[PATCH] D59553: [LLD][ELF][DebugInfo] llvm-symbolizer shows incorrect source line info if --gc-sections used

[Alexey Lapshin via Phabricator via llvm-commits]

avl added a comment.

In D59553#1460790 <https://reviews.llvm.org/D59553#1460790>, @ruiu wrote:

> In D59553#1459554 <https://reviews.llvm.org/D59553#1459554>, @avl wrote:
>
> > 1. on some platforms 0 is a valid address. i.e. image base starts from 0. In that case no way to differentiate valid address from invalid one just by checking for 0.
>
>
> But this is also true to UINT64_MAX-2 because you can put your code or data there, no?


yes. that`s true. But using UINT64_MAX-1  to assign addresses looks more like an artificial case. It is natural to start assign addresses from start of address space(which could be 0). It is much less common practice(AFAIK) to assign addresses in the very end of address space.

In the very first version of that path there was a code which analyzes sections for assigned addresses and select an address which does not clash with already assigned module`s ranges. It was suggested to simplify this processing to use UINT64_MAX-1 instead of analyzing address ranges. Assuming that it is really rare case when UINT64_MAX-1 would clash with existing module`s ranges. So using UINT64_MAX-1 is "simple, fast, good enough" solution which works better than 0. i.e. there would be much less cases when addresses would clash for UINT64_MAX-1 than for 0.


CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D59553/new/

https://reviews.llvm.org/D59553