-
-
Notifications
You must be signed in to change notification settings - Fork 510
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
[FEATURE REQUEST] http/2 support #947
Comments
howdy! can you elaborate on freezing? it never proceeds/finishes? It proceeds after the robots.txt timeout? something else? |
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions. |
This is happening to me as well on v2.10.0. It says operation timed out for the target url but I can access that url just fine from any other application. Weirdly, if I proxy through burp it works though. I thought it might be something to do with http/2 but I couldn't confirm that. |
hey @devx00 could you provide the command you used? is it sporadic, or consistent? if you have a consistently bugged target that you can share, that'd be super helpful |
Sure let me gather some more data for you. It was happening consistently once it started but prior to that I was able to scan the target. It could have been some option I started using. I will try and find the specific argument if I can. Just updating my computer right now in case it was something at the OS level but I doubt it is since everything else could access it fine. I'll try again in a few minutes. |
no hurry, thanks for helping run this down! |
Ok so I just tried again and I actually can't get it to work at all for this target. The minimal command that fails for me is feroxbuster -u https://www.rei.com/ -a 'Firefox/118.0' The curl https://www.rei.com/ # fails
curl https://www.rei.com/ -H 'User-Agent: Firefox/118.0' # succeeds |
Does rei have safe harbor? |
Ya Im scanning for a BBP. |
perfect, ill look into it |
I may see the problem but I'm not positive. It looks like it doesnt use the supplied headers when performing the connectivity check here Lines 106 to 111 in e3ec3ae
|
A little more context for why I was looking into the connectivity check
|
good thought, but i dont think so. if you check burp, it should go something like
all of them (minus maybe update check) should have the supplied headers |
it looks like a HTTP/2 problem. if i use the rustls backend instead of native-tls, i don't get any timeouts. performance goes down significantly tho. and i know i looked into different tls backends at some point with another guy that was submitting a PR. There were reasons we landed on native, but id have to dig up the convo to remember exatly why |
Whats weird is if I proxy it through burp. Literally just add
|
yea, i suspect burp is transparently handling the http/2 part of the conversation on our behalf |
Ya that was my initial thought. Any idea why performance decreases when you try to use http/2 natively? I have a hard time believing that Burp's implementation of http in Java could outperform rust. |
i don't off hand, i use the reqwest library to handle the actual web stuff. i changed otu the backend to (probably) confirm it's http/2 at fault. There may be a way to get http/2 support without a perf hit. I'll need to dig up that old convo (may be on the feroxfuzz repo as well) and look at reqwest's http/2 capabilities. |
actually, nevermind on performance. I was comparing debug build to debug build, which is typically ok as a sniff test. I believe since i was using native-tls in debug, those are already compiled C bindings and reasonably fast. The rustls would have been compiled without optimizations, making it a lot slower |
built with --release and it performs as expected |
ok, here's the thread i was thinking of. rustls probably isn't a good fit for ferox. I'll see what other options are availalbe |
i made the following change and everything seems to work. rei.com scans fine, even withotu the firefox header
|
also confirmed it doesn't break mTLS |
@all-contributors add @RavySena for idea |
I've put up a pull request to add @RavySena! 🎉 |
hi, i'm new here. sorry for bothering. i just found the same issue when i try to bruteforce the hackthebox machine "nibbles", i notice the HTTP/2 problem you guys just mentioned, but this page uses HTTP/1.1 i dont know where goes wrong thus come here searching for help |
Describe the bug
Sometimes on some sites the program freezes right at the beginning of the scan, maybe it's the request for robots.txt that freezes?
To Reproduce
Steps to reproduce the behavior:
Traceback / Error Output
If applicable, add error output to help explain your problem.
Environment:
The text was updated successfully, but these errors were encountered: