Re: [PATCH v4 1/2] protocol: learn to describe fd passing in messages

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



> 
> Add a new type, "unix_fd", used to describe file descriptor sharing via
> socket ancillary data (these messages are local only).
> 
> The marshaller/demarshaller can't serialize this in memory (consume_fd
> implementation is empty), so it is the responsability of the marshaller
> user to handle sending and receiving the handles, which are appended at
> the end of the message with an extra stream byte (because some Unix
> requires sending at least a byte with ancillary data).
> 
> Even if there is no fd to send (or if the fd is invalid etc), the
> receiver side expects an extra byte anyway.
> 
> Signed-off-by: Marc-André Lureau <marcandre.lureau@xxxxxxxxxx>
> ---
>  python_modules/demarshal.py    | 4 ++++
>  python_modules/ptypes.py       | 9 +++++++++
>  python_modules/spice_parser.py | 3 ++-
>  3 files changed, 15 insertions(+), 1 deletion(-)
> 
> diff --git a/python_modules/demarshal.py b/python_modules/demarshal.py
> index 209eafc..2252f37 100644
> --- a/python_modules/demarshal.py
> +++ b/python_modules/demarshal.py
> @@ -72,6 +72,10 @@ def write_parser_helpers(writer):
>              writer.statement("return val")
>              writer.end_block()
>  
> +    writer.function("SPICE_GNUC_UNUSED consume_fd", "int", "uint8_t **ptr",
> True)
> +    writer.statement("return -1")
> +    writer.end_block()
> +
>      writer.newline()
>      writer.statement("typedef struct PointerInfo PointerInfo")
>      writer.statement("typedef void (*message_destructor_t)(uint8_t
>      *message)")
> diff --git a/python_modules/ptypes.py b/python_modules/ptypes.py
> index 7ab2771..9c10b57 100644
> --- a/python_modules/ptypes.py
> +++ b/python_modules/ptypes.py
> @@ -1119,6 +1119,14 @@ class ProtocolType(Type):
>  
>          return self
>  
> +class FdType(IntegerType):
> +
> +    def primitive_type(self):
> +        return "fd"
> +
> +    def c_type(self):
> +        return "int"
> +
>  int8 = IntegerType(8, True)
>  uint8 = IntegerType(8, False)
>  int16 = IntegerType(16, True)
> @@ -1127,3 +1135,4 @@ int32 = IntegerType(32, True)
>  uint32 = IntegerType(32, False)
>  int64 = IntegerType(64, True)
>  uint64 = IntegerType(64, False)
> +unix_fd = FdType(1, True)

Why not overiding the __init__ and just calling a FdType().
Is not clear why 1 and True and actually is not 1 bit but it consumes
0 bytes on the marshalling.
Why signed was True? Not that actually it matters.

So

class FdType(IntegerType):
    def __init__(self):
        IntegerType.__init__(0, False)
        self.name = "fd"

    def c_type(self):
        return "int"

and


unix_fd = FdType()

> diff --git a/python_modules/spice_parser.py b/python_modules/spice_parser.py
> index 97af8b2..db3cc8d 100644
> --- a/python_modules/spice_parser.py
> +++ b/python_modules/spice_parser.py
> @@ -56,6 +56,7 @@ def SPICE_BNF():
>          uint32_    =
>          Keyword("uint32").setParseAction(replaceWith(ptypes.uint32))
>          int64_     =
>          Keyword("int64").setParseAction(replaceWith(ptypes.int64))
>          uint64_    =
>          Keyword("uint64").setParseAction(replaceWith(ptypes.uint64))
> +        unix_fd_   =
> Keyword("unix_fd").setParseAction(replaceWith(ptypes.unix_fd))
>  
>          # keywords
>          enum32_    = Keyword("enum32").setParseAction(replaceWith(32))
> @@ -108,7 +109,7 @@ def SPICE_BNF():
>  
>          # have to use longest match for type, in case a user-defined type
>          name starts with a keyword type, like "channel_type"
>          typeSpec << ( structSpec ^ int8_ ^ uint8_ ^ int16_ ^ uint16_ ^
> -                     int32_ ^ uint32_ ^ int64_ ^ uint64_ ^
> +                     int32_ ^ uint32_ ^ int64_ ^ uint64_ ^ unix_fd_ ^
>                       typename).setName("type")
>  
>          flagsBody = enumBody = Group(lbrace + delimitedList(Group (enumname
>          + Optional(equals + integer))) + Optional(comma) + rbrace)
> --
> 2.5.0
> 
> 

Frediano
_______________________________________________
Spice-devel mailing list
Spice-devel@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/spice-devel




[Index of Archives]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [ECOS]     [Asterisk Internet PBX]     [Linux API]     [Monitors]