llvm optimizations - c

List of available LLVM passes are desribed here. I am interested in knowing which passes are included for different optimization levels. For example, for -O2 which passes are included. Is there any site or document which provides that information.

It's visible in PassManagerBuilder.cpp file. Take a look at populateModulePassManager method where OptLevel is integer that follows -O option.

from the command line you can use:
echo "" | opt -O3 -disable-output -debug-pass=Arguments

Related

when is 'f ' needed in gcc option [duplicate]

Summary: What do the -f and -m in gcc and clang compiler options stand for?
Details:
When using clang I've noticed that many compiler options start with -f and others start with -m. I assume that there is some historical reason for this and I was curious so I looked at the gcc help and saw the following:
Options starting with -g, -f, -m, -O, -W, or --param are automatically
passed on to the various sub-processes invoked by gcc. In order to
pass other options on to these processes the -W options must
be used.
If I had to guess I think that -f might stand for frontend and -m for machine. But I'd be interested to hear a more comprehensive answer, possibly including the other sub-processes that gcc invokes.
I don't have specific sources that state what 'f' and 'm' mean, but we can infer based on usage patterns found in documentation.
'f' stands for 'flag'.
Flags are on if specified via '-fFLAG' and off via '-fno-FLAG'
ex:
-fpic # flag to set position independent code
-fno-builtin # don't recognize build in functions ...
The technical definition is that 'f' defines "Control the interface conventions used in code generation".
Src:
https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html
(e.g -fpci, when this flag is set)
'm' stands for mode. One general characteristic is that it sometimes has parameters. e.g
-mabi=name #abi mode = name
-mcpu=cpu
Src:
https://gcc.gnu.org/onlinedocs/gccint/Standard-Names.html (e.g ... when this mode...)
According to gcc onlinedocs, options of the form -ffoo and -fno-foo stand for machine independent code generation conventions.
Examples: fpic, -fno-pic
-m options stand for machine dependent options.
eg: -mcpu, -march, -matomic
https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html#Code-Gen-Options or https://home.cs.colorado.edu/~main/cs1300/doc/gnu/gcc_2.html#SEC43
https://gcc.gnu.org/onlinedocs/gcc/Submodel-Options.html#Submodel-Options or https://home.cs.colorado.edu/~main/cs1300/doc/gnu/gcc_2.html#SEC16

How to trace specific functions/files in C?

I have already known that the GCC's argument -finstrument-functions can hook the functions and the argument -finstrument-functions-exclude-file(functions)-list can exclude some files/functions to be traced.
But now I have a lot of files to be compiled and only some of them need to be traced. I wonder if I can include some specific functions/files to be traced, such as something like -finstrument-functions-include-file(functions)-list?
Thanks a lot!
GCC does not support this out-of-the-box (it's more a task for your build system). One common hack to achieve what you want is to write a shell wrapper which replaces GCC and adds flags where needed:
$ cat path/to/fake/gcc
#!/bin/sh
FLAGS=
if echo "$*" | grep -q 'myfile1.c'; then
FLAGS=-finstrument-functions
fi
exec /usr/bin/gcc "$#" $FLAGS
$ export PATH="path/to/fake:$PATH"
If you use cmake to build your project you may benefit from adding COMPILE_OPTIONS at a specific level. Use
add_compile_options()
for directory-wide settings
target_compile_options()
for target-specific settings and
set_source_files_properties()
for file specific settings.
In your case
set_source_files_properties(
myfile1.cc PROPERTIES COMPILE_FLAGS -finstrument-functions)
Recent GCC compilers can be extended by GCC plugins.
But now I have a lot of files to be compiled and only some of them need to be traced.
You should consider writing your own GCC plugin to do that job. See also this draft report.
You may configure your build automation tool (e.g. GNU make or ninja) to help you.
At last, some of your C code (e.g. #include-ed files) could be generated. Think of meta-programming approaches (e.g. with SWIG or ANTLR or Bison or GPP or your own C code generator), perhaps using X-macros.

Tracking compiler flags within a Shake Action

I need to track compiler flags used as part of a rule, e.g when supplied as arguments to a function. Does Shake have a way to track such flags as inputs in the same vein as need? As a trivial example, I would like for Shake to rebuild all .o files when the rule changes to pass -O2 to the C compiler instead of -O0.
There are two approaches to tracking things like whether to use optimisation or not.
1) Use Oracles
Oracles very closely match what you are asking for. To track something like
-O0 vs -O2 you'd need an oracle that tracks optimisation level:
newtype OptLevel = OptLevel ()
deriving (Show,Typeable,Eq,Hashable,Binary,NFData)
type instance RuleResult OptLevel = String
rules = do
addOracle $ \(OptLevel _) -> return $
if <whatever you use to decide> then "-O0" else "-O2"
"foo.o" %> \_ -> do
level <- askOracle $ OptLevel ()
cmd "gcc" level ...
Now the optimisation level is a tracked dependency which will update if anything changes. This example is based on the docs for addOracle.
2) Use different output files
For compiler flags another approach is to use different build directories, i.e. so build/opt (and build/opt/obj etc) has binaries and .o files built with -O, build/debug without, build/profile with profiling flags, and build/test with test flags. Some others, like build/doc and build/hsc, have generated files that aren't dependent on compiler flags.
The advantage of this approach is that you can keep all the files cached at once, and refreshing the debug or test build doesn't destroy the opt one.
The disadvantage is that it's just for a set of hardcoded
flags. But it's also not hard to add a new mode, you just need a new
(directory, flags) pair for it.

Can the object files output by gcc vary between compilations of the same source with the same options?

Does the gcc output of the object file (C language) vary between compilations? There is no time-specific information, no change in compilation options or the source code. No change in linked libraries, environmental variables either. This is a VxWorks MIPS64 cross compiler, if that helps. I personally think it shouldn't change. But I observe that sometimes randomly, the instructions generated changes. I don't know what's the reason. Can anyone throw some light on this?
How is this built? For example, if I built the very same Linux kernel, it includes a counter that is incremented each build. GCC has options to use profiler information to guide code generation, if the profiling information changes, so will the code.
What did you analyze? The generated assembly, an objdump of object files or the executable? How did you compare the different versions? Are you sure you looked at executable code, not compiler/assembler/linker timestamps?
Did anything change in the environment? New libraries (and header files/declarations/macro definitions!)? New compiler, linker? New kernel (yes, some header files originate with the kernel source and are shipped with it)?
Any changes in environment variables (another user doing the compiling, different machine, different hookup to the net gives a different IP address that makes it's way into the build)?
I'd try tracing the build process in detail (run a build and capture the output in a file, and do so again; compare those).
Completely mystified...
I had a similar problem with g++. Pre 4.3 versions produced exactly the same object files each time. With 4.3 (and later?) some of the mangled symbol names are different for each run - even without -g or other recordings. Perhaps the use a time stamp or random number (I hope not). Obviously some of those symbols make it into the .o symbol table and you get a difference.
Stripping the object file(s) makes them equal again (wrt. binary comparison).
g++ -c file.C ; strip file.o; cmp file.o origfile.o
Why should it vary? It is the same result always. Try this:
for i in `seq 1000`; do gcc 1.c; md5sum a.out; done | sort | uniq | wc -l
The answer is always 1. Replace 1.c and a.out to suit your needs.
The above counts how many different executables are generated by gcc when compiling the same source for 1000 times.
I've found that in at least some environments, the same source may yield a different executable if the source tree for the subsequent build is located in a different directory. Example:
Checkout a pristine copy of your project to dir1. Do a full rebuild from scratch.
Then, with the same user on the same machine, checkout the same exact copy of your source code to dir2 (dir1 != dir2). Do another full rebuild from scratch.
These builds are minutes apart, with no change in the toolchain or any 3rd party libs or code. Binary comparison of source code is the same. However, the executable in dir1 has different md5sum than the executable in dir2.
If I compare the different executables in BeyondCompare's hex editor, the difference is not just some tiny section that could plausibly be a timestamp.
I do get the same executable if I build in dir1, then rebuild again in dir1. Same if I keep building the same source over and over from dir2.
My only guess is that some sort of absolute paths of the include hierarchy are embedded in the executable.
My gcc sometimes produces different code for exactly the same Input. The output object files differ in exactly one byte.
Sometimes this causes linker Errors, because one possible object file is invalid. Recompiling another version usually fixes the linker error.
The gcc Version is 4.3.4 on Suse Linux Enterprise.
The gcc Parameters are:
cc -std=c++0x -Wall -fno-builtin -march=native -g -I<path1> -I<path2> -I<path3> -o obj/file.o -c file.cpp
If someone experiences the same effect, then please let me know.

gsoap client compile/link error

Now I am writing a program to call a web service. I write testMain.c. The others are generated by wsdl2h and soapcpp2.
My compiling command is like this:
gcc -Wall -g -c -L. soapC.c soapClient.c stdsoap2.c testMain.c
gcc -o testMain -L/usr/lib -lgsoap -lgsoapck -lgsoapssl soapC.o soapClient.o stdsoap2.o testMain.o
And I get these errors. Please help me.
stdsoap2.o: In function `soap_print_fault':
/test/stdsoap2.c:16279: undefined reference to `soap_check_faultsubcode'
/test/stdsoap2.c:16281: undefined reference to `soap_check_faultdetail'
stdsoap2.o: In function `soap_sprint_fault':
/test/stdsoap2.c:16341: undefined reference to `soap_check_faultdetail'
collect2: ld returned 1 exit status
Recent versions of GCC/ld/the GNU toolchain require that the object and library files be specified in a certain order, so that symbols can be found by the linker in the same order they depend on each other. This means that libraries should go to the end of the command line; your second line (when you're linking) should be
gcc -o testMain -L/usr/lib soapC.o soapClient.o stdsoap2.o testMain.o -lgsoap -lgsoapck -lgsoapssl
instead.
I search the web, and found a post which is very similar with my problem. I use this solution and have solved the problem. http://www.mail-archive.com/gsoap#yahoogroups.com/msg01022.html
You should not need to link stdsoap2.o to your project because it's already included in libgsoap (given through the gcc linker option -lgsoap). Try to exclude stdsoap2.c from your project. From the gSOAP FAQ:
I get a link error with gcc/g++ (GNU GCC). What should I do? For C
apps: use soapcpp2 option -c to generate C code, use only the
package's .c files, link with libgsoap.a (-lgsoap) or use the lib's
source stdsoap2.c (and dom.c when applicable).
I had the same problem with gsoap-2.8.16 compiled from source. (That version was shipped with CentOS 6.)
First I checked for a missing library. According to nm used on all static libraries provided by gsoap-2.8.16:
for X in /usr/local/lib/libgsoap*.a ; do echo $X; nm $X | grep soap_check_faultdetail; done`
it turned out that none of the libraries provided the missing symbols.
A brief look at the source code revealed that the expected return type of both methods soap_check_faultdetail and soap_check_faultsubcode was const char*, and that these were used to generate error messages.
It looked to me as if these are meant to be callbacks that the client must provide. Maybe their implementation is WSDL-dependent and would be supplied by the gsoap code generation utilities - that I don't know, see the answer from #ChristianAmmer above or below.
Anyway, since I knew the symbols were nowhere supplied, and that null-terminated strings were probably acceptable here, I just supplied my own no-op implementation:
// gsoap-missing-symbols.cpp
extern "C" {
const char* soap_check_faultdetail() { return 0; }
const char* soap_check_faultsubcode() { return 0; }
}
This is a brute-force solution. If you follow this solution, you should maybe check for linker warnings in the future; maybe some mechanism (eg. from the gsoap code generator) will supply conflicting implementations later during development.
For later versions of gsoap, I believe these symbols are no longer used and can be dropped (or renamed), see soap_check_faultX in https://www.genivia.com/changelog.html.

Resources