on the ngspice development list one of the developer
find a likely cygwin bug, see description an test case
--- Dom 12/9/10, Robert Larice ha scritto:
> Hello All,
>
> I need some help with a cygwin Problem.
> Trying to find a bug reported for ngspice,
> I finally could boil the problem down
> to what seems to be a plain cygwin bug.
>
> The attached very small C file, fails
> after some iterations to open a
> certain /proc/.. entry
>
[cut]
>
> Here I get:
> malloc -> 0x7fd30008
> malloc -> 0x7fab0008
> malloc -> 0x7f830008
> malloc -> 0x7f470008
> malloc -> 0x7f1f0008
> malloc -> 0x7ef70008
> /proc/3044/statm: No error
> malloc -> 0x7ecf0008
> /proc/3044/statm: No error
> ...
>
> { No error, nevertheless means,
> the thing can't be read ... }
>
> Thanks,
> Robert Larice
I checked and both 1.7.7 and the latest 20100912
show the issue.
Are we assuming something wrong and missing something else
or is it a bzero bug ?
Marco
// (compile "make CFLAGS='-Wall -O2' t3 && ./t3.exe")
// (man "proc(5)")
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <unistd.h>
#include <fcntl.h>
static void
foo(const char *path)
{
int length = 320000*8;
void *blob = malloc(length);
fprintf(stderr,"malloc -> %p\n", blob);
/* writing to the alloced space causes failure to open /proc/$pid/statm */
if(1)
bzero(blob, length);
{
int hndl = open(path, O_RDONLY);
if(hndl < 0)
perror(path);
else
close(hndl);
}
}
int
main(int argc, char **argv)
{
char proc_statm[100];
int i;
#if 0
// just some info, for debugging purpose
sprintf(proc_statm, "cat /proc/%d/maps | sort", getpid());
system(proc_statm);
// stat doesn't fail
sprintf(proc_statm, "/proc/%d/stat", getpid());
#endif
// statm will fail
sprintf(proc_statm, "/proc/%d/statm", getpid());
for(i=0; i<60; i++)
foo(proc_statm);
exit(0);
}
--
Problem reports: http://cygwin.com/problems.html
FAQ: http://cygwin.com/faq/
Documentation: http://cygwin.com/docs.html
Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple
