On Thu, 15 May 1997, T. William Wells wrote:
> System is FreeBSD2.1.5, exim 1.62. The pipe transport is including
> the exim group in the list of groups. Specifying user *and*
> initgroups in the transport is a workaround.
Rats. FreeBSD is one of the two OS that (in a recent set of tests) do
not recognize
setgroups(0, NULL)
The other was BSDI. That is why the code in Exim now reads
if (setgroups(0, NULL) != 0)
{
if (setgroups(1, group_list) != 0)
{
fprintf(stderr, "exim: setgroups() failed: %s\n", strerror(errno));
exit(EXIT_FAILURE);
}
}
I thought that the experiments showed that on such systems, the first
group in the group list is the current group, and that a subsequent use
of setgid() would change it. My test program explicitly checked for
that. Your report seems to suggest that it does not. Here's the test
program again:
#include <stdio.h>
#include <unistd.h>
#include <limits.h>
#include <errno.h>
#include <stdlib.h>
int main(void)
{
int group_count, i;
gid_t group_list[NGROUPS_MAX];
printf("count = %d\n", group_count = getgroups(NGROUPS_MAX, group_list));
for (i = 0; i < group_count; i++) printf("%d ", group_list[i]);
printf("\n");
if (setgroups(0, NULL) != 0)
{
printf("setgroups(0,NULL) failed: %s\n", strerror(errno));
if (setgroups(1, group_list) != 0)
{
printf("setgroups(1, group_list) failed: %s\n", strerror(errno));
return 1;
}
else printf("setgroups(1,grouplist) succeeded\n");
}
else printf("setgroups(0,NULL) succeeded\n");
printf("count = %d\n", group_count = getgroups(NGROUPS_MAX, group_list));
for (i = 0; i < group_count; i++) printf("%d ", group_list[i]);
printf("\n");
if (setgid(8) != 0)
{
printf("setgid(8) failed: %s\n", strerror(errno));
return 1;
}
printf("gid = %d\n", getgid());
printf("count = %d\n", group_count = getgroups(NGROUPS_MAX, group_list));
for (i = 0; i < group_count; i++) printf("%d ", group_list[i]);
printf("\n");
return 0;
}
Could you run that as root and let me see the output, please?
--
Philip Hazel University Computing Service,
ph10@??? New Museums Site, Cambridge CB2 3QG,
P.Hazel@??? England. Phone: +44 1223 334714
