[PATCH] D71272: [OpenCL] Pretty print __private addr space
Anastasia Stulova via Phabricator via cfe-commits
cfe-commits at lists.llvm.org
Fri Dec 13 05:21:54 PST 2019
Anastasia marked an inline comment as done.
Anastasia added inline comments.
================
Comment at: clang/test/SemaOpenCL/access-qualifier.cl:28
kernel void k1(img1d_wo img) {
- myRead(img); // expected-error {{passing 'img1d_wo' (aka '__write_only image1d_t') to parameter of incompatible type '__read_only image1d_t'}}
+ myRead(img); // expected-error {{passing '__private img1d_wo' (aka '__private __write_only image1d_t') to parameter of incompatible type '__read_only image1d_t'}}
}
----------------
AlexeySotkin wrote:
> Minor. An error message like this looks a bit confusing to me. User might wonder whether parameters are incompatible because of address space or because of access qualifiers. Should it print `passing '__private img1d_wo' (aka '__private __write_only image1d_t') to parameter of incompatible type '__private __read_only image1d_t'`? In this case it is clear that there is a mismatch in access qualifiers.
Yes, I agree. However, we are printing the full `QualType` in both places in this diagnostics, so the problem is that the addr space is either not deduced or being dropped from `QualType` somewhere. I don't think we have got addr spaces working yet for all cases correctly and printing `__private` have revealed a number of such issues. I suggest however to fix them in isolation case by case as we discover them. This commit is already pretty big and I don't want to expand it even more. I have opened a bug to track this issues: https://bugs.llvm.org/show_bug.cgi?id=44294
Hopefully we can fix it asap. Does it make sense?
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D71272/new/
https://reviews.llvm.org/D71272
More information about the cfe-commits
mailing list