1 RFC for the chip specification architecture 2 3\begin{abstract} 4At the end of this document is the original message that motivated the 5change. 6\end{abstract} 7 8\section{Scope} 9This document defines how LinuxBIOS programmers can specify chips that 10are used, specified, and initalized. The current scope is for superio 11chips, but the architecture should allow for specification of other chips such 12as southbridges. Multiple chips of same or different type are supported. 13 14\section{Goals} 15The goals of the new chip architecture are these: 16\begin{itemize} 17\item seperate implementation details from specification in the Config file 18(translation: no more C code in Config files) 19\item make the specification easier for people to use and understand 20\item remove private details of a given chip to the chip file as much 21as possible 22\item allow unique register-set-specifiers for each chip 23\end{itemize} 24 25\section{Specification in the Config file} 26The specification looks like this: 27\begin{verbatim} 28chip <name> [path=<path>] ["<configuration>"] 29\end{verbatim} 30The name is in the standard LinuxBIOS form of type/vendor/name, e.g. 31"southbridge/intel/piix4e" or "superio/ite/it8671f". The class of the 32chip is derived from the first pathname component of the name, and the chip 33configuration is derived from the following components. 34 35The path defines the access mechanism to the chip. 36It is optional. If present, it overrides the default path to the chip. 37 38The configuration defines chip-specific configuration details, and is also 39optional. Note that an empty configuration will leave the chip with 40no enabled resources. This may be desirable in some cases. 41 42\section{Results of specifying a chip} 43 44When one or more chips are specified, the data about the chips 45is saved until the entire file is parsed. At this point, the config tool 46creates a file in the build directory called chip.c This file contains 47a common struct containing information about 48each individual chip and an array of pointers to these structures. 49 50For each chip, there are two structures. The structures contain control 51information for the chip, and register initialization information. The 52names of the structures are derived by ``flattening'' the chip name, 53as in the current linuxbios. For example, superio/ite/xyz uses 54two structs, one called superio_ite_xyz_control and one called 55superio_ite_xyz_init. The control struct is initialized from the 56chip name and path information, and has a pointer to the 57config struct. The config struct is initialized from the quote string 58 59\begin{verbatim} 60From rminnich@lanl.gov Fri May 16 10:34:13 2003 61Date: Tue, 13 May 2003 08:11:46 -0600 (MDT) 62From: ron minnich <rminnich@lanl.gov> 63To: linuxbios@clustermatic.org 64Subject: RFC:new superio proposal 65 66Abstract: 67 The superio architecture for linuxbios has worked for the last 2 68years but is being stretched to the limit by the changes in superio chips. 69The architecture depended on superio resources being relatively constant 70between chips, but this assumption no longer holds. In this document we 71propose several alternatives and solicit comments. 72 73Overview: 74The superio architecture in linuxbios was developed over time, and 75modified as circumstances required. In the beginning it was relatively 76simple and assumed only one superio per mainboard. The latest version 77allows an arbitrary number of superios per mainboard, and allows complete 78specification of the superio base I/O address along with the specification 79of reasonable default valures for both the base I/O address and the 80superio parameters such as serial enable, baud rate, and so on. 81 82Specification of superio control parameters is done by a configuration 83line such as: 84 85nsuperio sis/950 com1={1} floppy=1 lpt=1 86 87This fragment sets the superio type to sis/950; sets com1, floppy, and lpt 88to enabled; and leaves the defaults to com1 (baud rate, etc.) to the 89default values. 90 91While it is not obvious, these configuration parameters are fragments of a 92C initializer. The initializers are used to build a statically initialized 93structure of this type: 94 95struct superio { 96 struct superio_control *super; // the ops for the device. 97 unsigned int port; // if non-zero, overrides the default port 98 // com ports. This is not done as an array (yet). 99 // We think it's easier to set up from python if it is not an 100 // array. 101 struct com_ports com1, com2, com3, com4; 102 // DMA, if it exists. 103 struct lpt_ports lpt1, lpt2; 104 /* flags for each device type. Unsigned int. */ 105 // low order bit ALWAYS means enable. Next bit means to enable 106 // LPT is in transition, so we leave this here for the moment. 107 // The winbond chips really stretched the way this works. 108 // so many functions! 109 unsigned int ide, floppy, lpt; 110 unsigned int keyboard, cir, game; 111 unsigned int gpio1, gpio2, gpio3; 112 unsigned int acpi,hwmonitor; 113}; 114 115These structures are, in turn, created and statically initialized by a 116config-tool-generated structure that defines all the superios. This file 117is called nsuperio.c, is created for each mainboard you build, only 118appears in the build directory, and looks like this: 119 120=== 121extern struct superio_control superio_winbond_w83627hf_control; 122 123struct superio superio_winbond_w83627hf= { 124 &superio_winbond_w83627hf_control, 125 .com1={1}, .com2={1}, .floppy=1, .lpt=1, .keyboard=1, .hwmonitor=1}; 126 127struct superio *all_superio[] = {&superio_winbond_w83627hf, 128}; 129 130unsigned long nsuperio = 1; 131=== 132 133This example shows a board with one superio (nsuperio). The superio 134consists of a winbond w83627hf, with com1, com2, floppy, lpt, keyboard, 135and hwmonitor enabled. Note that this structure also allows for 136over-riding the default superio base, although that capability is rarely 137used. 138 139The control structure is used to define how to access the superio for 140purposes of control. It looks like this: 141=== 142struct superio_control { 143 void (*pre_pci_init)(struct superio *s); 144 void (*init)(struct superio *s); 145 void (*finishup)(struct superio *s); 146 unsigned int defaultport; /* the defaultport. Can be overridden 147 * by commands in config 148 */ 149 // This is the print name for debugging 150 char *name; 151}; 152=== 153 154There are three methods for stages of hardwaremain. First is pre_pci_init 155(for chips like the acer southbridge that require you to enable some 156resources BEFORE pci scan); init, called during the 'middle' phase of 157hardwaremain; and finishup, called before the payload is loaded. 158 159This approach was inspired by and borrows heavily on the Plan 9 kernel 160configuration tools. 161 162The problem: 163 164When the first version of the superio structure came out it was much 165smaller. It has grown and in the limit this structure is the union of all 166possibly superio chips. Obviously, in the long term, this is not 167practical: we can not anticipate all possible superio chips for all time. 168 169The common PC BIOS solution to this type of problem is to continue with 170binary structures but add version numbers to them, so that all code that 171uses a given structure has to check the version number. Personally, I find 172this grotesque and would rather not work this way. 173 174Using textual strings for configuration is something I find far more 175attractive. Plan 9 has shown that this approach has no real limits and 176suffices for configuration tasks. The Linux kernel does more limited use 177of strings for configuration, but still depends on them. Strings are 178easier to read and work with than binary structures, and more important, a 179lot easier to deal with when things start going wrong. 180 181The proposed solution: 182 183What follows are three possible ideas for specifying superio resources and 184their settings. 185 186A common part of the new idea is to eliminate the common superio 187structure, due to the many variations in chips, and make it invisible 188outside a given superio source file -- the superio structure is now 189private to a given superio. Thus, sis/950/superio.c would contain its own 190superio structure definitions, and also might contain more than once 191instance of these structures (consider a board with 2 sis 950 chips). 192 193The control structure would change as follows: 194struct superio_control { 195 int (*create)(struct superio *s); 196 void (*pre_pci_init)(struct superio *s); 197 void (*init)(struct superio *s); 198 void (*finishup)(struct superio *s); 199 unsigned int defaultport; /* the defaultport. Can be overridden 200 * by commands in config 201 */ 202 // This is the print name for debugging 203 char *name; 204}; 205 206I.e. we add a new function for creating the superio. 207 208Communication of superio settings from linuxbios to the superio would be 209via textual strings. The superio structure becomes this: 210 211struct superio { 212 struct superio_control *super; // the ops for the device. 213 unsigned int port; // if non-zero, overrides the default port 214 struct configuration *config; 215}; 216 217 218So now the question becomes, what is the configuration structure? 219There are several choices. The simplest, from my point of view, are 220keyword-value pairs: 221struct configuration { 222 const char *keyword; 223 const char *value; 224}; 225 226These get filled in by the config tool as before. The linuxbios libary can 227then provide a generic parsing function for the superios to use. 228 229The remaining question is how should the superio command look in 230freebios2? 231 232superio sis/950 "com1=115200,8n1 lpt=1 com2=9600" 233 234or 235 236superio sis/950 "com1baud=115200 lpt=1 com1chars=8n1" 237 238or 239 240superio sis/950 ((com1 115200 8n1) (lpt 1)) 241 242So, my questions: 243 2441. Does this new scheme look workable. If not, what needs to change? 2452. What should the 'struct configuration' be? does keyword/value work? 2463. what should the superio command look like? 247 248Comments welcome. 249 250I'd like to adopt this "RFC" approach for freebios2 as much as we can. 251There was a lot of give-and-take in the early days of linuxbios about 252structure and it proved useful. There's a lot that will start happening in 253freebios2 now, and we need to try to make sure it will work for everyone. 254 255Those of you who are doing mainboards, please look at freebios2 and see 256how it looks for you. There's a lot of good work that has been done (not 257by me so far, thanks Eric and Stefan), and more that needs to be done. 258Consider trying out romcc as an "assembly code killer". See how it fits 259together and if you can work with it or need changes. Bring comments back 260to this list. 261 262thanks 263 264ron 265 266\end{verbatim} 267