[lldb-dev] mmx registers on Linux 32-bit inferior

[Todd Fiala]

I'll go ahead and cook up and add some tests for these.


On Thu, Feb 27, 2014 at 3:28 PM, Todd Fiala <tfiala at google.com> wrote:

> Hey all,
>
> I just put up a patch earlier today to support lldb debugging 32-bit EXEs
> on a 32-bit host (thanks to Matthew Gardiner for all the investigation and
> motivation!)  In the process of fixing that, I saw what looked like it
> might be the root cause of MMX registers being read incorrectly for a Linux
> 32-bit EXE debugged by a 64-bit host lldb.
>
> Is there anybody out there who has a valid MMX debug scenario for a 32-bit
> Linux app that they can test in these two scenarios?
> (1) debugging on a 32-bit Linux host with top of tree lldb.
> (2) debugging on a 64-bit Linux host with top of tree lldb.
>
> Those two code paths will be different now.  I suspect it will work
> correctly in (1) [32-bit exe debugged on 32-bit host].  But it might fail
> on #2.  If it works on #1 and not on #2, I suspect I can now change the
> 32-bit exe debugging path on 64 bit host to mirror the 32-bit host path
> now, which may fix it.  But I don't want to fiddle with it without a valid
> test scenario.
>
> In lieu of actually running the test, I'd be happy to consume somebody's
> MMX code that they know works and can tell me what the values should be at
> certain points in the code and I can run the tests over here.
>
> Thanks!
> --
> Todd Fiala | Software Engineer |  tfiala at google.com |  650-943-3180
>



-- 
Todd Fiala | Software Engineer | tfiala at google.com | 650-943-3180
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/lldb-dev/attachments/20140228/2fa01521/attachment.html>