05 Jan, 2012
2 commits
-
When testing a kernel that has warnings, ktest.pl will fail the test
when it sees the warning. If you need to test the the kernel and want
to ignore the errors that are produced, the option IGNORE_ERRORS has
been added. When IGNORE_ERRORS is set to something other than 0, it will
ignore call traces due to WARN_ON().Signed-off-by: Steven Rostedt
-
The REBOOT_TYPE may be either grub or script, if it is script
it is expected that a REBOOT_SCRIPT is defined.With the SWITCH_TO_TEST which is the complement of SWITCH_TO_GOOD,
which does basically the same thing as REBOOT_SCRIPT and but for
both grub and script, the REBOOT_SCRIPT does not need to be mandatory
anymore.Do not require the REBOOT_SCRIPT and always run the reboot code
for both grub and script.Signed-off-by: Steven Rostedt
23 Dec, 2011
12 commits
-
It becomes quite annoying when you go to run a test and then
realize that you typed an option name wrong, and the test starts
doing the default action and not what you expected it to do.It is even more annoying when you wake up the next day after
running the test over night when you discover this.By testing if all options specified in a config file are
used by either ktest or were used in one of the option's values
we can see if there are any dangling options that were not used.
In such a case, show the user the options that were not used
and ask them if they want to continue or not.The option IGNORE_UNUSED was also added to allow the user to
override this feature.Signed-off-by: Steven Rostedt
-
Currently the patchcheck, bisect, and config_bisect variables
are only able to be set per test. You can not set a default
value for them.By letting default values be set, it makes some config files
a bit easier, and also makes it easier to find typos in the
option names.Signed-off-by: Steven Rostedt
-
Initializing each default value by specifying the hash name is
ugly. This is one of the rare cases that the "perl way" is actually
much cleaner and easier to read.Signed-off-by: Steven Rostedt
-
For machines that do no use grub, it may be needed to update an
external image (tftp) before doing a reboot into either the
test image or the known good image.The option SWITCH_TO_GOOD is added, where if it is defined, the
command that is specified as its value will be executed before
doing a reboot into a known good image.The option SWITCH_TO_TEST is added, where if it is defined, the
command that is specified as its value will be executed before
doing a reboot into the test image.Signed-off-by: Steven Rostedt
-
When running the ktest git bisect test, if the BISECT_TYPE is "test",
the bisect is determined to be good or bad based off of the error
code of the test that is run. Currently, if the test returns 0,
it is considered a pass (good), a non-zero is considered a fail (bad).But it has been requested to add more options, and also change
the meanings of the error codes of the test. For example, one may
want the test to detect if the commit is not good or bad,
(maybe the bisect came to a point where the code in question
does not exist). The test could report an error code that should tell
ktest to skip the commit.Also, a test could detect that something is horribly wrong and the
biscet should just be aborted.The new options:
BISECT_RET_GOOD
BISECT_RET_BAD
BISECT_RET_SKIP
BISECT_RET_ABORT
BISECT_RET_DEFAULThave been added. The first 4 take an integer value that will
represent if the test should be considered a pass, fail, neither
good nor bad, or abort respectively.The BISECT_RET_DEFAULT will bo whatever is not defined by the
above codes. If only BISECT_RET_DEFAULT is defined, then all tests
will do the default.Signed-off-by: Steven Rostedt
-
All options can take variables "${var}". Before doing any processing
or decision making on the content of an option, evaluate it incase
there are variables that may change the outcome.Signed-off-by: Steven Rostedt
-
The install process may also need to know what the kernel version
is, to add it to the name. Evaluate it for both install and
post install.Signed-off-by: Steven Rostedt
-
If all the tests are only for build or install, do not ask
for options not needed to do the install, if the options do
not exist.Signed-off-by: Steven Rostedt
-
When creating a new config, ask for the BUILD_OPTIONS variable
that lets users add things like -j20 to the make.Signed-off-by: Steven Rostedt
-
When creating a ktest config or if te config only has build only
tests, some of the manditory config options are not needed.Do not ask for them if all tests in the config file are just build
tests.Suggested-by: Darren Hart
Signed-off-by: Steven Rostedt -
When no argument is supplied to ktest, or the config applied does
not exist and a new config is being created, instead of just using
the default test type, give the user an option to pick the test type
of either 'build, install, or boot'. Other options may be added later
but then those would require more questions as they require more
fields. But that's for another release of ktest to add that feature.Signed-off-by: Steven Rostedt
-
If a bisect is killed for some reason, have ktest detect that a bisect
is in progress and if so, allow the user to start the bisect where
it left off.Signed-off-by: Steven Rostedt
24 Nov, 2011
1 commit
-
Typing in a full path when you know that the path exists within
the directory your are running is tedious and unnecessary.Allow the user to use ${PWD} if they want a dynamic path name
which will be the path that ktest.pl is executed from
or use ${THIS_DIR} which is a variable assigned `pwd` and
the the variable will exist within the config, allowing the user
to change it and affect all other paths using this variable as wellSigned-off-by: Steven Rostedt
22 Nov, 2011
1 commit
-
When a user runs ktest without an argument, or the argument given
is not a config file that exists, ktest will ask the user a few
questions to create a simple ktest config file.A few of the questions should have a default value set, that if anything
it will make it easier for the user to know what is suppose to
be in that value.These new values are:
SSH_USER, BUILD_TARGET and TARGET_IMAGE
Signed-off-by: Steven Rostedt
19 Nov, 2011
3 commits
-
Add a STORE_SUCCESSES option, to allow success logs to be stored, for
example to double-check or otherwise post-process the test logs.Link: http://lkml.kernel.org/r/1321616131-21352-3-git-send-email-rabin@rab.in
Signed-off-by: Rabin Vincent
Signed-off-by: Steven Rostedt -
The test output may contain useful information; save it along with the
already-saved buildlog, dmesg, and .config.Link: http://lkml.kernel.org/r/1321616131-21352-1-git-send-email-rabin@rab.in
Signed-off-by: Rabin Vincent
Signed-off-by: Steven Rostedt -
Let's say we have "OUTPUT_DIR = build/${TEST_NAME}", and we're iterating
a test. In the second iteration of a test, the TEST_NAME of the test
we're repeating is not used. Instead, ${TEST_NAME} appears literally:touch /home/rabin/kernel/test/build/${TEST_NAME}/.config ... SUCCESS
Fix this by making __eval_option() check the parent test options
for a repeated test.Link: http://lkml.kernel.org/r/1321616131-21352-2-git-send-email-rabin@rab.in
Signed-off-by: Rabin Vincent
Signed-off-by: Steven Rostedt
28 Oct, 2011
1 commit
-
When ktest.pl is called without any arguments, or if the config
file does not exist, ktest.pl will ask the user for some information.
Some of these questions are code paths. Allowing the user to type
${PWD} for the current directory greatly simplifies these entries.Add variable processing to the entered values.
Signed-off-by: Steven Rostedt
22 Oct, 2011
2 commits
-
Adding the variable ${PWD} that equals `pwd` makes the config files
much simpler.Signed-off-by: Steven Rostedt
-
On some tests that do multiple boots (patchcheck, bisect, etc), the build
of the next kernel to run may finish before the stable kernel has finished
booting. Then the install of the new kernel will fail when it tries to connect
as the machine has not finished the boot process.Do one more monitor flush to make sure the machine is up and running before
trying to connect to it again.Signed-off-by: Steven Rostedt
20 Oct, 2011
1 commit
-
When setting the next kernel to boot to with grub, do not opencode
the reboot operation. The normal reboot operation can be modified by
config options (namely POWERCYCLE_AFTER_REBOOT). This needs to affect
all reboots. Remove the opencoded reboot to make sure that any changes
to the reboot code also affect all reboots.Signed-off-by: Steven Rostedt
17 Oct, 2011
17 commits
-
The IF statements for DEFAULTS and TEST_START sections now handle
complex statements (&&,||)Example:
TEST_START IF (DEFINED ALL_TESTS || ${MYTEST} == boottest) && ${MACHINE} == gandalf
Signed-off-by: Steven Rostedt
-
The order for some of the keywords on a section line
(TEST_START or DEFAULTS) does not really matter. Simply need
to remove the keyword from the line as we process it and
evaluate the next keyword in the line. By removing the keywords
as we find them, we do not need to keep track of where on the
line they were found.Signed-off-by: Steven Rostedt
-
The make_min_config test will turn off one config at a time and check
if the config boots or not, and if it does, it will remove that config
plus any config that depended on that config.ktest already looks if a config has a dependency and will try the
dependency config first. But by sorting the configs and trying the
config with the most configs dependent on it, we can shrink the
minconfig faster.Signed-off-by: Steven Rostedt
-
Have IF statements process if a config variable or option has been
defined or not. Can use NOT DEFINED in the case for telling if
a variable or option has not been defined.DEFAULTS IF NOT DEFINED SSH_USER
SSH_USER = rootSigned-off-by: Steven Rostedt
-
The OVERRIDE keyword will allow options defined in the given
DEFAULTS section to override options defined in previous DEFAULT
sections.Normally, options will error if they were previous defined.
The OVERRIDE keyword allows options that have been previously
defined to be changed in the given section.Note, the same option can not be defined in the same DEFAULT section
even if that section is marked as OVERRIDE.Signed-off-by: Steven Rostedt
-
The code that handles parsing the TEST_TYPE and DEFAULT code share
a lot of common functionality. Combine the two and add a if statement
that does what is different between them.Signed-off-by: Steven Rostedt
-
Have the reading of the config file allow reading of other config
files using the INCLUDE keyword. This allows multiple config files
to share config options.Signed-off-by: Steven Rostedt
-
Allow ==, !=, =, to be used in IF statements
to compare if a section should be processed or not.For example:
BITS := 32
DEFAULTS IF ${BITS} == 32
MIN_CONFIG = ${CONFIG_DIR}/config-32
ELSE
MIN_CONFIG = ${CONFIG_DIR}/config-64Signed-off-by: Steven Rostedt
-
Add IF keyword to sections within the config. Also added an ELSE
keyword that allows different config options to be set for a given
section.For example:
TYPE := 1
STATUS := 0DEFAULTS IF ${TYPE}
[...]
ELSE IF ${STATUS}
[...]
ELSE
[...]The above will process the first section as $TYPE is true. If it
was false, it would process the last section as $STATUS is false.Signed-off-by: Steven Rostedt
-
Even if REBOOT_ON_ERROR is set, it becomes annoying that the target
machine is rebooted when a config option is incorrect or a build
fails. There's no reason to reboot the target for host only issues.Signed-off-by: Steven Rostedt
-
When ktest.pl reboots, it will usuall wait SLEEP_TIME seconds of idle
console before starting the next test. By setting the
REBOOT_SUCCESS_LINE, ktest will not wait SLEEP_TIME when it detects the
line while rebooting to a new kernel.Signed-off-by: Steven Rostedt
-
There's cases where running the same kernel over and over again
is useful, and being able to not install the same kernel can
save time between tests.Add a NO_INSTALL option that tells ktest.pl to not install the
new kernel.Signed-off-by: Steven Rostedt
-
Currently if the grub menu that is supplied is not found, it will
just boot into the last grub menu in menu.lst. Fail instead of
confusing the user why their kernel is not booting.Signed-off-by: Steven Rostedt
-
Several places that call reboot do the same thing with respect to the
monitor. By adding this code into the reboot code, redundant code is
removed and it paves the way for the the reset time patch.Signed-off-by: Andrew Jones
Link: http://lkml.kernel.org/r/1313155932-20092-4-git-send-email-drjones@redhat.com
Signed-off-by: Steven Rostedt -
Only save the .config file if we're doing mrproper
Signed-off-by: Andrew Jones
Link: http://lkml.kernel.org/r/1313155932-20092-3-git-send-email-drjones@redhat.com
Signed-off-by: Steven Rostedt -
Signed-off-by: Andrew Jones
Link: http://lkml.kernel.org/r/1313155932-20092-2-git-send-email-drjones@redhat.com
Signed-off-by: Steven Rostedt -
In testing one of my boxes, I found that I only wanted to build and
install the kernel. I wanted to manually reboot the box and test it.
Adding a TEST_TYPE option "install" allows this to happen.Signed-off-by: Steven Rostedt
