As of 4/24/97 there have been about 300 bugs pulled from swdriver. 5/19: 2 swdriver bugs more to date 2 bug fixes in IOGEN-SCSI-CONFIG.BLI 1 bug fix in LOAD_DRIVER.BLI Also present IOGEN-CONNECT.BLI SWCTL.C SYSMANIOGEN.C MSCP.MAR IOSUBNPAG.MAR No errs found in these but they build now. Functions: SWDRIVER - switching driver SWCTL - setup, control, and report. Now has code for all its design functions. MSCP - Respond to dev$m_noclu when deciding whether to allow a connection to a device. SWdriver varies this depending on whether its path is local. SYSMANIOGEN - Used in sysman io connect. Now must issue SCSI INQUIRY and set HSZ alloc. class if needed for driver load. IOGEN* and LOAD_DRIVER - Look at inquiry data for autoconfig and set the HSZ alloc class if needed. Must be done prior to load driver call. Load_driver is altered so that an alloc class biased by 40000000 hex is treated as an alloc class without the bias for setup of DDT alloc class and NOT setting up the PAC bit. We must ensure a DDT is created if needed in cases of an HSZ70 etc. on the bus. IOGEN-SCSI-CONNECT has most of the new work in this area to get the HSZ alloc class out of INQUIRY data if present. 5/55 4 bugs iogen-ssi-config.bli (around char literals). Still seems to have some probs not finding swa0 the 2nd time thru. Note that some of the PAC naming issues may affect the HSZ devices also where collisions with other bus names are concerned. Thus it may be needed to flag device alloc. class separately and get into more of the stuff HallyB did... 5/23 2 bugs in load_driver, 1 in IOGEN-SCSI-CONFIG.BLI. 5/27 1 bug in IOGEN-SCSI-CONFIG.BLI 5/30 2 more bugs in iogen and in adding use of IRPE instead of sw-based queue for i/o context info I got rid of 20 errors in the course of getting the code working. (16 were type system related, 4 were fixing C compiler scaling.) 6/2 2 bugs removed in swdriver 6/18 - another 12 bugs in swdriver, 4 in swctl, 2 in testswitch (routines new in iosubnpag split off are testswitch) removed. Also removed 4 bugs in iogen-scsi-config in the process of ensuring the load code is present for CPU models that use the asynch.routines. Removed bug in printf calls in swctl also. At this point swserver exists and compiles without error and builds ok. 6/24 6 bugs out of swdriver (outstdg count not maintained everywhere same, mainly. Also needed to post with _nocnt routine). Added device name report to swctl and added packack call after chnl switch in swctl. 5 bugs out. 6/25 3 bugs out of swdriver. Moved UCB around so vicddt is 8byte aligned, fixed IRP ownership & posting in startio(). SWdriver interoperates with JFdriver indicating nested intercepts are ok. 7/2 another 5 bugs in swdriver, 2 in swctl squashed out. ; total 381 est so far