Date: Wed, 26 Apr 1995 01:31:20 -0300 (ADT) From: Bill Davidson Subject: Re: Chaining real-mode interrupts To: Eli Zaretskii Cc: djgpp AT sun DOT soe DOT clarkson DOT edu On Tue, 25 Apr 1995, Eli Zaretskii wrote: > I will try to make it more clear, but please also consider what, if > anything, needs amendment after you've solved your problems. If you > cannot do it right by reading the FAQ and the docs, something *must* be > missing. > Thanks for the pointer to pctime12.zip. I examined that code and noted that the only significant difference from mine was that the _go32_dpmi_registers struture whose address is passed to _go32_dpmi_simulate_fcall_iret() was a global, whereas I was allocating an auto in my handler (which is how the info page looks). I changed that, and it worked! Re-ran the program a couple of times to be sure, and the results were a bit erratic :-(. Still, it wasn't seg faulting, so I applied the same fix to my kbd handler and un-commented-out (commented in?) the line that set the RM kbd handler. Program segfaulted on start-up (never even got to the start-up screen). :-( The program sits in a pretty tight event loop, and I noticed that with 'topline' in my GO32 variable the 'P' and 'R' at top left both sat there all the time, just barely flickering. Should this tell me something? It looks like the processor is mode-switching very fast (maybe 18.2 times a second?). Incidentally, just to make sure that the compiler/library were OK I recompiled the pctime12 sources, and the program ran fine, which was really irritating! "My" program is one I am porting, and I am beginning to suspect that there is some weird interaction going on with some other function, although I can't imagine where. When it segfaults I always get _two_ "Segmentation Violation" messages (sometimes more), and symify can't make any sense of the tracebacks (which either show just one address or just two, the bottom one in that case being NULL). *Sigh* I have the depressing feeling that I am going to find the problem here some day, and it is going to be really simple... Thanks for all your help. Bill Davidson PS -- Have you been seeing multiple copies of my/your messages? I just received a "new" copy of my post from Monday.