[llvm] [BOLT] Fix LongJmp layout estimation (PR #215416)

Adam Bzowski via llvm-commits llvm-commits at lists.llvm.org
Mon Aug 17 02:30:22 PDT 2026


adam-bzowski-arm wrote:

> > Thank you for the review, @rafaelauler, but, unfortunately, I have realized that the patch does not fix the original issue I was investigating.
> > **Original issue**
> > The original motivation for this PR was processing a Chromium build with llvm-bolt, which resulted in several BOLT errors. All observed failures were caused by misestimating the 128 MiB range of AArch64 branch26/call26 instructions. Depending on when the out-of-range value is resolved, the failure appears either during MC emission:
> > ```
> > <unknown>:0: error: fixup value out of range
> > BOLT-ERROR: Emission failed.
> > ```
> > 
> > 
> >     
> >       
> >     
> > 
> >       
> >     
> > 
> >     
> >   
> > or during JITLink:
> > ```
> > BOLT-ERROR: JITLink failed: In graph in-memory object file, section .text: relocation target 0x8400048 (<anonymous symbol>:0x400040 + 0x8000008) is out of range of Branch26PCRel fixup at address 0x400040 ($x, 0x400000 + 0x4)
> > ```
> 
> @adam-bzowski-arm If you have time to debug this bug, here's how I fixed a similar problem: Branch26PCRel related to call (bl target) instruction. You can add more log to dump Target and SignedValue, calculate call site for this relocation.
> 
> ```
> static uint64_t adjustFixupValue(const MCFixup &Fixup, const MCValue &Target,
>                                 uint64_t Value, MCContext &Ctx,
>                                 const Triple &TheTriple, bool IsResolved) {
> .....
> 
>  case AArch64::fixup_aarch64_pcrel_branch26:
>  case AArch64::fixup_aarch64_pcrel_call26:
>    // Signed 28-bit immediate
>    if (SignedValue > 134217727 || SignedValue < -134217728)
>      Ctx.reportError(Fixup.getLoc(), "fixup value out of range");
>    // Low two bits are not encoded (4-byte alignment assumed).
>    if (Value & 0x3)
>      Ctx.reportError(Fixup.getLoc(), "fixup not sufficiently aligned");
>    return (Value >> 2) & 0x3ffffff;
> 
> .....
> }
> ```
> 
> In my case the issue was in getSymbolAddress, the target was in the middle of the target function.
> 
> I can try to help you with debug if you can share reproducer.

Thanks for the comment. I already have all the debugging tools, in fact I was wondering if we shouldn't add some extra data to MC emission and JIT Linker to show more than `<unknown>:0: error: fixup value out of range`.

I have the issue sorted out already, I'm just making sure that the solution matches BinaryEmitter one-to-one. And I'm adding LIT tests, which is a little painful.

https://github.com/llvm/llvm-project/pull/215416


More information about the llvm-commits mailing list