Paths

Table of Contentst

-
head/share/man/man8/
-
share/
-
man/
-
man8/
-
rc.subr.8

Differential D14928

Improve documentation of ${name}_limits and ${name}_login_class in rc.subr.8
ClosedPublic
Actions

Authored by 0mp on Apr 2 2018, 12:52 AM.

Details

Reviewers

jilles

Group Reviewers

manpages
Doc Committers

Commits

rS331955: rc.subr.8: Improve documentation of ${name}_limits and ${name}_login_class

Summary

Since the description of ${name}_limits did not mention anything about the default value I decided to improve it. Also, I documented the ${name}_login_class variable, which seems to be used by ${name}_limits only.

${name}_limits has been introduced in D14015.

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

0mp created this revision.Apr 2 2018, 12:52 AM

Herald added a subscriber: imp. · View Herald TranscriptApr 2 2018, 12:52 AM

0mp added a reviewer: Doc Committers.Apr 2 2018, 8:01 PM

0mp added a project: Doc Committers.

Thanks for writing missing documentation.

share/man/man8/rc.subr.8
590 ↗	(On Diff #40976)	I know this is out of scope, strictly speaking, but referring to limits(1) twice is a bit ugly. The topmost reference can just be "Resource limits".
594 ↗	(On Diff #40976)	This looks a bit strange comparing to the code in `rc.subr` which appends the _limits to `-C` _login_class. However, per the limits(1) man page, this is indeed how it works: manual resource specifications override a `-C` or `-U` setting. Resource limits from a login class can be tweaked by specifying `-a`. I'm not sure this (`-a`) detail need be in the man page though.
599 ↗	(On Diff #40976)	This should be `.Li` (or possibly `.Dq Li`) instead of `.Em`. The two `.Em` usages in the description of oomprotect below are also wrong.

@jilles Thank you for your feedback!

I've applied you suggestions. I did not change the description of the ${name}_limits mechanism, however. I don't know if we should include a note about the -a flag. Isn't the reference to limits(1) enough?

0mp marked 2 inline comments as done.Apr 2 2018, 10:33 PM

jilles accepted this revision.Apr 3 2018, 7:46 PM

This revision is now accepted and ready to land.Apr 3 2018, 7:46 PM

Closed by commit rS331955: rc.subr.8: Improve documentation of ${name}_limits and ${name}_login_class (authored by jilles). · Explain WhyApr 3 2018, 9:45 PM

This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

Path

Size

head/

share/

man/

man8/

rc.subr.8

16 lines

Diff 41062

View Options

head/share/man/man8/rc.subr.8

This file was added.

				#include <atf-c.h>
				#include <alias.h>
				#include <signal.h>
				#include <stdio.h>
				#include <stdlib.h>

				static int randcmp(const void a, const void b) {
				int res, r = rand();

				kpUnsubmitted Done Inline Actions I wonder if it's worth printing the random seed somewhere, or even using a fixed seed. If this triggers a bug it may end up being nondeterministic (and thus difficult to debug). kp: I wonder if it's worth printing the random seed somewhere, or even using a fixed seed. If this…
				(void)a;
				(void)b;
				res = (r/4 < RAND_MAX/9) ? 1
				: (r/5 < RAND_MAX/9) ? 0
				: -1;
				return (res);
				}

				ATF_TC(destroynull);
				ATF_TC_HEAD(destroynull, env)
				{
				atf_tc_set_md_var(env, "descr", "Destroy the NULL instance");
				}
				ATF_TC_BODY(destroynull, dummy)
				{
				atf_tc_expect_signal(SIGSEGV, "Code expects valid pointer.");
				LibAliasUninit(NULL);
				}

				ATF_TC(singleinit);
				ATF_TC_HEAD(singleinit, env)
				{
				atf_tc_set_md_var(env, "descr", "Create an instance");
				}
				ATF_TC_BODY(singleinit, dummy)
				{
				struct libalias *la;

				la = LibAliasInit(NULL);
				ATF_CHECK_MSG(la != NULL, "Creating an instance failed.");
				LibAliasUninit(la);
				}

				ATF_TC(multiinit);
				ATF_TC_HEAD(multiinit, env)
				{
				atf_tc_set_md_var(env, "descr", "Recreate an instance multiple times");
				}
				ATF_TC_BODY(multiinit, dummy)
				{
				struct libalias *la;
				int i;

				la = LibAliasInit(NULL);
				for(i = 1; i < 30; i++) {
				struct libalias *lo = la;

				la = LibAliasInit(la);
				ATF_CHECK_MSG(la == lo, "Recreating moved the instance around: %d", i);
				}
				LibAliasUninit(la);
				}

				ATF_TC(multiinstance);
				ATF_TC_HEAD(multiinstance, env)
				{
				atf_tc_set_md_var(env, "descr", "Create and destoy multiple instances.");
				}
				ATF_TC_BODY(multiinstance, dummy)
				{
				struct libalias *la[300];
				int const num_instances = sizeof(la) / sizeof(*la);
				int i;

				for (i = 0; i < num_instances; i++) {
				la[i] = LibAliasInit(NULL);
				ATF_CHECK_MSG(la[i] != NULL, "Creating instance %d failed.", i);
				}

				qsort(la, num_instances, sizeof(*la), randcmp);

				for (i = 0; i < num_instances; i++)
				LibAliasUninit(la[i]);
				}

				ATF_TP_ADD_TCS(aliasdb)
				{
				/* Use "dd if=/dev/random bs=2 count=1 \| od -x" to reproduce */
				srand(0x5ac4);

				ATF_TP_ADD_TC(aliasdb, destroynull);
				ATF_TP_ADD_TC(aliasdb, singleinit);
				ATF_TP_ADD_TC(aliasdb, multiinit);
				ATF_TP_ADD_TC(aliasdb, multiinstance);

				return atf_no_error();
				}

Improve documentation of ${name}_limits and ${name}_login_class in rc.subr.8ClosedPublicActions