Migrate control_flow.md to the new docs tree

Summary: Move control_flow.md into the new docs tree, update references, and add it to the main table of contents.

Differential Revision: D50612692

Author: dbort

docs/source/getting-started-architecture.md

@@ -32,7 +32,7 @@ Starting from the program source code, below are the steps you would go through
 
 
 * Like all PyTorch use cases, ExecuTorch starts from model authoring, where standard `nn.Module` eager mode PyTorch programs are created.
-* Export-specific helpers are used to represent advanced features like [control flow](https://github.com/pytorch/executorch/blob/main/docs/website/docs/ir_spec/control_flow.md) (for example, helper functions to trace both branches of if-else) and [dynamic shapes](https://pytorch.org/get-started/pytorch-2.0/#pytorch-2x-faster-more-pythonic-and-as-dynamic-as-ever) (for example, data dependent dynamic shape constraint).
+* Export-specific helpers are used to represent advanced features like [control flow](./ir-control-flow.md) (for example, helper functions to trace both branches of if-else) and [dynamic shapes](https://pytorch.org/get-started/pytorch-2.0/#pytorch-2x-faster-more-pythonic-and-as-dynamic-as-ever) (for example, data dependent dynamic shape constraint).
 
 
 ### Export

docs/source/index.rst

@@ -126,6 +126,7 @@ Topics in this section will help you get started with ExecuTorch.
 
    ir-exir
    ir-ops-set-definition
+   ir-control-flow
 
 .. toctree::
    :glob:

docs/website/docs/ir_spec/control_flow.md renamed to docs/source/ir-control-flow.md

@@ -1,10 +1,10 @@
-# (Prototype) Control Flow
+# IR Control Flow
 
 EXIR has a couple of special operators used to help specify control flow within
 some code, similar to jax's control flow operators. Currently these operators
 are only supported for inference.
 
-## torch.ops.higher_order.cond
+## `torch.ops.higher_order.cond`
 
 The `cond` function represents an “if” statement in other programming languages.
 It can logically be seen as implemented as follows:
@@ -105,10 +105,7 @@ class GraphModule(torch.nn.Module):
             return sin
 ```
 
-**See examples of advanced usage of `cond()` operator in ExportDB: [cond tag](https://www.internalfb.com/intern/staticdocs/exportdb/cond.html)**
-
-
-## torch.ops.higher_order.map
+## `torch.ops.higher_order.map`
 
 The `map` function is similar to Python's builtin `map`, where it represents
 applying an operation in the first dimension of a tensor.
@@ -171,31 +168,30 @@ class GraphModule(torch.nn.Module):
             return [%add]
 ```
 
-**See examples of advanced usage of `map()` operator in ExportDB: [map tag](https://www.internalfb.com/intern/staticdocs/exportdb/map.html)**
+## `torch.ops.while`
 
-## torch.ops.while
+A while loop is another control flow construct representing a repeated action.
 
-TODO
-<!-- A while loop is another control flow construct representing a repeated action.
+TODO: Add more details.
 
+<!--
+TODO:
 
 #### Representation in FX
 
-
 ```
 %name = call_function[target = exir.while_loop](args = (condition, body, init_val), kwargs = {})
 ```
 
-
 Above, both condition and body are “functions” represented by GraphModule. The semantics of this node is interpreted as, while `condition` is true, keep executing `body`. The return value of this node is the last val produced by `body` (i.e. the one that failed `cond`).
 
 The implementation of exir.while_loop matches this description:
 
-
 ```
 def while_loop(condition, body, init_val):
     val = init_val
     while cond(*val):
        val = body(*val)
     return val
-``` -->
+```
+-->

docs/website/docs/export/00_export_manual.md

@@ -26,6 +26,6 @@ issues, please file an issue on Github and tag "export".
   - [Modules and Entrypoints](./modules_and_entrypoints.md)
   - [Constraints API](./constraint_apis.md)
   - [Exported Program](../ir_spec/00_exir.md#exportedprogram)
-  - [Control Flow Operators](../ir_spec/control_flow.md)
+  - [Control Flow Operators](https://pytorch.org/executorch/stable/ir-control-flow.html)
   - [Custom Operators](./custom_operators.md)
 - [ExportDB](./exportdb.md)

docs/website/docs/export/background.md

@@ -39,7 +39,8 @@ Python feature, we strive to provide a workaround as part of the error message.
 Over time, we expect to fill in such gaps. On the other hand, not being able to
 decide which control flow path to continue tracing on is a necessary limitation
 of the compiler. You are required to use special operators to unblock such
-cases. See [Control Flow Operators](../ir_spec/control_flow.md).
+cases. See [Control Flow
+Operators](https://pytorch.org/executorch/stable/ir-control-flow.html).
 
 ## Shapes
 

docs/website/docs/export/export_api_reference.md

@@ -2,5 +2,5 @@
 
 - [Modules and Entrypoints](./modules_and_entrypoints.md)
 - [Constraints API](./constraint_apis.md)
-- [Control Flow Operators](../ir_spec/control_flow.md)
+- [Control Flow Operators](https://pytorch.org/executorch/stable/ir-control-flow.html)
 - [Custom Operators](./custom_operators.md)

docs/website/docs/tutorials/exporting_to_executorch.md

@@ -38,10 +38,10 @@ The entrypoint to ExecuTorch is through the `torch._export.capture_pre_autograd_
 to fully capture a PyTorch Model (either `torch.nn.Module` or a callable) into a
 `torch.fx` graph representation.
 
-In order for the model to
-be fully captured into a graph, it may require minor modifications to the user
-code, such as using [control flow operators](../ir_spec/control_flow.md) in
-place of if-statements to capture dynamic behavior. To learn more about what
+In order for the model to be fully captured into a graph, it may require
+minor modifications to the user code, such as using [control flow
+operators](https://pytorch.org/executorch/stable/ir-control-flow.html) in place
+of if-statements to capture dynamic behavior. To learn more about what
 is supported in the export flow, you can take a look at the examples in
 [exportdb](https://pytorch.org/docs/main/generated/exportdb/index.html).
 [Custom operators](./custom_ops.md) can also be traced into the graph