similar to: [LLVMdev] modifiy the address of GlobalVariable emitted by JIT

On Friday 11 July 2008 12:12, David Greene wrote: > On Friday 11 July 2008 12:05, Devang Patel wrote: > > > Ok, I've mostly got a mechanism to do what I want: > > > > > > 1. As each function comes in for op/codegen, clone it and save off > > > the clone and its associated ValueMap (I call these clones > > > "pristine" > > >

I have a simple C++ class that looks like this: struct Foo { Thing *first; Blob *second; unsigned long third; }; Then I have a derived C++ class that look like this: struct Bar : Foo { long long fourth; } I generate JIT code to access 'fourth' from a Foo * as follows: BarPtr = CreateBitCast(FooPtr, BarPtrTy); FourthPtr = CreateConstGEP2_32(BarPtr, 0, 3); FourthValue =

On Jul 23, 2009, at 12:10 PM, robert muth wrote: > Bob: > > Thanks for cleaning this up. I like the new patch much better than > the old one. > Teaching the (abstract) ConstantValue class about jumptable indices > is a little > bit ugly but I do not see any better solution without massive > refactoring. > I have added TODOs here and elsewhere and plan to address

On Tue, Jul 14, 2009 at 6:48 PM, Bob Wilson <bob.wilson at apple.com> wrote: > > On Jul 2, 2009, at 10:48 AM, robert muth wrote: > > I spend over a day trying to follow your suggestion. In the end I was not > successful. Here is what Iearned: > > After setting > > ARMJITInfo::hasCustomJumpTables -> true > setOperationAction for ISD::BR_JT -> Expand >

On Friday 11 July 2008 12:05, Devang Patel wrote: > > Ok, I've mostly got a mechanism to do what I want: > > > > 1. As each function comes in for op/codegen, clone it and save off > > the clone and its associated ValueMap (I call these clones > > "pristine" > > functions. > > > > 2. After all processing is done, clone the resulting

On Jul 11, 2008, at 9:59 AM, David Greene wrote: > On Wednesday 09 July 2008 13:49, David Greene wrote: > >>> then it seems you're doing >>> >>> for each function >>> generate_ir >>> convert_to_llvm_ir >>> optimize_llvm_ir >> >> Yep. > > Ok, I've mostly got a mechanism to do what I want: > > 1. As

I want to add just one more function to an already installed, locally developed site-library package. Is there a good tutorial on this laying about? From what I see the new function would touch several files and directories in the existing package. Tools to automate that ?

Hi, to install current PC models (with new Intel NICs) via Kickstart/PXE, I wanted to add the newest e1000e-Treiber to initrd.img. With this modified image, the Kickstart kernel crashes with the following error messages: [...] md: Autodetecting RAID arrays. md: autorun ... md: ... autorun DONE. RAMDISK: Compressed image found at block 0 RAMDISK: incomplete write (20480 != 32768) 6062080

Hi Vivien, I'm not familiar with Windows development, but I believe you could dump the object file to disk then run LINK.EXE to produce the PDB. Alternatively, I think LLD can also produce PDB files (Rui -- is that right?), in which case you may be able to use that. Is the aim to be able to debug JIT'd code? Which debugger do you plan to use? -- Lang. On Thu, Mar 8, 2018 at 4:23 AM,

When constant folding a MAD operation, we first fold the multiply and generate an ADD. However we do so without making sure that the immediate can be handled in the saturate case. If it can't, load the immediate in a separate instruction. Reported-by: Tiziano Bacocco <tizbac2 at gmail.com> Signed-off-by: Ilia Mirkin <imirkin at alum.mit.edu> Cc: "10.2 10.3"

Hi everybody, I'm toying with the Kaleidoscope example from the documentation (with llvm 2.5) and I want to print out on stderr, the native code (X86) emitted with the JIT. I have found the "createMachineFunctionPrinterPass" function. But when I add it to the FunctionPassManager ( OurFPM.add(createMachineFunctionPrinterPass(cerr, "")); ). I get an error at runtime :

All, I'm working on a new scheduler and have encountered this assert isNew && "Node emitted out of order - early" I believe that it occurs when I clone a compare/test instruction. I've looked at the SUnit graph and the units are scheduled correctly. Any advice is appreciated. Aran

Hi Lorenzo, > Thanks to everyone for the quick replies! I spent some time looking > into the issue. It turns out that llvm-dis crashes on CLANG-generated > bytecode if LLVM is compiled for a 64-bit architecture. The problem > disappears when compiling for a 32-bit architecture. Should I file a > bug report? it still sounds like LLVM is being miscompiled to me. What compiler did

I'm just getting started changing the emitted code. Suppose I want to just add one new instruction to the function prologue at the top. Where would I do this? Does it depend on the backend? I'm targeting RISCV64. Gry

Is it possible to write a .pdb file containing the debug info stored in a COFFObjectFile once the NotifyObjectEmitted is triggered by the JIT compiler ? -------------- next part -------------- An HTML attachment was scrubbed... URL: <http://lists.llvm.org/pipermail/llvm-dev/attachments/20180308/3c25cddd/attachment.html>

Hi All, For various obscure reasons I'd like to detect the condition when X86 CALL instruction immediately precedes a function epilogue in the final emitted code, and insert a NOP between them if that happens. My initial attempt at it looked like this: MachineBasicBlock& MBB; MachineBasicBlock::iterator MBBI; <-- points to where the epilogue would be inserted if (MBBI != MBB.begin()

Hi Gordon, You're talking about Thumb mode code here, right? As I understand it, the ARM mode MOV instruction is valid for everything ARMv4 and up. (nit picky side note: CPY is an obsolete mnemonic as of the introduction of unified syntax) Assuming so, using a high register causes headaches, as R8+ are call preserved registers, and the Thumb push/pop instructions can't reference them.

Hi, thanks for the answers, i tried to use LINK.EXE to produce a pdb, but unfortunately, it produces an empty pdb (i have debug symbols in the dumped .obj because I can see them with PEViewer). It is probably because i have undefined symbols, which is logical because i'm dumping machine code which points to running native code. I still dont understand why LINK.EXE is not able to produce a

Hi, I've been thinking about ways to get around this in the short term for some time, http://llvm.org/bugs/show_bug.cgi?id=1388 An end-user workaround is possibly to use at least one register > r7 for the MOV form of the instruction. In that case, what is listed in the bug as the CPY instruction (which is the ARMv6 version generated if both Rd and Rm are <= r7) will become a valid MOV

On Jul 2, 2009, at 10:48 AM, robert muth wrote: > I spend over a day trying to follow your suggestion. In the end I > was not successful. Here is what Iearned: > > After setting > > ARMJITInfo::hasCustomJumpTables -> true > setOperationAction for ISD::BR_JT -> Expand > > I needed to add a "brind" definition to ARMInstrInfo.td > I picked