Request for approval: Allow Java's ELF symtab reader to use separate debuginfo files

Tom Rodriguez Thomas.Rodriguez at Sun.COM
Thu Mar 4 10:54:49 PST 2010


The changes look good to me.  I created 6932270 for this and applied your changes to our workspace.  I'll run some SA specific tests on it first and then push it using JPRT assuming the results are clean.  The webrev of the changes are http://cr.openjdk.java.net/~never/6932270.

tom

On Mar 3, 2010, at 7:18 AM, Andrew Haley wrote:

> On 01/20/2010 05:35 PM, Tom Rodriguez wrote:
>> 
>> On Jan 20, 2010, at 2:16 AM, Andrew Haley wrote:
>> 
>>> On 01/19/2010 07:43 PM, Tom Rodriguez wrote:
>>> 
>>>> So do you want to pursue making the SA work with a stripped
>>>> libjvm.so or is requiring debug info to be installed acceptable?
>>>> Sorry if I made this more complicated than you wanted.  I think your
>>>> original change is fine, but if the norm is being stripped then
>>>> making these tools work in that situation would be good.
>>> 
>>> Clearly it would be.  Java programmers on Linux systems aren't used to
>>> installing debuginfo, so it would be better all round if these tools
>>> were to work without it.  (And besides, the idea that the Windows port
>>> might be better in any way irks me!)  I'll have a look.
>> 
>> So the brute force approach is to make all vtables visible in the
>> dynamic symbol table.
> 
> This patch does everything, I hope.  It adds all vtables to the
> dynamic symtab, it changes the ELF reader to use the dynamic symtab,
> and it adds a debuginfo reader that works with every GNU/Linux system
> I know of.
> 
> Webrev at http://cr.openjdk.java.net/~aph/100126-hotspot-webrev-3/
> 
> OK to commit?  I guess I'll need a bug ID in the Sun database too.
> 
> This is https://bugs.openjdk.java.net/show_bug.cgi?id=100126
> 
> Andrew.



More information about the hotspot-dev mailing list