Disclaimer : This handbook has been replaced by a newer version and is not maintained anymore. |
Content:
First of all, welcome to Gentoo. You are about to enter the world of choices and performance. Gentoo is all about choices. When installing Gentoo, this is made clear to you several times -- you can choose how much you want to compile yourself, how to install Gentoo, what system logger you want, etc.
Gentoo is a fast, modern metadistribution with a clean and flexible design. Gentoo is built around free software and doesn't hide from its users what is beneath the hood. Portage, the package maintenance system which Gentoo uses, is written in Python, meaning you can easily view and modify the source code. Gentoo's packaging system uses source code (although support for precompiled packages is included too) and configuring Gentoo happens through regular textfiles. In other words, openness everywhere.
It is very important that you understand that choices are what makes Gentoo run. We try not to force you onto anything you don't like. If you feel like we do, please bugreport it.
How is the Installation Structured?
The Gentoo Installation can be seen as a 10-step procedure, corresponding to chapters 2 - 11. Every step results in a certain state:
When you are given a certain choice, we try our best to explain what the pros and cons are. We will continue then with a default choice, identified by "Default: " in the title. The other possibilities are marked by "Alternative: ". Do not think that the default is what we recommend. It is however what we believe most users will use.
Sometimes you can pursue an optional step. Such steps are marked as "Optional: " and are therefore not needed to install Gentoo. However, some optional steps are dependant on a previous decision you made. We will inform you when this happens, both when you make the decision, and right before the optional step is described.
You can install Gentoo in many different ways. You can download and install from one of our Installation CDs, from an existing distribution, from a bootable CD (such as Knoppix), from a netbooted environment, from a rescue floppy, etc.
This document covers the installation using the Universal Installation CD, a bootable CD that contains everything you need to get Gentoo Linux up and running. You can optionally use one of our Package CDs as well to install a complete system in a matter of minutes after having installed the Gentoo base system.
This installation approach however does not immediately use the latest version of the available packages; if you want this you should check out the Installation Instructions inside our Gentoo Linux Handbooks.
For help on the other installation approaches, please read our Alternative Installation Guide. We also provide a Gentoo Installation Tips & Tricks document that might be useful to read as well. If you feel that the current installation instructions are too elaborate, feel free to use our Quick Installation Guide available from our Documentation Resources if your architecture has such a document available.
If you find a problem in the installation (or in the installation documentation), please check the errata from our Gentoo Release Engineering Project, visit our bugtracking system and check if the bug is known. If not, please create a bugreport for it so we can take care of it. Do not be afraid of the developers who are assigned to (your) bugs -- they generally don't eat people.
Note though that, although the document you are now reading is architecture-specific, it will contain references to other architectures as well. This is due to the fact that large parts of the Gentoo Handbook use source code that is common for all architectures (to avoid duplication of efforts and starvation of development resources). We will try to keep this to a minimum to avoid confusion.
If you are uncertain if the problem is a user-problem (some error you made despite having read the documentation carefully) or a software-problem (some error we made despite having tested the installation/documentation carefully) you are free to join #gentoo on irc.freenode.net. Of course, you are welcome otherwise too :)
If you have a question regarding Gentoo, check out our Frequently Asked Questions, available from the Gentoo Documentation. You can also view the FAQs on our forums. If you can't find the answer there ask on #gentoo, our IRC-channel on irc.freenode.net. Yes, several of us are freaks who sit on IRC :-)
1.b. Fast Installation using the Gentoo Reference Platform
What is the Gentoo Reference Platform?
The Gentoo Reference Platform, from now on abbreviated to GRP, is a snapshot of prebuilt packages users (that means you!) can install during the installation of Gentoo to speed up the installation process. The GRP consists of all packages required to have a fully functional Gentoo installation. They are not just the ones you need to have a base installation up to speed in no time, but all lengthier builds (such as KDE, xorg-x11, GNOME, OpenOffice, Mozilla, ...) are available as GRP packages too.
However, these prebuilt packages aren't maintained during the lifetime of the Gentoo distribution. They are snapshots released at every Gentoo release and make it possible to have a functional environment in a short amount of time. You can then upgrade your system in the background while working in your Gentoo environment.
How Portage Handles GRP Packages
Your Portage tree - the collection of ebuilds (files that contain all information about a package, such as its description, homepage, sourcecode URLs, compilation instructions, dependencies, etc.) - must be synchronised with the GRP set: the versions of the available ebuilds and their accompanying GRP packages must match.
For this reason you can only benefit from the GRP packages Gentoo provides while performing the current installation approach. GRP is not available for those interested in performing an installation using the latest versions of all available packages.
Not all architectures provide GRP packages. That doesn't mean GRP isn't supported on the other architectures, but it means that we don't have the resources to build and test the GRP packages.
At present we provide GRP packages for the following architectures:
If your architecture (or subarchitecture) isn't on this list, you are not able to opt for a GRP installation.
Now that this introduction is over, let's continue with Booting the Universal Installation CD.
Before we start, we first list what hardware requirements you need to successfully install Gentoo on your box.
| Sparc System | Please check the UltraLinux FAQ |
| CPU | Although sparc64 is the only officially supported platform, experimental support for sparc32 is available as well |
| Memory | 64 MB |
| Diskspace | 1.5 GB (excluding swap space) |
| Swap space | At least 256 MB |
We currently only provide Installation CDs for the sparc64 architecture. Users of sparc32 can use the experimental netboot images to install Gentoo from. More information about netbooting can be found in our Gentoo Linux based Netboot HOWTO.
2.b. The Gentoo Universal Installation CD
Gentoo Linux can be installed using one of three stage tarball files. A stage file is a tarball (compressed archive) that contains a minimal environment.
We will opt for a stage3 installation throughout this document. If you want to perform a Gentoo installation using the stage1 or stage2 files, please use the installation instructions in the Gentoo Handbook. They do require a working Internet connection though.
Gentoo Universal Installation CD
An Installation CD is a bootable medium which contains a self-sustained Gentoo environment. It allows you to boot Linux from the CD. During the boot process your hardware is detected and the appropriate drivers are loaded. The Gentoo Installation CDs are maintained by Gentoo developers.
There currently are two Installation CDs available:
Gentoo also provides a Package CD. This is no Installation CD but an additional resource that you can exploit during the installation of your Gentoo system. It contains prebuilt packages (the so-called GRP set) that allows you to easily and quickly install additional applications (such as OpenOffice.org, KDE, GNOME, ...) immediately after the Gentoo installation and right before you update your Portage tree.
The use of the Package CD is covered later in this document.
2.c. Download, Burn and Boot a Gentoo Installation CD
Downloading and Burning the Installation CDs
You can download the Universal Installation CD (and, if you want to, the Packages CD as well) from one of our mirrors. The Installation CDs are located in the releases/sparc/2005.0/installcd/sparc64 directory; the Package CDs are located in the releases/sparc/2005.0/packagecd/sparc64 directory.
Inside those directories you'll find so-called ISO-files. Those are full CD images which you can write on a CD-R.
After downloading the file, you can verify its integrity to see if it is corrupted or not:
To fetch our public key using the GnuPG application, run the following command:
Code Listing 3.1: Obtaining the public key |
$ gpg --keyserver pgp.mit.edu --recv-keys 17072058
|
Now verify the signature:
Code Listing 3.2: Verify the cryptographic signature |
$ gpg --verify <signature file> <downloaded iso>
|
To burn the downloaded ISO(s), you have to select raw-burning. How you do this is highly program-dependent. We will discuss cdrecord and K3B here; more information can be found in our Gentoo FAQ.
Booting the Universal Installation CD
Insert the Gentoo Installation CD in the CD-ROM and boot your system. During startup, press Stop-A to enter OpenBootPROM (OBP). Once you are in the OBP, boot from the CD-ROM:
Code Listing 3.3: Booting the Installation CD |
ok boot cdrom
|
You will be greeted by the SILO boot manager (on the Installation CD). Type in gentoo-2.4 (single-CPU kernel) or gentoo-2.4-smp (multi-CPU kernel) and press enter to continue booting the system. In the following example we'll boot the gentoo-2.4 kernel.
Code Listing 3.4: Continue booting from the Installation CD |
boot: gentoo-2.4
|
Once the Installation CD is booted, you will be automatically logged on to the system.
You should have a root ("#") prompt on the current console and can also switch to other consoles by pressing Alt-F2, Alt-F3 and Alt-F4. Get back to the one you started on by pressing Alt-F1. You will also find a root prompt on the serial console (ttyS0).
Continue with Extra Hardware Configuration.
If not all hardware is supported out-of-the-box, you will need to load the appropriate kernel modules.
In the next example we try to load the 8139too module (support for certain kinds of network interfaces):
Code Listing 3.5: Loading kernel modules |
# modprobe 8139too
|
If you plan on giving other people access to your installation environment or you want to chat using irssi without root privileges (for security reasons), you need to create the necessary user accounts and change the root password.
To change the root password, use the passwd utility:
Code Listing 3.6: Changing the root password |
# passwd New password: (Enter your new password) Re-enter password: (Re-enter your password) |
To create a user account, we first enter their credentials, followed by its password. We use useradd and passwd for these tasks. In the next example, we create a user called "john".
Code Listing 3.7: Creating a user account |
# useradd -m -G users john # passwd john New password: (Enter john's password) Re-enter password: (Re-enter john's password) |
You can change your user id from root to the newly created user by using su:
Code Listing 3.8: Changing user id |
# su - john
|
Optional: Viewing Documentation while Installing
If you want to view the Gentoo Handbook (either from-CD or online) during the installation, make sure you have created a user account (see Optional: User Accounts). Then press Alt-F2 to go to a new terminal and log in.
If you want to view the documentation on the CD you can immediately run links2 to read it:
Code Listing 3.9: Viewing the on-CD documentation |
# links2 /mnt/cdrom/docs/handbook/html/index.html
|
However, it is preferred that you use the online Gentoo Handbook as it will be more recent than the one provided on the CD. You can view it using links2 as well, but only after having completed the Configuring your Network chapter (otherwise you won't be able to go on the Internet to view the document):
Code Listing 3.10: Viewing the Online Documentation |
# links2 http://www.gentoo.org/doc/en/handbook/handbook-sparc.xml
|
You can go back to your original terminal by pressing Alt-F1.
Optional: Starting the SSH Daemon
If you want to allow other users to access your computer during the Gentoo installation (perhaps because those users are going to help you install Gentoo, or even do it for you), you need to create a user account for them and perhaps even provide them with your root password (only do that if you fully trust that user).
To fire up the SSH daemon, execute the following command:
Code Listing 3.11: Starting the SSH daemon |
# /etc/init.d/sshd start
|
To be able to use sshd, you first need to set up your networking. Continue with the chapter on Configuring your Network.
Generally, you don't need a working network connection to install Gentoo using the Universal Installation CD. However, there are some circumstances where you do want to have a working Internet connection:
To find out if the stage3 file for your architecture is available, take a look inside /mnt/cdrom/stages and check if one of the available stages matches your architecture. If not, you can still opt for a stage3 file of an architecture compatible with yours.
If you on the other hand want to use a stage3 file optimized for your architecture and the stage3 file of your choice is not available, then you will need networking to download the appropriate stage3 file.
So, if you don't need networking, you can skip the rest of this chapter and continue with Preparing the Disks. Otherwise, continue with the networking configuration sections below.
3.b. Automatic Network Detection
If your system is plugged into an Ethernet network with a DHCP server, it is very likely that your networking configuration has already been set up automatically for you. If so, you should be able to take advantage of the many included network-aware commands on the Installation CD such as ssh, scp, ping, irssi, wget and links, among others.
If networking has been configured for you, the /sbin/ifconfig command should list some network interfaces besides lo, such as eth0:
Code Listing 2.1: /sbin/ifconfig for a working network configuration |
# /sbin/ifconfig (...) eth0 Link encap:Ethernet HWaddr 00:50:BA:8F:61:7A inet addr:192.168.0.2 Bcast:192.168.0.255 Mask:255.255.255.0 inet6 addr: fe80::50:ba8f:617a/10 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:1498792 errors:0 dropped:0 overruns:0 frame:0 TX packets:1284980 errors:0 dropped:0 overruns:0 carrier:0 collisions:1984 txqueuelen:100 RX bytes:485691215 (463.1 Mb) TX bytes:123951388 (118.2 Mb) Interrupt:11 Base address:0xe800 |
If you access the Internet through a proxy, you might need to set up proxy information during the installation. It is very easy to define a proxy: you just need to define a variable which contains the proxy server information.
In most cases, you can just define the variables using the server hostname. As an example, we assume the proxy is called proxy.gentoo.org and the port is 8080.
Code Listing 2.2: Defining proxy servers |
(If the proxy filters HTTP traffic) # export http_proxy="http://proxy.gentoo.org:8080" (If the proxy filters FTP traffic) # export ftp_proxy="ftp://proxy.gentoo.org:8080" (If the proxy filters RSYNC traffic) # export RSYNC_PROXY="proxy.gentoo.org:8080" |
If your proxy requires a username and password, you should use the following syntax for the variable:
Code Listing 2.3: Adding username/password to the proxy variable |
http://username:password@proxy.gentoo.org:8080 |
You may want to try pinging your ISP's DNS server (found in /etc/resolv.conf) and a Web site of choice, just to make sure that your packets are reaching the net, DNS name resolution is working correctly, etc..
Code Listing 2.4: Further network testing |
# ping -c 3 www.yahoo.com
|
Are you able to use your network? If so, you can skip the rest of this section and continue with Preparing the Disks. If not, bad luck, you'll have to work on it a bit more.
3.c. Automatic Network Configuration
If the network doesn't work immediately, some installation media allow you to use net-setup (for regular or wireless networks), adsl-setup (for ADSL-users) or pptp (for PPTP-users - only available on x86).
If your installation medium does not contain any of these tools or your network doesn't function yet, continue with Manual Network Configuration.
The simplest way to set up networking if it didn't get configured automatically is to run the net-setup script:
Code Listing 3.1: Running the net-setup script |
# net-setup eth0
|
net-setup will ask you some questions about your network environment. When all is done, you should have a working network connection. Test your network connection as stated before. If the tests are positive, congratulations! You are now ready to install Gentoo. Skip the rest of this section and continue with Preparing the Disks.
If your network still doesn't work, continue with Manual Network Configuration.
Assuming you need PPPoE to connect to the internet, the Installation CD (any version) has made things easy for you by including rp-pppoe. Use the provided adsl-setup script to configure your connection. You will be prompted for the ethernet device that is connected to your adsl modem, your username and password, the IPs of your DNS servers and if you need a basic firewall or not.
Code Listing 3.2: Using rp-pppoe |
# adsl-setup # adsl-start |
If something goes wrong, double-check that you correctly typed your username and password by looking at /etc/ppp/pap-secrets or /etc/ppp/chap-secrets and make sure you are using the right ethernet device. If your ethernet device doesn't exist, you will have to load the appropriate network modules. In that case you should continue with Manual Network Configuration as we explain how to load the appropriate network modules there.
If everything worked, continue with Preparing the Disks.
Note: PPTP is only available for the x86 architecture. |
If you need PPTP support, you can use pptpclient which is provided by our Installation CDs. But first you need to make sure that your configuration is correct. Edit /etc/ppp/pap-secrets or /etc/ppp/chap-secrets so it contains the correct username/password combination:
Code Listing 3.3: Editing /etc/ppp/chap-secrets |
# nano -w /etc/ppp/chap-secrets
|
Then adjust /etc/ppp/options.pptp if necessary:
Code Listing 3.4: Editing /etc/ppp/options.pptp |
# nano -w /etc/ppp/options.pptp
|
When all that is done, just run pptp (along with the options you couldn't set in options.pptp) to connect the server:
Code Listing 3.5: Connection to a dial-in server |
# pptp <server ip>
|
Now continue with Preparing the Disks.
3.d. Manual Network Configuration
Loading the Appropriate Network Modules
When the Installation CD boots, it tries to detect all your hardware devices and loads the appropriate kernel modules (drivers) to support your hardware. In the vast majority of cases, it does a very good job. However, in some cases, it may not auto-load the kernel modules you need.
If net-setup or adsl-setup failed, then it is possible that your networkcard wasn't found immediately. This means you may have to load the appropriate kernel modules manually.
To find out what kernel modules we provide for networking, use ls:
Code Listing 4.1: Searching for provided modules |
# ls /lib/modules/`uname -r`/kernel/drivers/net
|
If you find a driver for your network card, use modprobe to load the kernel module:
Code Listing 4.2: Using modprobe to load a kernel module |
(As an example, we load the pcnet32 module) # modprobe pcnet32 |
To check if your network card is now detected, use ifconfig. A detected network card would result in something like this:
Code Listing 4.3: Testing availability of your network card, successful |
# ifconfig eth0
eth0 Link encap:Ethernet HWaddr FE:FD:00:00:00:00
BROADCAST NOARP MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
|
If however you receive the following error, the network card is not detected:
Code Listing 4.4: Testing availability of your network card, failed |
# ifconfig eth0
eth0: error fetching interface information: Device not found
|
If you have multiple network cards in your system they are named eth0, eth1, etc. Make sure that the network card you want to use works well and remember to use the correct naming throughout this document. We will assume that the network card eth0 is used.
Assuming that you now have a detected network card, you can retry net-setup or adsl-setup again (which should work now), but for the hardcore people amongst you we explain how to configure your network manually.
Select one of the following sections based on your network setup:
DHCP (Dynamic Host Configuration Protocol) makes it possible to automatically receive networking information (IP address, netmask, broadcast address, gateway, nameservers etc.). This only works if you have a DHCP server in your network (or if your provider provides a DHCP service). To have a network interface receive this information automatically, use dhcpcd:
Code Listing 4.5: Using dhcpcd |
# dhcpcd eth0 Some network admins require that you use the hostname and domainname provided by the DHCP server. In that case, use # dhcpcd -HD eth0 |
If this works (try pinging some internet server, like Google), then you are all set and ready to continue. Skip the rest of this section and continue with Preparing the Disks.
Note: Support for the iwconfig command is only available on x86, amd64 and ppc Installation CDs. You can still get the extensions working otherwise by following the instructions of the linux-wlan-ng project. |
If you are using a wireless (802.11) card, you may need to configure your wireless settings before going any further. To see the current wireless settings on your card, you can use iwconfig. Running iwconfig might show something like:
Code Listing 4.6: Showing the current wireless settings |
# iwconfig eth0
eth0 IEEE 802.11-DS ESSID:"GentooNode"
Mode:Managed Frequency:2.442GHz Access Point: 00:09:5B:11:CC:F2
Bit Rate:11Mb/s Tx-Power=20 dBm Sensitivity=0/65535
Retry limit:16 RTS thr:off Fragment thr:off
Power Management:off
Link Quality:25/10 Signal level:-51 dBm Noise level:-102 dBm
Rx invalid nwid:5901 Rx invalid crypt:0 Rx invalid frag:0 Tx
excessive retries:237 Invalid misc:350282 Missed beacon:84
|
Note: Some wireless cards may have a device name of wlan0 or ra0 instead of eth0. Run iwconfig without any command-line parameters to determine the correct device name. |
For most users, there are only two settings that might be important to change, the ESSID (aka wireless network name) or the WEP key. If the ESSID and Access Point address listed are already that of your access point and you are not using WEP, then your wireless is working. If you need to change your ESSID, or add a WEP key, you can issue the following commands:
Code Listing 4.7: Changing ESSID and/or adding WEP key |
(This sets the network name to "GentooNode") # iwconfig eth0 essid GentooNode (This sets a hex WEP key) # iwconfig eth0 key 1234123412341234abcd (This sets an ASCII key - prefix it with "s:") # iwconfig eth0 key s:some-password |
You can then confirm your wireless settings again by using iwconfig. Once you have wireless working, you can continue configuring the IP level networking options as described in the next section (Understanding Network Terminology) or use the net-setup tool as described previously.
Understanding Network Terminology
Note: If you know your IP address, broadcast address, netmask and nameservers, then you can skip this subsection and continue with Using ifconfig and route. |
If all above fails, you will have to configure your network manually. Have no fear, it is far from difficult. But we are going to explain a certain amount of networking to you as you will need it to be able to configure your network to your satisfaction. When you're done reading this, you will know what a gateway is, what a netmask serves for, how a broadcast address is formed and why you need nameservers.
In a network, hosts are identified by their IP address (Internet Protocol address). Such an address is a combination of four numbers between 0 and 255. Well, at least that is how we perceive it. In reality, such an IP address consists of 32 bits (ones and zeros). Let's view an example:
Code Listing 4.8: Example of an IP address |
IP Address (numbers): 192.168.0.2
IP Address (bits): 11000000 10101000 00000000 00000010
-------- -------- -------- --------
192 168 0 2
|
Such an IP address is unique to a host as far as all accessible networks are concerned (i.e. all hosts that you are able to reach must have unique IP addresses). To be able to make a distinction between hosts inside a network, and hosts outside a network, the IP address is divided in two parts: the network part and the host part.
The separation is written down with the netmask, a collection of ones followed by a collection of zeros. The part of the IP that can be mapped on the ones is the network-part, the other one is the host-part. As usual, the netmask can be written down as an IP-address.
Code Listing 4.9: Example of network/host separation |
IP-address: 192 168 0 2
11000000 10101000 00000000 00000010
Netmask: 11111111 11111111 11111111 00000000
255 255 255 0
+--------------------------+--------+
Network Host
|
In other words, 192.168.0.14 is still part of our example network, but 192.168.1.2 is not.
The broadcast address is an IP-address with the same network-part as your network, but with only ones as host-part. Every host on your network listens to this IP address. It is truly meant for broadcasting packets.
Code Listing 4.10: Broadcast address |
IP-address: 192 168 0 2
11000000 10101000 00000000 00000010
Broadcast: 11000000 10101000 00000000 11111111
192 168 0 255
+--------------------------+--------+
Network Host
|
To be able to surf on the internet, you must know which host shares the Internet connection. This host is called the gateway. Since it is a regular host, it has a regular IP address (for instance 192.168.0.1).
We previously stated that every host has its own IP address. To be able to reach this host by a name (instead of an IP address) you need a service that translates a name (such as dev.gentoo.org) to an IP address (such as 64.5.62.82). Such a service is called a name service. To use such a service, you must define the necessary name servers in /etc/resolv.conf.
In some cases, your gateway also serves as nameserver. Otherwise you will have to enter the nameservers provided by your ISP.
To summarise, you will need the following information before continuing:
| Network Item | Example |
| Your IP address | 192.168.0.2 |
| Netmask | 255.255.255.0 |
| Broadcast | 192.168.0.255 |
| Gateway | 192.168.0.1 |
| Nameserver(s) | 195.130.130.5, 195.130.130.133 |
Setting up your network consists of three steps. First we assign ourselves an IP address using ifconfig. Then we set up routing to the gateway using route. Then we finish up by placing the nameserver IPs in /etc/resolv.conf.
To assign an IP address, you will need your IP address, broadcast address and netmask. Then execute the following command, substituting ${IP_ADDR} with your IP address, ${BROADCAST} with your broadcast address and ${NETMASK} with your netmask:
Code Listing 4.11: Using ifconfig |
# ifconfig eth0 ${IP_ADDR} broadcast ${BROADCAST} netmask ${NETMASK} up
|
Now set up routing using route. Substitute ${GATEWAY} with your gateway IP address:
Code Listing 4.12: Using route |
# route add default gw ${GATEWAY}
|
Now open /etc/resolv.conf with your favorite editor (in our example, we use nano):
Code Listing 4.13: Creating /etc/resolv.conf |
# nano -w /etc/resolv.conf
|
Now fill in your nameserver(s) using the following as a template. Make sure you substitute ${NAMESERVER1} and ${NAMESERVER2} with the appropriate nameserver addresses:
Code Listing 4.14: /etc/resolv.conf template |
nameserver ${NAMESERVER1}
nameserver ${NAMESERVER2}
|
That's it. Now test your network by pinging some Internet server (like Google). If this works, congratulations then. You are now ready to install Gentoo. Continue with Preparing the Disks.
4.a. Introduction to Block Devices
We'll take a good look at some of the disk-oriented aspects of Gentoo Linux and Linux in general, including Linux filesystems, partitions, and block devices. Then, once you're familiar with the ins and outs of disks and filesystems, you'll be guided through the process of setting up partitions and filesystems for your Gentoo Linux installation.
To begin, we introduce block devices. The most typical block device is probably the one that represents the first SCSI hard disk in a Linux system, namely /dev/sda.
Block devices represent an abstract interface to the disk. User programs can use these block devices to interact with your disk without worrying about whether your drives are IDE, SCSI, or something else. The program can simply address the storage on the disk as a bunch of contiguous, randomly-accessible 512-byte blocks.
Block devices show up as entries in /dev/. Typically, the first SCSI drive is named /dev/sda, the second /dev/sdb, and so on. IDE drives are named similarly, however, they are prefixed by hd- instead of sd-. If you are using IDE drives, the first one will be named /dev/hda, the second /dev/hdb, and so on.
Although it is theoretically possible to use the entire disk to house your Linux system, this is almost never done in practice. Instead, full disk block devices are split up in smaller, more manageable block devices. These are known as partitions or slices.
The first partition on the first SCSI disk is /dev/sda1, the second /dev/sda2 and so on. Similarly, the first two partitions on the first IDE disk are /dev/hda1 and /dev/hda2.
The third partition on Sun systems is set aside as a special "whole disk" slice. This partition must not contain a file system.
Users who are used to the DOS partitioning scheme should note that Sun disklabels do not have "primary" and "extended" partitions. Instead, up to eight partitions are available per drive, with the third of these being reserved.
4.b. Designing a Partitioning Scheme
If you are not interested in drawing up a partitioning scheme, the table below suggests a suitable starting point for most systems. For IDE-based systems, substitute hda for sda in the following.
Note that a separate /boot partition is generally not recommended on SPARC, as it complicates the bootloader configuration.
| Partition | Filesystem | Size | Mount Point | Description |
| /dev/sda1 | ext3 | <2 GByte | / | Root partition. For all sparc32 systems, and sparc64 systems with older OBP versions, this must be less than 2 GBytes in size, and the first partition on the disk. |
| /dev/sda2 | swap | 512 MBytes | none | Swap partition. For bootstrap and certain larger compiles, at least 512 MBytes of RAM (including swap) is required. |
| /dev/sda3 | none | Whole disk | none | Whole disk partition. This is required on SPARC systems. |
| /dev/sda4 | ext3 | at least 2 GBytes | /usr | /usr partition. Applications are installed here. By default this partition is also used for Portage data (which takes around 500 Mbyte excluding source code). |
| /dev/sda5 | ext3 | at least 1GByte | /var | /var partition. Used for program-generated data. By default Portage uses this partition for temporary space whilst compiling. Certain larger applications such as Mozilla and OpenOffice.org can require over 1 GByte of temporary space here when building. |
| /dev/sda6 | ext3 | remaining space | /home | /home partition. Used for users' home directories. |
4.c. Using fdisk to Partition your Disk
The following parts explain how to create the example partition layout described previously, namely:
| Partition | Description |
| /dev/sda1 | / |
| /dev/sda2 | swap |
| /dev/sda3 | whole disk slice |
| /dev/sda4 | /usr |
| /dev/sda5 | /var |
| /dev/sda6 | /home |
Change the partition layout as required. Remember to keep the root partition entirely within the first 2 GBytes of the disk for older systems. There is also a 15-partition limit for SCSI and SATA.
Start fdisk with your disk as argument:
Code Listing 3.1: Starting fdisk |
# fdisk /dev/sda
|
You should be greeted with the fdisk prompt:
Code Listing 3.2: The fdisk prompt |
Command (m for help): |
To view the available partitions, type in p:
Code Listing 3.3: Listing available partitions |
Command (m for help): p
Disk /dev/sda (Sun disk label): 64 heads, 32 sectors, 8635 cylinders
Units = cylinders of 2048 * 512 bytes
Device Flag Start End Blocks Id System
/dev/sda1 0 488 499712 83 Linux native
/dev/sda2 488 976 499712 82 Linux swap
/dev/sda3 0 8635 8842240 5 Whole disk
/dev/sda4 976 1953 1000448 83 Linux native
/dev/sda5 1953 2144 195584 83 Linux native
/dev/sda6 2144 8635 6646784 83 Linux native
|
Note the Sun disk label in the output. If this is missing, the disk is using the DOS-partitioning, not the Sun partitioning. In this case, use s to ensure that the disk has a sun partition table:
Code Listing 3.4: Creating a Sun Disklabel |
Command (m for help): s Building a new sun disklabel. Changes will remain in memory only, until you decide to write them. After that, of course, the previous content won't be recoverable. Drive type ? auto configure 0 custom (with hardware detected defaults) a Quantum ProDrive 80S b Quantum ProDrive 105S c CDC Wren IV 94171-344 d IBM DPES-31080 e IBM DORS-32160 f IBM DNES-318350 g SEAGATE ST34371 h SUN0104 i SUN0207 j SUN0327 k SUN0340 l SUN0424 m SUN0535 n SUN0669 o SUN1.0G p SUN1.05 q SUN1.3G r SUN2.1G s IOMEGA Jaz Select type (? for auto, 0 for custom): 0 Heads (1-1024, default 64): Using default value 64 Sectors/track (1-1024, default 32): Using default value 32 Cylinders (1-65535, default 8635): Using default value 8635 Alternate cylinders (0-65535, default 2): Using default value 2 Physical cylinders (0-65535, default 8637): Using default value 8637 Rotation speed (rpm) (1-100000, default 5400): 10000 Interleave factor (1-32, default 1): Using default value 1 Extra sectors per cylinder (0-32, default 0): Using default value 0 |
You can find the correct values in your disk's documentation. The 'auto configure' option does not usually work.
It's time to delete any existing partitions. To do this, type d and hit Enter. You will then be prompted for the partition number you would like to delete. To delete a pre-existing /dev/sda1, you would type:
Code Listing 3.5: Deleting a partition |
Command (m for help): d Partition number (1-4): 1 |
You should not delete partition 3 (whole disk). This is required. If this partition does not exist, follow the "Creating a Sun Disklabel" instructions above.
After deleting all partitions except the Whole disk slice, you should have a partition layout similar to the following:
Code Listing 3.6: View an empty partition scheme |
Command (m for help): p
Disk /dev/sda (Sun disk label): 64 heads, 32 sectors, 8635 cylinders
Units = cylinders of 2048 * 512 bytes
Device Flag Start End Blocks Id System
/dev/sda3 0 8635 8842240 5 Whole disk
|
We're ready to create the root partition. To do this, type n to create a new partition, then type 1 to create the partition. When prompted for the first cylinder, hit enter. When prompted for the last cylinder, type +512M to create a partition 512MBytes in size. Make sure that the entire root partition must be contained entirely within the first 2GBytes of the disk. You can see output from these steps below:
Code Listing 3.7: Creating a root partition |
Command (m for help): n Partition number (1-8): 1 First cylinder (0-8635): (press Enter) Last cylinder or +size or +sizeM or +sizeK (0-8635, default 8635): +512M |
Now, when you type p, you should see the following partition printout:
Code Listing 3.8: Listing the partition layout |
Command (m for help): p
Disk /dev/sda (Sun disk label): 64 heads, 32 sectors, 8635 cylinders
Units = cylinders of 2048 * 512 bytes
Device Flag Start End Blocks Id System
/dev/sda1 0 488 499712 83 Linux native
/dev/sda3 0 8635 8842240 5 Whole disk
|
Next, let's create the swap partition. To do this, type n to create a new partition, then 2 to create the second partition, /dev/sda2 in our case. When prompted for the first cylinder, hit enter. When prompted for the last cylinder, type +512M to create a partition 512MB in size. After you've done this, type t to set the partition type, and then type in 82 to set the partition type to "Linux Swap". After completing these steps, typing p should display a partition table that looks similar to this:
Code Listing 3.9: Listing of available partitions |
Command (m for help): p
Disk /dev/sda (Sun disk label): 64 heads, 32 sectors, 8635 cylinders
Units = cylinders of 2048 * 512 bytes
Device Flag Start End Blocks Id System
/dev/sda1 0 488 499712 83 Linux native
/dev/sda2 488 976 499712 82 Linux swap
/dev/sda3 0 8635 8842240 5 Whole disk
|
Creating the /usr, /var and /home partitions
Finally, let's create the /usr, /var and /home partitions. As before, type n to create a new partition, then type 4 to create the third partition, /dev/sda4 in our case. When prompted for the first cylinder, hit enter. When prompted for the last cylinder, enter +2048M to create a partition 2 GBytes in size. Repeat this process for sda5 and sda6, using the desired sizes. Once you're done, you should see something like this:
Code Listing 3.10: Listing complete partition table |
Command (m for help): p
Disk /dev/sda (Sun disk label): 64 heads, 32 sectors, 8635 cylinders
Units = cylinders of 2048 * 512 bytes
Device Flag Start End Blocks Id System
/dev/sda1 0 488 499712 83 Linux native
/dev/sda2 488 976 499712 82 Linux swap
/dev/sda3 0 8635 8842240 5 Whole disk
/dev/sda4 976 1953 1000448 83 Linux native
/dev/sda5 1953 2144 195584 83 Linux native
/dev/sda6 2144 8635 6646784 83 Linux native
|
To save your partition layout and exit fdisk, type w:
Code Listing 3.11: Save and exit fdisk |
Command (m for help): w
|
Now that your partitions are created, you can now continue with Creating Filesystems.
Now that your partitions are created, it is time to place a filesystem on them. If you don't care about what filesystem to choose and are happy with what is used as default in this handbook, continue with Applying a Filesystem to a Partition. Otherwise, read on to learn about the available filesystems...
Several filesystems are available, some are known to be stable on the SPARC architecture. Ext2 and ext3, for example, are known to work well. Alternate filesystems may not function correctly.
ext2 is the tried-and-true Linux filesystem. It does not support journaling, which means that periodic checks of ext2 filesystems at startup can be quite time-consuming. There is quite a selection of newer-generation journaled filesystems that can be checked for consistency very quickly at startup, and are therefore generally preferred over their non-journaled counterparts. In general, journaled filesystems prevent long delays when a system is booted and the filesystem is in an inconsistent state.
ext3 is the journaled version of the ext2 filesystem. It provides metadata journaling for fast recovery as well as other enhanced journaling modes like full-data and ordered-data journaling. Ext3 has an additional hashed b-tree indexing option that enables high performance in almost all situations. You can enable this indexing by adding -O dir_index to the mke2fs command. Ext3 makes an excellent and reliable alternative to ext2.
Applying a Filesystem to a Partition
To create a filesystem on a partition or volume, tools specific to the chosen filesystem are available:
| Filesystem | Creation Command |
| ext2 | mke2fs |
| ext3 | mke2fs -j |
| ext3 with hashed b-tree indexing (2.6 kernels only) | mke2fs -j -O dir_index |
For instance, to create the root partition (/dev/sda1 in our example) as ext2, and the /usr, /var, and /home partitions (/dev/sda4, 5 and 6 in our example, respectively) as ext3, you would use:
Code Listing 4.1: Applying a filesystem on a partition |
# mke2fs /dev/sda1 # mke2fs -j /dev/sda4 # mke2fs -j /dev/sda5 # mke2fs -j /dev/sda6 |
mkswap is the command used to initialize swap partitions:
Code Listing 4.2: Creating a Swap signature |
# mkswap /dev/sda2
|
To activate the swap partition, use swapon:
Code Listing 4.3: Activating the swap partition |
# swapon /dev/sda2
|
Create and activate the swap now.
Now that your partitions are initialized and are housing a filesystem, it is time to mount them using the mount command. Don't forget to first create the necessary mount directories for every partition you created. For example:
Warning: Due to a bug in the e2fsprogs package, you need to explicitly use the mount -t ext3 option if you are using an ext3 filesystem. |
Code Listing 5.1: Mounting partitions |
# mount /dev/sda1 /mnt/gentoo (For an ext3 partition:) # mount -t ext3 /dev/sda1 /mnt/gentoo # mkdir /mnt/gentoo/usr # mount /dev/sda4 /mnt/gentoo/usr # mkdir /mnt/gentoo/var # mount /dev/sda5 /mnt/gentoo/var # mkdir /mnt/gentoo/home # mount /dev/sda6 /mnt/gentoo/home |
Note: If you want your /tmp to reside on a separate partition, be sure to change its permissions after mounting: chmod 1777 /mnt/gentoo/tmp. This also holds for /var/tmp. |
We will also have to mount the proc filesystem (a virtual interface with the kernel) on /proc. But first we will need to place our files on the partitions.
Continue with Installing the Gentoo Installation Files.
5.a. Installing a Stage Tarball
Before you continue you need to check your date/time and update it. A misconfigured clock may lead to strange results in the future!
To verify the current date/time, run date:
Code Listing 1.1: Verifying the date/time |
# date
Fri Mar 29 16:21:18 CEST 2005
|
If the date/time displayed is wrong, update it using the date MMDDhhmmYYYY syntax (Month, Day, hour, minute and Year). For instance, to set the date to Mar 29th, 16:21 in the year 2005:
Code Listing 1.2: Setting the date/time |
# date 032916212005
|
If you have configured networking because you need to download a stage3 file for your architecture, continue with Alternative: Using a Stage3 from the Internet. Otherwise read Default: Using a Stage3 from the Installation CD.
5.b. Default: Using a Stage from the Installation CD
The stages on the CD reside in the /mnt/cdrom/stages directory. To see a listing of available stages, use ls:
Code Listing 2.1: List all available stages |
# ls /mnt/cdrom/stages
|
If the system replies with an error, you may need to mount the CD-ROM first:
Code Listing 2.2: Mounting the CD-ROM |
# ls /mnt/cdrom/stages ls: /mnt/cdrom/stages: No such file or directory # mount /dev/cdroms/cdrom0 /mnt/cdrom # ls /mnt/cdrom/stages |
Now go into your Gentoo mountpoint (usually /mnt/gentoo):
Code Listing 2.3: Changing directory to /mnt/gentoo |
# cd /mnt/gentoo
|
We will now extract the stage tarball of your choice. We will do this with the GNU tar tool. Make sure you use the same options (-xvjpf)! The x stands for Extract, the v for Verbose to see what happens during the extraction process (this one is optional), the j for Decompress with bzip2, the p for Preserve permissions and the f to denote that we want to extract a file, not standard input. In the next example, we extract the stage tarball stage3-<subarch>-2005.0.tar.bz2. Be sure to substitute the tarball filename with your stage.
Code Listing 2.4: Extracting the stage tarball |
# tar -xvjpf /mnt/cdrom/stages/stage3-<subarch>-2005.0.tar.bz2
|
Now that the stage is installed, continue with Installing Portage.
5.c. Alternative: Using a Stage from the Internet
Go to the Gentoo mountpoint at which you mounted your filesystems (most likely /mnt/gentoo):
Code Listing 3.1: Going to the Gentoo mountpoint |
# cd /mnt/gentoo
|
Depending on your installation medium, you have a couple of tools available to download a stage. If you have links2 available, then you can immediately surf to the Gentoo mirrorlist and choose a mirror close to you.
If you don't have links2 available you should have lynx at your disposal. If you need to go through a proxy, export the http_proxy and ftp_proxy variables:
Code Listing 3.2: Setting proxy information for lynx |
# export http_proxy="http://proxy.server.com:port" # export ftp_proxy="http://proxy.server.com:port" |
We will now assume that you have links2 at your disposal.
Pick the releases/ directory, followed by your architecture (for instance x86/) and the Gentoo version (2005.0/) to finish up with the stages/ directory. There you should see all available stage files for your architecture (they might be stored within subdirectories named to the individual sub architectures). Select one and press D to download. When you're finished, press Q to quit the browser.
Code Listing 3.3: Surfing to the mirror listing with links2 |
# links2 http://www.gentoo.org/main/en/mirrors.xml (If you need proxy support with links2:) # links2 -http-proxy proxy.server.com:8080 http://www.gentoo.org/main/en/mirrors.xml |
If you want to check the integrity of the downloaded stage tarball, use md5sum and compare the output with the MD5 checksum provided on the mirror. For instance, to check the validity of the x86 stage tarball:
Code Listing 3.4: Example checking integrity of a stage tarball |
# md5sum -c stage3-x86-2005.0.tar.bz2.md5
stage3-x86-2005.0.tar.bz2: OK
|
Now unpack your downloaded stage onto your system. We use GNU's tar to proceed as it is the easiest method:
Code Listing 3.5: Unpacking the stage |
# tar -xvjpf stage3-*.tar.bz2
|
Make sure that you use the same options (-xvjpf). The x stands for Extract, the v for Verbose to see what happens during the extraction process (this one is optional), the j for Decompress with bzip2, the p for Preserve permissions and the f to denote that we want to extract a file, not standard input.
Now that the stage is installed, continue with Installing Portage.
You now have to install a Portage snapshot, a collection of files that inform Portage what software titles you can install, which profiles are available, etc.
Unpack the Snapshot from the Installation CD
To install the snapshot, take a look inside /mnt/cdrom/snapshots/ to see what snapshot is available:
Code Listing 4.1: Checking the /mnt/cdrom/snapshots content |
# ls /mnt/cdrom/snapshots
|
Now extract the snapshot using the following construct. Again, make sure you use the correct options with tar. Also, the -C is with a capital C, not c. In the next example we use portage-<date>.tar.bz2 as the snapshot filename. Be sure to substitute with the name of the snapshot that is on your Installation CD.
Code Listing 4.2: Extracting a Portage snapshot |
# tar -xvjf /mnt/cdrom/snapshots/portage-<date>.tar.bz2 -C /mnt/gentoo/usr
|
You also need to copy over all source code from the Universal Installation CD.
Code Listing 4.3: Copy over source code |
# mkdir /mnt/gentoo/usr/portage/distfiles # cp /mnt/cdrom/distfiles/* /mnt/gentoo/usr/portage/distfiles/ |
5.e. Configuring the Compile Options
To optimize Gentoo, you can set a couple of variables which impact Portage behaviour. All those variables can be set as environment variables (using export) but that isn't permanent. To keep your settings, Portage provides you with /etc/make.conf, a configuration file for Portage. It is this file we will edit now.
Note: A commented listing of all possible variables can be found in /mnt/gentoo/etc/make.conf.example. For a successful Gentoo installation you'll only need to set the variables which are mentioned beneath. |
Fire up your favorite editor (in this guide we use nano) so we can alter the optimization variables we will discuss hereafter.
Code Listing 5.1: Opening /etc/make.conf |
# nano -w /mnt/gentoo/etc/make.conf
|
As you probably noticed, the make.conf.example file is structured in a generic way: commented lines start with "#", other lines define variables using the VARIABLE="content" syntax. The make.conf file uses the same syntax. Several of those variables are discussed next.
Warning: Do not make any modifications to the USE variable if you are performing a stage3 with GRP installation. You can alter the USE variable after having installed the packages you want. Gremlins are known to attack your system if you ignore this warning! |
Warning: Although it might be tempting for non-stage1 users, they should not change the CHOST setting in make.conf. Doing so might render their system unusable. Again: only change this variable if you use a stage1 installation. |
The networkless installation approach only supports stage3 installations. Do not touch the CHOST variable!
The CFLAGS and CXXFLAGS variables define the optimization flags for the gcc C and C++ compiler respectively. Although we define those generally here, you will only have maximum performance if you optimize these flags for each program separately. The reason for this is because every program is different.
In make.conf you should define the optimization flags you think will make your system the most responsive generally. Don't place experimental settings in this variable; too much optimization can make programs behave bad (crash, or even worse, malfunction).
We will not explain all possible optimization options. If you want to know them all, read the GNU Online Manual(s) or the gcc info page (info gcc -- only works on a working Linux system). The make.conf.example file itself also contains lots of examples and information; don't forget to read it too.
A first setting is the -march= flag, which specifies the name of the target architecture. Possible options are described in the make.conf.example file (as comments). For instance, for the x86 Athlon XP architecture:
Code Listing 5.2: The GCC march setting |
# AMD64 users who want to use a native 64 bit system should use -march=k8
-march=athlon-xp
|
A second one is the -O flag (that is a capital O, not a zero), which specifies the gcc optimization class flag. Possible classes are s (for size-optimized), 0 (zero - for no optimizations), 1, 2 or 3 for more speed-optimization flags (every class has the same flags as the one before, plus some extras). For instance, for a class-2 optimization:
Code Listing 5.3: The GCC O setting |
-O2 |
Another popular optimization flag is -pipe (use pipes rather than temporary files for communication between the various stages of compilation).
Mind you that using -fomit-frame-pointer (which doesn't keep the frame pointer in a register for functions that don't need one) might have serious repercussions on the debugging of applications!
When you define the CFLAGS and CXXFLAGS, you should combine several optimization flags, like in the following example:
Code Listing 5.4: Defining the CFLAGS and CXXFLAGS variable |
CFLAGS="-march=athlon-xp -pipe -O2" # AMD64 users use march=k8 CXXFLAGS="${CFLAGS}" # Use the same settings for both variables |
With MAKEOPTS you define how many parallel compilations should occur when you install a package. A good choice is the number of CPUs in your system plus one, but this guideline isn't always perfect.
Code Listing 5.5: MAKEOPTS for a regular, 1-CPU system |
MAKEOPTS="-j2" |
Update your /mnt/gentoo/etc/make.conf to your own preference and save (nano users would hit Ctrl-X). You are now ready to continue with Chrooting into the Gentoo Base System.
Mount the /proc filesystem on /mnt/gentoo/proc to allow the installation to use the kernel-provided information even within the chrooted environment.
Code Listing 1.1: Mounting /proc |
# mount -t proc none /mnt/gentoo/proc
|
Optional: Copy over DNS Information
If you configured your network to fetch the appropriate stage file later on from the Internet, you need to copy over the DNS information stored in /etc/resolv.conf to /mnt/gentoo/etc/resolv.conf. This file contains the nameservers your system will use to resolve names to IP addresses.
Code Listing 1.2: Copy over DNS Information |
# cp -L /etc/resolv.conf /mnt/gentoo/etc/resolv.conf
|
Now that all partitions are initialized and the base environment installed, it is time to enter our new installation environment by chrooting into it. This means that we change from the current installation environment to your installation system (namely the initialized partitions).
This chrooting is done in three steps. First we will change the root from / (on the installation medium) to /mnt/gentoo (on your partitions) using chroot. Then we will create a new environment using env-update, which essentially creates environment variables. Finally, we load those variables into memory using source.
Code Listing 1.3: Chrooting into the new environment |
# chroot /mnt/gentoo /bin/bash # env-update * Caching service dependencies... # source /etc/profile |
Congratulations! You are now inside your own Gentoo Linux environment. Of course it is far from finished, which is why the installation still has some sections left :-)
6.b. Configuring the USE Variable
USE is one of the most powerful variables Gentoo provides to its users. Several programs can be compiled with or without optional support for certain items. For instance, some programs can be compiled with gtk-support, or with qt-support. Others can be compiled with or without SSL support. Some programs can even be compiled with framebuffer support (svgalib) instead of X11 support (X-server).
Most distributions compile their packages with support for as much as possible, increasing the size of the programs and startup time, not to mention an enormous amount of dependencies. With Gentoo you can define what options a package should be compiled with. This is where USE comes into play.
In the USE variable you define keywords which are mapped onto compile-options. For instance, ssl will compile ssl-support in the programs that support it. -X will remove X-server support (note the minus sign in front). gnome gtk -kde -qt will compile your programs with gnome (and gtk) support, and not with kde (and qt) support, making your system fully tweaked for GNOME.
Warning: Do not make any modifications to the USE variable yet if you plan to use our prebuilt packages (GRP set). You can alter the USE variable after having installed the packages you want. Gremlins are known to attack your system if you ignore this warning! |
The default USE settings are placed in /etc/make.profile/make.defaults. What you place in /etc/make.conf is calculated against these defaults settings. If you add something to the USE setting, it is added to the default list. If you remove something from the USE setting (by placing a minus sign in front of it) it is removed from the default list (if it was in the default list at all). Never alter anything inside the /etc/make.profile directory; it gets overwritten when you update Portage!
A full description on USE can be found in the second part of the Gentoo Handbook, USE flags. A full description on the available USE flags can be found on your system in /usr/portage/profiles/use.desc.
Code Listing 2.1: Viewing available USE flags |
# less /usr/portage/profiles/use.desc (You can scroll using your arrow keys, exit by pressing 'q') |
As an example we show a USE setting for a KDE-based system with DVD, ALSA and CD Recording support:
Code Listing 2.2: Opening /etc/make.conf |
# nano -w /etc/make.conf
|
Code Listing 2.3: USE setting |
USE="-gtk -gnome qt kde dvd alsa cdr" |
You first need to select your timezone so that your system knows where it is located. Look for your timezone in /usr/share/zoneinfo, then make a symlink to /etc/localtime using ln:
Code Listing 1.1: Setting the timezone information |
# ls /usr/share/zoneinfo (Suppose you want to use GMT) # ln -sf /usr/share/zoneinfo/GMT /etc/localtime |
The core around which all distributions are built is the Linux kernel. It is the layer between the user programs and your system hardware. Gentoo provides its users several possible kernel sources. A full listing with description is available at the Gentoo Kernel Guide.
For sparc-based systems we have sparc-sources (kernel source optimized for SPARC users) and vanilla-sources (the default kernel source as developed by the linux-kernel developers).
Choose your kernel source and install it using emerge.
In the next example we install the sparc-sources. Of course substitute with your choice of sources, this is merely an example:
Code Listing 2.1: Installing a kernel source |
# emerge sparc-sources
|
When you take a look in /usr/src you should see a symlink called linux pointing to your kernel source:
Code Listing 2.2: Viewing the kernel source symlink |
# ls -l /usr/src/linux
lrwxrwxrwx 1 root root 12 Oct 13 11:04 /usr/src/linux -> linux-2.4.29-sparc
|
If this isn't the case (i.e. the symlink points to a different kernel source) change the symlink before you continue:
Code Listing 2.3: Changing the kernel source symlink |
# rm /usr/src/linux # cd /usr/src # ln -s linux-2.4.29-sparc linux |
Now it is time to configure and compile your kernel source.
Manually configuring a kernel is often seen as the most difficult procedure a Linux user ever has to perform. Nothing is less true -- after configuring a couple of kernels you don't even remember that it was difficult ;)
However, one thing is true: you must know your system when you start configuring a kernel manually. Most information can be gathered by emerging pciutils (emerge pciutils) which contains lspci. You will now be able to use lspci within the chrooted environment. You may safely ignore any pcilib warnings (like pcilib: cannot open /sys/bus/pci/devices) that lspci throws out. Alternatively, you can run lspci from a non-chrooted environment. The results are the same. You can also run lsmod to see what kernel modules the Installation CD uses (it might provide you with a nice hint on what to enable).
Now go to your kernel source directory and execute make menuconfig. This will fire up an ncurses-based configuration menu.
Code Listing 3.1: Invoking menuconfig |
# cd /usr/src/linux # make menuconfig |
You will be greeted with several configuration sections. We'll first list some options you must activate (otherwise Gentoo will not function, or not function properly without additional tweaks).
First of all, activate the use of development and experimental code/drivers. You need this, otherwise some very important code/drivers won't show up:
Code Listing 3.2: Selecting experimental code/drivers |
Code maturity level options ---> [*] Prompt for development and/or incomplete code/drivers |
Now go to File Systems and select support for the filesystems you use. Don't compile them as modules, otherwise your Gentoo system will not be able to mount your partitions. Also select Virtual memory, /proc file system, /dev file system + Automatically mount at boot:
Code Listing 3.3: Selecting necessary file systems |
File systems --->
[*] Virtual memory file system support (former shm fs)
[*] /proc file system support
[*] /dev file system support (EXPERIMENTAL)
[*] Automatically mount at boot
[ ] /dev/pts file system for Unix98 PTYs
(Select one or more of the following options as needed by your system)
<*> Ext3 journalling file system support
<*> Second extended fs support
|
If you are using PPPoE to connect to the Internet or you are using a dial-up modem, you will need the following options in the kernel:
Code Listing 3.4: Selecting PPPoE necessary drivers |
Network device support ---> <*> PPP (point-to-point protocol) support <*> PPP support for async serial ports <*> PPP support for sync tty ports |
The two compression options won't harm but are not definitely needed, neither does the PPP over Ethernet option, that might only be used by rp-pppoe when configured to do kernel mode PPPoE.
Now activate the correct bus-support:
Code Listing 3.5: Activating SBUS/UPA |
Console drivers --->
Frame-buffer support --->
[*] SBUS and UPA framebuffers
[*] Creator/Creator3D support (Only for UPA slot adapter used in many Ultras)
[*] CGsix (GX,TurboGX) support (Only for SBUS slot adapter used in many SPARCStations)
|
Of course you want support for the OBP:
Code Listing 3.6: Activating OBP Support |
Misc Linux/SPARC drivers ---> [*] /dev/openprom device support |
You will also need SCSI-specific support:
Code Listing 3.7: Activating SCSI-specific support |
SCSI support --->
SCSI low-level drivers --->
<*> Sparc ESP Scsi Driver (Only for SPARC ESP on-board SCSI adapter)
<*> PTI Qlogic, ISP Driver (Only for SBUS SCSI controllers from PTI or QLogic)
<*> SYM53C8XX Version 2 SCSI support (Only for Ultra 60 on-board SCSI adapter)
|
To support your network card, select one of the following:
Code Listing 3.8: Activating networking support |
Network device support --->
Ethernet (10 or 100Mbit) --->
<*> Sun LANCE support (Only for SPARCStation, older Ultra systems, and as Sbus option)
<*> Sun Happy Meal 10/100baseT support (Only for Ultra; also supports "qfe" quad-ethernet on PCI and Sbus)
<*> DECchip Tulip (dc21x4x) PCI support (For some Netras, like N1)
Ethernet (1000Mbit) --->
<*> Broadcom Tigon3 support (Modern Netra, Sun Fire machines)
|
When you have a 4-port Ethernet machine (10/100 or 10/100/1000) the port order is different from the one used by Solaris. You can use sys-apps/ethtool or mii-tool to check the port link status.
When you're done configuring your kernel, continue with Compiling and Installing. However, after having compiled the kernel, check its size:
Code Listing 3.9: Check kernel size |
# ls -lh vmlinux
-rw-r--r-- 1 root root 2.4M Oct 25 14:38 vmlinux
|
If the (uncompressed) size is bigger than 2.5 MB (for Sparc32) or 3.5 MB (for Sparc64), reconfigure your kernel until it doesn't exceed these limits. One way of accomplishing this is by having most kernel drivers compiled as modules. Ignoring this can lead to a non-booting kernel.
Also, if your kernel is just a tad too big, you can try stripping it using the strip command:
Code Listing 3.10: Stripping the kernel |
# strip -R .comment -R .note vmlinux
|
Now that your kernel is configured, it is time to compile and install it. Exit the configuration and start the compilation process:
Code Listing 3.11: Compiling the kernel |
(sparc32) # make dep && make clean vmlinux modules modules_install (sparc64) # make dep && make clean vmlinux image modules modules_install |
When the kernel has finished compiling, copy the kernel image to /boot.
Code Listing 3.12: Installing the kernel |
(sparc32) # cp vmlinux /boot/kernel-2.4.29 (sparc64) # cp arch/sparc64/boot/image /boot/kernel-2.4.29 (Don't forget to copy over the System.map file) # cp System.map /boot |
It is also wise to copy over your kernel configuration file to /boot, just in case :)
Code Listing 3.13: Backing up your kernel configuration |
# cp .config /boot/config-2.4.29
|
Now continue with Installing Separate Kernel Modules.
7.d. Installing Separate Kernel Modules
You should list the modules you want automatically loaded in /etc/modules.autoload.d/kernel-2.4. You can add extra options to the modules too if you want.
To view all available modules, run the following find command. Don't forget to substitute "<kernel version>" with the version of the kernel you just compiled:
Code Listing 4.1: Viewing all available modules |
# find /lib/modules/<kernel version>/ -type f -iname '*.o' -or -iname '*.ko'
|
For instance, to automatically load the 3c59x.o module, edit the kernel-2.4 file and enter the module name in it.
Code Listing 4.2: Editing /etc/modules.autoload.d/kernel-2.4 |
# nano -w /etc/modules.autoload.d/kernel-2.4
|
Code Listing 4.3: /etc/modules.autoload.d/kernel-2.4 |
3c59x |
Continue the installation with Configuring your System.
Under Linux, all partitions used by the system must be listed in /etc/fstab. This file contains the mountpoints of those partitions (where they are seen in the file system structure), how they should be mounted and with what special options (automatically or not, whether users can mount them or not, etc.)
/etc/fstab uses a special syntax. Every line consists of six fields, separated by whitespace (space(s), tabs or a mixture). Each field has its own meaning:
The default /etc/fstab file provided by Gentoo is no valid fstab file, so start nano (or your favorite editor) to create your /etc/fstab:
Code Listing 1.1: Opening /etc/fstab |
# nano -w /etc/fstab
|
Let us take a look at how we write down the options for the /boot partition. This is just an example, so if your architecture doesn't require a /boot partition (such as PPC), don't copy it verbatim.
In our default x86 partitioning example /boot is the /dev/hda1 partition, with ext2 as filesystem. It needs to be checked during boot, so we would write down:
Code Listing 1.2: An example /boot line for /etc/fstab |
/dev/hda1 /boot ext2 defaults 1 2 |
Some users don't want their /boot partition to be mounted automatically to improve their system's security. Those people should substitute defaults with noauto. This does mean that you need to manually mount this partition every time you want to use it.
Now, to improve performance, most users would want to add the noatime option as mountoption, which results in a faster system since access times aren't registered (you don't need those generally anyway):
Code Listing 1.3: An improved /boot line for /etc/fstab |
/dev/hda1 /boot ext2 defaults,noatime 1 2 |
If we continue with this, we would end up with the following three lines (for /boot, / and the swap partition):
Code Listing 1.4: Three /etc/fstab lines |
/dev/hda1 /boot ext2 defaults,noatime 1 2 /dev/hda2 none swap sw 0 0 /dev/hda3 / ext3 noatime 0 1 |
To finish up, you should add a rule for /proc, tmpfs (required) and for your CD-ROM drive (and of course, if you have other partitions or drives, for those too):
Code Listing 1.5: A full /etc/fstab example |
/dev/hda1 /boot ext2 defaults,noatime 1 2 /dev/hda2 none swap sw 0 0 /dev/hda3 / ext3 noatime 0 1 none /proc proc defaults 0 0 none /dev/shm tmpfs nodev,nosuid,noexec 0 0 /dev/cdroms/cdrom0 /mnt/cdrom auto noauto,user 0 0 |
auto makes mount guess for the filesystem (recommended for removable media as they can be created with one of many filesystems) and user makes it possible for non-root users to mount the CD.
Now use the above example to create your /etc/fstab. If you are a SPARC-user, you should add the following line to your /etc/fstab too:
Code Listing 1.6: Adding openprom filesystem to /etc/fstab |
none /proc/openprom openpromfs defaults 0 0 |
Double-check your /etc/fstab, save and quit to continue.
One of the choices the user has to make is name his/her PC. This seems to be quite easy, but lots of users are having difficulties finding the appropriate name for their Linux-pc. To speed things up, know that any name you choose can be changed afterwards. For all we care, you can just call your system tux and domain homenetwork.
We use these values in the next examples. First we set the hostname:
Code Listing 2.1: Setting the hostname |
# echo tux > /etc/hostname
|
Second we set the domainname:
Code Listing 2.2: Setting the domainname |
# echo homenetwork > /etc/dnsdomainname
|
If you have a NIS domain (if you don't know what that is, then you don't have one), you need to define that one too:
Code Listing 2.3: Setting the NIS domainname |
# echo nis.homenetwork > /etc/nisdomainname
|
Now add the domainname script to the default runlevel:
Code Listing 2.4: Adding domainname to the default runlevel |
# rc-update add domainname default
|
Before you get that "Hey, we've had that already"-feeling, you should remember that the networking you set up in the beginning of the gentoo installation was just for the installation. Right now you are going to configure networking for your Gentoo system permanently.
All networking information is gathered in /etc/conf.d/net. It uses a straightforward yet not intuitive syntax if you don't know how to set up networking manually. But don't fear, we'll explain everything :)
First open /etc/conf.d/net with your favorite editor (nano is used in this example):
Code Listing 2.5: Opening /etc/conf.d/net for editing |
# nano -w /etc/conf.d/net
|
The first variable you'll find is iface_eth0. It uses the following syntax:
Code Listing 2.6: iface_eth0 syntaxis |
iface_eth0="<your ip address> broadcast <your broadcast address> netmask <your netmask>" |
If you use DHCP (automatic IP retrieval), you should just set iface_eth0 to dhcp. If you use rp-pppoe (e.g. for ADSL), set it to up. If you need to set up your network manually and you're not familiar with all the above terms, please read the section on Understanding Network Terminology if you haven't done so already.
So let us give three examples; the first one uses DHCP, the second one a static IP (192.168.0.2) with netmask 255.255.255.0, broadcast 192.168.0.255 and gateway 192.168.0.1 while the third one just activates the interface for rp-pppoe usage:
Code Listing 2.7: Examples for /etc/conf.d/net |
(For DHCP) iface_eth0="dhcp" # Some network admins require that you use the # hostname and domainname provided by the DHCP server. # In that case, add the following to let dhcpcd use them. # That will override your own hostname and domainname definitions. dhcpcd_eth0="-HD" # If you intend on using NTP to keep your machine clock synchronized, use # the -N option to prevent dhcpcd from overwriting your /etc/ntp.conf file dhcpcd_eth0="-N" (For static IP) iface_eth0="192.168.0.2 broadcast 192.168.0.255 netmask 255.255.255.0" gateway="eth0/192.168.0.1" (For rp-pppoe) iface_eth0="up" |
If you have several network interfaces, create extra iface_eth variables, like iface_eth1, iface_eth2 etc. The gateway variable shouldn't be reproduced as you can only set one gateway per computer.
Now save the configuration and exit to continue.
Automatically Start Networking at Boot
To have your network interfaces activated at boot, you need to add them to the default runlevel. If you have PCMCIA interfaces you should skip this action as the PCMCIA interfaces are started by the PCMCIA init script.
Code Listing 2.8: Adding net.eth0 to the default runlevel |
# rc-update add net.eth0 default
|
If you have several network interfaces, you need to create the appropriate net.eth1, net.eth2 etc. initscripts for those. You can use ln to do this:
Code Listing 2.9: Creating extra initscripts |
# cd /etc/init.d # ln -s net.eth0 net.eth1 # rc-update add net.eth1 default |
Writing Down Network Information
You now need to inform Linux about your network. This is defined in /etc/hosts and helps in resolving hostnames to IP addresses for hosts that aren't resolved by your nameserver. For instance, if your internal network consists of three PCs called jenny (192.168.0.5), benny (192.168.0.6) and tux (192.168.0.7 - this system) you would open /etc/hosts and fill in the values:
Code Listing 2.10: Opening /etc/hosts |
# nano -w /etc/hosts
|
Code Listing 2.11: Filling in the networking information |
127.0.0.1 localhost 192.168.0.5 jenny.homenetwork jenny 192.168.0.6 benny.homenetwork benny 192.168.0.7 tux.homenetwork tux |
If your system is the only system (or the nameservers handle all name resolution) a single line is sufficient. For instance, if you want to call your system tux:
Code Listing 2.12: /etc/hosts for lonely or fully integrated PCs |
127.0.0.1 localhost tux |
Save and exit the editor to continue.
If you don't have PCMCIA, you can now continue with System Information. PCMCIA-users should read the following topic on PCMCIA.
Note: pcmcia-cs is only available for x86, amd64 and ppc platforms. |
PCMCIA-users should first install the pcmcia-cs package. This also includes users who will be working with a 2.6 kernel (even though they won't be using the PCMCIA drivers from this package).
Code Listing 2.13: Installing pcmcia-cs |
# emerge pcmcia-cs
|
When pcmcia-cs is installed, add pcmcia to the default runlevel:
Code Listing 2.14: Adding pcmcia to the default runlevel |
# rc-update add pcmcia default
|
First we set the root password by typing:
Code Listing 3.1: Setting the root password |
# passwd
|
If you want root to be able to log on through the serial console, add tts/0 to /etc/securetty:
Code Listing 3.2: Adding tts/0 to /etc/securetty |
# echo "tts/0" >> /etc/securetty
|
Gentoo uses /etc/rc.conf for general, system-wide configuration. Open up /etc/rc.conf and enjoy all the comments in that file :)
Code Listing 3.3: Opening /etc/rc.conf |
# nano -w /etc/rc.conf
|
As you can see, this file is well commented to help you set up the necessary configuration variables. Take special care with the KEYMAP setting: if you select the wrong KEYMAP you will get weird results when typing on your keyboard.
Note: Users of USB-based SPARC systems and SPARC clones might need to select an i386 keymap (such as "us") instead of "sunkeymap". |
PPC uses x86 keymaps on most systems. Users who want to be able to use ADB keymaps on boot have to enable ADB keycode sendings in their kernel and have to set a mac/ppc keymap in rc.conf.
If your hardware clock is not using UTC, you need to add CLOCK="local" to the file. Otherwise you will notice some clock skew.
When you're finished configuring /etc/rc.conf, save and exit, then continue with Installing Necessary System Tools.
The first tool you need to decide on has to provide logging facilities for your system. Unix and Linux have an excellent history of logging capabilities -- if you want you can log everything that happens on your system in logfiles. This happens through the system logger.
Gentoo offers several system loggers to choose from. There are sysklogd, which is the traditional set of system logging daemons, syslog-ng, an advanced system logger, and metalog which is a highly-configurable system logger. Others might be available through Portage as well - our number of available packages increases on a daily basis.
If you plan on using sysklogd or syslog-ng you might want to install logrotate afterwards as those system loggers don't provide any rotation mechanism for the log files.
To install the system logger of your choice, emerge it and have it added to the default runlevel using rc-update. The following example installs syslog-ng. Of course substitute with your system logger:
Code Listing 1.1: Installing a system logger |
# emerge syslog-ng # rc-update add syslog-ng default |
Next is the cron daemon. Although it is optional and not required for your system, it is wise to install one. But what is a cron daemon? A cron daemon executes scheduled commands. It is very handy if you need to execute some command regularly (for instance daily, weekly or monthly).
We only provide vixie-cron for networkless installations. If you want another cron daemon you can wait and install it later on.
Code Listing 2.1: Installing a cron daemon |
# emerge vixie-cron # rc-update add vixie-cron default |
If you want to index your system's files so you are able to quickly locate them using the locate tool, you need to install sys-apps/slocate.
Code Listing 3.1: Installing slocate |
# emerge slocate
|
Depending on what file systems you are using, you need to install the necessary file system utilities (for checking the filesystem integrity, creating additional file systems etc.).
The following table lists the tools you need to install if you use a certain file system. Not all filesystems are available for each and every architecture though.
| File System | Tool | Install Command |
| XFS | xfsprogs | emerge xfsprogs |
| ReiserFS | reiserfsprogs | emerge reiserfsprogs |
| JFS | jfsutils | emerge jfsutils |
If you don't require any additional networking-related tools (such as rp-pppoe or a dhcp client) continue with Configuring the Bootloader.
Optional: Installing a DHCP Client
If you require Gentoo to automatically obtain an IP address for your network interface(s), you need to install dhcpcd (or any other DHCP Client) on your system. If you don't do this now, you might not be able to connect to the internet after the installation!
Code Listing 5.1: Installing dhcpcd |
# emerge dhcpcd
|
Optional: Installing a PPPoE Client
If you need rp-pppoe to connect to the net, you need to install it.
Code Listing 5.2: Installing rp-pppoe |
# USE="-X" emerge rp-pppoe
|
The USE="-X" will prohibit xorg-x11 to be installed as a dependency (rp-pppoe has graphical tools; if you want those enabled, you can recompile rp-pppoe later on or have xorg-x11 installed now -- which takes a long time to compile).
Now continue with Configuring the Bootloader.
Now that your kernel is configured and compiled and the necessary system configuration files are filled in correctly, it is time to install a program that will fire up your kernel when you start the system. Such a program is called a bootloader.
10.b. Installing the SPARC Bootloader: SILO
It is now time to install and configure SILO, the Sparc Improved boot LOader.
Code Listing 2.1: Installing SILO |
# emerge silo
|
Now open up your favorite editor (we use nano as an example) and create /etc/silo.conf.
Code Listing 2.2: Creating /etc/silo.conf |
# nano -w /etc/silo.conf
|
Below you'll find an example silo.conf file. It uses the partitioning scheme we use throughout this book and kernel-2.4.29 as kernelimage.
Code Listing 2.3: Example /etc/silo.conf |
partition = 1 # Boot partition (= root partition) root = /dev/sda1 # Root partition timeout = 150 # Wait 15 seconds before booting the default section image = /boot/kernel-2.4.29 label = linux |
If you use the example silo.conf delivered by Portage, be sure to comment out all lines that you do not need.
If the physical disk on which you want to install SILO (as bootloader) differs from the physical disk on which /etc/silo.conf resides, you must copy over /etc/silo.conf to a partition on that disk. Assuming that /boot is a separate partition on that disk, copy over the configuration file to /boot and run /sbin/silo:
Code Listing 2.4: Only if /boot and the SILO destination are on the same disk |
# cp /etc/silo.conf /boot # /sbin/silo -C /boot/silo.conf /boot/silo.conf appears to be valid |
Otherwise just run /sbin/silo:
Code Listing 2.5: Run silo |
# /sbin/silo
/etc/silo.conf appears to be valid
|
Now continue with Rebooting the System.
Exit the chrooted environment and unmount all mounted partitions. Then type in that one magical command you have been waiting for: reboot.
Code Listing 3.1: Exiting the chroot, unmounting all partitions and rebooting |
# exit cdimage ~# cd cdimage ~# umount /mnt/gentoo/boot /mnt/gentoo/proc /mnt/gentoo cdimage ~# reboot |
Of course, don't forget to remove the bootable CD, otherwise the CD will be booted again instead of your new Gentoo system.
Once rebooted in your Gentoo installation, finish up with Finalizing your Gentoo Installation.
Working as root on a Unix/Linux system is dangerous and should be avoided as much as possible. Therefore it is strongly recommended to add a user for day-to-day use.
The groups the user is member of define what activities the user can perform. The following table lists a number of important groups you might wish to use:
| Group | Description |
| audio | be able to access the audio devices |
| cdrom | be able to directly access optical devices |
| floppy | be able to directly access floppy devices |
| games | be able to play games |
| usb | be able to access USB devices |
| video | be able to access video capturing hardware and doing hardware acceleration |
| wheel | be able to use su |
For instance, to create a user called john who is member of the wheel, users and audio groups, log in as root first (only root can create users) and run useradd:
Code Listing 1.1: Adding a user for day-to-day use |
Login: root Password: (Your root password) # useradd -m -G users,wheel,audio -s /bin/bash john # passwd john Password: (Enter the password for john) Re-enter password: (Re-enter the password to verify) |
If a user ever needs to perform some task as root, they can use su - to temporarily receive root privileges. Another way is to use the sudo package which is, if correctly configured, very secure.
11.b. Optional: Install GRP Packages
Important: This part is for GRP users only. Other users should skip this part and continue with Where to go from here?. |
Now that your system is booted, log on as the user you created (for instance, john) and use su - to gain root privileges:
Code Listing 2.1: Gaining root privileges |
$ su - Password: (Enter your root password) |
Now we need to change the Portage configuration to look for the prebuilt binaries from the second CD (Gentoo Packages CD). First mount this CD:
Code Listing 2.2: Mount the Packages CD |
(Put the Gentoo Packages CD in the CD tray) # mount /mnt/cdrom |
Now configure Portage to use /mnt/cdrom for its prebuilt packages:
Code Listing 2.3: Configuring Portage to use /mnt/cdrom |
# ls /mnt/cdrom (If there is a /mnt/cdrom/packages directory:) # export PKGDIR="/mnt/cdrom/packages" (Otherwise:) # export PKGDIR="/mnt/cdrom" |
Now install the packages you want. The Packages CD contains several prebuilt binaries, for instance KDE:
Code Listing 2.4: Installing KDE |
# emerge --usepkg kde
|
Be sure to install the binaries now. When you do an emerge --sync to update Portage (as you will learn later), the prebuilt binaries might not match against the ebuilds in your updated Portage. You can try to circumvent this by using emerge --usepkgonly instead of emerge --usepkg.
Congratulations, your system is now fully equipped! Continue with Where to go from here? to learn more about Gentoo.
Congratulations! You now have a working Gentoo system. But where to go from here? What are your options now? What to explore first? Gentoo provides its users with lots of possibilities, and therefore lots of documented (and less documented) features.
You should definitely take a look at the next part of the Gentoo Handbook entitled Working with Gentoo which explains how to keep your software up to date, how to install more software, what USE flags are, how the Gentoo Init system works, etc.
If you are interested in optimizing your system for desktop-use, or you want to learn how to configure your system to be a full working desktop system, consult our extensive Gentoo Desktop Documentation Resources.
We also have a Gentoo Security Handbook which is worth reading.
For a full listing of all our available documentation check out our Documentation Resources page.
You are of course always welcome on our Gentoo Forums or on one of our many Gentoo IRC channels.
We also have several mailinglists open to all our users. Information on how to join is contained in that page.
We'll shut up now and let you enjoy your installation :)
12.c. Gentoo Changes since 2005.0
Gentoo is a fast-moving target. The following sections describe important changes that affect a Gentoo installation. We only list those that have anything in common with the installation, not with package changes that did not occur during the installation.
The following changes need to happen right after having updated your system (and before rebooting):
Portage is probably Gentoo's most notable innovation in software management. With its high flexibility and enormous amount of features it is frequently seen as the best software management tool available for Linux.
Portage is completely written in Python and Bash and therefore fully visible to the users as both are scripting languages.
Most users will work with Portage through the emerge tool. This chapter is not meant to duplicate the information available from the emerge man page. For a complete rundown of emerge's options, please consult the man page:
Code Listing 1.1: Reading the emerge man page |
$ man emerge
|
When we talk about packages, we often mean software titles that are available to the Gentoo users through the Portage tree. The Portage tree is a collection of ebuilds, files that contain all information Portage needs to maintain software (install, search, query, ...). These ebuilds reside in /usr/portage by default.
Whenever you ask Portage to perform some action regarding software titles, it will use the ebuilds on your system as a base. It is therefore important that you regularly update the ebuilds on your system so Portage knows about new software, security updates, etc.
The Portage tree is usually updated with rsync, a fast incremental file transfer utility. Updating is fairly simple as the emerge command provides a front-end for rsync:
Code Listing 2.1: Updating the Portage tree |
# emerge --sync
|
If you are unable to rsync due to firewall restrictions you can still update your Portage tree by using our daily generated Portage tree snapshots. The emerge-webrsync tool automatically fetches and installs the latest snapshot on your system:
Code Listing 2.2: Running emerge-webrsync |
# emerge-webrsync
|
To search through the Portage tree after software titles, you can use emerge built-in search capabilities. By default, emerge --search returns the names of packages whose title matches (either fully or partially) the given search term.
For instance, to search for all packages who have "pdf" in their name:
Code Listing 3.1: Searching for pdf-named packages |
$ emerge --search pdf
|
If you want to search through the descriptions as well you can use the --searchdesc (or -S) switch:
Code Listing 3.2: Searching for pdf-related packages |
$ emerge --searchdesc pdf
|
When you take a look at the output, you'll notice that it gives you a lot of information. The fields are clearly labelled so we won't go further into their meanings:
Code Listing 3.3: Example 'emerge --search' output |
* net-print/cups-pdf
Latest version available: 1.5.2
Latest version installed: [ Not Installed ]
Size of downloaded files: 15 kB
Homepage: http://cip.physik.uni-wuerzburg.de/~vrbehr/cups-pdf/
Description: Provides a virtual printer for CUPS to produce PDF files.
License: GPL-2
|
Once you've found a software title to your liking, you can easily install it with emerge: just add the package name. For instance, to install gnumeric:
Code Listing 3.4: Installing gnumeric |
# emerge gnumeric
|
Since many applications depend on each other, any attempt to install a certain software package might result in the installation of several dependencies as well. Don't worry, Portage handles dependencies well. If you want to find out what Portage would install when you ask it to install a certain package, add the --pretend switch. For instance:
Code Listing 3.5: Pretend to install gnumeric |
# emerge --pretend gnumeric
|
When you ask Portage to install a package, it will download the necessary source code from the internet (if necessary) and store it by default in /usr/portage/distfiles. After this it will unpack, compile and install the package. If you want Portage to only download the sources without installing them, add the --fetchonly option to the emerge command:
Code Listing 3.6: Download the sourcecode for gnumeric |
# emerge --fetchonly gnumeric
|
Finding Installed Package Documentation
Many packages come with their own documentation. Sometimes, the doc USE flag determines whether the package documentation should be installed or not. You can check the existence of a doc USE flag with the emerge -vp <package name> command.
Code Listing 3.7: Checking the existence of a doc USE flag |
(alsa-lib is just an example, of course.) # emerge -vp alsa-lib [ebuild N ] media-libs/alsa-lib-1.0.14_rc1 -debug +doc 698 kB |
The best way of enabling the doc USE flag is doing it on a per-package basis via /etc/portage/package.use, so that you get documentation only for packages that you are interested in. Enabling this flag globally is known to cause problems with circular dependencies. For more information, please read the USE Flags chapter.
Once the package installed, its documentation is generally found in a subdirectory named after the package under the /usr/share/doc directory. You can also list all installed files with the equery tool which is part of the app-portage/gentoolkit package.
Code Listing 3.8: Locating package documentation |
# ls -l /usr/share/doc/alsa-lib-1.0.14_rc1 total 28 -rw-r--r-- 1 root root 669 May 17 21:54 ChangeLog.gz -rw-r--r-- 1 root root 9373 May 17 21:54 COPYING.gz drwxr-xr-x 2 root root 8560 May 17 21:54 html -rw-r--r-- 1 root root 196 May 17 21:54 TODO.gz (Alternatively, use equery to locate interesting files:) # equery files alsa-lib | less media-libs/alsa-lib-1.0.14_rc1 * Contents of media-libs/alsa-lib-1.0.14_rc1: /usr /usr/bin /usr/bin/alsalisp (Output truncated) |
When you want to remove a software package from your system, use emerge --unmerge. This will tell Portage to remove all files installed by that package from your system except the configuration files of that application if you have altered those after the installation. Leaving the configuration files allows you to continue working with the package if you ever decide to install it again.
However, a big warning applies: Portage will not check if the package you want to remove is required by another package. It will however warn you when you want to remove an important package that breaks your system if you unmerge it.
Code Listing 3.9: Removing gnumeric from the system |
# emerge --unmerge gnumeric
|
When you remove a package from your system, the dependencies of that package that were installed automatically when you installed the software are left. To have Portage locate all dependencies that can now be removed, use emerge's --depclean functionality. We will talk about this later on.
To keep your system in perfect shape (and not to mention install the latest security updates) you need to update your system regularly. Since Portage only checks the ebuilds in your Portage tree you first have to update your Portage tree. When your Portage tree is updated, you can update your system with emerge --update world. In the next example, we'll also use the --ask switch which will tell Portage to display the list of packages it wants to upgrade and ask you if you want to continue:
Code Listing 3.10: Updating your system |
# emerge --update --ask world
|
Portage will then search for newer version of the applications you have installed. However, it will only verify the versions for the applications you have explicitly installed (the applications listed in /var/lib/portage/world) - it does not thoroughly check their dependencies. If you want to update every single package on your system, add the --deep argument:
Code Listing 3.11: Updating your entire system |
# emerge --update --deep world
|
Since security updates also happen in packages you have not explicitly installed on your system (but that are pulled in as dependencies of other programs), it is recommended to run this command once in a while.
If you have altered any of your USE flags lately you might want to add --newuse as well. Portage will then verify if the change requires the installation of new packages or recompilation of existing ones:
Code Listing 3.12: Performing a full update |
# emerge --update --deep --newuse world
|
Some packages in the Portage tree don't have any real content but are used to install a collection of packages. For instance, the kde package will install a complete KDE environment on your system by pulling in various KDE-related packages as dependencies.
If you ever want to remove such a package from your system, running emerge --unmerge on the package won't have much effect as the dependencies remain on the system.
Portage has the functionality to remove orphaned dependencies as well, but since the availability of software is dynamically dependent you first need to update your entire system fully, including the new changes you applied when changing USE flags. After this you can run emerge --depclean to remove the orphaned dependencies. When this is done, you need to rebuild the applications that were dynamically linked to the now-removed software titles but don't require them anymore.
All this is handled with the following three commands:
Code Listing 3.13: Removing orphaned dependencies |
# emerge --update --deep --newuse world # emerge --depclean # revdep-rebuild |
revdep-rebuild is provided by the gentoolkit package; don't forget to emerge it first:
Code Listing 3.14: Installing the gentoolkit package |
# emerge gentoolkit
|
1.d. When Portage is Complaining...
About SLOTs, Virtuals, Branches, Architectures and Profiles
As we stated before, Portage is extremely powerful and supports many features that other software management tools lack. To understand this, we explain a few aspects of Portage without going into too much detail.
With Portage different versions of a single package can coexist on a system. While other distributions tend to name their package to those versions (like freetype and freetype2) Portage uses a technology called SLOTs. An ebuild declares a certain SLOT for its version. Ebuilds with different SLOTs can coexist on the same system. For instance, the freetype package has ebuilds with SLOT="1" and SLOT="2".
There are also packages that provide the same functionality but are implemented differently. For instance, metalogd, sysklogd and syslog-ng are all system loggers. Applications that rely on the availability of "a system logger" cannot depend on, for instance, metalogd, as the other system loggers are as good a choice as any. Portage allows for virtuals: each system logger provides virtual/syslog so that applications can depend on virtual/syslog.
Software in the Portage tree can reside in different branches. By default your system only accepts packages that Gentoo deems stable. Most new software titles, when committed, are added to the testing branch, meaning more testing needs to be done before it is marked as stable. Although you will see the ebuilds for those software in the Portage tree, Portage will not update them before they are placed in the stable branch.
Some software is only available for a few architectures. Or the software doesn't work on the other architectures, or it needs more testing, or the developer that committed the software to the Portage tree is unable to verify if the package works on different architectures.
Each Gentoo installation adheres to a certain profile which contains, amongst other information, the list of packages that are required for a system to function normally.
Code Listing 4.1: Portage warning about blocked packages (with --pretend) |
[blocks B ] mail-mta/ssmtp (is blocking mail-mta/postfix-2.2.2-r1) |
Code Listing 4.2: Portage warning about blocked packages (without --pretend) |
!!! Error: the mail-mta/postfix package conflicts with another package. !!! both can't be installed on the same system together. !!! Please use 'emerge --pretend' to determine blockers. |
Ebuilds contain specific fields that inform Portage about its dependencies. There are two possible dependencies: build dependencies, declared in DEPEND and run-time dependencies, declared in RDEPEND. When one of these dependencies explicitly marks a package or virtual as being not compatible, it triggers a blockage.
To fix a blockage, you can choose to not install the package or unmerge the conflicting package first. In the given example, you can opt not to install postfix or to remove ssmtp first.
You may also see blocking packages with specific atoms, such as <media-video/mplayer-bin-1.0_rc1-r2. In this case, updating to a more recent version of the blocking package would remove the block.
It is also possible that two packages that are yet to be installed are blocking each other. In this rare case, you should find out why you need to install both. In most cases you can do with one of the packages alone. If not, please file a bug on Gentoo's bugtracking system.
Code Listing 4.3: Portage warning about masked packages |
!!! all ebuilds that could satisfy "bootsplash" have been masked. |
Code Listing 4.4: Portage warning about masked packages - reason |
!!! possible candidates are: - gnome-base/gnome-2.8.0_pre1 (masked by: ~x86 keyword) - lm-sensors/lm-sensors-2.8.7 (masked by: -sparc keyword) - sys-libs/glibc-2.3.4.20040808 (masked by: -* keyword) - dev-util/cvsd-1.0.2 (masked by: missing keyword) - games-fps/unreal-tournament-451 (masked by: package.mask) - sys-libs/glibc-2.3.2-r11 (masked by: profile) |
When you want to install a package that isn't available for your system, you will receive this masking error. You should try installing a different application that is available for your system or wait until the package is put available. There is always a reason why a package is masked:
Code Listing 4.5: Portage warning about missing dependency |
emerge: there are no ebuilds to satisfy ">=sys-devel/gcc-3.4.2-r4". !!! Problem with ebuild sys-devel/gcc-3.4.2-r2 !!! Possibly a DEPEND/*DEPEND problem. |
The application you are trying to install depends on another package that is not available for your system. Please check bugzilla if the issue is known and if not, please report it. Unless you are mixing branches this should not occur and is therefore a bug.
Code Listing 4.6: Portage warning about ambiguous ebuild names |
!!! The short ebuild name "aterm" is ambiguous. Please specify
!!! one of the following fully-qualified ebuild names instead:
dev-libs/aterm
x11-terms/aterm
|
The application you want to install has a name that corresponds with more than one package. You need to supply the category name as well. Portage will inform you of possible matches to choose from.
Code Listing 4.7: Portage warning about circular dependencies |
!!! Error: circular dependencies: ebuild / net-print/cups-1.1.15-r2 depends on ebuild / app-text/ghostscript-7.05.3-r1 ebuild / app-text/ghostscript-7.05.3-r1 depends on ebuild / net-print/cups-1.1.15-r2 |
Two (or more) packages you want to install depend on each other and can therefore not be installed. This is most likely a bug in the Portage tree. Please resync after a while and try again. You can also check bugzilla if the issue is known and if not, report it.
Code Listing 4.8: Portage warning about fetch failed |
!!! Fetch failed for sys-libs/ncurses-5.4-r5, continuing...
(...)
!!! Some fetch errors were encountered. Please see above for details.
|
Portage was unable to download the sources for the given application and will try to continue installing the other applications (if applicable). This failure can be due to a mirror that has not synchronised correctly or because the ebuild points to an incorrect location. The server where the sources reside can also be down for some reason.
Retry after one hour to see if the issue still persists.
Code Listing 4.9: Portage warning about profile-protected package |
!!! Trying to unmerge package(s) in system profile. 'sys-apps/portage' !!! This could be damaging to your system. |
You have asked to remove a package that is part of your system's core packages. It is listed in your profile as required and should therefore not be removed from the system.
Sometimes, when you attempt to emerge a package, it will fail with the message:
Code Listing 4.10: Digest verification failure |
>>> checking ebuild checksums !!! Digest verification failed: |
This is a sign that something is wrong with the Portage tree -- often, it is because a developer may have made a mistake when committing a package to the tree.
When the digest verification fails, do not try to re-digest the package yourself. Running ebuild foo manifest will not fix the problem; it will almost certainly make it worse!
Instead, wait an hour or two for the tree to settle down. It's likely that the error was noticed right away, but it can take a little time for the fix to trickle down the Portage tree. While you're waiting, check Bugzilla and see if anyone has reported the problem yet. If not, go ahead and file a bug for the broken package.
Once you see that the bug has been fixed, you may want to re-sync to pick up the fixed digest.
Important: This does not mean that you can re-sync your tree multiple times! As stated in the rsync policy (when you run emerge --sync), users who sync too often will be banned! In fact, it's better to just wait until your next scheduled sync, so that you don't overload the rsync servers. |
When you are installing Gentoo (or any other distribution, or even operating system for that matter) you make choices depending on the environment you are working with. A setup for a server differs from a setup for a workstation. A gaming workstation differs from a 3D rendering workstation.
This is not only true for choosing what packages you want to install, but also what features a certain package should support. If you don't need OpenGL, why would you bother installing OpenGL and build OpenGL support in most of your packages? If you don't want to use KDE, why would you bother compiling packages with KDE support if those packages work flawlessly without?
To help users in deciding what to install/activate and what not, we wanted the user to specify his/her environment in an easy way. This forces the user into deciding what they really want and eases the process for Portage, our package management system, to make useful decisions.
Enter the USE flags. Such a flag is a keyword that embodies support and dependency-information for a certain concept. If you define a certain USE flag, Portage will know that you want support for the chosen keyword. Of course this also alters the dependency information for a package.
Let us take a look at a specific example: the kde keyword. If you do not have this keyword in your USE variable, all packages that have optional KDE support will be compiled without KDE support. All packages that have an optional KDE dependency will be installed without installing the KDE libraries (as dependency). If you have defined the kde keyword, then those packages will be compiled with KDE support, and the KDE libraries will be installed as dependency.
By correctly defining the keywords you will receive a system tailored specifically to your needs.
There are two types of USE flags: global and local USE flags.
A list of available global USE flags can be found online or locally in /usr/portage/profiles/use.desc.
A list of available local USE flags can be found locally in /usr/portage/profiles/use.local.desc.
In the hope you are convinced of the importance of USE flags we will now inform you how to declare USE flags.
As previously mentioned, all USE flags are declared inside the USE variable. To make it easy for users to search and pick USE flags, we already provide a default USE setting. This setting is a collection of USE flags we think are commonly used by the Gentoo users. This default setting is declared in the make.defaults files part of your profile.
The profile your system listens to is pointed to by the /etc/make.profile symlink. Each profile works on top of another, larger profile, the end result is therefore the sum of all profiles. The top profile is the base profile (/usr/portage/profiles/base).
Let us take a look at this default setting for the 2004.3 profile:
Code Listing 2.1: Cumulative make.defaults USE variable for the 2004.3 profile |
(This example is the sum of the settings in base, default-linux,
default-linux/x86 and default-linux/x86/2004.3)
USE="x86 oss apm arts avi berkdb bitmap-fonts crypt cups encode fortran f77
foomaticdb gdbm gif gpm gtk imlib jpeg kde gnome libg++ libwww mad
mikmod motif mpeg ncurses nls oggvorbis opengl pam pdflib png python qt
quicktime readline sdl spell ssl svga tcpd truetype X xml2 xmms xv zlib"
|
As you can see, this variable already contains quite a lot of keywords. Do not alter any make.defaults file to tailor the USE variable to your needs: changes in this file will be undone when you update Portage!
To change this default setting, you need to add or remove keywords to the USE variable. This is done globally by defining the USE variable in /etc/make.conf. In this variable you add the extra USE flags you require, or remove the USE flags you don't want. This latter is done by prefixing the keyword with the minus-sign ("-").
For instance, to remove support for KDE and QT but add support for ldap, the following USE can be defined in /etc/make.conf:
Code Listing 2.2: An example USE setting in /etc/make.conf |
USE="-kde -qt3 -qt4 ldap" |
Declaring USE flags for individual packages
Sometimes you want to declare a certain USE flag for one (or a couple) of applications but not system-wide. To accomplish this, you will need to create the /etc/portage directory (if it doesn't exist yet) and edit /etc/portage/package.use. This is usually a single file, but can also be a directory; see man portage for more information. The following examples assume package.use is a single file.
For instance, if you don't want berkdb support globally but you do want it for mysql, you would add:
Code Listing 2.3: /etc/portage/package.use example |
dev-db/mysql berkdb |
You can of course also explicitly disable USE flags for a certain application. For instance, if you don't want java support in PHP:
Code Listing 2.4: /etc/portage/package.use 2nd example |
dev-php/php -java |
Sometimes you want to set a certain USE setting only once. Instead of editing /etc/make.conf twice (to do and undo the USE changes) you can just declare the USE variable as environment variable. Remember that, when you re-emerge or update this application (either explicitly or as part of a system update) your changes will be lost!
As an example we will temporarily remove java from the USE setting during the installation of seamonkey.
Code Listing 2.5: Using USE as environment variable |
# USE="-java" emerge seamonkey
|
Of course there is a certain precedence on what setting has priority over the USE setting. You don't want to declare USE="-java" only to see that java is still used due to a setting that has a higher priority. The precedence for the USE setting is, ordered by priority (first has lowest priority):
To view the final USE setting as seen by Portage, run emerge --info. This will list all relevant variables (including the USE variable) with the content used by Portage.
Code Listing 2.6: Running emerge --info |
# emerge --info
|
Adapting your Entire System to New USE Flags
If you have altered your USE flags and you wish to update your entire system to use the new USE flags, use emerge's --newuse option:
Code Listing 2.7: Rebuilding your entire system |
# emerge --update --deep --newuse world
|
Next, run Portage's depclean to remove the conditional dependencies that were emerged on your "old" system but that have been obsoleted by the new USE flags.
Warning: Running emerge --depclean is a dangerous operation and should be handled with care. Double-check the provided list of "obsoleted" packages to make sure it doesn't remove packages you need. In the following example we add the -p switch to have depclean only list the packages without removing them. |
Code Listing 2.8: Removing obsoleted packages |
# emerge -p --depclean
|
When depclean has finished, run revdep-rebuild to rebuild the applications that are dynamically linked against shared objects provided by possibly removed packages. revdep-rebuild is part of the gentoolkit package; don't forget to emerge it first.
Code Listing 2.9: Running revdep-rebuild |
# revdep-rebuild
|
When all this is accomplished, your system is using the new USE flag settings.
2.c. Package specific USE flags
Let us take the example of seamonkey: what USE flags does it listen to? To find out, we use emerge with the --pretend and --verbose options:
Code Listing 3.1: Viewing the used USE flags |
# emerge --pretend --verbose seamonkey
These are the packages that I would merge, in order:
Calculating dependencies ...done!
[ebuild R ] www-client/seamonkey-1.0.7 USE="crypt gnome java -debug -ipv6
-ldap -mozcalendar -mozdevelop -moznocompose -moznoirc -moznomail -moznopango
-moznoroaming -postgres -xinerama -xprint" 0 kB
|
emerge isn't the only tool for this job. In fact, we have a tool dedicated to package information called equery which resides in the gentoolkit package. First, install gentoolkit:
Code Listing 3.2: Installing gentoolkit |
# emerge gentoolkit
|
Now run equery with the uses argument to view the USE flags of a certain package. For instance, for the gnumeric package:
Code Listing 3.3: Using equery to view used USE flags |
# equery --nocolor uses =gnumeric-1.6.3 -a
[ Searching for packages matching =gnumeric-1.6.3... ]
[ Colour Code : set unset ]
[ Legend : Left column (U) - USE flags from make.conf ]
[ : Right column (I) - USE flags packages was installed with ]
[ Found these USE variables for app-office/gnumeric-1.6.3 ]
U I
- - debug : Enable extra debug codepaths, like asserts and extra output.
If you want to get meaningful backtraces see
http://www.gentoo.org/proj/en/qa/backtraces.xml .
+ + gnome : Adds GNOME support
+ + python : Adds support/bindings for the Python language
- - static : !!do not set this during bootstrap!! Causes binaries to be
statically linked instead of dynamically
|
Portage has several additional features that makes your Gentoo experience even better. Many of these features rely on certain software tools that improve performance, reliability, security, ...
To enable or disable certain Portage features you need to edit /etc/make.conf's FEATURES variable which contains the various feature keywords, separated by white space. In several cases you will also need to install the additional tool on which the feature relies.
Not all features that Portage supports are listed here. For a full overview, please consult the make.conf man page:
Code Listing 1.1: Consulting the make.conf man page |
$ man make.conf
|
To find out what FEATURES are default set, run emerge --info and search for the FEATURES variable or grep it out:
Code Listing 1.2: Finding out the FEATURES that are already set |
$ emerge --info | grep FEATURES
|
distcc is a program to distribute compilations across several, not necessarily identical, machines on a network. The distcc client sends all necessary information to the available distcc servers (running distccd) so they can compile pieces of source code for the client. The net result is a faster compilation time.
You can find more information about distcc (and how to have it work with Gentoo) in our Gentoo Distcc Documentation.
Distcc ships with a graphical monitor to monitor tasks that your computer is sending away for compilation. If you use Gnome then put 'gnome' in your USE variable. However, if you don't use Gnome and would still like to have the monitor then you should put 'gtk' in your USE variable.
Code Listing 2.1: Installing distcc |
# emerge distcc
|
Add distcc to the FEATURES variable inside /etc/make.conf. Next, edit the MAKEOPTS variable to your liking. A known guideline is to fill in "-jX" with X the number of CPUs that run distccd (including the current host) plus one, but you might have better results with other numbers.
Now run distcc-config and enter the list of available distcc servers. For a simple example we assume that the available DistCC servers are 192.168.1.102 (the current host), 192.168.1.103 and 192.168.1.104 (two "remote" hosts):
Code Listing 2.2: Configuring distcc to use three available distcc servers |
# distcc-config --set-hosts "192.168.1.102 192.168.1.103 192.168.1.104"
|
Don't forget to run the distccd daemon as well:
Code Listing 2.3: Starting the distccd daemons |
# rc-update add distccd default # /etc/init.d/distccd start |
ccache is a fast compiler cache. When you compile a program, it will cache intermediate results so that, whenever you recompile the same program, the compilation time is greatly reduced. In common compilations this can result in 5 to 10 times faster compilation times.
If you are interested in the ins and outs of ccache, please visit the ccache homepage.
To install ccache, run emerge ccache:
Code Listing 3.1: Installing ccache |
# emerge ccache
|
Open /etc/make.conf and add ccache to the FEATURES variable. Next, add a new variable called CCACHE_SIZE and set it to "2G":
Code Listing 3.2: Editing CCACHE_SIZE in /etc/make.conf |
CCACHE_SIZE="2G" |
To check if ccache functions, ask ccache to provide you with its statistics. Because Portage uses a different ccache home directory, you need to set the CCACHE_DIR variable as well:
Code Listing 3.3: Viewing ccache statistics |
# CCACHE_DIR="/var/tmp/ccache" ccache -s
|
The /var/tmp/ccache location is Portage' default ccache home directory; if you want to alter this setting you can set the CCACHE_DIR variable in /etc/make.conf.
However, if you would run ccache, it would use the default location of ${HOME}/.ccache, which is why you needed to set the CCACHE_DIR variable when asking for the (Portage) ccache statistics.
Using ccache for non-Portage C Compiling
If you would like to use ccache for non-Portage compilations, add /usr/lib/ccache/bin to the beginning of your PATH variable (before /usr/bin). This can be accomplished by editing .bash_profile in your user's home directory. Using .bash_profile is one way to define PATH variables.
Code Listing 3.4: Editing .bash_profile |
PATH="/usr/lib/ccache/bin:/opt/bin:${PATH}"
|
Portage supports the installation of prebuilt packages. Even though Gentoo does not provide prebuilt packages by itself (except for the GRP snapshots) Portage can be made fully aware of prebuilt packages.
To create a prebuilt package you can use quickpkg if the package is already installed on your system, or emerge with the --buildpkg or --buildpkgonly options.
If you want Portage to create prebuilt packages of every single package you install, add buildpkg to the FEATURES variable.
More extended support for creating prebuilt package sets can be obtained with catalyst. For more information on catalyst please read the Catalyst Frequently Asked Questions.
Although Gentoo doesn't provide one, you can create a central repository where you store prebuilt packages. If you want to use this repository, you need to make Portage aware of it by having the PORTAGE_BINHOST variable point to it. For instance, if the prebuilt packages are on ftp://buildhost/gentoo:
Code Listing 4.1: Setting PORTAGE_BINHOST in /etc/make.conf |
PORTAGE_BINHOST="ftp://buildhost/gentoo" |
When you want to install a prebuilt package, add the --getbinpkg option to the emerge command alongside of the --usepkg option. The former tells emerge to download the prebuilt package from the previously defined server while the latter asks emerge to try to install the prebuilt package first before fetching the sources and compiling it.
For instance, to install gnumeric with prebuilt packages:
Code Listing 4.2: Installing the gnumeric prebuilt package |
# emerge --usepkg --getbinpkg gnumeric
|
More information about emerge's prebuilt package options can be found in the emerge man page:
Code Listing 4.3: Reading the emerge man page |
$ man emerge
|
When you are emerging a series of packages, Portage can fetch the source files for the next package in the list even while it is compiling another package, thus shortening compile times. To make use of this capability, add "parallel-fetch" to your FEATURES.
When Portage is run as root, FEATURES="userfetch" will allow Portage to drop root privileges while fetching package sources. This is a small security improvement.
When you boot your system, you will notice lots of text floating by. If you pay close attention, you will notice this text is the same every time you reboot your system. The sequence of all these actions is called the boot sequence and is (more or less) statically defined.
First, your boot loader will load the kernel image you have defined in the boot loader configuration into memory after which it tells the CPU to run the kernel. When the kernel is loaded and run, it initializes all kernel-specific structures and tasks and starts the init process.
This process then makes sure that all filesystems (defined in /etc/fstab) are mounted and ready to be used. Then it executes several scripts located in /etc/init.d, which will start the services you need in order to have a successfully booted system.
Finally, when all scripts are executed, init activates the terminals (in most cases just the virtual consoles which are hidden beneath Alt-F1, Alt-F2, etc.) attaching a special process called agetty to it. This process will then make sure you are able to log on through these terminals by running login.
Now init doesn't just execute the scripts in /etc/init.d randomly. Even more, it doesn't run all scripts in /etc/init.d, only the scripts it is told to execute. It decides which scripts to execute by looking into /etc/runlevels.
First, init runs all scripts from /etc/init.d that have symbolic links inside /etc/runlevels/boot. Usually, it will start the scripts in alphabetical order, but some scripts have dependency information in them, telling the system that another script must be run before they can be started.
When all /etc/runlevels/boot referenced scripts are executed, init continues with running the scripts that have a symbolic link to them in /etc/runlevels/default. Again, it will use the alphabetical order to decide what script to run first, unless a script has dependency information in it, in which case the order is changed to provide a valid start-up sequence.
Of course init doesn't decide all that by itself. It needs a configuration file that specifies what actions need to be taken. This configuration file is /etc/inittab.
If you remember the boot sequence we have just described, you will remember that init's first action is to mount all filesystems. This is defined in the following line from /etc/inittab:
Code Listing 1.1: The system initialisation line in /etc/inittab |
si::sysinit:/sbin/rc sysinit |
This line tells init that it must run /sbin/rc sysinit to initialize the system. The /sbin/rc script takes care of the initialisation, so you might say that init doesn't do much -- it delegates the task of initialising the system to another process.
Second, init executed all scripts that had symbolic links in /etc/runlevels/boot. This is defined in the following line:
Code Listing 1.2: The system initialisation, continued |
rc::bootwait:/sbin/rc boot |
Again the rc script performs the necessary tasks. Note that the option given to rc (boot) is the same as the subdirectory of /etc/runlevels that is used.
Now init checks its configuration file to see what runlevel it should run. To decide this, it reads the following line from /etc/inittab:
Code Listing 1.3: The initdefault line |
id:3:initdefault: |
In this case (which the majority of Gentoo users will use), the runlevel id is 3. Using this information, init checks what it must run to start runlevel 3:
Code Listing 1.4: The runlevel definitions |
l0:0:wait:/sbin/rc shutdown l1:S1:wait:/sbin/rc single l2:2:wait:/sbin/rc nonetwork l3:3:wait:/sbin/rc default l4:4:wait:/sbin/rc default l5:5:wait:/sbin/rc default l6:6:wait:/sbin/rc reboot |
The line that defines level 3, again, uses the rc script to start the services (now with argument default). Again note that the argument of rc is the same as the subdirectory from /etc/runlevels.
When rc has finished, init decides what virtual consoles it should activate and what commands need to be run at each console:
Code Listing 1.5: The virtual consoles definition |
c1:12345:respawn:/sbin/agetty 38400 tty1 linux c2:12345:respawn:/sbin/agetty 38400 tty2 linux c3:12345:respawn:/sbin/agetty 38400 tty3 linux c4:12345:respawn:/sbin/agetty 38400 tty4 linux c5:12345:respawn:/sbin/agetty 38400 tty5 linux c6:12345:respawn:/sbin/agetty 38400 tty6 linux |
You have seen that init uses a numbering scheme to decide what runlevel it should activate. A runlevel is a state in which your system is running and contains a collection of scripts (runlevel scripts or initscripts) that must be executed when you enter or leave a runlevel.
In Gentoo, there are seven runlevels defined: three internal runlevels, and four user-defined runlevels. The internal runlevels are called sysinit, shutdown and reboot and do exactly what their names imply: initialize the system, powering off the system and rebooting the system.
The user-defined runlevels are those with an accompanying /etc/runlevels subdirectory: boot, default, nonetwork and single. The boot runlevel starts all system-necessary services which all other runlevels use. The remaining three runlevels differ in what services they start: default is used for day-to-day operations, nonetwork is used in case no network connectivity is required, and single is used when you need to fix the system.
The scripts that the rc process starts are called init scripts. Each script in /etc/init.d can be executed with the arguments start, stop, restart, pause, zap, status, ineed, iuse, needsme, usesme or broken.
To start, stop or restart a service (and all depending services), start, stop and restart should be used:
Code Listing 1.6: Starting Postfix |
# /etc/init.d/postfix start
|
Note: Only the services that need the given service are stopped or restarted. The other depending services (those that use the service but don't need it) are left untouched. |
If you want to stop a service, but not the services that depend on it, you can use the pause argument:
Code Listing 1.7: Stopping Postfix but keep the depending services running |
# /etc/init.d/postfix pause
|
If you want to see what status a service has (started, stopped, paused, ...) you can use the status argument:
Code Listing 1.8: Status information for postfix |
# /etc/init.d/postfix status
|
If the status information tells you that the service is running, but you know that it is not, then you can reset the status information to "stopped" with the zap argument:
Code Listing 1.9: Resetting status information for postfix |
# /etc/init.d/postfix zap
|
To also ask what dependencies the service has, you can use iuse or ineed. With ineed you can see the services that are really necessary for the correct functioning of the service. iuse on the other hand shows the services that can be used by the service, but are not necessary for the correct functioning.
Code Listing 1.10: Requesting a list of all necessary services on which Postfix depends |
# /etc/init.d/postfix ineed
|
Similarly, you can ask what services require the service (needsme) or can use it (usesme):
Code Listing 1.11: Requesting a list of all services that require Postfix |
# /etc/init.d/postfix needsme
|
Finally, you can ask what dependencies the service requires that are missing:
Code Listing 1.12: Requesting a list of missing dependencies for Postfix |
# /etc/init.d/postfix broken
|
Gentoo's init system uses a dependency-tree to decide what service needs to be started first. As this is a tedious task that we wouldn't want our users to have to do manually, we have created tools that ease the administration of the runlevels and init scripts.
With rc-update you can add and remove init scripts to a runlevel. The rc-update tool will then automatically ask the depscan.sh script to rebuild the dependency tree.
You have already added init scripts to the "default" runlevel during the installation of Gentoo. At that time you might not have had a clue what the "default" is for, but now you should. The rc-update script requires a second argument that defines the action: add, del or show.
To add or remove an init script, just give rc-update the add or del argument, followed by the init script and the runlevel. For instance:
Code Listing 2.1: Removing Postfix from the default runlevel |
# rc-update del postfix default
|
The rc-update -v show command will show all the available init scripts and list at which runlevels they will execute:
Code Listing 2.2: Receiving init script information |
# rc-update -v show
|
You can also run rc-update show (without -v) to just view enabled init scripts and their runlevels.
Why the Need for Extra Configuration?
Init scripts can be quite complex. It is therefore not really desirable to have the users edit the init script directly, as it would make it more error-prone. It is however important to be able to configure such a service. For instance, you might want to give more options to the service itself.
A second reason to have this configuration outside the init script is to be able to update the init scripts without the fear that your configuration changes will be undone.
Gentoo provides an easy way to configure such a service: every init script that can be configured has a file in /etc/conf.d. For instance, the apache2 initscript (called /etc/init.d/apache2) has a configuration file called /etc/conf.d/apache2, which can contain the options you want to give to the Apache 2 server when it is started:
Code Listing 3.1: Variable defined in /etc/conf.d/apache2 |
APACHE2_OPTS="-D PHP5" |
Such a configuration file contains variables and variables alone (just like /etc/make.conf), making it very easy to configure services. It also allows us to provide more information about the variables (as comments).
No, writing an init script is usually not necessary as Gentoo provides ready-to-use init scripts for all provided services. However, you might have installed a service without using Portage, in which case you will most likely have to create an init script.
Do not use the init script provided by the service if it isn't explicitly written for Gentoo: Gentoo's init scripts are not compatible with the init scripts used by other distributions!
The basic layout of an init script is shown below.
Code Listing 4.1: Basic layout of an init script |
#!/sbin/runscript
depend() {
(Dependency information)
}
start() {
(Commands necessary to start the service)
}
stop() {
(Commands necessary to stop the service)
}
restart() {
(Commands necessary to restart the service)
}
|
Any init script requires the start() function to be defined. All other sections are optional.
There are two dependencies you can define: use and need. As we have mentioned before, the need dependency is more strict than the use dependency. Following this dependency type you enter the service you depend on, or the virtual dependency.
A virtual dependency is a dependency that a service provides, but that is not provided solely by that service. Your init script can depend on a system logger, but there are many system loggers available (metalogd, syslog-ng, sysklogd, ...). As you cannot need every single one of them (no sensible system has all these system loggers installed and running) we made sure that all these services provide a virtual dependency.
Let us take a look at the dependency information for the postfix service.
Code Listing 4.2: Dependency information for Postfix |
depend() {
need net
use logger dns
provide mta
}
|
As you can see, the postfix service:
In some cases you might not require a service, but want your service to be started before (or after) another service if it is available on the system (note the conditional - this is no dependency anymore) and run in the same runlevel (note the conditional - only services in the same runlevel are involved). You can provide this information using the before or after settings.
As an example we view the settings of the Portmap service:
Code Listing 4.3: The depend() function in the Portmap service |
depend() {
need net
before inetd
before xinetd
}
|
You can also use the "*" glob to catch all services in the same runlevel, although this isn't advisable.
Code Listing 4.4: Running an init script as first script in the runlevel |
depend() {
before *
}
|
If your service must write to local disks, it should need localmount. If it places anything in /var/run such as a pidfile, then it should start after bootmisc:
Code Listing 4.5: Example depend() function |
depend() {
need localmount
after bootmisc
}
|
Next to the depend() functionality, you also need to define the start() function. This one contains all the commands necessary to initialize your service. It is advisable to use the ebegin and eend functions to inform the user about what is happening:
Code Listing 4.6: Example start() function |
start() {
ebegin "Starting my_service"
start-stop-daemon --start --exec /path/to/my_service \
--pidfile /path/to/my_pidfile
eend $?
}
|
Both --exec and --pidfile should be used in start and stop functions. If the service does not create a pidfile, then use --make-pidfile if possible, though you should test this to be sure. Otherwise, don't use pidfiles. You can also add --quiet to the start-stop-daemon options, but this is not recommended unless the service is extremely verbose. Using --quiet may hinder debugging if the service fails to start.
Note: Make sure that --exec actually calls a service and not just a shell script that launches services and exits -- that's what the init script is supposed to do. |
If you need more examples of the start() function, please read the source code of the available init scripts in your /etc/init.d directory.
Other functions you can define are: stop() and restart(). You are not obliged to define these functions! Our init system is intelligent enough to fill these functions by itself if you use start-stop-daemon.
Although you do not have to create a stop() function, here is an example:
Code Listing 4.7: Example stop() function |
stop() {
ebegin "Stopping my_service"
start-stop-daemon --stop --exec /path/to/my_service \
--pidfile /path/to/my_pidfile
eend $?
}
|
If your service runs some other script (for example, bash, python, or perl), and this script later changes names (for example, foo.py to foo), then you will need to add --name to start-stop-daemon. You must specify the name that your script will be changed to. In this example, a service starts foo.py, which changes names to foo:
Code Listing 4.8: A service that starts the foo script |
start() {
ebegin "Starting my_script"
start-stop-daemon --start --exec /path/to/my_script \
--pidfile /path/to/my_pidfile --name foo
eend $?
}
|
start-stop-daemon has an excellent man page available if you need more information:
Code Listing 4.9: Getting the man page for start-stop-daemon |
$ man start-stop-daemon
|
Gentoo's init script syntax is based on the Bourne Again Shell (bash) so you are free to use bash-compatible constructs inside your init script.
If you want your init script to support more options than the ones we have already encountered, you should add the option to the opts variable, and create a function with the same name as the option. For instance, to support an option called restartdelay:
Code Listing 4.10: Supporting the restartdelay option |
opts="${opts} restartdelay"
restartdelay() {
stop
sleep 3 # Wait 3 seconds before starting again
start
}
|
Service Configuration Variables
You don't have to do anything to support a configuration file in /etc/conf.d: if your init script is executed, the following files are automatically sourced (i.e. the variables are available to use):
Also, if your init script provides a virtual dependency (such as net), the file associated with that dependency (such as /etc/conf.d/net) will be sourced too.
4.e. Changing the Runlevel Behaviour
Many laptop users know the situation: at home you need to start net.eth0 while you don't want to start net.eth0 while you're on the road (as there is no network available). With Gentoo you can alter the runlevel behaviour to your own will.
For instance you can create a second "default" runlevel which you can boot that has other init scripts assigned to it. You can then select at boottime what default runlevel you want to use.
First of all, create the runlevel directory for your second "default" runlevel. As an example we create the offline runlevel:
Code Listing 5.1: Creating a runlevel directory |
# mkdir /etc/runlevels/offline
|
Add the necessary init scripts to the newly created runlevels. For instance, if you want to have an exact copy of your current default runlevel but without net.eth0:
Code Listing 5.2: Adding the necessary init scripts |
(Copy all services from default runlevel to offline runlevel) # cd /etc/runlevels/default # for service in *; do rc-update add $service offline; done (Remove unwanted service from offline runlevel) # rc-update del net.eth0 offline (Display active services for offline runlevel) # rc-update show offline (Partial sample Output) acpid | offline domainname | offline local | offline net.eth0 | |
Even though net.eth0 has been removed from the offline runlevel, udev will still attempt to start any devices it detects and launch the appropriate services. Therefore, you will need to add each network service you do not want started (as well as services for any other devices that may be started by udev) to /etc/conf.d/rc as shown.
Code Listing 5.3: Disabling device initiated services in /etc/conf.d/rc |
RC_COLDPLUG="yes"
(Next, specify the services you do not want automatically started)
RC_PLUG_SERVICES="!net.eth0"
|
Note: For more information on device initiated services, please see the comments inside /etc/conf.d/rc. |
Now edit your bootloader configuration and add a new entry for the offline runlevel. For instance, in /boot/grub/grub.conf:
Code Listing 5.4: Adding an entry for the offline runlevel |
title Gentoo Linux Offline Usage
root (hd0,0)
kernel (hd0,0)/kernel-2.4.25 root=/dev/hda3 softlevel=offline
|
Voilà, you're all set now. If you boot your system and select the newly added entry at boot, the offline runlevel will be used instead of the default one.
Using bootlevel is completely analogous to softlevel. The only difference here is that you define a second "boot" runlevel instead of a second "default" runlevel.
An environment variable is a named object that contains information used by one or more applications. Many users (and especially those new to Linux) find this a bit weird or unmanageable. However, this is a mistake: by using environment variables one can easily change a configuration setting for one or more applications.
The following table lists a number of variables used by a Linux system and describes their use. Example values are presented after the table.
| Variable | Description |
| PATH | This variable contains a colon-separated list of directories in which your system looks for executable files. If you enter a name of an executable (such as ls, rc-update or emerge) but this executable is not located in a listed directory, your system will not execute it (unless you enter the full path as command, such as /bin/ls). |
| ROOTPATH | This variable has the same function as PATH, but this one only lists the directories that should be checked when the root-user enters a command. |
| LDPATH | This variable contains a colon-separated list of directories in which the dynamical linker searches through to find a library. |
| MANPATH | This variable contains a colon-separated list of directories in which the man command searches for the man pages. |
| INFODIR | This variable contains a colon-separated list of directories in which the info command searches for the info pages. |
| PAGER | This variable contains the path to the program used to list the contents of files through (such as less or more). |
| EDITOR | This variable contains the path to the program used to change the contents of files with (such as nano or vi). |
| KDEDIRS | This variable contains a colon-separated list of directories which contain KDE-specific material. |
| CONFIG_PROTECT | This variable contains a space-delimited list of directories which should be protected by Portage during updates. |
| CONFIG_PROTECT_MASK | This variable contains a space-delimited list of directories which should not be protected by Portage during updates. |
Below you will find an example definition of all these variables:
Code Listing 1.1: Example definitions |
PATH="/bin:/usr/bin:/usr/local/bin:/opt/bin:/usr/games/bin"
ROOTPATH="/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/sbin:/usr/local/bin"
LDPATH="/lib:/usr/lib:/usr/local/lib:/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.3"
MANPATH="/usr/share/man:/usr/local/share/man"
INFODIR="/usr/share/info:/usr/local/share/info"
PAGER="/usr/bin/less"
EDITOR="/usr/bin/vim"
KDEDIRS="/usr"
CONFIG_PROTECT="/usr/X11R6/lib/X11/xkb /opt/tomcat/conf \
/usr/kde/3.1/share/config /usr/share/texmf/tex/generic/config/ \
/usr/share/texmf/tex/platex/config/ /usr/share/config"
CONFIG_PROTECT_MASK="/etc/gconf"
|
5.b. Defining Variables Globally
To centralise the definitions of these variables, Gentoo introduced the /etc/env.d directory. Inside this directory you will find a number of files, such as 00basic, 05gcc, etc. which contain the variables needed by the application mentioned in their name.
For instance, when you installed gcc, a file called 05gcc was created by the ebuild which contains the definitions of the following variables:
Code Listing 2.1: /etc/env.d/05gcc |
PATH="/usr/i686-pc-linux-gnu/gcc-bin/3.2" ROOTPATH="/usr/i686-pc-linux-gnu/gcc-bin/3.2" MANPATH="/usr/share/gcc-data/i686-pc-linux-gnu/3.2/man" INFOPATH="/usr/share/gcc-data/i686-pc-linux-gnu/3.2/info" CC="gcc" CXX="g++" LDPATH="/usr/lib/gcc-lib/i686-pc-linux-gnu/3.2.3" |
Other distributions tell you to change or add such environment variable definitions in /etc/profile or other locations. Gentoo on the other hand makes it easy for you (and for Portage) to maintain and manage the environment variables without having to pay attention to the numerous files that can contain environment variables.
For instance, when gcc is updated, the /etc/env.d/05gcc file is updated too without requesting any user-interaction.
This not only benefits Portage, but also you, as user. Occasionally you might be asked to set a certain environment variable system-wide. As an example we take the http_proxy variable. Instead of messing about with /etc/profile, you can now just create a file (/etc/env.d/99local) and enter your definition(s) in it:
Code Listing 2.2: /etc/env.d/99local |
http_proxy="proxy.server.com:8080" |
By using the same file for all your variables, you have a quick overview on the variables you have defined yourself.
Several files in /etc/env.d define the PATH variable. This is not a mistake: when you run env-update, it will append the several definitions before it updates the environment variables, thereby making it easy for packages (or users) to add their own environment variable settings without interfering with the already existing values.
The env-update script will append the values in the alphabetical order of the /etc/env.d files. The file names must begin with two decimal digits.
Code Listing 2.3: Update order used by env-update |
00basic 99kde-env 99local
+-------------+----------------+-------------+
PATH="/bin:/usr/bin:/usr/kde/3.2/bin:/usr/local/bin"
|
The concatenation of variables does not always happen, only with the following variables: KDEDIRS, PATH, LDPATH, MANPATH, INFODIR, INFOPATH, ROOTPATH, CONFIG_PROTECT, CONFIG_PROTECT_MASK, PRELINK_PATH and PRELINK_PATH_MASK. For all other variables the latest defined value (in alphabetical order of the files in /etc/env.d) is used.
When you run env-update, the script will create all environment variables and place them in /etc/profile.env (which is used by /etc/profile). It will also extract the information from the LDPATH variable and use that to create /etc/ld.so.conf. After this, it will run ldconfig to recreate the /etc/ld.so.cache file used by the dynamical linker.
If you want to notice the effect of env-update immediately after you run it, execute the following command to update your environment. Users who have installed Gentoo themselves will probably remember this from the installation instructions:
Code Listing 2.4: Updating the environment |
# env-update && source /etc/profile
|
Note: The above command only updates the variables in your current terminal, new consoles, and their children. Thus, if you are working in X11, you will need to either type source /etc/profile in every new terminal you open or restart X so that all new terminals source the new variables. If you use a login manager, become root and type /etc/init.d/xdm restart. If not, you will need to logout and log back in for X to spawn children with the new variable values. |
Important: You cannot use shell variables when defining other variables. This means things like FOO="$BAR" (where $BAR is another variable) are forbidden. |
5.c. Defining Variables Locally
You do not always want to define an environment variable globally. For instance, you might want to add /home/my_user/bin and the current working directory (the directory you are in) to the PATH variable but don't want all other users on your system to have that in their PATH too. If you want to define an environment variable locally, you should use ~/.bashrc or ~/.bash_profile:
Code Listing 3.1: Extending PATH for local usage in ~/.bashrc |
(A colon followed by no directory is treated as the current working directory)
PATH="${PATH}:/home/my_user/bin:"
|
When you relogin, your PATH variable will be updated.
Sometimes even stricter definitions are requested. You might want to be able to use binaries from a temporary directory you created without using the path to the binaries themselves or editing ~/.bashrc for the short time you need it.
In this case, you can just define the PATH variable in your current session by using the export command. As long as you don't log out, the PATH variable will be using the temporary settings.
Code Listing 3.2: Defining a session-specific environment variable |
# export PATH="${PATH}:/home/my_user/tmp/usr/bin"
|
Portage comes with a default configuration stored in /etc/make.globals. When you take a look at it, you'll notice that all Portage configuration is handled through variables. What variables Portage listens to and what they mean are described later.
Since many configuration directives differ between architectures, Portage also has default configuration files which are part of your profile. Your profile is pointed to by the /etc/make.profile symlink; Portage' configurations are set in the make.defaults files of your profile and all parent profiles. We'll explain more about profiles and the /etc/make.profile directory later on.
If you're planning on changing a configuration variable, don't alter /etc/make.globals or make.defaults. Instead use /etc/make.conf which has precedence over the previous files. You'll also find a /usr/share/portage/config/make.conf.example. As the name implies, this is merely an example file - Portage does not read in this file.
You can also define a Portage configuration variable as an environment variable, but we don't recommend this.
We've already encountered the /etc/make.profile directory. Well, this isn't exactly a directory but a symbolic link to a profile, by default one inside /usr/portage/profiles although you can create your own profiles elsewhere and point to them. The profile this symlink points to is the profile to which your system adheres.
A profile contains architecture-specific information for Portage, such as a list of packages that belong to the system corresponding with that profile, a list of packages that don't work (or are masked-out) for that profile, etc.
When you need to override Portage's behaviour regarding the installation of software, you will end up editing files within /etc/portage. You are highly recommended to use files within /etc/portage and highly discouraged to override the behaviour through environment variables!
Within /etc/portage you can create the following files:
These don't have to be files; they can also be directories that contain one file per package. More information about the /etc/portage directory and a full list of possible files you can create can be found in the Portage man page:
Code Listing 1.1: Reading the Portage man page |
$ man portage
|
Changing Portage File & Directory Locations
The previously mentioned configuration files cannot be stored elsewhere - Portage will always look for those configuration files at those exact locations. However, Portage uses many other locations for various purposes: build directory, source code storage, Portage tree location, ...
All these purposes have well-known default locations but can be altered to your own taste through /etc/make.conf. The rest of this chapter explains what special-purpose locations Portage uses and how to alter their placement on your filesystem.
This document isn't meant to be used as a reference though. If you need 100% coverage, please consult the Portage and make.conf man pages:
Code Listing 1.2: Reading the Portage and make.conf man pages |
$ man portage $ man make.conf |
The Portage tree default location is /usr/portage. This is defined by the PORTDIR variable. When you store the Portage tree elsewhere (by altering this variable), don't forget to change the /etc/make.profile symbolic link accordingly.
If you alter the PORTDIR variable, you might want to alter the following variables as well since they will not notice the PORTDIR change. This is due to how Portage handles variables: PKGDIR, DISTDIR, RPMDIR.
Even though Portage doesn't use prebuilt binaries by default, it has extensive support for them. When you ask Portage to work with prebuilt packages, it will look for them in /usr/portage/packages. This location is defined by the PKGDIR variable.
Application source code is stored in /usr/portage/distfiles by default. This location is defined by the DISTDIR variable.
Portage stores the state of your system (what packages are installed, what files belong to which package, ...) in /var/db/pkg. Do not alter these files manually! It might break Portage's knowledge of your system.
The Portage cache (with modification times, virtuals, dependency tree information, ...) is stored in /var/cache/edb. This location really is a cache: you can clean it if you are not running any portage-related application at that moment.
Portage's temporary files are stored in /var/tmp by default. This is defined by the PORTAGE_TMPDIR variable.
If you alter the PORTAGE_TMPDIR variable, you might want to alter the following variables as well since they will not notice the PORTAGE_TMPDIR change. This is due to how Portage handles variables: BUILD_PREFIX.
Portage creates specific build directories for each package it emerges inside /var/tmp/portage. This location is defined by the BUILD_PREFIX variable.
By default Portage installs all files on the current filesystem (/), but you can change this by setting the ROOT environment variable. This is useful when you want to create new build images.
Portage can create per-ebuild logfiles, but only when the PORT_LOGDIR variable is set to a location that is writable by Portage (the portage user). By default this variable is unset. If you don't set PORT_LOGDIR, then you won't receive any build logs with the current logging system, though you may receive some logs from the new elog. If you do have PORT_LOGDIR defined and you use elog, you will receive build logs and any logs saved by elog, as explained below.
Portage offers fine-grained control over logging through the use of elog:
Important: If you used enotice with Portage-2.0.*, you must completely remove enotice, as it is incompatible with elog. |
As noted previously, Portage is configurable through many variables which you should define in /etc/make.conf. Please refer to the make.conf man page for more and complete information:
Code Listing 1.1: Reading the make.conf man page |
$ man make.conf
|
Configure and Compiler Options
When Portage builds applications, it passes the contents of the following variables to the compiler and configure script:
The USE variable is also used during configure and compilations but has been explained in great detail in previous chapters.
When Portage has merged a newer version of a certain software title, it will remove the obsoleted files of the older version from your system. Portage gives the user a 5 second delay before unmerging the older version. These 5 seconds are defined by the CLEAN_DELAY variable.
You can tell emerge to use certain options every time it is run by setting EMERGE_DEFAULT_OPTS. Some useful options would be --ask, --verbose, --tree, and so on.
2.c. Configuration File Protection
Portage overwrites files provided by newer versions of a software title if the files aren't stored in a protected location. These protected locations are defined by the CONFIG_PROTECT variable and are generally configuration file locations. The directory listing is space-delimited.
A file that would be written in such a protected location is renamed and the user is warned about the presence of a newer version of the (presumable) configuration file.
You can find out about the current CONFIG_PROTECT setting from the emerge --info output:
Code Listing 3.1: Getting the CONFIG_PROTECT setting |
$ emerge --info | grep 'CONFIG_PROTECT='
|
More information about Portage's Configuration File Protection is available in the CONFIGURATION FILES section of the emerge manpage:
Code Listing 3.2: More information about Configuration File Protection |
$ man emerge
|
To 'unprotect' certain subdirectories of protected locations you can use the CONFIG_PROTECT_MASK variable.
When the requested information or data is not available on your system, Portage will retrieve it from the Internet. The server locations for the various information and data channels are defined by the following variables:
A third setting involves the location of the rsync server which you use when you update your Portage tree:
The GENTOO_MIRRORS and SYNC variables can be set automatically through the mirrorselect application. You need to emerge mirrorselect first before you can use it. For more information, see mirrorselect's online help:
Code Listing 4.1: More information about mirrorselect |
# mirrorselect --help
|
If your environment requires you to use a proxy server, you can use the http_proxy, ftp_proxy and RSYNC_PROXY variables to declare a proxy server.
When Portage needs to fetch source code, it uses wget by default. You can change this through the FETCHCOMMAND variable.
Portage is able to resume partially downloaded source code. It uses wget by default, but this can be altered through the RESUMECOMMAND variable.
Make sure that your FETCHCOMMAND and RESUMECOMMAND stores the source code in the correct location. Inside the variables you should use \${URI} and \${DISTDIR} to point to the source code location and distfiles location respectively.
You can also define protocol-specific handlers with FETCHCOMMAND_HTTP, FETCHCOMMAND_FTP, RESUMECOMMAND_HTTP, RESUMECOMMAND_FTP, and so on.
You cannot alter the rsync command used by Portage to update the Portage tree, but you can set some variables related to the rsync command:
For more information on these options and others, please read man rsync.
You can change your default branch with the ACCEPT_KEYWORDS variable. It defaults to your architecture's stable branch. More information on Gentoo's branches can be found in the next chapter.
You can activate certain Portage features through the FEATURES variable. The Portage Features have been discussed in previous chapters, such as Portage Features.
With the PORTAGE_NICENESS variable you can augment or reduce the nice value Portage runs with. The PORTAGE_NICENESS value is added to the current nice value.
For more information about nice values, see the nice man page:
Code Listing 6.1: More information about nice |
$ man nice
|
The NOCOLOR, which defaults to "false", defines if Portage should disable the use of coloured output.
The ACCEPT_KEYWORDS variable defines what software branch you use on your system. It defaults to the stable software branch for your architecture, for instance x86.
We recommend that you only use the stable branch. However, if you don't care about stability this much and you want to help out Gentoo by submitting bugreports to http://bugs.gentoo.org, read on.
If you want to use more recent software, you can consider using the testing branch instead. To have Portage use the testing branch, add a ~ in front of your architecture.
The testing branch is exactly what it says - Testing. If a package is in testing, it means that the developers feel that it is functional but has not been thoroughly tested. You could very well be the first to discover a bug in the package in which case you could file a bugreport to let the developers know about it.
Beware though, you might notice stability issues, imperfect package handling (for instance wrong/missing dependencies), too frequent updates (resulting in lots of building) or broken packages. If you do not know how Gentoo works and how to solve problems, we recommend that you stick with the stable and tested branch.
For example, to select the testing branch for the x86 architecture, edit /etc/make.conf and set:
Code Listing 1.1: Setting the ACCEPT_KEYWORDS variable |
ACCEPT_KEYWORDS="~x86" |
If you update your system now, you will find out that lots of packages will be updated. Mind you though: when you have updated your system to use the testing branch there is usually no easy way back to the stable, official branch (except for using backups of course).
3.b. Mixing Stable with Testing
You can ask Portage to allow the testing branch for particular packages but use the stable branch for the rest of the system. To achieve this, add the package category and name you want to use the testing branch of in /etc/portage/package.keywords. You can also create a directory (with the same name) and list the package in the files under that directory. For instance, to use the testing branch for gnumeric:
Code Listing 2.1: /etc/portage/package.keywords setting for gnumeric, full line |
app-office/gnumeric ~x86 |
If you want to use a specific software version from the testing branch but you don't want Portage to use the testing branch for subsequent versions, you can add in the version in the package.keywords location. In this case you must use the = operator. You can also enter a version range using the <=, <, > or >= operators.
In any case, if you add version information, you must use an operator. If you leave out version information, you cannot use an operator.
In the following example we ask Portage to accept gnumeric-1.2.13:
Code Listing 2.2: Enabling a particular gnumeric test version |
=app-office/gnumeric-1.2.13 ~x86 |
The Gentoo developers do not support the use of this location. Please exercise due caution when doing so. Support requests related to package.unmask and/or package.mask will not be answered. You have been warned.
When a package has been masked by the Gentoo developers and you still want to use it despite the reason mentioned in the package.mask file (situated in /usr/portage/profiles by default), add the exact same line in the /etc/portage/package.unmask file (or in a file in that directory if it is a directory).
For instance, if =net-mail/hotwayd-0.8 is masked, you can unmask it by adding the exact same line in the package.unmask location:
Code Listing 3.1: /etc/portage/package.unmask |
=net-mail/hotwayd-0.8 |
When you don't want Portage to take a certain package or a specific version of a package into account you can mask it yourself by adding an appropriate line to the /etc/portage/package.mask location (either in that file or in a file in this directory).
For instance, if you don't want Portage to install newer kernel sources than gentoo-sources-2.6.8.1, you add the following line at the package.mask location:
Code Listing 3.2: /etc/portage/package.mask example |
>sys-kernel/gentoo-sources-2.6.8.1 |
dispatch-conf is a tool that aids in merging the ._cfg0000_<name> files. ._cfg0000_<name> files are generated by Portage when it wants to overwrite a file in a directory protected by the CONFIG_PROTECT variable.
With dispatch-conf, you are able to merge updates to your configuration files while keeping track of all changes. dispatch-conf stores the differences between the configuration files as patches or by using the RCS revision system. This means that if you make a mistake when updating a config file, you can revert to the previous version of your config file at any time.
When using dispatch-conf, you can ask to keep the configuration file as-is, use the new configuration file, edit the current one or merge the changes interactively. dispatch-conf also has some nice additional features:
Make certain you edit /etc/dispatch-conf.conf first and create the directory referenced by the archive-dir variable.
Code Listing 1.1: Running dispatch-conf |
# dispatch-conf
|
When running dispatch-conf, you'll be taken through each changed config file, one at a time. Press u to update (replace) the current config file with the new one and continue to the next file. Press z to zap (delete) the new config file and continue to the next file. Once all config files have been taken care of, dispatch-conf will exit. You can also press q to exit any time.
For more information, check out the dispatch-conf man page. It tells you how to interactively merge current and new config files, edit new config files, examine differences between files, and more.
Code Listing 1.2: Reading the dispatch-conf man page |
$ man dispatch-conf
|
You can also use etc-update to merge config files. It's not as simple to use as dispatch-conf, nor as featureful, but it does provide an interactive merging setup and can also auto-merge trivial changes.
However, unlike dispatch-conf, etc-update does not preserve the old versions of your config files. Once you update the file, the old version is gone forever! So be very careful, as using etc-update is significantly less safe than using dispatch-conf.
Code Listing 2.1: Running etc-update |
# etc-update
|
After merging the straightforward changes, you will be prompted with a list of protected files that have an update waiting. At the bottom you are greeted by the possible options:
Code Listing 2.2: etc-update options |
Please select a file to edit by entering the corresponding number.
(-1 to exit) (-3 to auto merge all remaining files)
(-5 to auto-merge AND not use 'mv -i'):
|
If you enter -1, etc-update will exit and discontinue any further changes. If you enter -3 or -5, all listed configuration files will be overwritten with the newer versions. It is therefore very important to first select the configuration files that should not be automatically updated. This is simply a matter of entering the number listed to the left of that configuration file.
As an example, we select the configuration file /etc/pear.conf:
Code Listing 2.3: Updating a specific configuration file |
Beginning of differences between /etc/pear.conf and /etc/._cfg0000_pear.conf
[...]
End of differences between /etc/pear.conf and /etc/._cfg0000_pear.conf
1) Replace original with update
2) Delete update, keeping original as is
3) Interactively merge original with update
4) Show differences again
|
You can now see the differences between the two files. If you believe that the updated configuration file can be used without problems, enter 1. If you believe that the updated configuration file isn't necessary, or doesn't provide any new or useful information, enter 2. If you want to interactively update your current configuration file, enter 3.
There is no point in further elaborating the interactive merging here. For completeness sake, we will list the possible commands you can use while you are interactively merging the two files. You are greeted with two lines (the original one, and the proposed new one) and a prompt at which you can enter one of the following commands:
Code Listing 2.4: Commands available for the interactive merging |
ed: Edit then use both versions, each decorated with a header. eb: Edit then use both versions. el: Edit then use the left version. er: Edit then use the right version. e: Edit a new version. l: Use the left version. r: Use the right version. s: Silently include common lines. v: Verbosely include common lines. q: Quit. |
When you have finished updating the important configuration files, you can now automatically update all the other configuration files. etc-update will exit if it doesn't find any more updateable configuration files.
With quickpkg you can create archives of the packages that are already merged on your system. These archives can be used as prebuilt packages. Running quickpkg is straightforward: just add the names of the packages you want to archive.
For instance, to archive curl, arts and procps:
Code Listing 3.1: Example quickpkg usage |
# quickpkg curl arts procps
|
The prebuilt packages will be stored in $PKGDIR/All (/usr/portage/packages/All by default). Symbolic links pointing to these packages are placed in $PKGDIR/<category>.
5.a. Using a Portage Tree Subset
You can selectively update certain categories/packages and ignore the other categories/packages. We achieve this by having rsync exclude categories/packages during the emerge --sync step.
You need to define the name of the file that contains the exclude patterns in the --exclude-from variable in your /etc/make.conf.
Code Listing 1.1: Defining the exclude file in /etc/make.conf |
PORTAGE_RSYNC_EXTRA_OPTS="--exclude-from=/etc/portage/rsync_excludes" |
Code Listing 1.2: Excluding all games in /etc/portage/rsync_excludes |
games-*/* |
Note however that this may lead to dependency issues since new, allowed packages might depend on new but excluded packages.
5.b. Adding Unofficial Ebuilds
Defining a Portage Overlay Directory
You can ask Portage to use ebuilds that are not officially available through the Portage tree. Create a new directory (for instance /usr/local/portage) in which you store the 3rd-party ebuilds. Use the same directory structure as the official Portage tree!
Then define PORTDIR_OVERLAY in /etc/make.conf and have it point to the previously defined directory. When you use Portage now, it will take those ebuilds into account as well without removing/overwriting those ebuilds the next time you run emerge --sync.
For the powerusers who develop on several overlays, test packages before they hit the Portage tree or just want to use unofficial ebuilds from various sources, the app-portage/gentoolkit-dev package brings you gensync, a tool to help you keep the overlay repositories up to date.
With gensync you can update all the repositories at once, or select just a few of them. Each repository should have a .syncsource file in the /etc/gensync/ configuration directory which contains the repository location, name, ID, etc.
Suppose you have two additional repositories called java (for the in-development java ebuilds) and entapps (for the applications developed in-house for your enterprise). You can update those repositories with the following command:
Code Listing 2.1: Using gensync to update a few repositories |
# gensync java entapps
|
5.c. Non-Portage Maintained Software
Using Portage with Self-Maintained Software
In some cases you want to configure, install and maintain software yourself without having Portage automate the process for you, even though Portage can provide the software titles. Known cases are kernel sources and nvidia drivers. You can configure Portage so it knows that a certain package is manually installed on your system. This process is called injecting and supported by Portage through the /etc/portage/profile/package.provided file.
For instance, if you want to inform Portage about gentoo-sources-2.6.11.6 which you've installed manually, add the following line to /etc/portage/profile/package.provided:
Code Listing 3.1: Example line for package.provided |
sys-kernel/gentoo-sources-2.6.11.6 |
Note: This document assumes that you have correctly configured your kernel, its modules for your hardware and you know the interface name of your hardware. We also assume that you are configuring eth0, but it could also be eth1, wlan0, etc. |
Note: This document requires you to run baselayout-1.11.11 or better. |
To get started configuring your network card, you need to tell the Gentoo RC system about it. This is done by creating a symbolic link from net.lo to net.eth0 in /etc/init.d.
Code Listing 1.1: Symlinking net.eth0 to net.lo |
# cd /etc/init.d # ln -s net.lo net.eth0 |
Gentoo's RC system now knows about that interface. It also needs to know how to configure the new interface. All the network interfaces are configured in /etc/conf.d/net. Below is a sample configuration for DHCP and static addresses.
Code Listing 1.2: Examples for /etc/conf.d/net |
# For DHCP config_eth0=( "dhcp" ) # For static IP using CIDR notation config_eth0=( "192.168.0.7/24" ) routes_eth0=( "default via 192.168.0.1" ) # For static IP using netmask notation config_eth0=( "192.168.0.7 netmask 255.255.255.0" ) routes_eth0=( "default via 192.168.0.1" ) |
Note: If you do not specify a configuration for your interface then DHCP is assumed. |
Note: CIDR stands for Classless InterDomain Routing. Originally, IPv4 addresses were classified as A, B, or C. The early classification system did not envision the massive popularity of the Internet, and is in danger of running out of new unique addresses. CIDR is an addressing scheme that allows one IP address to designate many IP addresses. A CIDR IP address looks like a normal IP address except that it ends with a slash followed by a number; for example, 192.168.0.0/16. CIDR is described in RFC 1519. |
Now that we have configured our interface, we can start and stop it using the following commands:
Code Listing 1.3: Starting and stopping network scripts |
# /etc/init.d/net.eth0 start # /etc/init.d/net.eth0 stop |
Important: When troubleshooting networking, it is recommended to set RC_VERBOSE="yes" in /etc/conf.d/rc so that you get more information about what's happening. |
Now that you have successfully started and stopped your network interface, you may wish to get it to start when Gentoo boots. Here's how to do this. The last "rc" command instructs Gentoo to start any scripts in the current runlevel that have not yet been started.
Code Listing 1.4: Configuring a network interface to load at boot time |
# rc-update add net.eth0 default # rc |
The config_eth0 variable is the heart of an interface configuration. It's a high level instruction list for configuring the interface (eth0 in this case). Each command in the instruction list is performed sequentially. The interface is deemed OK if at least one command works.
Here's a list of built-in instructions.
| Command | Description |
| null | Do nothing |
| noop | If the interface is up and there is an address then abort configuration successfully |
| an IPv4 or IPv6 address | Add the address to the interface |
| dhcp, adsl or apipa (or a custom command from a 3rd party module) | Run the module which provides the command. For example dhcp will run a module that provides DHCP which can be one of either dhcpcd, dhclient or pump. |
If a command fails, you can specify a fallback command. The fallback has to match the config structure exactly.
You can chain these commands together. Here are some real world examples.
Code Listing 1.1: Configuration examples |
# Adding three IPv4 addresses config_eth0=( "192.168.0.2/24" "192.168.0.3/24" "192.168.0.4/24" ) # Adding an IPv4 address and two IPv6 addresses config_eth0=( "192.168.0.2/24" "4321:0:1:2:3:4:567:89ab" "4321:0:1:2:3:4:567:89ac" ) # Keep our kernel assigned address, unless the interface goes # down so assign another via DHCP. If DHCP fails then add a # static address determined by APIPA config_eth0=( "noop" "dhcp" ) fallback_eth0=( "null" "apipa" ) |
Note: When using the ifconfig module and adding more than one address, interface aliases are created for each extra address. So with the above two examples you will get interfaces eth0, eth0:1 and eth0:2. You cannot do anything special with these interfaces as the kernel and other programs will just treat eth0:1 and eth0:2 as eth0. |
Important: The fallback order is important! If we did not specify the null option then the apipa command would only be run if the noop command failed. |
Init scripts in /etc/init.d can depend on a specific network interface or just net. net can be defined in /etc/conf.d/rc to mean different things using the RC_NET_STRICT_CHECKING variable.
| Value | Description |
| none | The net service is always considered up |
| no | This basically means that at least one net.* service besides net.lo must be up. This can be used by notebook users that have a WIFI and a static NIC, and only wants one up at any given time to have the net service seen as up. |
| lo | This is the same as the no option, but net.lo is also counted. This should be useful to people that do not care about any specific interface being up at boot. |
| yes | For this ALL network interfaces MUST be up for the net service to be considered up. |
But what about net.br0 depending on net.eth0 and net.eth1? net.eth1 may be a wireless or PPP device that needs configuration before it can be added to the bridge. This cannot be done in /etc/init.d/net.br0 as that's a symbolic link to net.lo.
The answer is making your own depend() function in /etc/conf.d/net.
Code Listing 2.1: net.br0 dependency in /etc/conf.d/net |
# You can use any dependency (use, after, before) as found in current scripts
depend_br0() {
need net.eth0 net.eth1
}
|
For a more detailed discussion about dependency, consult the section Writing Init Scripts in the Gentoo Handbook.
2.c. Variable names and values
Variable names are dynamic. They normally follow the structure of variable_${interface|mac|essid|apmac}. For example, the variable dhcpcd_eth0 holds the value for dhcpcd options for eth0 and dhcpcd_essid holds the value for dhcpcd options when any interface connects to the ESSID "essid".
However, there is no hard and fast rule that states interface names must be ethx. In fact, many wireless interfaces have names like wlanx, rax as well as ethx. Also, some user defined interfaces such as bridges can be given any name, such as foo. To make life more interesting, wireless Access Points can have names with non alpha-numeric characters in them - this is important because you can configure networking parameters per ESSID.
The downside of all this is that Gentoo uses bash variables for networking - and bash cannot use anything outside of English alpha-numerics. To get around this limitation we change every character that is not an English alpha-numeric into a _ character.
Another downside of bash is the content of variables - some characters need to be escaped. This can be achived by placing the \ character in front of the character that needs to be escaped. The following list of characters needs to be escaped in this way: ", ' and \.
In this example we use wireless ESSID as they can contain the widest scope of characters. We shall use the ESSID My "\ NET:
Code Listing 3.1: variable name example |
(This does work, but the domain is invalid) dns_domain_My____NET="My \"\\ NET" (The above sets the dns domain to My "\ NET when a wireless card connects to an AP whose ESSID is My "\ NET) |
We now support modular networking scripts, which means we can easily add support for new interface types and configuration modules while keeping compatibility with existing ones.
Modules load by default if the package they need is installed. If you specify a module here that doesn't have its package installed then you get an error stating which package you need to install. Ideally, you only use the modules setting when you have two or more packages installed that supply the same service and you need to prefer one over the other.
Note: All settings discussed here are stored in /etc/conf.d/net unless otherwise specified. |
Code Listing 1.1: Module preference |
# Prefer iproute2 over ifconfig modules=( "iproute2" ) # You can also specify other modules for an interface # In this case we prefer pump over dhcpcd modules_eth0=( "pump" ) # You can also specify which modules not to use - for example you may be # using a supplicant or linux-wlan-ng to control wireless configuration but # you still want to configure network settings per ESSID associated with. modules=( "!iwconfig" ) |
We provide two interface handlers presently: ifconfig and iproute2. You need one of these to do any kind of network configuration.
ifconfig is the current Gentoo default and it's included in the system profile. iproute2 is a more powerful and flexible package, but it's not included by default.
Code Listing 2.1: To install iproute2 |
# emerge sys-apps/iproute2 # To prefer iproute2 over ifconfig if both are installed modules=( "iproute2" ) |
As both ifconfig and iproute2 do very similar things we allow their basic configuration to work with each other. For example both the below code snippet work regardless of which module you are using.
Code Listing 2.2: ifconfig and iproute2 examples |
config_eth0=( "192.168.0.2/24" )
config_eth0=( "192.168.0.2 netmask 255.255.255.0" )
# We can also specify broadcast
config_eth0=( "192.168.0.2/24 brd 192.168.0.255" )
config_eth0=( "192.168.0.2 netmask 255.255.255.0 broadcast 192.168.0.255" )
|
DHCP is a means of obtaining network information (IP address, DNS servers, Gateway, etc) from a DHCP server. This means that if there is a DHCP server running on the network, you just have to tell each client to use DHCP and it sets up the network all by itself. Of course, you will have to configure for other things like wireless, PPP or other things if required before you can use DHCP.
DHCP can be provided by dhclient, dhcpcd, or pump. Each DHCP module has its pros and cons - here's a quick run down.
| DHCP Module | Package | Pros | Cons |
| dhclient | net-misc/dhcp | Made by ISC, the same people who make the BIND DNS software. Very configurable | Configuration is overly complex, software is quite bloated, cannot get NTP servers from DHCP, does not send hostname by default |
| dhcpcd | net-misc/dhcpcd | Long time Gentoo default, no reliance on outside tools, actively developed by Gentoo | Can be slow at times, does not yet daemonize when lease is infinite |
| pump | net-misc/pump | Lightweight, no reliance on outside tools | No longer maintained upstream, unreliable, especially over modems, cannot get NIS servers from DHCP |
If you have more than one DHCP client installed, you need to specify which one to use - otherwise we default to dhcpcd if available.
To send specific options to the DHCP module, use module_eth0="..." (change module to the DHCP module you're using - i.e. dhcpcd_eth0).
We try and make DHCP relatively agnostic - as such we support the following commands using the dhcp_eth0 variable. The default is not to set any of them:
Code Listing 3.1: Sample DHCP configuration in /etc/conf.d/net |
# Only needed if you have more than one DHCP module installed modules=( "dhcpcd" ) config_eth0=( "dhcp" ) dhcpcd_eth0="-t 10" # Timeout after 10 seconds dhcp_eth0="release nodns nontp nonis" # Only get an address |
Note: dhcpcd and pump send the current hostname to the DHCP server by default so you don't need to specify this anymore. |
First we need to install the ADSL software.
Code Listing 4.1: Install the ppp package |
# emerge net-dialup/ppp
|
Note: If you need PPPoA, then make sure to use >=baselayout-1.12.x. |
Second, create the PPP net script and the net script for the ethernet interface to be used by PPP:
Code Listing 4.2: Creating the PPP and ethernet scripts |
# ln -s /etc/init.d/net.lo /etc/init.d/net.ppp0 # ln -s /etc/init.d/net.lo /etc/init.d/net.eth0 |
Be sure to set RC_NET_STRICT_CHECKING="yes" in /etc/conf.d/rc.
Now we need to configure /etc/conf.d/net.
Code Listing 4.3: A basic PPPoE setup |
config_eth0=( null ) (Specify your ethernet interface) config_ppp0=( "ppp" ) link_ppp0="eth0" (Specify your ethernet interface) plugins_ppp0=( "pppoe" ) username_ppp0='user' password_ppp0='password' pppd_ppp0=( "noauth" "defaultroute" "usepeerdns" "holdoff 3" "child-timeout 60" "lcp-echo-interval 15" "lcp-echo-failure 3" noaccomp noccp nobsdcomp nodeflate nopcomp novj novjccomp ) depend_ppp0() { need net.eth0 } |
You can also set your password in /etc/ppp/pap-secrets.
Code Listing 4.4: Sample /etc/ppp/pap-secrets |
# The * is important
"username" * "password"
|
If you use PPPoE with a USB modem you'll need to emerge br2684ctl. Please read /usr/portage/net-dialup/speedtouch-usb/files/README for information on how to properly configure it.
Important: Please carefully read the section on ADSL and PPP in /etc/conf.d/net.example. It contains many more detailed explanations of all the settings your particular PPP setup will likely need. |
3.e. APIPA (Automatic Private IP Addressing)
APIPA tries to find a free address in the range 169.254.0.0-169.254.255.255 by arping a random address in that range on the interface. If no reply is found then we assign that address to the interface.
This is only useful for LANs where there is no DHCP server and you don't connect directly to the internet and all other computers use APIPA.
For APIPA support, emerge net-misc/iputils or net-analyzer/arping.
Code Listing 5.1: APIPA configuration in /etc/conf.d/net |
# Try DHCP first - if that fails then fallback to APIPA config_eth0=( "dhcp" ) fallback_eth0=( "apipa" ) # Just use APIPA config_eth0=( "apipa" ) |
For link bonding/trunking emerge net-misc/ifenslave.
Bonding is used to increase network bandwidth. If you have two network cards going to the same network, you can bond them together so your applications see just one interface but they really use both network cards.
Code Listing 6.1: bonding configuration in /etc/conf.d/net |
# To bond interfaces together slaves_bond0="eth0 eth1 eth2" # You may not want to assign an IP to the bonded interface config_bond0=( "null" ) # Depend on eth0, eth1 and eth2 as they may require extra configuration depend_bond0() { need net.eth0 net.eth1 net.eth2 } |
3.g. Bridging (802.1d support)
For bridging support emerge net-misc/bridge-utils.
Bridging is used to join networks together. For example, you may have a server that connects to the internet via an ADSL modem and a wireless access card to enable other computers to connect to the internet via the ADSL modem. You could create a bridge to join the two interfaces together.
Code Listing 7.1: Bridge configuration in /etc/conf.d/net |
# Configure the bridge - "man brctl" for more details brctl_br0=( "setfd 0" "sethello 0" "stp off" ) # To add ports to bridge br0 bridge_br0="eth0 eth1" # You need to configure the ports to null values so dhcp does not get started config_eth0=( "null" ) config_eth1=( "null" ) # Finally give the bridge an address - you could use DHCP as well config_br0=( "192.168.0.1/24" ) # Depend on eth0 and eth1 as they may require extra configuration depend_br0() { need net.eth0 net.eth1 } |
Important: For using some bridge setups, you may need to consult the variable name documentation. |
You don't need to emerge anything for changing the MAC address of your interface if you have sys-apps/baselayout-1.11.14 or newer and want to change to a specific MAC address. However, if you need to change to a random MAC address or have a baselayout older than the version mentioned above, you have to emerge net-analyzer/macchanger to be able to make use of this feature.
Code Listing 8.1: MAC Address change example |
# To set the MAC address of the interface mac_eth0="00:11:22:33:44:55" # To randomize the last 3 bytes only mac_eth0="random-ending" # To randomize between the same physical type of connection (e.g. fibre, # copper, wireless) , all vendors mac_eth0="random-samekind" # To randomize between any physical type of connection (e.g. fibre, copper, # wireless) , all vendors mac_eth0="random-anykind" # Full randomization - WARNING: some MAC addresses generated by this may # NOT act as expected mac_eth0="random-full" |
You don't need to emerge anything for tunnelling as the interface handler can do it for you.
Code Listing 9.1: Tunnelling configuration in /etc/conf.d/net |
# For GRE tunnels iptunnel_vpn0="mode gre remote 207.170.82.1 key 0xffffffff ttl 255" # For IPIP tunnels iptunnel_vpn0="mode ipip remote 207.170.82.2 ttl 255" # To configure the interface config_vpn0=( "192.168.0.2 peer 192.168.1.1" ) |
For VLAN support, emerge net-misc/vconfig.
Virtual LAN is a group of network devices that behave as if they were connected to a single network segment - even though they may not be. VLAN members can only see members of the same VLAN even though they may share the same physical network.
Code Listing 10.1: VLAN configuration in /etc/conf.d/net |
# Specify the VLAN numbers for the interface like so # Please ensure your VLAN IDs are NOT zero-padded vlans_eth0="1 2" # You can also configure the VLAN # see for vconfig man page for more details vconfig_eth0=( "set_name_type VLAN_PLUS_VID_NO_PAD" ) vconfig_vlan1=( "set_flag 1" "set_egress_map 2 6" ) # Configure the interface as usual config_vlan1=( "172.16.3.1 netmask 255.255.254.0" ) config_vlan2=( "172.16.2.1 netmask 255.255.254.0" ) |
Important: For using some VLAN setups, you may need to consult the variable name documentation. |
Currently we support wireless setup either by wireless-tools or wpa_supplicant. The important thing to remember is that you configure for wireless networks on a global basis and not an interface basis.
wpa_supplicant is the best choice, but it does not support all drivers. For a list of supported drivers, read the wpa_supplicant site. Also, wpa_supplicant can currently only connect to SSIDs that you have configured.
wireless-tools supports nearly all cards and drivers, but it cannot connect to WPA only Access Points.
Warning: The linux-wlan-ng driver is not supported by baselayout at this time. This is because linux-wlan-ng have its own setup and configuration which is completely different to everyone else's. The linux-wlan-ng developers are rumoured to be changing their setup over to wireless-tools, so when this happens you may use linux-wlan-ng with baselayout. |
WPA Supplicant is a package that allows you to connect to WPA enabled access points. Its setup is fairly fluid as it is still in beta - however it works fine for the most part.
Code Listing 2.1: Install wpa_supplicant |
# emerge net-wireless/wpa_supplicant
|
Important: You have to have CONFIG_PACKET enabled in your kernel for wpa_supplicant to work. |
Now we have to configure /etc/conf.d/net to so that we prefer wpa_supplicant over wireless-tools (if both are installed, wireless-tools is the default).
Code Listing 2.2: configure /etc/conf.d/net for wpa_supplicant |
# Prefer wpa_supplicant over wireless-tools modules=( "wpa_supplicant" ) # It's important that we tell wpa_supplicant which driver we should # be using as it's not very good at guessing yet wpa_supplicant_eth0="-Dmadwifi" |
Note: If you're using the host-ap driver you will need to put the card in Managed mode before it can be used with wpa_supplicant correctly. You can use iwconfig_eth0="mode managed" to achieve this in /etc/conf.d/net. |
That was simple, wasn't it? However, we still have to configure wpa_supplicant itself which is a bit more tricky depending on how secure the Access Points are that you are trying to connect to. The below example is taken and simplified from /usr/share/doc/wpa_supplicant-<version>/wpa_supplicant.conf.gz which ships with wpa_supplicant.
Code Listing 2.3: an example /etc/wpa_supplicant/wpa_supplicant.conf |
# The below line not be changed otherwise we refuse to work ctrl_interface=/var/run/wpa_supplicant # Ensure that only root can read the WPA configuration ctrl_interface_group=0 # Let wpa_supplicant take care of scanning and AP selection ap_scan=1 # Simple case: WPA-PSK, PSK as an ASCII passphrase, allow all valid ciphers network={ ssid="simple" psk="very secret passphrase" # The higher the priority the sooner we are matched priority=5 } # Same as previous, but request SSID-specific scanning (for APs that reject # broadcast SSID) network={ ssid="second ssid" scan_ssid=1 psk="very secret passphrase" priority=2 } # Only WPA-PSK is used. Any valid cipher combination is accepted network={ ssid="example" proto=WPA key_mgmt=WPA-PSK pairwise=CCMP TKIP group=CCMP TKIP WEP104 WEP40 psk=06b4be19da289f475aa46a33cb793029d4ab3db7a23ee92382eb0106c72ac7bb priority=2 } # Plaintext connection (no WPA, no IEEE 802.1X) network={ ssid="plaintext-test" key_mgmt=NONE } # Shared WEP key connection (no WPA, no IEEE 802.1X) network={ ssid="static-wep-test" key_mgmt=NONE # Keys in quotes are ASCII keys wep_key0="abcde" # Keys specified without quotes are hex keys wep_key1=0102030405 wep_key2="1234567890123" wep_tx_keyidx=0 priority=5 } # Shared WEP key connection (no WPA, no IEEE 802.1X) using Shared Key # IEEE 802.11 authentication network={ ssid="static-wep-test2" key_mgmt=NONE wep_key0="abcde" wep_key1=0102030405 wep_key2="1234567890123" wep_tx_keyidx=0 priority=5 auth_alg=SHARED } # IBSS/ad-hoc network with WPA-None/TKIP network={ ssid="test adhoc" mode=1 proto=WPA key_mgmt=WPA-NONE pairwise=NONE group=TKIP psk="secret passphrase" } |
Initial setup and Managed Mode
Wireless Tools provide a generic way to configure basic wireless interfaces up to the WEP security level. While WEP is a weak security method it's also the most prevalent.
Wireless Tools configuration is controlled by a few main variables. The sample configuration file below should describe all you need. One thing to bear in mind is that no configuration means "connect to the strongest unencrypted Access Point" - we will always try and connect you to something.
Code Listing 3.1: Install wireless-tools |
# emerge net-wireless/wireless-tools
|
Note: Although you can store your wireless settings in /etc/conf.d/wireless this guide recommends you store them in /etc/conf.d/net. |
Important: You will need to consult the variable name documentation. |
Code Listing 3.2: sample iwconfig setup in /etc/conf.d/net |
# Prefer iwconfig over wpa_supplicant modules=( "iwconfig" ) # Configure WEP keys for Access Points called ESSID1 and ESSID2 # You may configure up to 4 WEP keys, but only 1 can be active at # any time so we supply a default index of [1] to set key [1] and then # again afterwards to change the active key to [1] # We do this incase you define other ESSID's to use WEP keys other than 1 # # Prefixing the key with s: means it's an ASCII key, otherwise a HEX key # # enc open specified open security (most secure) # enc restricted specified restricted security (least secure) key_ESSID1="[1] s:yourkeyhere key [1] enc open" key_ESSID2="[1] aaaa-bbbb-cccc-dd key [1] enc restricted" # The below only work when we scan for available Access Points # Sometimes more than one Access Point is visible so we need to # define a preferred order to connect in preferred_aps=( "ESSID1" "ESSID2" ) |
Fine tune Access Point Selection
You can add some extra options to fine-tune your Access Point selection, but these are not normally required.
You can decide whether we only connect to preferred Access Points or not. By default if everything configured has failed and we can connect to an unencrypted Access Point then we will. This can be controlled by the associate_order variable. Here's a table of values and how they control this.
| Value | Description |
| any | Default behaviour |
| preferredonly | We will only connect to visible APs in the preferred list |
| forcepreferred | We will forceably connect to APs in the preferred order if they are not found in a scan |
| forcepreferredonly | Do not scan for APs - instead just try to connect to each one in order |
| forceany | Same as forcepreferred + connect to any other available AP |
Finally we have some blacklist_aps and unique_ap selection. blacklist_aps works in a similar way to preferred_aps. unique_ap is a yes or no value that says if a second wireless interface can connect to the same Access Point as the first interface.
Code Listing 3.3: blacklist_aps and unique_ap example |
# Sometimes you never want to connect to certain access points blacklist_aps=( "ESSID3" "ESSID4" ) # If you have more than one wireless card, you can say if you want # to allow each card to associate with the same Access Point or not # Values are "yes" and "no" # Default is "yes" unique_ap="yes" |
If you want to set yourself up as an Ad-Hoc node if you fail to connect to any Access Point in managed mode, you can do that too.
Code Listing 3.4: fallback to ad-hoc mode |
adhoc_essid_eth0="This Adhoc Node" |
What about connecting to Ad-Hoc networks or running in Master mode to become an Access Point? Here's a configuration just for that! You may need to specify WEP keys as shown above.
Code Listing 3.5: sample ad-hoc/master configuration |
# Set the mode - can be managed (default), ad-hoc or master # Not all drivers support all modes mode_eth0="ad-hoc" # Set the ESSID of the interface # In managed mode, this forces the interface to try and connect to the # specified ESSID and nothing else essid_eth0="This Adhoc Node" # We use channel 3 if you don't specify one channel_eth0="9" |
Important: The below is taken verbatim from the BSD wavelan documentation found at the NetBSD documentation. There are 14 channels possible; We are told that channels 1-11 are legal for North America, channels 1-13 for most of Europe, channels 10-13 for France, and only channel 14 for Japan. If in doubt, please refer to the documentation that came with your card or access point. Make sure that the channel you select is the same channel your access point (or the other card in an ad-hoc network) is on. The default for cards sold in North America and most of Europe is 3; the default for cards sold in France is 11, and the default for cards sold in Japan is 14. |
Troubleshooting Wireless Tools
There are some more variables you can use to help get your wireless up and running due to driver or environment problems. Here's a table of other things you can try.
| Variable | Default Value | Description |
| iwconfig_eth0 | See the iwconfig man page for details on what to send iwconfig | |
| iwpriv_eth0 | See the iwpriv man page for details on what to send iwpriv | |
| sleep_scan_eth0 | 0 | The number of seconds to sleep before attempting to scan. This is needed when the driver/firmware needs more time to active before it can be used. |
| sleep_associate_eth0 | 5 | The number of seconds to wait for the interface to associate with the Access Point before moving onto the next one |
| associate_test_eth0 | MAC | Some drivers do not reset the MAC address associated with an invalid one when they lose or attempt association. Some drivers do not reset the quality level when they lose or attempt association. Valid settings are MAC, quality and all. |
| scan_mode_eth0 | Some drivers have to scan in ad-hoc mode, so if scanning fails try setting ad-hoc here | |
| iwpriv_scan_pre_eth0 | Sends some iwpriv commands to the interface before scanning. See the iwpriv man page for more details. | |
| iwpriv_scan_post_eth0 | Sends some iwpriv commands to the interface after scanning. See the iwpriv man page for more details. |
4.d. Defining network configuration per ESSID
Sometimes, you need a static IP when you connect to ESSID1 and you need DHCP when you connect to ESSID2. In fact, most module variables can be defined per ESSID. Here's how we do this.
Note: These work if you're using WPA Supplicant or Wireless Tools. |
Important: You will need to consult the variable name documentation. |
Code Listing 4.1: override network settings per ESSID |
config_ESSID1=( "192.168.0.3/24 brd 192.168.0.255" ) routes_ESSID1=( "default via 192.168.0.1" ) config_ESSID2=( "dhcp" ) fallback_ESSID2=( "192.168.3.4/24" ) fallback_route_ESSID2=( "default via 192.168.3.1" ) # We can define nameservers and other things too # NOTE: DHCP will override these unless it's told not too dns_servers_ESSID1=( "192.168.0.1" "192.168.0.2" ) dns_domain_ESSID1="some.domain" dns_search_domains_ESSID1="search.this.domain search.that.domain" # You override by the MAC address of the Access Point # This handy if you goto different locations that have the same ESSID config_001122334455=( "dhcp" ) dhcpcd_001122334455="-t 10" dns_servers_001122334455=( "192.168.0.1" "192.168.0.2" ) |
Four functions can be defined which will be called surrounding the start/stop operations. The functions are called with the interface name first so that one function can control multiple adapters.
The return values for the preup() and predown() functions should be 0 (success) to indicate that configuration or deconfiguration of the interface can continue. If preup() returns a non-zero value, then interface configuration will be aborted. If predown() returns a non-zero value, then the interface will not be allowed to continue deconfiguration.
The return values for the postup() and postdown() functions are ignored since there's nothing to do if they indicate failure.
${IFACE} is set to the interface being brought up/down. ${IFVAR} is ${IFACE} converted to variable name bash allows.
Code Listing 1.1: pre/post up/down function examples |
preup() {
# Test for link on the interface prior to bringing it up. This
# only works on some network adapters and requires the ethtool
# package to be installed.
if ethtool ${IFACE} | grep -q 'Link detected: no'; then
ewarn "No link on ${IFACE}, aborting configuration"
return 1
fi
# Remember to return 0 on success
return 0
}
predown() {
# The default in the script is to test for NFS root and disallow
# downing interfaces in that case. Note that if you specify a
# predown() function you will override that logic. Here it is, in
# case you still want it...
if is_net_fs /; then
eerror "root filesystem is network mounted -- can't stop ${IFACE}"
return 1
fi
# Remember to return 0 on success
return 0
}
postup() {
# This function could be used, for example, to register with a
# dynamic DNS service. Another possibility would be to
# send/receive mail once the interface is brought up.
return 0
}
postdown() {
# This function is mostly here for completeness... I haven't
# thought of anything nifty to do with it yet ;-)
return 0
}
|
5.b. Wireless Tools function hooks
Note: This will not work with WPA Supplicant - but the ${ESSID} and ${ESSIDVAR} variables are available in the postup() function. |
Two functions can be defined which will be called surrounding the associate function. The functions are called with the interface name first so that one function can control multiple adapters.
The return values for the preassociate() function should be 0 (success) to indicate that configuration or deconfiguration of the interface can continue. If preassociate() returns a non-zero value, then interface configuration will be aborted.
The return value for the postassociate() function is ignored since there's nothing to do if it indicates failure.
${ESSID} is set to the exact ESSID of the AP you're connecting to. ${ESSIDVAR} is ${ESSID} converted to variable name bash allows.
Code Listing 2.1: pre/post association functions |
preassociate() {
# The below adds two configuration variables leap_user_ESSID
# and leap_pass_ESSID. When they are both configured for the ESSID
# being connected to then we run the CISCO LEAP script
local user pass
eval user=\"\$\{leap_user_${ESSIDVAR}\}\"
eval pass=\"\$\{leap_pass_${ESSIDVAR}\}\"
if [[ -n ${user} && -n ${pass} ]]; then
if [[ ! -x /opt/cisco/bin/leapscript ]]; then
eend "For LEAP support, please emerge net-misc/cisco-aironet-client-utils"
return 1
fi
einfo "Waiting for LEAP Authentication on \"${ESSID//\\\\//}\""
if /opt/cisco/bin/leapscript ${user} ${pass} | grep -q 'Login incorrect'; then
ewarn "Login Failed for ${user}"
return 1
fi
fi
return 0
}
postassociate() {
# This function is mostly here for completeness... I haven't
# thought of anything nifty to do with it yet ;-)
return 0
}
|
Note: ${ESSID} and ${ESSIDVAR} are unavailable in predown() and postdown() functions. |
If you and your computer are always on the move, you may not always have an ethernet cable or plugged in or an access point available. Also, you may want networking to automatically work when an ethernet cable is plugged in or an access point is found.
Here you can find some tools that help you manage this.
Note: This document only talks about ifplugd, but there are alternatives such as netplug. netplug is a lightweight alternative to ifplugd, but it relies on your kernel network drivers working correctly, and many drivers do not. |
ifplugd is a daemon that starts and stops interfaces when an ethernet cable is inserted or removed. It can also manage detecting association to Access Points or when new ones come in range.
Code Listing 2.1: Installing ifplugd |
# emerge sys-apps/ifplugd
|
Configuration for ifplugd is fairly straightforward too. The configuration file is held in /etc/conf.d/net. Run man ifplugd for details on the available variables. Also, see /etc/conf.d/net.example for more examples.
Code Listing 2.2: Sample ifplug configuration |
(Replace eth0 with the interface to be monitored) ifplugd_eth0="..." (To monitor a wireless interface) ifplugd_eth0="--api-mode=wlan" |
In addition to managing multiple network connections, you may want to add a tool that makes it easy to work with multiple DNS servers and configurations. This is very handy when you receive your IP address via DHCP. Simply emerge openresolv.
Code Listing 2.3: Installing openresolv |
# emerge openresolv
|
See man resolvconf to learn more about its features.
The contents of this document are licensed under the Creative Commons - Attribution / Share Alike license.