Inconsistencies in (de?)referencing buffers for fd_socket_tx writes (send_80211_* functions) about hcxdumptool HOT 6 CLOSED

It looks like these were part of a big change 7 months ago ( 311e093 ) and would explain my (perceived) change in hcxdumptool attack performance around that time. If it was sending "garbage" ( just data/pointers in the GOT? ) instead of actual packet data, that would probably massively impact things.

from hcxdumptool.

You're right. Thanks.
I'll fix that.

from hcxdumptool.

Ugly (nasty) copy and paste error.
Should be fixed, now:
ac555d3

from hcxdumptool.

I ended up doing some tests to see which was the right one... & or not. Both seem to have the same behavior and output (strace to verify). I ended up comparing the output bytecode... and the compiler appears to be smart and results in the same exact code for both situations(&wltxbuffer and wltxbuffer). I would not have expected that.

Anyway, the same issue also appears on some of the memcpys before the writes. For example in send_80211_authenticationresponse (

from hcxdumptool.

Looks like gcc is a way too smart.
But anyway. that need a fix. I'll check that.

from hcxdumptool.

Should be fixed by latest commit, now.

from hcxdumptool.