Hi Ivan
I'm also not a fan of the DIAG implementation of X'75'. And that's why I'm using Jason's original implementation as a non privileged instruction of its own for TK4- Hercules. So, at least, it doesn't flaw DIAG.
Otherwise, I'm completely with you: It doesn't only break all virtualization rules, even in native MVS operations it requires users to play nice, or else beware ;-). But nonetheless: It's currently the only way to bring IP connectivity to MVS 3.8, and given the vast majority of MVS 3.8 systems are in fact single user systems, it's simply good enough for the time being (or, if you so wish, it's _not_ bad enough to require the implementation of a full blown IP stack on MVS 3.8 ;-)).
Cheers
JÃŒrgen
---In H390-***@yahoogroups.com, <***@...> wrote :
Le 7/11/2018 à 5:52 PM, ***@... mailto:***@... [H390-MVS] a écrit :
Hi Ivan
exactly my saying: I _don't_ want it to be able to listen at well known ports of its host, regardless whether this is achieved through not being root, or lacking the required capability. At least, it should not be able to do this without a conscious decision of the user to allow this.
Remember, the OP asked, why FTPD was _not_ able to listen at port 21. And my answer was, "because Hercules isn't able to".
JÃŒrgen,
You are right !
But I always thought myself that DIAG 75 is flawed to start with (it makes DIAG non privileged, has no configuration statement, break any and all virtualization principle)... It's just a makeshift solution...
Only my take, not everyone think the same.
--Ivan