stap samples: tweak process keyword & nd_syscall usage
* process/thread-business.stp: use 'name' from nd-syscall tapset
instead of clumsy computation by hand
* process/syscalls_by_*: use nd_syscall.* instead of syscall.*
David Smith [Thu, 26 Jun 2014 16:32:07 +0000 (11:32 -0500)]
PR16716 partial fix: Fix types in 'syscall.{fdatasync,fstat,fstatat,fsync}'.
* tapset/linux/aux_syscalls.stp (_at_flag_str): Update and improve.
* tapset/linux/syscalls.stp: Fix types in
syscall.{fdatasync,fstat,fsync}. Fixed types and 32-on-64 support in
syscall.fstatat and added 'dirfd_str', 'flags', and 'flags_str'
convenience variables.
* tapset/linux/nd_syscalls.stp: Fixed 32-on-64 bit support in
nd_syscall.fstatat and added 'dirfd_str', 'flags', and 'flags_str'
convenience variables.
* testsuite/buildok/syscalls-detailed.stp: Added compile test for new
syscall.fstatat convenience variables.
* testsuite/buildok/nd_syscalls-detailed.stp: Ditto.
* testsuite/systemtap.syscall/futimes.c: Updated test.
* testsuite/systemtap.syscall/link.c: Ditto.
* testsuite/systemtap.syscall/stat.c: Ditto.
* testsuite/systemtap.syscall/sync.c: Ditto.
Abegail Jakop [Tue, 24 Jun 2014 15:21:33 +0000 (11:21 -0400)]
stmt_counting.exp: nested foreach runtime flag
stmt_counting.exp: combine the four cases for each test
(no flags, -u, low MAXACTION, high MAXACTION) by using
nested foreach loop for the different runtime flags.
Abegail Jakop [Tue, 24 Jun 2014 13:29:46 +0000 (09:29 -0400)]
stmt_counting*: statement counting tests
stmt_counting_*.stp: basic tests containing loops, recursion,
straight through function bodies and straight through probe bodies.
stmt_counting.exp: runs the tests with and without -u, and with a
low and a high MAXACTION.
Mark Wielaard [Sun, 22 Jun 2014 19:23:21 +0000 (21:23 +0200)]
Handle DW_TAG_restrict_type.
C99 keyword restrict is just like const or volatile for pointers. Like
DW_TAG_const_type and DW_TAG_volatile_type just follow through to the
underlying type.
Josh Stone [Fri, 20 Jun 2014 16:45:46 +0000 (09:45 -0700)]
runtime: Add a universal stp_synchronize_sched
The kernel has had synchronize_sched for a while. However, in the RHEL5
era this was merely a #define to synchronize_rcu, which our exportconf
wouldn't detect. RHEL4 only had synchronize_kernel.
Now stp_synchronize_sched covers all of these options, otherwise issuing
an #error message.
Abegail Jakop [Thu, 19 Jun 2014 20:32:14 +0000 (16:32 -0400)]
::functioncall_visitor tracks seen and nested call
this change is for functioncall_traversing_visitor
elaborate.cxx: changed instances of *traversed* to *seen*
staptree.cxx: changed ::functioncall_visitor in
functioncall_traversing_visitor to keep track of nested function calls
and kept the old functionality of keeping track of function calls seen.
staptree.h: remove traversed set from functioncall_traversing_visitor,
replacing it with a seen set, and a nested set.
translate.cxx: edited comment to reflect change in visitor
David Smith [Thu, 19 Jun 2014 17:33:52 +0000 (12:33 -0500)]
Minor testsuite updates.
* testsuite/lib/compile_flags.exp: Added a bit more aarch64 support.
* testsuite/systemtap.base/bz1027459.exp: Only run on make
installcheck. Clean up better after a 'spawn'.
* testsuite/systemtap.base/task_dentry_path.exp: Only run on make
installcheck.
* testsuite/systemtap.base/wordexp.exp: Ditto.
* testsuite/systemtap.server/server_locale.exp: Ditto.
Abegail Jakop [Wed, 18 Jun 2014 14:15:05 +0000 (10:15 -0400)]
translate.cxx: statement counting for functions
translate.cxx: added counting statements to be executed, and a check
against that number with MAXACTION to determine if statement counting
within a function's generated c code gets suppressed. allow users to
turn this off by running script in unoptimized mode.
Abegail Jakop [Tue, 17 Jun 2014 20:16:28 +0000 (16:16 -0400)]
translate.cxx: unoptimized statement counting
translate.cxx: allow for users to turn off the optimized statement
counting, reverting to the statement counting being in generated
c code, through unoptimize mode.
Abegail Jakop [Tue, 17 Jun 2014 14:41:59 +0000 (10:41 -0400)]
translate.cxx: suppress action counting in c code
translate.cxx: moved the statement counting portion to ::emit_probe().
added in a check against MAXACTION, which determines if statement
counting within the generated c code gets suppressed.
Abegail Jakop [Fri, 13 Jun 2014 18:35:07 +0000 (14:35 -0400)]
translate.cxx: fix counting in max_action_info
translate.cxx: removed the creation of temporary visitors to visit
the thenblock and elseblock in visit_if_statement in max_action_info
because of errors with recursive calls. now uses only one visitor
to visit blocks.
Josh Stone [Fri, 13 Jun 2014 23:12:34 +0000 (16:12 -0700)]
runtime: Make sure to free special message buffers too
These special buffers were formerly kept aside, and never released back
to the memory pool, so they weren't freed either. DEBUG_MEM did report
this leak properly. Now these are freed too, and a new debug_mem.exp
test ensures that a simple script doesn't leak any memory.
Josh Stone [Fri, 13 Jun 2014 21:47:54 +0000 (14:47 -0700)]
PR17052: Use RCU-sched for freeing contexts
A context is grabbed from the _stp_ctl_work_callback timer, but it
doesn't actually need it for anything other than preventing a lock
recursion. Since this timer is still running while we are freeing the
contexts, there's a chance to race on the existence of contexts.
Using rcu_assign_pointer, rcu_dereference_sched, and synchronize_sched,
we can make sure that anyone who might have seen the existing contexts
has finished before we free them.
Jonathan Lebon [Fri, 13 Jun 2014 17:24:36 +0000 (13:24 -0400)]
initscript: check script names
Because script names are the basis for some variable names (e.g.
script_name_REQ), they need to adhere to bash variable naming rules.
These are almost the same rules that we enforce for stap -m, so we just
fail 'earlier'. See also RHBZ1109084.
This patch ensures that scripts with invalid names are skipped and
logged.
Jonathan Lebon [Fri, 13 Jun 2014 14:45:03 +0000 (10:45 -0400)]
initscript: use SYSTEMCTL_SKIP_REDIRECT
The SYSTEMCTL_SKIP_REDIRECT variable allows us to avoid being 'handled'
by systemd when users use the legacy 'service' command. This is
necessary in our case because systemctl omits the parameters passed to
the 'service' command for standard service actions such as start/stop.
This commit was initially required because execve would not pass on
RHEL6, which relied on kernel_string() rather than user_string_quoted().
With the new kernel_string_quoted() function, this test should pass on
all platforms.
Jonathan Lebon [Thu, 12 Jun 2014 17:34:06 +0000 (13:34 -0400)]
conversions.stp: new kernel_string_quoted() function
Just like user_string_quoted(), the new kernel_string_quoted() function
returns the resulting string in quotes if it successfully retrieved the
string and the pointer address otherwise.
Abegail Jakop [Thu, 12 Jun 2014 14:13:18 +0000 (10:13 -0400)]
count max statements in probes in translate.cxx
translate.cxx: counts the maximum number of statements expected
to be executed for each probe, in pass 3. results are printed when
verbosity is > 1. probes with for/foreach statements have a
reported statement count of ~0.
Jonathan Lebon [Wed, 11 Jun 2014 18:06:07 +0000 (14:06 -0400)]
listing_mode.exp: use stap itself to get addr of vfs_read
On older platforms, the addresses obtained from /proc/kallsyms show up
as 0 for non-root users. So instead, we use stap itself to retrieve the
address of vfs_read which is later used for testing.
Jonathan Lebon [Wed, 11 Jun 2014 17:34:36 +0000 (13:34 -0400)]
lib/systemtap.exp: add callee_probes_p proc
Rather than checking for the conditions for callee support in various
testcases, add a callee_probes_p proc in lib/systemtap.exp and use that
instead where needed.
Jonathan Lebon [Wed, 11 Jun 2014 17:28:51 +0000 (13:28 -0400)]
lib/systemtap.exp: rename hw_breakpoint_p to hwbkpt_probes_p
To be more consistent with the other procs in lib/systemtap.exp which
check whether certain types of probes are supported, rename
hw_breakpoint_p to hwbkpt_probes_p.
Jonathan Lebon [Wed, 11 Jun 2014 15:15:14 +0000 (11:15 -0400)]
vars.exp: only parse out from first line of grep
There can be multiple _stp_snprintf lines, in which case, the resulting
values for vars and parms can hold the same variables multiple times,
thus throwing off the test (see RHBZ1099757#c4).
Fix this by only using the first line output from grep.
Using a process.function(0xabcd) may still cause prologue searching to
occur on older platforms (for GCC < 4.5) due to possibly bad debuginfo
in the prologue (see e.g. commit 277a2b9 and GDB's PR13777).
Thus, the resulting probe from a process.function(0xabcd) might not
match the exact address given if it has been adjusted by
prologue-searching. So we relax the requirement by simply checking that
the address printed in a proper hex format.
Jonathan Lebon [Tue, 10 Jun 2014 21:38:05 +0000 (17:38 -0400)]
listing_mode.exp: don't test HW bkpt probes if unsupported
We check if HW breakpoint probes are available on the machine during
testsuite setup and use the new hw_breakpoint_p proc in listing_mode.exp
to check if these probes should be tested.
Jonathan Lebon [Wed, 4 Jun 2014 21:46:03 +0000 (17:46 -0400)]
statement.exp: remove mentions of RANGE
Since commit f01d4ff, we now no longer have a RANGE lineno_type. So
let's remove mentions of it to avoid future confusion and replace it by
the new ENUMERATED lineno type. Also clear up subtest names.
Jonathan Lebon [Tue, 10 Jun 2014 21:36:39 +0000 (17:36 -0400)]
stmt_rel.exp: relax requirement for PASS
We previously required that all linenos accessible through wildcards
were the same set as all linenos individually accessible through
relative numbering.
The issue is that two linenos may correspond to the same address (e.g.
due to optimization), in which case a wildcard probe would only pick up
the first lineno (since the second lineno would be eliminated due to
being the same address), which two individual relative probes for these
two linenos would both resolve.
So instead, this test now simply checks that all the linenos which were
found in the wildcard probe were also accessible through relative lineno
(i.e. a subtest test instead of equality test).
David Smith [Tue, 10 Jun 2014 19:58:28 +0000 (14:58 -0500)]
PR16716 partial fix: Fix types in syscall.{fchdir,mkdir,mkdirat}.
* tapset/linux/syscalls.stp (syscall.fchdir): Fixed types.
(syscall.mkdir): Ditto.
(syscall.mkdirat): Fixed types and added 'dirfd_str' convenience
variable, to match syscall.{faccessat,fchmodat,fchownat,etc.}.
* tapset/linux/nd_syscalls.stp: Similar changes.
* testsuite/systemtap.syscall/dir.c: Added more tests.
* testsuite/buildok/syscalls-detailed.stp: Added test for 'dirfd_str'
variable of syscall.mkdirat.
* testsuite/buildok/nd_syscalls-detailed.stp: Ditto.
Jonathan Lebon [Tue, 10 Jun 2014 17:41:28 +0000 (13:41 -0400)]
loc2c: make u_fetch_register() respect CONFIG_COMPAT
If CONFIG_COMPAT is defined, then it is possible to be running 32/31-bit
tasks on a 64-bit kernel. In such cases, we need to ensure that the
values returned from u_fetch_register() are 32-bit friendly.
Thus, we mask whatever pt_regs_fetch_register() returns.
David Smith [Mon, 9 Jun 2014 14:11:13 +0000 (09:11 -0500)]
Fix BZ1099555 by reworking gtod.exp.
* testsuite/systemtap.base/gtod.exp: Rework so that we don't care about
what order the output comes. All we care about now is the variance
between the kernel's time and systemtap's time.
* testsuite/systemtap.base/gtod.c (main): Print the iteration number first.
* testsuite/systemtap.base/gtod.stp: Ditto.
stap -V version string: include build-time elfutils version
We have some translator facilities that are based on _ELFUTILS_PREREQ,
a compile-time elfutils version check. Others might change behavior
based on the run-time elfutils version. These versions may differ,
as the system elfutils libraries are upgraded without stap being
rebuilt. This patch changes stap -V to report both run-time and
build-time elfutils versions (if they are different).
Jonathan Lebon [Thu, 5 Jun 2014 15:09:11 +0000 (11:09 -0400)]
PR17027: adapt pr16806.exp for RHEL6
This test failed on RHEL6 for two reasons:
- The 'timeout' program doesn't support non-integer arguments
- We assumed the uprobes module was already inserted.
Fix the first by changing the timeout to 1s (I verified that this
doesn't change the lethality of the test under the right conditions).
Fix the second by doing a sample run first to allow stap the chance to
self-compile and pass the uprobes module to staprun -u.
Mark Wielaard [Thu, 5 Jun 2014 09:13:31 +0000 (11:13 +0200)]
systemtap_session::modules_must_be_signed might use uninitialized memory.
statm >> status might not actually assign anything to status.
operator >> on ifstream doesn't assign anything the underlying file
doesn't exist. So status will just have random contents that happened
to be on the stack. Make sure to initialize status first.
Mark Wielaard [Thu, 5 Jun 2014 08:12:57 +0000 (10:12 +0200)]
Extra details to explain the error or provide alternatives to the user.
Each one printed after the main error message and tokens aligned on
separate lines. Just push_back anything you want that better explains
the error to the user (not meant for extra verbose developer messages).
Use it for the new explanations of the "not accessible at this address"
errors. Split in die_location_as_string, die_location_as_function_string
and suggested_locations_string.
Mark Wielaard [Wed, 4 Jun 2014 18:48:23 +0000 (20:48 +0200)]
Report address locations where things are accessible in error message.
Instead of just saying "not accessible at this address" tell at which
addresses the location is accessible. Uses dwarf_getlocations which
requires elfutils 0.158+ (if not available, just don't suggest anything).
Mark Wielaard [Wed, 4 Jun 2014 13:24:41 +0000 (15:24 +0200)]
Improve not accessible at this address error message.
Add dwflpp::pc_die_location_as_string and dwflpp::pc_die_location_as_string
to provide more human readable PC and DWARF DIE offset names and source
locations (including which function was inlined where) to help track down
why an argument or variable isn't accessible at a certain address.
David Smith [Tue, 3 Jun 2014 16:05:20 +0000 (11:05 -0500)]
Fixed BZ1099791 by updating the runtime to correctly find the vdso.
* translate.cxx (find_vdso): Add comment mentioning the logic in
_stp_vma_match_vdso needs to match.
* runtime/vma.c (_stp_vma_match_vdso): Add logic from find_vdso() to parse
the path, looking for the vdso.
Jonathan Lebon [Sun, 1 Jun 2014 19:58:48 +0000 (15:58 -0400)]
dwflpp.cxx: check addr returned by lower_bound()
This is a follow-up to commit 71b8e66. The lower_bound() function
returns an iterator to the first addr which does not compare less than
the entrypc. This means that the returned addr is either equal or
greater than the entrypc (depending on whether the latter is in the
vector).
This patch simply checks that the addr returned by lower_bound() is
equal, not just greater or equal.
Jonathan Lebon [Fri, 30 May 2014 19:42:00 +0000 (15:42 -0400)]
dwflpp.cxx: use lower_bound() instead of manual binary search
This is a follow-up to commit 1d50099. We previously implemented our own
binary search to find the line record associated with the function's
entrypc. To be on the safe side, we instead rely on lower_bound() for
binary searching, which will pick the first element in a range of equal
addresses, thus avoiding issues such as BZ1099133.
Jonathan Lebon [Fri, 30 May 2014 15:47:51 +0000 (11:47 -0400)]
BZ1099133: prologue searching: check previous line too
We perform binary search to retrieve the line record matching the
entrypc. However, we may by chance fall on the second of two line
records for the same entrypc. Check for this after binary search is
done.
Jonathan Lebon [Fri, 30 May 2014 14:26:01 +0000 (10:26 -0400)]
BZ1099825: probe_by_pid.exp: small tweak in C program
Since PR6941 (commit 277a2b9), prologue-searching is skipped for
optimized code. However, we do not skip it for GCC < 4.5, which may
provide bad location info (see GDB's PR13777).
On RHEL6, this testcase was failing because prologue-searching was not
skipped, and fell inside the loop in main(). To allow for a safe landing
zone, we slightly tweak the C program so that there is a safe
instruction to land on post-prologue that is not in the loop.
Stan Cox [Thu, 29 May 2014 20:04:13 +0000 (16:04 -0400)]
Relax static-probes.d input file syntax.
* dtrace.in (dtrace_bnf): Allow probe_ident to start with any alphanumeric.
Make trailing probe declaration ';' optional.
(_PypProvider::probe_write): Improve error messages.