Messages in this thread Patch in this message | | | From | "George Spelvin" <> | Date | Sun, 17 Feb 2008 20:21:03 -0500 | Subject | IO_DELAY documentation |
| |
"make oldconfig" doesn't know how to display the per-item help, so something like this is needed. It was a squeeze to make all the help text, plus option prompt, fit into 24 lines. If you think that's not a concern, some of the wording could be improved.
(Added text placed into the public domain; copyright abandoned.)
diff --git a/arch/x86/Kconfig.debug b/arch/x86/Kconfig.debug index 864affc..8e2ceb6 100644 --- a/arch/x86/Kconfig.debug +++ b/arch/x86/Kconfig.debug @@ -157,6 +157,24 @@ config IO_DELAY_TYPE_NONE choice prompt "IO delay type" default IO_DELAY_0XED + help + The original IBM PC required short pauses between consecutive + accesses to its peripheral chips. This can be provided four ways: + + * A dummy write to port 0x80 has been used for 15 years, but + crashes the AMD64-based HP dv9000z laptop. The safest option. + * Port 0xED is also believed safe, and testing is requested. + * Udelay is awkward to implement in the kernel due to processor + speed variations, but should be extremely safe. + * All but the very oldest PCs use emulations of the original + peripheral chips built into a "Super I/O" chip, and usually + do no need the delay at all. The most efficient option. + + On the majority of machines, all four options should work. + If you have a computer on which any of the four options + (especially _0XED or _NONE) does not work, please report it, + with a detailed description of the motherboard (and "lspci -nn" + and "dmidecode" output), to linux-kernel@vger.kernel.org. config IO_DELAY_0X80 bool "port 0x80 based port-IO delay [recommended]"
| |