]> sourceware.org Git - systemtap.git/log
systemtap.git
5 years agoConditionally define __NR_bdflush in systemtap runtime.
Martin Cermak [Tue, 12 Feb 2019 15:34:14 +0000 (16:34 +0100)]
Conditionally define __NR_bdflush in systemtap runtime.

The kernel-headers-4.17.0-0.rc6.1.el8+7.aarch64.rpm dropped
__NR_bdflush define, so let's define it conditionally in
compat_unistd.h so that stap scripts can -p4 without
complaints on those kernels.

5 years agotestsuite/systemtap.bpf :: couple of bpf_tests clarifications
Serhei Makarov [Mon, 11 Feb 2019 22:02:56 +0000 (17:02 -0500)]
testsuite/systemtap.bpf :: couple of bpf_tests clarifications

5 years agofix embedded-c tag memoization thinko
Frank Ch. Eigler [Mon, 11 Feb 2019 00:47:19 +0000 (19:47 -0500)]
fix embedded-c tag memoization thinko

The memoization needs to track both haystack and needle to compute &
cache the appropriate result.  Much badness happens without.

5 years agopass-2 optimization: embedded-c code pragma searching
Frank Ch. Eigler [Sun, 10 Feb 2019 19:27:34 +0000 (14:27 -0500)]
pass-2 optimization: embedded-c code pragma searching

It turns out that scripts like "--example errsnoop.stp" involves tens
of millions (!) of substring searches, as combinatorial & nesting
factors cause many embedded-c function/statement bodies to be searched
for "/* pragma */" type tags.

We probably do this too much - see the number (166417+70699) of times
varuse_collecting_visitor::visit_embedded* end up being called in
optimization/relaxation loops.  But even without delving into that, we
can improve the constant factor: the actual string searching speed.

This patch adds a little memoization widget to interned_string
substring searching.  It takes pass-2 runtime from 71s to 22s on my
workstation and costs practically no extra memory.  (Storing copies of
interned_strings is cheap.)

5 years agoSet up a BPF related testcase blacklist for regression testsing purposes.
Martin Cermak [Fri, 8 Feb 2019 09:58:37 +0000 (10:58 +0100)]
Set up a BPF related testcase blacklist for regression testsing purposes.

5 years agolinux api update: compat_time* structs
Frank Ch. Eigler [Thu, 7 Feb 2019 00:29:40 +0000 (18:29 -0600)]
linux api update: compat_time* structs

Several 32-bit time compatibility structs have been gradually
deprecated, being moved into new headers or renamed.

5 years agolinux api update: FAN_ALL_CLASS_BITS
Frank Ch. Eigler [Wed, 6 Feb 2019 15:51:10 +0000 (09:51 -0600)]
linux api update: FAN_ALL_CLASS_BITS

Above macro went into uapi-only header file in commit 23c9deeb3.

5 years agounwinder: increase MAX_CFI to 8192 after a census of CFI sizes on Fedora 29
Frank Ch. Eigler [Tue, 5 Feb 2019 20:44:35 +0000 (15:44 -0500)]
unwinder: increase MAX_CFI to 8192 after a census of CFI sizes on Fedora 29

O(5000) seen in the wild.

5 years agoruntime: adapt to uapi/linux/mount.h changes
Frank Ch. Eigler [Tue, 5 Feb 2019 18:29:26 +0000 (12:29 -0600)]
runtime: adapt to uapi/linux/mount.h changes

MS_* flags were moved to a new header file back in linux commit e262e32d.

5 years agodummy commit to force buildbot rebuilds
Serhei Makarov [Mon, 4 Feb 2019 20:09:35 +0000 (15:09 -0500)]
dummy commit to force buildbot rebuilds

5 years agobuildrun.cxx: older g++ compatibility: no std::string.back()
Frank Ch. Eigler [Mon, 4 Feb 2019 18:40:33 +0000 (13:40 -0500)]
buildrun.cxx: older g++ compatibility: no std::string.back()

Oh well, use [] indexing instead.

5 years agoHandle installation without stapusr group
David Ward [Wed, 30 Jan 2019 05:37:56 +0000 (00:37 -0500)]
Handle installation without stapusr group

Do not cause "make install" to return an error if the stapusr group
cannot be found or created (even as root); continue without setting
the ownership or mode of the installed executables. This may happen
when building distribution packages using fakeroot (it was observed
on Arch Linux). This step is often performed directly in the build
files of the distribution package instead (such as systemtap.spec).

5 years agoSimplify creation of groups during installation
David Ward [Wed, 30 Jan 2019 05:37:55 +0000 (00:37 -0500)]
Simplify creation of groups during installation

Use the "-f" option of "groupadd", rather than calling it a second
time if the desired GID is already in use.

Do not call "getent" twice. We know that a group exists if the first
call to "getent" returned successfully, or otherwise if "groupadd"
returned successfully.

5 years agoFix get_user_pages() autoconf test for 4.4.y kernel
Hou Tao [Fri, 1 Feb 2019 11:22:57 +0000 (19:22 +0800)]
Fix get_user_pages() autoconf test for 4.4.y kernel

The following kernel commit has been back-ported to linux-4.4.y kernel:

    commit 8e50b8b07f462ab4b91bc1491b1c91bd75e4ad40
    Author: Lorenzo Stoakes <lstoakes@gmail.com>
    Date:   Thu Oct 13 01:20:16 2016 +0100

mm: replace get_user_pages() write/force parameters with gup_flags

commit 768ae309a96103ed02eb1e111e838c87854d8b51 upstream.

This removes the 'write' and 'force' from get_user_pages() and replaces
them with 'gup_flags' to make the use of FOLL_FORCE explicit in callers
as use of this flag can result in surprising behaviour (and hence bugs)
within the mm subsystem.

And it changes the function signature of get_user_pages(), so introduce
an extra flag STAPCONF_GET_USER_PAGES_FLAGS and the corresponding test program
to fix it.

Signed-off-by: Hou Tao <houtao1@huawei.com>
5 years agolkm-runtime time.c: prefer 4.15+ ktime_get_real_fast_ns
Frank Ch. Eigler [Mon, 4 Feb 2019 16:15:44 +0000 (11:15 -0500)]
lkm-runtime time.c: prefer 4.15+ ktime_get_real_fast_ns

Comment out the whole time.c implementation if this function is
available.  It's NMI-safe and appears to do all we need it to.

5 years agolinux 5.0-rc3 adaptations
Frank Ch. Eigler [Sat, 2 Feb 2019 20:50:32 +0000 (14:50 -0600)]
linux 5.0-rc3 adaptations

Adapt to access_ok() kapi change in commit 96d4f267e40.
Adapt to changes in kbuild dependency automation.
Adapt to a gcc9 false-positive warning about empty array iteration.

5 years agoparser: allow larger arrays
Frank Ch. Eigler [Thu, 31 Jan 2019 14:30:29 +0000 (09:30 -0500)]
parser: allow larger arrays

We shouldn't enforce arbitrary array size limits at parse time.
Let actual resource limits kick in at run time instead.

5 years agostap man page: document histogram per-bucket access/iteration
Frank Ch. Eigler [Tue, 29 Jan 2019 19:23:30 +0000 (14:23 -0500)]
stap man page: document histogram per-bucket access/iteration

We've had this feature forever, but haven't documented it much.

5 years agoPR16406 fixup :: correct comparison in _stp_module_check
Serhei Makarov [Fri, 25 Jan 2019 16:44:24 +0000 (11:44 -0500)]
PR16406 fixup :: correct comparison in _stp_module_check

5 years agoPR10280 initial fix: force vermagic for guru-mode scripts
Serhei Makarov [Wed, 16 Jan 2019 18:56:23 +0000 (13:56 -0500)]
PR10280 initial fix: force vermagic for guru-mode scripts

Current version checking based on kernel ABI and build-ids is not
strict enough to prevent launching a stap module on a kernel version
it wasn't compiled for. This has the potential to crash a running
kernel, since ABI compatibility may not give a sufficient guarantee of
real compatibility.

Initial fix, should investigate what other scenarios should have
tighter checking.

5 years agoMake testcase at_var.exp work with stap-4.1+ and kernel-4.17+
Martin Cermak [Tue, 8 Jan 2019 16:06:41 +0000 (17:06 +0100)]
Make testcase at_var.exp work with stap-4.1+ and kernel-4.17+

This testcase update tries to reflect following changes:
- 4.17 breaks syscalls tapset
  https://sourceware.org/bugzilla/show_bug.cgi?id=23160
- $foo->bar differs from @var("foo")->bar
  https://bugzilla.redhat.com/show_bug.cgi?id=1660889#c4 .

5 years agostap-prep: switch typo ($NEEDED => $package)
Serhei Makarov [Thu, 3 Jan 2019 17:48:49 +0000 (12:48 -0500)]
stap-prep: switch typo ($NEEDED => $package)

$NEEDED is not actually defined at that point of the script.
I got an empty `apt install -y` command as a result.

5 years agoPR16596: add support for DW_OP_GNU_entry_value in location expressions
Jafeer Uddin [Mon, 10 Dec 2018 20:11:26 +0000 (15:11 -0500)]
PR16596: add support for DW_OP_GNU_entry_value in location expressions

5 years agoMake sysc_bdflush.stp compatible with 4.17+ kernels.
Martin Cermak [Thu, 6 Dec 2018 12:52:20 +0000 (13:52 +0100)]
Make sysc_bdflush.stp compatible with 4.17+ kernels.

The bdflush syscall itself appears to be obsolete since 2.6,
but this way we at least won't end up with pass 1 "resolution
failed in alias expansion builder" when randomly probing for it.
There is an old customer rh bz 544960 related to bdflush.

5 years agoconfigury: tolerate dyninst10's need for -lboost_system
Frank Ch. Eigler [Tue, 4 Dec 2018 19:22:33 +0000 (13:22 -0600)]
configury: tolerate dyninst10's need for -lboost_system

Some shlib references are generated from compiling stapdyn.cxx.

5 years agoRHBZ1655631: systemtap.spec: use Recommends: for kernel-*-devel
Frank Ch. Eigler [Mon, 3 Dec 2018 20:26:52 +0000 (15:26 -0500)]
RHBZ1655631: systemtap.spec: use Recommends: for kernel-*-devel

5 years agoruntime/unwind: Allow to increase MAX_CFI
Torsten Polle [Tue, 27 Nov 2018 21:15:44 +0000 (22:15 +0100)]
runtime/unwind: Allow to increase MAX_CFI

As the limit MAX_CFI is an arbitrary value, the value might need to be
changed by an external definition (e.g. -DMAX_CFI=8192).

Acked-by: Torsten Polle <Torsten.Polle@gmx.de>
5 years agosystemtap.spec: merge unbotch - missing %changelog
Frank Ch. Eigler [Tue, 27 Nov 2018 03:36:55 +0000 (22:36 -0500)]
systemtap.spec: merge unbotch - missing %changelog

5 years agosystemtap.spec: importing fedora rawhide changes
Peter Robinson [Mon, 26 Nov 2018 23:50:25 +0000 (18:50 -0500)]
systemtap.spec: importing fedora rawhide changes

Crash and virt host supported on ppc64le and aarch64,
Move chkconfig to traditional initscripts deps (not required for systemd),
Minor spec cleanups

5 years agostaprun/mainloop.c :: fix for loop initial decls
Serhei Makarov [Mon, 26 Nov 2018 20:35:19 +0000 (15:35 -0500)]
staprun/mainloop.c :: fix for loop initial decls

5 years agoPR23866 part: expose raw syscall tracepoint to bpf
Frank Ch. Eigler [Mon, 26 Nov 2018 15:36:28 +0000 (10:36 -0500)]
PR23866 part: expose raw syscall tracepoint to bpf

Support TRACE_EVENT_FN() type tracepoints in bpf target, since those
too define trace-event structures that bpf callbacks have access to.
(We do not yet decode nor explose Synthetic cooked per-syscall
trace-event structs.  Those aren't statically extractable from headers,
as the kernel builds them at boot time.)

5 years agoPR23891: Make sure stap and staprun respond to SIGTERM when stderr/stdout are blocked
Yichun Zhang (agentzh) [Thu, 22 Nov 2018 07:10:14 +0000 (23:10 -0800)]
PR23891: Make sure stap and staprun respond to SIGTERM when stderr/stdout are blocked

When stderr/stdout are blocked (the write buffers are full), write()
syscalls in stap's signal handler and staprun's stp_main_loop() might
prevent these processes from responding to signals like SIGTERM.

Also make staprun respond to SIGPIPE just like SIGTERM.

We introduce the kill_relayfs() function to kill the reader thread in
staprun without waiting for the readers (which might alreayd be blocked
on writing to stdout).

Our local stress tests have confirmed that this patch indeed fixes the
hanging issues in stap and staprun.

5 years agoFix segfaults in dwarf array pointer subscripting when -vvv is specified
Yichun Zhang (agentzh) [Thu, 22 Nov 2018 07:27:30 +0000 (23:27 -0800)]
Fix segfaults in dwarf array pointer subscripting when -vvv is specified

location_context::new_symref() forgot to initialize sym->tok which led
to NULL e->tok pointers.

Thanks to Mozilla rr for greatly simplifying debugging this issue.

Added some tests to cover this fix.

5 years agoMake opeartor @var() no longer assume @entry() in return probes.
Yichun Zhang (agentzh) [Thu, 22 Nov 2018 07:52:41 +0000 (23:52 -0800)]
Make opeartor @var() no longer assume @entry() in return probes.

The old behavior would yield stale values when the function being probed
changes the global variables being read via @var() in the return probe
handler.

Added tests to cover this fix (and the old behavior for compatibility).

5 years agoAdd the ucalls.stp to the systemtap examples.
William Cohen [Tue, 20 Nov 2018 19:50:41 +0000 (14:50 -0500)]
Add the ucalls.stp to the systemtap examples.

5 years agoAdjust tcp_trace.stp example to work with newer Linux kernel's timers
William Cohen [Mon, 19 Nov 2018 20:17:10 +0000 (15:17 -0500)]
Adjust tcp_trace.stp example to work with newer Linux kernel's timers

The newer Linux kernels removed the data field from struct timer_list
and now derive that equivalent information using the container_of
macro to find the data structure the struct timer_list is embedded in.
This patch makes tcp_trace.stp flexible and allows it to get the
needed information when the struct timer_list data field and the
associated $data target variable for timer functions are not
available.

5 years agoAdjust the vfs_open to provide cred variable with 4.18 kernels
William Cohen [Mon, 19 Nov 2018 19:23:15 +0000 (14:23 -0500)]
Adjust the vfs_open to provide cred variable with 4.18 kernels

The kernel's git commit ae2bb293a3e8adbc54d08cede5afc22929030c03
removed the cred argument from the vfs_open.  Thus, there is no $cred
target variable available.  This missing target variable lives on as a
field in the $file target variable.  The patch makes the tapset use
that field if the $cred target variable is not available.  Fixing this
allows the slowvfs.stp example to work with newer linux 4.18 kernels.

5 years agoUserspace_probing.xml: change '-ldd' to '--ldd' command line option.
Mark Wielaard [Sun, 18 Nov 2018 13:33:47 +0000 (14:33 +0100)]
Userspace_probing.xml: change '-ldd' to '--ldd' command line option.

https://sourceware.org/bugzilla/show_bug.cgi?id=23894

5 years agobpf-translate.cxx: don't clobber any earlier value of c
Serhei Makarov [Fri, 16 Nov 2018 17:12:07 +0000 (12:12 -0500)]
bpf-translate.cxx: don't clobber any earlier value of c

5 years agobpf-translate.cxx: quiet a compiler warning
Frank Ch. Eigler [Fri, 16 Nov 2018 03:02:18 +0000 (22:02 -0500)]
bpf-translate.cxx: quiet a compiler warning

It's a false positive with rawhide gcc, but easy to quiet anyway.

5 years agoPR23890 bonus: show nicer messages upon a buildid mismatch
Frank Ch. Eigler [Fri, 16 Nov 2018 01:22:34 +0000 (20:22 -0500)]
PR23890 bonus: show nicer messages upon a buildid mismatch

Instead of producing only a one-byte error, we now compute the entire
builds into hex text strings, and report the whole shebang on an
error.  (Also, ditch some 2.6.27 kernel-bug compatibiltiy fossil
in the area.)

5 years agoPR23890: tolerate f29+ style ELF files
Frank Ch. Eigler [Thu, 15 Nov 2018 21:27:58 +0000 (16:27 -0500)]
PR23890: tolerate f29+ style ELF files

Reported by kenj@pcp, with mjw et al.'s help, we found out why
systemtap on fedora 29+ routinely fails to verify build-ids for
userspace programs.  F29 adds a separate loadable segment with the
relevante .note's, before the main text segment.  The runtime code
that listens to mmaps-in-progress now accepts this configuration.
As long as the .note section is loaded (time-wise and space-wise)
before the .text one(s), we're good.

5 years agoPR23747: tolerate symbols with odd section#s
Mark Wielaard [Wed, 14 Nov 2018 18:28:04 +0000 (13:28 -0500)]
PR23747: tolerate symbols with odd section#s

In f29+ kernels, some note-related symbols are not found in our
usual section# search, when collecting unwind/symbol data.  As
these symbols are now ignored instead of causing an error.

5 years agosystemtap.bpf/asm_tests :: fix documentation / expected result
Serhei Makarov [Tue, 13 Nov 2018 19:11:15 +0000 (14:11 -0500)]
systemtap.bpf/asm_tests :: fix documentation / expected result

5 years agoPR23875 bpf_unparser::visit_foreach_loop(): temporarily disable string key iteration
Serhei Makarov [Tue, 13 Nov 2018 18:13:14 +0000 (13:13 -0500)]
PR23875 bpf_unparser::visit_foreach_loop(): temporarily disable string key iteration

5 years agoPR23875: another testcase that loops indefinitely
Serhei Makarov [Tue, 13 Nov 2018 18:07:52 +0000 (13:07 -0500)]
PR23875: another testcase that loops indefinitely

5 years agoPR23875 bpf.exp: fail testcase on 'stack smashing'
Serhei Makarov [Tue, 13 Nov 2018 17:27:05 +0000 (12:27 -0500)]
PR23875 bpf.exp: fail testcase on 'stack smashing'

5 years agoPR23860 bpf_interpret() :: NEG should not fall through to DIV
Serhei Makarov [Tue, 13 Nov 2018 16:42:46 +0000 (11:42 -0500)]
PR23860 bpf_interpret() :: NEG should not fall through to DIV

5 years agoAdjust the periodic.stp example to work with newer Linux kernels
William Cohen [Mon, 12 Nov 2018 22:24:44 +0000 (17:24 -0500)]
Adjust the periodic.stp example to work with newer Linux kernels

The data field in the timer_list struct was removed in newer kernels.
The various functions executed when a timer expires now use
container_of macros to find the struct that the timer_list was
embedded in.  The periodic.stp script has been modified to use
container_of when the data field is not available.

5 years agostap -vv: also print kernel build-tree path name
Frank Ch. Eigler [Sun, 11 Nov 2018 02:52:32 +0000 (21:52 -0500)]
stap -vv: also print kernel build-tree path name

This aids diagnostics.

5 years agotestcase for PR23875
Serhei Makarov [Fri, 9 Nov 2018 21:42:21 +0000 (16:42 -0500)]
testcase for PR23875

This triggers a 'stack smashing' error in the userspace interpreter.

* testsuite/systemtap.bpf/bpf_tests/pr23875.stp: New testcase.

5 years agoPR23860: reduce stack pressure from format strings
Serhei Makarov [Fri, 9 Nov 2018 21:24:09 +0000 (16:24 -0500)]
PR23860: reduce stack pressure from format strings

Reduce stack pressure created by the earlier commits by allocating
format strings in a predictable location in the top half of the stack
[-BPF_MAXSTRINGLEN*2..0) as long as they fit in there. This works
since only one format string is active at a time and no ordinary
strings are being allocated in that region of the stack now.

* bpf-opt.cxx (alloc_literal_str): Store format_str in top half.

5 years agoPR23860: additional ugly stack/clobber protection for strings
Serhei Makarov [Fri, 9 Nov 2018 19:36:19 +0000 (14:36 -0500)]
PR23860: additional ugly stack/clobber protection for strings

In addition to prior commit, emit_string_copy() does not work for
overlapping source/destination. So, make sure strings are not
allocated in a way which overlaps map key/value arguments.

This increases space pressure, inducing a couple of bpf-asm.exp
testcase failures.

* bpf-internal.h (value::format_str): New flag.
(value::value): Take format_str flag.
(value::mk_str): Take format_str flag.
(program::format_map): New field, caches format_str separately.
(program::new_str): Take format_str flag.
* bpf-base.cxx (program::new_str): Cache format_str separately.
* bpf-opt.cxx (alloc_literal_str): Store non-format str in lower half.
* bpf-translate.cxx (emit_string_copy): Comment -- doesn't support overlap.
(emit_string_copy): DEBUG_CODEGEN -- identify if zero-padding was done.
(emit_print_format): Set format_str flag.

5 years agoPR23860: additional stack protection for strings
Serhei Makarov [Thu, 8 Nov 2018 21:40:40 +0000 (16:40 -0500)]
PR23860: additional stack protection for strings

Fixes for verifier rejection of some cases requiring string copy,
since the verifier would reject string copy code extending beyond the
end of the string even if it was not reachable.

* bpf-opt.cxx (alloc_literal_str): make sure the offset for a short
  string is at least BPF_MAXSTRINGLEN.
(zero_stack): New function.
(program::generate): Use zero_stack() to zero temporary area and
  prevent verifier complaints.
* testsuite/systemtap.bpf/asm_tests/pr23860.stp: New testcase.

5 years agoPR23860 bugfix: incorrect comparison direction in string_copy()
Serhei Makarov [Fri, 9 Nov 2018 21:19:17 +0000 (16:19 -0500)]
PR23860 bugfix: incorrect comparison direction in string_copy()

(Turns out this branch was flipped and it was a root cause of the havoc.)

* bpf-translate.cxx (bpf_unparser::emit_string_copy): Correct
  direction of JEQ(all_nz,0) jump instruction.

5 years agoPR23507: add new command-line option to disable automatic unread global variable...
Jafeer Uddin [Thu, 8 Nov 2018 21:19:14 +0000 (16:19 -0500)]
PR23507: add new command-line option to disable automatic unread global variable display

5 years agopr23860 verifier workaround :: be sure to delete all mov rN,rN
Serhei Makarov [Wed, 7 Nov 2018 18:07:51 +0000 (13:07 -0500)]
pr23860 verifier workaround :: be sure to delete all mov rN,rN

An apparent bug in the eBPF verifier fails to preserve register state
when MOVing a register to itself, marking rN as 'unknown scalar'.

Previously bpf-opt.cxx failed to remove spurious MOVs if they were the
final instruction in a basic block. This would fail verification if
the register holds a pointer.

5 years agoPR23507: add underscores to global @this variables
Jafeer Uddin [Wed, 7 Nov 2018 16:52:33 +0000 (11:52 -0500)]
PR23507: add underscores to global @this variables

5 years agoPR23761: generalized @entry
Jafeer Uddin [Wed, 7 Nov 2018 14:41:45 +0000 (09:41 -0500)]
PR23761: generalized @entry

With the changes to the syscall tapset for kernel 4.17+, it is now
possible for non-[uk]retprobes to trap return events. This means that
the @entry mechanism to access entry probe target variables in return
probes is not guaranteed to work. To get around this issue, a collection
of 8 global variables have been added to the tapset which can be used
to save variables in the entry probe and retrieve them later in return
probes. The global variables can be accessed using the @this1, .., @this8
macros.

5 years agostandardize ktime_get_ns() across lkm, bpf runtimes
Frank Ch. Eigler [Tue, 6 Nov 2018 21:26:57 +0000 (16:26 -0500)]
standardize ktime_get_ns() across lkm, bpf runtimes

Make sure ktime_get_ns() is available across runtimes.  In the case of
bpf, add a userspace helper to implement the function.  Add test case.
Add a systemtap.bpf/nobpf.exp test driver, which runs all the
bpf_tests but specifically without "--bpf", in the hope that all those
scripts should run on the normal backend too.  PR23866 blocks some of
that at the moment.

5 years agobpf behind-the-scenes :: useful DEBUG_CODEGEN diagnostic
Serhei Makarov [Tue, 6 Nov 2018 17:32:38 +0000 (12:32 -0500)]
bpf behind-the-scenes :: useful DEBUG_CODEGEN diagnostic

5 years agoAlways use nssInit for http and nss server.
Stan Cox [Tue, 6 Nov 2018 17:09:58 +0000 (12:09 -0500)]
Always use nssInit for http and nss server.

* nsscommon.h (db_init_types): Add db_init_types
* nsscommon.cxx (add_client_cert):  Use it to differentiate type of
  db init.  Change all callers.
* client-http.cxx (fill_in_server_info):  Use http server default
  port if none specified.

5 years agoPR23860 partial fix: fix BPF_NEG opcode generation.
Serhei Makarov [Mon, 5 Nov 2018 21:58:21 +0000 (16:58 -0500)]
PR23860 partial fix: fix BPF_NEG opcode generation.

Plus some improved diagnostics on malformed code.

* bpf-base.cxx (value::print): Don't abort() on unknown operand.
(opcode_name): Don't abort() on unknown opcode.
(insn::print): Don't abort() on malformed insn.
(program::mk_binary): Ensure BPF_NEG src==dest, don't use BPF_X.

5 years agoPR23829 :: fallback defines __BPF_FUNC_MAPPER and BPF_J{LT,LE,SLT,SLE} for older...
Serhei Makarov [Fri, 2 Nov 2018 20:49:23 +0000 (16:49 -0400)]
PR23829 :: fallback defines __BPF_FUNC_MAPPER and BPF_J{LT,LE,SLT,SLE} for older kernels

5 years agotapset/bpf/task.stp :: rudiment of task tapset
Serhei Makarov [Fri, 2 Nov 2018 16:22:01 +0000 (12:22 -0400)]
tapset/bpf/task.stp :: rudiment of task tapset

5 years agoPR23849 -- temporarily disable stapbpf script caching
Serhei Makarov [Thu, 1 Nov 2018 20:13:07 +0000 (16:13 -0400)]
PR23849 -- temporarily disable stapbpf script caching

5 years agoprometheus-exporter samples: change reported metric name
Frank Ch. Eigler [Fri, 26 Oct 2018 15:24:05 +0000 (11:24 -0400)]
prometheus-exporter samples: change reported metric name

When prometheus scrapes metrics, by default it'll simply preserve
the incoming names.  This doesn't work well when many different
stap scripts use the same metric name ("count"), or if the name
happens to be a reserved keyword in the promql language ("count").

So rename them to something more collision-proof.

5 years agostap-exporter: drop initial demo scripts under .examples; not used
Frank Ch. Eigler [Thu, 25 Oct 2018 16:32:33 +0000 (12:32 -0400)]
stap-exporter: drop initial demo scripts under .examples; not used

5 years agopowerpc64: add missing system call defines
Victor Kamensky [Wed, 31 Oct 2018 06:15:16 +0000 (23:15 -0700)]
powerpc64: add missing system call defines

A set of system call defines like __NR_pkey_alloc, __NR_pkey_free
__NR_pkey_mprotect are missing in powerpc64 kernel as of 4.14 kernel
version.

Add corresponding definitions so system call related probes
would compile. Tested with nd_syscalls-all-probes.stp.

Signed-off-by: Victor Kamensky <kamensky@cisco.com>
5 years agoaarch64: add missing system call defines
Victor Kamensky [Wed, 31 Oct 2018 06:15:15 +0000 (23:15 -0700)]
aarch64: add missing system call defines

A set of system call defines like __NR_alarm, __NR_ioperm,
__NR_modify_ldt, __NR_time, __NR_utime is missing in aarch64
kernel as of 4.18 kernel version.

Add corresponding definitions so system call related probes
would compile. Tested with nd_syscalls-all-probes.stp.

Signed-off-by: Victor Kamensky <kamensky@cisco.com>
5 years agotapset/bpf/context.stp :: add execname(), triage other functions
Serhei Makarov [Tue, 30 Oct 2018 21:29:46 +0000 (17:29 -0400)]
tapset/bpf/context.stp :: add execname(), triage other functions

* tapset/bpf/context.stp: Notes on other functions that could be added.
(execname): New tapset function.

* tapset/linux/context.stp: Move pexecname() to a more logical location.

5 years agotapset/bpf/conversions.stp bugfix :: helper name in kernel_string(addr, err_msg)
Serhei Makarov [Tue, 30 Oct 2018 21:13:28 +0000 (17:13 -0400)]
tapset/bpf/conversions.stp bugfix :: helper name in kernel_string(addr, err_msg)

5 years agobpf-asm.exp bugfix :: bad_output does occur
Serhei Makarov [Tue, 30 Oct 2018 21:11:45 +0000 (17:11 -0400)]
bpf-asm.exp bugfix :: bad_output does occur

5 years agobpf-translate.cxx :: fix segfault with malformed register
Serhei Makarov [Tue, 30 Oct 2018 21:10:53 +0000 (17:10 -0400)]
bpf-translate.cxx :: fix segfault with malformed register

5 years agoFix miscellaneous errors/typos in syscall tapset
Jafeer Uddin [Tue, 30 Oct 2018 20:07:05 +0000 (16:07 -0400)]
Fix miscellaneous errors/typos in syscall tapset

* tapset/linux/sysc_pkey_*.stp: properly cast target variables in dw_syscall probe
* runtime/linux/compat_unistd.h: fix typos and errors in __NR_* definitions
* testsuite/systemtap.syscall/pkey.c: remove empty first line in file

5 years agoOn aarch64 Linux system calls related SystemTap scripts
Victor Kamensky [Tue, 30 Oct 2018 19:58:24 +0000 (15:58 -0400)]
On aarch64 Linux system calls related SystemTap scripts
compilation fail with "__NR_compat_[exit|read|write] redefined"
errors after following two commits:

 7abf0aee9 PR23160,PR14690: remove references to ia32 and x86 to make sysc_* files as arch-independent as possible
 cd84aedca PR23160,PR14690: adapt 13 more syscalls for 4.17 __ARCH_sys_FOO and sys_enter/exit

aarch64 kernel defines __NR_compat_[exit|read|write] after
 a1ae65b21941 arm64: add seccomp support

aarch64 kernel define __NR_compat_restart_syscall after
 f3e5c847ec3d arm64: Add __NR_* definitions for compat syscalls

Fix by adding proper conditional compilation based on current
architecture and kernel version.

5 years agoAdjust the BPF translate error report formatting to work on 32-bit architectures
William Cohen [Tue, 30 Oct 2018 18:20:46 +0000 (14:20 -0400)]
Adjust the BPF translate error report formatting to work on 32-bit architectures

The 32-bit architectures such as arm and i686 had arguments in the
error reporting that did not match up with the %lu or %ld formatting.
Used type casting and %llu and %lld to avoid variation between 32-bit
and 64-bit architectures.

5 years agosession.cxx :: enable caching for bpf backend
Serhei Makarov [Fri, 26 Oct 2018 15:15:13 +0000 (11:15 -0400)]
session.cxx :: enable caching for bpf backend

5 years agobpf-translate.cxx :: plug an exception gap in is_numeric()
Serhei Makarov [Thu, 25 Oct 2018 16:19:53 +0000 (12:19 -0400)]
bpf-translate.cxx :: plug an exception gap in is_numeric()

5 years agoMerge branch 'serhei/bpf_asm' -- kernel_string() tapset and experimental bpf assembler
Serhei Makarov [Wed, 24 Oct 2018 20:04:30 +0000 (16:04 -0400)]
Merge branch 'serhei/bpf_asm' -- kernel_string() tapset and experimental bpf assembler

Note the big comment in bpf-translate.cxx explaining the new assembler.

Major changes:
- Embedded-code assembler
- TODO Embedded-code tapset function call support is incomplete, only enabled for exit()
- TODO Token adjustment for assembler diagnostics needs work.
- Refactor bpf_unparser
- Improved metadata about helpers
- String handling changes
- Misc cleanup/notes

Tapset for kernel_string:

* tapset/bpf/conversions.stp: New file.
(kernel_string): New function.
(kernel_string): New function (err_msg version), in assembly.
(kernel_string_n): New function, in assembly.
* testsuite/systemtap.bpf/bpf_tests/context_vars3.stp: New testcase.

Embedded-code assembler:

* bpf-translate.cxx (bpf_unparser::parse_imm): New function.
(bpf_unparser::parse_asm_stmt): New function.
(bpf_unparser::emit_asm_arg): New function.
(bpf_unparser::parse_reg): Removed.
(bpf_unparser::emit_asm_reg): New function.
(bpf_unparser::get_asm_reg): New function.
(bpf_unparser::emit_asm_opcode): New function.
(bpf_unparser::visit_embeddedcode): Process new assembly format.
(BPF_ASM_DEBUG): New (disabled) macro for diagnostics.
(struct asm_stmt): New structure.
(operator <<): New function -- print logic for asm_stmt.
(is_numeric): New function.
* testsuite/systemtap.bpf/asm_tests/*: New testcases for embedded-code assembler.
* testsuite/systemtap.bpf/bpf-asm.exp: TODO Initial test driver for embedded-code assembler.

TODO Embedded-code tapset function call support is incomplete, only enabled for exit():

* bpf-translate.cxx (translate_bpf_pass): Pass systemtap_session to assembler globals.
* bpf-internal.h (globals::session): New field to pass systemtap_session to assembler.

TODO Token adjustment for assembler diagnostics needs work:

* parse.h (token::adjust_location): New function.

Refactor bpf_unparser:

* bpf-translate.cxx (bpf_unparser::emit_functioncall): New function.
(bpf_unparser::visit_functioncall): Use emit_functioncall.
(print_format_add_tag): New function on std::string.
(bpf_unparser::emit_print_format): New function.
(bpf_unparser::visit_print_format): Use print_format_add_tag, emit_print_format.

Improved metadata about helpers:

* bpf-base.cxx (bpf_func_name_map): New structure -- id->name map.
(bpf_func_id_map): New structure -- name->id map.
(init_bpf_helper_tables): New function -- populate name->id and id->name map.
(bpf_function_name): Change to use the maps.
(bpf_function_id): New function -- map from name to helper id.
(bpf_function_nargs): TODO Still need to expand the list of helpers.
* bpf-translate.cxx (translate_bpf_pass): Call init_bpf_helper_tables to populate info.
* bpf-internal.h (init_bpf_helper_table): New function.
(bpf_function_id): New function.
(__STAPBPF_FUNC_MAPPER): New macro -- like __BPF_FUNC_MAPPER for userspace-only helpers.

String handling changes:

* bpf-translate.cxx (bpf_unparser::emit_literal_str): New function.
(bpf_unparser::visit_literal_str): Use emit_literal_str.
(emit_simple_literal_str): Renamed from emit_literal_str.
(bpf_unparser::emit_string_copy): Renamed from emit_copied_str;
rename emit_literal_str to emit_simple_literal_str.
(bpf_unparser::emit_str_arg): Rename emit_copied_str to emit_string_copy.
(translate_escapes): Takes a const string now.
* bpf-opt.cxx (alloc_literal_str): Rename emit_literal_str to emit_simple_literal_str.
* bpf-internal.h (emit_simple_literal_str): Renamed from emit_literal_str.

Misc cleanup/notes:

* bpf-internal.h (BPF_MAXSTRINGLEN): TODO Someday this will be increased.
(program::use_tmp_space): Assert to catch miscalulations.
* tapset/logging.stp (abort): TODO Could abort immediately with assembly in future.

5 years agotestsuite/systemtap.bpf :: diagnose a bug in print_format("%s%s", ...)
Serhei Makarov [Wed, 24 Oct 2018 19:56:44 +0000 (15:56 -0400)]
testsuite/systemtap.bpf :: diagnose a bug in print_format("%s%s", ...)

5 years agostapbpf assembler WIP #8 :: bpf-asm.exp driver and more testcases serhei/bpf_asm
Serhei Makarov [Wed, 24 Oct 2018 17:30:29 +0000 (13:30 -0400)]
stapbpf assembler WIP #8 :: bpf-asm.exp driver and more testcases

5 years agostapbpf assembler WIP #7 :: fixed kernel_string() tapset and testcase
Serhei Makarov [Wed, 24 Oct 2018 16:46:55 +0000 (12:46 -0400)]
stapbpf assembler WIP #7 :: fixed kernel_string() tapset and testcase

* tapset/bpf/conversions.stp (kernel_string_n): enable error path.
* tapset/logging.stp (abort): note future work.
* testsuite/systemtap.bpf/bpf_tests/context_vars3.stp: new testcase.

5 years agostapbpf assembler WIP #6 :: other call functions ({s}printf and tapset)
Serhei Makarov [Tue, 23 Oct 2018 17:35:08 +0000 (13:35 -0400)]
stapbpf assembler WIP #6 :: other call functions ({s}printf and tapset)

Only very limited support for tapset functions (restricted to exit()
for now) due to the difficulty of resolving symbols after the semantic
pass is already completed. Could address this in the future.

* bpf_internal.h (program::use_tmp_space): check for overflow.
(globals::session): new field for systemtap_session (used by function lookup).

* bpf_translate.cxx (asm_stmt::has_jmp_target): new field.
(operator <<): printing rules for alloc, call.
(bpf_unparser::parse_asm_stmt): remove printf/error, BUGFIX alloc, call, string literal.
Also calculate has_jmp_target in the resulting stmt.
(bpf_unparser::visit_embeddedcode): handle printf, sprintf and exit().
Also fix the way fallthrough fields are populated to avoid spurious extra jump.
(bpf_unparser::emit_functioncall): new function. Factors out non-staptree code.
(bpf_unparser::visit_functioncall): use new emit_functioncall().
(print_format_add_tag): new function on std::string. Factors out string operations.
(bpf_unparser::emit_print_format): new function. Factors out non-staptree code.
(bpf_unparser::visit_print_format): use new emit_print_format().
(translate_bpf_pass): store session in globals.

5 years agoPR21080: support added for new pkey_* syscalls
Jafeer Uddin [Tue, 23 Oct 2018 19:19:29 +0000 (15:19 -0400)]
PR21080: support added for new pkey_* syscalls

* sysc_pkey_*.stp: new syscall probes
* aux_syscalls.stp: add new function to convert init_val to PKEY_DISABLE_[ACCESS|WRITE]
* compat_unistd.h: add new syscall numbers
* pkey.c: tests for new syscall

5 years agostap-exporter examples: use symlinks rather than copies
Frank Ch. Eigler [Tue, 23 Oct 2018 18:19:35 +0000 (14:19 -0400)]
stap-exporter examples: use symlinks rather than copies

Some of the .stp examples were accidentally replaced by
copies of the EXAMPLE driver script.  Back to symlinks.

6 years agoUse NSS_InitContext instead of NSS_Init.
Stan Cox [Tue, 23 Oct 2018 02:29:32 +0000 (22:29 -0400)]
Use NSS_InitContext instead of NSS_Init.

* nsscommon.cxx (nssInitContext): New function which allows
   multiple nss invocations.  Change all callers except where
   write access is required.
  (nssCleanup): Add context parameter.  Change all callers.
* client-http.cxx (download): Add cleanup parameter to choose
   curl_easy_cleanup.  Change all callers.
  (download_pem_cert): If CURLINFO_CERTINFO fails then retrieve cert from server.
  (find_and_connect_to_server): If the database cert fails then try again
   with retrieved server cert.
  (fill_in_server_info): Likewise.

* testsuite/systemtap.http_server/server_trust.exp:  New test.

6 years agoprometheus tapset: add dump_array_*_unquoted variants
Frank Ch. Eigler [Mon, 22 Oct 2018 20:26:15 +0000 (16:26 -0400)]
prometheus tapset: add dump_array_*_unquoted variants

For demos like also_ran.stp, the incoming strings are usually already
quoted.  Re-quoting them for prometheus labeling is counterproductive,
so we now offer an option to bypass that string_quoted() wrapping.
also_ran.stp updated.

before:

count{path="\"/lib64/libdl.so.2\""} 69
count{path="\"/lib64/libpthread.so.0\""} 76
count{path="\"/usr/bin/expr\""} 89

after:

count{path="/lib64/libdl.so.2"} 69
count{path="/lib64/libpthread.so.0"} 76
count{path="/usr/bin/expr"} 89

6 years agonfsd tapset: adapt nfsd.proc4.commit probe to different kernel versions
Frank Ch. Eigler [Fri, 19 Oct 2018 23:31:31 +0000 (19:31 -0400)]
nfsd tapset: adapt nfsd.proc4.commit probe to different kernel versions

Here too an if(@defined($u)) is appropriate, just like in
every other case.  Audited remainder.

6 years agoPR23799 - sprint_ustack() always returns empty string values
Yichun Zhang (agentzh) [Fri, 19 Oct 2018 21:55:08 +0000 (14:55 -0700)]
PR23799 - sprint_ustack() always returns empty string values

The tapset function sprint_ustack() forgot to actually return
the computed string value. Alas.

6 years agonfsd tapset: adapt nfsd.proc4.read probe to different kernel versions
Frank Ch. Eigler [Fri, 19 Oct 2018 21:36:04 +0000 (17:36 -0400)]
nfsd tapset: adapt nfsd.proc4.read probe to different kernel versions

Here too an if(@defined($u)) is appropriate, just like in the .write case.

6 years agostep-prep: on debian/ubuntu machines, attempt "apt-get -y install"
Frank Ch. Eigler [Thu, 18 Oct 2018 19:46:09 +0000 (15:46 -0400)]
step-prep: on debian/ubuntu machines, attempt "apt-get -y install"

If we're running as root, and the apt cache already knows our desired
package, actually run apt-get to install it.

6 years agoUse cast to make c->cycles_sum aways match the %lld format.
William Cohen [Fri, 19 Oct 2018 18:59:27 +0000 (14:59 -0400)]
Use cast to make c->cycles_sum aways match the %lld format.

On aarch64 and ppc64le cycles_t is a slightly different type from the
x86_64 and does not match up with the %lld format.  Cast c->cyles_sum
to always be (long long) to avoid the compile failing on aarch64 and
ppc64le with the following message:

**** failed systemtap kernel-devel smoke test:

/tmp/stapzubPmR/stap_e418199b88a6f8adf13a14e064ae79da_1403_src.c: In function '_stp_hrtimer_notify_function':
/tmp/stapzubPmR/stap_e418199b88a6f8adf13a14e064ae79da_1403_src.c:477:45: error: format '%lld' expects argument of type 'long long int', but argument 2 has type 'cycles_t' {aka 'long unsigned int'} [-Werror=format=]
             _stp_error ("probe overhead (%lld cycles) exceeded threshold (%lld cycles) in last %lld cycles", c->cycles_sum, STP_OVERLOAD_THRESHOLD, STP_OVERLOAD_INTERVAL);
                                          ~~~^                                                                ~~~~~~~~~~~~~
                                          %ld
cc1: all warnings being treated as errors

6 years agoAdd more quantitative data to error message when probes exceed threshold
Jafeer Uddin [Fri, 19 Oct 2018 14:12:52 +0000 (10:12 -0400)]
Add more quantitative data to error message when probes exceed threshold

6 years agoAvoid using target variable in target_set.stp for syscall.* probes.
William Cohen [Thu, 18 Oct 2018 17:44:03 +0000 (13:44 -0400)]
Avoid using target variable in target_set.stp for syscall.* probes.

6 years agoAvoid using target variables for syscall.write in print_user_buffer.stp.
William Cohen [Thu, 18 Oct 2018 17:31:28 +0000 (13:31 -0400)]
Avoid using target variables for syscall.write in print_user_buffer.stp.

6 years agoAvoid using target variables in signal.stp for syscall.* based probe points.
William Cohen [Thu, 18 Oct 2018 17:05:25 +0000 (13:05 -0400)]
Avoid using target variables in signal.stp for syscall.* based probe points.

6 years agoAdd buildok/syscall_any.stp to list of tests dyninst will not run.
William Cohen [Thu, 18 Oct 2018 16:05:56 +0000 (12:05 -0400)]
Add buildok/syscall_any.stp to list of tests dyninst will not run.

6 years agokprocess.exec_complete should avoid using $return from syscall.execve.return
William Cohen [Thu, 18 Oct 2018 15:47:50 +0000 (11:47 -0400)]
kprocess.exec_complete should avoid using $return from syscall.execve.return

6 years agostapbpf assembler WIP #5 :: basic kernel_string() implementation
Serhei Makarov [Tue, 16 Oct 2018 22:17:25 +0000 (18:17 -0400)]
stapbpf assembler WIP #5 :: basic kernel_string() implementation

This page took 0.067362 seconds and 5 git commands to generate.