david@ci-dandelion.UUCP (David M. Watson) (08/27/87)
Before we converted our 11/750 to ULTRIX-32 v2.0, we ran two vs100's without problems, although DEC did not, and still does not support that device for ULTRIX. From the first time we booted ULTRIX 2.0, the autoconf has reported that the vs's didn't interrupt, so I put a couple of printf's in the driver probe routine and found that the autoconf appears to lie -- the probe is returning normally. Therefore, I assume that the autoconf is looking for a different sort of return value from the probe than the value which indicated success in 1.2 (sizeof(struct vsdevice)). Is this true? What value is autoconf looking for as an indication that the device actually did interrupt? Or have I jumped to some wrong conclusions? -David Watson Cognition, Inc. (617) 667-4800 900 Technology Park Drive Billerica, MA 01821
chris@mimsy.UUCP (08/28/87)
In article <3244@ci-dandelion.UUCP> david@ci-dandelion.UUCP (David M. Watson) writes: >... From the first time we booted ULTRIX 2.0, the autoconf has reported >that the vs's didn't interrupt, so I put a couple of printf's in the >driver probe routine and found that the autoconf appears to lie -- the >probe is returning normally. `did not interrupt' means not that probe() returned zero, but rather that `cvec' was not altered. cvec is a global variable in Ultrix 2.0, so code that says register int br, cvec; br = 0x15; cvec = <expr>; will fail until the declarations for br and cvec are removed. -- In-Real-Life: Chris Torek, Univ of MD Comp Sci Dept (+1 301 454 7690) Domain: chris@mimsy.umd.edu Path: seismo!mimsy!chris
das@eplunix.UUCP (David Steffens) (08/28/87)
in article <8223@mimsy.UUCP>, chris@mimsy.UUCP (Chris Torek) says: > ... cvec is a global variable in Ultrix [2.0] ... OK, I'll bite... anyone know why the change was made? Just another example of creeping DECism or is there a *good* reason? -- {harvard,mit-eddie,think}!eplunix!das David Allan Steffens 243 Charles St., Boston, MA 02114 Eaton-Peabody Laboratory (617) 573-3748 Mass. Eye & Ear Infirmary
wyatt@cfa.harvard.EDU (Bill Wyatt) (08/29/87)
>> ... cvec is a global variable in Ultrix [2.0] ... > > OK, I'll bite... anyone know why the change was made? > Just another example of creeping DECism or is there a *good* reason? I can't resist putting in my 2 cents worth... I have no idea why this change was made, but I always though the older way (arrange for R10 and R11 to have the cvec & br) was pretty kludgey. The new way is more in keeping with proper C, and perhaps easier to move to other architectures. -- Bill UUCP: {seismo|ihnp4}!harvard!cfa!wyatt Wyatt ARPA: wyatt@cfa.harvard.edu (or) wyatt%cfa@harvard.harvard.edu BITNET: wyatt@cfa2 SPAN: 17410::wyatt (this will change in June)