Re: [PATCH] tracer for sys_open() - sreadahead

["Frédéric Weisbecker" <fweisbec@gmail.com>]

2009/1/29 Ingo Molnar <mingo@elte.hu>:
>
> * Frédéric Weisbecker <fweisbec@gmail.com> wrote:
>
>> 2009/1/29 Ingo Molnar <mingo@elte.hu>:
>> >>
>> >> Several people talked me about utrace and gave some examples about it in
>> >> this discussion. The Api is very convenient to fetch syscall numbers,
>> >> arguments and return values. And the hooks are done in the generic core
>> >> code, so it is arch independent.
>> >>
>> >> The only drawback I can see is that it is not yet merged upstream, in
>> >> need of in-kernel users. If it only depends on this condition, we could
>> >> be these users...
>> >>
>> >> What do you think?
>> >
>> > sure - how do the minimal bits/callbacks look like which enable syscall
>> > tracing?
>> >
>> >        Ingo
>>
>>
>> There is a very straightforward example provided by Ananth in there:
>> http://lkml.org/lkml/2009/1/28/59
>
> I mean, how does the infrastructure patch look like - what code does this
> add to the kernel - just to get the syscall tracing bits. Lets get some
> progress here - it's clear that tracing syscalls is good, we just need to
> do it and look at actual patches.
>
>        Ingo
>

The latest snapshot version I've found is here:
http://people.redhat.com/roland/utrace/2.6-current/utrace.patch
This is mostly independent core code and a good number of hooks inside ptrace.

But I don't know much about the overhead it potentially brings on ptrace.