OSDev.org

In high level programming languages IDEs , usually there are two common debug commands :
1)trace into
2)trace over .
"trace over" is used to skip the getting into procedure code...
in bochs debugger we have "s" to simulate "trace into" , is there a way to simualte "trace over" , so jumping over "jmp" or "call" instructions

thanks

firas981 wrote: In high level programming languages IDEs , usually there are two common debug commands :
1)trace into
2)trace over .
"trace over" is used to skip the getting into procedure code...
in bochs debugger we have "s" to simulate "trace into" , is there a way to simualte "trace over" , so jumping over "jmp" or "call" instructions

Set a breakpoint after the jmp/call (check for the size to the right of it, the bytes are listed) and continue. Note, known issue, doesn't always break. If you want to be sure, set a number of extra breakpoints at byte intervals after that too.

well , if i didn't mis understand you , yuo say that when returning from call or jmp to some breakpoint then the bochs debugger may not stop at this breakpoint.
yes , this what actually happened with me , although i followed your advice about putting several breakpoints .
is there a solution ?
thanks

firas981 wrote: yes , this what actually happened with me , although i followed your advice about putting several breakpoints .

Not that I know of. It's a bug in bochs debugger, and I've run into some others (only 9 or 10 breakpoints?).

Report it on the bochs bug DB and mark it as high as you can, then expect to not hear anything for half a year.

Or mail it to Christophe Bothamy at [email protected] (not public email addy, remove all the spam and smpa words), who is a bochs developer.

As an aside, bochs has more bugs

In debugging, when you look at a backward short jump (byte EB) it's followed by a signed byte. It's decoded unsigned by the decoder, but executed signed. All backward jumps look weird in disassemblies