diff options
Diffstat (limited to 'fi/hardware/supported-peripherals.xml')
| -rw-r--r-- | fi/hardware/supported-peripherals.xml | 187 |
1 files changed, 187 insertions, 0 deletions
diff --git a/fi/hardware/supported-peripherals.xml b/fi/hardware/supported-peripherals.xml new file mode 100644 index 000000000..f3ab5db99 --- /dev/null +++ b/fi/hardware/supported-peripherals.xml @@ -0,0 +1,187 @@ +<!-- retain these comments for translator revision tracking --> +<!-- original version: 29467 untranslated --> + + <sect1 id="supported-peripherals"> + <title>Peripherals and Other Hardware</title> +<para arch="not-s390"> + +Linux supports a large variety of hardware devices such as mice, +printers, scanners, PCMCIA and USB devices. However, most of these +devices are not required while installing the system. + +</para><para arch="i386"> + +USB hardware generally works fine, only some +USB keyboards may require additional configuration +(see <xref linkend="usb-keyboard-config"/>). + +</para><para arch="i386"> + +Again, see the +<ulink url="&url-hardware-howto;">Linux Hardware Compatibility HOWTO</ulink> +to determine whether your specific hardware is supported by Linux. + +</para><para arch="s390"> + +Package installations from XPRAM and tape are not supported by this +system. All packages that you want to install need to be available on a +DASD or over the network using NFS, HTTP or FTP. + +</para><para arch="mips"> + +The Broadcom BCM91250A evaluation board offers standard 3.3v 32 bit and 64 +bit PCI slots as well as USB connectors. + +</para><para arch="mipsel"> + +The Broadcom BCM91250A evaluation board offers standard 3.3v 32 bit and 64 +bit PCI slots as well as USB connectors. The Cobalt RaQ has no support for +additional devices but the Qube has one PCI slot. + +</para> +</sect1> + + <sect1 arch="not-s390"><title>Purchasing Hardware Specifically for GNU/Linux</title> + +<para> + +There are several vendors, who ship systems with Debian or other +distributions of GNU/Linux +<ulink url="&url-pre-installed;">pre-installed</ulink>. You might pay more +for the privilege, but it does buy a level of peace of mind, since you can +be sure that the hardware is well-supported by GNU/Linux. + +</para><para arch="m68k"> + +Unfortunately, it's quite rare to find any vendor shipping +new &arch-title; machines at all. + +</para><para arch="i386"> + +If you do have to buy a machine with Windows bundled, carefully read +the software license that comes with Windows; you may be able to +reject the license and obtain a rebate from your vendor. Searching +the Internet for <quote>windows refund</quote> may get you some useful +information to help with that. + +</para><para> + +Whether or not you are purchasing a system with Linux bundled, or even +a used system, it is still important to check that your hardware is +supported by the Linux kernel. Check if your hardware is listed in +the references found above. Let your salesperson (if any) know that +you're shopping for a Linux system. Support Linux-friendly hardware +vendors. + +</para> + + <sect2><title>Avoid Proprietary or Closed Hardware</title> +<para> + +Some hardware manufacturers simply won't tell us how to write drivers +for their hardware. Others won't allow us access to the documentation +without a non-disclosure agreement that would prevent us from +releasing the Linux source code. + +</para><para arch="m68k"> + +Another example is the proprietary hardware in the older +Macintosh line. In fact, no specifications or documentation have ever +been released for any Macintosh hardware, most notably the ADB +controller (used by the mouse and keyboard), the floppy controller, +and all acceleration and CLUT manipulation of the video hardware +(though we do now support CLUT manipulation on nearly all internal +video chips). In a nutshell, this explains why the Macintosh Linux +port lags behind other Linux ports. + +</para><para> + +Since we haven't been granted access to the documentation on these +devices, they simply won't work under Linux. You can help by asking +the manufacturers of such hardware to release the documentation. If +enough people ask, they will realize that the free software community +is an important market. + +</para> +</sect2> + + + <sect2 arch="i386"><title>Windows-specific Hardware</title> +<para> + +A disturbing trend is the proliferation of Windows-specific modems and +printers. In some cases these are specially designed to be operated by +the Microsoft Windows operating system and bear the legend <quote>WinModem</quote> +or <quote>Made especially for Windows-based computers</quote>. This +is generally done by removing the embedded processors of the hardware +and shifting the work they do over to a Windows driver that is run by +your computer's main CPU. This strategy makes the hardware less +expensive, but the savings are often <emphasis>not</emphasis> passed on to the +user and this hardware may even be more expensive than equivalent +devices that retain their embedded intelligence. + +</para><para> + +You should avoid Windows-specific hardware for two reasons. The first +is that the manufacturers do not generally make the resources +available to write a Linux driver. Generally, the hardware and +software interface to the device is proprietary, and documentation is +not available without a non-disclosure agreement, if it is available +at all. This precludes its being used for free software, since free +software writers disclose the source code of their programs. The +second reason is that when devices like these have had their embedded +processors removed, the operating system must perform the work of the +embedded processors, often at <emphasis>real-time</emphasis> priority, +and thus the CPU is not available to run your programs while it is +driving these devices. Since the typical Windows user does not +multi-process as intensively as a Linux user, the manufacturers hope +that the Windows user simply won't notice the burden this hardware +places on their CPU. However, any multi-processing operating system, +even Windows 2000 or XP, suffers from degraded performance when +peripheral manufacturers skimp on the embedded processing power of +their hardware. + +</para><para> + +You can help this situation by encouraging these manufacturers to +release the documentation and other resources necessary for us to +program their hardware, but the best strategy is simply to avoid this +sort of hardware until it is listed as working in the +<ulink url="&url-hardware-howto;">Linux Hardware Compatibility HOWTO</ulink>. + +</para> +</sect2> + + + <sect2 id="Parity-RAM"> + <title>Fake or <quote>Virtual</quote> Parity RAM</title> +<para> + +If you ask for Parity RAM in a computer store, you'll probably get +<emphasis>virtual parity</emphasis> memory modules instead of +<emphasis>true parity</emphasis> ones. Virtual parity SIMMs can often +(but not always) be distinguished because they only have one more chip +than an equivalent non-parity SIMM, and that one extra chip is smaller +than all the others. Virtual-parity SIMMs work exactly like non-parity +memory. They can't tell you when you have a single-bit RAM error the +way true-parity SIMMs do in a motherboard that implements +parity. Don't ever pay more for a virtual-parity SIMM than a +non-parity one. Do expect to pay a little more for true-parity SIMMs, +because you are actually buying one extra bit of memory for every 8 +bits. + +</para><para> + +If you want complete information on &arch-title; RAM issues, and what +is the best RAM to buy, see the +<ulink url="&url-pc-hw-faq;">PC Hardware FAQ</ulink>. + +</para><para arch="alpha"> + +Most, if not all, Alpha systems require true-parity RAM. + +</para> + + </sect2> + + </sect1> |