Masahiro Yamada | cc85b7b | 2015-07-26 02:46:26 +0900 | [diff] [blame] | 1 | menu "Generic Driver Options" |
| 2 | |
Masahiro Yamada | 57ad8ee | 2014-10-23 22:26:09 +0900 | [diff] [blame] | 3 | config DM |
| 4 | bool "Enable Driver Model" |
Masahiro Yamada | 57ad8ee | 2014-10-23 22:26:09 +0900 | [diff] [blame] | 5 | help |
Simon Glass | d8b771d | 2015-02-05 21:41:35 -0700 | [diff] [blame] | 6 | This config option enables Driver Model. This brings in the core |
| 7 | support, including scanning of platform data on start-up. If |
| 8 | CONFIG_OF_CONTROL is enabled, the device tree will be scanned also |
| 9 | when available. |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 10 | |
| 11 | config SPL_DM |
| 12 | bool "Enable Driver Model for SPL" |
| 13 | depends on DM && SPL |
| 14 | help |
| 15 | Enable driver model in SPL. You will need to provide a |
| 16 | suitable malloc() implementation. If you are not using the |
| 17 | full malloc() enabled by CONFIG_SYS_SPL_MALLOC_START, |
| 18 | consider using CONFIG_SYS_MALLOC_SIMPLE. In that case you |
Andy Yan | 1fa20e4d | 2017-07-24 17:43:34 +0800 | [diff] [blame] | 19 | must provide CONFIG_SPL_SYS_MALLOC_F_LEN to set the size. |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 20 | In most cases driver model will only allocate a few uclasses |
| 21 | and devices in SPL, so 1KB should be enable. See |
Andy Yan | 1fa20e4d | 2017-07-24 17:43:34 +0800 | [diff] [blame] | 22 | CONFIG_SPL_SYS_MALLOC_F_LEN for more details on how to enable it. |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 23 | |
Simon Glass | c833c00 | 2017-04-02 09:50:31 -0600 | [diff] [blame] | 24 | config TPL_DM |
| 25 | bool "Enable Driver Model for TPL" |
| 26 | depends on DM && TPL |
| 27 | help |
| 28 | Enable driver model in TPL. You will need to provide a |
| 29 | suitable malloc() implementation. If you are not using the |
| 30 | full malloc() enabled by CONFIG_SYS_SPL_MALLOC_START, |
| 31 | consider using CONFIG_SYS_MALLOC_SIMPLE. In that case you |
Andy Yan | 1fa20e4d | 2017-07-24 17:43:34 +0800 | [diff] [blame] | 32 | must provide CONFIG_SPL_SYS_MALLOC_F_LEN to set the size. |
Simon Glass | c833c00 | 2017-04-02 09:50:31 -0600 | [diff] [blame] | 33 | In most cases driver model will only allocate a few uclasses |
| 34 | and devices in SPL, so 1KB should be enough. See |
Andy Yan | 1fa20e4d | 2017-07-24 17:43:34 +0800 | [diff] [blame] | 35 | CONFIG_SPL_SYS_MALLOC_F_LEN for more details on how to enable it. |
Simon Glass | c833c00 | 2017-04-02 09:50:31 -0600 | [diff] [blame] | 36 | Disable this for very small implementations. |
| 37 | |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 38 | config DM_WARN |
| 39 | bool "Enable warnings in driver model" |
Masahiro Yamada | cfc2f7d | 2015-02-24 22:26:21 +0900 | [diff] [blame] | 40 | depends on DM |
| 41 | default y |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 42 | help |
Simon Glass | 85bf4f2 | 2020-10-03 11:31:26 -0600 | [diff] [blame] | 43 | Enable this to see warnings related to driver model. |
| 44 | |
| 45 | Warnings may help with debugging, such as when expected devices do |
| 46 | not bind correctly. If the option is disabled, dm_warn() is compiled |
| 47 | out - it will do nothing when called. |
| 48 | |
| 49 | config SPL_DM_WARN |
| 50 | bool "Enable warnings in driver model wuth SPL" |
| 51 | depends on SPL_DM |
| 52 | help |
| 53 | Enable this to see warnings related to driver model in SPL |
| 54 | |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 55 | The dm_warn() function can use up quite a bit of space for its |
| 56 | strings. By default this is disabled for SPL builds to save space. |
Simon Glass | 85bf4f2 | 2020-10-03 11:31:26 -0600 | [diff] [blame] | 57 | |
| 58 | Warnings may help with debugging, such as when expected devices do |
| 59 | not bind correctly. If the option is disabled, dm_warn() is compiled |
| 60 | out - it will do nothing when called. |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 61 | |
Masahiro Yamada | f70f39f | 2017-09-29 12:31:20 +0900 | [diff] [blame] | 62 | config DM_DEBUG |
| 63 | bool "Enable debug messages in driver model core" |
| 64 | depends on DM |
| 65 | help |
| 66 | Say Y here if you want to compile in debug messages in DM core. |
| 67 | |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 68 | config DM_DEVICE_REMOVE |
| 69 | bool "Support device removal" |
Masahiro Yamada | cfc2f7d | 2015-02-24 22:26:21 +0900 | [diff] [blame] | 70 | depends on DM |
| 71 | default y |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 72 | help |
| 73 | We can save some code space by dropping support for removing a |
Jean-Jacques Hiblot | 00b2988 | 2018-12-07 14:50:53 +0100 | [diff] [blame] | 74 | device. |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 75 | |
Hans de Goede | 246ba26 | 2015-07-01 20:52:59 +0200 | [diff] [blame] | 76 | Note that this may have undesirable results in the USB subsystem as |
| 77 | it causes unplugged devices to linger around in the dm-tree, and it |
| 78 | causes USB host controllers to not be stopped when booting the OS. |
| 79 | |
Jean-Jacques Hiblot | 00b2988 | 2018-12-07 14:50:53 +0100 | [diff] [blame] | 80 | config SPL_DM_DEVICE_REMOVE |
| 81 | bool "Support device removal in SPL" |
| 82 | depends on SPL_DM |
| 83 | default n |
| 84 | help |
| 85 | We can save some code space by dropping support for removing a |
| 86 | device. This is not normally required in SPL, so by default this |
| 87 | option is disabled for SPL. |
| 88 | |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 89 | config DM_STDIO |
| 90 | bool "Support stdio registration" |
Masahiro Yamada | cfc2f7d | 2015-02-24 22:26:21 +0900 | [diff] [blame] | 91 | depends on DM |
| 92 | default y |
Simon Glass | 93e414d | 2015-02-05 21:41:36 -0700 | [diff] [blame] | 93 | help |
| 94 | Normally serial drivers register with stdio so that they can be used |
| 95 | as normal output devices. In SPL we don't normally use stdio, so |
| 96 | we can omit this feature. |
Simon Glass | 1bd3deb | 2015-02-27 22:06:30 -0700 | [diff] [blame] | 97 | |
| 98 | config DM_SEQ_ALIAS |
| 99 | bool "Support numbered aliases in device tree" |
| 100 | depends on DM |
| 101 | default y |
| 102 | help |
| 103 | Most boards will have a '/aliases' node containing the path to |
| 104 | numbered devices (e.g. serial0 = &serial0). This feature can be |
Nathan Rossi | 7fa6ff4 | 2016-01-08 03:00:45 +1000 | [diff] [blame] | 105 | disabled if it is not required. |
| 106 | |
| 107 | config SPL_DM_SEQ_ALIAS |
| 108 | bool "Support numbered aliases in device tree in SPL" |
Adam Ford | ac4d80e | 2019-08-24 13:50:34 -0500 | [diff] [blame] | 109 | depends on SPL_DM |
Nathan Rossi | 7fa6ff4 | 2016-01-08 03:00:45 +1000 | [diff] [blame] | 110 | default n |
| 111 | help |
| 112 | Most boards will have a '/aliases' node containing the path to |
| 113 | numbered devices (e.g. serial0 = &serial0). This feature can be |
Simon Glass | 1bd3deb | 2015-02-27 22:06:30 -0700 | [diff] [blame] | 114 | disabled if it is not required, to save code space in SPL. |
Simon Glass | 8e5efa0 | 2015-07-17 09:22:07 -0600 | [diff] [blame] | 115 | |
Simon Glass | 39f1d28 | 2020-12-16 17:25:06 -0700 | [diff] [blame] | 116 | config SPL_DM_INLINE_OFNODE |
| 117 | bool "Inline some ofnode functions which are seldom used in SPL" |
| 118 | depends on SPL_DM |
| 119 | default y |
| 120 | help |
| 121 | This applies to several ofnode functions (see ofnode.h) which are |
| 122 | seldom used. Inlining them can help reduce code size. |
| 123 | |
| 124 | config TPL_DM_INLINE_OFNODE |
| 125 | bool "Inline some ofnode functions which are seldom used in TPL" |
| 126 | depends on TPL_DM |
| 127 | default y |
| 128 | help |
| 129 | This applies to several ofnode functions (see ofnode.h) which are |
| 130 | seldom used. Inlining them can help reduce code size. |
| 131 | |
Nicolas Saenz Julienne | a2c7ead | 2021-01-12 13:55:24 +0100 | [diff] [blame] | 132 | config DM_DMA |
| 133 | bool "Support per-device DMA constraints" |
| 134 | depends on DM |
| 135 | default n |
| 136 | help |
| 137 | Enable this to extract per-device DMA constraints, only supported on |
| 138 | device-tree systems for now. This is needed in order translate |
| 139 | addresses on systems where different buses have different views of |
| 140 | the physical address space. |
| 141 | |
Simon Glass | 8e5efa0 | 2015-07-17 09:22:07 -0600 | [diff] [blame] | 142 | config REGMAP |
| 143 | bool "Support register maps" |
| 144 | depends on DM |
| 145 | help |
| 146 | Hardware peripherals tend to have one or more sets of registers |
| 147 | which can be accessed to control the hardware. A register map |
| 148 | models this with a simple read/write interface. It can in principle |
| 149 | support any bus type (I2C, SPI) but so far this only supports |
| 150 | direct memory access. |
| 151 | |
huang lin | dd8515e | 2015-11-17 14:20:13 +0800 | [diff] [blame] | 152 | config SPL_REGMAP |
| 153 | bool "Support register maps in SPL" |
Philipp Tomsich | cdda349 | 2017-06-29 01:37:10 +0200 | [diff] [blame] | 154 | depends on SPL_DM |
huang lin | dd8515e | 2015-11-17 14:20:13 +0800 | [diff] [blame] | 155 | help |
| 156 | Hardware peripherals tend to have one or more sets of registers |
| 157 | which can be accessed to control the hardware. A register map |
| 158 | models this with a simple read/write interface. It can in principle |
| 159 | support any bus type (I2C, SPI) but so far this only supports |
| 160 | direct memory access. |
| 161 | |
Philipp Tomsich | 03cdf69 | 2017-06-29 01:38:49 +0200 | [diff] [blame] | 162 | config TPL_REGMAP |
| 163 | bool "Support register maps in TPL" |
| 164 | depends on TPL_DM |
| 165 | help |
| 166 | Hardware peripherals tend to have one or more sets of registers |
| 167 | which can be accessed to control the hardware. A register map |
| 168 | models this with a simple read/write interface. It can in principle |
| 169 | support any bus type (I2C, SPI) but so far this only supports |
| 170 | direct memory access. |
| 171 | |
Simon Glass | 8e5efa0 | 2015-07-17 09:22:07 -0600 | [diff] [blame] | 172 | config SYSCON |
| 173 | bool "Support system controllers" |
| 174 | depends on REGMAP |
| 175 | help |
| 176 | Many SoCs have a number of system controllers which are dealt with |
| 177 | as a group by a single driver. Some common functionality is provided |
| 178 | by this uclass, including accessing registers via regmap and |
| 179 | assigning a unique number to each. |
Masahiro Yamada | 8b15b16 | 2015-07-25 21:52:35 +0900 | [diff] [blame] | 180 | |
huang lin | dd8515e | 2015-11-17 14:20:13 +0800 | [diff] [blame] | 181 | config SPL_SYSCON |
| 182 | bool "Support system controllers in SPL" |
Philipp Tomsich | cdda349 | 2017-06-29 01:37:10 +0200 | [diff] [blame] | 183 | depends on SPL_REGMAP |
huang lin | dd8515e | 2015-11-17 14:20:13 +0800 | [diff] [blame] | 184 | help |
| 185 | Many SoCs have a number of system controllers which are dealt with |
| 186 | as a group by a single driver. Some common functionality is provided |
| 187 | by this uclass, including accessing registers via regmap and |
| 188 | assigning a unique number to each. |
| 189 | |
Philipp Tomsich | 03cdf69 | 2017-06-29 01:38:49 +0200 | [diff] [blame] | 190 | config TPL_SYSCON |
| 191 | bool "Support system controllers in TPL" |
| 192 | depends on TPL_REGMAP |
| 193 | help |
| 194 | Many SoCs have a number of system controllers which are dealt with |
| 195 | as a group by a single driver. Some common functionality is provided |
| 196 | by this uclass, including accessing registers via regmap and |
| 197 | assigning a unique number to each. |
| 198 | |
Masahiro Yamada | 029bfca | 2015-07-25 21:52:37 +0900 | [diff] [blame] | 199 | config DEVRES |
| 200 | bool "Managed device resources" |
| 201 | depends on DM |
| 202 | help |
| 203 | This option enables the Managed device resources core support. |
| 204 | Device resources managed by the devres framework are automatically |
| 205 | released whether initialization fails half-way or the device gets |
| 206 | detached. |
| 207 | |
| 208 | If this option is disabled, devres functions fall back to |
| 209 | non-managed variants. For example, devres_alloc() to kzalloc(), |
| 210 | devm_kmalloc() to kmalloc(), etc. |
| 211 | |
Masahiro Yamada | 8b15b16 | 2015-07-25 21:52:35 +0900 | [diff] [blame] | 212 | config DEBUG_DEVRES |
Masahiro Yamada | f929c0b | 2015-07-25 21:52:38 +0900 | [diff] [blame] | 213 | bool "Managed device resources debugging functions" |
Masahiro Yamada | 029bfca | 2015-07-25 21:52:37 +0900 | [diff] [blame] | 214 | depends on DEVRES |
Masahiro Yamada | 8b15b16 | 2015-07-25 21:52:35 +0900 | [diff] [blame] | 215 | help |
| 216 | If this option is enabled, devres debug messages are printed. |
Masahiro Yamada | f929c0b | 2015-07-25 21:52:38 +0900 | [diff] [blame] | 217 | Also, a function is available to dump a list of device resources. |
Masahiro Yamada | 8b15b16 | 2015-07-25 21:52:35 +0900 | [diff] [blame] | 218 | Select this if you are having a problem with devres or want to |
| 219 | debug resource management for a managed device. |
| 220 | |
| 221 | If you are unsure about this, Say N here. |
Masahiro Yamada | cc85b7b | 2015-07-26 02:46:26 +0900 | [diff] [blame] | 222 | |
Marek Vasut | 298e608 | 2015-08-03 01:15:48 +0200 | [diff] [blame] | 223 | config SIMPLE_BUS |
| 224 | bool "Support simple-bus driver" |
| 225 | depends on DM && OF_CONTROL |
| 226 | default y |
| 227 | help |
| 228 | Supports the 'simple-bus' driver, which is used on some systems. |
| 229 | |
| 230 | config SPL_SIMPLE_BUS |
| 231 | bool "Support simple-bus driver in SPL" |
| 232 | depends on SPL_DM && SPL_OF_CONTROL |
Michal Simek | b32467f | 2015-12-01 08:37:16 +0100 | [diff] [blame] | 233 | default y |
Marek Vasut | 298e608 | 2015-08-03 01:15:48 +0200 | [diff] [blame] | 234 | help |
| 235 | Supports the 'simple-bus' driver, which is used on some systems |
| 236 | in SPL. |
| 237 | |
Bin Meng | 1a43b23 | 2021-03-14 20:15:02 +0800 | [diff] [blame^] | 238 | config SIMPLE_BUS_CORRECT_RANGE |
| 239 | bool "Decode the 'simple-bus' <range> by honoring the #address-cells and #size-cells" |
| 240 | depends on SIMPLE_BUS |
| 241 | help |
| 242 | Decoding the 'simple-bus' <range> by honoring the #address-cells |
| 243 | and #size-cells of parent/child bus. If unset, #address-cells of |
| 244 | parent bus is assumed to be 1, #address-cells and #size-cells of |
| 245 | child bus is also assumed to be 1, to save some spaces of using |
| 246 | an advanced API to decode the <range>, which benefits SPL image |
| 247 | builds that have size limits. |
| 248 | |
| 249 | If you are unsure about this, Say N here. |
| 250 | |
Sean Anderson | b786054 | 2020-06-24 06:41:12 -0400 | [diff] [blame] | 251 | config SIMPLE_PM_BUS |
| 252 | bool "Support simple-pm-bus driver" |
| 253 | depends on DM && OF_CONTROL && CLK && POWER_DOMAIN |
| 254 | help |
| 255 | Supports the 'simple-pm-bus' driver, which is used for busses that |
| 256 | have power domains and/or clocks which need to be enabled before use. |
| 257 | |
Stefan Roese | adc0905 | 2015-09-02 07:41:12 +0200 | [diff] [blame] | 258 | config OF_TRANSLATE |
| 259 | bool "Translate addresses using fdt_translate_address" |
| 260 | depends on DM && OF_CONTROL |
| 261 | default y |
| 262 | help |
| 263 | If this option is enabled, the reg property will be translated |
| 264 | using the fdt_translate_address() function. This is necessary |
| 265 | on some platforms (e.g. MVEBU) using complex "ranges" |
| 266 | properties in many nodes. As this translation is not handled |
| 267 | correctly in the default simple_bus_translate() function. |
| 268 | |
| 269 | If this option is not enabled, simple_bus_translate() will be |
| 270 | used for the address translation. This function is faster and |
| 271 | smaller in size than fdt_translate_address(). |
| 272 | |
Dario Binacchi | b574d68 | 2020-12-30 00:16:21 +0100 | [diff] [blame] | 273 | config OF_TRANSLATE_ZERO_SIZE_CELLS |
| 274 | bool "Enable translation for zero size cells" |
| 275 | depends on OF_TRANSLATE |
| 276 | default n |
| 277 | help |
| 278 | The routine used to translate an FDT address into a physical CPU |
| 279 | address was developed by IBM. It considers that crossing any level |
| 280 | with #size-cells = <0> makes translation impossible, even if it is |
| 281 | not the way it was specified. |
| 282 | Enabling this option makes translation possible even in the case |
| 283 | of crossing levels with #size-cells = <0>. |
| 284 | |
Stefan Roese | adc0905 | 2015-09-02 07:41:12 +0200 | [diff] [blame] | 285 | config SPL_OF_TRANSLATE |
Stefan Roese | f05eeb2 | 2015-11-26 13:38:01 +0100 | [diff] [blame] | 286 | bool "Translate addresses using fdt_translate_address in SPL" |
Stefan Roese | adc0905 | 2015-09-02 07:41:12 +0200 | [diff] [blame] | 287 | depends on SPL_DM && SPL_OF_CONTROL |
| 288 | default n |
| 289 | help |
| 290 | If this option is enabled, the reg property will be translated |
| 291 | using the fdt_translate_address() function. This is necessary |
| 292 | on some platforms (e.g. MVEBU) using complex "ranges" |
| 293 | properties in many nodes. As this translation is not handled |
| 294 | correctly in the default simple_bus_translate() function. |
| 295 | |
| 296 | If this option is not enabled, simple_bus_translate() will be |
| 297 | used for the address translation. This function is faster and |
| 298 | smaller in size than fdt_translate_address(). |
| 299 | |
Stefan Roese | 85bddff | 2019-04-12 16:42:28 +0200 | [diff] [blame] | 300 | config TRANSLATION_OFFSET |
| 301 | bool "Platforms specific translation offset" |
| 302 | depends on DM && OF_CONTROL |
| 303 | help |
| 304 | Some platforms need a special address translation. Those |
| 305 | platforms (e.g. mvebu in SPL) can configure a translation |
| 306 | offset by enabling this option and setting the translation_offset |
| 307 | variable in the GD in their platform- / board-specific code. |
| 308 | |
Paul Burton | dbef2f0 | 2016-05-17 07:43:24 +0100 | [diff] [blame] | 309 | config OF_ISA_BUS |
| 310 | bool |
| 311 | depends on OF_TRANSLATE |
| 312 | help |
| 313 | Is this option is enabled then support for the ISA bus will |
| 314 | be included for addresses read from DT. This is something that |
| 315 | should be known to be required or not based upon the board |
Chris Packham | 3b6b462 | 2019-01-13 22:13:24 +1300 | [diff] [blame] | 316 | being targeted, and whether or not it makes use of an ISA bus. |
Paul Burton | dbef2f0 | 2016-05-17 07:43:24 +0100 | [diff] [blame] | 317 | |
| 318 | The bus is matched based upon its node name equalling "isa". The |
| 319 | busses #address-cells should equal 2, with the first cell being |
| 320 | used to hold flags & flag 0x1 indicating that the address range |
| 321 | should be accessed using I/O port in/out accessors. The second |
| 322 | cell holds the offset into ISA bus address space. The #size-cells |
| 323 | property should equal 1, and of course holds the size of the |
| 324 | address range used by a device. |
| 325 | |
| 326 | If this option is not enabled then support for the ISA bus is |
| 327 | not included and any such busses used in DT will be treated as |
| 328 | typical simple-bus compatible busses. This will lead to |
| 329 | mistranslation of device addresses, so ensure that this is |
| 330 | enabled if your board does include an ISA bus. |
| 331 | |
Simon Glass | 34ef0c2 | 2017-05-18 20:09:03 -0600 | [diff] [blame] | 332 | config DM_DEV_READ_INLINE |
| 333 | bool |
| 334 | default y if !OF_LIVE |
| 335 | |
Simon Glass | 9fb9e9b | 2020-04-09 10:27:38 -0600 | [diff] [blame] | 336 | config ACPIGEN |
| 337 | bool "Support ACPI table generation in driver model" |
Simon Glass | 55f3cde | 2020-07-16 21:22:39 -0600 | [diff] [blame] | 338 | default y if SANDBOX || (GENERATE_ACPI_TABLE && !QEMU) |
Simon Glass | 9fb9e9b | 2020-04-09 10:27:38 -0600 | [diff] [blame] | 339 | help |
| 340 | This option enables generation of ACPI tables using driver-model |
| 341 | devices. It adds a new operation struct to each driver, to support |
| 342 | things like generating device-specific tables and returning the ACPI |
| 343 | name of a device. |
| 344 | |
Simon Glass | 5f78ee7 | 2020-09-10 20:21:25 -0600 | [diff] [blame] | 345 | config BOUNCE_BUFFER |
| 346 | bool "Include bounce buffer API" |
| 347 | help |
| 348 | Some peripherals support DMA from a subset of physically |
| 349 | addressable memory only. To support such peripherals, the |
| 350 | bounce buffer API uses a temporary buffer: it copies data |
| 351 | to/from DMA regions while managing cache operations. |
| 352 | |
| 353 | A second possible use of bounce buffers is their ability to |
| 354 | provide aligned buffers for DMA operations. |
| 355 | |
Masahiro Yamada | cc85b7b | 2015-07-26 02:46:26 +0900 | [diff] [blame] | 356 | endmenu |