- 23 May, 2008 40 commits
-
-
Ingo Molnar authored
Signed-off-by:
Ingo Molnar <mingo@elte.hu> Signed-off-by:
Thomas Gleixner <tglx@linutronix.de>
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
now that we have a kbuild method for notrace, no need to pollute the C code with the annotations. Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Steven Rostedt authored
Most archs define the string and memory compare functions in assembly. Some do not. But these functions may be used in some archs at early boot up. Since most archs define this code in assembly and they are not usually traced, there's no need to trace them when they are not defined in assembly. This patch removes the -pg from the CFLAGS for lib/string.o. This prevents the string functions use in either vdso or early bootup from crashing the system. Signed-off-by:
Steven Rostedt <srostedt@redhat.com> Signed-off-by:
-
Steven Rostedt authored
PowerPC is very fragile when it comes to use of function names and function addresses. ftrace needs to either use all function addresses or function names (i.e. my_func as suppose to &my_func). This patch chooses to use the names and not the addresses, and makes ftrace consistent. Signed-off-by:
-
Ingo Molnar authored
clean up recent code. Signed-off-by:
-
Steven Rostedt authored
A new check was added in the ftrace function that wont trace if the CPU trace buffer is disabled. Unfortunately, other tracers used ftrace() to write to the buffer after they disabled it. The new disable check makes these calls into a nop. This patch changes the __ftrace that is called without the check into a new api for the other tracers to use, called "trace_function". The other tracers use this interface instead when the trace CPU buffer is already disabled. Signed-off-by:
-
Soeren Sandmann Pedersen authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
no need to backmerge, only affects ftrace-enabled kernels. (which is not the default) Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Steven Rostedt authored
This patch is the correct way to have tracing enabled by default. Signed-off-by:
-
Steven Rostedt authored
Since ftrace touches practically every function. If we detect any anomaly, we want to fully disable ftrace. This patch adds code to try shutdown ftrace as much as possible without doing any more harm is something is detected not quite correct. This only kills ftrace, this patch does have checks for other parts of the tracer (irqsoff, wakeup, etc.). Signed-off-by:
-
Steven Rostedt authored
The ftrace dynamic function update allocates a record to store the instruction pointers that are being modified. If the modified instruction pointer fails to update, then the record is marked as failed and nothing more is done. Worse, if the modification fails, but the record ip function is still called, it will allocate a new record and try again. In just a matter of time, will this cause a serious memory leak and crash the system. This patch plugs this memory leak. When a record fails, it is included back into the pool of records to be used. Now a record may fail over and over again, but the number of allocated records will not increase. Signed-off-by:
-
Steven Rostedt authored
Some fixes for better output with the trace pipe. Signed-off-by:
-
Ingo Molnar authored
should be an irq-safe lock ... Signed-off-by:
-
Ingo Molnar authored
for ad-hoc tracing. Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
Signed-off-by:
-
Ingo Molnar authored
remove idx syncing - it's expensive on SMP. Signed-off-by:
-
Ingo Molnar authored
implement globally synchronized, fast and scalable time source for tracing. Signed-off-by:
-
Ingo Molnar authored
rename and uninline now() to ftrace_now(). Signed-off-by:
-
Ingo Molnar authored
no code changed. Signed-off-by:
-
Ingo Molnar authored
we can hold all cpu trace buffer locks at once - put each into a separate lock class. Signed-off-by:
-
Steven Rostedt authored
This patch creates a file called trace_pipe in the tracing debug directory. This file is a consumer of the trace buffers. This means that reads of this file consumes the entries from the trace buffers so that they will not be read a second time, as contrast to the static buffers latency_trace and trace. Reading from the trace_pipe will remove the entries from trace and latency_trace too. The advantage that trace_pipe has is that it can record live traces. It will block when there is nothing in the buffer, and read the entries as they are entered. An EOF happens when tracing is disabled (tracing_enabled = 0). Signed-off-by:
-
Steven Rostedt authored
Later patches will need to print the same things as the seq output does. But those outputs will not use the seq utility. This patch adds a buffer to the iterator, that can be used by either the seq utility or other output. Signed-off-by:
-
Steven Rostedt authored
This patch changes the way the CPU trace buffers are handled. Instead of always starting from the trace page head, the logic is changed to a producer consumer logic. This allows for the buffers to be drained while they are alive. Signed-off-by:
-
Steven Rostedt authored
The tests may leave stuff in the buffers. This resets the buffers after each test is run. If a test fails, it does not reset the buffer to avoid touching a buffer that is corrupted. Signed-off-by:
-
Steven Rostedt authored
If the trace buffer is detected to be corrupted, then we disable all tracers. Signed-off-by:
-
Ingo Molnar authored
fix time offset calculations and ordering, plus make code more consistent. Signed-off-by:
-
