>>>>> "Roeland" == Roeland Th Jansen writes: Roeland> who (GNU sh-utils) 2.0 Written by Joseph Arceneaux and David Roeland> MacKenzie. [snip] Roeland> who seems to loose interest in the fact that tty1 is also Roeland> being used. [snip] Roeland> I have seen this with many, if not all 2.4 kernels. this is Roeland> 2.4.6. tty1 _is_ active : I have discussed this with Roeland offline and we are both using XFree 4.02. I start my system with runlevel 3 (and so does Roeland?) and then we run `startx &'. It is only after this that the `utmp' file gets stomped. I am not quite sure how this could be a kernel problem. I don't know where the code would be that touches `utmp'. I have sh-utils versin 2.0 (freshly compiled) with Linux 2.4.7. I also wrote this code, #include #include #include #include #include #include #include int main(int argc, char *argv[]) { int fd; struct utmp loginEntry; fd = open(_PATH_UTMP,O_RDONLY); if(fd >= 0) { while(read(fd,&loginEntry,sizeof(loginEntry))) { printf("type %d, ", loginEntry.ut_type); printf("pid %d, ", loginEntry.ut_pid); printf("line %s, ", loginEntry.ut_line); printf("id %s, ", loginEntry.ut_id); printf("user %s, ", loginEntry.ut_user); printf("host %s\n", loginEntry.ut_host); } close(fd); } else { printf("Cann't open " _PATH_UTMP); } return 0; } I have attached the output of this code and my `ps -ef'. You can see from the ps output that there is a user logged on tty1. I use lkml, but not the sh-utils... feel free to contact me if you wish additional information. I always see this after running `X', never before. regards, Bill Pringlemeir.