Support for truncate primitive #157
Conversation
| @@ -48,6 +48,8 @@ extern packet_out { | |||
| void emit<T>(in T hdr); | |||
| } | |||
|
|
|||
| extern void truncate(in bit<32> length); | |||
There was a problem hiding this comment.
This should probably be in an architecture-specific file, not in core.p4.
There was a problem hiding this comment.
The question is, is this something that should be the same in every architecture that supports it (like other things in core), or something that might be different in different architectures?
|
My main problem is that the notion of "payload" is not well-defined. It makes sense only after the parser has run - the payload starts at the point where the parser has stopped extracting. However, where in the pipeline the parser runs is a question of architecture. Some architectures may have multiple parsers. What happens if you call truncate in the parser? |
|
Perhaps truncate should be a method of packet_out then? But that would
require having packet_out as an argument to other than the deparser, or
having a way of doing things other than emit in a deparser.
…On Wed, Nov 30, 2016 at 10:25 AM, Mihai Budiu ***@***.***> wrote:
My main problem is that the notion of "payload" is not well-defined. It
makes sense only after the parser has run - the payload starts at the point
where the parser has stopped extracting. However, where in the pipeline the
parser runs is a question of architecture. Some architectures may have
multiple parsers.
What happens if you call truncate in the parser?
—
You are receiving this because you authored the thread.
Reply to this email directly, view it on GitHub
<#157 (comment)>, or mute
the thread
<https://github.com/notifications/unsubscribe-auth/AD4c8ekyGSL6zo35Fu5w4bdwsP3_9mxcks5rDb-QgaJpZM4K7Pm2>
.
|
|
After I merge this I plan to move truncate to v1model.p4. |
No description provided.