johanw@ttds.UUCP (Johan Wide'n) (05/06/85)
I have recently started to run 2.9BSD on a PDP-11/45. Everything seems to work all right except for ps and adb. I seems as if ps and the kernel have different ideas on how the user structure looks, or as if the user structure does not reside on a 64 byte boundary. Adb and the kernel do not agree on where the registers reside, adb places the contents of 'sp' in 'r1' and so on. I have NOT defined UCB_NET. Has anyone seen this problem before? Any help appreciated. mcvax!enea!ttds!johanw Johan Widen
johanw@ttds.UUCP (Johan Wide'n) (05/18/85)
>I have recently started to run 2.9BSD on a PDP-11/45. Everything seems to >work all right except for ps and adb. I seems as if ps and the kernel have >different ideas on how the user structure looks, or as if the user structure >does not reside on a 64 byte boundary. Adb and the kernel do not agree on >where the registers reside, adb places the contents of 'sp' in 'r1' and so >on. I found the cause of these problems. It turned out that /usr/include/whoami.h was out of date (it was the original file from the boot-tape). Doing cp /usr/sys/MYMACHINE/whoami.h /usr/include and then recompiling adb and ps made the problems go away. As some of the responders pointed out: Make sure that /usr/sys/MYMACHINE/whoami.h is identical to /usr/include/whoami.h and that /usr/sys/MYMACHINE/localopts.h is identical to /usr/include/sys/localopts.h My thanks to all who responded. BTW: We previously run a more or less vanilla v7 on this machine. 2.9BSD makes quite a difference. My compliments to Berkeley for producing this marvelous piece of software. My only gripe so far is that identifiers in C are still only significant in the first 7 characters. On the other hand: you get job control and UCB_SCRIPT and symbolic links and overlays and ... Hot dog! mcvax!enea!ttds!johanw Johan Widen