blob: 6c58cd8bbbae71ad86592b4627be483879374a9c [file] [log] [blame]
Steven Rostedta75fece2010-11-02 14:58:27 -04001#
Steven Rostedta57419b2010-11-02 15:13:54 -04002# Config file for ktest.pl
Steven Rostedta75fece2010-11-02 14:58:27 -04003#
4# Note, all paths must be absolute
5#
6
Steven Rostedta57419b2010-11-02 15:13:54 -04007# Options set in the beginning of the file are considered to be
8# default options. These options can be overriden by test specific
9# options, with the following exceptions:
Steven Rostedta75fece2010-11-02 14:58:27 -040010#
Steven Rostedta75fece2010-11-02 14:58:27 -040011# LOG_FILE
12# CLEAR_LOG
13# POWEROFF_ON_SUCCESS
14# REBOOT_ON_SUCCESS
15#
Steven Rostedta57419b2010-11-02 15:13:54 -040016# Test specific options are set after the label:
17#
18# TEST_START
19#
20# The options after a TEST_START label are specific to that test.
21# Each TEST_START label will set up a new test. If you want to
22# perform a test more than once, you can add the ITERATE label
23# to it followed by the number of times you want that test
24# to iterate. If the ITERATE is left off, the test will only
25# be performed once.
26#
27# TEST_START ITERATE 10
28#
29# You can skip a test by adding SKIP (before or after the ITERATE
30# and number)
31#
32# TEST_START SKIP
33#
34# TEST_START SKIP ITERATE 10
35#
36# TEST_START ITERATE 10 SKIP
37#
38# The SKIP label causes the options and the test itself to be ignored.
39# This is useful to set up several different tests in one config file, and
40# only enabling the ones you want to use for a current test run.
41#
42# You can add default options anywhere in the file as well
43# with the DEFAULTS tag. This allows you to have default options
44# after the test options to keep the test options at the top
45# of the file. You can even place the DEFAULTS tag between
46# test cases (but not in the middle of a single test case)
47#
48# TEST_START
49# MIN_CONFIG = /home/test/config-test1
50#
51# DEFAULTS
52# MIN_CONFIG = /home/test/config-default
53#
54# TEST_START ITERATE 10
55#
56# The above will run the first test with MIN_CONFIG set to
57# /home/test/config-test-1. Then 10 tests will be executed
58# with MIN_CONFIG with /home/test/config-default.
59#
60# You can also disable defaults with the SKIP option
61#
62# DEFAULTS SKIP
63# MIN_CONFIG = /home/test/config-use-sometimes
64#
65# DEFAULTS
66# MIN_CONFIG = /home/test/config-most-times
67#
68# The above will ignore the first MIN_CONFIG. If you want to
69# use the first MIN_CONFIG, remove the SKIP from the first
70# DEFAULTS tag and add it to the second. Be careful, options
71# may only be declared once per test or default. If you have
72# the same option name under the same test or as default
73# ktest will fail to execute, and no tests will run.
74#
Steven Rostedt3d1cc412011-09-30 22:14:21 -040075# DEFAULTS OVERRIDE
76#
77# Options defined in the DEFAULTS section can not be duplicated
78# even if they are defined in two different DEFAULT sections.
79# This is done to catch mistakes where an option is added but
80# the previous option was forgotten about and not commented.
81#
82# The OVERRIDE keyword can be added to a section to allow this
83# section to override other DEFAULT sections values that have
84# been defined previously. It will only override options that
85# have been defined before its use. Options defined later
86# in a non override section will still error. The same option
87# can not be defined in the same section even if that section
88# is marked OVERRIDE.
89#
Steven Rostedtab7a3f52011-09-30 20:24:07 -040090#
91#
Steven Rostedt45d73a52011-09-30 19:44:53 -040092# Both TEST_START and DEFAULTS sections can also have the IF keyword
93# The value after the IF must evaluate into a 0 or non 0 positive
94# integer, and can use the config variables (explained below).
95#
96# DEFAULTS IF ${IS_X86_32}
97#
98# The above will process the DEFAULTS section if the config
99# variable IS_X86_32 evaluates to a non zero positive integer
100# otherwise if it evaluates to zero, it will act the same
101# as if the SKIP keyword was used.
102#
103# The ELSE keyword can be used directly after a section with
104# a IF statement.
105#
106# TEST_START IF ${RUN_NET_TESTS}
107# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-network
108#
109# ELSE
110#
111# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-normal
112#
113#
114# The ELSE keyword can also contain an IF statement to allow multiple
115# if then else sections. But all the sections must be either
116# DEFAULT or TEST_START, they can not be a mixture.
117#
118# TEST_START IF ${RUN_NET_TESTS}
119# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-network
120#
121# ELSE IF ${RUN_DISK_TESTS}
122# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-tests
123#
124# ELSE IF ${RUN_CPU_TESTS}
125# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-cpu
126#
127# ELSE
128# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-network
129#
Steven Rostedtab7a3f52011-09-30 20:24:07 -0400130# The if statement may also have comparisons that will and for
131# == and !=, strings may be used for both sides.
132#
133# BOX_TYPE := x86_32
134#
135# DEFAULTS IF ${BOX_TYPE} == x86_32
136# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-32
137# ELSE
138# BUILD_TYPE = useconfig:${CONFIG_DIR}/config-64
139#
Steven Rostedt9900b5d2011-09-30 22:41:14 -0400140# The DEFINED keyword can be used by the IF statements too.
141# It returns true if the given config variable or option has been defined
142# or false otherwise.
143#
144#
145# DEFAULTS IF DEFINED USE_CC
146# CC := ${USE_CC}
147# ELSE
148# CC := gcc
149#
150#
151# As well as NOT DEFINED.
152#
153# DEFAULTS IF NOT DEFINED MAKE_CMD
154# MAKE_CMD := make ARCH=x86
155#
156#
Steven Rostedt8d735212011-10-17 11:36:44 -0400157# And/or ops (&&,||) may also be used to make complex conditionals.
158#
159# TEST_START IF (DEFINED ALL_TESTS || ${MYTEST} == boottest) && ${MACHINE} == gandalf
160#
Jesper Juhl11c38b72012-04-16 19:40:24 +0200161# Notice the use of parentheses. Without any parentheses the above would be
Steven Rostedt8d735212011-10-17 11:36:44 -0400162# processed the same as:
163#
164# TEST_START IF DEFINED ALL_TESTS || (${MYTEST} == boottest && ${MACHINE} == gandalf)
165#
166#
Steven Rostedt2ed3b162011-09-30 21:00:00 -0400167#
168# INCLUDE file
169#
170# The INCLUDE keyword may be used in DEFAULT sections. This will
171# read another config file and process that file as well. The included
172# file can include other files, add new test cases or default
173# statements. Config variables will be passed to these files and changes
174# to config variables will be seen by top level config files. Including
175# a file is processed just like the contents of the file was cut and pasted
176# into the top level file, except, that include files that end with
177# TEST_START sections will have that section ended at the end of
178# the include file. That is, an included file is included followed
179# by another DEFAULT keyword.
180#
181# Unlike other files referenced in this config, the file path does not need
182# to be absolute. If the file does not start with '/', then the directory
183# that the current config file was located in is used. If no config by the
184# given name is found there, then the current directory is searched.
185#
186# INCLUDE myfile
187# DEFAULT
188#
189# is the same as:
190#
191# INCLUDE myfile
192#
193# Note, if the include file does not contain a full path, the file is
194# searched first by the location of the original include file, and then
195# by the location that ktest.pl was executed in.
196#
Steven Rostedta75fece2010-11-02 14:58:27 -0400197
Steven Rostedt77d942c2011-05-20 13:36:58 -0400198#### Config variables ####
199#
200# This config file can also contain "config variables".
201# These are assigned with ":=" instead of the ktest option
202# assigment "=".
203#
204# The difference between ktest options and config variables
205# is that config variables can be used multiple times,
206# where each instance will override the previous instance.
207# And that they only live at time of processing this config.
208#
209# The advantage to config variables are that they can be used
210# by any option or any other config variables to define thing
211# that you may use over and over again in the options.
212#
213# For example:
214#
215# USER := root
216# TARGET := mybox
217# TEST_CASE := ssh ${USER}@${TARGET} /path/to/my/test
218#
219# TEST_START
220# MIN_CONFIG = config1
221# TEST = ${TEST_CASE}
222#
223# TEST_START
224# MIN_CONFIG = config2
225# TEST = ${TEST_CASE}
226#
227# TEST_CASE := ssh ${USER}@${TARGET} /path/to/my/test2
228#
229# TEST_START
230# MIN_CONFIG = config1
231# TEST = ${TEST_CASE}
232#
233# TEST_START
234# MIN_CONFIG = config2
235# TEST = ${TEST_CASE}
236#
237# TEST_DIR := /home/me/test
238#
239# BUILD_DIR = ${TEST_DIR}/linux.git
240# OUTPUT_DIR = ${TEST_DIR}/test
241#
242# Note, the config variables are evaluated immediately, thus
243# updating TARGET after TEST_CASE has been assigned does nothing
244# to TEST_CASE.
245#
246# As shown in the example, to evaluate a config variable, you
247# use the ${X} convention. Simple $X will not work.
248#
249# If the config variable does not exist, the ${X} will not
250# be evaluated. Thus:
251#
252# MAKE_CMD = PATH=/mypath:${PATH} make
253#
254# If PATH is not a config variable, then the ${PATH} in
255# the MAKE_CMD option will be evaluated by the shell when
256# the MAKE_CMD option is passed into shell processing.
Steven Rostedta57419b2010-11-02 15:13:54 -0400257
Steven Rostedt2a625122011-05-20 15:48:59 -0400258#### Using options in other options ####
259#
260# Options that are defined in the config file may also be used
261# by other options. All options are evaulated at time of
262# use (except that config variables are evaluated at config
263# processing time).
264#
265# If an ktest option is used within another option, instead of
266# typing it again in that option you can simply use the option
267# just like you can config variables.
268#
269# MACHINE = mybox
270#
271# TEST = ssh root@${MACHINE} /path/to/test
272#
273# The option will be used per test case. Thus:
274#
275# TEST_TYPE = test
276# TEST = ssh root@{MACHINE}
277#
278# TEST_START
279# MACHINE = box1
280#
281# TEST_START
282# MACHINE = box2
283#
284# For both test cases, MACHINE will be evaluated at the time
285# of the test case. The first test will run ssh root@box1
286# and the second will run ssh root@box2.
287
Steven Rostedta57419b2010-11-02 15:13:54 -0400288#### Mandatory Default Options ####
289
290# These options must be in the default section, although most
291# may be overridden by test options.
Steven Rostedta75fece2010-11-02 14:58:27 -0400292
293# The machine hostname that you will test
294#MACHINE = target
295
296# The box is expected to have ssh on normal bootup, provide the user
297# (most likely root, since you need privileged operations)
298#SSH_USER = root
299
300# The directory that contains the Linux source code
301#BUILD_DIR = /home/test/linux.git
302
303# The directory that the objects will be built
304# (can not be same as BUILD_DIR)
305#OUTPUT_DIR = /home/test/build/target
306
307# The location of the compiled file to copy to the target
308# (relative to OUTPUT_DIR)
309#BUILD_TARGET = arch/x86/boot/bzImage
310
311# The place to put your image on the test machine
312#TARGET_IMAGE = /boot/vmlinuz-test
313
314# A script or command to reboot the box
Steven Rostedta57419b2010-11-02 15:13:54 -0400315#
Steven Rostedta75fece2010-11-02 14:58:27 -0400316# Here is a digital loggers power switch example
317#POWER_CYCLE = wget --no-proxy -O /dev/null -q --auth-no-challenge 'http://admin:admin@power/outlet?5=CCL'
Steven Rostedta57419b2010-11-02 15:13:54 -0400318#
Steven Rostedta75fece2010-11-02 14:58:27 -0400319# Here is an example to reboot a virtual box on the current host
320# with the name "Guest".
Steven Rostedta57419b2010-11-02 15:13:54 -0400321#POWER_CYCLE = virsh destroy Guest; sleep 5; virsh start Guest
Steven Rostedta75fece2010-11-02 14:58:27 -0400322
323# The script or command that reads the console
Steven Rostedta57419b2010-11-02 15:13:54 -0400324#
Steven Rostedta75fece2010-11-02 14:58:27 -0400325# If you use ttywatch server, something like the following would work.
326#CONSOLE = nc -d localhost 3001
Steven Rostedta57419b2010-11-02 15:13:54 -0400327#
Steven Rostedta75fece2010-11-02 14:58:27 -0400328# For a virtual machine with guest name "Guest".
Steven Rostedta57419b2010-11-02 15:13:54 -0400329#CONSOLE = virsh console Guest
Steven Rostedta75fece2010-11-02 14:58:27 -0400330
Steven Rostedt (Red Hat)298a0d12013-12-03 11:57:23 -0500331# Signal to send to kill console.
332# ktest.pl will create a child process to monitor the console.
333# When the console is finished, ktest will kill the child process
334# with this signal.
335# (default INT)
336#CLOSE_CONSOLE_SIGNAL = HUP
337
Steven Rostedta75fece2010-11-02 14:58:27 -0400338# Required version ending to differentiate the test
339# from other linux builds on the system.
340#LOCALVERSION = -test
341
Steven Rostedta15ba912012-11-13 14:30:37 -0500342# For REBOOT_TYPE = grub2, you must specify where the grub.cfg
343# file is. This is the file that is searched to find the menu
344# option to boot to with GRUB_REBOOT
345#GRUB_FILE = /boot/grub2/grub.cfg
346
347# The tool for REBOOT_TYPE = grub2 to set the next reboot kernel
348# to boot into (one shot mode).
349# (default grub2_reboot)
350#GRUB_REBOOT = grub2_reboot
351
Steven Rostedta75fece2010-11-02 14:58:27 -0400352# The grub title name for the test kernel to boot
Steven Rostedta15ba912012-11-13 14:30:37 -0500353# (Only mandatory if REBOOT_TYPE = grub or grub2)
Steven Rostedta75fece2010-11-02 14:58:27 -0400354#
Steven Rostedta57419b2010-11-02 15:13:54 -0400355# Note, ktest.pl will not update the grub menu.lst, you need to
356# manually add an option for the test. ktest.pl will search
357# the grub menu.lst for this option to find what kernel to
358# reboot into.
359#
Steven Rostedta75fece2010-11-02 14:58:27 -0400360# For example, if in the /boot/grub/menu.lst the test kernel title has:
361# title Test Kernel
Steven Rostedta57419b2010-11-02 15:13:54 -0400362# kernel vmlinuz-test
Steven Rostedta15ba912012-11-13 14:30:37 -0500363#
364# For grub2, a search of top level "menuentry"s are done. No
365# submenu is searched. The menu is found by searching for the
366# contents of GRUB_MENU in the line that starts with "menuentry".
367# You may want to include the quotes around the option. For example:
368# for: menuentry 'Test Kernel'
369# do a: GRUB_MENU = 'Test Kernel'
370# For customizing, add your entry in /etc/grub.d/40_custom.
371#
Steven Rostedta75fece2010-11-02 14:58:27 -0400372#GRUB_MENU = Test Kernel
373
Steven Rostedt77869542012-12-11 17:37:41 -0500374# For REBOOT_TYPE = syslinux, the name of the syslinux executable
375# (on the target) to use to set up the next reboot to boot the
376# test kernel.
377# (default extlinux)
378#SYSLINUX = syslinux
379
380# For REBOOT_TYPE = syslinux, the path that is passed to to the
381# syslinux command where syslinux is installed.
382# (default /boot/extlinux)
383#SYSLINUX_PATH = /boot/syslinux
384
385# For REBOOT_TYPE = syslinux, the syslinux label that references the
386# test kernel in the syslinux config file.
387# (default undefined)
388#SYSLINUX_LABEL = "test-kernel"
389
Steven Rostedta75fece2010-11-02 14:58:27 -0400390# A script to reboot the target into the test kernel
Steven Rostedt96f6a0d2011-12-23 00:24:51 -0500391# This and SWITCH_TO_TEST are about the same, except
392# SWITCH_TO_TEST is run even for REBOOT_TYPE = grub.
393# This may be left undefined.
394# (default undefined)
Steven Rostedta75fece2010-11-02 14:58:27 -0400395#REBOOT_SCRIPT =
396
397#### Optional Config Options (all have defaults) ####
398
Steven Rostedta57419b2010-11-02 15:13:54 -0400399# Start a test setup. If you leave this off, all options
400# will be default and the test will run once.
401# This is a label and not really an option (it takes no value).
402# You can append ITERATE and a number after it to iterate the
403# test a number of times, or SKIP to ignore this test.
404#
405#TEST_START
406#TEST_START ITERATE 5
407#TEST_START SKIP
Steven Rostedta75fece2010-11-02 14:58:27 -0400408
Steven Rostedtdc895682010-11-02 15:22:53 -0400409# Have the following options as default again. Used after tests
410# have already been defined by TEST_START. Optionally, you can
411# just define all default options before the first TEST_START
412# and you do not need this option.
413#
414# This is a label and not really an option (it takes no value).
415# You can append SKIP to this label and the options within this
416# section will be ignored.
417#
418# DEFAULTS
419# DEFAULTS SKIP
420
Steven Rostedt921ed4c2012-07-19 15:18:27 -0400421# If you want to execute some command before the first test runs
422# you can set this option. Note, it can be set as a default option
423# or an option in the first test case. All other test cases will
424# ignore it. If both the default and first test have this option
425# set, then the first test will take precedence.
426#
427# default (undefined)
428#PRE_KTEST = ${SSH} ~/set_up_test
429
430# If you want to execute some command after all the tests have
431# completed, you can set this option. Note, it can be set as a
432# default or any test case can override it. If multiple test cases
433# set this option, then the last test case that set it will take
434# precedence
435#
436# default (undefined)
437#POST_KTEST = ${SSH} ~/dismantle_test
438
Steven Rostedta75fece2010-11-02 14:58:27 -0400439# The default test type (default test)
440# The test types may be:
Steven Rostedtcd8e3682011-08-18 16:35:44 -0400441# build - only build the kernel, do nothing else
442# install - build and install, but do nothing else (does not reboot)
443# boot - build, install, and boot the kernel
444# test - build, boot and if TEST is set, run the test script
Steven Rostedta57419b2010-11-02 15:13:54 -0400445# (If TEST is not set, it defaults back to boot)
Steven Rostedta75fece2010-11-02 14:58:27 -0400446# bisect - Perform a bisect on the kernel (see BISECT_TYPE below)
447# patchcheck - Do a test on a series of commits in git (see PATCHCHECK below)
448#TEST_TYPE = test
449
Steven Rostedta57419b2010-11-02 15:13:54 -0400450# Test to run if there is a successful boot and TEST_TYPE is test.
451# Must exit with 0 on success and non zero on error
452# default (undefined)
453#TEST = ssh user@machine /root/run_test
454
455# The build type is any make config type or special command
Steven Rostedta75fece2010-11-02 14:58:27 -0400456# (default randconfig)
457# nobuild - skip the clean and build step
Steven Rostedta57419b2010-11-02 15:13:54 -0400458# useconfig:/path/to/config - use the given config and run
459# oldconfig on it.
460# This option is ignored if TEST_TYPE is patchcheck or bisect
Steven Rostedta75fece2010-11-02 14:58:27 -0400461#BUILD_TYPE = randconfig
462
463# The make command (default make)
464# If you are building a 32bit x86 on a 64 bit host
465#MAKE_CMD = CC=i386-gcc AS=i386-as make ARCH=i386
466
Steven Rostedtdc895682010-11-02 15:22:53 -0400467# Any build options for the make of the kernel (not for other makes, like configs)
468# (default "")
469#BUILD_OPTIONS = -j20
470
Steven Rostedte5c2ec12012-07-19 15:22:05 -0400471# If you need to do some special handling before installing
472# you can add a script with this option.
473# The environment variable KERNEL_VERSION will be set to the
474# kernel version that is used.
475#
476# default (undefined)
477#PRE_INSTALL = ssh user@target rm -rf '/lib/modules/*-test*'
478
Steven Rostedt8b37ca82010-11-02 14:58:33 -0400479# If you need an initrd, you can add a script or code here to install
480# it. The environment variable KERNEL_VERSION will be set to the
Steven Rostedta57419b2010-11-02 15:13:54 -0400481# kernel version that is used. Remember to add the initrd line
482# to your grub menu.lst file.
483#
484# Here's a couple of examples to use:
Steven Rostedt8b37ca82010-11-02 14:58:33 -0400485#POST_INSTALL = ssh user@target /sbin/mkinitrd --allow-missing -f /boot/initramfs-test.img $KERNEL_VERSION
Steven Rostedta57419b2010-11-02 15:13:54 -0400486#
487# or on some systems:
488#POST_INSTALL = ssh user@target /sbin/dracut -f /boot/initramfs-test.img $KERNEL_VERSION
Steven Rostedt8b37ca82010-11-02 14:58:33 -0400489
Steven Rostedte0a87422011-09-30 17:50:48 -0400490# If for some reason you just want to boot the kernel and you do not
491# want the test to install anything new. For example, you may just want
492# to boot test the same kernel over and over and do not want to go through
493# the hassle of installing anything, you can set this option to 1
494# (default 0)
495#NO_INSTALL = 1
496
Steven Rostedt921ed4c2012-07-19 15:18:27 -0400497# If there is a command that you want to run before the individual test
498# case executes, then you can set this option
499#
500# default (undefined)
501#PRE_TEST = ${SSH} reboot_to_special_kernel
502
503# If there is a command you want to run after the individual test case
504# completes, then you can set this option.
505#
506# default (undefined)
507#POST_TEST = cd ${BUILD_DIR}; git reset --hard
508
Steven Rostedt0bd6c1a2011-06-14 20:39:31 -0400509# If there is a script that you require to run before the build is done
510# you can specify it with PRE_BUILD.
511#
512# One example may be if you must add a temporary patch to the build to
513# fix a unrelated bug to perform a patchcheck test. This will apply the
514# patch before each build that is made. Use the POST_BUILD to do a git reset --hard
515# to remove the patch.
516#
517# (default undef)
518#PRE_BUILD = cd ${BUILD_DIR} && patch -p1 < /tmp/temp.patch
519
520# To specify if the test should fail if the PRE_BUILD fails,
521# PRE_BUILD_DIE needs to be set to 1. Otherwise the PRE_BUILD
522# result is ignored.
523# (default 0)
524# PRE_BUILD_DIE = 1
525
526# If there is a script that should run after the build is done
527# you can specify it with POST_BUILD.
528#
529# As the example in PRE_BUILD, POST_BUILD can be used to reset modifications
530# made by the PRE_BUILD.
531#
532# (default undef)
533#POST_BUILD = cd ${BUILD_DIR} && git reset --hard
534
535# To specify if the test should fail if the POST_BUILD fails,
536# POST_BUILD_DIE needs to be set to 1. Otherwise the POST_BUILD
537# result is ignored.
538# (default 0)
539#POST_BUILD_DIE = 1
540
Steven Rostedta75fece2010-11-02 14:58:27 -0400541# Way to reboot the box to the test kernel.
Steven Rostedt77869542012-12-11 17:37:41 -0500542# Only valid options so far are "grub", "grub2", "syslinux" and "script"
Steven Rostedta75fece2010-11-02 14:58:27 -0400543# (default grub)
544# If you specify grub, it will assume grub version 1
545# and will search in /boot/grub/menu.lst for the title $GRUB_MENU
546# and select that target to reboot to the kernel. If this is not
547# your setup, then specify "script" and have a command or script
548# specified in REBOOT_SCRIPT to boot to the target.
Steven Rostedta57419b2010-11-02 15:13:54 -0400549#
Steven Rostedta15ba912012-11-13 14:30:37 -0500550# For REBOOT_TYPE = grub2, you must define both GRUB_MENU and
551# GRUB_FILE.
552#
Steven Rostedt77869542012-12-11 17:37:41 -0500553# For REBOOT_TYPE = syslinux, you must define SYSLINUX_LABEL, and
554# perhaps modify SYSLINUX (default extlinux) and SYSLINUX_PATH
555# (default /boot/extlinux)
556#
Steven Rostedta57419b2010-11-02 15:13:54 -0400557# The entry in /boot/grub/menu.lst must be entered in manually.
558# The test will not modify that file.
Steven Rostedta75fece2010-11-02 14:58:27 -0400559#REBOOT_TYPE = grub
560
Steven Rostedtbc7c5802011-12-22 16:29:10 -0500561# If you are using a machine that doesn't boot with grub, and
562# perhaps gets its kernel from a remote server (tftp), then
563# you can use this option to update the target image with the
564# test image.
565#
566# You could also do the same with POST_INSTALL, but the difference
567# between that option and this option is that POST_INSTALL runs
568# after the install, where this one runs just before a reboot.
569# (default undefined)
570#SWITCH_TO_TEST = cp ${OUTPUT_DIR}/${BUILD_TARGET} ${TARGET_IMAGE}
571
572# If you are using a machine that doesn't boot with grub, and
573# perhaps gets its kernel from a remote server (tftp), then
574# you can use this option to update the target image with the
575# the known good image to reboot safely back into.
576#
577# This option holds a command that will execute before needing
578# to reboot to a good known image.
579# (default undefined)
580#SWITCH_TO_GOOD = ssh ${SSH_USER}/${MACHINE} cp good_image ${TARGET_IMAGE}
581
Steven Rostedta75fece2010-11-02 14:58:27 -0400582# The min config that is needed to build for the machine
Steven Rostedta57419b2010-11-02 15:13:54 -0400583# A nice way to create this is with the following:
Steven Rostedta75fece2010-11-02 14:58:27 -0400584#
Steven Rostedta57419b2010-11-02 15:13:54 -0400585# $ ssh target
586# $ lsmod > mymods
587# $ scp mymods host:/tmp
588# $ exit
589# $ cd linux.git
590# $ rm .config
591# $ make LSMOD=mymods localyesconfig
592# $ grep '^CONFIG' .config > /home/test/config-min
593#
594# If you want even less configs:
595#
596# log in directly to target (do not ssh)
597#
598# $ su
599# # lsmod | cut -d' ' -f1 | xargs rmmod
600#
601# repeat the above several times
602#
603# # lsmod > mymods
604# # reboot
605#
606# May need to reboot to get your network back to copy the mymods
607# to the host, and then remove the previous .config and run the
608# localyesconfig again. The CONFIG_MIN generated like this will
609# not guarantee network activity to the box so the TEST_TYPE of
610# test may fail.
611#
612# You might also want to set:
Steven Rostedta75fece2010-11-02 14:58:27 -0400613# CONFIG_CMDLINE="<your options here>"
614# randconfig may set the above and override your real command
615# line options.
Steven Rostedta57419b2010-11-02 15:13:54 -0400616# (default undefined)
Steven Rostedta75fece2010-11-02 14:58:27 -0400617#MIN_CONFIG = /home/test/config-min
618
619# Sometimes there's options that just break the boot and
620# you do not care about. Here are a few:
621# # CONFIG_STAGING is not set
622# Staging drivers are horrible, and can break the build.
623# # CONFIG_SCSI_DEBUG is not set
624# SCSI_DEBUG may change your root partition
625# # CONFIG_KGDB_SERIAL_CONSOLE is not set
626# KGDB may cause oops waiting for a connection that's not there.
627# This option points to the file containing config options that will be prepended
628# to the MIN_CONFIG (or be the MIN_CONFIG if it is not set)
Steven Rostedta57419b2010-11-02 15:13:54 -0400629#
630# Note, config options in MIN_CONFIG will override these options.
631#
632# (default undefined)
Steven Rostedta75fece2010-11-02 14:58:27 -0400633#ADD_CONFIG = /home/test/config-broken
634
Steven Rostedtdc895682010-11-02 15:22:53 -0400635# The location on the host where to write temp files
Steven Rostedt48920632011-06-14 20:42:19 -0400636# (default /tmp/ktest/${MACHINE})
637#TMP_DIR = /tmp/ktest/${MACHINE}
Steven Rostedtdc895682010-11-02 15:22:53 -0400638
639# Optional log file to write the status (recommended)
640# Note, this is a DEFAULT section only option.
641# (default undefined)
642#LOG_FILE = /home/test/logfiles/target.log
643
644# Remove old logfile if it exists before starting all tests.
645# Note, this is a DEFAULT section only option.
646# (default 0)
647#CLEAR_LOG = 0
648
649# Line to define a successful boot up in console output.
650# This is what the line contains, not the entire line. If you need
651# the entire line to match, then use regural expression syntax like:
652# (do not add any quotes around it)
653#
654# SUCCESS_LINE = ^MyBox Login:$
655#
656# (default "login:")
657#SUCCESS_LINE = login:
658
Steven Rostedt2b803362011-09-30 18:00:23 -0400659# To speed up between reboots, defining a line that the
660# default kernel produces that represents that the default
661# kernel has successfully booted and can be used to pass
662# a new test kernel to it. Otherwise ktest.pl will wait till
663# SLEEP_TIME to continue.
664# (default undefined)
665#REBOOT_SUCCESS_LINE = login:
666
Steven Rostedt1c8a6172010-11-09 12:55:40 -0500667# In case the console constantly fills the screen, having
668# a specified time to stop the test after success is recommended.
669# (in seconds)
670# (default 10)
671#STOP_AFTER_SUCCESS = 10
672
673# In case the console constantly fills the screen, having
674# a specified time to stop the test after failure is recommended.
675# (in seconds)
676# (default 60)
677#STOP_AFTER_FAILURE = 60
678
Steven Rostedt2d01b262011-03-08 09:47:54 -0500679# In case the console constantly fills the screen, having
680# a specified time to stop the test if it never succeeds nor fails
681# is recommended.
682# Note: this is ignored if a success or failure is detected.
683# (in seconds)
684# (default 600, -1 is to never stop)
685#STOP_TEST_AFTER = 600
686
Steven Rostedtdc895682010-11-02 15:22:53 -0400687# Stop testing if a build fails. If set, the script will end if
688# a failure is detected, otherwise it will save off the .config,
689# dmesg and bootlog in a directory called
690# MACHINE-TEST_TYPE_BUILD_TYPE-fail-yyyymmddhhmmss
691# if the STORE_FAILURES directory is set.
692# (default 1)
693# Note, even if this is set to zero, there are some errors that still
694# stop the tests.
695#DIE_ON_FAILURE = 1
696
697# Directory to store failure directories on failure. If this is not
698# set, DIE_ON_FAILURE=0 will not save off the .config, dmesg and
699# bootlog. This option is ignored if DIE_ON_FAILURE is not set.
700# (default undefined)
701#STORE_FAILURES = /home/test/failures
702
Rabin Vincentde5b6e32011-11-18 17:05:31 +0530703# Directory to store success directories on success. If this is not
704# set, the .config, dmesg and bootlog will not be saved if a
705# test succeeds.
706# (default undefined)
707#STORE_SUCCESSES = /home/test/successes
708
Steven Rostedtdc895682010-11-02 15:22:53 -0400709# Build without doing a make mrproper, or removing .config
710# (default 0)
711#BUILD_NOCLEAN = 0
712
713# As the test reads the console, after it hits the SUCCESS_LINE
714# the time it waits for the monitor to settle down between reads
715# can usually be lowered.
716# (in seconds) (default 1)
717#BOOTED_TIMEOUT = 1
718
719# The timeout in seconds when we consider the box hung after
720# the console stop producing output. Be sure to leave enough
721# time here to get pass a reboot. Some machines may not produce
722# any console output for a long time during a reboot. You do
723# not want the test to fail just because the system was in
724# the process of rebooting to the test kernel.
725# (default 120)
726#TIMEOUT = 120
727
728# In between tests, a reboot of the box may occur, and this
729# is the time to wait for the console after it stops producing
730# output. Some machines may not produce a large lag on reboot
731# so this should accommodate it.
732# The difference between this and TIMEOUT, is that TIMEOUT happens
733# when rebooting to the test kernel. This sleep time happens
734# after a test has completed and we are about to start running
735# another test. If a reboot to the reliable kernel happens,
736# we wait SLEEP_TIME for the console to stop producing output
737# before starting the next test.
Steven Rostedt2b803362011-09-30 18:00:23 -0400738#
739# You can speed up reboot times even more by setting REBOOT_SUCCESS_LINE.
Steven Rostedtdc895682010-11-02 15:22:53 -0400740# (default 60)
741#SLEEP_TIME = 60
742
743# The time in between bisects to sleep (in seconds)
744# (default 60)
745#BISECT_SLEEP_TIME = 60
746
Steven Rostedt407b95b2012-07-19 16:05:42 -0400747# The max wait time (in seconds) for waiting for the console to finish.
748# If for some reason, the console is outputting content without
749# ever finishing, this will cause ktest to get stuck. This
750# option is the max time ktest will wait for the monitor (console)
751# to settle down before continuing.
752# (default 1800)
753#MAX_MONITOR_WAIT
754
Steven Rostedt27d934b2011-05-20 09:18:18 -0400755# The time in between patch checks to sleep (in seconds)
756# (default 60)
757#PATCHCHECK_SLEEP_TIME = 60
758
Steven Rostedtdc895682010-11-02 15:22:53 -0400759# Reboot the target box on error (default 0)
760#REBOOT_ON_ERROR = 0
761
762# Power off the target on error (ignored if REBOOT_ON_ERROR is set)
763# Note, this is a DEFAULT section only option.
764# (default 0)
765#POWEROFF_ON_ERROR = 0
766
767# Power off the target after all tests have completed successfully
768# Note, this is a DEFAULT section only option.
769# (default 0)
770#POWEROFF_ON_SUCCESS = 0
771
772# Reboot the target after all test completed successfully (default 1)
773# (ignored if POWEROFF_ON_SUCCESS is set)
774#REBOOT_ON_SUCCESS = 1
775
776# In case there are isses with rebooting, you can specify this
777# to always powercycle after this amount of time after calling
778# reboot.
779# Note, POWERCYCLE_AFTER_REBOOT = 0 does NOT disable it. It just
780# makes it powercycle immediately after rebooting. Do not define
781# it if you do not want it.
782# (default undefined)
783#POWERCYCLE_AFTER_REBOOT = 5
784
785# In case there's isses with halting, you can specify this
786# to always poweroff after this amount of time after calling
787# halt.
788# Note, POWEROFF_AFTER_HALT = 0 does NOT disable it. It just
789# makes it poweroff immediately after halting. Do not define
790# it if you do not want it.
791# (default undefined)
792#POWEROFF_AFTER_HALT = 20
793
794# A script or command to power off the box (default undefined)
795# Needed for POWEROFF_ON_ERROR and SUCCESS
796#
797# Example for digital loggers power switch:
798#POWER_OFF = wget --no-proxy -O /dev/null -q --auth-no-challenge 'http://admin:admin@power/outlet?5=OFF'
799#
800# Example for a virtual guest call "Guest".
801#POWER_OFF = virsh destroy Guest
802
Steven Rostedt (Red Hat)4283b162013-01-30 18:37:47 -0500803# To have the build fail on "new" warnings, create a file that
804# contains a list of all known warnings (they must match exactly
805# to the line with 'warning:', 'error:' or 'Error:'. If the option
806# WARNINGS_FILE is set, then that file will be read, and if the
807# build detects a warning, it will examine this file and if the
808# warning does not exist in it, it will fail the build.
809#
810# Note, if this option is defined to a file that does not exist
811# then any warning will fail the build.
812# (see make_warnings_file below)
813#
814# (optional, default undefined)
815#WARNINGS_FILE = ${OUTPUT_DIR}/warnings_file
816
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -0500817# The way to execute a command on the target
818# (default ssh $SSH_USER@$MACHINE $SSH_COMMAND";)
819# The variables SSH_USER, MACHINE and SSH_COMMAND are defined
820#SSH_EXEC = ssh $SSH_USER@$MACHINE $SSH_COMMAND";
821
Steven Rostedt02ad2612012-03-21 08:21:24 -0400822# The way to copy a file to the target (install and modules)
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -0500823# (default scp $SRC_FILE $SSH_USER@$MACHINE:$DST_FILE)
Steven Rostedt02ad2612012-03-21 08:21:24 -0400824# The variables SSH_USER, MACHINE are defined by the config
825# SRC_FILE and DST_FILE are ktest internal variables and
826# should only have '$' and not the '${}' notation.
827# (default scp $SRC_FILE ${SSH_USER}@${MACHINE}:$DST_FILE)
828#SCP_TO_TARGET = echo skip scp for $SRC_FILE $DST_FILE
829
830# If install needs to be different than modules, then this
831# option will override the SCP_TO_TARGET for installation.
832# (default ${SCP_TO_TARGET} )
833#SCP_TO_TARGET_INSTALL = scp $SRC_FILE tftp@tftpserver:$DST_FILE
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -0500834
835# The nice way to reboot the target
836# (default ssh $SSH_USER@$MACHINE reboot)
837# The variables SSH_USER and MACHINE are defined.
838#REBOOT = ssh $SSH_USER@$MACHINE reboot
839
Steven Rostedtf1a5b962011-06-13 10:30:00 -0400840# The way triple faults are detected is by testing the kernel
841# banner. If the kernel banner for the kernel we are testing is
842# found, and then later a kernel banner for another kernel version
843# is found, it is considered that we encountered a triple fault,
844# and there is no panic or callback, but simply a reboot.
845# To disable this (because it did a false positive) set the following
846# to 0.
847# (default 1)
848#DETECT_TRIPLE_FAULT = 0
849
Steven Rostedt9cc9e092011-12-22 21:37:22 -0500850# All options in the config file should be either used by ktest
851# or could be used within a value of another option. If an option
852# in the config file is not used, ktest will warn about it and ask
853# if you want to continue.
854#
855# If you don't care if there are non-used options, enable this
856# option. Be careful though, a non-used option is usually a sign
857# of an option name being typed incorrectly.
858# (default 0)
859#IGNORE_UNUSED = 1
860
Steven Rostedtbe405f92012-01-04 21:51:59 -0500861# When testing a kernel that happens to have WARNINGs, and call
862# traces, ktest.pl will detect these and fail a boot or test run
863# due to warnings. By setting this option, ktest will ignore
864# call traces, and will not fail a test if the kernel produces
865# an oops. Use this option with care.
866# (default 0)
867#IGNORE_ERRORS = 1
868
Steven Rostedta75fece2010-11-02 14:58:27 -0400869#### Per test run options ####
Steven Rostedta57419b2010-11-02 15:13:54 -0400870# The following options are only allowed in TEST_START sections.
871# They are ignored in the DEFAULTS sections.
Steven Rostedta75fece2010-11-02 14:58:27 -0400872#
Steven Rostedta57419b2010-11-02 15:13:54 -0400873# All of these are optional and undefined by default, although
874# some of these options are required for TEST_TYPE of patchcheck
875# and bisect.
Steven Rostedta75fece2010-11-02 14:58:27 -0400876#
Steven Rostedta57419b2010-11-02 15:13:54 -0400877#
878# CHECKOUT = branch
Steven Rostedta75fece2010-11-02 14:58:27 -0400879#
880# If the BUILD_DIR is a git repository, then you can set this option
881# to checkout the given branch before running the TEST. If you
882# specify this for the first run, that branch will be used for
Steven Rostedta57419b2010-11-02 15:13:54 -0400883# all preceding tests until a new CHECKOUT is set.
Steven Rostedta75fece2010-11-02 14:58:27 -0400884#
Steven Rostedta57419b2010-11-02 15:13:54 -0400885#
Steven Rostedt9064af52011-06-13 10:38:48 -0400886# TEST_NAME = name
887#
888# If you want the test to have a name that is displayed in
889# the test result banner at the end of the test, then use this
890# option. This is useful to search for the RESULT keyword and
891# not have to translate a test number to a test in the config.
Steven Rostedta57419b2010-11-02 15:13:54 -0400892#
893# For TEST_TYPE = patchcheck
Steven Rostedta75fece2010-11-02 14:58:27 -0400894#
895# This expects the BUILD_DIR to be a git repository, and
Steven Rostedta57419b2010-11-02 15:13:54 -0400896# will checkout the PATCHCHECK_START commit.
Steven Rostedta75fece2010-11-02 14:58:27 -0400897#
Steven Rostedta57419b2010-11-02 15:13:54 -0400898# The option BUILD_TYPE will be ignored.
Steven Rostedta75fece2010-11-02 14:58:27 -0400899#
Steven Rostedta57419b2010-11-02 15:13:54 -0400900# The MIN_CONFIG will be used for all builds of the patchcheck. The build type
901# used for patchcheck is oldconfig.
Steven Rostedta75fece2010-11-02 14:58:27 -0400902#
Steven Rostedta57419b2010-11-02 15:13:54 -0400903# PATCHCHECK_START is required and is the first patch to
904# test (the SHA1 of the commit). You may also specify anything
905# that git checkout allows (branch name, tage, HEAD~3).
906#
907# PATCHCHECK_END is the last patch to check (default HEAD)
908#
Steven Rostedt (Red Hat)23a0e162014-09-19 20:10:39 -0400909# PATCHCHECK_CHERRY if set to non zero, then git cherry will be
910# performed against PATCHCHECK_START and PATCHCHECK_END. That is
911#
912# git cherry ${PATCHCHECK_START} ${PATCHCHECK_END}
913#
914# Then the changes found will be tested.
915#
916# Note, PATCHCHECK_CHERRY requires PATCHCHECK_END to be defined.
917# (default 0)
918#
Steven Rostedta57419b2010-11-02 15:13:54 -0400919# PATCHCHECK_TYPE is required and is the type of test to run:
Steven Rostedta75fece2010-11-02 14:58:27 -0400920# build, boot, test.
921#
922# Note, the build test will look for warnings, if a warning occurred
Steven Rostedt19902072011-06-14 20:46:25 -0400923# in a file that a commit touches, the build will fail, unless
924# IGNORE_WARNINGS is set for the given commit's sha1
925#
926# IGNORE_WARNINGS can be used to disable the failure of patchcheck
927# on a particuler commit (SHA1). You can add more than one commit
928# by adding a list of SHA1s that are space delimited.
Steven Rostedta75fece2010-11-02 14:58:27 -0400929#
930# If BUILD_NOCLEAN is set, then make mrproper will not be run on
931# any of the builds, just like all other TEST_TYPE tests. But
932# what makes patchcheck different from the other tests, is if
933# BUILD_NOCLEAN is not set, only the first and last patch run
934# make mrproper. This helps speed up the test.
935#
936# Example:
Steven Rostedta57419b2010-11-02 15:13:54 -0400937# TEST_START
938# TEST_TYPE = patchcheck
939# CHECKOUT = mybranch
940# PATCHCHECK_TYPE = boot
941# PATCHCHECK_START = 747e94ae3d1b4c9bf5380e569f614eb9040b79e7
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -0500942# PATCHCHECK_END = HEAD~2
Steven Rostedt19902072011-06-14 20:46:25 -0400943# IGNORE_WARNINGS = 42f9c6b69b54946ffc0515f57d01dc7f5c0e4712 0c17ca2c7187f431d8ffc79e81addc730f33d128
Steven Rostedta75fece2010-11-02 14:58:27 -0400944#
945#
Steven Rostedta75fece2010-11-02 14:58:27 -0400946#
Steven Rostedta57419b2010-11-02 15:13:54 -0400947# For TEST_TYPE = bisect
Steven Rostedta75fece2010-11-02 14:58:27 -0400948#
Steven Rostedta57419b2010-11-02 15:13:54 -0400949# You can specify a git bisect if the BUILD_DIR is a git repository.
950# The MIN_CONFIG will be used for all builds of the bisect. The build type
951# used for bisecting is oldconfig.
952#
953# The option BUILD_TYPE will be ignored.
954#
955# BISECT_TYPE is the type of test to perform:
Steven Rostedta75fece2010-11-02 14:58:27 -0400956# build - bad fails to build
957# boot - bad builds but fails to boot
958# test - bad boots but fails a test
959#
Steven Rostedta57419b2010-11-02 15:13:54 -0400960# BISECT_GOOD is the commit (SHA1) to label as good (accepts all git good commit types)
961# BISECT_BAD is the commit to label as bad (accepts all git bad commit types)
Steven Rostedta75fece2010-11-02 14:58:27 -0400962#
963# The above three options are required for a bisect operation.
964#
Steven Rostedta57419b2010-11-02 15:13:54 -0400965# BISECT_REPLAY = /path/to/replay/file (optional, default undefined)
Steven Rostedta75fece2010-11-02 14:58:27 -0400966#
967# If an operation failed in the bisect that was not expected to
968# fail. Then the test ends. The state of the BUILD_DIR will be
Steven Rostedta57419b2010-11-02 15:13:54 -0400969# left off at where the failure occurred. You can examine the
Steven Rostedta75fece2010-11-02 14:58:27 -0400970# reason for the failure, and perhaps even find a git commit
971# that would work to continue with. You can run:
972#
973# git bisect log > /path/to/replay/file
974#
Steven Rostedta57419b2010-11-02 15:13:54 -0400975# The adding:
Steven Rostedta75fece2010-11-02 14:58:27 -0400976#
Steven Rostedta57419b2010-11-02 15:13:54 -0400977# BISECT_REPLAY= /path/to/replay/file
Steven Rostedta75fece2010-11-02 14:58:27 -0400978#
Steven Rostedta57419b2010-11-02 15:13:54 -0400979# And running the test again. The test will perform the initial
980# git bisect start, git bisect good, and git bisect bad, and
981# then it will run git bisect replay on this file, before
982# continuing with the bisect.
983#
984# BISECT_START = commit (optional, default undefined)
985#
986# As with BISECT_REPLAY, if the test failed on a commit that
Steven Rostedta75fece2010-11-02 14:58:27 -0400987# just happen to have a bad commit in the middle of the bisect,
Steven Rostedta57419b2010-11-02 15:13:54 -0400988# and you need to skip it. If BISECT_START is defined, it
989# will checkout that commit after doing the initial git bisect start,
990# git bisect good, git bisect bad, and running the git bisect replay
991# if the BISECT_REPLAY is set.
Steven Rostedta75fece2010-11-02 14:58:27 -0400992#
Steven Rostedtc23dca72011-03-08 09:26:31 -0500993# BISECT_SKIP = 1 (optional, default 0)
994#
995# If BISECT_TYPE is set to test but the build fails, ktest will
996# simply fail the test and end their. You could use BISECT_REPLAY
997# and BISECT_START to resume after you found a new starting point,
998# or you could set BISECT_SKIP to 1. If BISECT_SKIP is set to 1,
999# when something other than the BISECT_TYPE fails, ktest.pl will
1000# run "git bisect skip" and try again.
1001#
Steven Rostedt3410f6f2011-03-08 09:38:12 -05001002# BISECT_FILES = <path> (optional, default undefined)
1003#
1004# To just run the git bisect on a specific path, set BISECT_FILES.
1005# For example:
1006#
1007# BISECT_FILES = arch/x86 kernel/time
1008#
1009# Will run the bisect with "git bisect start -- arch/x86 kernel/time"
1010#
Steven Rostedta57419b2010-11-02 15:13:54 -04001011# BISECT_REVERSE = 1 (optional, default 0)
Steven Rostedta75fece2010-11-02 14:58:27 -04001012#
1013# In those strange instances where it was broken forever
1014# and you are trying to find where it started to work!
Steven Rostedta57419b2010-11-02 15:13:54 -04001015# Set BISECT_GOOD to the commit that was last known to fail
1016# Set BISECT_BAD to the commit that is known to start working.
1017# With BISECT_REVERSE = 1, The test will consider failures as
1018# good, and success as bad.
Steven Rostedta75fece2010-11-02 14:58:27 -04001019#
Steven Rostedtc960bb92011-03-08 09:22:39 -05001020# BISECT_MANUAL = 1 (optional, default 0)
1021#
1022# In case there's a problem with automating the bisect for
1023# whatever reason. (Can't reboot, want to inspect each iteration)
1024# Doing a BISECT_MANUAL will have the test wait for you to
1025# tell it if the test passed or failed after each iteration.
1026# This is basicall the same as running git bisect yourself
1027# but ktest will rebuild and install the kernel for you.
1028#
Steven Rostedta57419b2010-11-02 15:13:54 -04001029# BISECT_CHECK = 1 (optional, default 0)
Steven Rostedta75fece2010-11-02 14:58:27 -04001030#
1031# Just to be sure the good is good and bad is bad, setting
Steven Rostedta57419b2010-11-02 15:13:54 -04001032# BISECT_CHECK to 1 will start the bisect by first checking
1033# out BISECT_BAD and makes sure it fails, then it will check
1034# out BISECT_GOOD and makes sure it succeeds before starting
1035# the bisect (it works for BISECT_REVERSE too).
Steven Rostedta75fece2010-11-02 14:58:27 -04001036#
Steven Rostedta57419b2010-11-02 15:13:54 -04001037# You can limit the test to just check BISECT_GOOD or
1038# BISECT_BAD with BISECT_CHECK = good or
1039# BISECT_CHECK = bad, respectively.
Steven Rostedta75fece2010-11-02 14:58:27 -04001040#
Steven Rostedt (Red Hat)961d9ca2014-01-18 19:52:13 -05001041# BISECT_TRIES = 5 (optional, default 1)
1042#
1043# For those cases that it takes several tries to hit a bug,
1044# the BISECT_TRIES is useful. It is the number of times the
1045# test is ran before it says the kernel is good. The first failure
1046# will stop trying and mark the current SHA1 as bad.
1047#
1048# Note, as with all race bugs, there's no guarantee that if
1049# it succeeds, it is really a good bisect. But it helps in case
1050# the bug is some what reliable.
1051#
1052# You can set BISECT_TRIES to zero, and all tests will be considered
1053# good, unless you also set BISECT_MANUAL.
1054#
Steven Rostedtc5dacb82011-12-22 12:43:57 -05001055# BISECT_RET_GOOD = 0 (optional, default undefined)
1056#
1057# In case the specificed test returns something other than just
1058# 0 for good, and non-zero for bad, you can override 0 being
1059# good by defining BISECT_RET_GOOD.
1060#
1061# BISECT_RET_BAD = 1 (optional, default undefined)
1062#
1063# In case the specificed test returns something other than just
1064# 0 for good, and non-zero for bad, you can override non-zero being
1065# bad by defining BISECT_RET_BAD.
1066#
1067# BISECT_RET_ABORT = 255 (optional, default undefined)
1068#
1069# If you need to abort the bisect if the test discovers something
1070# that was wrong, you can define BISECT_RET_ABORT to be the error
1071# code returned by the test in order to abort the bisect.
1072#
1073# BISECT_RET_SKIP = 2 (optional, default undefined)
1074#
1075# If the test detects that the current commit is neither good
1076# nor bad, but something else happened (another bug detected)
1077# you can specify BISECT_RET_SKIP to an error code that the
1078# test returns when it should skip the current commit.
1079#
1080# BISECT_RET_DEFAULT = good (optional, default undefined)
1081#
1082# You can override the default of what to do when the above
1083# options are not hit. This may be one of, "good", "bad",
1084# "abort" or "skip" (without the quotes).
1085#
1086# Note, if you do not define any of the previous BISECT_RET_*
1087# and define BISECT_RET_DEFAULT, all bisects results will do
1088# what the BISECT_RET_DEFAULT has.
1089#
1090#
Steven Rostedta75fece2010-11-02 14:58:27 -04001091# Example:
Steven Rostedta57419b2010-11-02 15:13:54 -04001092# TEST_START
1093# TEST_TYPE = bisect
1094# BISECT_GOOD = v2.6.36
1095# BISECT_BAD = b5153163ed580e00c67bdfecb02b2e3843817b3e
1096# BISECT_TYPE = build
1097# MIN_CONFIG = /home/test/config-bisect
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001098#
1099#
1100#
1101# For TEST_TYPE = config_bisect
1102#
1103# In those cases that you have two different configs. One of them
1104# work, the other does not, and you do not know what config causes
1105# the problem.
1106# The TEST_TYPE config_bisect will bisect the bad config looking for
1107# what config causes the failure.
1108#
1109# The way it works is this:
1110#
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001111# You can specify a good config with CONFIG_BISECT_GOOD, otherwise it
1112# will use the MIN_CONFIG, and if that's not specified, it will use
1113# the config that comes with "make defconfig".
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001114#
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001115# It runs both the good and bad configs through a make oldconfig to
1116# make sure that they are set up for the kernel that is checked out.
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001117#
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001118# It then reads the configs that are set, as well as the ones that are
1119# not set for both the good and bad configs, and then compares them.
1120# It will set half of the good configs within the bad config (note,
1121# "set" means to make the bad config match the good config, a config
1122# in the good config that is off, will be turned off in the bad
1123# config. That is considered a "set").
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001124#
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001125# It tests this new config and if it works, it becomes the new good
1126# config, otherwise it becomes the new bad config. It continues this
1127# process until there's only one config left and it will report that
1128# config.
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001129#
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001130# The "bad config" can also be a config that is needed to boot but was
1131# disabled because it depended on something that wasn't set.
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001132#
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001133# During this process, it saves the current good and bad configs in
1134# ${TMP_DIR}/good_config and ${TMP_DIR}/bad_config respectively.
1135# If you stop the test, you can copy them to a new location to
1136# reuse them again.
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001137#
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001138# Although the MIN_CONFIG may be the config it starts with, the
1139# MIN_CONFIG is ignored.
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001140#
1141# The option BUILD_TYPE will be ignored.
1142#
1143# CONFIG_BISECT_TYPE is the type of test to perform:
1144# build - bad fails to build
1145# boot - bad builds but fails to boot
1146# test - bad boots but fails a test
1147#
Steven Rostedt30f75da2011-06-13 10:35:35 -04001148# CONFIG_BISECT is the config that failed to boot
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001149#
Steven Rostedt30f75da2011-06-13 10:35:35 -04001150# If BISECT_MANUAL is set, it will pause between iterations.
1151# This is useful to use just ktest.pl just for the config bisect.
1152# If you set it to build, it will run the bisect and you can
1153# control what happens in between iterations. It will ask you if
1154# the test succeeded or not and continue the config bisect.
1155#
1156# CONFIG_BISECT_GOOD (optional)
1157# If you have a good config to start with, then you
1158# can specify it with CONFIG_BISECT_GOOD. Otherwise
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001159# the MIN_CONFIG is the base, if MIN_CONFIG is not set
1160# It will build a config with "make defconfig"
Steven Rostedtc960bb92011-03-08 09:22:39 -05001161#
Steven Rostedtb0918612012-07-19 15:26:00 -04001162# CONFIG_BISECT_CHECK (optional)
1163# Set this to 1 if you want to confirm that the config ktest
1164# generates (the bad config with the min config) is still bad.
1165# It may be that the min config fixes what broke the bad config
1166# and the test will not return a result.
Steven Rostedt (Red Hat)4c16b1d2014-04-23 22:30:00 -04001167# Set it to "good" to test only the good config and set it
1168# to "bad" to only test the bad config.
Steven Rostedtb0918612012-07-19 15:26:00 -04001169#
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001170# Example:
1171# TEST_START
1172# TEST_TYPE = config_bisect
1173# CONFIG_BISECT_TYPE = build
Al Virod36b6912011-12-29 17:09:01 -05001174# CONFIG_BISECT = /home/test/config-bad
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001175# MIN_CONFIG = /home/test/config-min
Steven Rostedtc960bb92011-03-08 09:22:39 -05001176# BISECT_MANUAL = 1
Steven Rostedtd1fbd7e2010-11-08 17:41:37 -05001177#
Steven Rostedt4c4ab122011-07-15 21:16:17 -04001178#
1179#
1180# For TEST_TYPE = make_min_config
1181#
1182# After doing a make localyesconfig, your kernel configuration may
1183# not be the most useful minimum configuration. Having a true minimum
1184# config that you can use against other configs is very useful if
1185# someone else has a config that breaks on your code. By only forcing
1186# those configurations that are truly required to boot your machine
1187# will give you less of a chance that one of your set configurations
1188# will make the bug go away. This will give you a better chance to
1189# be able to reproduce the reported bug matching the broken config.
1190#
1191# Note, this does take some time, and may require you to run the
1192# test over night, or perhaps over the weekend. But it also allows
1193# you to interrupt it, and gives you the current minimum config
1194# that was found till that time.
1195#
1196# Note, this test automatically assumes a BUILD_TYPE of oldconfig
1197# and its test type acts like boot.
1198# TODO: add a test version that makes the config do more than just
1199# boot, like having network access.
1200#
Steven Rostedtb9066f62011-07-15 21:25:24 -04001201# To save time, the test does not just grab any option and test
1202# it. The Kconfig files are examined to determine the dependencies
1203# of the configs. If a config is chosen that depends on another
1204# config, that config will be checked first. By checking the
1205# parents first, we can eliminate whole groups of configs that
1206# may have been enabled.
1207#
1208# For example, if a USB device config is chosen and depends on CONFIG_USB,
1209# the CONFIG_USB will be tested before the device. If CONFIG_USB is
1210# found not to be needed, it, as well as all configs that depend on
1211# it, will be disabled and removed from the current min_config.
1212#
Steven Rostedt4c4ab122011-07-15 21:16:17 -04001213# OUTPUT_MIN_CONFIG is the path and filename of the file that will
1214# be created from the MIN_CONFIG. If you interrupt the test, set
1215# this file as your new min config, and use it to continue the test.
1216# This file does not need to exist on start of test.
1217# This file is not created until a config is found that can be removed.
Steven Rostedt35ce5952011-07-15 21:57:25 -04001218# If this file exists, you will be prompted if you want to use it
1219# as the min_config (overriding MIN_CONFIG) if START_MIN_CONFIG
1220# is not defined.
Steven Rostedt4c4ab122011-07-15 21:16:17 -04001221# (required field)
1222#
1223# START_MIN_CONFIG is the config to use to start the test with.
1224# you can set this as the same OUTPUT_MIN_CONFIG, but if you do
1225# the OUTPUT_MIN_CONFIG file must exist.
1226# (default MIN_CONFIG)
1227#
1228# IGNORE_CONFIG is used to specify a config file that has configs that
1229# you already know must be set. Configs are written here that have
1230# been tested and proved to be required. It is best to define this
1231# file if you intend on interrupting the test and running it where
1232# it left off. New configs that it finds will be written to this file
1233# and will not be tested again in later runs.
1234# (optional)
1235#
Steven Rostedtccc513b2012-05-21 17:13:40 -04001236# MIN_CONFIG_TYPE can be either 'boot' or 'test'. With 'boot' it will
1237# test if the created config can just boot the machine. If this is
1238# set to 'test', then the TEST option must be defined and the created
1239# config will not only boot the target, but also make sure that the
1240# config lets the test succeed. This is useful to make sure the final
1241# config that is generated allows network activity (ssh).
1242# (optional)
1243#
Steven Rostedt43de3312012-05-21 23:35:12 -04001244# USE_OUTPUT_MIN_CONFIG set this to 1 if you do not want to be prompted
1245# about using the OUTPUT_MIN_CONFIG as the MIN_CONFIG as the starting
1246# point. Set it to 0 if you want to always just use the given MIN_CONFIG.
1247# If it is not defined, it will prompt you to pick which config
1248# to start with (MIN_CONFIG or OUTPUT_MIN_CONFIG).
1249#
Steven Rostedt4c4ab122011-07-15 21:16:17 -04001250# Example:
1251#
1252# TEST_TYPE = make_min_config
1253# OUTPUT_MIN_CONFIG = /path/to/config-new-min
1254# START_MIN_CONFIG = /path/to/config-min
1255# IGNORE_CONFIG = /path/to/config-tested
Steven Rostedtccc513b2012-05-21 17:13:40 -04001256# MIN_CONFIG_TYPE = test
1257# TEST = ssh ${USER}@${MACHINE} echo hi
Steven Rostedt4c4ab122011-07-15 21:16:17 -04001258#
Steven Rostedt (Red Hat)4283b162013-01-30 18:37:47 -05001259#
1260#
1261#
1262# For TEST_TYPE = make_warnings_file
1263#
1264# If you want the build to fail when a new warning is discovered
1265# you set the WARNINGS_FILE to point to a file of known warnings.
1266#
1267# The test "make_warnings_file" will let you create a new warnings
1268# file before you run other tests, like patchcheck.
1269#
1270# What this test does is to run just a build, you still need to
1271# specify BUILD_TYPE to tell the test what type of config to use.
1272# A BUILD_TYPE of nobuild will fail this test.
1273#
1274# The test will do the build and scan for all warnings. Any warning
1275# it discovers will be saved in the WARNINGS_FILE (required) option.
1276#
1277# It is recommended (but not necessary) to make sure BUILD_NOCLEAN is
1278# off, so that a full build is done (make mrproper is performed).
1279# That way, all warnings will be captured.
1280#
1281# Example:
1282#
1283# TEST_TYPE = make_warnings_file
1284# WARNINGS_FILE = ${OUTPUT_DIR}
1285# BUILD_TYPE = useconfig:oldconfig
1286# CHECKOUT = v3.8
1287# BUILD_NOCLEAN = 0
1288#