<div>Hi Jason,</div>
<div> </div>
<div>Thanks for your reply. </div>
<div> </div>
<div>I am more interested in kdb because,</div>
<div> </div>
<div>It does not required any tools[except for heperterminal] on host side [kgdb requires a linux pc]</div>
<div>Will enter automatically on kernel crash/panic and gives option to dump some memory area.</div>
<div> </div>
<div>And to me </div>
<div>Single stepping is not on the list. </div>
<div>Dis-assembly is of lower priority.</div>
<div> </div>
<div>And </div>
<div>Kdump requires configuration of another kernel.</div>
<div>So kdb and lkcd look to suit our situation. And I have started with kdb and I am yet investigate more on lkcd.</div>
<div> </div>
<div>Thanks once again for your reply.</div>
<div> </div>
<div>Regards</div>
<div>Basavaraj</div>
<div> </div>
<div> </div>
<div> </div>
<div><br> </div>
<div class="gmail_quote">On Fri, Jun 12, 2009 at 11:16 PM, Jason Wessel <span dir="ltr"><<a href="mailto:jason.wessel@windriver.com">jason.wessel@windriver.com</a>></span> wrote:<br>
<blockquote class="gmail_quote" style="PADDING-LEFT: 1ex; MARGIN: 0px 0px 0px 0.8ex; BORDER-LEFT: #ccc 1px solid">
<div class="im">Martin Hicks wrote:<br>> On Fri, Jun 12, 2009 at 06:00:46PM +0900, Basavaraj Dengi wrote:<br>><br>>> Hi,<br>>><br>>> After applying below patches<br>>> kdb-v4.4-2.6.24-common-2.bz2<br>
>> kdb-v4.3-2.4.20-xscale-1.bz2<br>>><br>><br>> Are you really applying these two different patches?<br>> I really don't know what major changes happened between v4.3 and 4.4,<br>> but I'm not really shocked that things don't work.<br>
><br>> You might be on your own to get this working again.<br>> Jason Wessel has been working on getting KB to function as a front-end<br>> for KGDB. He posted patches to LKML about a month ago. They might be<br>
> of interest to you.<br>><br>><br><br></div>I doubt the whole armv7 instruction set was in the old xscale patch.<br><br>In terms of the patch set Martin mentioned, the kdb front end code does<br>happen to work on ARM, but I had extra patches which I did not publish<br>
which contained the logic for software single stepping on ARM. The<br>reason these were never published is that whole mess deserves a complete<br>re-write. There are traps that get placed and removed and ugly extra<br>exception logic in place.<br>
<br>The right way to implement single stepping is to use the concept of out<br>of line instruction emulation such that you can execute the instruction<br>equivalent on a particular kernel thread and the final instruction would<br>
be a trap to indicate the end of the out of line instruction execution.<br><br>This allows you to do things like leave a breakpoint planted or truly<br>single step an individual kernel thread while not planting a breakpoint<br>
in shared kernel code. This is a back burner project however at the<br>current time, unless someone else wants to contribute this logic or port<br>it from gdb.<br><br>You might be better served to consider using kgdboc + gdb, if you want<br>
to do some instruction disassembly.<br><font color="#888888"><br>Jason.<br></font></blockquote></div><br>