<div dir="ltr"><div dir="ltr">On Tue, Jun 18, 2019 at 4:45 AM Eli Friedman <<a href="mailto:efriedma@quicinc.com">efriedma@quicinc.com</a>> wrote:<br></div><div class="gmail_quote"><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">> -----Original Message-----<br>
> From: Peter Smith <<a href="mailto:peter.smith@linaro.org" target="_blank">peter.smith@linaro.org</a>><br>
> Sent: Monday, June 17, 2019 3:33 AM<br>
> To: Eli Friedman <<a href="mailto:efriedma@qualcomm.com" target="_blank">efriedma@qualcomm.com</a>><br>
> Cc: llvm-dev <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>><br>
> Subject: [EXT] Re: [llvm-dev] lld symbol choice for symbol present in both a<br>
> shared and a static library, with and without LTO<br>
> <br>
> On Fri, 14 Jun 2019 at 20:58, Eli Friedman via llvm-dev<br>
> <<a href="mailto:llvm-dev@lists.llvm.org" target="_blank">llvm-dev@lists.llvm.org</a>> wrote:<br>
> ><br>
> ><br>
> ><br>
> > If “obj.o” is built with LTO enabled, and the function is specifically a runtime<br>
> function, the behavior is different. For example, suppose the IR contains a call<br>
> to “llvm.memcpy”, and the generated code eventually calls “memcpy”. Or<br>
> suppose the IR contains a “resume” instruction, and the generated code<br>
> eventually calls “_Unwind_Resume”. In this case, the choice is different: lld<br>
> always chooses the “memcpy” or “_Unwind_Resume” from the shared library,<br>
> ignoring the order the files are specified on the command-line. Is this the<br>
> expected behavior?<br>
> <br>
> As I understand it, there is no more selection of members from static<br>
> libraries after the LTO code-generator has run. In the example from<br>
> the PR there is no other object with a reference to memcpy so the<br>
> member containing the static definition is not loaded, leaving only<br>
> the shared library to match against. I would expect if there were<br>
> another reference to memcpy from a bitcode file or another ELF file<br>
> and the static library was before the shared then it would match<br>
> against that.<br>
> <br>
> As to whether this is expected or not, I don't know for certain. One<br>
> desirable property of not selecting more objects from static libraries<br>
> is that you are guaranteed not to load any more bitcode files from<br>
> static libraries, which would either need compiling separately from<br>
> the other bitcode files, or have the whole compilation done again with<br>
> the new objects, which could cause more bitcode files to be loaded<br>
> etc.<br>
<br>
For runtime functions defined in bitcode, we avoid the "double-LTO" scenario you describe by including them in the LTO link even if we can't prove they will be used. This is the handleLibcall code you pointed out. (<a href="https://github.com/llvm-mirror/lld/blob/master/ELF/Driver.cpp#L1733" rel="noreferrer" target="_blank">https://github.com/llvm-mirror/lld/blob/master/ELF/Driver.cpp#L1733</a>). As the comment there describes, we don't do this for runtime functions which are not defined in bitcode, to avoid other side-effects; instead we resolve those symbols after LTO.<br>
<br>
For the scenario I'm describing, though, it looks like the key decision here is made in SymbolTable::addShared, before handleLibcall and LTO. If a symbol is defined in both a static library and a shared library, and we haven't seen a reference to the static library's symbol at that point, we throw away the record of the symbol defined in the static library. <br>
<br>
Ultimately, I guess the question is what alternatives are possible, without breaking the scenarios handleLibcall is supposed to handle. I see a few possibilities here:<br>
<br>
1. Whenever we see any bitcode file, treat it as referencing every possible runtime function, even those defined in non-bitcode static libraries. Then we try to resolve the __sync_val_compare_and_swap_8 issue from <a href="https://reviews.llvm.org/D50475" rel="noreferrer" target="_blank">https://reviews.llvm.org/D50475</a> some other way.<br></blockquote><div><br></div><div>That seems technically doable, but how do we know the names of all possible runtime functions?</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
2. Change the symbol resolution that runs after LTO to use a different symbol resolution rules from normal non-LTO/before-LTO symbol resolution, so it finds the function from the static library instead of the shared library.<br></blockquote><div><br></div><div>We don't actually do any symbol resolution after LTO. We merge a result of LTO to other object files, but no new symbols are expected to appear after LTO. We can change that assumption of course, but that's perhaps too much.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
3. Change symbol resolution in general to prefer "lazy" symbols from static libraries over symbols from shared libraries, even outside LTO. So "static.a shared.so object.o" picks the symbol from static.a, instead of shared.so like it does now.<br></blockquote><div><br></div><div>This change seems risky.</div><div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
4. We WONTFIX <a href="https://bugs.llvm.org/show_bug.cgi?id=42273" rel="noreferrer" target="_blank">https://bugs.llvm.org/show_bug.cgi?id=42273</a> .</blockquote><div><br></div><div>The other option I can think of is to add a command line option to force loading a file from a static archive. With `-u`, we can force loading a member file when a specified name remains undefined after name resolution. That doesn't work for this case because after LTO `memcpy` is not an undefined symbol but a library symbol. So, maybe we can define a new option `-U` to insert a given name as an undefined symbol from the beginning, which forces the linker to load a member file immediately after it finds one.</div></div></div>