Re: [Qemu-riscv] [PATCH 1/3] decodetree: Allow !function with no input bits

[Peter Maydell]

On Fri, 9 Aug 2019 at 16:41, Richard Henderson
<address@hidden> wrote:
>
> Call this form a "parameter", returning a value extracted
> from the DisasContext.
>
> Signed-off-by: Richard Henderson <address@hidden>
> ---
>  docs/devel/decodetree.rst         |  8 ++++-
>  scripts/decodetree.py             | 54 ++++++++++++++++++++++++-------
>  tests/decode/err_field6.decode    |  5 +++
>  tests/decode/succ_function.decode |  6 ++++
>  4 files changed, 60 insertions(+), 13 deletions(-)
>  create mode 100644 tests/decode/err_field6.decode
>  create mode 100644 tests/decode/succ_function.decode
>
> diff --git a/docs/devel/decodetree.rst b/docs/devel/decodetree.rst
> index 44ac621ea8..ce7f52308f 100644
> --- a/docs/devel/decodetree.rst
> +++ b/docs/devel/decodetree.rst
> @@ -23,7 +23,7 @@ Fields
>
>  Syntax::
>
> -  field_def     := '%' identifier ( unnamed_field )+ ( !function=identifier 
> )?
> +  field_def     := '%' identifier ( unnamed_field )* ( !function=identifier 
> )?
>    unnamed_field := number ':' ( 's' ) number
>
>  For *unnamed_field*, the first number is the least-significant bit position
> @@ -34,6 +34,12 @@ present, they are concatenated.  In this way one can 
> define disjoint fields.
>  If ``!function`` is specified, the concatenated result is passed through the
>  named function, taking and returning an integral value.
>
> +One may use ``!function`` with zero ``unnamed_fields``.  This case is called
> +a *parameter*, and the named function is only passed the ``DisasContext``
> +and returns an integral value extracted from there.
> +
> +A field with no ``unnamed_fields`` and no ``!function`` is in error.
> +
>  FIXME: the fields of the structure into which this result will be stored
>  is restricted to ``int``.  Which means that we cannot expand 64-bit items.
>
> diff --git a/scripts/decodetree.py b/scripts/decodetree.py
> index d7a59d63ac..a2490aeb74 100755
> --- a/scripts/decodetree.py
> +++ b/scripts/decodetree.py
> @@ -195,7 +195,10 @@ class MultiField:
>      """Class representing a compound instruction field"""
>      def __init__(self, subs, mask):
>          self.subs = subs
> -        self.sign = subs[0].sign
> +        if len(subs):
> +            self.sign = subs[0].sign
> +        else:
> +            self.sign = 0
>          self.mask = mask
>
>      def __str__(self):

Is this change to MultiField left over from the previous
implementation ? I was expecting a zero-length MultiField
to now be a bug in this python script (ie we should never
try to create one).

> @@ -245,7 +248,7 @@ class ConstField:
>
>
>  class FunctionField:
> -    """Class representing a field passed through an expander"""
> +    """Class representing a field passed through a function"""
>      def __init__(self, func, base):
>          self.mask = base.mask
>          self.sign = base.sign
> @@ -266,6 +269,27 @@ class FunctionField:
>  # end FunctionField
>
>
> +class ParameterField:
> +    """Class representing a psuedo-field read from a function"""

"pseudo"

> +    def __init__(self, func):
> +        self.mask = 0
> +        self.sign = 0
> +        self.func = func
> +
> +    def __str__(self):
> +        return self.func
> +
> +    def str_extract(self):
> +        return self.func + '(ctx)'
> +
> +    def __eq__(self, other):
> +        return self.func == other.func
> +
> +    def __ne__(self, other):
> +        return not self.__eq__(other)
> +# end FunctionField
>

thanks
-- PMM