Unreliable stepping

[Nukem]

#!

[16:01] <ljames> the issue is basically caused by this commit https://bitbucket.org/mrexodia/titanengine-update/commits/a815753c524780b79d6ba126ea4550cb56a51b7f
[16:01] <ljames> it assumes that
[16:01] <ljames> control-flow is linear, or next EIP is EIP+sizeof(insn)
[16:02] <ljames> so if you have for example
[16:02] <ljames> jmp something
[16:02] <ljames> pushf
[16:02] <ljames> then it will bp after pushf
[16:02] <ljames> and run
[16:02] <ljames> thus if you F7/step into the jmp
[16:02] <ljames> it just runs
[16:02] <ljames> the commit tries to avoid an antidebug trick, but opens up the debugger to a worse exploit

Starting EIP/Step: 77D54167

[Duncan]

Hello,

Thanks to ljames for reporting this bug. I will look into it.

Greetings

[Duncan]

resolved issue #44 (by removing the push ss, pop ss 'fix') + fixed SingleStep

→ <<cset c541b9d764b1>>