[cfe-dev] RFC: Remove uninteresting debug locations at -O0

Adrian Prantl via cfe-dev cfe-dev at lists.llvm.org
Thu Apr 30 20:42:08 PDT 2020



> On Apr 29, 2020, at 2:58 PM, David Blaikie <dblaikie at gmail.com> wrote:
> 
> 
> 
> On Wed, Apr 29, 2020 at 2:52 PM Reid Kleckner <rnk at google.com <mailto:rnk at google.com>> wrote:
> Sure, I can try to elaborate, but I am assuming a bit about what the users desired stepping behavior is. I am assuming in this case that the user is doing the equivalent of `s` or `n` in gdb, and they want to get from one semicolon to the next. If that is not the case, if the user wants the debugger to stop at the location of both getFoo() calls in your example, my suggestion isn't helpful.
> 
> However, since the last dev meeting, I have been thinking about having IR intrinsics that mark the IR position where a statement begins. The intrinsic would carry a source location. The next instruction emitted with that source location would carry the DWARF is_stmt line table flag.
> 
> Making this idea work with optimizations is harder, since the instructions that make up a statement may all be removed or hoisted. To make this work, we would have to establish which instructions properly belong to the statement. This could be done by adding a level to the DIScope hierarchy. The statement markers would remain in place, and code motion would happen around them.
> 
> In this case there would be no intrinsic, just a new kind of "statement" DIScope? yeah, that's the best idea I've had/heard of/discussed for statement grouping so far. Certainly seems imminently prototype-able & then get a sense for just how expensive scoping everything like that would be.
>  
> Late in the codegen pipeline, the first instruction belonging to the most recently activated statement will be emitted with the is_stmt flag.
> 
> On Wed, Apr 29, 2020 at 9:36 AM Adrian Prantl <aprantl at apple.com <mailto:aprantl at apple.com>> wrote:
> Thanks to all of you for sharing your perspective! You all brought up important aspects that I hadn't considered. The argument that really clicked for me was Pavel's that you might want to change the value before the load happens.
> 
> One of my worries here is that what I called "less interesting" locations might delete more interesting ones when instructions and their locations are merged. But if we indeed consider the stack slot loads to be interesting then that is really the best we can do. After all, storing more than source location per instruction would be a UI design nightmare for a debugger.
> 
> Actually I'd kind of love this - I think it'd be really great to describe source ranges instead of source locations - probably super size costly though (& intermediate compiler memory usage, etc). Being able to describe nesting, etc - the same way Clang does with expression/subexpression highlighting I think would be wonderful - imagine if every instruction weren't just attributed to one location, but to a source range with a preferred location (instead of just pointing to the "+" for an add, be able to highlight the LHS and RHS of that plus expression, etc).

I've been thinking about this, too. An extreme point of view is that today's editors are so syntax-aware that they wouldn't need the debugger to tell it where an expression ends, as long as the start point is unambiguous. For example, LLDB already uses clang for syntax highlighting. But having it pre-encoded makes a lot things easier and unambiguous.

For the "+" infix operator example, you really need a third value in addition to the (start, length) that is otherwise sufficient. But in the normal case the length would be very short, so we could probably come up with a cheap encoding for it.

-- adrian 


>  
> 
> 
> Reid, I would like to learn more about what you mean by statement tracking. I'm thinking of a statements as the expressions separated by semicolons, but since my whole example was a single statement I'm assuming you have something more fine-grained in mind. Perhaps you can post an example to illustrate what you have in mind?
> 
> 
> thanks!
> -- adrian

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20200430/85d49726/attachment.html>


More information about the cfe-dev mailing list