Chapter 16 Miscellaneous Questions

FreeBSD only appears to use more swap than Linux. In actual fact, it does not. The main difference between FreeBSD and Linux in this regard is that FreeBSD will proactively move entirely idle, unused pages of main memory into swap in order to make more main memory available for active use. Linux tends to only move pages to swap as a last resort. The perceived heavier use of swap is balanced by the more efficient use of main memory.

Note that while FreeBSD is proactive in this regard, it does not arbitrarily decide to swap pages when the system is truly idle. Thus you will not find your system all paged out when you get up in the morning after leaving it idle overnight.

The simple answer is that free memory is wasted memory. Any memory that your programs do not actively allocate is used within the FreeBSD kernel as disk cache. The values shown by top(1) labeled as Inact, Cache, and Buf are all cached data at different aging levels. This cached data means the system does not have to access a slow disk again for data it has accessed recently, thus increasing overall performance. In general, a low value shown for Free memory in top(1) is good, provided it is not very low.

To understand why FreeBSD uses the ELF format, you must first know a little about the 3 currently ``dominant'' executable formats for Unix:

• a.out(5)

  The oldest and ``classic'' Unix object format. It uses a short and compact header with a magic number at the beginning that is often used to characterize the format (see a.out(5) for more details). It contains three loaded segments: .text, .data, and .bss plus a symbol table and a string table.

• COFF

  The SVR3 object format. The header now comprises a section table, so you can have more than just .text, .data, and .bss sections.

• ELF

  The successor to COFF, featuring Multiple sections and 32-bit or 64-bit possible values. One major drawback: ELF was also designed with the assumption that there would be only one ABI per system architecture. That assumption is actually quite incorrect, and not even in the commercial SYSV world (which has at least three ABIs: SVR4, Solaris, SCO) does it hold true.

  FreeBSD tries to work around this problem somewhat by providing a utility for branding a known ELF executable with information about the ABI it is compliant with. See the manual page for brandelf(1) for more information.

FreeBSD comes from the ``classic'' camp and has traditionally used the a.out(5) format, a technology tried and proven through many generations of BSD releases. Though it has also been possible for some time to build and run native ELF binaries (and kernels) on a FreeBSD system, FreeBSD initially resisted the ``push'' to switch to ELF as the default format. Why? Well, when the Linux camp made their painful transition to ELF, it was not so much to flee the a.out executable format as it was their inflexible jump-table based shared library mechanism, which made the construction of shared libraries very difficult for vendors and developers alike. Since the ELF tools available offered a solution to the shared library problem and were generally seen as ``the way forward'' anyway, the migration cost was accepted as necessary and the transition made.

In FreeBSD's case, our shared library mechanism is based more closely on Sun's SunOS-style shared library mechanism and, as such, is very easy to use. However, starting with 3.0, FreeBSD officially supports ELF binaries as the default format. Even though the a.out executable format has served us well, the GNU people, who author the compiler tools we use, have dropped support for the a.out format. This has forced us to maintain a divergent version of the compiler and linker, and has kept us from reaping the benefits of the latest GNU development efforts. Also the demands of ISO-C++, notably constructors and destructors, has also led to native ELF support in future FreeBSD releases.

Back in the dim, dark past, there was simple hardware. This simple hardware supported a simple, small system. a.out was completely adequate for the job of representing binaries on this simple system (a PDP-11). As people ported Unix from this simple system, they retained the a.out format because it was sufficient for the early ports of Unix to architectures like the Motorola 68k, VAXen, etc.

Then some bright hardware engineer decided that if he could force software to do some sleazy tricks, then he would be able to shave a few gates off the design and allow his CPU core to run faster. While it was made to work with this new kind of hardware (known these days as RISC), a.out was ill-suited for this hardware, so many formats were developed to get to a better performance from this hardware than the limited, simple a.out format could offer. Things like COFF, ECOFF, and a few obscure others were invented and their limitations explored before things seemed to settle on ELF.

In addition, program sizes were getting huge and disks (and physical memory) were still relatively small so the concept of a shared library was born. The VM system also became more sophisticated. While each one of these advancements was done using the a.out format, its usefulness was stretched more and more with each new feature. In addition, people wanted to dynamically load things at run time, or to junk parts of their program after the init code had run to save in core memory and swap space. Languages became more sophisticated and people wanted code called before main automatically. Lots of hacks were done to the a.out format to allow all of these things to happen, and they basically worked for a time. In time, a.out was not up to handling all these problems without an ever increasing overhead in code and complexity. While ELF solved many of these problems, it would be painful to switch from the system that basically worked. So ELF had to wait until it was more painful to remain with a.out than it was to migrate to ELF.

However, as time passed, the build tools that FreeBSD derived their build tools from (the assembler and loader especially) evolved in two parallel trees. The FreeBSD tree added shared libraries and fixed some bugs. The GNU folks that originally write these programs rewrote them and added simpler support for building cross compilers, plugging in different formats at will, etc. Since many people wanted to build cross compilers targeting FreeBSD, they were out of luck since the older sources that FreeBSD had for as and ld were not up to the task. The new gnu tools chain (binutils) does support cross compiling, ELF, shared libraries, C++ extensions, etc. In addition, many vendors are releasing ELF binaries, and it is a good thing for FreeBSD to run them. And if it is running ELF binaries, why bother having a.out any more? It is a tired old horse that has proven useful for a long time, but it is time to turn him out to pasture for his long, faithful years of service.

ELF is more expressive than a.out and will allow more extensibility in the base system. The ELF tools are better maintained, and offer cross compilation support, which is important to many people. ELF may be a little slower than a.out, but trying to measure it can be difficult. There are also numerous details that are different between the two in how they map pages, handle init code, etc. None of these are very important, but they are differences. In time support for a.out will be moved out of the GENERIC kernel, and eventually removed from the kernel once the need to run legacy a.out programs is past.

Symlinks do not have permissions, and by default, chmod(1) will not follow symlinks to change the permissions on the target file. So if you have a file, foo, and a symlink to that file, bar, then this command will always succeed.

    % chmod g-w bar

However, the permissions on foo will not have changed.

You have to use either -H or -L together with the -R option to make this work. See the chmod(1) and symlink(7) manual pages for more info.

You would think it would be easy enough to change UT_NAMESIZE and rebuild the whole world, and everything would just work. Unfortunately there are often scads of applications and utilities (including system tools) that have hard-coded small numbers (not always 8 or 9, but oddball ones like 15 and 20) in structures and buffers. Not only will this get you log files which are trashed (due to variable-length records getting written when fixed records were expected), but it can break Suns NIS clients and potentially cause other problems in interacting with other Unix systems.

In FreeBSD 3.0 and later, the maximum name length has been increased to 16 characters and those various utilities with hard-coded name sizes have been found and fixed. The fact that this touched so many areas of the system is why, in fact, the change was not made until 3.0.

If you are absolutely confident in your ability to find and fix these sorts of problems for yourself when and if they pop up, you can increase the login name length in earlier releases by editing /usr/include/utmp.h and changing UT_NAMESIZE accordingly. You must also update MAXLOGNAME in /usr/include/sys/param.h to match the UT_NAMESIZE change. Finally, if you build from sources, do not forget that /usr/include is updated each time! Change the appropriate files in /usr/src/.. instead.

Yes, starting with version 3.0 you can use BSDI's doscmd DOS emulation which has been integrated and enhanced. Send mail to the FreeBSD-emulation mailing list <freebsd-emulation@FreeBSD.org> if you are interested in joining this ongoing effort!

For pre-3.0 systems, there is a neat utility called pcemu in the ports collection which emulates an 8088 and enough BIOS services to run DOS text mode applications. It requires the X Window System (provided as XFree86).

See the Translation FAQ in the FreeBSD Documentation Project Primer.

The FreeBSD.org mail system implements some of the stricter Postfix checks on incoming mail and rejects mail that is either misconfigured or is potential spam. Your mail might bounce for one of the following reasons:

• The email is being sent from a known spam domain or IP block.

  The FreeBSD mail servers reject email from known spam sources. If you have service through a company or domain who generates or relays spam, please switch to a service provider who does not.

• The body of the email only contains HTML.

  Mail should be sent in plain text only. Please configure your mail user agent to send plain text.

• The mailer at FreeBSD.org cannot resolve the IP address of the connecting host back to a symbolic name.

  Working reverse DNS is a standard requirement for accepting mail from a host. Set up reverse DNS for your mail server's IP address. Many home services (DSL, cable, dialup, etc.) will not give you this option. In this case, relay your email through your service provider's mail server.

• The hostname given in the EHLO/HELO part of the SMTP exchange cannot be resolved to an IP address.

  A fully qualified, resolvable host name is necessary in this part of the SMTP dialogue before mail will be accepted. If you do not have a host name that is registered in the DNS, then you should use your service provider's mail server to relay your mail.

• Your message had a message ID ending with the string ``localhost''.

  Some mail user agents generate bad message IDs which will not be accepted. You will need to persuade your mail user agent to generate a valid message ID or else configure your mail transfer agent to rewrite them.

While FreeBSD does not provide open access to any of their servers, others do provide open access Unix systems. The charge varies and limited services may be available.

Arbornet, Inc, also known as M-Net, has been providing open access to Unix systems since 1983. Starting on an Altos running System III, the site switched to BSD/OS in 1991. In June of 2000, the site switched again to FreeBSD. M-Net can be accessed via telnet and SSH and provides basic access to the entire FreeBSD software suite. However, network access is limited to members and patrons who donate to the system, which is run as a non-profit organization. M-Net also provides an bulletin board system and interactive chat.

Grex provides a site very similar to M-Net including the same bulletin board and interactive chat software. However, the machine is a Sun 4M and is running SunOS

SUP stands for Software Update Protocol, and was developed by CMU for keeping their development trees in sync. We used it to keep remote sites in sync with our central development sources.

SUP is not bandwidth friendly, and has been retired. The current recommended method to keep your sources up to date is CVSup

He does not have one, and is just called ``the BSD daemon''. If you insist upon using a name, call him ``beastie''. Note that ``beastie'' is pronounced ``BSD''.

You can learn more about the BSD daemon on his home page.

Perhaps. The BSD daemon is copyrighted by Marshall Kirk McKusick. You will want to check his Statement on the Use of the BSD Daemon Figure for detailed usage terms.

In summary, you are free to use the image in a tasteful manner, for personal use, so long as appropriate credit is given. If you want to use him commercially, you must contact Kirk McKusick. More details are available on the BSD Daemon's home page.

You will find eps and Xfig drawings under /usr/share/examples/BSD_daemon/.

MFC is an acronym for ``Merged From -CURRENT''. It is used in the CVS logs to denote when a change was migrated from the CURRENT to the STABLE branches.

It stands for something in a secret language that only members can know. It does not translate literally but its ok to tell you that BSD's translation is something between, ``Formula-1 Racing Team'', ``Penguins are tasty snacks'', and ``We have a better sense of humor than Linux''. :-)

Seriously, BSD is an acronym for ``Berkeley Software Distribution'', which is the name the Berkeley CSRG (Computer Systems Research Group) chose for their Unix distribution way back when.

Principle of Least Astonishment. It means that as FreeBSD evolves, changes visible to the user should be kept as unsurprising as possible. For example, arbitrarily rearranging system startup variables in /etc/defaults/rc.conf violates POLA. Developers consider POLA when contemplating user-visible system changes.

A repo-copy (which is a short form of ``repository copy'') refers to the direct copying of files within the CVS repository.

Without a repo-copy, if a file needed to be copied or moved to another place in the repository, the committer would run cvs add to put the file in its new location, and then cvs rm on the old file if the old copy was being removed.

The disadvantage of this method is that the history (i.e. the entries in the CVS logs) of the file would not be copied to the new location. As the FreeBSD Project considers this history very useful, a repository copy is often used instead. This is a process where one of the repository meisters will copy the files directly within the repository, rather than using the cvs(1) program.

The really, really short answer is that you should not. The somewhat longer answer is that just because you are capable of building a bikeshed does not mean you should stop others from building one just because you do not like the color they plan to paint it. This is a metaphor indicating that you need not argue about every little feature just because you know enough to do so. Some people have commented that the amount of noise generated by a change is inversely proportional to the complexity of the change.

The longer and more complete answer is that after a very long argument about whether sleep(1) should take fractional second arguments, Poul-Henning Kamp <phk@FreeBSD.org> posted a long message entitled ``A bike shed (any colour will do) on greener grass...''. The appropriate portions of that message are quoted below.