[TensorExpr] Remove dtype_ and add buf_ fields to CodeGen::BufferArg.

[ZolotukhinM]

Stack from ghstack:

• [TensorExpr] Add methods for inspecting generated code in TensorExprKernel. #57074 [TensorExpr] Add methods for inspecting generated code in TensorExprKernel.
• [TensorExpr] Remove dtype_ and add buf_ fields to CodeGen::BufferArg. #57382 [TensorExpr] Remove dtype_ and add buf_ fields to CodeGen::BufferArg.
• [JIT] Add a pass for removing a first (self) argument from a graph if it is unused. #57169 [JIT] Add a pass for removing a first (self) argument from a graph if it is unused.
• [JIT] Add a pass for annotating graph with input types derived from sample inputs. #57076 [JIT] Add a pass for annotating graph with input types derived from sample inputs.

BufferArg is used to describe parameters passed to the codegen: it
indicates whether the parameter is a var or a buf and holds a pointer to
the corresponding var/buf. Both var and buf contain dtype, and thus
duplicating it in BufferArg is unnecessary - we can always get it from
the var/buf. Hence we're removing dtype_ field from BufferArg in this
PR. We're also adding a buf_ field here: this is done so that
BufferArg truly has all the info about the parameter.

Differential Revision: D28128329

[facebook-github-bot]

💊 CI failures summary and remediations

As of commit f81f2a4 (more details on the Dr. CI page):

---

• 3/3 failures possibly* introduced in this PR
  • 1/3 non-scanned failure(s)

---

🕵️ 1 new failure recognized by patterns

The following CI failures do not appear to be due to upstream breakages:

pytorch_xla_linux_bionic_py3_6_clang9_test (1/1)

Step: "Run tests" (full log | diagnosis details | 🔁 rerun)

May 03 23:20:26 ERROR [1.705s]: test_clamp_xla_int64 (__main__.TestDevicePrecisionXLA)

May 03 23:20:26 	PyRun_FileExFlags
May 03 23:20:26 	PyRun_SimpleFileExFlags
May 03 23:20:26 	Py_Main
May 03 23:20:26 	main
May 03 23:20:26 	__libc_start_main
May 03 23:20:26 	
May 03 23:20:26 *** End stack trace ***
May 03 23:20:26 
May 03 23:20:26 
May 03 23:20:26 ======================================================================
May 03 23:20:26 ERROR [1.705s]: test_clamp_xla_int64 (__main__.TestDevicePrecisionXLA)
May 03 23:20:26 ----------------------------------------------------------------------
May 03 23:20:26 Traceback (most recent call last):
May 03 23:20:26   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_device_type.py", line 297, in instantiated_test
May 03 23:20:26     raise rte
May 03 23:20:26   File "/opt/conda/lib/python3.6/site-packages/torch/testing/_internal/common_device_type.py", line 292, in instantiated_test
May 03 23:20:26     result = test_fn(self, *args)
May 03 23:20:26   File "/var/lib/jenkins/workspace/xla/test/../../test/test_torch.py", line 7500, in test_clamp
May 03 23:20:26     actual = x[..., :1].clamp(lb, ub)
May 03 23:20:26 RuntimeError: /var/lib/jenkins/workspace/xla/third_party/tensorflow/bazel-tensorflow/tensorflow/compiler/xla/xla_client/debug_macros.h:27 : Check failed: status.status() == ::tensorflow::Status::OK() (Invalid argument: Input dimension should be either 1 or equal to the output dimension it is broadcasting into; the 1th operand dimension is 50, the 1th output dimension is 1. vs. OK)
May 03 23:20:26 *** Begin stack trace ***

---

1 failure not recognized by patterns:

Job	Step	Action
pytorch_linux_xenial_py3_6_gcc5_4_test	Run tests	🔁 rerun

---

This comment was automatically generated by Dr. CI (expand for details).

Follow this link to opt-out of these comments for your Pull Requests.

Please report bugs/suggestions to the (internal) Dr. CI Users group.

Click here to manually regenerate this comment.

[facebook-github-bot]

@ZolotukhinM merged this pull request in 030692c.