This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [Patch] Fix cycle detection and initialization order in dynamicloader
On 6/21/2012 11:35 AM, Jeff Law wrote:
>> There are performance implications for the undefined case of cycles,
>> and I'll mention them here, but it really doesn't matter since
>> good performance would only be achieved by rewriting this code to
>> use a sensibly studied and proven algorithm (or avoid the undefined
>> case of libraries with circular dependencies).
> Realistically I don't think we can completely avoid the cycle problem.
No, I agree. Even if a cycle produces an undefined total ordering it
should not stop the dynamic linker from producing a sensible and stable
ordering given the same set of input.
Cheers,
Carlos.
--
Carlos O'Donell
Mentor Graphics / CodeSourcery
carlos_odonell@mentor.com
carlos@codesourcery.com
+1 (613) 963 1026