Event starve. Topic is solved

[Endre]

Hi,

If a core runs a code something like this:

Code: Select all

while (1) {
  select {
    case ifaceA.put(char ch):
      ...
      break;
    case ifaceB.put(char ch):
      ...
      break;
  }
}

There are 2 other cores: A and B. A continuously calls ifaceA.put(), B continuously calls ifaceB.put().
In this case what happens? Can one case block/starve the other? Is there some scheduling implemented in event processing?

[mon2]

http://www.xmos.com/zh/support/examples/AN10016

[Endre]

AN10016 seems to solve this issue but it also adds performance penalty and a little bit bloated xC code.
Isn't it planed in the future to add hardware support for Round-robin like event scheduling?
I guess it would be also good to execute all involved "case"-es, one after the other, when the "select" is unblocked by multiple events.