[LLVMdev] [lld] Implementing the aliasing feature

[Nick Kledzik]

On Sep 13, 2013, at 3:35 PM, Shankar Easwaran <shankare at codeaurora.org> wrote:
> This would work only if an alias is another name for the same symbol(weak symbols).
I don’t know what that means. Can you clarify?

> 
> If what is being aliased is another function definition, which is a non zero sized atom, aliasing will not work.
That is the exact scenario I think it *will* work in.  What do you think won’t work.


> 
> I was thinking to model this for ELF for the below functionalities :-
> 
> a) __wrap
> 
> For example : --wrap fn
> 
> What I plan to do here is,
> 
> create a undefined function fn atom
> create a defined weak atom fn
> create a alias reference to __wrap_fn which is a undefined atom.
I don’t see how those steps will achieve wrapping functionality.  Say you are wrapping malloc.  There will be a malloc seen at build time from libc, and all references to malloc will bind to it.  Adding alternate names won’t stop that binding.

Another way to do wrapping (using malloc as the example) is to:
1) introduce (pre-Resolver) an atom (call it X) that is named __real_malloc and has references: 1)to undefined _malloc and 2)undefined __wrap_malloc.
2) add a Pass which get X and looks at the reference to finds what the two references now point to (malloc and __wrap_malloc).  It then walks the whole graph and changes references that point to malloc to point to __wrap_malloc instead, and any references to __real_malloc to malloc. 


> 
> b) --defsym option
> 
> For example : --defsym A = B
> 
> the defsym option is going to have a undefined atom for A
> create a defined weak atom "B" of size 0
> create a alias reference to the expression atom, that defines the expression.
> 
> Both (a), (b) would work if there is an actual function fn, and a symbol A that can override.
> 
> If there is no such symbol, whatever is being aliased to will take preference.
> 
> Thanks
> 
> Shankar Easwaran
> 
> On 9/13/2013 5:26 PM, Nick Kledzik wrote:
>> Can an “alias” atom just be a zero sized atom with one reference of kindLayoutBefore to its target?  The layout engine should then place the alias atom right before the real atom, so you wind up with two symbols at one address.
>> 
>> -Nick
>> 
>> On Sep 12, 2013, at 3:02 PM, Shankar Easwaran <shankare at codeaurora.org> wrote:
>> 
>>> Hi Nick,
>>> 
>>> In addition to what you mentioned, I think there needs a special *AliasReference* that need to be created which the DefinedAtom points to.
>>> 
>>> Thanks
>>> 
>>> Shankar Easwaran
>>> 
>>> 
>>> -------- Original Message --------
>>> Subject:	Re: [PATCH] Add a fallback mechanism for undefined atom.
>>> Date:	Thu, 29 Aug 2013 15:15:49 -0700
>>> From:	kledzik at apple.com <kledzik at apple.com>
>>> Reply-To:	reviews+D1550+public+25c98da67f89e22f at llvm-reviews.chandlerc.com
>>> To:	shankarke at gmail.com, kledzik at apple.com, bigcheesegs at gmail.com, ruiu at google.com
>>> CC:	llvm-commits at cs.uiuc.edu
>>> 
>>> 
>>>   Interesting feature.
>>> 
>>>   What happens if a.o has an undefine foo which says its fallback is bar, but in b.o an undefine for foo says its fallback is baz?  Is that an error, or does each fallback apply to its original usage.  The problem is that the Resolver is currently coalescing UndefinedAtoms based on their name. I does not know that fallbacks need to match too.
>>> 
>>>   Another variant is to have Undefined return a alternate name as a string, rather than returning a new UndefinedAtom object.
>>> 
>>>   Another approach is to use weak aliases.  That is if a.o has an undefine for foo with a fallback of bar, that when parsing that .o file into atoms produces an UndefinedAtom for foo, but also a DefinedAtom with name=foo, weak (mergeAsWeak), hidden (scopeLinkageUnit), size=0, isAlias=true which aliases to bar. So, it any definition of foo does show up, it will override the weak alias.  If not the weak alias will be used and change the reference to bar.
>>> 
>>>   I don't think the isAlias support is all wired up in lld.  But weak aliases are something common in ELF. So would should  get it working, and doing so may provide a way to implement this COFF feature.
>>> 
>>>   One issue is how this interacts with static libraries.  The weak alias for foo make look like there is a definition, so any static archives will not be searched for foo.  That seems different than what you want.  I'm not sure what the gnu linker does about this with weak aliases.  We may need a new resolver option as to how to handle this.
>>> 
>>> 
>>> ================
>>> Comment at: lib/Core/Resolver.cpp:214
>>> @@ +213,3 @@
>>> +      // for COFF "weak external" symbol.
>>> +      const UndefinedAtom *fallbackUndefAtom = undefAtom->fallback();
>>> +      if (fallbackUndefAtom) {
>>> ----------------
>>> You need to re-test:
>>>   if (!_symbolTable.isDefined(undefName)
>>> because searchLibraries() may have already found a definition.
>>> 
>>> 
>>> http://llvm-reviews.chandlerc.com/D1550
>>> _______________________________________________
>>> llvm-commits mailing list
>>> llvm-commits at cs.uiuc.edu
>>> http://lists.cs.uiuc.edu/mailman/listinfo/llvm-commits
>>> 
>>> 
>>> 
>>> 
>>> 
>> 
> 
> 
> -- 
> Qualcomm Innovation Center, Inc. is a member of Code Aurora Forum, hosted by the Linux Foundation