Several people have reported that dirconc sometimes crashes in its
cleanup phase with a TLB miss on address 0x3, 0x5, or thereabouts.
This seems to be a bug in error-path code that probably wasn't ever
exercised while we were watching. So if you're seeing this problem,
please apply the following change.
Then you get to figure out why your system is issuing unexpected
errors :-)
Index: dirconc.c
===================================================================
RCS file: /disk/disk0/cs161/CVSREPO/os161/src/testbin/dirconc/dirconc.c,v
retrieving revision 1.1
diff -u -r1.1 dirconc.c
--- dirconc.c 15 May 2002 21:20:25 -0000 1.1
+++ dirconc.c 21 Apr 2003 20:49:42 -0000
@@ -133,7 +133,7 @@
case ENOENT:
break;
default:
- say("cleanup: rmdir %s: %s\n",
+ say("cleanup (pid %d): rmdir %s: %s\n",
getpid(), name, strerror(errno));
break;
}
--
- David A. Holland / dholland(a)eecs.harvard.edu
attached now includes up through asst3. asst1 stdev was changed to more
correctly reflect the class distro.
enjoy!
-gwa-
geoffrey werner-allen : 617-694-7261 : www.eecs.harvard.edu/~werner
cs161 : spring 2003 : grade stats : asst0 - asst2
mean median stdev
-----------------------------
asst0 91.3 92.5 6.6
asst1 69.9 84.5 23.0
asst2 82.5 85 8.2
asst3 88.5 90 6.7
Brave cs161ers,
'Here comes brave Sir Robin' (with coconuts in the background)...
Traditionally, there has been a t-shirt made to honor and celebrate the
students who take 161 each year. One of the TFs can coordinate the actual
purchase of the t-shirts, but we would like to poll you to see if anyone
would be interesting in coming up with a design for the shirt. If you are
interested, let us know. Thanks!
-jonathan
161ers,
I am going to reschedule my office hours for 4-6pm today instead of 6-8pm.
They will be in the same location (Science Center lab).
Sasha will have her office hours at her regular time, 8-10pm.
My apologies for the late notice. If this causes you difficulty,
particularly if you are in my section, please let me know and I will
arrange to meet with you individually.
-jonathan
hello cs161ers:
those of you lucky enough to be in jon and i's section are meeting outside
today... walk around MD till you see a little patch of grass that isn't roped
off, that's where we'll be.
see you soon.
-gwa-
I'm on the verge of losing my voice, which would make teaching kind of
difficult. It's not yet clear if this will be a problem tomorrow or
not, but in case it is, if you ordinarily attend my section you might
want to consider making plans to attend another section as well or
instead. (This is meant strictly as a heads-up notice.)
I also just happened to notice a silly bug in the base sfs code. In
sfs_domount, if it finds a bad magic number, it frees the sfs
structure first and then accesses it to report the wrong magic number
it saw. This is clearly broken; the free should be moved after the
kprintf.
--
- David A. Holland / dholland(a)eecs.harvard.edu
range: 50-70
average: 61.8
standard deviation: 6.16
median: 63
This applies to FAS people only - I don't have the extension data on
hand.
--
- David A. Holland / dholland(a)eecs.harvard.edu