durham_2@husc4.HARVARD.EDU (peter durham) (07/07/87)
>By the way, I created a shell on /t1 with the command "shell i=/t1 &" since >we don't have a login program with "Personal OS-9." Just out of curiosity >I tried running the Level I tsmon and login programs. I once disassembled >login and discovered that it changes the user id number by writing to an >address in low memory using absolute addressing!!! Needless to say this >totally violates the OS-9 philosophy and creates an incredible security hole. >(Well, who ever said OS-9 was very secure in the first place). Of course, on >Level II this hack doesn't work since even absolute addresses don't >necessarily reference the same physical memory locations. I don't know >what it IS changing, but I don't intend to find out the hard way by leaving >myself open to random system crashes. I'm not sure exactly what tsmon >does and whether it could have adverse effects. Walt, Tsmon works fine on level II (though you may want to make one teensy patch: change the parameter length given to the F$Fork call that starts login from 0 to 1; the length of zero seems to confuse C programs, so when I wrote a C version of login this caused interesting but harmless things to happen; anyway this patch is very optional). Login will not, as you said, work on Level II. What login does is sneak into the direct page and directly change the "user number" or "owner" field in its own process descriptor! You will want to patch the code that does this (since you have disassembled it it shouldn't be too hard to find), replacing this hack with a real Level II F$SUser system call. This doesn't require too much changing (I think it's only a very few bytes); unfortunately I don't have the patch with me (it's on CIS, Delphi, and in Kev Darling's book). Hope this helps! Pete Durham UNIX: durham_2@husc4.harvard.edu COMPUSERVE: 73177,1215 seismo!harvard!husc4!durham_2 DELPHI: PEDXING ______ / / / / Tandy / / / / Color Computer 3 and Microware's OS-9 Level II /_/_/_/