RR(S): JDK-7127191 SA JSDB does not display native symbols correctly for transported Linux cores

Staffan Larsen staffan.larsen at oracle.com
Mon Jan 27 05:38:00 PST 2014


Dmitry,

agent/src/os/linux/libproc_impl.c
line 35: it would be good to initialize alt_root to something for clarity.
line 41-46: it looks like this will cause getenv() to be called for every call to pathmap_open if SA_ALTROOT is set. Perhaps better to add an alt_root_initialized variable instead of trying to reuse the alt_root variable for two purposes.
nit: there are some missing spaces after commas and after “while”
line 55: nit: extra space at beginning of line
line 53: rename ptr to alt_path_end?

agent/src/share/classes/sun/jvm/hotspot/debugger/linux/LinuxCDebugger.java
line 69: shouldn’t the second part of the condition be pc.lessThan(base.addOffsetTo(size)) ?

agent/src/share/classes/sun/jvm/hotspot/utilities/soql/sa.js
no comments

Thanks,
/Staffan

On 20 jan 2014, at 17:00, Dmitry Samersoff <dmitry.samersoff at oracle.com> wrote:

> Hi Everyone,
> 
> Please review the fix.
> 
> http://cr.openjdk.java.net/~dsamersoff/JDK-7127191/webrev.01/
> 
> This fix doesn't solve all problems with symbol lookup in SA, but
> address the problem mentioned in bug reports.
> 
> 1. I change algorithm of pathmap_open. Now, it tries to find library
> hardly.
> 
> 2. I decided not to fix broken binary search in loadObjectContainingPC,
> because with less then 20 DSO's we typically have here performance of
> just linear search is approx the same as load, sort, convert to array
> and do binary search.
> 
> -Dmitry
> 
> -- 
> Dmitry Samersoff
> Oracle Java development team, Saint Petersburg, Russia
> * I would love to change the world, but they won't give me the sources.



More information about the serviceability-dev mailing list