When going over procfs, I used a deprecated method. I'll go over the current way to access the proc on Friday.
xtime is also deprecated. I'll update the writeup Wednesday, but you should look in linux/timekeeping.h and use struct timespec and current_kernel_time().
Wednesday, September 16th
The deadline for the first project has been pushed back a week to Monday, September 28th. This will not affect when the other projects are due.
The systems group is having trouble with our class taking down linprog due to too many processes
Add "limit maxproc 15" to your .tcshrc file
If you are having problems creating new processes
create a new connection to linprog
use "ps -u $user" to see your processes
issue "killall -STOP pid1 pid2 ..." for each of those processes
issue "killall -KILL pid1 pid2 ..." for the same processes
Monday, September 14th
Apparently, MCH 202 has a class at 1:30pm. Since there's a conflict there and in LOV 301, this recitation slot will not meet and the hour will be absorbed as office hours. For those in this slot, try to go to one of the other recitations if possible. If not, there's always my office hours and email. Sorry for the trouble.
Monday, September 7th
I've made some errors in the writup/lectures:
First, you should not pathname expand each argument in the command. I said this because I was trying to save you the hassle of having a special case for cd and other built-in commands. Keep in mind, though, that you do need to look for environmental variables in all cases.
For external commands, the argument paths will resolve automatically if needed, but you do need to resolve the command pathname based on the rules mentioned in class.
"ls -la" should expand to "/bin/ls", "-la", NULL.
For cd, you do not resolve the pathname of the command, but you do need to handle expanding the path argument.
"cd ~/assignments" should expand to "cd", "/home/faculty/cop4610t/assignments", NULL.
For other built-in commands, you do not resolve the pathname of the command, and you do need to handle expanding the path of the arguments.
"etime sleep 1" should expand to "etime", "sleep" "1", NULL.
Second, the time is seconds and microseconds and the padding is 6 digits instead of 9. This is just a case of me forgetting what the granularity of gettimeofday() was.
Friday, September 4th
Workshops meet at the same times but are now in MCH 202. This is due to a scheduling conflict with CDA 3101.