https://homeostasis.scs.carleton.ca/wiki/index.php?action=history&feed=atom&title=Operating_Systems_2021F_Lecture_23 2026-04-06T06:18:41Z Revision history for this page on the wiki MediaWiki 1.42.1 https://homeostasis.scs.carleton.ca/wiki/index.php?title=Operating_Systems_2021F_Lecture_23&diff=23579&oldid=prev 2021-12-09T18:45:14Z

<p>Created page with &quot;==Video== Video from the lecture given on December 9, 2021 is now available: * [https://homeostasis.scs.carleton.ca/~soma/os-2021f/lectures/comp3000-2021f-lec23-20211209.m4v...&quot;</p> <p><b>New page</b></p><div>==Video==<br /> <br /> Video from the lecture given on December 9, 2021 is now available:<br /> * [https://homeostasis.scs.carleton.ca/~soma/os-2021f/lectures/comp3000-2021f-lec23-20211209.m4v video]<br /> * [https://homeostasis.scs.carleton.ca/~soma/os-2021f/lectures/comp3000-2021f-lec23-20211209.cc.vtt auto-generated captions]<br /> Video is also available through Brightspace (Resources-&gt;Class zoom meetings-&gt;Cloud Recordings tab)<br /> <br /> ==Notes==<br /> <br /> &lt;pre&gt;<br /> Lecture 23<br /> ----------<br /> <br /> Final is on December 16th, 7-9 PM<br /> - if you have accommodations, the timing might be a bit different<br /> <br /> I can potentially have the final earlier in the day for some, let me know<br /> - but I need it to be fixed time<br /> <br /> Final is same format as midterm<br /> - no proctoring software<br /> - but there will be interviews, ideally the week of Dec. 27th,<br /> but if necessary the week of January 3rd.<br /> - so everything will be graded by then<br /> - if week of Jan 3rd, final grades will be submitted,<br /> but I can submit a change of grade if things go up<br /> <br /> Practice questions<br /> - that&#039;s the assignments and the midterm<br /> - I suggest making up your own<br /> <br /> I will have an optional review session on Dec. 14th<br /> - will be recorded, but I will have no prepared material<br /> <br /> Final exam is open book/open note/open internet<br /> - just NO COLLABORATION<br /> <br /> But here&#039;s a hint - online isn&#039;t your best choice<br /> - if you look up more than a few questions, you are doing it wrong<br /> - takes up too much time, and will often mislead you<br /> <br /> Look to the textbook when the tutorials/assignments feel unclear<br /> - they will provide lots of context<br /> <br /> atomic - something that is &quot;indivisible&quot;<br /> - so an atomic operation can&#039;t be divided, it all happens or none of it happens<br /> - in databases, we make things indivisible with transactions<br /> - but that just means we undo changes if we can do them all<br /> <br /> What is the heart of concurrency problems? TOCTTOU<br /> - &quot;time of check to time of use&quot;<br /> - data changes between the time you check it to the time you modify/use<br /> it<br /> <br /> - Anywhere you check data and then do something based on that test<br /> is where concurrency issues can bite you<br /> - which is all over the place<br /> <br /> single-threaded processes really minimize concurrency issues<br /> - minimal concurrency issue inside a process (except signals)<br /> - between process concurrency is managed through system calls<br /> - which from a process perspective are mostly atomic<br /> - and there are mechanisms to enfore mutual exclusion for communication, e.g. pipes<br /> <br /> What is ioctl?<br /> - a system call that is device specific<br /> - basically, a way to send custom messages to a device driver<br /> - NO uniformity, so can make it whatever you want<br /> - very old fashioned, not preferred nowadays but still used<br /> <br /> Mental model: ioctl lets you control a device<br /> - e.g., if it is a tape drive, lets you eject a tape<br /> - things that don&#039;t fit into a classic file API model<br /> <br /> You may volunteer for final interviews<br /> - will have plenty of time!<br /> - may have to extend into Winter term<br /> &lt;/pre&gt;</div>

Soma