Page MenuHomePhabricator

[llvm] Opt-in flag for X86DiscriminateMemOps
ClosedPublic

Authored by mtrofin on Jan 25 2019, 1:11 PM.

Details

Summary

Currently, if an instruction with a memory operand has no debug information, X86DiscriminateMemOps will generate one based on the first line of the enclosing function, or the last seen debug info.

This may cause confusion in certain debugging scenarios. The long term approach would be to use the line number '0' in such cases, however, that brings in challenges: the base discriminator value range is limited (4096 values).

For the short term, adding an opt-in flag for this feature.

See bug 40319 (https://bugs.llvm.org/show_bug.cgi?id=40319)

Diff Detail

Repository
rL LLVM

Event Timeline

mtrofin created this revision.Jan 25 2019, 1:11 PM
mtrofin retitled this revision from [llvm] Opt-in flag for X86DiscriminateMemOps Currently, if an instruction with a memory operand has no debug information, X86DiscriminateMemOps will generate one based on the first line of the enclosing function, or the last seen debug info. This... to [llvm] Opt-in flag for X86DiscriminateMemOpsCurrently, if an instruction with a memory operand has no debug information, X86DiscriminateMemOps will generate one based on the first line of the enclosing function, or the last seen debug info. This....Jan 25 2019, 1:13 PM
mtrofin edited the summary of this revision. (Show Details)
mtrofin added reviewers: dblaikie, jmorse, gbedwell.
mtrofin retitled this revision from [llvm] Opt-in flag for X86DiscriminateMemOpsCurrently, if an instruction with a memory operand has no debug information, X86DiscriminateMemOps will generate one based on the first line of the enclosing function, or the last seen debug info. This... to [llvm] Opt-in flag for X86DiscriminateMemOps.
mtrofin edited the summary of this revision. (Show Details)
mtrofin updated this revision to Diff 183608.Jan 25 2019, 1:16 PM

git commit message update

dblaikie accepted this revision.Jan 25 2019, 1:27 PM

Looks good - thanks!

This revision is now accepted and ready to land.Jan 25 2019, 1:27 PM
This revision was automatically updated to reflect the committed changes.

This review happened without -commits list being subscribed.
(can that be fixes please, so many reviews recently bypass lists, it's raising questions)

This review happened without -commits list being subscribed.
(can that be fixes please, so many reviews recently bypass lists, it's raising questions)

@klimek - Manuel, any idea why this is? I thought generally this had been fixed for a while now.

This review happened without -commits list being subscribed.
(can that be fixes please, so many reviews recently bypass lists, it's raising questions)

@klimek - Manuel, any idea why this is? I thought generally this had been fixed for a while now.

Ugh, i *just* submitted a mail about this, let me add more CC's.