The due date for design document for Assignment 4 has been moved to
WEDNESDAY, April 10, at midnight. We have decided to move the date, so
that you have a chance to go to the Asst4 section before writing the design doc.
-- Sasha
Just a reminder for those of you who are not in my section but need to
come. My section has moved to Sever 212. Check the web page for a map.
thanks,
-mike
The solution set for assignment 3 has been released and can be gotten
from the assignments area of the course web page.
It's not the world's fastest or most elegant, but it should be
reasonably comprehensible and it seems quite stable.
--
- David A. Holland / dholland(a)eecs.harvard.edu
There seems to have been an off-by-one error in the lhd driver that
kept it from using the last sector on the disk.
Patch follows. (This will be posted on the web site, and will also be
included in the upcoming sol3 release.)
Index: src/kern/dev/lamebus/lhd.c
===================================================================
RCS file: /disk/disk0/cs161/CVSREPO/os161/src/kern/dev/lamebus/lhd.c,v
retrieving revision 1.8
diff -u -r1.8 lhd.c
--- lhd.c 2001/09/21 20:47:59 1.8
+++ lhd.c 2002/04/02 05:28:34
@@ -187,7 +187,7 @@
}
/* Don't allow I/O past the end of the disk. */
- if (sector+len >= lh->lh_dev.d_blocks) {
+ if (sector+len > lh->lh_dev.d_blocks) {
return EINVAL;
}
--
- David A. Holland / dholland(a)eecs.harvard.edu
Greetings --
Midterm is on Thursday! Just a reminder for those who may have been
incapacitated in lecture.
There will be a review section tonight at 8:30 pm in MD 223. We will be
going over the 1999 midterm and fielding general questions.
Office Hours for this week will be tomorrow. They have been posted on the
web site. They are as follows:
Drop-In: Dave, MD 213/209
4 - 5 pm: Sasha, MD 233
6 - 7 pm: Georgi, 2nd Floor
7 - 8 pm: Nick, 2nd Floor
8 - 9 pm: Mike, MD 208
Also, the LFS and FFS papers have been posted, as have the 1999 and 2001
midterms.
-mike
Yesterday, during the midterm review session we gave you an incorrect
explanation of the question why external fragmentation can occur in FFS.
This was one of the questions on last year's midterm. We told you that
external fragmentation can occur because you pre-allocate a chunk of space
for inodes, and then you cannot use this space for data blocks, even if it
is not entirely used by inodes.
Here is the correct explanation (quoting Margo):
> External fragmentation has nothing
> to do with inodes. It's exactly what happens on news partitions.
> You end up allocating a lot of fragments and are unable to allocate
> blocks.
Sorry about the confusion. See, the exam is hard! You have to study.
--Sasha
Some of you mentioned that you don't have the course reader. I just posted
the old midterm exams from 1999 and 2001 on the site under handouts. The
one from 2001 is probably quite similar in format to the one you will do on
Thursday. Looking over the other one is also useful though.
-Georgi
Here is the official word on how much the midterm will cover. Those of you
who don't have a hard copy of the FFS paper can find it at
http://citeseer.nj.nec.com/mckusick84fast.html.
-Georgi
>Date: Mon, 1 Apr 2002 20:51:02 -0500 (EST)
>From: Margo Seltzer <margo(a)eecs.harvard.edu>
>To: matev(a)fas.harvard.edu
>Subject: Re: How much does the midterm cover
>
>Midterm covers everything in class up until break. I.e., FFS but not
>LFS.
>
>- M
As you remember this week there is a midterm. We will be holding one
review session today from 6 to 7:30 PM in MD319, instead of regular Monday
sections. Another sessions will be held on Tuesday, we will announce the
time shortly.
-- Sasha