>The biggest obstacle for privileged programs is that SRB mode is more >restrictive than task mode. Not all services are allowed in SRB mode; many >functions require a TCB. Error recovery for an SRB is nowhere near as flexible >as for TCB. Also, debugging is not as easy for SRB as for TCB.
I can't really count the times when I had to explain srb-to-task-percolation to some vendor scheduling srbs and not being able to debug, much less fix their 0C4s! >There is not even >a control block equivalent of the TCB where the operating system keeps >information about an SRB while it's running! I don't understand. Suspended SRBs all have their SSRB, and there's the IHSA. The state of the SRB while it is actually dispatched can be seen in the cpu status and PSA, reliable with regards to storage only in an sadump. Or has that substantially changed while I wasn't looking? >In light of these restrictions, I suppose one could argue that more difficult >programming means less stable programs. Exactly. And less people knowing how to figure out what went wrong or even to imagine what might go wrong! Just look at all the discussions here on ibmmain about this exact topic. Best regards, Barbara Nitz ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
