Re: [klee-dev] One question about external dispatcher
Thanks. You are right. I solved the problem.
Best regards,
Qiuping Yi
Parasol Laboratory
Department of Computer Science and Engineering
Texas A University
College Station
TX 77843
On Sat, Jan 28, 2017 at 4:12 AM, Dan Liew <d...@su-root.co.uk> wrote:
> On 27 January 2017 at 04:18, Qiuping Yi <yiqiup...@gmail.com> wrote:
> > Dear all,
> >
> > I encountered a strange problem when testing the next code snippet:
> >
> > 1 if (pw = getpwuid(getuid()) == NULL)
> > 2 return ;
> >
> > 3 .. = pw->pw_dir;
>
> Please use the correct mailing list (klee-dev@imperial.ac.uk) instead
> of the old klee-...@keeda.stanford.edu mailing list.
>
> It would be better if you provided a small complete example. Like this.
>
> ```
> #include
> #include
> #include
> #include
> #include
>
> int main(int arc, char** argv) {
> struct passwd* pw;
> uid_t uid = getuid();
> printf("uid is %d\n", uid);
> if (pw = getpwuid(getuid()) == NULL) {
> printf("Failed\n");
> return 1;
> }
> assert(pw && "pw cannot be NULL");
>
> char* pw_dir = pw->pw_dir;
> printf("pw_dir: %s\n", pw_dir);
> return 0;
> }
> ```
>
> Your code is wrong.
>
> if (pw = getpwuid(getuid()) == NULL)
>
> is doing this
>
> if ( pw = ( getpwuid(getuid()) == NULL )
>
> so a pointer is returned by `getpwuid()` and then we compare with NULL
> which is false so then `pw` gets assigned the value zero.
>
> However once I fix your code to
>
> if ((pw = getpwuid(getuid())) == NULL) {
>
> then I can reproduce the problem if I just run `klee program.bc`
>
> I suspect it's to do with the fact `getpwuid()` returns a pointer to
> "real memory" which does not point to anything in KLEE's own model of
> the memory (i.e. the address space of the program under).
>
> To fix this you need not call `getpwuid()` as an external function but
> instead call it from klee-uclibc so that it can be symbolically
> executed.
>
> If you run
>
> ```
> klee -libc=uclibc program.bc
> ```
>
> no out of bounds access is reported.
>
> HTH,
> Dan.
>
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
[klee-dev] One question about external dispatcher
Dear all, I encountered a strange problem when testing the next code snippet: 1 if (pw = getpwuid(getuid()) == NULL) 2 return ; 3 .. = pw->pw_dir; When handling line 1, KLEE firstly invokes *externalDispatcher->executeCall* which will invoke *runProtectedCall* to execute the external function and store the result to a given memory location. Then, it will invoke *fromMemory* to get the return value from the location. However, it encounters an "out of bound" error when handling line 3. I printed the value of variable *pw* at line 1, and got something like *139894903382656*, which definitely does not equal to ' *null*', thus it will arrive line 3. Actually, *139894903382656* seems an invalid address. So why does this strange situation happen? Thank you all in advance. Best regards, Qiuping Yi Parasol Laboratory Department of Computer Science and Engineering Texas A University College Station TX 77843 ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
Re: [klee-dev] Strange behavior of KLEE when evaluating NULL pointer
Hi, Paul
Next is my whole test code:
1void test2(int *p) {
2 if (p == NULL)
3p = malloc(sizeof(*p));
4 *p = 1;
}
5 int main() {
6 int *p;
7 test(p);
8 return 0;
}
This code was executed without any error after compiled with gcc. However, when
I applied KLEE to this code, it encountered an 'out of bound pointer'
error. If I replace line 6 to 'int *p = NULL', no error happened. So KLEE
needs to explicitly initialize each variable, containing the pointers,
right?
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
On Tue, May 13, 2014 at 6:50 PM, Paul Thomson pault...@gmail.com wrote:
Please can you provide the code that calls test?
Or, please try using something like:
int main()
{
int *p = NULL;
test(p);
return 0;
}
Thanks,
Paul
On 13 May 2014 11:09, Qiuping Yi yiqiup...@gmail.com wrote:
Hi, everyone
I found a strange behavior of KLEE.
When I applied KLEE to the next code snippet, a out-of-bound-pointer
error happened at line 3. However, this code snippet explicitly allocates
space for pointer p at line 2 when it is evaluated to NULL. So what's wrong?
0 void test (int *p) {
1if (p == NULL)
2p = malloc(sizeof(*p));
3 *p = 2;
}
Best Regards!
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
Re: [klee-dev] Strange behavior of KLEE when evaluating NULL pointer
Hi Oscar, Hongxu,
Thank you for you comment.
Yes, you are right. In practice, my test code have an uninitialized error.
But now I am considering some variables from external environment. These
variables may be uninitialized, so I use pointer p in function 'test' to
describe such a simple situation. I want to use the if-statement at line 2
to judge whether the external pointer p is NULL, if it is I want to
explicitly allocate some space to it at line 3. But now I know KLEE cannot
judge on an uninitialized pointer.
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
On Wed, May 14, 2014 at 10:51 PM, Oscar Soria Dustmann
oscar.soriadustm...@comsys.rwth-aachen.de wrote:
Hi Qiuping,
your code exhibits undefined behaviour as you read from uninitialised
memory. The reason it doesn't fail when run natively is implementation
defined and pure luck. It's not that KLEE needs the initialisation; The
C standard demands it (reason behind it: Stack variables are not
default-initialised for performance reasons).
I'd actually consider it desired behaviour for a testtool to throw you
an error.
Cheers,
Oscar
On 14/05/14 15:43, Qiuping Yi wrote:
Hi, Paul
Next is my whole test code:
1void test2(int *p) {
2 if (p == NULL)
3p = malloc(sizeof(*p));
4 *p = 1;
}
5 int main() {
6 int *p;
7 test(p);
8 return 0;
}
This code was executed without any error after compiled with gcc.
However, when
I applied KLEE to this code, it encountered an 'out of bound pointer'
error. If I replace line 6 to 'int *p = NULL', no error happened. So KLEE
needs to explicitly initialize each variable, containing the pointers,
right?
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
On Tue, May 13, 2014 at 6:50 PM, Paul Thomson pault...@gmail.com
wrote:
Please can you provide the code that calls test?
Or, please try using something like:
int main()
{
int *p = NULL;
test(p);
return 0;
}
Thanks,
Paul
On 13 May 2014 11:09, Qiuping Yi yiqiup...@gmail.com wrote:
Hi, everyone
I found a strange behavior of KLEE.
When I applied KLEE to the next code snippet, a out-of-bound-pointer
error happened at line 3. However, this code snippet explicitly
allocates
space for pointer p at line 2 when it is evaluated to NULL. So what's
wrong?
0 void test (int *p) {
1if (p == NULL)
2p = malloc(sizeof(*p));
3 *p = 2;
}
Best Regards!
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
[klee-dev] Invitation to connect on LinkedIn
LinkedIn I'd like to add you to my professional network on LinkedIn. - Qiuping Qiuping Yi -- China Confirm that you know Qiuping Yi: https://www.linkedin.com/e/fwmsg6-hv7cmrb6-2i/isd/5872510063670738944/1temtXms/?hs=falsetok=1O_Jrw8fFq4mg1 -- You are receiving Invitation to Connect emails. Click to unsubscribe: http://www.linkedin.com/e/fwmsg6-hv7cmrb6-2i/eFPC4J5Rz7QQtXHa3b6Rb1PEH6bj6NMYmSof/goo/klee-dev%40imperial%2Eac%2Euk/20061/I7078372816_1/?hs=falsetok=0uOyfM9vZq4mg1 (c) 2012 LinkedIn Corporation. 2029 Stierlin Ct, Mountain View, CA 94043, USA. ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
[klee-dev] Invitation to connect on LinkedIn
LinkedIn I'd like to add you to my professional network on LinkedIn. - Qiuping Qiuping Yi -- China Confirm that you know Qiuping Yi: https://www.linkedin.com/e/-8mnz76-hv7cmk5x-63/isd/5872510076471758848/Kko2r6sn/?hs=falsetok=2gU_XTtn5q4mg1 -- You are receiving Invitation to Connect emails. Click to unsubscribe: http://www.linkedin.com/e/-8mnz76-hv7cmk5x-63/vqX3hpfgkANPlqlepZu0pdf9_QcPAz40Hmv3R5zl8S/goo/klee-dev%40keeda%2Estanford%2Eedu/20061/I7078372851_1/?hs=falsetok=01Z8LeSqpq4mg1 (c) 2012 LinkedIn Corporation. 2029 Stierlin Ct, Mountain View, CA 94043, USA. ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
[klee-dev] Strange behavior of KLEE when evaluating NULL pointer
Hi, everyone
I found a strange behavior of KLEE.
When I applied KLEE to the next code snippet, a out-of-bound-pointer error
happened at line 3. However, this code snippet explicitly allocates space
for pointer p at line 2 when it is evaluated to NULL. So what's wrong?
0 void test (int *p) {
1if (p == NULL)
2p = malloc(sizeof(*p));
3 *p = 2;
}
Best Regards!
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
___
klee-dev mailing list
klee-dev@imperial.ac.uk
https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
Re: [klee-dev] Strange error when compiling 'bash-4.0' using klee-gcc
Hi, Dan Liew, Thank you for your advice. Now I can use *wllvm *to generate bash.bc successfully based on the next steps: $ export LLVM_COMPILER=llvm-gcc $ export LLVM_GCC_PREFIX=llvm- $ cd bash-4.0 $ ./configure --disable-nls CFLAGS=-g $ make CC=wllvm $ extract-bc bash however, I cannot run klee on the generated 'bash.bc'. when I carry out: $ klee *--libc=uclibc --posix-runtime *./bash.sh test.sh I got the next error messag: KLEE: ERROR: unable to load symbol(PC) while initializing globals. Could you tell me what should I do then? Thank you very much. Qiuping Yi Institute Of Software Chinese Academy of Sciences On Mon, Mar 24, 2014 at 7:49 PM, Daniel Liew daniel.l...@imperial.ac.ukwrote: /home/guest/installed/klee/scripts/klee-gcc -DPROGRAM='bash' -DCONF_HOSTTYPE='x86_64' -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='x86_64-unknown-linux-gnu' -DCONF_VENDOR='unknown' -DLOCALEDIR='/usr/local/share/locale' -DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H -I. -I. -I./include -I./lib -g -o mksyntax ./mksyntax.c /tmp/cc1H53KR.o: file not recognized: File format not recognized collect2: ld returned 1 exit status make: *** [mksyntax] Error 1 It looks like the script is trying to invoke your system's native linker to build a native executable. That won't work because /tmp/cc1H53KR.o is probably an LLVM bitcode file so the linker does not know what to do with it. You probably should be using the '-c' flag so you do not invoke the linker. This approach is really only feasible with single file programs. Furthermore, it's strange that the not recognized file keeping changed if I execute 'make' again and again. For example, the second time I encounter the next error message: /tmp/ccaJGvTn.o: file not recognized: File format not recognized It's not strange at all, the compiler is creating temporary files during the compilation process. BUT I found both file /tmp/cc1H53KR.o and /tmp/ccaJGvTn.o do not exist. These files are deleted automatically by the compiler because they are temporary intermediate files. What's wrong? What should I do? I've already explained above what is wrong. I do not know much about bash's build system but I would expect it to be more complex that coreutil's. klee-gcc is a bit of a hack so you should probably either - Use wllvm [1]. This is also a hack but is considerably better than klee-gcc and can link together multi file programs. - Use LLVM gold [2] so you can link together a multi-file program to a bitcode module. [1] https://github.com/klee/whole-program-llvm [2] http://llvm.org/docs/GoldPlugin.html Thanks, Dan Liew. ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
[klee-dev] Strange error when compiling 'bash-4.0' using klee-gcc
Hi, everyone I want to compile 'bash-4.0' with klee-gcc following the next commands (like compiling coreutils http://klee.github.io/klee/TestingCoreutils.html ): $ ../configure --disable-nls CFLAGS=-g $ make CC=/home/guest/installed/klee/scripts/klee-gcc However, I encountered an error when 'make', which the following error message: /home/guest/installed/klee/scripts/klee-gcc -DPROGRAM='bash' -DCONF_HOSTTYPE='x86_64' -DCONF_OSTYPE='linux-gnu' -DCONF_MACHTYPE='x86_64-unknown-linux-gnu' -DCONF_VENDOR='unknown' -DLOCALEDIR='/usr/local/share/locale' -DPACKAGE='bash' -DSHELL -DHAVE_CONFIG_H -I. -I. -I./include -I./lib -g -o mksyntax ./mksyntax.c */tmp/cc1H53KR.o: file not recognized: File format not recognized* collect2: ld returned 1 exit status make: *** [mksyntax] Error 1 Furthermore, it's strange that the not recognized file keeping changed if I execute 'make' again and again. For example, the second time I encounter the next error message: */tmp/ccaJGvTn.o: file not recognized: File format not recognized* BUT I found both file */tmp/cc1H53KR.o and **/tmp/ccaJGvTn.o *do not exist. What's wrong? What should I do? Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
[klee-dev] why different traces are generated under the same input?
Hi, I ran klee on 'findutils' under the same input without any symbolic variable, and want to get a determinate execution. However, I get two different executions randomly(one path with a bigger probability). Why? How can I get a determinate execution? Thank you very much. Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
[klee-dev] How to preserve program variables?
I am using klee-gcc to compile 'findutils', but I don't know why the variables are all replaced with temp variables(such as %10), how can I generate the bytecode and reserve the program variables. I used the next commands without any optimization options: 1. $ ../configure --disable-nls CFLAGS=-g 2. $ make CC=/home/guest/installed/klee/scripts/klee-gcc Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
Re: [klee-dev] How to preserve program variables?
*I found the bytecode before linking had program variables, but the generated target object had non. why?* * * *Because during the 'make' step, I encountered an error: * /home/guest/installed/klee/scripts/klee-gcc -Wdeclaration-after-statement -Wall -g -o find ftsfind.o ./libfindtools.a ../lib/libfind.a ../gnulib/lib/libgnulib.a -lrt -lm -lm /usr/bin/ld.bfd.real: cannot find crt1.o: No such file or directory /usr/bin/ld.bfd.real: cannot find crti.o: No such file or directory ftsfind.o: file not recognized: File format not recognized collect2: ld returned 1 exit status *so I use the next command by hand instead:* llvm-ld -o find ftsfind.o ./libfindtools.a ../lib/libfind.a ../gnulib/lib/libgnulib.a -L /usr/lib32/ -lrt -lm -lm *when '-disable-opt' was added, the same output was generated. * llvm-ld -disable-opt -o find ftsfind.o ./libfindtools.a ../lib/libfind.a ../gnulib/lib/libgnulib.a -L /usr/lib32/ -lrt -lm -lm *Does some error in my llvm-ld command? * Qiuping Yi Institute Of Software Chinese Academy of Sciences On Fri, Nov 1, 2013 at 4:20 PM, Qiuping Yi yiqiup...@gmail.com wrote: I am using klee-gcc to compile 'findutils', but I don't know why the variables are all replaced with temp variables(such as %10), how can I generate the bytecode and reserve the program variables. I used the next commands without any optimization options: 1. $ ../configure --disable-nls CFLAGS=-g 2. $ make CC=/home/guest/installed/klee/scripts/klee-gcc Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
Re: [klee-dev] how to compile gnu utility 'find' using klee-gcc?
Thank you, we have solved this problem following your instructions. Qiuping Yi Institute Of Software Chinese Academy of Sciences On Thu, Oct 31, 2013 at 10:04 AM, Hongxu Chen leftcopy@gmail.comwrote: Hi Yi, I guess you need to read the original makefile and do some wrapper yourself instead of using klee-gcc. From KLEE's coreutils case study page, Step 2( http://ccadar.github.io/klee/TestingCoreutils.html), There are such sentences: It depends on the actual project what the best way to do this is. For coreutils, we use a helper script klee-gcc, which acts like llvm-gcc but adds additional arguments to cause it to emit LLVM bitcode files and to call llvm-ld to link executables. This is *not* a general solution, and your mileage may vary. klee-gcc is a python script specified for coreutils bitcode generation. To build bc instead of native code for general purpose, I think you can try wllvm. Thanks, Hongxu Chen On Thu, Oct 31, 2013 at 9:28 AM, Yi Zhou zhouyi198...@gmail.com wrote: Hi, I want to compile gnu utility 'find' to get an executed *.bc file, so I use the following commands: ./configure CC=/path/to/klee-gcc LD=/path/to/llvm-ld make make install However, at the end I did not get any *.bc file, and the generated file 'find' is said to Invalid bitcode signature,but I can use the similar configure command to compile bosybox successfully. where is the error, and what should I do? Thank you all. Yi Zhou Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev ___ klee-dev mailing list klee-dev@imperial.ac.uk https://mailman.ic.ac.uk/mailman/listinfo/klee-dev
Re: [klee-dev] how to get a new copy of execution state?
Thank you for your replay! I want the constraints collected in b don't appear in a, and the changes on the constraints of b don't affect a. I try the branch(), fork(), and its constructors, but they don't work. Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@keeda.stanford.edu http://keeda.Stanford.EDU/mailman/listinfo/klee-dev
[klee-dev] strange update on refExpr
Hello, everyone I got lost on the management of refExpr, my problem is followed: Now I want to replace a kid of an Expr object with a new Expr, so I implemented a new method Expr::setKid(int i, refExpr e), which replace the i-th kid with e. It worked all right at first. But when I want to update a Expr use its own value, for example, when i is increased with 1, I want to update the old Expr (i+1) to ((i+1)+1): (i+1)- (Add w32 1 (ReadLSB W32 0 i) ) to ((i+1)+1)--- (Add w32 1 (Add w32 1 (ReadLSB W32 0 i) ), But I got a wrong Expr N0:(Add w32 1 N0) Why I can not get the right Expr through function setKid(). It seems that Expr (i+1) (which was updated) and Expr (i+1) (the new value of i) shared the same memory, how can I explicitly make these two refExpr have different memory to solve this question. Thank you! Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@keeda.stanford.edu http://keeda.Stanford.EDU/mailman/listinfo/klee-dev
[klee-dev] how to get a new copy of execution state?
Hi, Who can tell me how to clone a new ExecutionState b of an exist ExecutionState a, and the change on b won't affect a? when I tried to use ExecutionState *b = new ExecutionState(a); to get a copy b of a, and did some change on b, I found a is also changed. Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@keeda.stanford.edu http://keeda.Stanford.EDU/mailman/listinfo/klee-dev
Re: [klee-dev] error in Coreutils Case Study
Hi,
yes, you are right!
I rebuilt my llvm and klee, now it's OK. Thank you very much!
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
2012/9/14 Delcypher delcyp...@gmail.com
Hi,
I'm not sure if the fact that your target triple i386-pc-linux-gnu is a
problem as x86_64 as far as I know is a super set of i386. However it is
surprising that your version of llvm-gcc is doing this which leads me to
believe you might of downloaded the 32-bit version of llvm-gcc by mistake.
If you run
$ llvm-gcc -dumpmachine
x86_64-unknown-linux-gnu
And don't get x86_64-unknown-linux-gnu you know you have the wrong version
(you can also run llvm-gcc --verbose to find out what options your
compiler was built with too). The x86_64 version of llvm-gcc is here...
http://llvm.org/releases/2.9/llvm-gcc4.2-2.9-x86_64-linux.tar.bz2
If it is the case that you had the 32-bit version of llvm-gcc I would
advise rebuilding uclibc, llvm and KLEE from scratch using your 64-bit
llvm-gcc compiler just to be sure nothing gets messed up.
Hope that helps,
Regards,
Dan Liew.
On 14 September 2012 03:28, Qiuping Yi yiqiup...@gmail.com wrote:
Dear Dan Liew,
Thanks for your detailed info.
I followed you instructions, but I still have the error. And I got the
next information.
$ llvm-ld --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
Built Aug 12 2012 (10:31:59).
Host: x86_64-unknown-linux-gnu
Host CPU: penryn
Registered Targets:
(none)
$ lli --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
Built Aug 12 2012 (10:31:59).
Host: x86_64-unknown-linux-gnu
Host CPU: penryn
Registered Targets:
x86- 32-bit X86: Pentium-Pro and above
x86-64 - 64-bit X86: EM64T and AMD64
*I checked the readable form of cat.bc, and I found the target triple is
i386-pc-linux-gnu not **x86_64-unknown-linux-gnu*. I think it is the
problem, how can I generate cat whose target triple is *
x86_64-unknown-linux-gnu*?
$ file /sbin/init
/sbin/init: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV),
dynamically linked (uses shared libs), for GNU/Linux 2.6.24,
BuildID[sha1]=0x029d6e45c81ad4eee2627f676b7580c9ed5fde47, stripped
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
2012/9/12 Delcypher delcyp...@gmail.com
Hi,
Could you make sure you reply to the mailing rather than just me, else
no one else but me will see your problem!
I'm sorry but I don't know why you're getting this new error message. I
never had that problem but you could try the following
0. Verify that you don't have multiple versions of LLVM installed. So
verify that your output is similar to this...
$ llvm-gcc --version
llvm-gcc (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2.9)
$ llvm-ldd --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
$ lli --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
Built Jul 20 2012 (11:49:59).
Host: x86_64-unknown-linux-gnu
Host CPU: penryn
Registered Targets:
x86- 32-bit X86: Pentium-Pro and above
x86-64 - 64-bit X86: EM64T and AMD64
If lli doesn't have any registered targets then you will have
problems! You will also have problems if your machines architecture is not
listed.
1. delete the obj-llvm directory
2. Make sure LLVM_LIB_SEARCH_PATH is set
3. Now make the obj-llvm directory. Make sure this folder is NOT in
the source tree for core-utils, in the instructions it seems to be. I
prefer having something like core-utils/src and core-utils/obj-llvm as the
source code and build directories respectively.
$ mkdir obj-llvm
$ cd obj-llvm/
$ ../src/configure --disable-nls CFLAGS=-g
4. Run make
$ make CC=/path/to/klee-gcc
5. Now try again running cat again.
If that fails inspect the produced bit code files (*.bc)
$ llvm-dis cat.bc
This will produce a cat.ll file. Which is the Bitcode file in human
readable form, take a look and check your target triple is something like
x86_64-unknown-linux-gnu
Other than that I have nothing more to suggest because I don't know much
about LLVM.
Thanks,
Dan Liew.
On 11 September 2012 15:28, Qiuping Yi yiqiup...@gmail.com wrote:
yes, you are right. But after I executed *make
CC=/full/path/to/klee/scripts/klee-gcc* successfully, I cannot
successfully execute */cat --version* following the guide, and I got
the next error information.
qiu@qiu:~/installed/coreutils-6.11/obj-llvm/src$ ./cat --version
lli: JITEmitter.cpp:1141: void
{anonymous}::JITEmitter::emitJumpTableInfo(llvm::MachineJumpTableInfo*):
Assertion `MJTI-getEntrySize(*TheJIT-getTargetData()) == sizeof(void*)
Cross JIT'ing?' failed.
0 lli
Re: [klee-dev] error in Coreutils Case Study
Dear Dan Liew,
Thanks for your detailed info.
I followed you instructions, but I still have the error. And I got the next
information.
$ llvm-ld --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
Built Aug 12 2012 (10:31:59).
Host: x86_64-unknown-linux-gnu
Host CPU: penryn
Registered Targets:
(none)
$ lli --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
Built Aug 12 2012 (10:31:59).
Host: x86_64-unknown-linux-gnu
Host CPU: penryn
Registered Targets:
x86- 32-bit X86: Pentium-Pro and above
x86-64 - 64-bit X86: EM64T and AMD64
*I checked the readable form of cat.bc, and I found the target triple is
i386-pc-linux-gnu not **x86_64-unknown-linux-gnu*. I think it is the
problem, how can I generate cat whose target triple is *
x86_64-unknown-linux-gnu*?
$ file /sbin/init
/sbin/init: ELF 64-bit LSB shared object, x86-64, version 1 (SYSV),
dynamically linked (uses shared libs), for GNU/Linux 2.6.24,
BuildID[sha1]=0x029d6e45c81ad4eee2627f676b7580c9ed5fde47, stripped
Qiuping Yi
Institute Of Software
Chinese Academy of Sciences
2012/9/12 Delcypher delcyp...@gmail.com
Hi,
Could you make sure you reply to the mailing rather than just me, else no
one else but me will see your problem!
I'm sorry but I don't know why you're getting this new error message. I
never had that problem but you could try the following
0. Verify that you don't have multiple versions of LLVM installed. So
verify that your output is similar to this...
$ llvm-gcc --version
llvm-gcc (GCC) 4.2.1 (Based on Apple Inc. build 5658) (LLVM build 2.9)
$ llvm-ldd --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
$ lli --version
Low Level Virtual Machine (http://llvm.org/):
llvm version 2.9
Optimized build with assertions.
Built Jul 20 2012 (11:49:59).
Host: x86_64-unknown-linux-gnu
Host CPU: penryn
Registered Targets:
x86- 32-bit X86: Pentium-Pro and above
x86-64 - 64-bit X86: EM64T and AMD64
If lli doesn't have any registered targets then you will have problems!
You will also have problems if your machines architecture is not listed.
1. delete the obj-llvm directory
2. Make sure LLVM_LIB_SEARCH_PATH is set
3. Now make the obj-llvm directory. Make sure this folder is NOT in the
source tree for core-utils, in the instructions it seems to be. I prefer
having something like core-utils/src and core-utils/obj-llvm as the source
code and build directories respectively.
$ mkdir obj-llvm
$ cd obj-llvm/
$ ../src/configure --disable-nls CFLAGS=-g
4. Run make
$ make CC=/path/to/klee-gcc
5. Now try again running cat again.
If that fails inspect the produced bit code files (*.bc)
$ llvm-dis cat.bc
This will produce a cat.ll file. Which is the Bitcode file in human
readable form, take a look and check your target triple is something like
x86_64-unknown-linux-gnu
Other than that I have nothing more to suggest because I don't know much
about LLVM.
Thanks,
Dan Liew.
On 11 September 2012 15:28, Qiuping Yi yiqiup...@gmail.com wrote:
yes, you are right. But after I executed *make
CC=/full/path/to/klee/scripts/klee-gcc* successfully, I cannot
successfully execute */cat --version* following the guide, and I got
the next error information.
qiu@qiu:~/installed/coreutils-6.11/obj-llvm/src$ ./cat --version
lli: JITEmitter.cpp:1141: void
{anonymous}::JITEmitter::emitJumpTableInfo(llvm::MachineJumpTableInfo*):
Assertion `MJTI-getEntrySize(*TheJIT-getTargetData()) == sizeof(void*)
Cross JIT'ing?' failed.
0 lli 0x00b6edff
1 lli 0x00b6f309
2 libpthread.so.0 0x7f3595437cb0
3 libc.so.6 0x7f359468d445 gsignal + 53
4 libc.so.6 0x7f3594690bab abort + 379
5 libc.so.6 0x7f359468610e
6 libc.so.6 0x7f35946861b2
7 lli 0x0078e8a6
8 lli 0x0054dfc0
9 lli 0x00b00acf
llvm::FPPassManager::runOnFunction(llvm::Function) + 591
10 lli 0x00b0150d
llvm::FunctionPassManagerImpl::run(llvm::Function) + 141
11 lli 0x00b016cd
llvm::FunctionPassManager::run(llvm::Function) + 173
12 lli 0x00780d8e
llvm::JIT::jitTheFunction(llvm::Function*, llvm::MutexGuard const) + 46
13 lli 0x0078114c
llvm::JIT::runJITOnFunctionUnlocked(llvm::Function*, llvm::MutexGuard
const) + 28
14 lli 0x00781440
llvm::JIT::getPointerToFunction(llvm::Function*) + 576
15 lli 0x0077f677
llvm::JIT::runFunction(llvm::Function*, std::vectorllvm::GenericValue,
std::allocatorllvm::GenericValue const) + 71
16 lli 0x00a0f292
llvm::ExecutionEngine::runFunctionAsMain(llvm::Function
[klee-dev] error in Coreutils Case Study
Dear everyone, I am a new user of klee, today I began to test the coreutils case study following the instructions in klee.llvm.org/TestingCoreutils.html. In the second step (build coreutils with llvm), I got the next error information when I executed *obj-llvm$ make CC=/full/path/to/klee/scripts/klee-gcc* . What should I do to correct the error? Thank you all very much! /home/yq/installed/klee/scripts/klee-gcc -g -Wl,--as-needed -o chmod chmod.o ../lib/libcoreutils.a ../lib/libcoreutils.a /home/yq/installed/klee/scripts/klee-gcc -I. -I../../src -I../lib -I../../lib -g -MT cp.o -MD -MP -MF .deps/cp.Tpo -c -o cp.o ../../src/cp.c mv -f .deps/cp.Tpo .deps/cp.Po /home/yq/installed/klee/scripts/klee-gcc -I. -I../../src -I../lib -I../../lib -g -MT copy.o -MD -MP -MF .deps/copy.Tpo -c -o copy.o ../../src/copy.c mv -f .deps/copy.Tpo .deps/copy.Po /home/yq/installed/klee/scripts/klee-gcc -I. -I../../src -I../lib -I../../lib -g -MT cp-hash.o -MD -MP -MF .deps/cp-hash.Tpo -c -o cp-hash.o ../../src/cp-hash.c mv -f .deps/cp-hash.Tpo .deps/cp-hash.Po /home/yq/installed/klee/scripts/klee-gcc -g -Wl,--as-needed -o cp cp.o copy.o cp-hash.o ../lib/libcoreutils.a ../lib/libcoreutils.a /home/yq/installed/klee/scripts/klee-gcc -I. -I../../src -I../lib -I../../lib -g -MT dd.o -MD -MP -MF .deps/dd.Tpo -c -o dd.o ../../src/dd.c mv -f .deps/dd.Tpo .deps/dd.Po /home/yq/installed/klee/scripts/klee-gcc -g -Wl,--as-needed -o dd dd.o ../lib/libcoreutils.a ../lib/libcoreutils.a -lrt llvm-ld: error: Cannot find library 'rt' make[3]: *** [dd] Error 1 make[3]: Leaving directory `/home/qiu/installed/coreutils-6.11/obj-llvm/src' make[2]: *** [all] Error 2 make[2]: Leaving directory `/home/qiu/installed/coreutils-6.11/obj-llvm/src' make[1]: *** [all-recursive] Error 1 make[1]: Leaving directory `/home/qiu/installed/coreutils-6.11/obj-llvm' Qiuping Yi Institute Of Software Chinese Academy of Sciences ___ klee-dev mailing list klee-dev@keeda.stanford.edu http://keeda.Stanford.EDU/mailman/listinfo/klee-dev
