[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. Thanks for clarification and merging! Repository: rL LLVM CHANGES SINCE LAST ACTION https://reviews.llvm.org/D54258/new/ https://reviews.llvm.org/D54258 ___ cfe-commits mailing list cfe-commits@lists.llvm.org h

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

Anastasia added a comment. In D54258#1528872 , @richardmembarth wrote: > Merging in two weeks is fine for me. No worries. I have just committed this. > My assumption was that accepted patches are merged into upstream in a timely > manner. > Maybe thi

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

This revision was automatically updated to reflect the committed changes. Closed by commit rL362623: [Clang] Fix pretty printing of CUDA address spaces (authored by stulova, committed by ). Herald added a project: LLVM. Herald added a subscriber: llvm-commits. Changed prior to commit: https://r

Re: [PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

On Tue, Jun 4, 2019 at 5:37 AM Richard Membarth via Phabricator wrote: > > richardmembarth added a comment. > > Merging in two weeks is fine for me. > > My assumption was that accepted patches are merged into upstream in a timely > manner. > Maybe this is not how it works? Your assumption is cor

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. Merging in two weeks is fine for me. My assumption was that accepted patches are merged into upstream in a timely manner. Maybe this is not how it works? Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D54258/new/ https://reviews.l

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

aaron.ballman added a comment. In D54258#1524643 , @richardmembarth wrote: > Do you know when this will be merged? I apologize, I wasn't aware you needed this merged on your behalf. I normally would be happy to do so, but I'm currently traveling. Maybe

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. Herald added a subscriber: ebevhan. Herald added a project: clang. Do you know when this will be merged? Repository: rC Clang CHANGES SINCE LAST ACTION https://reviews.llvm.org/D54258/new/ https://reviews.llvm.org/D54258 _

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

aaron.ballman accepted this revision. aaron.ballman added a comment. This revision is now accepted and ready to land. In https://reviews.llvm.org/D54258#1297687, @richardmembarth wrote: > > In https://reviews.llvm.org/D54258#1297191, @Anastasia wrote: > > > > I agree with the change itself... bu

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. > In https://reviews.llvm.org/D54258#1297191, @Anastasia wrote: > > I agree with the change itself... but it's quite annoying that such things > can't be tested. :( Yes, that's a pity :( Is there anything missing so that this can be merged? Repository: rC C

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

Anastasia added a comment. In https://reviews.llvm.org/D54258#1296706, @richardmembarth wrote: > CUDA maps `__shared__` internally also to `__attribute__((shared))`: > > #define __annotate__(a) \ > __attribute__((a)) > #define __location__(a) \ > __annotate__(a) > ... >

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. CUDA maps `__shared__` internally also to `__attribute__((shared))`: #define __annotate__(a) \ __attribute__((a)) #define __location__(a) \ __annotate__(a) ... #define __shared__ \ __location__(shared) My guess is that Clang

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

aaron.ballman added a comment. In https://reviews.llvm.org/D54258#1295158, @richardmembarth wrote: > There are external tools (e.g. hipacc ) that > generate Clang AST. This AST uses `LangAS` annotations and emits incorrect > memory space specifiers for CUDA when pretty-

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. There are external tools (e.g. hipacc ) that generate Clang AST. This AST uses `LangAS` annotations and emits incorrect memory space specifiers for CUDA when pretty-printed. Repository: rC Clang https://reviews.llvm.org/D54258 ___

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

aaron.ballman added a comment. In https://reviews.llvm.org/D54258#1292129, @richardmembarth wrote: > I think it's not so easy to provide such tests for CUDA. > CUDA memory space specifiers are implemented via attributes, e.g. `#define > __shared__ __attribute__((shared))`. > As a result of thi

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. Same problem here: The CUDA memory space specifiers are represented via attributes, e.g. `CUDASharedAttr` and only converted in CodeGen to `LangAS::cuda_shared`. We would need a different frontend that annotates `LangAS::cuda_shared`. Repository: rC Clang ht

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

Anastasia added a comment. Could this be tested using diagnostics that prints the type? Like in test/SemaOpenCL/address-spaces.cl. Repository: rC Clang https://reviews.llvm.org/D54258 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http:

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth added a comment. I think it's not so easy to provide such tests for CUDA. CUDA memory space specifiers are implemented via attributes, e.g. `#define __shared__ __attribute__((shared))`. As a result of this, they are pretty-printed via a different code path. In my example, I call `

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

aaron.ballman added a comment. Can you add tests for this change? We typically have these in Misc by passing `-ast-print`. Repository: rC Clang https://reviews.llvm.org/D54258 ___ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.

[PATCH] D54258: [Clang] Fix pretty printing of CUDA address spaces

richardmembarth created this revision. richardmembarth added reviewers: Anastasia, aaron.ballman. Herald added a subscriber: cfe-commits. The current pretty-printer emits OpenCL-style memory spaces specifiers: __device, __constant, and __shared. The correct CUDA memory space specifiers are: __dev