Add perf_output support for high rate events

Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Incoporating feedback from bgregg:
Make per-cpu nature of output buffer be hidden and implicit. As such,
incoporate some rewriting into the definition of the PERF_OUTPUT.
Create two different macros to distinguish the perf_array (hardware
counters) from the perf_output (ring buffer) use cases.

Also, rename perf_output to perf_submit.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

This adds support for the bpf_perf_event_output command. This is
intended for per-process events from bpf to userspace at high rate. The
events from the bpf program can be completely customized.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Reorganize examples into subdirectories

Examples directory has been growing, so add a bit of organization.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Bump ulimit for locked memory in test wrapper

This follows from a change in upstream kernel for counting bpf maps
against the locked memory limit.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Python: fix detach_kprobe()

As open_kprobes contains not fd but pointer to struct perf_reader,
it should call lib.perf_reader_free() instead of os.close()

perf_reader: allow more than one reader to poll

When more than one events are used with attach_kprobe, the
perf_reader_poll() will segfault by buffer overflow like below.

*** buffer overflow detected ***: python terminated
======= Backtrace: =========
/lib/x86_64-linux-gnu/libc.so.6(+0x7338f)[0x7fcc971fb38f]
/lib/x86_64-linux-gnu/libc.so.6(__fortify_fail+0x5c)[0x7fcc97292c9c]
/lib/x86_64-linux-gnu/libc.so.6(+0x109b60)[0x7fcc97291b60]
/lib/x86_64-linux-gnu/libc.so.6(+0x10ac04)[0x7fcc97292c04]
/usr/lib/x86_64-linux-gnu/libbcc.so(perf_reader_poll+0x42)[0x7fcc93969cc2]
/usr/lib/x86_64-linux-gnu/libffi.so.6(ffi_call_unix64+0x4c)[0x7fcc95c2badc]
/usr/lib/x86_64-linux-gnu/libffi.so.6(ffi_call+0x1fc)[0x7fcc95c2b40c]
/usr/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so(_ctypes_callproc+0x48e)[0x7fcc95e425fe]
/usr/lib/python2.7/lib-dynload/_ctypes.x86_64-linux-gnu.so(+0x15f9e)[0x7fcc95e43f9e]
python(PyEval_EvalFrameEx+0x965)[0x499be5]
python(PyEval_EvalCodeEx+0x2ac)[0x4a090c]
python(PyEval_EvalFrameEx+0x7d2)[0x499a52]
python[0x4a1634]
python(PyRun_FileExFlags+0x92)[0x44e4a5]
python(PyRun_SimpleFileExFlags+0x2ee)[0x44ec9f]
python(Py_Main+0xb5e)[0x44f904]
/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xf5)[0x7fcc971a9ec5]
python[0x578c4e]

softirq and hardirq

tcpconnect for IPv4 and IPv6, and make tcpv4connect a trimmed example

perf_reader : add timeout argument to perf_reader_poll

tcpaccept

Fixes for deb/rpm packaging

These fixes are manually included in the released 0.1.7 packages, but
will help with future builds.
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Update tag to v0.1.7

Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

handle return value in kretprobe handler for tcpv4connect

Fix an uninitialized variable access in test_clang

Signed-off-by: Yonghong Song <yhs@plumgrid.com>

Signed-off-by: Yonghong Song <yhs@plumgrid.com>

Make KeyboardInterrupt catch more aggressive

Fixes: #186
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

Fixes for table indexing and clear()

In the case that python process tries to clear() a table object while
the bpf program is still running and doing lookup_or_init, the python
clean() routine will race with the bpf program (and bpf will win in the
case of frequent kprobe hits) causing a hang. Instead, first get the
full list of keys and then delete them rather than rely on the default
popitem implementation.

Fixes: #233
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>

In the case that 0-filled keys are valid in the table, the previous
algorithm did not properly iterate.

The API of the bpf_get_next_key routine is such that the iteration
over a map should start with an invalid key. When a 0 key is valid, this
causes iteration to start anywhere inside the hash table, skipping some
entries. So, add logic to the Iter object to test if the init key is
invalid. If otherwise, try a few alternatives until an invalid key is
found. If none found, raise an exception.

Also adds a test for indexing arrays from 0, which nows works with this
too.

Fixes: #260
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>