Show notes
We read the FreeBSD Q3 status report, explore good and bad syscalls, list GOG Games for OpenBSD, and show you what devmatch can do. This episode was brought to you by Headlines FreeBSD Q3 Status Report 2017 FreeBSD Team Reports FreeBSD Release Engineering Team Ports Collection The FreeBSD Core Team The FreeBSD Foundation Projects FreeBSD CI Kernel Intel 10G iflib Driver Update Intel iWARP Support pNFS Server Plan B Architectures AMD Zen (family 17h) support Userland Programs Updates to GDB Ports FreeBSDDesktop OpenJFX 8 Puppet Documentation Absolute FreeBSD, 3rd Edition Manual Pages Third-Party Projects The nosh Project ####FreeBSD Foundation Q4 Update *** ###11 syscalls that rock the world 0. read > You cannot go wrong with a read. You can barely EFAULT it! On Linux amd64 it is syscall zero. If all its arguments are zero it returns zero. Cool! 1. pipe > The society for the preservation of historic calling conventions is very fond of pipe, as in many operating systems and architectures it preserves the fun feature of returning both of the file descriptors as return values. At least Linux MIPS does, and NetBSD does even on x86 and amd64. Multiple return values are making a comeback in languages like Lua and Go, but C has always had a bit of a funny thing about them, but they have long been supported in many calling conventions, so let us use them in syscalls! Well, one syscall. 2. kqueue > When the world went all C10K on our ass, and scaleable polling was a thing, Linux went epoll, the BSDs went kqueue and Solaris went /dev/poll. The nicest interface was kqueue, while epoll is some mix of edge and level triggered semantics and design errors so bugs are still being found. 3. unshare > Sounds like a selfish syscall, but this generous syscall call is the basis of Linux namespaces, allowing a process to isolate its resources. Containers are built from unshares. 4. setns > If you liked unshare, its younger but cooler friend takes file descriptors for namespaces. Pass it down a unix socket to another process, or stash it for later, and do that namespace switching. All the best system calls take file descriptors. 5. execveat > Despite its somewhat confusing name (FreeBSD has the saner fexecve, but other BSDs do not have support last time I checked), this syscall finally lets you execute a program just given a file descriptor for the file. I say finally, as Linux only implemented this in 3.19, which means it is hard to rely on it (yeah, stop using those stupid old kernels folks). Before that Glibc had a terrible userspace implementation that is basically useless. Perfect for creating sandboxes, as you can sandbox a program into a filesystem with nothing at all in, or with a totally controlled tree, by opening the file to execute before chroot or changing the namespace. 6. pdfork > Too cool for Linux, you have to head out to FreeBSD for this one. Like fork, but you get a file descriptor for the process not a pid. Then you can throw it in the kqueue or send it to another process. Once you have tried process descriptors you will never go back. 7. signalfd > You might detect a theme here, but if you have ever written traditional 1980s style signal handlers you know how much they suck. How about turning your signals into messages that you can read on, you guessed it, file descriptors. Like, usable. 8. wstat > This one is from Plan 9. It does the opposite of stat and writes the same structure. Simples. Avoids having chmod, chown, rename, utime and so on, by the simple expedient of making the syscall symmetric. Why not? 9. clonefile > The only cool syscall on OSX, and only supported on the new APFS filesystem. Copies whole files or directories on a single syscall using copy on write for all the data. Look on my works, copy_file_range and despair. 10. pledge > The little sandbox that worked. OpenBSD only here, they managed to make a simple sandbox that was practical for real programs, like the base OpenBSD system. Capsicum form FreeBSD (and promised for Linux for years but no sign) is a lovely design, and gave us pdfork, but its still kind of difficult and intrusive to implement. Linux has, well, seccomp, LSMs, and still nothing that usable for the average program. ###Eleven syscalls that suck 0. ioctl > It can‘t decide if it‘s arguments are integers, strings, or some struct that is lost in the midst of time. Make up your mind! Plan 9 was invented to get rid of this. 1. fcntl > Just like ioctl but for some different miscellaneous operations, because one miscelleny is not enough. 2. tuxcall > Linux put a web server in the kernel! To win a benchmark contest with Microsoft! It had it‘s own syscall! My enum tux_reactions are YUK! Don‘t worry though, it was a distro patch (thanks Red Hat!) and never made it upstream, so only the man page and reserved number survive to taunt you and remind you that the path of the righteous is beset by premature optmization! 3. io_setup > The Linux asynchronous IO syscalls are almost entirely useless! Almost nothing works! You have to use O_DIRECT for a start. And then they still barely work! They have one use, benchmarking SSDs, to show what speed you could get if only there was a usable API. Want async IO in kernel? Use Windows! 4. stat, and its friends and relatives > Yes this one is useful, but can you find the data structure it uses? We have oldstat, oldfstat, ustat, oldlstat, statfs, fstatfs, stat, lstat, fstat, stat64, lstat64, fstat64, statfs64, fstatfs64, fstatat64 for stating files and links and filesystems in Linux. A new bunch will be along soon for Y2038. Simplify your life, use a BSD, where they cleaned up the mess as they did the cooking! Linux on 32 bit platforms is just sucky in comparison, and will get worse. And don’t even look at MIPS, where the padding is wrong. 5. Linux on MIPS > Not a syscall, a whole implemntation of the Linux ABI. Unlike the lovely clean BSDs, Linux is different on each architecture, system calls randomly take arguments in different orders, and constants have different values, and there are special syscalls. But MIPS takes the biscuit, the whole packet of biscuits. It was made to be binary compatible with old SGI machines that don’t even exist, and has more syscall ABIs than I have had hot dinners. Clean it up! Make a new sane MIPS ABI and deprecate the old ones, nothing like adding another variant. So annoying I think I threw out all my MIPS machines, each different. 6. inotify, fanotify and friends > Linux has no fewer than three file system change notification protocols. The first, dnotify hopped on ioctl‘s sidekick fcntl, while the two later ones, inotify and fanotify added a bunch more syscalls. You can use any of them, and they still will not provide the notification API you want for most applications. Most people use the second one, inotify and curse it. Did you know kqueue can do this on the BSDs? 7. personality > Oozing in personality, but we just don’t get along. Basically obsolete, as the kernel can decide what kind of system emulation to do from binaries directly, it stays around with some use cases in persuading ./configure it is running on a 32 bit system. But it can turn off ASLR, and let the CVEs right into your system. We need less persoanlity! 8. gettimeofday > Still has an obsolete timezone value from an old times when people thought timezones should go all the way to the kernel. Now we know that your computer should not know. Set its clock to UTC. Do the timezones in the UI based on where the user is, not the computer. You should use clock_gettime now. Don’t even talk to me about locales. This syscall is fast though, don’t use it for benchmarking, its in the VDSO. 9. splice and tee > These, back in 2005 were a quite nice idea, although Linux said then “it is incomplete, the interfaces are ugly, and it will oops the system if anything goes wrong”. It won’t oops your system now, but usage has not taken off. The nice idea from Linus was that a pipe is just a ring buffer in the kernel, that can have a more general API and use cases for performant code, but a decade on it hasn’t really worked out. It was also supposed to be a more general sendfile, which in many ways was the successor of that Tux web server, but I think sendfile is still more widely used. 10. userfaultfd > Yes, I like file descriptors. Yes CRIU is kind of cool. But userspace handling page faults? Is nothing sacred? I get that you can do this badly with a SIGSEGV handler, but talk about lipstick on a pig. *** ###OpenBSD 6.0 on an iMac G3 from 1999 > A while ago I spent $50 for an iMac G3 (aka the iMac,1). This iconic model restored Apple's fortunes in the late '90s. Since the iMac G3 can still boot Mac OSes 8 and 9, I mostly use the machine to indulge a nostalgia for childhood schooldays spent poking at the operating system and playing Escape Velocity. But before I got around to that, I decided to try out the software that the previous owner had left on the machine. The antiquated OSX 10.2 install and 12 year old versions of Safari and Internet Explorer were too slow and old to use for anything. Updating to newer software was almost impossible; a later OSX is required to run the little PowerPC-compatible software still languishing in forgotten corners of the Internet. This got me thinking: could this machine be used, really used, nowadays? Lacking a newer OSX disc, I decided to try the most recent OpenBSD release. (And, since then, to re-try with each new OpenBSD release.) Below are the results of this experiment (plus a working xorg.conf file) and a few background notes. Background > This iMac is a Revision D iMac G3 in grape. It's part of the iMac,1 family of computers. This family includes all tray-loading iMac G3s. (Later iMac G3s had a slot-loading CD drive and different components.) Save for a slightly faster processor, a dedicated graphics card, and cosmetic tweaks to the case, my iMac is identical to the prior year's line-launching Bondi Blue iMac. My machine has had its memory upgraded from 32 MB to 320 MB. Thank Goodness. > The Revision D iMac G3 shipped with Mac OS 8.5. It can run up to Mac OS 9.2.2 or OSX 10.3.9. Other operating systems that tout support for the iMac,1 include NetBSD, OpenBSD, and a shrinking number of Linux distributions. > OpenBSD is simple (by design) and well-maintained. In contrast, NetBSD seems rather more complex and featureful, and I have heard grumbling that despite its reputation for portability, NetBSD really only works well on amd64. I'd test that assertion if OpenBSD's macppc installation instructions didn't seem much simpler than NetBSD's. Linux is even more complicated, although most distros are put together in a way that you can mostly ignore that complexity (until you can't). In the end I went with OpenBSD because I am familiar with it and because I like it. Installing OpenBSD on the iMac,1 > Installing OpenBSD on this iMac was simple. It's the same procedure as installing OpenBSD on an amd64 rig. You put in the installation disc; you tell the machine to boot from it; and then you answer a few prompts, most of which simply ask you to press enter. In this case, OpenBSD recognizes all machine's hardware just fine, including sound and networking, though I had a little trouble with video. > The OpenBSD documentation says video should just work and that an xorg.conf file isn't necessary. As such, it no longer ships with an xorg.conf file. Though that's never posed a problem on my other OpenBSD machines, it does here. Video doesn't work out of the box on my iMac,1. startx just blanks the screen. Fortunately, because the BSDs use a centralized development model where each operating system is stored in one repository, OpenBSD's website provides a web interface to the source code going back to the early days. I was able to find the last version of the sample xorg.conf that used to ship on macppc. With a little tweaking, I transformed that file into this one, with which video works just fine. Just drop it into your iMac's /etc/X11 directory. You'll also need to remember to set the machdep.allowaperture sysctl to 2 (e.g., as root run sysctl machdep.allowaperture=2), although the installer will do that automatically if you answer yes to the question about whether you plan to run X. > All that being said, video performance is pretty poor. I am either doing something wrong, or OpenBSD doesn't have accelerated video for this iMac, or this machine is just really old! I will discuss performance below. Running OpenBSD on the iMac,1 > The machine performs okay under OpenBSD. You can expect to ably run minimalistic software under minimalistic window managers. I tried dillo, mrxvt, and cmus under cwm and fvwm. Performance here was just fine. I also tried Firefox 26, 33, and 34 under fvwm and cwm. Firefox ran, but "modern," Javascript-heavy sites were an exercise in frustration; the 2015 version of CNN.com basically froze Firefox for 30 seconds or more. A lighter browser like dillo is doable. > You'll notice that I used the past-tense to talk about Firefox. Firefox currently doesn't build on PowerPC on OpenBSD. Neither does Chromium. Neither do a fair number of applications. But whatever -- there's still a lot of lighter applications available, and it's these you'll use day-to-day on a decades-old machine. > Lightweight window managers work okay, as you'd expect. You can even run heavier desktop environments, such as xfce, though you'll give up a lot of performance. > I ran the Ubench benchmark on this iMac and two more modern machines also running OpenBSD. The benchmark seems like an old one; I don't know how (if at all) it accounts for hardware changes in the past 13 years. That is, I don't know if the difference in score accurately measures the difference in real-world performance. Here are the results anyway: Conclusion > Except for when I check to see if OpenBSD still works, I run Mac OS9 on this rig. I have faster and better machines for running OpenBSD. If I didn't -- if this rig were, improbably, all I had left, and I was waiting on the rush delivery of something modern -- then I would use OpenBSD on my iMac,1. I'd have to stick to lightweight applications, but at least they'd be up-to-date and running on a simple, stable, OS. *** ##News Roundup ###34th Chaos Communication Congress Schedule Many talks are streamed live, a good mixture of english and german talks May contain DTraces of FreeBSD Are all BSDs created equally? library operating systems Hardening Open Source Development *** ###OpenBSD 6.2 + CDE > If you've noticed a disruption in the time-space continuum recently, it is likely because I have finally been able to compile and install the Common Desktop Environment (CDE) in a current and actively-developed operating system (OpenBSD 6.2 in this case). > This comes after so many attempts (across multiple platforms) that ended up with the build process prematurely stopping itself in its own tracks for a variety of infinitesimal reasons that were beyond my comprehension as a non-programmer, or when there was success it was not without some broken parts. As for the latter, I've been able to build CDE on OpenIndiana Hipster, but with an end product where I'm unable to change the color scheme in dtstyle (because "useColorObj" is set to "False"), with a default color scheme that is low-res and unpleasant. As for changing "useColorObj" to "True", I tried every recommended trick I could find online, but nothing worked. > My recent attempts at installing CDE on OpenBSD (version 6.1) saw the process stop due to a number of errors that are pure gibberish to these naive eyes. While disappointing, it was par for th…
Full show notes at the publisher 
Our TOPPODCAST Picks 
Stay Connected
