"successful preparation means successful execution" guarantee neglects bad operation inputs
Under certain circumstances, we guarantee that a prepared model can be
executed successfully. In describing those circumstances, we
neglected to specify that operation input operands must have legal
values for the guarantee to hold. For example, the guarantee doesn't
hold if an ADD operation has an activation input that is not one of
the defined values; or if a RESHAPE operation has a shape input in
which two or more components are -1.
This change modifies the guarantee to apply only when operation input
operands have legal values. It also documents this guarantee for
burst execution.
Note that if an operation has an input operand that can be proven to
have an illegal value at preparation time (e.g., a constant value that
is illegal), model preparation might (but is not required to) fail for
that reason.
Bug: 135933040
Test: $ cd neuralnetworks ; mma
Change-Id: I8b421550dd89e4bbbdae899e7cb5e9e88a46d2fb
diff --git a/neuralnetworks/1.0/IPreparedModel.hal b/neuralnetworks/1.0/IPreparedModel.hal
index 5320050..3dc3202 100644
--- a/neuralnetworks/1.0/IPreparedModel.hal
+++ b/neuralnetworks/1.0/IPreparedModel.hal
@@ -49,11 +49,14 @@
* must not change the content of any of the data objects corresponding to
* 'request' inputs.
*
- * If the prepared model was prepared from a model wherein all
- * tensor operands have fully specified dimensions, and the inputs
- * to the function are valid, then the execution should launch
- * and complete successfully (ErrorStatus::NONE). There must be
- * no failure unless the device itself is in a bad state.
+ * If the prepared model was prepared from a model wherein all tensor
+ * operands have fully specified dimensions, and the inputs to the function
+ * are valid, then:
+ * - the execution should launch successfully (ErrorStatus::NONE): There
+ * must be no failure unless the device itself is in a bad state.
+ * - if at execution time every operation's input operands have legal
+ * values, the execution should complete successfully (ErrorStatus::NONE):
+ * There must be no failure unless the device itself is in a bad state.
*
* Multiple threads can call the execute function on the same IPreparedModel
* object concurrently with different requests.