similar to: [LLVMdev] Can I modify C syntax by using Clang and LLVM?

Hi, Inserting a call instruction is a bit of a pain. The only way I know how to do it is to write a bunch of code like the following: std::vector<const Type*> formalArgs; formalArgs.push_back(arg1->getType()); formalArgs.push_back(arg2->getType()); ... formalArgs.push_back(argn->getType()); std::vector<Value*> args; args.push_back(arg1);

I'm trying to insert some instructions before each function calls (before arguments push): lea %EAX, label ----- new instructions mov [ESP+stacksize], %EAX ----- new instructions push arg1 push arg2 ... push argn call callee_name I am a newbie to LLVM. I tried to use buildMI() to insert the instructions in the lowercall() function. But I couldn't

How do you handle this during codegen? One problem is avoid stack changes (like spills). Another is coordinating things that are using allocas and those that are not but end up in the stack. Consider void foo(int arg1, int arg2, int arg3, ....CXXTypeWithCopyConstructor argn, int argp1...) You will need an alloca for argn, but the ABI also requires it to be next to the plain integers that

So one way might look like this: IRBuilder<> Builder(&*BB); // BB = Function::iterator OR IRBuilder<> Builder(CallInst->getParent()); Builder.SetInsertPoint(CallInst); InstructionClass *YourNewInstruction = builder.CreateInstructionClass(.....); // InstructionClass = type of instruction you are inserting -Ryan On Sat, Sep 3, 2016 at 6:04 PM, Ryan Taylor <ryta1203 at

On Sun, Sep 4, 2016 at 7:44 AM, Ryan Taylor <ryta1203 at gmail.com> wrote: > Mehdi, > > Sorry, I misread his original post. > > So something like: > > XXXInsrtInfo *XII; // target instruction info > MachineBasicBlock::iterator MI = MachineBasicBlock(YourCallInst); > MachineBasicBlock *MBB = YourCallInst->getParent(); // basic block > location of

When I run this now on the Target branch alone, it finds 22,000 errors from just simple things like lines too long, tabs and whitespace at the end of lines. More features to come. Usage: style_check.py arg1 arg2 ... argn [options] Options: -h, --help show this help message and exit --no-check-for-tabs don't check for tabs --no-check-line-length

Hi LLVM folks, To properly implement pass-by-value in the Microsoft C++ ABI, we need to be able to take the address of an outgoing call argument slot. This is http://llvm.org/PR5064 . Problem ------- On Windows, C structs are pushed right onto the stack in line with the other arguments. In LLVM, we use byval to model this, and it works for C structs. However, C++ records are also passed this

We do something very similar here for XRay, and I would think the approach would be similar. What XRay does are the following: - Find the machine instructions in a MachineFunctionPass that look interesting from the XRay perspective. Theses turn out to be: the beginning of the function (not really an instruction but a location), tail calls, and returns. I suspect you can very simply find the call

Hi, I have a situation where a DTrace script is printing out extra characters, despite the copyin() call giving a specific length. Can anyone think of why this might be? It''s fine the first time all of the probes fire, but on a second run of my generating operations, I get junk in there. For example: set setop length 5, FOUND KEY, STORED set setop length 5, FOUND KEY, STORED get

The loads, stores and float arithmetic in attached function should be completely vectorizable. The bb-vectorizer does a good job at first, but from instruction %96 on it messes up by adding unnecessary vectorshuffles. (The function was designed so that no shuffle would be needed in order to vectorize it). I tested this with llvm 3.6 with the following command:

Hi, I was tracking down a performance regression and noticed that LoopStrengthReduce hoists loop invariants (e.g., the initial formulae of indvars) in the reverse order of how they appear in the loop. This reverse order creates troubles for the StraightLineStrengthReduce pass I recently add. While I understand ultimately SLSR should be able to sort independent candidates in an optimal order,

Hello, I'm using ROCR to plot ROC Curves and I want to automate the saving of plots into PNG files using a custom function. My data frames are named like test1, test2, test3. Each data frame has three variables: method1, method2, goldstandard. Right now, for each plot I have to run: png('test1_method1.png') plot(performance(prediction(test1$method1, test1$goldstandard),

I am writing a program where non-technical R users will read in a config file and the config file will then parse the arguments found within the config and pass them to respective functions. I'm having trouble (efficiently) writing a piece of code to retain quotation marks around the argument which requires it as input, as found in the example function below, myFuncton1. Below is a minimal,

I seem to have problem to get the SLP vectorizer to make use of the full 8 floats available in a SIMD vector on a Sandy Bridge CPU with AVX. The function is attached, the CPU flags are: flags : fpu vme de pse tsc msr pae mce cx8 apic mtrr pge mca cmov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon pebs bts rep_good

Hello R-helpers, I'm trying to add labels to points in xyplot graphs, but I want lo label only those points which have a certain level of my grouping variable, and have encountered a few problems. An example dataframe that goes with the following code is at the end of this message. 1st step, adding labels (from another column in the dataframe) to my chosen points : I've figures this

Hello All: How can I determing the "types" of args passed to an R function? For example, given the following: calculate <- function(...) { args <- list(...) argName = names(args) if (arg1 == character) cat("arg1 is a character") else cat("arg1 is numeric") if (arg2 == character)

Dear R community, I am struggling a bit with a probably fairly simple task. I need to use some already existing functions as argument for a new function that I am going to create. 'dataset' is an argument, and it comprises objects named 'mean_test', 'sd_test', 'kurt_test' and so on. 'arg1' tells what object I want (mean, sd, kurt) while 'arg2' tells

Harold: As a general rule, if you are using eval(parse(...)) you are doing it poorly in R; cf library("fortunes") fortune(106) Why is something like this not suitable: fun1 <- function(a1,a2,a3 = c("hi","by")) { cat(a3,a1+a2,"\n") } > fun1 (1,2) hi by 3 > fun1(1,2, a3 = "whoopee") whoopee 3 ... or, if you want to include the

Hi all, This is a simple concept, however I'm not entirely comfortable with available applications and functions available to me to make this happen. I have a simple dialout macro such as the following: [macro-dialout]; arg1 = callerid number; arg2 = phone numberl exten => s,1,Set(CALLERID(number)=${ARG1}) exten => s,2,GotoIf($[${LEN(${ARG2})} = 10]?3:4) exten =>

Please CC as I''m not subscribed to valgrind-developers. Under Xen the toolstack is responsible for managing the domains in the system, e.g. creating, destroying, and otherwise manipulating them. To do this it uses a number of ioctls on the /proc/xen/privcmd device. Most of these (the MMAPBATCH ones) simply set things up such that a subsequenct mmap call will map the desired guest