, I did runs with a locally modified clang that enabled the integrated assembler by default for my target (ARM/Darwin at the time). The second piece is checking the