Fix structured kernel codegen

[smessmer]

Stack from ghstack:

• Fix structured kernel codegen #49244 Fix structured kernel codegen

see https://fb.quip.com/ceEdANd5iVsO

RegisterMkldnnCPU kernels incorrectly used makeUnboxedOnly() calls to register add_.Tensor kernels. This is because the codegen incorrectly thought they're not c10-full.
This PR fixes that.

Differential Revision: D25500246

[facebook-github-bot]

💊 CI failures summary and remediations

As of commit 7790021 (more details on the Dr. CI page):

---

• 2/2 failures introduced in this PR

---

🕵️ 2 new failures recognized by patterns

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

pytorch_linux_bionic_py3_6_clang9_test (1/2)

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

Dec 11 20:25:33 AssertionError: mypy failed: tools/codegen/gen.py:388:9: error: Missing return statement

Dec 11 20:25:12   test_run_mypy (__main__.TestTypeHints) ... ok (57.181s)
Dec 11 20:25:15   test_run_mypy_strict (__main__.TestTypeHints) ... FAIL (2.631s)
Dec 11 20:25:33   test_type_hint_examples (__main__.TestTypeHints) ... ok (18.650s)
Dec 11 20:25:33 
Dec 11 20:25:33 ======================================================================
Dec 11 20:25:33 FAIL [2.631s]: test_run_mypy_strict (__main__.TestTypeHints)
Dec 11 20:25:33 ----------------------------------------------------------------------
Dec 11 20:25:33 Traceback (most recent call last):
Dec 11 20:25:33   File "test_type_hints.py", line 239, in test_run_mypy_strict
Dec 11 20:25:33     self.fail(f"mypy failed: {stdout} {stderr}")
Dec 11 20:25:33 AssertionError: mypy failed: tools/codegen/gen.py:388:9: error: Missing return statement
Dec 11 20:25:33 Found 1 error in 1 file (checked 11 source files)
Dec 11 20:25:33  
Dec 11 20:25:33 
Dec 11 20:25:33 ----------------------------------------------------------------------
Dec 11 20:25:33 Ran 4 tests in 88.745s
Dec 11 20:25:33 
Dec 11 20:25:33 FAILED (failures=1)
Dec 11 20:25:33 
Dec 11 20:25:33 Generating XML reports...
Dec 11 20:25:33 Generated XML report: test-reports/dist-gloo/TEST-TestTypeHints-20201211202405.xml

pytorch_linux_backward_compatibility_check_test (2/2)

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

Dec 11 19:19:49 The PR is introducing backward incompatible changes to the operator library. Please contact PyTorch team to confirm whether this change is wanted or not.

Dec 11 19:19:49 processing existing schema:  gather(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor[] _1, Tensor _2, int _3) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  scatter(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor _1, Tensor[] _2, int _3) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  reduce_scatter(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor _1, Tensor[] _2) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  alltoall_base(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor _1, Tensor _2, int[] _3, int[] _4) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  alltoall(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor[] _1, Tensor[] _2) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  send(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor[] _1, int _2, int _3) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  recv(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor[] _1, int _2, int _3) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  recv_anysource(__torch__.torch.classes.dist_c10d.ProcessGroup _0, Tensor[] _1, int _2) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  barrier(__torch__.torch.classes.dist_c10d.ProcessGroup _0) -> (__torch__.torch.classes.dist_c10d.Work _0)
Dec 11 19:19:49 processing existing schema:  __init__(__torch__.torch.classes.dist_rpc.WorkerInfo _0, str _1, int _2) -> (None _0)
Dec 11 19:19:49 The PR is introducing backward incompatible changes to the operator library. Please contact PyTorch team to confirm whether this change is wanted or not. 
Dec 11 19:19:49 
Dec 11 19:19:49 Broken ops: [
Dec 11 19:19:49 	aten::nanquantile.scalar_out(Tensor self, float q, int? dim=None, str interpolation="linear", bool keepdim=False, *, Tensor(a!) out) -> (Tensor(a!))
Dec 11 19:19:49 	aten::nanquantile.scalar(Tensor self, float q, int? dim=None, str interpolation="linear", bool keepdim=False) -> (Tensor)
Dec 11 19:19:49 	aten::nanquantile.out(Tensor self, Tensor q, int? dim=None, str interpolation="linear", bool keepdim=False, *, Tensor(a!) out) -> (Tensor(a!))
Dec 11 19:19:49 	aten::nanquantile(Tensor self, Tensor q, int? dim=None, str interpolation="linear", bool keepdim=False) -> (Tensor)
Dec 11 19:19:49 	aten::quantile.scalar_out(Tensor self, float q, int? dim=None, str interpolation="linear", bool keepdim=False, *, Tensor(a!) out) -> (Tensor(a!))
Dec 11 19:19:49 	aten::quantile.scalar(Tensor self, float q, int? dim=None, str interpolation="linear", bool keepdim=False) -> (Tensor)
Dec 11 19:19:49 	aten::quantile.out(Tensor self, Tensor q, int? dim=None, str interpolation="linear", bool keepdim=False, *, Tensor(a!) out) -> (Tensor(a!))
Dec 11 19:19:49 	aten::quantile(Tensor self, Tensor q, int? dim=None, str interpolation="linear", bool keepdim=False) -> (Tensor)

---

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.

This comment has been revised 5 times.

[bhosmer]

@ezyang calling back to thread about explicit plumbing vs. dynamic scoping like I imagine you figured I would