Page MenuHomePhabricator

pxli168 (Xiuli PAN)Email Not Verified
User

Projects

User does not belong to any projects.

User Details

User Since
Nov 5 2015, 8:52 PM (185 w, 12 h)

Recent Activity

May 3 2017

pxli168 updated the diff for D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

Fix attribute test case.

May 3 2017, 10:36 PM

Apr 20 2017

pxli168 added a comment to D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

@Anastasia is this test added OK?

Apr 20 2017, 8:36 PM

Apr 19 2017

pxli168 updated the diff for D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

Add attribute test.

Apr 19 2017, 11:50 PM
pxli168 accepted D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

I found that there is a newly added patch about attribute that will fail the test, and I am working on it.
https://reviews.llvm.org/D30009
https://reviews.llvm.org/D32176

Apr 19 2017, 11:25 PM

Apr 10 2017

pxli168 updated the diff for D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

Add duplicate warning test case.

Apr 10 2017, 11:46 PM

Mar 27 2017

pxli168 updated the diff for D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

Fix some varibale up case problems.

Mar 27 2017, 2:39 AM

Mar 19 2017

pxli168 added inline comments to D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.
Mar 19 2017, 7:27 PM

Mar 16 2017

pxli168 updated the diff for D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

Refine some comments.

Mar 16 2017, 12:50 AM
pxli168 added inline comments to D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.
Mar 16 2017, 12:46 AM

Mar 14 2017

pxli168 updated the diff for D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.

Fix some minor issues from bader.

Mar 14 2017, 8:19 PM

Mar 9 2017

pxli168 retitled D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support from intel_reqd_sub_group_size to [OpenCL] Add intel_reqd_sub_group_size attribute support .
Mar 9 2017, 7:19 PM
pxli168 created D30805: [OpenCL] Add intel_reqd_sub_group_size attribute support.
Mar 9 2017, 7:13 PM

Jul 11 2016

pxli168 accepted D21744: [OpenCL] Fix code generation of kernel pipe parameters..

Sorry for the late reply, I was busy with some backend problem.
LGTM, thanks for the fix!

Jul 11 2016, 1:44 AM

Jun 11 2016

pxli168 added a comment to D20133: [OpenCL] Fix __builtin_astype for vec3 types..

I was on a vecation. LGTM, thanks!

Jun 11 2016, 8:28 PM

May 29 2016

pxli168 updated the diff for D20090: [OPENCL] Fix wrongly vla error for OpenCL array..

Make all tests in OpenCL 2.0.

May 29 2016, 10:42 PM
pxli168 updated subscribers of D20682: Expose target-specific pre-linking passes in TargetMachine.
May 29 2016, 8:32 PM
pxli168 updated subscribers of D20681: Add target-specific pre-linking passes to Clang.
May 29 2016, 8:32 PM
pxli168 added a comment to D17438: [OpenCL] Add Sema checks for atomics and implicit declaration.

Yes, I found although the khronos have make a clarify with implicit declarations but they sames to be useful with some program link. And I found some test case about link could not pass by this reason.
But how should we handle the new added spec?

May 29 2016, 7:32 PM
pxli168 accepted D20444: [OpenCL] Include opencl-c.h by default as a clang module.

LGTM!
Thanks!

May 29 2016, 7:29 PM

May 26 2016

pxli168 added inline comments to D20090: [OPENCL] Fix wrongly vla error for OpenCL array..
May 26 2016, 7:41 PM

May 19 2016

pxli168 accepted D17578: [OpenCL]Allowing explicit conversion of "0" to event_t type.

Sorry about late reply.
LGTM!

May 19 2016, 9:49 PM
pxli168 accepted D18369: [OpenCL] Upstreaming khronos OpenCL header file..

LGTM!
I think we may add optimization on this base later.

May 19 2016, 9:48 PM

May 15 2016

pxli168 added a comment to D18369: [OpenCL] Upstreaming khronos OpenCL header file..
May 15 2016, 7:23 PM

May 12 2016

pxli168 accepted D19932: [OpenCL] Add to_{global|local|private} builtin functions..

LGTM

May 12 2016, 2:05 AM

May 10 2016

pxli168 added inline comments to D20090: [OPENCL] Fix wrongly vla error for OpenCL array..
May 10 2016, 7:08 PM
pxli168 added inline comments to D19932: [OpenCL] Add to_{global|local|private} builtin functions..
May 10 2016, 7:02 PM
pxli168 retitled D20090: [OPENCL] Fix wrongly vla error for OpenCL array. from to [OPENCL] Fix wrongly vla error for OpenCL array..
May 10 2016, 1:03 AM
pxli168 added a comment to D19932: [OpenCL] Add to_{global|local|private} builtin functions..

You can update the SUMMARY of this diff as we are now using generic codegen output.
And there are some inline comments about spec.

May 10 2016, 12:05 AM

May 8 2016

pxli168 added a comment to D19932: [OpenCL] Add to_{global|local|private} builtin functions..

The pointer type seems to be not that important, we can just cast it to any type we want.

May 8 2016, 9:16 PM
pxli168 accepted D19780: Output OpenCL version in Clang diagnostics.

LGTM, thanks!

May 8 2016, 9:15 PM
pxli168 added a comment to D18369: [OpenCL] Upstreaming khronos OpenCL header file..

If we want to save some space, could we use some macro to expand the gentype or some script to expand the gentype into each types when the clang is build?

We need to balance between space and readability. When I absorbed attribute((overloadable)) into __const_func to save space, it did not sacrifice readability. However using macro with gentype will cause the header file more difficult to read.

May 8 2016, 8:33 PM

May 4 2016

pxli168 added a comment to D18369: [OpenCL] Upstreaming khronos OpenCL header file..

If we want to save some space, could we use some macro to expand the gentype or some script to expand the gentype into each types when the clang is build?

May 4 2016, 10:25 PM
pxli168 added a comment to D19932: [OpenCL] Add to_{global|local|private} builtin functions..

Could we output a generic function in CodeGen?
This seems to have no big difference to have a lot of declaration in an opencl c header file.

May 4 2016, 10:11 PM

May 2 2016

pxli168 updated subscribers of D19697: SROA assertion: creating bitcast between ptr types with different addr spaces.
May 2 2016, 7:52 PM

Apr 26 2016

pxli168 updated the diff for D19524: [OpenCL] Fix pipe type dump..

Add test and fix some old test.

Apr 26 2016, 9:57 PM
pxli168 added a comment to D19524: [OpenCL] Fix pipe type dump..

I will add a test and send a new version.

Apr 26 2016, 7:50 PM
pxli168 retitled D19524: [OpenCL] Fix pipe type dump. from to [OpenCL] Fix pipe type dump..
Apr 26 2016, 1:41 AM

Apr 25 2016

pxli168 accepted D19478: Remove assert mandating you can only use SPIR target with OpenCL.

LGTM.

Apr 25 2016, 7:56 PM

Apr 13 2016

pxli168 added inline comments to D18369: [OpenCL] Upstreaming khronos OpenCL header file..
Apr 13 2016, 8:42 PM
pxli168 added inline comments to D19071: [OpenCL] Add predefined macros..
Apr 13 2016, 8:21 PM

Apr 11 2016

pxli168 added a comment to D18369: [OpenCL] Upstreaming khronos OpenCL header file..

I think merge them into one file is a good idea.
And the layout by sepc order is OK, is will makes the review easy.

Apr 11 2016, 12:50 AM

Apr 5 2016

pxli168 accepted D17412: PR19957: [OpenCL] incorrectly accepts implicit address space conversion with ternary operator.

LGTM!
Thanks!

Apr 5 2016, 8:29 PM

Apr 4 2016

pxli168 added a comment to D17412: PR19957: [OpenCL] incorrectly accepts implicit address space conversion with ternary operator.
Apr 4 2016, 11:19 PM

Mar 27 2016

pxli168 added a comment to D17438: [OpenCL] Add Sema checks for atomics and implicit declaration.

Hi Anastasia,
https://cvs.khronos.org/bugzilla/show_bug.cgi?id=15603
Now we got clarify from khronos. I will continue on this patch.
BTW, https://cvs.khronos.org/bugzilla/show_bug.cgi?id=15599 is also fixed in the same revision.
Thanks
Xiuli

Mar 27 2016, 11:27 PM

Mar 24 2016

pxli168 updated the diff for D17955: [OpenCL] Fix pipe builtin bugs.

Add test for correct case.

Mar 24 2016, 10:15 PM

Mar 21 2016

pxli168 updated the diff for D17596: [OpenCL] Add ocl and spir version for spir target.

Add test for spir64.

Mar 21 2016, 9:28 PM

Mar 20 2016

pxli168 updated the diff for D17955: [OpenCL] Fix pipe builtin bugs.

Refine other check using the changed diag.
Change test as well.

Mar 20 2016, 11:37 PM
pxli168 added a comment to D17596: [OpenCL] Add ocl and spir version for spir target.
Mar 20 2016, 3:16 AM
pxli168 added inline comments to D17596: [OpenCL] Add ocl and spir version for spir target.
Mar 20 2016, 3:16 AM
pxli168 updated the diff for D17596: [OpenCL] Add ocl and spir version for spir target.
Mar 20 2016, 3:16 AM

Mar 15 2016

pxli168 added inline comments to D17412: PR19957: [OpenCL] incorrectly accepts implicit address space conversion with ternary operator.
Mar 15 2016, 1:44 AM

Mar 9 2016

pxli168 added a comment to D17412: PR19957: [OpenCL] incorrectly accepts implicit address space conversion with ternary operator.

The logic is still to complex and I hope it can be optimized.

Mar 9 2016, 8:03 PM

Mar 8 2016

pxli168 retitled D17955: [OpenCL] Fix pipe builtin bugs from [OpenCL] Fix piep builtin bug to [OpenCL] Fix pipe builtin bugs.
Mar 8 2016, 5:59 AM
pxli168 retitled D17955: [OpenCL] Fix pipe builtin bugs from to [OpenCL] Fix piep builtin bug.
Mar 8 2016, 3:51 AM

Mar 7 2016

pxli168 accepted D17300: llvm-config: fix --libs on Linux.

LGTM!
It fixed the problem that our project could not built with llvm 3.9.
Hope this fix can merge in soon.
Thanks.

Mar 7 2016, 6:53 PM

Mar 3 2016

pxli168 closed D16876: [OpenCL] Refine pipe builtin support.
Mar 3 2016, 11:15 PM
pxli168 updated subscribers of D17821: [OpenCL] Complete image types support.
Mar 3 2016, 11:00 PM
pxli168 updated the diff for D16876: [OpenCL] Refine pipe builtin support.
Mar 3 2016, 9:51 PM

Feb 25 2016

pxli168 added inline comments to D17412: PR19957: [OpenCL] incorrectly accepts implicit address space conversion with ternary operator.
Feb 25 2016, 9:14 PM
pxli168 closed D16040: [OpenCL] Refine OpenCLImageAccessAttr to OpenCLAccessAttr.
Feb 25 2016, 7:17 PM
pxli168 added a comment to D17596: [OpenCL] Add ocl and spir version for spir target.

I think this is just a small change to help identify the which standard llvm-ir is using.

Feb 25 2016, 6:12 PM
pxli168 updated the diff for D17596: [OpenCL] Add ocl and spir version for spir target.
Feb 25 2016, 6:00 PM
pxli168 added a comment to D17438: [OpenCL] Add Sema checks for atomics and implicit declaration.

https://cvs.khronos.org/bugzilla/show_bug.cgi?id=15603
Bug reported.

Feb 25 2016, 1:04 AM
pxli168 updated subscribers of D17596: [OpenCL] Add ocl and spir version for spir target.
Feb 25 2016, 12:48 AM
pxli168 retitled D17596: [OpenCL] Add ocl and spir version for spir target from to [OpenCL] Add ocl and spir version for spir target.
Feb 25 2016, 12:47 AM
pxli168 added inline comments to D17578: [OpenCL]Allowing explicit conversion of "0" to event_t type.
Feb 25 2016, 12:43 AM

Feb 24 2016

pxli168 added a comment to D17578: [OpenCL]Allowing explicit conversion of "0" to event_t type.
foo((event_t)0);

Is above use have been hint by some test cases?

Feb 24 2016, 10:42 PM
pxli168 updated the diff for D16876: [OpenCL] Refine pipe builtin support.
  1. Make new indent and leave space for the incoming OpenCL C++.
  2. Check for the index to see if they are integers.
Feb 24 2016, 9:57 PM
pxli168 removed a reviewer for D16876: [OpenCL] Refine pipe builtin support: rsmith.
Feb 24 2016, 9:32 PM
pxli168 updated the diff for D16040: [OpenCL] Refine OpenCLImageAccessAttr to OpenCLAccessAttr.
Feb 24 2016, 9:28 PM
pxli168 added a comment to D16040: [OpenCL] Refine OpenCLImageAccessAttr to OpenCLAccessAttr.

Remove test case for access quilifier in test/SemaOpenCL/invalid-kernel-attrs.cl.
Due to the patch http://reviews.llvm.org/D17437.
read_only can only be used in parameters with pipe and image type.

Feb 24 2016, 9:21 PM
pxli168 closed D17437: [OpenCL] Add Sema checks for types.
Feb 24 2016, 7:38 PM

Feb 23 2016

pxli168 added a comment to D17438: [OpenCL] Add Sema checks for atomics and implicit declaration.

It now gives a warning for both C99 and OpenCL.
But for target SPIR it gives an err only because it treat unprototyped function as varidic function:

/// \brief Checks whether the given calling convention supports variadic
/// calls. Unprototyped calls also use the variadic call rules.
inline bool sthiupportsVariadicCall(CallingConv CC) {...}

and in OpenCL s6.9.e it says

Feb 23 2016, 6:34 PM
pxli168 removed a reviewer for D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block: pekka.jaaskelainen.
Feb 23 2016, 6:19 PM
pxli168 updated the diff for D17437: [OpenCL] Add Sema checks for types.
Feb 23 2016, 6:19 PM
pxli168 removed a reviewer for D17437: [OpenCL] Add Sema checks for types: pekka.jaaskelainen.
Feb 23 2016, 6:17 PM

Feb 22 2016

pxli168 added inline comments to D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block.
Feb 22 2016, 8:58 PM
pxli168 updated the diff for D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block.
Feb 22 2016, 8:57 PM
pxli168 added a child revision for D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block: D17437: [OpenCL] Add Sema checks for types.
Feb 22 2016, 8:57 PM
pxli168 added a parent revision for D17437: [OpenCL] Add Sema checks for types: D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block.
Feb 22 2016, 8:57 PM
pxli168 updated the diff for D17437: [OpenCL] Add Sema checks for types.

Add array check for types.

Feb 22 2016, 8:56 PM
pxli168 added a comment to D17438: [OpenCL] Add Sema checks for atomics and implicit declaration.

It seems this patch is useless.
The spec does not tell about implicit declaration of function, but now clang with -triple spir will output err if there is implicit declaration of function.
I have read about spir and opencl spec but could not find anything talk about that.
If this is a clang bug then this patch is useless.

Feb 22 2016, 6:46 PM
pxli168 accepted D16928: [OpenCL] Apply missing restrictions for Blocks in OpenCL v2.0.

Then there will be no problem.
I will change my patches as well.
LGTM!
Thanks.

Feb 22 2016, 5:51 PM
pxli168 added inline comments to D17437: [OpenCL] Add Sema checks for types.
Feb 22 2016, 12:46 AM
pxli168 updated the diff for D17437: [OpenCL] Add Sema checks for types.

Refine comments and diag.

Feb 22 2016, 12:46 AM

Feb 21 2016

pxli168 abandoned D16047: [OpenCL] Add Sema checks for OpenCL 2.0.
Feb 21 2016, 10:40 PM
pxli168 added inline comments to D16040: [OpenCL] Refine OpenCLImageAccessAttr to OpenCLAccessAttr.
Feb 21 2016, 10:12 PM
pxli168 updated the diff for D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block.
Feb 21 2016, 10:09 PM
pxli168 added a comment to D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block.

Block is an OpenCL v2.0 feature, I think all test should be handled only for CL2.0 or newer version.

Feb 21 2016, 10:09 PM
pxli168 requested changes to D16928: [OpenCL] Apply missing restrictions for Blocks in OpenCL v2.0.

It seems this patch will check block for CL1.2 or eailer? But the spec reference is for OpenCL v2.0.

Feb 21 2016, 9:02 PM

Feb 20 2016

pxli168 accepted D16928: [OpenCL] Apply missing restrictions for Blocks in OpenCL v2.0.

LGTM!
Thanks

Feb 20 2016, 6:41 PM
pxli168 added a comment to D16928: [OpenCL] Apply missing restrictions for Blocks in OpenCL v2.0.
Feb 20 2016, 6:40 PM

Feb 19 2016

pxli168 retitled D17438: [OpenCL] Add Sema checks for atomics and implicit declaration from to [OpenCL] Add Sema checks for atomics and implicit declaration.
Feb 19 2016, 12:34 AM
pxli168 added reviewers for D17437: [OpenCL] Add Sema checks for types: Anastasia, pekka.jaaskelainen, yaxunl.
Feb 19 2016, 12:18 AM
pxli168 retitled D17437: [OpenCL] Add Sema checks for types from to [OpenCL] Add Sema checks for image and pipe.
Feb 19 2016, 12:16 AM

Feb 18 2016

pxli168 updated D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block.
Feb 18 2016, 11:54 PM
pxli168 retitled D17436: [OpenCL] Add Sema checks for OpenCL 2.0 block from to [OpenCL] Add Sema checks for OpenCL 2.0 block.
Feb 18 2016, 11:53 PM
pxli168 updated the diff for D16040: [OpenCL] Refine OpenCLImageAccessAttr to OpenCLAccessAttr.

Refine the pipe parse to solve the problem the attribute for pipe will be handled twice within Declarator

Feb 18 2016, 8:28 PM
pxli168 planned changes to D16040: [OpenCL] Refine OpenCLImageAccessAttr to OpenCLAccessAttr.

It seems it is related with the pipe type. I am still working on how to fix this problem.

Feb 18 2016, 1:23 AM

Feb 17 2016

pxli168 added a comment to D16040: [OpenCL] Refine OpenCLImageAccessAttr to OpenCLAccessAttr.

Hi Anastasia,
It seems the access qualifier attribute may need to be handled like the you have mentioned in clang-commit that to be in type.
What is you opinion? I can write a patch for that too.

Feb 17 2016, 6:49 PM