hardenedlinux / harbian-qa

Bug hunting through fuzzer/*-sanitizer/etc...
GNU General Public License v3.0
130 stars 19 forks source link

Mismatch of the code example with "Types of branch" #13

Closed daydayup40 closed 5 years ago

daydayup40 commented 5 years ago

The analysis in the following link is very good! https://github.com/hardenedlinux/harbian-qa/tree/master/syzkaller/kstat_demo#types-of-branch But when I follow into the code example, I think that there may be some mismatch with the code and what you mean. Right? If I misunderstand something, please let me know.

Bins94 commented 5 years ago

Could you point out which link is mismatch?

daydayup40 commented 5 years ago

For the first type of branch, the example link https://elixir.bootlin.com/linux/v4.20/source/net/ipv4/tcp.c#L1188 is not pointed to msg_flags. For the second type of branch, the example link https://elixir.bootlin.com/linux/v4.20/source/net/ipv4/tcp.c#L1189 is not pointed to sk_state. And https://elixir.bootlin.com/linux/v4.20/source/net/ipv4/tcp.c#L1231 is not pointed to repair_queue. For the third type of branch, I can't get which local variable is changed by which line form the link https://elixir.bootlin.com/linux/v4.20/source/net/ipv4/tcp.c#L1346 and https://elixir.bootlin.com/linux/v4.20/source/net/ipv4/tcp.c#L1330. Thanks for your help!

Bins94 commented 5 years ago

https://elixir.bootlin.com/linux/v4.20/source/net/ipv4/tcp.c#L1188 is determined by msg_flags, Right? It seems there is no mismatch.

daydayup40 commented 5 years ago

Sorry. It's due to my 360 browser. Its results are not correct. Chrome's results are correct. Thanks!