For "pretty" V=0 builds, change 'make check' to run 'runtests' in
non-verbose mode, so we just get a summary of test results, instead of
the raw test output. Without V=0, the default is unchanged, so we
still print out all test output by default.
Change-Id: I554f9d32ed5a9cd27e83fef6245af589d91e801f
Reviewed-on: https://gerrit.openafs.org/14619
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Cheyenne Wills <cwills@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Currently 'make check' always runs all tests. We can run individual
tests manually, but doing so is a bit cumbersome to do under the same
environment as 'make check', since doing so means running something
like this:
$ MAKECHECK=1 $(abs_top_srcdir)/tests/libwrap @TOP_OBJDIR@/lib \
./runtests opr/fmt util/ktime
To make it easier to run single tests introduce a way of calling 'make
check' like this:
$ make check TESTS='opr/fmt util/ktime'
Which will run the same commands as 'make check', but will run
runtests with only the specified tests, instead of running the default
list.
Some makefiles currently use a "TESTS" or "tests" variable to list
their test binaries; rename them all to "BINS" to avoid conflicting
with this new use for "TESTS" and to make our makefiles a little more
consistent.
Change-Id: I427f83be0d4571794644a97123bcd1f32427bd05
Reviewed-on: https://gerrit.openafs.org/14317
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
There is some code in tests/rpctestlib that appears to be for testing
fileserver RPCs and callback processing, added in commit 262a678d (An
RPC test dispatch library for vice). However, it has never been used,
it seems unlikely that it will be used anytime soon, and it's not
clear if it even works (it contains many hard-coded references to
interfaces and IP addresses).
Just remove the unused code, and some references to rpctestlib. It can
always be added back if needed (or more likely, reimplemented to be
more in line with our current test framework in tests/ ).
Change-Id: Ied3be7474581d8ee75ae000815bb7364d143fd31
Reviewed-on: https://gerrit.openafs.org/14617
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Running clang's static analysis revealed a possible divide by zero
condition.
There is a random chance of the divide by zero.
- it has to be in the first pass of the main loop testing events
(counter = 0)
- 90% chance path : if (counter < (NUMEVENTS -1) &&
random() % 10 == 0) -- needs to be false
- 25% chance path: if (random() % 4 == 0) -- needs to be true
if the above conditions are met, the statement
int victim = random() % counter
is a divide by zero.
Add a check to ensure the counter is greater than zero. Add a comment
to document that only events prior to the current event are randomly
selected.
Change-Id: I4b4e73fa324842bb504bcc952079af15aea8a6a3
Reviewed-on: https://gerrit.openafs.org/14501
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Each server process can log a couple of different warnings about the
server keys found on disk:
- If afsconf_GetLatestKey() returns success (indicating a single-DES
key is present), we call LogDesWarning().
- If afsconf_CountKeys() returns 0 (indicating there are no keys at
all on disk), we log a warning that all authenticated access will
fail.
Currently, the code to do these checks and log the relevant warning is
duplicated across the startup code for nearly every server process. To
avoid this duplication, and to make sure the checks aren't
accidentally skipped for anyone, move these checks to
afsconf_BuildServerSecurityObjects, which every server process calls.
We must add an additional parameter to
afsconf_BuildServerSecurityObjects to handle the different logging
mechanism these servers use, but afsconf_BuildServerSecurityObjects is
declared in a public header (cellconfig.h), and is exported in a
public library (libafsauthent). So to avoid changing a public symbol,
introduce a new variant of the function, called
afsconf_BuildServerSecurityObjects_int. Declare this in a new internal
header, authcon.h.
We don't have easily-usable logging functions for upserver and butc,
so just don't log the warnings for those. For ubik servers, don't
update ubik_SetServerSecurityProcs to use the new function; the
initial call to afsconf_BuildServerSecurityObjects_int in the server's
startup code will cover logging the warning on startup.
Change-Id: I5d5fceefdaf907f96db9f1c0d21ceb6957299a59
Reviewed-on: https://gerrit.openafs.org/10831
Tested-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
The SOURCE and BUILD environment variables have been changed to
C_TAP_SOURCE and C_TAP_BUILD in the new version of c-tap-harness. The
runtests command syntax has changed as well.
Convert all of the old SOURCE and BUILD environment variables to the new
C_TAP_SOURCE and C_TAP_BUILD names.
Add the required -l command line option to specify the test list.
Add the new runtests -v option to run the tests in verbose mode to make
it easier to see which tests failed.
Change-Id: I209a6dc13d6cd1507519234fce1564fc4641e70b
Reviewed-on: https://gerrit.openafs.org/14295
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
tests/auth/writekeyfile.c contains some code used to generate
tests/auth/KeyFile, which is used to test code interpreting the
old-style KeyFile format. This code currently has a few problems:
- We don't check the results of afstest_mkdtemp, which could allow
symlink attacks from other users on the system.
- We duplicate some logic from afstest_BuildTestConfig, in order to
build a temporary config dir.
- writekeyfile isn't built or run by default (it only exists to
generate KeyFile, so it's almost never run), so eventual bitrot is
quite likely, and the existing code already generates warnings.
To avoid this, change writekeyfile.c to use the existing
afstest_BuildTestConfig to generate a local config dir. To ensure we
avoid bitrot, build writekeyfile by default, and create a test to run
it, to make sure it can generate a KeyFile as expected.
Note that the KeyFile.short we test against is different than the
KeyFile currently in the tree. The existing KeyFile was generated from
an older OpenAFS release, which always generated 100-byte KeyFiles,
even if we only have a few keys. The current codebase only writes out
as much key data as needed, so the generated KeyFiles are shorter (but
still understandable by older OpenAFS releases).
Keep the old 100-byte KeyFile around, since that's what older OpenAFS
would generate, and create a new KeyFile.short to test against, to
make sure our code for generating KeyFiles doesn't change any further.
Change-Id: Ibe9246c6dd808ed2b2225dd7be2b27bbdee072fd
Reviewed-on: https://gerrit.openafs.org/14246
Reviewed-by: Cheyenne Wills <cwills@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Commit "Build tests by default" 68f406436c
changes the build so tests are always built.
On Solaris 10 the build fails because nanosleep is in librt, which we do
not link against.
Replace nanosleep with usleep. This avoids introducing extra configure
tests just for Solaris 10.
Note that with Solaris 11 nanosleep was moved from librt to libc, the
standard C library.
Change-Id: I6639f32bb8c8ace438e0092a866f06561dad54f1
Reviewed-on: https://gerrit.openafs.org/14244
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Commit "Build tests by default" 68f406436c
changes the build so tests are always built.
On Solaris 10 Update 10 and earlier the build fails because the mkdtemp
function is not available.
Introduce a wrapper 'afstest_mkdtemp' that uses mkdtemp if available,
otherwise uses mktemp/mkdir.
Change-Id: I0118f838ed9a89927e2ddac4cad822574601558a
Reviewed-on: https://gerrit.openafs.org/14243
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Currently, the rx/event-t tests schedule a bunch of events up to 3
seconds in the future, and then we sleep for 3 seconds to give them a
chance to run. Since we're cutting it so close, this can rarely result
in a few events not being run (observed occasionally on FreeBSD 12.1,
where we failed to run about 3 events out of 10000).
To avoid this, just sleep for 4 seconds instead of 3. Also print out a
little more info regarding the number of fired/cancelled events, so we
can see the event count when it's wrong.
Change-Id: I6269bea2c245aeed00c129ff638423d0fa81ad23
Reviewed-on: https://gerrit.openafs.org/14160
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Currently, 'softsig-helper -buserror' causes a SIGBUS on most
platforms, but can result in SIGSEGV on FreeBSD by default (at least
on 11.3-RELEASE). Skip the test on FreeBSD, until we can provide a
more reliable way to generate SIGBUS.
Note that when the sysctl machdep.prot_fault_translation is set to 1,
'softsig-helper -buserror' generates a SIGBUS instead of SIGSEGV,
suggesting that generating a SIGBUS here is the old 'compat' behavior.
When machdep.prot_fault_translation is 0 (the default), the code path
in the FreeBSD kernel that dictates whether to send a SIGBUS or
SIGSEGV in this situation depends on some autodetection heuristics,
and so may produce different results depending on FreeBSD releases or
even compiler settings (due to detection of ABI based on some ELF
notes in the relevant binary).
For some details on this sysctl, see
<https://www.freebsd.org/news/status/report-2019-07-2019-09.html#Signals-delivered-on-unhandled-Page-Faults>
or the FreeBSD source code. In 11.3-RELEASE, the decision to issue a
SIGBUS or SIGSEGV can be found around sys/amd64/amd64/trap.c:355.
Change-Id: Ib75b43cc12302532ee87a3744fc364424f2a3ca6
Reviewed-on: https://gerrit.openafs.org/14145
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
The 'perl' binary may not be /usr/bin/perl, depending on the system.
For example, on modern FreeBSD it tends to be /usr/local/bin/perl
instead.
To avoid relying on perl to be in a specific location, just run via
/usr/bin/env instead, so we pick up perl from $PATH instead.
Change-Id: Ic8dc247c82342ff79dfa80426c489ccb8e3e1450
Reviewed-on: https://gerrit.openafs.org/14144
Tested-by: Andrew Deason <adeason@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Currently, opr_cache_init requires that opts->n_buckets is a power of
2 (since our underlying opr_dict requires this). However, callers may
want to pick a number of buckets based on some other value. Requiring
each caller to calculate the nearest power-of-2 is annoying, so
instead just have opr_cache_init itself calculate a nearby power of 2.
That is, with this commit, opts->n_buckets is allowed to not be a
power of 2; when it's not a power of 2, opr_cache_init will calculate
the next highest power of 2 and use that as the number of buckets.
Change-Id: Icd3c56c1fe0733e3dac964ea9a98ff7b436254e6
Reviewed-on: https://gerrit.openafs.org/14122
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Marcio Brito Barbosa <mbarbosa@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
The auth/superuser-t test runs an Rx server and client in two child
processes. If the client process tries to contact the server before
the server has started listening on its port, some tests involving
RPCs can fail (notably test 39, "Can run a simple RPC").
Normally if we try to contact a server that's not there, Rx will try
resending its packets a few times, but on Linux with AFS_RXERRQ_ENV,
if the port isn't open at all, we can get an ICMP_PORT_UNREACH error,
which causes the relevant Rx call to die immediately with
RX_CALL_DEAD.
This means that if the auth/superuser-t client is only just a bit
faster than the server starting up, tests can fail, since the server's
port is not open yet.
To avoid this, we can wait until the server's port is open before
starting the client process. To do this, have the server process send
a SIGUSR1 to the parent after rx_Init() is called, and have the parent
process wait for the SIGUSR1 (waiting for a max of 5 seconds before
failing). This should guarantee that the server's port will be open by
the time the client starts running.
Note that before commit 086d1858 (LINUX: Include linux/time.h for
linux/errqueue.h), AFS_RXERRQ_ENV was mistakenly disabled on Linux
3.17+, so this issue was probably not possible on recent Linux before
that commit.
Change-Id: I0032a640b83c24f72c03e7bea100df5bc3d9ed4c
Reviewed-on: https://gerrit.openafs.org/14109
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Cheyenne Wills <cwills@sinenomine.net>
Commit 48fbb45 (opr: Introduce opr_cache) added a new test (cache-t),
but did not update the .gitignore file for it.
Change-Id: I6de6130257a62f495ac942c05937eb109ce84a75
Reviewed-on: https://gerrit.openafs.org/14102
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
opr/softsig-t can produce a core file as part of its test.
Change-Id: I3bc7e587151e5915038e31887018889a7ffa6993
Reviewed-on: https://gerrit.openafs.org/14101
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Ever since commit f0774acd (Introduce TAP tests of man pages for
command_subcommand), we've had tests to check that we have man pages
for every subcommand in a command suite. This was done for several
command suites, including 'bos', and 'fs', but the bos and fs tests were
never added to the TESTS file.
Add them, so the tests run by default in a 'make check'. Fortunately,
the tests still pass today.
Change-Id: I90c006845d054fa3e795203bb1deff675e558622
Reviewed-on: https://gerrit.openafs.org/14073
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Add a simple general-purpose in-memory cache implementation, called
opr_cache. Keys and values are simple flat opaque buffers (no complex
nested structures allowed), hashing is done with jhash, and cache
eviction is mostly random with some LRU bias.
Partly based off a different implementation by
mbarbosa@sinenomine.net.
Change-Id: I16b5988947ff603dfe31613cd7be3908a69264e5
Reviewed-on: https://gerrit.openafs.org/13884
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
The vos tests start a temporary vlserver process, which is problematic
when the local system already has an installed vlserver. Attempt to
temporarily bind a socket to the vlserver port, and if unable to bind
with an EADDRINUSE error, assume the vlserver is already running and
skip these tests.
Change-Id: I1dd3bc4c7ebcd2c7bffc8aca422222a50058090e
Reviewed-on: https://gerrit.openafs.org/14021
Reviewed-by: Cheyenne Wills <cwills@sinenomine.net>
Reviewed-by: Andrew Deason <adeason@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
The vos-t test adds a set of 10.* test addresses to a test vlserver and
runs vos to read them back. When the test is run in an environment
where hosts have been assigned in the 10.* internal network, vos will resolve
the addresses to hostnames and the test fails. Pass the -noresolve
option to vos for this test when checking for the expected list of
addresses.
Example test output before this commit:
./vos-t
...
# seen: 10.0.0.0
10.0.0.1
myhost.example.com
10.0.0.3
...
not ok 5 - vos output matches
Change-Id: Ief43fe180a0dfff211f28d5f47be6224270907a3
Reviewed-on: https://gerrit.openafs.org/14020
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Currently, if we encounter an error and 'goto out' after starting the
test vlserver, we'll exit without stopping the test vlserver. This can
confuse the test harness, causing 'runtests' to hang forever.
To avoid this, move the afstest_StopServer() call to also run when
we're bailing out, but only if the server has actally started of
course.
Change-Id: Ice5a56c20bc8d2eac85b3e760850c4d85e4601a8
Reviewed-on: https://gerrit.openafs.org/13992
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Currently, in tests/volser/vos-t.c we call afs_com_err as
"authname-t", which is clearly a mistake during some code refactoring
(introduced in commit 2ce3fdc5, "tests: Abstract out code to produce a
Ubik client").
We could just change this to "vos-t", but instead of specifying
constant strings everywhere, change this to figure out what the
current command is called, and just use that. Put this code into a new
function, afstest_GetProgname, and convert existing tests to use that
instead of hard-coding the program name given to afs_com_err.
Change-Id: I3ed02c89f93798568783c7d717e8fb2e39dcce14
Reviewed-on: https://gerrit.openafs.org/13991
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Currently, the volser/vos test starts a local vlserver to communicate
with. If the vlserver dies during startup, the spawned 'vos'
subprocesses take forever to run, since we need to wait for our Rx
calls to timeout for every operation.
To make it less annoying to detect and investigate errors that might
cause the vlserver to fail during startup, check if the vlserver dies
right away. We already sleep for 5 seconds when starting the vlserver,
so just check if the pid still exists after those 5 seconds.
Change-Id: I6c33059542fa975e4cb389b718f9da190cd13289
Reviewed-on: https://gerrit.openafs.org/13942
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Cheyenne Wills <cwills@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
The manpage tests have a couple of problems when running for objdir
builds:
- We try to specify './tests-lib/perl5' as a directory to find our
helper library. However, the cwd when we're running the tests is in
an objdir build, where the helper library is in the srcdir. Fix this
by using the SOURCE env var specified by the tests wrapper.
- All of these tests specify the directory in which to find the man
pages in a subdir of BUILD, but our manpages are located in the src
dir (since they are built by regen.sh, not by configure/make). Fix
this by specifying a SOURCE-based directory instead.
To avoid needing to make the same change for each of these tests, also
refactor the manpage tests so each test only needs to specify the
subdirectory and command name, and get rid of some of the common
boilerplate.
Change-Id: I96be199b1dec8db0545ae3cf19d2595c4afe4cdd
Reviewed-on: https://gerrit.openafs.org/13940
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Fix a few miscellaneous issues with building and running our tests in
objdir builds:
- Our C tests use -I$(srcdir)/../.. in the CFLAGS, so we can #include
<tests/tap/basic.h>. However, basic.h actually gets copied from
src/external/c-tap-harness/tests/tap/ to tests/tap/ during the
build, and so basic.h is available in the objdir, not srcdir. For
objdir builds, this causes building the tests to fail with failing
to find basic.h. Fix this to use TOP_OBJDIR as the include path
instead.
- Our 'make check' in tests/ tries to run ./libwrap; but our cwd will
be in the objdir for objdir builds, and libwrap is a script in our
srcdir. Fix this to run libwrap from the srcdir path.
- In tests/opr/softsig-t, it tries to find the 'softsig-helper' binary
in the same dir as 'softsig-t'. However, softsig-t is just a script
in the srcdir, but softsig-helper is a binary built in the objdir.
Fix this to use the BUILD env var provided by the tests wrapper, by
default.
Change-Id: Iff642613bfc88d0d7e348660dc62f59e6fa8af75
Reviewed-on: https://gerrit.openafs.org/13939
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
There is a perhaps-surprisingly large amount of code disabled behind
directives like '#if 0', '#ifdef notdef', and '#ifdef notyet'. At
best, this code is clutter, and at worst some of it is
confusing/outdated, and/or confusingly nested inside other
preprocessor conditionals. Sometimes this disabled code shows up when
grepping the tree, and causes a nuisance when refactoring related
areas of code.
Get rid of all of it. If anyone ever wants this code back, it can
always be restored by reverting portions of this commit.
Also delete some comments that clearly refer to the disabled code, and
in some cases, adjust the adjacent comments to make sense accordingly.
This commit doesn't touch any files in src/external/.
Change-Id: If260a41257e8d107930bd3c177eddb8ab336f0d1
Reviewed-on: https://gerrit.openafs.org/13683
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: Benjamin Kaduk <kaduk@mit.edu>
This lock-free library toolkit is intriguing and may be the subject
of future work, but currently nothing uses this code, and these files
are just clutter.
Remove src/mcas and stop mentioning it in SOURCE-MAP; don't reference
it in the rpctests, either.
Reviewed-on: https://gerrit.openafs.org/12682
Tested-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Stephan Wiesand <stephan.wiesand@desy.de>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
(cherry picked from commit bfc5d1ada2)
Change-Id: I98bec6f0a91e4aad05846a6791719cac63050f02
Reviewed-on: https://gerrit.openafs.org/13538
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
In afsconf_BuildServerSecurityObjects, create a server security object
for rxgk. Currently, this will only accept printed rxgk tokens, not
tokens negotiated via GSSNegotiate. Future commits will add
functionality to handle user-negotiated tokens, fileserver-specific
creds, etc.
Change-Id: Ie2bbef0d591641e80bb85240316c4ee5f9f8ff05
Reviewed-on: https://gerrit.openafs.org/12941
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
The last references to these objects were removed with commit
3828c257ae
"dead-code-and-prototyes-20060214".
A few mentions of CBS and BBS are left in the documentation as
historical references:
- doc/man-pages/pod1/rxgen.pod
- src/kauth/AuthServer.mss
Change-Id: Ia24eef7bb1509ff10d11de5c51e688e27f69417a
Reviewed-on: https://gerrit.openafs.org/13324
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Add unit tests for the utility functions to convert between partition
names and partition ids.
Change-Id: I4b12f9d611cb9f3ce49909cda5cbcedd3e6c3d10
Reviewed-on: https://gerrit.openafs.org/13176
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: Benjamin Kaduk <kaduk@mit.edu>
Use the NUMEVENTS symbol which defines the array size instead of an
incorrect hard coded number when checking if a second event can be added
to be fired at the same time. This fixes a potential out of bounds
access of the event test array.
Also update the comment which incorrectly mentions the incorrect number
of events in the test.
Change-Id: I4f993b42e53e7e6a42fa31302fd1baa70e9f5041
Reviewed-on: https://gerrit.openafs.org/12762
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Go over all consumers of the rx event framework and normalize its usage
according to the following principles:
rxevent_Post() is used to create an event, and it returns an event
handle (with a reference on the event structure) that can be used
to cancel the event before its timeout fires. (There is also an
additional reference on the event held by the global event tree.)
In all(*) usage within the tree, that event handle is stored within
either an rx_connection or an rx_call. Reads/writes to the member variable
that holds the event handle require either the conn_data_lock or call
lock, respectively -- that means that in most cases, callers of
rxevent_Post() and rxevent_Cancel() will be holding one of those
aforementioned locks. The event handlers themselves will need to
modify the call/connection object according to the nature of the
event, which requires holding those same locks, and also a guarantee
that the call/connection is still a live object and has not been
deallocated! Whether or not rxevent_Cancel() succeeds in cancelling
the event before it fires, whenever passed a non-NULL event structure
it will NULL out the supplied pointer and drop a reference on the
event structure. This is the correct behavior, since the caller
has asked to cancel the event and has no further use for the event
handle or its reference on the event structure. The caller of
rxevent_Cancel() must check its return value to know whether or
not the event was cancelled before its handler was able to run.
The interaction window between the call/connection lock and the lock
protecting the red/black tree of pending events opens up a somewhat
problematic race window. Because the application thread is expected
to hold the call/connection lock around rxevent_Cancel() (to protect
the write to the field in the call/connection structure that holds
an event handle), and rxevent_Cancel() must take the lock protecting
the red/black tree of events, this establishes a lock order with the
call/connection lock taken before the eventTree lock. This is in
conflict with the event handler thread, which must take the eventTree
lock first, in order to select an event to run (and thus know what
additional lock would need to be taken, by virtue of what handler
function is to be run). The conflict is easy to resolve in the
standard way, by having a local pointer to the event that is obtained
while the event is removed from the red/black tree under the eventTree
lock, and then the eventTree lock can be dropped and the event run
based on the local variable referring to it. The race window occurs
when the caller of rxevent_Cancel() holds the call/connection lock,
and rxevent_Cancel() obtains the eventTree lock just after the event
handler thread drops it in order to run the event. The event handler
function begins to execute, and immediately blocks trying to obtain
the call/connection lock. Now that rxevent_Cancel() has the eventTree
lock it can proceed to search the tree, fail to find the indicated event
in the tree, clear out the event pointer from the call/connection
data structure, drop its caller's reference to the event structure,
and return failure (the event was not cancelled). Only then does the
caller of rxevent_Cancel() drop the call/connection lock and allow
the event handler to make progress.
This race is not necessarily problematic if appropriate care is taken,
but in the previous code such was not the case. In particular, it
is a common idiom for the firing event to call rxevent_Put() on itself,
to release the handle stored in the call/connection that could have
been used to cancel the event before it fired. Failing to do so would
result in a memory leak of event structures; however, rxevent_Put() does
not check for a NULL argument, so a segfault (NULL dereference) was
observed in the test suite when the race occurred and the event handler
tried to rxevent_Put() the reference that had already been released by
the unsuccessful rxevent_Cancel() call. Upon inspection, many (but not
all) of the uses in rx.c were susceptible to a similar race condition
and crash.
The test suite also papers over a related issue in that the event handler
in the test suite always knows that the data structure containing the
event handle will remain live, since it is a global array that is allocated
for the entire scope of the test. In rx.c, events are associated with
calls and connections that have a finite lifetime, so we need to take care
to ensure that the call/connection pointer stored in the event remains
valid for the duration of the event's lifecycle. In particular, even an
attempt to take the call/connection lock to check whether the corresponding
event field is NULL is fraught with risk, as it could crash if the lock
(and containing call/connection) has already been destroyed! There are
several potential ways to ensure the liveness of the associated
call/connection while the event handler runs, most notably to take care
in the call/connection destruction path to ensure that all associated
events are either successfully cancelled or run to completion before
tearing down the call/connection structure, and to give the pending event
its own reference on the associated call/connection. Here, we opt for
the latter, acknowledging that this may result in the event handler thread
doing the full call/connection teardown and delay the firing of subsequent
events. This is deemed acceptable, as pending events are for intentionally
delayed tasks, and some extra delay is probably acceptable. (The various
keepalive events and the challenge event could delay the user experience
and/or security properties if significantly delayed, but I do not believe
that this change admits completely unbounded delay in the event handler
thread, so the practical risk seems minimal.)
Accordingly, this commit attempts to ensure that:
* Each event holds a formal reference on its associated call/connection.
* The appropriate lock is held for all accesses to event pointers in
call/connection structures.
* Each event handler (after taking the appropriate lock) checks whether
it raced with rxevent_Cancel() and only drops the call/connection's
reference to the event if the race did not occur.
* Each event handler drops its reference to the associated call/connection
*after* doing any actions that might access/modify the call/connection.
* The per-event reference on the associated call/connection is dropped by
the thread that removes the event from the red/black tree. That is,
the event handler function if the event runs, or by the caller of
rxevent_Cancel() when the cancellation succeed.
* No non-NULL event handles remain in a call/connection being destroyed,
which would indicate a refcounting error.
(*) There is an additional event used in practice, to reap old connections,
but it is effectively a background task that reschedules itself
periodically, with no handle to the event retained so as to be able
to cancel it. As such, it is unaffected by the concerns raised here.
While here, standardize on the rx_GetConnection() function for incrementing
the reference count on a connection object, instead of inlining the
corresponding mutex lock/unlock and variable access.
Also enable refcount checking unconditionally on unix, as this is a
rather invasive change late in the 1.8.0 release process and we want
to get as much sanity checking coverage as possible.
Change-Id: I27bcb932ec200ff20364fb1b83ea811221f9871c
Reviewed-on: https://gerrit.openafs.org/12756
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
We currently do not properly handle the case where a thread runs
rxevent_Cancel() in parallel with the event-handler thread attempting
to fire that event, but the test suite only picked up on this issue
in a handful of the Debian automated builds (somewhat less-resourced
ones, perhaps).
Modify the event scheduling algorithm in the test so as to create a
larger chunk of events scheduled to fire "right away" and thereby
exercise the race condition more often when we proceed to cancel
a quarter of events "right away".
Change-Id: I50f55fd532901147cfda1a5f40ef949bf3270401
Reviewed-on: https://gerrit.openafs.org/12755
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Mark Vitale <mvitale@sinenomine.net>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
If the build directory happened to contain shell metacharacters, like
the ~ in /build/openafs-vb8tid/openafs-1.8.0~pre1 used by the Debian
builders, Perl was running softsig-helper via an intermediate sh -c,
which would then intercept the signals we tried to send to
softsig-helper. Use the list syntax to avoid this sh -c.
Change-Id: I054b9c8f606e197accb414bfe3f89719255c62c4
Reviewed-on: https://gerrit.openafs.org/12488
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
No need to leave the shell process hanging around.
In particular, if we are manually running softsig-helper under
libwrap to debug test failures, the child process of the shell is
another shell, which interprets some signals that we wanted to
be passed through, like SIGTERM. On the other hand, once the
softsig-helper is exec()'d, you basically need another shell to
terminate it, which is a different problem....
Change-Id: Iff7c519886a018cb68e692746d40c427b6299457
Reviewed-on: https://gerrit.openafs.org/12490
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Anders Kaseorg <andersk@mit.edu>
Tested-by: Anders Kaseorg <andersk@mit.edu>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Fix the loop condition when scanning the signal number to name table to
convert a signal number to a name. Instead of looping sizeof(size_t)
times, loop for the number of elements in the table.
This bug was masked on 64 bit-platforms, since the signal number to name
table table currently has 8 elements, which is coincidently the same as
sizeof(size_t) on 64-bit platforms. The bug becomes apparent on 32-bit
systems; only the first 4 elements of the table are checked.
Example error output before this fix:
$ cd tests
$ ./libwrap ../lib ./runtests -o opr/softsig
1..11
ok 1
ok 2
ok 3
ok 4
ok 5
not ok 6
# Failed test in ./opr/softsig-t at line 57.
# got: 'Received UNK
# '
# expected: 'Received TERM
# '
not ok 7
# Failed test in ./opr/softsig-t at line 60.
# got: 'Received UNK
# '
# expected: 'Received USR1
# '
not ok 8
# Failed test in ./opr/softsig-t at line 63.
# got: 'Received UNK
# '
# expected: 'Received USR2
# '
ok 9 - Helper exited on KILL signal.
ok 10 - Helper exited on SEGV signal.
ok 11 # skip Skipping buserror test; SIGBUS constant is not defined.
# Looks like you failed 3 tests of 11.
Change-Id: I863cc9f3650c4a5e9ac9159d90e063b986a8460a
Reviewed-on: https://gerrit.openafs.org/12367
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
This fixes test failures observed on new Debian build servers that no
longer install tzdata by default. As the tests expect, EST is defined
as UTC−05:00 with no daylight saving time.
Change-Id: Ida8cb33687b5d87761cb0422e446afd99246d47a
Reviewed-on: https://gerrit.openafs.org/12414
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Some libc implementations will crash when NULL string arguments are given to
*printf. Avoid passing NULL string arguments in the make check tests that did
so, and pass the string "(null)" instead.
Change-Id: I65f11a3eef88d1c7b210c867ae0c40018160f55a
Reviewed-on: https://gerrit.openafs.org/12377
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Older versions of the perl POSIX module do not define the SIGBUS symbol, which
causes the opr/softsig-t perl test to fail to compile. Instead of trying to
defined SIGBUS, which may be platform dependent, skip the buserror unit test on
these older platforms.
Change-Id: Ib8cfd77215ea43566e9d47b501d4989556b83734
Reviewed-on: http://gerrit.openafs.org/12186
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
In order to start the softsig test helper properly,
the full path of this program is necessary.
FIXES 132246
Change-Id: I4e9ff1e62a0b82078338eeaf0d4368ac1b35dccc
Reviewed-on: http://gerrit.openafs.org/11977
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Michael Meffie <mmeffie@sinenomine.net>
Tested-by: Michael Meffie <mmeffie@sinenomine.net>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Resolves this warning:
keys-t.c: In function ‘copy’:
keys-t.c:63:6: warning: ignoring return value of ‘write’, declared with attribute warn_unused_result [-Wunused-result]
write(out, block, len);
^
Change-Id: If2427f2658b428091ffba3d11643ad95f193a67d
Reviewed-on: http://gerrit.openafs.org/11957
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Resolves this warning:
vos-t.c: In function ‘TestListAddrs’:
vos-t.c:60:5: warning: ignoring return value of ‘pipe’, declared with attribute warn_unused_result [-Wunused-result]
pipe(outpipe);
^
Change-Id: I7eb58a91b5a7d9df18a4952400f74c79299e857d
Reviewed-on: http://gerrit.openafs.org/11958
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Resolves this warning on 32-bit GCC:
jhash-t.c: In function ‘main’:
jhash-t.c:60:4: warning: this decimal constant is unsigned only in ISO C90
is_int(3704403432, opr_jhash(test, 2, 0),
^
jhash-t.c:62:4: warning: this decimal constant is unsigned only in ISO C90
is_int(3704403432, opr_jhash_int2(test[0], test[1], 0),
^
Change-Id: Ie3ab0f5aacdc719fa63f32e545b5863ec351f5eb
Reviewed-on: http://gerrit.openafs.org/11961
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Resolves this warning with clang:
time-t.c:46:8: warning: absolute value function 'abs' given an argument of type 'long' but has parameter of type 'int' which may cause
truncation of value [-Wabsolute-value]
ok(abs(osTime - osNow) < 2, "opr_time_Now returns a reasonable value");
^
time-t.c:46:8: note: use function 'labs' instead
ok(abs(osTime - osNow) < 2, "opr_time_Now returns a reasonable value");
^~~
labs
Change-Id: Ib98069e1349161d936c8ada0e69f9b33d2f71ce3
Reviewed-on: http://gerrit.openafs.org/11965
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Signals and pthreaded applications are a poor match. OpenAFS has had
the softsig system (currently in src/util/softsig.c) in an attempt to
alleviate some of these problems. However, that implementation itself
has a number of problems. It uses signal functions that are unsafe in
pthreaded applications, and uses pthread_kill within its signal
handlers. Over the years it has been responsible for a number of
portability bugs.
The old implementation continues to receive signals in the main thread
of the application. However, the handler code is run within a seperate
signal handler thread. When the main thread receives a signal a stub
handler is invoked, which simply pthread_kill()s the signal handler
thread.
The new implementation simplifies things by only receiving signals in
the handler thread. It uses only pthread-compatible signal functions,
and invokes no code from within async signal handlers.
A complete test suite is supplied.
Change-Id: I4bac68c2f853f1e7578b54ddced3833a97dd3f82
Reviewed-on: http://gerrit.openafs.org/6947
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Chas Williams <3chas3@gmail.com>
Reviewed-by: Daria Brashear <shadow@your-file-system.com>
The MT_LIBS library list already includes XLIBS, so there's no need
to specify both on a link line.
Change-Id: I8594b1b6e1a16af741b40822cbce49e846b26f49
Reviewed-on: http://gerrit.openafs.org/8904
Reviewed-by: Daria Brashear <shadow@your-file-system.com>
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
The return value of asprintf() is the number of bytes printed, or -1 if there
was an error allocating a large enough buffer. In the latter case, the value
of the result string is undefined, and so it cannot be counted on to be NULL.
This change fixes numerous places where the result of asprintf is checked
incorrectly (by examining the output pointer and not the return value) or not
at all.
Change-Id: I9fef14d60c096795d59c42798f3906041fb18c86
Reviewed-on: http://gerrit.openafs.org/9978
Reviewed-by: Benjamin Kaduk <kaduk@mit.edu>
Reviewed-by: D Brashear <shadow@your-file-system.com>
Tested-by: BuildBot <buildbot@rampaginggeek.com>
'make clean' and 'make maintainer-clean' still leave around a fair
number of droppings, prior to this commit.
We were not descending into the 'tests' top-level directory while
cleaning. Furthermore, tests/opr/Makefile needed $(LT_CLEAN), and
tests/rx/Makefile needed to spell it correctly.
The libtoolization places a lot of files to be removed in the
'pristine' target.
The processing used to implement the =include directive in the pod
sources for the man pages leaves around the non-.in versions of
files; we should clean that up in the 'pristine' target as well.
The 'pristine' target should likewise remove the man pages which
are generated from the pod files.
Additionally, the documentation build uses a Doxyfile which is
output by configure; that should be removed (if present) by the
'distclean' target.
When hcrypto was converted to libtool, the use of ${OBJECTS} in
the clean target was missed, so we were leaving around most of the
actual object files -- $(LT_CLEAN) does not handle this for us.
Change the rule to remove *.o as is done elsewhere.
The conversion of libafsrpc to libtool added a convenience library
libafsrpc_sys.la, and changed how syscall.o was generated on
most architectures, to be the result of compiling an empty .c file
(instead of just an empty .o file). This introduced a new
intermediate file, syscall.c, which must be cleaned up.
tvolser was only listing volserver and not vos in its list of
executables to remove while cleaning.
The conversion of venus/test to libtool was not done quite right.
Makefile.libtool and the .lo suffix are only needed when libtool
is being used to link *libraries*; just Makefile.pthread suffices
when libtool is being used to link executables. As such, remove
the inclusion of Makefile.libtool, and change the .lo targets back
to regular .o ones, and add back *.o to the list of files to remove
in the 'clean' target (it was needed there even without the
other changes to that Makefile).
Change-Id: Ifbc3eee4ad2dce54df991301bc5edd11eb29a24a
Reviewed-on: http://gerrit.openafs.org/11532
Tested-by: BuildBot <buildbot@rampaginggeek.com>
Reviewed-by: Chas Williams - CONTRACTOR <chas@cmf.nrl.navy.mil>
Reviewed-by: Jeffrey Altman <jaltman@your-file-system.com>
