There seems to be a reasonably simple solution to this: we don't
guarantee that the test program passes.
While I still don't see it failing, I now at least know why others see
it fail, and I can only say that it's impossible to get the behaviour
people want. When Linux suspends a process, it does so from the signal
handler context, and as such the reader pipe has been closed at that
time.
There is no way this is going to get "fixed" for 2.2.
As such, we can only approach the POSIX people and tell them so, and ask
that the program be dismissed from the test suite, or that Linux gets
dispensation. This is not something special: as far as I know every
single major UNIX vendor has "notes", ie things that they do subtly
differently.
Linus
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/