Hi,
Revision 1.81 from kern_pledge.c still doesn't solve the issue, miblen
should be changed to 2 since machine.c from top(1) code is:
268 } else {
269 int cp_time_mib[] = {CTL_KERN, KERN_CPTIME};
270 long cp_time_tmp[CPUSTATES];
271
272 size = sizeof(cp_time_tmp);
273 if (sysctl(cp_time_mib, 2, cp_time_tmp, &size, NULL, 0) < 0)
274 warn("sysctl kern.cp_time failed");
By applying the patch below top(1) works again:
Index: kern_pledge.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_pledge.c,v
retrieving revision 1.81
diff -u -p -u -r1.81 kern_pledge.c
--- kern_pledge.c 28 Oct 2015 02:12:54 -0000 1.81
+++ kern_pledge.c 28 Oct 2015 07:44:20 -0000
@@ -908,7 +908,7 @@ pledge_sysctl_check(struct proc *p, int
if (miblen == 2 && /* kern.loadavg */
mib[0] == CTL_VM && mib[1] == VM_LOADAVG)
return (0);
- if (miblen == 3 && /* kern.cptime */
+ if (miblen == 2 && /* kern.cptime */
mib[0] == CTL_KERN && mib[1] == KERN_CPTIME)
return (0);
if (miblen == 3 && /* kern.cptime2 */
Best regards,
Ricardo Mestre
On 27/10/2015 19:37, Ricardo Mestre wrote:
Hi people,
I can confirm this regress, just updated the kernel and top and had
the same issue, but this diff seems to solve it, I just don't know if
it's the right place to put it or not:
Index: kern_pledge.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_pledge.c,v
retrieving revision 1.80
diff -u -p -u -r1.80 kern_pledge.c
--- kern_pledge.c 26 Oct 2015 17:52:19 -0000 1.80
+++ kern_pledge.c 27 Oct 2015 19:32:09 -0000
@@ -911,6 +911,9 @@ pledge_sysctl_check(struct proc *p, int
if (miblen == 3 && /*
kern.cptime2 */
mib[0] == CTL_KERN && mib[1] == KERN_CPTIME2)
return (0);
+ if (miblen == 2 && /* kern.cp_time */
+ mib[0] == CTL_KERN && mib[1] == KERN_CPTIME)
+ return (0);
}
if ((p->p_p->ps_pledge & PLEDGE_PS)) {
Best regards,
Ricardo Mestre
On 27/10/2015 19:00, Mike wrote:
OpenBSD 5.8-current (GENERIC) #1: Tue Oct 27 12:31:10 EDT 2015
m...@otest.24cl.home:/usr/src/sys/arch/amd64/compile/GENERIC
I didn't see anything in current.html that may affect this.
I downloaded the Oct 20 snapshot. Then I updated the source to current
this morning. After the build, top crashes immediately upon invocation.
# top
Abort (core dumped)
In messages.log, I see two lines:
Oct 27 14:52:22 otest /bsd: top(12603): sysctl 2: 1 40 -2129088583 -1
981777920 -255
Oct 27 14:52:22 otest /bsd: top(12603): syscall 202 "stdio"
which looks like it may be pledge output.
If I need to do something else to track this let me know. I can make
the core dump available to download, if needed.
thx.
