- 12 May, 2016 1 commit
-
-
Zaafar Ahmed authored
-
- 11 May, 2016 5 commits
-
-
4ast authored
Add new example with test case for for unroll, as requested.
-
Brenden Blanco authored
Revert "Force lua main.c to include luajit headers"
-
Brenden Blanco authored
This reverts commit c21b9c03. Instead, use `LUAJIT_DIR=/usr/include/luajit-2.0 cmake ..` or similar as required. Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
-
Valkum authored
-
Valkum authored
Add dns_matching example with test case for for unroll, as requested at iovisor-dev@lists.iovisor.org
-
- 06 May, 2016 13 commits
-
-
Brenden Blanco authored
Lua installation instructions.
-
Nicole Izumi authored
-
Brenden Blanco authored
[RFC] New USDT API
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
Vicent Marti authored
-
- 05 May, 2016 8 commits
-
-
4ast authored
Add check for number of arguments
-
Brenden Blanco authored
sys.exc_value and sys.exc_type should come from sys.exc_info() Convert (decode) char* return values from C functions to a native python str. Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
-
Brenden Blanco authored
There are two problems here: 1. bcc is leaving the extra function parameters in the prototype, when these really just serve as an annotation on the method being traced. They aren't really passed into the function. LLVM lowering phase later errors out on such functions. 2. bcc doesn't limit the size of the argument list, when currently it just supports up to the number of arguments held in registers (6). Fix 1. by rewriting the arguments out of the prototype and into the preamble where they are currently being initialized. Fix 2. by enforcing the limit and returning a more meaningful error message. Added a test case. Extra fix included for string type on python3. Fixes: #497 Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
-
Brenden Blanco authored
Introduce helpers to access pt_regs in an arch-independent manner
-
Naveen N. Rao authored
Convert some of the examples and tools to use the new helpers. Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
-
Brenden Blanco authored
fix llvm 3.9 build
-
Brenden Blanco authored
Signed-off-by: Brenden Blanco <bblanco@plumgrid.com>
-
4ast authored
Force lua main.c to include luajit headers
-
- 04 May, 2016 1 commit
-
-
Brenden Blanco authored
On my system, when both lua and luajit are installed, multiple headers exist in /usr/include. However, my lua installation is 5.3, which has some api-breaking changes, which means that the standard lua.h must not be included. Simply force lua.h to come from luajit-2.0 directory.
-
- 03 May, 2016 4 commits
-
-
4ast authored
Make style-check test non-blocking
-
Brenden Blanco authored
Since the codebase is as of yet inconsistent, disable per-commit style check enforcement. This should be turned back on once the codebase style is consistent.
-
Brendan Gregg authored
Adjust pid filtering/display in runqlat
-
Brenden Blanco authored
The filtering and display of pids in the runqlat tool was not correct. Internally, the kernel keeps pid and tgid, which correspond to thread-id and user process-id, respectively. The runqlat tool was filtering and displaying pid instead of tgid. Change -P and -p options to filter by tgid, and add a new option to give a breakdown by pid (thread-id). Update the docs with the -L option.
-
- 02 May, 2016 8 commits
-
-
4ast authored
Fixes for clang frontend bugs and misc
-
Brenden Blanco authored
-
4ast authored
Powerpc support v2
-
Naveen N. Rao authored
Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
-
Naveen N. Rao authored
__per_cpu_start has a different address on each architecture. Instead, use a arch-unique symbol for the test. Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
-
Naveen N. Rao authored
Basic support for building on powerpc. Tested on ppc64le. Signed-off-by: Naveen N. Rao <naveen.n.rao@linux.vnet.ibm.com>
-
Brenden Blanco authored
-
Brenden Blanco authored
Even after fixing the nested union issue in the preceding commits, this example was failing. The reason was that the kernel was filling in all bits of the bpf_tunnel_key, which in recent kernels includes non-zero fields such as ttl. That non-zero field would break the lookup in the tunkey2if table, which was populated assuming all extra bytes would be zero. Fix it by defining a simpler key structure, and copy the required fields out from the local variable. Fixes: #510
-