]> sourceware.org Git - systemtap.git/log
systemtap.git
11 years agoProperly check and use minidebuginfo
Lukas Berk [Fri, 13 Sep 2013 19:27:45 +0000 (15:27 -0400)]
Properly check and use minidebuginfo

dwflpp.cxx: add bool has_gnu_debuginfo() to check if section exists
dwflpp.h: declare bool has_gnu_debuginfo()
tapsets.cxx: check for .gnu_debuginfo as condition to query_module_symtab
     and listing functions in a process
minidebuginfo.stp: correct the binary to probe

11 years agoBZ915180: fix _stp_is_comat_task typo in s390 test case
Frank Ch. Eigler [Wed, 11 Sep 2013 10:12:52 +0000 (06:12 -0400)]
BZ915180: fix _stp_is_comat_task typo in s390 test case

11 years agoFix i[36]86 target matching in systemtap.base/sdt_misc.exp.
David Smith [Tue, 10 Sep 2013 20:16:42 +0000 (15:16 -0500)]
Fix i[36]86 target matching in systemtap.base/sdt_misc.exp.

11 years agoMake user_string_n_quoted.exp more robust.
David Smith [Mon, 9 Sep 2013 20:03:31 +0000 (15:03 -0500)]
Make user_string_n_quoted.exp more robust.

* testsuite/systemtap.string/user_string_n_quoted.exp: Several small
  changes. If we're not doing install testing, mark the test as
  untested. Increase the timeout value, in case we're operating with a
  cold cache. If we timeout, try to kill stap. Properly call "close" and
  "wait" after "spawn" call.
* testsuite/systemtap.string/user_string_n_quoted.stp: Change
  'nd_syscall.write' to 'syscall.write', since the 'syscall' variant has
  more support.

11 years agoFix minidebuginfo.exp testcase
Lukas Berk [Fri, 6 Sep 2013 17:41:39 +0000 (13:41 -0400)]
Fix minidebuginfo.exp testcase

The minidebuginfo testcase was based off of an incomplete documenation.  We need
to properly fix the creation of the executable to test in a standalone instance.

create-minidebug.sh - properly create standalone testcase, this includes
      removing all symbol, relocation and .debug_* sections.
      Tweak naming of files with binary. prefix for easier
      tracking

minidebuginfo.exp - updated file names to target and cleanup

11 years agominidebuginfo: tighten up create-minidebug.sh
Frank Ch. Eigler [Thu, 5 Sep 2013 19:16:41 +0000 (15:16 -0400)]
minidebuginfo: tighten up create-minidebug.sh

... so we remove .debug_* and .sym* stuff, leaving only the
new .gnu_debugdata.

11 years agoAdded s390x kfails when reading certain addresses that should error, but don't.
David Smith [Thu, 5 Sep 2013 16:53:40 +0000 (11:53 -0500)]
Added s390x kfails when reading certain addresses that should error, but don't.

* testsuite/systemtap.base/net-sanity.exp: Add s390x kfails.
* testsuite/systemtap.base/atomic.exp: Ditto.

11 years agoAdded a testcase for BZ1003976.
David Smith [Thu, 5 Sep 2013 16:49:08 +0000 (11:49 -0500)]
Added a testcase for BZ1003976.

* testsuite/systemtap.base/cmdline.exp: New file.
* testsuite/systemtap.base/cmdline.stp: Ditto.

11 years agoBZ1003976: fix function::cmdline_arg processing
Frank Ch. Eigler [Thu, 5 Sep 2013 15:48:39 +0000 (11:48 -0400)]
BZ1003976: fix function::cmdline_arg processing

* tapset/linux/context.stp (cmdline_arg): Simplify, correct.

11 years agoMinor fixes for java/Makefile
Josh Stone [Wed, 4 Sep 2013 22:52:52 +0000 (15:52 -0700)]
Minor fixes for java/Makefile

- Use the full path to refer to HelperSDT.class, so that make can
  actually tell when it needs to be rebuilt.

- libselinux and git_version.h are not used by this component.

11 years agoFixed PR15928 by making memory1.exp read kernel memory instead of user memory.
David Smith [Wed, 4 Sep 2013 18:25:05 +0000 (13:25 -0500)]
Fixed PR15928 by making memory1.exp read kernel memory instead of user memory.

* testsuite/systemtap.printf/memory1.exp: Build a custom kernel module to
  probe.
* testsuite/systemtap.printf/memory1.stp: Instead of probing
  'syscall.open', probe the custom kernel module instead.
* testsuite/systemtap.printf/memory1_module.Makefile: New file.
* testsuite/systemtap.printf/memory1_module.c: Ditto.

11 years agoDo better testing to determine whether the hw_watch-*.stp will work
William Cohen [Tue, 3 Sep 2013 21:26:00 +0000 (17:26 -0400)]
Do better testing to determine whether the hw_watch-*.stp will work

If perf doesn't support the hardware watchpoints, then systemtap hardware
watchpoints are not going to work either.

11 years agoSkip debugtypes.exp if we don't have uprobes.
David Smith [Tue, 3 Sep 2013 19:39:51 +0000 (14:39 -0500)]
Skip debugtypes.exp if we don't have uprobes.

11 years agoRemove PR15913 fix, since the problem isn't just with nd_syscall.execve.
David Smith [Tue, 3 Sep 2013 19:38:29 +0000 (14:38 -0500)]
Remove PR15913 fix, since the problem isn't just with nd_syscall.execve.

* tapset/linux/nd_syscalls.stp: Revert s390x using register r2 for the
  first argument.

11 years agoUpdate s390's syscall.getresgid16 and syscall.getresuid16 probe aliases.
David Smith [Tue, 3 Sep 2013 16:12:55 +0000 (11:12 -0500)]
Update s390's syscall.getresgid16 and syscall.getresuid16 probe aliases.

* tapset/linux/s390/syscalls.stp: Update syscall.getresgid16 and
  syscall.getresuid16 probe aliases to handle upstream renaming of
  arguments.

11 years agoFix PR15913 by handling nd_syscall.execve's 'filename' argument on s390x.
David Smith [Fri, 30 Aug 2013 20:52:52 +0000 (15:52 -0500)]
Fix PR15913 by handling nd_syscall.execve's 'filename' argument on s390x.

* tapset/linux/nd_syscalls.stp: On s390x, use register r2 for the filename
  argument.

11 years agoImprove the 'num_args' subtest of systemtap.context/context.exp.
David Smith [Fri, 30 Aug 2013 18:51:42 +0000 (13:51 -0500)]
Improve the 'num_args' subtest of systemtap.context/context.exp.

* testsuite/systemtap.context/num_args.stp: Improve s390 support by
  looking in r2 for the first argument.
* testsuite/systemtap.context/num_args.tcl: Fail if we don't get the
  proper string function arguments (instead of silently ignoring the
  error).

11 years agoSkip the 'utf' tests if we don't have uprobes.
David Smith [Fri, 30 Aug 2013 16:12:34 +0000 (11:12 -0500)]
Skip the 'utf' tests if we don't have uprobes.

* testsuite/systemtap.base/utf_pretty.exp: Don't run test if we're not
  doing install testing or if we don't have uprobes.
* testsuite/systemtap.base/utf_user.exp: Ditto.
* testsuite/systemtap.base/utf_user_trunc.exp: Ditto.

11 years agoProperly read user memory in target_set.stp.
David Smith [Fri, 30 Aug 2013 16:03:14 +0000 (11:03 -0500)]
Properly read user memory in target_set.stp.

* testsuite/systemtap.base/target_set.stp: Be sure to read user memory
  with user_long().

11 years agoDon't run target_set_thread.exp if we don't have uprobes.
David Smith [Fri, 30 Aug 2013 15:59:16 +0000 (10:59 -0500)]
Don't run target_set_thread.exp if we don't have uprobes.

11 years agoImprove skipping perf probe tests on unsupported systems.
David Smith [Fri, 30 Aug 2013 14:55:52 +0000 (09:55 -0500)]
Improve skipping perf probe tests on unsupported systems.

* testsuite/lib/systemtap.exp (perf_probes_p): New function.
* testsuite/systemtap.stress/conversions.exp: Use new 'perf_probes_p'
  function instead of an explicit test.
* testsuite/systemtap.base/perf.exp: If we don't have perf probes, mark
  test as untested.
* testsuite/systemtap.base/pt_user_mode.exp: Skip test if we're not doing
  an install test. We don't need to check for perf probe support.
* testsuite/systemtap.examples/profiling/perf.tcl: Skip the example if we
  don't have perf probes or uprobes.

11 years agoMark systemtap.base/pthread_stacks.exp untested when necessary.
David Smith [Fri, 30 Aug 2013 14:51:20 +0000 (09:51 -0500)]
Mark systemtap.base/pthread_stacks.exp untested when necessary.

* testsuite/systemtap.base/pthread_stacks.exp: Don't run test if we're not
  doing install testing or if we don't have uprobes.

11 years agoFix minidebuginfo.exp testcase problems.
David Smith [Fri, 30 Aug 2013 13:43:25 +0000 (08:43 -0500)]
Fix minidebuginfo.exp testcase problems.

* testsuite/systemtap.base/minidebuginfo.exp: If the tested system doesn't
  have uprobes, mark as untested.
* testsuite/systemtap.base/create-minidebug.sh: Change "/usr/bin/bash" to
  "/bin/bash", so older systems can run the script.

11 years agoMake systemtap.examples/check.exp look for an example's optional tcl file.
David Smith [Thu, 29 Aug 2013 21:28:27 +0000 (16:28 -0500)]
Make systemtap.examples/check.exp look for an example's optional tcl file.

* testsuite/systemtap.examples/check.exp: If the example has an associated
  tcl file, run it.
* testsuite/systemtap.examples/README: Mention new tcl file that can be
  optionally associated with an example.
* testsuite/systemtap.examples/stapgames/pingpong.meta: Add a
  'test_support' tag.
* testsuite/systemtap.examples/virtualization/kvm_service_time.meta: Make
  the 'test_support' tag more specific.
* testsuite/systemtap.examples/process/auditbt.tcl: Skip the example if we
  don't have uprobes.
* testsuite/systemtap.examples/process/mutex-contention.tcl: Ditto.
* testsuite/systemtap.examples/process/threadstacks.tcl: Ditto.
* testsuite/systemtap.examples/profiling/pf4.tcl: Ditto.
* testsuite/systemtap.examples/process/ltrace.tcl: Skip the example if we
  don't have .plt probes.
* testsuite/systemtap.examples/interrupt/scf.tcl: Skip the example if
  we're on ia64, which has no dwarf unwinding support.
* testsuite/systemtap.examples/io/io_submit.tcl: Ditto.
* testsuite/systemtap.examples/io/mbrwatch.tcl: Ditto.
* testsuite/systemtap.examples/process/chng_cpu.tcl: Ditto.
* testsuite/systemtap.examples/process/sleepingBeauties.tcl: Ditto.
* testsuite/systemtap.examples/profiling/latencytap.tcl: Ditto.
* testsuite/systemtap.examples/process/strace.tcl: Skip the example if
  we're on ia64, which doesn't have nd_syscall support.
* testsuite/systemtap.examples/process/thread-business.tcl: Ditto.

11 years agoPR15906: staprun: suppress loading of stap modules in FIPS mode
Frank Ch. Eigler [Thu, 29 Aug 2013 17:02:07 +0000 (13:02 -0400)]
PR15906: staprun: suppress loading of stap modules in FIPS mode

* staprun/staprun.c (insert_stap_module): Check & fear
  /proc/sys/crypto/fips_enabled.

11 years agoFix failures in uprobes_filtering.exp by improving as_non_root().
David Smith [Wed, 28 Aug 2013 20:39:29 +0000 (15:39 -0500)]
Fix failures in uprobes_filtering.exp by improving as_non_root().

* testsuite/lib/systemtap.exp (as_non_root): If we're using user 'nobody',
  override SYSTEMTAP_DIR to '/tmp/.systemtap-nobody'.

11 years agoman page error::pass2: mention stap-prep
Frank Ch. Eigler [Wed, 28 Aug 2013 19:48:28 +0000 (15:48 -0400)]
man page error::pass2: mention stap-prep

11 years agoImprove skipping .plt tests on unsupported systems.
David Smith [Wed, 28 Aug 2013 17:39:41 +0000 (12:39 -0500)]
Improve skipping .plt tests on unsupported systems.

* testsuite/lib/systemtap.exp (plt_probes_p): New function.
* testsuite/systemtap.base/list.exp: Uses 'plt_probes_p' to know to mark
  .plt probe tests as 'untested'.
* testsuite/systemtap.base/plt.exp: Uses 'plt_probes_p'.
* testsuite/systemtap.pass1-4/semok.exp: Ditto.
* testsuite/systemtap.unprivileged/unprivileged_myproc.exp: Ditto.
* testsuite/systemtap.unprivileged/unprivileged_probes.exp: Ditto.

11 years agoImproved spawned process cleanup in bz6850.exp and fib.exp.
David Smith [Wed, 28 Aug 2013 15:35:03 +0000 (10:35 -0500)]
Improved spawned process cleanup in bz6850.exp and fib.exp.

* testsuite/systemtap.base/bz6850.exp: Improved cleanup logic.
* testsuite/systemtap.context/fib.exp: Ditto.

11 years agoUpdate badname.stp example for newer kernels (with inlined may_create()).
David Smith [Wed, 28 Aug 2013 15:31:57 +0000 (10:31 -0500)]
Update badname.stp example for newer kernels (with inlined may_create()).

* testsuite/systemtap.examples/general/badname.stp: Since may_create() is
  now always inlined (and we can't find its arguments), probe
  may_create()'s callers instead.

11 years agoAdd some dyninst XFAILS in sdt_misc.exp.
David Smith [Tue, 27 Aug 2013 21:39:09 +0000 (16:39 -0500)]
Add some dyninst XFAILS in sdt_misc.exp.

* testsuite/systemtap.base/sdt_misc.exp: XFAIL tests that attach to
  processes when testing dyninst.

11 years agoAdd note to NEWS for PR15805.
David Smith [Tue, 27 Aug 2013 19:08:53 +0000 (14:08 -0500)]
Add note to NEWS for PR15805.

11 years agostaprun requires elfutils 0.142+, so always assume we have elf_getshdrstrndx().
David Smith [Tue, 27 Aug 2013 19:05:15 +0000 (14:05 -0500)]
staprun requires elfutils 0.142+, so always assume we have elf_getshdrstrndx().

* staprun/common.c (usage): Always assume we have elf_getshdrstrndx().
* staprun/staprun.c (init_staprun): Ditto.
* staprun/staprun_funcs.c: Ditto.
* configure.ac: For staprun, the lowest version of elfutils we require is
0.142+, which has elf_getshdrstrndx(). So, if elf_getshdrstrndx()
doesn't exist in libelf, fail.
* Makefile.in: Regenerated.
* aclocal.m4: Ditto.
* config.in: Ditto.
* configure: Ditto.
* doc/Makefile.in: Ditto.
* doc/SystemTap_Tapset_Reference/Makefile.in: Ditto.
* doc/beginners/Makefile.in: Ditto.
* java/Makefile.in: Ditto.
* man/Makefile.in: Ditto.
* stapdyn/Makefile.in: Ditto.
* staprun/Makefile.in: Ditto.
* NEWS: Mention staprun requiring elfuilts 0.142+.

11 years agoPR15900 weaken & simplify server_locale.exp test case
Frank Ch. Eigler [Tue, 27 Aug 2013 18:22:37 +0000 (14:22 -0400)]
PR15900 weaken & simplify server_locale.exp test case

The server_locale.exp test case need only confirm that selected
local-related environment variables are passed to the server.  A
smaller & simpler set of tests are sufficient for that.

This also reverts commit dcd8a2ebbcc00ab6aed227ebd721346d81c26a7b.

11 years agoPR15900 fix syntax errors in server_local.exp test case
Martin Cermak [Tue, 27 Aug 2013 15:15:32 +0000 (11:15 -0400)]
PR15900 fix syntax errors in server_local.exp test case

11 years agoBZ998595, systemtap.spec fix disagreement between %file attrs and explicit chmods
Frank Ch. Eigler [Tue, 27 Aug 2013 15:08:35 +0000 (11:08 -0400)]
BZ998595, systemtap.spec fix disagreement between %file attrs and explicit chmods

11 years agoMake pr13158.exp test results less random.
David Smith [Tue, 27 Aug 2013 14:58:22 +0000 (09:58 -0500)]
Make pr13158.exp test results less random.

* testsuite/systemtap.base/pr13158.exp: Make test result less random, by
  making sure the first copy of the module is loaded before trying to load
  the second copy.

11 years agoMake the 'systemtap-testsuite' rpm require 'nc' for ipaddr.exp.
David Smith [Mon, 26 Aug 2013 21:43:40 +0000 (16:43 -0500)]
Make the 'systemtap-testsuite' rpm require 'nc' for ipaddr.exp.

11 years agoMove %changelog pruning into diff check block
Lukas Berk [Fri, 23 Aug 2013 19:05:30 +0000 (15:05 -0400)]
Move %changelog pruning into diff check block

If there is no spec diff, then there's no need to even check for a
difference in the %changelog section.  Furthermore we should check
that there is even a %changelog section in the diff and only attempt
to prune accordingly

scripts/spin-rawhide - move pruning into file check statements

11 years agoTweak spin-rawhide script to stop overwriting the changelog
Lukas Berk [Fri, 23 Aug 2013 18:26:54 +0000 (14:26 -0400)]
Tweak spin-rawhide script to stop overwriting the changelog

scripts/spin-rawhide - when merging an upstream diff, sometimes the
       changelog can be overwritten, which shouldn't happen
       trim the changelog diffs and their context lines

11 years agoTestcase and documentation for minidebuginfo
Lukas Berk [Fri, 23 Aug 2013 15:31:48 +0000 (11:31 -0400)]
Testcase and documentation for minidebuginfo

man/warning::debuginfo.7stap - explain basics of minidebuginfo, note
       that it requires elfutils 0.156 and add
       a pointer to the fedora feature page
testsuite/systemtap.base/ files:
create-minidebuginfo.sh - script to create minidebuginfo
minidebug.c - simple c program with multiple function calls
minidebuginfo.exp  - compile minidebug.c and generate minidebuginfo,
  test that the functions are still listed with
  probefunc()
minidebuginfo.stp - script to probe process.function("*") and list
  probefunc()'s

11 years agoSimplify the testsuite tests for uprobes and utrace
Josh Stone [Thu, 22 Aug 2013 22:39:26 +0000 (15:39 -0700)]
Simplify the testsuite tests for uprobes and utrace

- Searching /proc/kallsyms is now standardized to grep_kallsyms.
- utrace and uprobes tests are broken into separate procs for the
  different variants, then a unified proc combines variants.

11 years agoFix PR15625 by making unprivileged_embedded_C.exp ignore stub functions.
David Smith [Thu, 22 Aug 2013 21:09:47 +0000 (16:09 -0500)]
Fix PR15625 by making unprivileged_embedded_C.exp ignore stub functions.

* testsuite/systemtap.unprivileged/unprivileged_embedded_C.exp: Ignore
  stub functions in registers.stp.

11 years agoFix PR15881 by fixing unprivileged_myproc.exp failures.
David Smith [Thu, 22 Aug 2013 20:31:04 +0000 (15:31 -0500)]
Fix PR15881 by fixing unprivileged_myproc.exp failures.

* testsuite/systemtap.unprivileged/unprivileged_myproc.exp: KFAIL
  '.absolute' and '.absolute.return' probe tests when using inode uprobes,
  since inode uprobes don't support those probe types. For
  'library.plt.statement(NUMBER)' probe tests, use the address of the
  library function, instead of the address of the main program
  function. With the last problem fixed, remove the XFAILS for those probe
  tests (an update to PR13373).

11 years agoAvoid problems for unprivileged users when getting MAXSTRINGLEN value.
David Smith [Thu, 22 Aug 2013 15:34:51 +0000 (10:34 -0500)]
Avoid problems for unprivileged users when getting MAXSTRINGLEN value.

* tapset/macros.stpm: New '@MAXSTRINGLEN' macro.
* tapset/linux/context-caller.stp: To avoid problems for unprivileged
  users, use new '@MAXSTRINGLEN' macro instead of '%{ MAXSTRINGLEN %}'.
* tapset/linux/context-symbols.stp: Ditto.
* tapset/linux/ucontext-symbols.stp: Ditto.
* tapset/linux/ucontext.stp: Ditto.

11 years agoActually add the i386 portions of the spec file fix
Lukas Berk [Thu, 22 Aug 2013 03:37:38 +0000 (23:37 -0400)]
Actually add the i386 portions of the spec file fix

systemtap.spec - add the i386 portions of the arch symlink fix.

11 years agoFix BZ999585 where spec file symlinks don't match java directories
Lukas Berk [Thu, 22 Aug 2013 03:31:20 +0000 (23:31 -0400)]
Fix BZ999585 where spec file symlinks don't match java directories

systemtap.spec - add clauses to catch cases (i386, ppc64) where the
         java directory we want to symlink to doesn't actually
 match the basename type

11 years agoFix PR15876 by making the ptridx.exp testcase more robust.
David Smith [Wed, 21 Aug 2013 21:16:53 +0000 (16:16 -0500)]
Fix PR15876 by making the ptridx.exp testcase more robust.

* testsuite/systemtap.base/ptridx.exp: Set up the timeout value to be 60
  seconds and kills stap if we do have a timeout.

11 years agoPR15074 update: Relaxed pthread_stacks.exp debuginfo check.
David Smith [Wed, 21 Aug 2013 19:55:33 +0000 (14:55 -0500)]
PR15074 update: Relaxed pthread_stacks.exp debuginfo check.

* testsuite/systemtap.base/pthread_stacks.exp: Relaxed debuginfo check a
  bit, since we only need "native" glibc debuginfo installed on multilib
  systems like x86_64.

11 years agoFix PR15875 by making sure 'strace' is on a system before running ptrace.exp.
David Smith [Wed, 21 Aug 2013 18:30:18 +0000 (13:30 -0500)]
Fix PR15875 by making sure 'strace' is on a system before running ptrace.exp.

* testsuite/systemtap.base/ptrace.exp: If 'strace' doesn't exist on the
  system, quit early.
* systemtap.spec: Make 'systemtap-testsuite' require 'strace', since
  ptrace.exp needs it.

11 years agoFix PR15874 by making systemtap.base/onoffprobe.exp more robust.
David Smith [Wed, 21 Aug 2013 17:42:27 +0000 (12:42 -0500)]
Fix PR15874 by making systemtap.base/onoffprobe.exp more robust.

* testsuite/systemtap.base/onoffprobe.exp: Make more robust.
* testsuite/systemtap.base/onoffprobe.stp: Call exit() when finished.

11 years agoFew typos in langref and tutorial
Miroslav Franc [Wed, 21 Aug 2013 11:41:25 +0000 (13:41 +0200)]
Few typos in langref and tutorial

11 years agoFix up 3 XPASS results in the semko.exp testcase.
David Smith [Tue, 20 Aug 2013 21:49:20 +0000 (16:49 -0500)]
Fix up 3 XPASS results in the semko.exp testcase.

* testsuite/systemtap.pass1-4/semko.exp: Mark the utrace.stp test as
  "untested" when using dyninst, since dyninst contains utrace-like
  functionality.
* testsuite/semko/badregex01.stp: Actually add an invalid regexp to the
  testcase.

11 years agoFix PR15866 by updating the signal.pending/signal.pending.return probe aliases.
David Smith [Tue, 20 Aug 2013 19:59:20 +0000 (14:59 -0500)]
Fix PR15866 by updating the signal.pending/signal.pending.return probe aliases.

* tapset/linux/signal.stp: Make the 'signal.pending' and
  'signal.pending.return' probe aliases use 'sys_rt_sigpending' and
  'compat_sys_rt_sigpending' if 'do_sigpending' is an inlined function.

11 years agoFixed PR15865 by updating nfs_proc.stp to handle removal of NFSv2.
David Smith [Tue, 20 Aug 2013 18:22:32 +0000 (13:22 -0500)]
Fixed PR15865 by updating nfs_proc.stp to handle removal of NFSv2.

* tapset/linux/nfs_proc.stp: Handle removal of NFSv2 functions from the
  kernel. Aliased all NFSv2 functions to 'never' if not present.

11 years agostrace example: add a note re. PR6762
Frank Ch. Eigler [Tue, 20 Aug 2013 10:11:30 +0000 (06:11 -0400)]
strace example: add a note re. PR6762

The reminder was motivated by coming across Jan Kratoch's much more
elaborate strace-in-systemtap prototype, which solves this problem:
http://git.jankratochvil.net/?p=staptrace.git

11 years agoobsolete regex tests: drop re2c-migrate and old systemtap.base/regex.*
Frank Ch. Eigler [Tue, 20 Aug 2013 01:48:47 +0000 (21:48 -0400)]
obsolete regex tests: drop re2c-migrate and old systemtap.base/regex.*

11 years agoregex tests: add as pure systemtap script
Frank Ch. Eigler [Tue, 20 Aug 2013 01:42:12 +0000 (21:42 -0400)]
regex tests: add as pure systemtap script

While at it, introduce simple runok.exp test family.

11 years agonew sample script: strace.stp; you'll never guess what it does
Frank Ch. Eigler [Mon, 19 Aug 2013 23:55:25 +0000 (19:55 -0400)]
new sample script: strace.stp; you'll never guess what it does

11 years agoUpdate systemtap.examples/process/threadstacks.stp for rawhide.
David Smith [Mon, 19 Aug 2013 21:07:14 +0000 (16:07 -0500)]
Update systemtap.examples/process/threadstacks.stp for rawhide.

* testsuite/systemtap.examples/process/threadstacks.stp: Updated for glibc
  2.17 changes to how pthread stacks are allocated.

11 years agoFix PR15805 by allocating map memory with vmalloc() instead of kmalloc().
David Smith [Fri, 16 Aug 2013 16:00:35 +0000 (11:00 -0500)]
Fix PR15805 by allocating map memory with vmalloc() instead of kmalloc().

* runtime/linux/map_runtime.h (_stp_map_del): Call _stp_vfree() to free
  all the node memory.
  (_stp_map_vzalloc): Renamed from _stp_map_kzalloc() and reworked to use
  vmalloc-based allocations instead of kmalloc-based allocations.
  (_stp_map_init): Allocate all the node memory in one big chunk using
  vmalloc(). Change all other allocations to use vmalloc().
* runtime/map.h: Add 'node_mem' variable to 'map_root' structure
  definition.
* runtime/linux/alloc.c (_stp_vzalloc): New function.
  (_stp_vzalloc_node): Ditto.

11 years agobelated spring cleaning: ditch spurious TODOXXX
Serguei Makarov [Wed, 7 Aug 2013 15:35:45 +0000 (11:35 -0400)]
belated spring cleaning: ditch spurious TODOXXX

11 years agobelated spring cleaning: imitate semantic_error chaining for macro feature
Serguei Makarov [Wed, 7 Aug 2013 15:28:27 +0000 (11:28 -0400)]
belated spring cleaning: imitate semantic_error chaining for macro feature

11 years agobelated spring cleaning: clarify error behavior for macro feature
Serguei Makarov [Wed, 7 Aug 2013 14:58:05 +0000 (10:58 -0400)]
belated spring cleaning: clarify error behavior for macro feature

11 years agobelated spring cleaning: missing i18n for macro feature
Serguei Makarov [Wed, 7 Aug 2013 14:30:08 +0000 (10:30 -0400)]
belated spring cleaning: missing i18n for macro feature

11 years agoMinor update for publican bug workaound in systemtap.spec.
David Smith [Mon, 5 Aug 2013 16:45:03 +0000 (11:45 -0500)]
Minor update for publican bug workaound in systemtap.spec.

* systemtap.spec: Move the BZ920216 workaround into the 'with_publican'
  section, since that is what it is related to.

11 years agoPR15757 Fix truncated 32bit backtraces on x86_64
Lukas Berk [Fri, 2 Aug 2013 22:13:34 +0000 (18:13 -0400)]
PR15757 Fix truncated 32bit backtraces on x86_64

runtime/unwind.c: Fix the variable size to be incremented by in the
  eh_frame and table_enc.

11 years agoFor xen kernels, handle cpufreq_register_notifier() failure.
Timo Juhani Lindfors [Fri, 2 Aug 2013 18:05:26 +0000 (13:05 -0500)]
For xen kernels, handle cpufreq_register_notifier() failure.

* runtime/time.c (_stp_kill_time): If cpufreq_register_notifier()
  suceeded, unregister the notifier.
  (_stp_init_time): On xen kernels, cpufreq_register_notifier() can
  fail. Just ignore the error.

11 years agoMake stap-stop-server wait (for up to 10 seconds) for the server to quit.
David Smith [Fri, 2 Aug 2013 15:35:25 +0000 (10:35 -0500)]
Make stap-stop-server wait (for up to 10 seconds) for the server to quit.

11 years agoFix reported tcl testsuite errors.
David Smith [Thu, 1 Aug 2013 14:10:27 +0000 (09:10 -0500)]
Fix reported tcl testsuite errors.

* testsuite/systemtap.base/implicitptr.exp: Fix misnamed variable.
* testsuite/systemtap.base/sdt_misc.exp: Change "/usr/bin/kill" to
  "/bin/kill" (for Debian systems).

11 years agobelated NEWS entry re. user_string_quoted() and friends
Frank Ch. Eigler [Tue, 30 Jul 2013 20:38:03 +0000 (16:38 -0400)]
belated NEWS entry re. user_string_quoted() and friends

11 years agoMade systemtap.server/server.exp handle inode uprobes properly.
David Smith [Tue, 30 Jul 2013 19:35:33 +0000 (14:35 -0500)]
Made systemtap.server/server.exp handle inode uprobes properly.

* testsuite/systemtap.server/server.exp: Handle systems with inode uprobes
  properly: don't try to use or look for uprobes.ko.

11 years agoMade systemtap.server/server_locale.exp more robust.
David Smith [Tue, 30 Jul 2013 18:41:49 +0000 (13:41 -0500)]
Made systemtap.server/server_locale.exp more robust.

* testsuite/systemtap.server/server_locale.exp: Updated tests for new
  warning/error messages. For French tests, made them more permissive - if
  we get anything in French, the test passes. This was done because we may
  not have a translation of a particular message.

11 years agoPR15803: netfilter: tolerate absent context when collecting $verdict
Frank Ch. Eigler [Tue, 30 Jul 2013 12:26:59 +0000 (08:26 -0400)]
PR15803: netfilter: tolerate absent context when collecting $verdict

It is possible to skip from the prologue epilogue through to the
prologue for reasons such as shortage of available stack space.
In these cases, the context* c; pointer could remain NULL, which
is considered generally inappropriate to dereference.

* tapset-netfilter.cxx (epilogue): Check for NULL c before extracting
  $verdict.

11 years agoMade systemtap.server/client.exp more robust.
David Smith [Mon, 29 Jul 2013 21:59:31 +0000 (16:59 -0500)]
Made systemtap.server/client.exp more robust.

* testsuite/systemtap.server/client.exp: When this test is run right after
  another server test, sometimes the server from the previous test hasn't
  quite died. So, when getting the initial list of online servers, make
  sure we get the same result twice. Simplify code by using new
  'array_compare' procedure.

11 years agoPR15793: numa addr_to_node(): protect __pa() call with virt_addr_valid()
Frank Ch. Eigler [Mon, 29 Jul 2013 18:16:01 +0000 (14:16 -0400)]
PR15793: numa addr_to_node(): protect __pa() call with virt_addr_valid()

Update test case to tolerate exception that may now be thrown by the
tapset function.

11 years agosession.cxx: make string params const
Jonathan Lebon [Fri, 26 Jul 2013 19:31:49 +0000 (15:31 -0400)]
session.cxx: make string params const

And change colorize() to take in a token pointer rather than a ref. This
is more in line with other funcs that take in a token and also ensures
that we don't have undefined behaviour if we're called with a
dereferenced NULL pointer (which coverity complained about).

11 years agoelaborate.cxx -- placate coverity (CTOR_DTOR_LEAK)
Serguei Makarov [Fri, 26 Jul 2013 19:39:14 +0000 (15:39 -0400)]
elaborate.cxx -- placate coverity (CTOR_DTOR_LEAK)

11 years agostapregex -- placate the gods of coverity (RESOURCE_LEAK)
Serguei Makarov [Fri, 26 Jul 2013 19:37:33 +0000 (15:37 -0400)]
stapregex -- placate the gods of coverity (RESOURCE_LEAK)

11 years agostapregex -- placate the gods of coverity (INVALIDATE_ITERATOR)
Serguei Makarov [Fri, 26 Jul 2013 19:36:19 +0000 (15:36 -0400)]
stapregex -- placate the gods of coverity (INVALIDATE_ITERATOR)

11 years agostapregex -- placate the gods of coverity (UNINIT_CTOR)
Serguei Makarov [Fri, 26 Jul 2013 19:35:47 +0000 (15:35 -0400)]
stapregex -- placate the gods of coverity (UNINIT_CTOR)

11 years agosession.cxx: add assert statement for benchmark opts
Jonathan Lebon [Fri, 26 Jul 2013 17:58:21 +0000 (13:58 -0400)]
session.cxx: add assert statement for benchmark opts

11 years agostaprun/common.c: add assert statement for -C, -S, -r
Jonathan Lebon [Fri, 26 Jul 2013 17:22:34 +0000 (13:22 -0400)]
staprun/common.c: add assert statement for -C, -S, -r

11 years agosystemtap.spec: prefer with_systemd on recent OS
Frank Ch. Eigler [Fri, 26 Jul 2013 14:35:03 +0000 (10:35 -0400)]
systemtap.spec: prefer with_systemd on recent OS

11 years agoversion bump for 2.4 development
David Smith [Fri, 26 Jul 2013 02:03:44 +0000 (21:03 -0500)]
version bump for 2.4 development

11 years agoRevert "testsuite: set -B CONFIG_DEBUG_INFO=y in testsuite rc to help debug crashes" release-2.3
Frank Ch. Eigler [Thu, 25 Jul 2013 20:21:14 +0000 (16:21 -0400)]
Revert "testsuite: set -B CONFIG_DEBUG_INFO=y in testsuite rc to help debug crashes"

This reverts commit 6bb6bf29ae.....
                and f1d4f4313.....

11 years agoRevert "PR15782 cont'd: PR_Read_Complete should initialize vars more"
Frank Ch. Eigler [Thu, 25 Jul 2013 20:03:25 +0000 (16:03 -0400)]
Revert "PR15782 cont'd: PR_Read_Complete should initialize vars more"

This reverts commit 59331ea0de1f10d0694b3389e0b4710db4da0798.

11 years agoPR15782 cont'd: PR_Read_Complete should initialize vars more
Frank Ch. Eigler [Thu, 25 Jul 2013 20:01:31 +0000 (16:01 -0400)]
PR15782 cont'd: PR_Read_Complete should initialize vars more

Use a clearer looping/accumulation structure.

11 years agoImplement PR_Read_Complete() in nsscommon.cxx.
Dave Brolley [Thu, 25 Jul 2013 19:20:50 +0000 (15:20 -0400)]
Implement PR_Read_Complete() in nsscommon.cxx.

A wrapper for PR_Read which ensures that all of the requested data
is read.
Updates callers of PR_Read() as appropriate.

11 years agoregex: warning cleanup as per gcc 3.4.6 on i686
Frank Ch. Eigler [Thu, 25 Jul 2013 18:30:02 +0000 (14:30 -0400)]
regex: warning cleanup as per gcc 3.4.6 on i686

11 years agostapregex rewrite: appropriate NEWS items
Serguei Makarov [Wed, 24 Jul 2013 21:30:30 +0000 (17:30 -0400)]
stapregex rewrite: appropriate NEWS items

11 years agostapregex rewrite: regenerate autoconf
Serguei Makarov [Wed, 24 Jul 2013 21:19:00 +0000 (17:19 -0400)]
stapregex rewrite: regenerate autoconf

11 years agostapregex rewrite -- prepare for rebase onto main branch for 2.3 release
Serguei Makarov [Wed, 24 Jul 2013 20:50:03 +0000 (16:50 -0400)]
stapregex rewrite -- prepare for rebase onto main branch for 2.3 release

11 years agostapregex REWRITE -- fixes for 100% (existing, non-subexpression) testsuite pass
Serguei Makarov [Wed, 24 Jul 2013 18:52:20 +0000 (14:52 -0400)]
stapregex REWRITE -- fixes for 100% (existing, non-subexpression) testsuite pass

11 years agostapregex REWRITE -- further bug fixes for engine (2 testsuite errors remaining)
Serguei Makarov [Wed, 24 Jul 2013 16:06:06 +0000 (12:06 -0400)]
stapregex REWRITE -- further bug fixes for engine (2 testsuite errors remaining)

11 years agostapregex REWRITE -- further bug fixes (esp. dealing with infinite loops)
Serguei Makarov [Tue, 23 Jul 2013 21:01:50 +0000 (17:01 -0400)]
stapregex REWRITE -- further bug fixes (esp. dealing with infinite loops)

11 years agostapregex REWRITE -- rudimentary emit implementation (no tags yet)
Serguei Makarov [Tue, 23 Jul 2013 16:01:25 +0000 (12:01 -0400)]
stapregex REWRITE -- rudimentary emit implementation (no tags yet)

11 years agostapregex REWRITE -- bug fix progress (fix deallocation with ^, fix duplicate-state...
Serguei Makarov [Mon, 22 Jul 2013 20:41:02 +0000 (16:41 -0400)]
stapregex REWRITE -- bug fix progress (fix deallocation with ^, fix duplicate-state-check, ditch debug output)

11 years agostapregex REWRITE -- bug fix progress (failure alternative, ^ anchoring)
Serguei Makarov [Mon, 22 Jul 2013 19:55:15 +0000 (15:55 -0400)]
stapregex REWRITE -- bug fix progress (failure alternative, ^ anchoring)

11 years agostapregex REWRITE -- bug fix progress (range creation off-by-one error)
Serguei Makarov [Mon, 22 Jul 2013 19:02:50 +0000 (15:02 -0400)]
stapregex REWRITE -- bug fix progress (range creation off-by-one error)

This page took 0.061094 seconds and 5 git commands to generate.