I need to use strerror_r to translate error numbers into human readable messages compiled with g++ on Linux Debian Bullseye. The man page notes:
int strerror_r(int errnum, char *buf, size_t buflen);
/* XSI-compliant */
char *strerror_r(int errnum, char *buf, size_t buflen);
/* GNU-specific */
strerror_r():
The XSI-compliant version is provided if:
(_POSIX_C_SOURCE >= 200112L) && ! _GNU_SOURCE
Otherwise, the GNU-specific version is provided.
We have two different types of the return value: int or char* depending on the version defined by _POSIX_C_SOURCE. I have this small test program:
~$ cat strerror_r.c
#include <string.h>
#include <stdio.h>
// #define _POSIX_C_SOURCE 200112L
// #undef _GNU_SOURCE
#define ERROR_BUFFER_LEN (size_t)256
int main(int argc, char **argv)
{
#if _POSIX_C_SOURCE < 200112L
char* ret;
#else
int ret;
#endif
char errorBuffer[ERROR_BUFFER_LEN];
int errno;
errno = 0;
ret = strerror_r(errno, errorBuffer, ERROR_BUFFER_LEN);
fprintf(stderr, "Error message by pointer = '%s'\n", ret);
fprintf(stderr, "Content of errorBuffer = '%s'\n", errorBuffer);
return 0;
}
If I compile it with gcc everything is as expected:
$ gcc strerror_r.c && ./a.out; rm a.out
Error message by pointer = '(null)'
Content of errorBuffer = 'Success'
If I compile it with g++ I get this:
$ g++ strerror_r.c && ./a.out; rm a.out
strerror_r.c: In function ‘int main(int, char**)’:
strerror_r.c:24:21: error: invalid conversion from ‘char*’ to ‘int’ [-fpermissive]
24 | ret = strerror_r(errno, errorBuffer, ERROR_BUFFER_LEN);
| ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| |
| char*
rm: cannot remove 'a.out': No such file or directory
If I try to force the needed version by uncommenting
#define _POSIX_C_SOURCE 200112L
#undef _GNU_SOURCE
I get:
$ g++ strerror_r.c && ./a.out; rm a.out
strerror_r.c:7: warning: "_POSIX_C_SOURCE" redefined
7 | #define _POSIX_C_SOURCE 200112L
|
In file included from /usr/include/x86_64-linux-gnu/bits/libc-header-start.h:33,
from /usr/include/string.h:26,
from strerror_r.c:3:
/usr/include/features.h:281: note: this is the location of the previous definition
281 | # define _POSIX_C_SOURCE 200809L
|
strerror_r.c: In function ‘int main(int, char**)’:
strerror_r.c:24:21: error: invalid conversion from ‘char*’ to ‘int’ [-fpermissive]
24 | ret = strerror_r(errno, errorBuffer, ERROR_BUFFER_LEN);
| ~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
| |
| char*
rm: cannot remove 'a.out': No such file or directory
What I'm missing here? Why g++ does not compile the default thread save version of strerror_r? I need that version. How can I fix it?
Reference
Feature Test Macros
You need to specify the #define and #undef directives before you include any header files, so the first few lines should look like this:
#define _POSIX_C_SOURCE 200112L
#undef _GNU_SOURCE
#include <string.h>
#include <stdio.h>
That's because those header files or internal header files they include need those values defined to choose the proper variant. If you define them after including the headers, the headers don't see the right values and they don't include the version you want.
Often people specify these values on the command like with the -D and -U arguments so they are always specified before header files are included.
Related
I'm trying to port some Solaris serial port code to Linux, however the XOn / XOff symbol typically found in termios.h seems to be missing.
// ...
#include <unistd.h>
#include <fcntl.h>
#include <errno.h>
#include <termios.h>
// ...
int config_port(int fd, int timeout)
{
struct termios options; /* Holds the port option flags */
int status; /* Holds return value of system calls */
int min_chars; /* Holds the minimum number of characters to read
* before returning. */
// ...
options.c_cflag &= ~( CRTSXOFF | CRTSCTS ); // <<-- HERE
Of course, Linux GCC doesn't know CRTSXOFF:
# gcc -c -g serial.c
serial.c: In function ‘config_port’:
serial.c:125:25: error: ‘CRTSXOFF’ undeclared (first use in this function)
125 | options.c_cflag &= ~( CRTSXOFF | CRTSCTS );
On Solaris, this symbol is certainly defined in termios.h:
# find /usr/include -iname \*.h -exec grep CRTSXOFF {} /dev/null \;
/usr/include/sys/termios.h:#define CRTSXOFF 010000000000
But on Linux the same command finds me nothing.
Is this symbol called something else on Linux?
Found it: under Linux CRTSXOFF is implemented as IXOFF.
mode_t mode = S_IRWXU | S_IRWXG | S_IRWXO;
shm_fd = shm_open("/ipc_shm", O_CREAT | O_RDWR, mode);
This works, returns 4 for shm_fd. The same process then calls a library function that also calls
fd = shm_open("/ipc_shm", O_RDWR, 0);
This one fails with errno set to 2, i.e. ENOENT (No such file or directory). There is no shm_unlink call in the middle. Any idea why the second call is failing. Appreciate your help.
my test.c:
#include <sys/mman.h>
#include <sys/stat.h> /* For mode constants */
#include <fcntl.h> /* For O_* constants */
int main (int argc, char *argv[])
{
mode_t mode = S_IRWXU | S_IRWXG | S_IRWXO;
int shm_fd = shm_open("/ipc_shm", O_CREAT | O_RDWR, mode);
int fd = shm_open("/ipc_shm", O_RDWR, 0);
return 0;
}
compiled with gcc test.c -Wall -lrt works as expected:
$strace ./a.out
....
statfs("/dev/shm/", {f_type=0x1021994, f_bsize=4096, f_blocks=22290, f_bfree=22290, f_bavail=22290, f_files=55725, f_ffree=55723, f_fsid={0, 0}, f_namelen=255, f_frsize=4096}) = 0
futex(0xb6f5d1c0, FUTEX_WAKE_PRIVATE, 2147483647) = 0
open("/dev/shm/ipc_shm", O_RDWR|O_CREAT|O_NOFOLLOW|O_CLOEXEC, 0777) = 3
fcntl64(3, F_GETFD) = 0x1 (flags FD_CLOEXEC)
open("/dev/shm/ipc_shm", O_RDWR|O_NOFOLLOW|O_CLOEXEC) = 4
exit_group(0)
Please run strace on your application, and search for all occurrences of ipc, and maybe chroot(). Maybe something unlinks the file?
I'm trying to create a FIFO named pipe using the mknod() command:
int main() {
char* file="pipe.txt";
int state;
state = mknod(file, S_IFIFO & 0777, 0);
printf("%d",state);
return 0;
}
But the file is not created in my current directory. I tried listing it by ls -l . State returns -1.
I found similar questions here and on other sites and I've tried the solution that most suggested:
int main() {
char* file="pipe.txt";
int state;
unlink(file);
state = mknod(file, S_IFIFO & 0777, 0);
printf("%d",state);
return 0;
}
This made no difference though and the error remains. Am I doing something wrong here or is there some sort of system intervention which is causing this problem?
Help.. Thanks in advance
You are using & to set the file type instead of |. From the docs:
The file type for path is OR'ed into the mode argument, and the
application shall select one of the following symbolic
constants...
Try this:
state = mknod(file, S_IFIFO | 0777, 0);
Because this works:
#include <stdio.h>
#include <stdlib.h>
#include <unistd.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <unistd.h>
int main() {
char* file="pipe.txt";
int state;
unlink(file);
state = mknod(file, S_IFIFO | 0777, 0);
printf("state %d\n", state);
return 0;
}
Compile it:
gcc -o fifo fifo.c
Run it:
$ strace -e trace=mknod ./fifo
mknod("pipe.txt", S_IFIFO|0777) = 0
state 0
+++ exited with 0 +++
See the result:
$ ls -l pipe.txt
prwxrwxr-x. 1 lars lars 0 Jul 16 12:54 pipe.txt
I am trying to execute "word count" command on file given by absolute path - "/home/aaa/xxzz.txt" . I have closed the stdin so as to take input from file but the program doesn't give any output .
Also if I add some statement after "execve" command, it is also getting executed . Shouldn't the program exit after execve ?
int main()
{
char *envp[]={NULL };
int fd=open("/home/aaa/xxzz.txt",O_RDONLY);
close(0);
dup(fd);
char *param[]={ "/bin/wc",NULL } ;
execve("/bin/wc",param,envp);
}
Probably wc does not live in /bin (except for some systems which symlink that to /usr/bin, because wc normally lives in the latter). If I change the path in your example to /usr/bin/wc, it works for me:
#include <unistd.h>
#include <fcntl.h>
int
main()
{
char *envp[] = {NULL};
int fd = open("/home/aaa/xxzz.txt", O_RDONLY);
close(0);
dup(fd);
char *program = "/usr/bin/wc";
char *param[] = {program,NULL};
execve(program, param, envp);
}
I was curious what kind of buffer sizes write() and read() could handle on Linux/OSX/FreeBSD, so I started playing around with dumb programs like the following:
#include <unistd.h>
#include <fcntl.h>
#include <stdio.h>
#include <string.h>
#include <sys/stat.h>
int main( void ) {
size_t s = 8*1024*1024 - 16*1024;
while( 1 ) {
s += 1024;
int f = open( "test.txt", O_CREAT | O_WRONLY | O_TRUNC, S_IRUSR | S_IWUSR | S_IXUSR );
char mem[s];
size_t written = write( f, &mem[0], s );
close( f );
printf( "(%ld) %lu\n", sizeof(size_t), written );
}
return 0;
}
This allowed me to test how close to a seeming "8MB barrier" I could get before segfaulting. Somewhere around the 8MB mark, my program dies, here's an example output:
(8) 8373248
(8) 8374272
(8) 8375296
(8) 8376320
(8) 8377344
(8) 8378368
(8) 8379392
(8) 8380416
(8) 8381440
(8) 8382464
Segmentation fault: 11
This is the same on OSX and Linux, however my FreeBSD VM is not only much faster at running this test, it also can go on for quite a ways! I've successfully tested it up to 511MB, which is just a ridiculous amount of data to write in one call.
What is it that makes the write() call segfault, and how can I figure out the maximum amount that I can possibly write() in a single call, without doing something ridiculous like I'm doing right now?
(Note, all three operating systems are 64-bit, OSX 10.7.3, Ubuntu 11.10, FreeBSD 9.0)
The fault isn't within write(), it's a stack overflow. Try this:
#include <stdlib.h>
#include <unistd.h>
#include <fcntl.h>
#include <stdio.h>
#include <string.h>
#include <sys/stat.h>
int main( void )
{
void *mem;
size_t s = 512*1024*1024 - 16*1024;
while( 1 )
{
s += 1024;
int f = open( "test.txt", O_CREAT | O_WRONLY | O_TRUNC, S_IRUSR | S_IWUSR | S_IXUSR );
mem = malloc(s);
size_t written = write( f, mem, s );
free(mem);
close( f );
printf( "(%ld) %lu\n", sizeof(size_t), written );
}
return 0;
}