Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sun-barr!lll-winken!lll-crg.llnl.gov
From: booloo@lll-crg.llnl.gov (Mark Boolootian)
Newsgroups: comp.unix.wizards
Subject: Where does dbx get addresses for use with ptrace()?
Message-ID: <92124@lll-winken.LLNL.GOV>
Date: 28 Feb 91 00:31:10 GMT
Sender: usenet@lll-winken.LLNL.GOV
Organization: Lawrence Livermore National Laboratory
Lines: 21
Nntp-Posting-Host: lll-crg.llnl.gov

Can anyone tell me how dbx comes up with addresses to hand to ptrace()?  For
example, when breakpointing at a given line number in a C program, how does dbx 
calculate the address of the first instruction for that line (which is what I
am assuming is handed to ptrace()).  

And, while I'm at it, am I correct to assume that dbx stuffs some illegal 
instruction in place of the instruction at the breakpoint location, replacing
it with the original instruction when a "continue" is issued to the debugger?

I'm asking about dbx on a BSD-derived system.  Is dbx on SYSV different?  I
know that SYSV has the /proc directory (used for reading/writing a process as I
understand it).  Does this correspond to the core image that ptrace() on BSD
systems work with?

Thanks in advance for any info.  I can take replies via email and summarize (and
will do so unless reponses are posted).  

cheers,
mb

booloo@lll-crg.llnl.gov