WIP Add Intel PT tracing support #2471
Conversation
Marcondiro
force-pushed
the
intel_pt
branch
2 times, most recently
from
615b87b
to
b4ae309
Compare
|
I'll rebase on this before making libafl_qemu changes in #2438. |
Marcondiro
force-pushed
the
intel_pt
branch
3 times, most recently
from
138002d
to
121d6ca
Compare
|
What's the benefit of the perf-event-open-sys crate thingy over bindgen? Just double checking / asking stupid questions :) |
|
No worries :)
It has a couple of syscall/ioctl wrappers already available and it
addresses an issue I had with Bindgen and `#define`s. They solved it
already in their `wrapper.h` adding an enum.
Let me know if you want me to go deeper than this answer :D
|
|
Nah sounds good, it's just always good to double-check if new crates are necessary. If it solves issues let's gooo |
| } | ||
|
|
||
| #[inline] | ||
| const fn next_page_aligned_addr(address: usize) -> usize { |
There was a problem hiding this comment.
we should avoid writing this kind of function here i think, we already have libafl_page_from_addr in the qemu repo in C. we should maybe start gathering these utils functions at the same place either in a c (so that it's usable in QEMU as well) or a rs file.
There was a problem hiding this comment.
you're linking them together, anyway, right? So you can write it in rust and still use them from C
There was a problem hiding this comment.
Do you guys know if we lose inlining this way?
There was a problem hiding this comment.
from rust to rust? No. From C to rust, probably not? maybe?
Marcondiro
force-pushed
the
intel_pt
branch
4 times, most recently
from
e65ba68
to
0a8d8ba
Compare
Marcondiro
force-pushed
the
intel_pt
branch
7 times, most recently
from
30489e9
to
354ad86
Compare
Marcondiro
force-pushed
the
intel_pt
branch
2 times, most recently
from
a2990ad
to
8d72014
Compare
Marcondiro
force-pushed
the
intel_pt
branch
from
d26447d
to
736ac50
Compare
Marcondiro
force-pushed
the
intel_pt
branch
from
1d1104f
to
40d434d
Compare
|
Hello @rmalmain @domenukk, if you wish you could start to have a look at the new code in libafl_bolts (interface to intelPT usable in both libafl and libafl_qemu) and at the IntelPT executor hook. I'm still working on cleaning the commandExecutor (and the fuzzer built on top of it), it might be removed from this PR depending on the 0.14 cutoff date |
WIP, any feedback is welcome