On Wed, 21 Nov 2007 09:41:06 EST "(IBM Mainframe Discussion List)" <[EMAIL PROTECTED]> wrote:
:>In a message dated 11/21/2007 5:31:09 A.M. Central Standard Time, :>[EMAIL PROTECTED] writes: :>:>The manual gives a further explanation: :>:>MVS does not guarantee preservation of the interrupt status of :>:>programs that explicitly disable for I/O and external interrupts :>:>through the STNSM instruction. :>:>What does 'does not guarantee' mean? :>I would presume that should the code be interrupted (PIC-10/11, restart, :>etc.) :>the masking bits in the old PSW may not be preserved and upon redispatch the :>PSW may be enabled for interrupts. :>Basically it means - under MVS, do not use STNSM - take the CPU lock. :>There must be more to it than that. The IBM module IOSVSSCQ (BALR-ed to in :>the expansion of the STARTIO macro) has a STNSM PSASYMSK as its 3rd :>instruction (preceded only by a BASR and AHI in the system I am using). I have :>studied the source code of this module extensively, and remember that later in this :>module the UCB lock for the device involved is obtained. What happens if an :>interrupt occurs between the STNM and the SETLOCK? I don't think that IOS :>would like it if its interrupt status was not preserved. The STARTIO macro :>can be issued by any authorized code whether it is enabled or disabled and :>whether it holds a lock or not. :>Another clue is that in the PSA DSECT the comment on the PSASYMSK byte says :>this: "This field will be used in conjunction with the STNSM instruction to :>place IOS channel scheduler into a disabled state ..." If you scan this :>whole DSECT for "STNSM" you will find several such fields that are used to hold :>the system mask upon doing a STNSM and even more STOSM instructions whose :>operands are the PSA bytes where various STNSMs saved the mask. :>Translation: IBM's use of the STNSM to disable interrupts explicitly within :>the nucleus is rife, so why do they externally document that it should not :>be done? Nucleus is not going to page fault. And if STARTIO fails before it gets the UCB lock, no harm - no foul - since nothing is "in the middle". Again, this is my opinion. -- Binyamin Dissen <[EMAIL PROTECTED]> http://www.dissensoftware.com Director, Dissen Software, Bar & Grill - Israel Should you use the mailblocks package and expect a response from me, you should preauthorize the dissensoftware.com domain. I very rarely bother responding to challenge/response systems, especially those from irresponsible companies. ---------------------------------------------------------------------- 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