Displaying 3 results from an estimated 3 matches for "getmergedlocation".
2017 Feb 10
2
Stripping Debug Locations on cross BB moves, part 2 (PR31891)
...would *not* be fine to use this flag for merged/combined instructions, because in those cases we do not have a correct "original" source location to preserve. In the latter case, we still need to use line-0 for calls, because there's no other correct thing to do.
>
> Probably getMergedLocation should optionally return line-0 instead of null; whether it's an optional don’t-return-null parameter or a separate getMergedLocationForCall API is an implementation detail.
We need something to this end to support getMergedLocationForCall for CallInstructions. I'm not yet sure what the be...
2017 Feb 09
2
Stripping Debug Locations on cross BB moves, part 2 (PR31891)
On Wed, Feb 8, 2017 at 10:32 AM Adrian Prantl <aprantl at apple.com> wrote:
> On Feb 8, 2017, at 10:17 AM, David Blaikie <dblaikie at gmail.com> wrote:
>
>
>
> On Wed, Feb 8, 2017 at 9:56 AM Adrian Prantl <aprantl at apple.com> wrote:
>
> On Feb 8, 2017, at 9:44 AM, David Blaikie <dblaikie at gmail.com> wrote:
>
>
>
> On Wed, Feb 8, 2017 at
2018 Nov 09
5
[RFC] Tablegen-erated GlobalISel Combine Rules
...hers. Here's the current thinking on it though.
Because, we're using MIR it becomes fairly easy to represent debug line information in the rules themselves. Here we expand on the above example to merge the line info for the G_ZEXT and the G_TRUNC into the resulting G_ZEXT using DILocation::getMergedLocation() by simply specifying both debug-location metadata on the new instruction.
def : GICombineRule<(defs root:$D, operand:$S),
(match [{MIR %0 = G_ZEXT %S, debug-location !1
%D = G_TRUNC %0, debug-location !2 }],...