Heinrich Schuchardt | 0445978 | 2018-07-29 13:45:47 +0200 | [diff] [blame] | 1 | .. SPDX-License-Identifier: GPL-2.0+ |
| 2 | |
| 3 | Linker-Generated Arrays |
| 4 | ======================= |
| 5 | |
| 6 | A linker list is constructed by grouping together linker input |
| 7 | sections, each containing one entry of the list. Each input section |
| 8 | contains a constant initialized variable which holds the entry's |
| 9 | content. Linker list input sections are constructed from the list |
| 10 | and entry names, plus a prefix which allows grouping all lists |
| 11 | together. Assuming _list and _entry are the list and entry names, |
| 12 | then the corresponding input section name is |
| 13 | |
| 14 | :: |
| 15 | |
| 16 | .u_boot_list_ + 2_ + @_list + _2_ + @_entry |
| 17 | |
| 18 | and the C variable name is |
| 19 | |
| 20 | :: |
| 21 | |
| 22 | _u_boot_list + _2_ + @_list + _2_ + @_entry |
| 23 | |
| 24 | This ensures uniqueness for both input section and C variable name. |
| 25 | |
| 26 | Note that the names differ only in the first character, "." for the |
| 27 | section and "_" for the variable, so that the linker cannot confuse |
| 28 | section and symbol names. From now on, both names will be referred |
| 29 | to as |
| 30 | |
| 31 | :: |
| 32 | |
| 33 | %u_boot_list_ + 2_ + @_list + _2_ + @_entry |
| 34 | |
| 35 | Entry variables need never be referred to directly. |
| 36 | |
| 37 | The naming scheme for input sections allows grouping all linker lists |
| 38 | into a single linker output section and grouping all entries for a |
| 39 | single list. |
| 40 | |
| 41 | Note the two '_2_' constant components in the names: their presence |
| 42 | allows putting a start and end symbols around a list, by mapping |
| 43 | these symbols to sections names with components "1" (before) and |
| 44 | "3" (after) instead of "2" (within). |
| 45 | Start and end symbols for a list can generally be defined as |
| 46 | |
| 47 | :: |
| 48 | |
| 49 | %u_boot_list_2_ + @_list + _1_... |
| 50 | %u_boot_list_2_ + @_list + _3_... |
| 51 | |
| 52 | Start and end symbols for the whole of the linker lists area can be |
| 53 | defined as |
| 54 | |
| 55 | :: |
| 56 | |
| 57 | %u_boot_list_1_... |
| 58 | %u_boot_list_3_... |
| 59 | |
| 60 | Here is an example of the sorted sections which result from a list |
| 61 | "array" made up of three entries : "first", "second" and "third", |
| 62 | iterated at least once. |
| 63 | |
| 64 | :: |
| 65 | |
| 66 | .u_boot_list_2_array_1 |
| 67 | .u_boot_list_2_array_2_first |
| 68 | .u_boot_list_2_array_2_second |
| 69 | .u_boot_list_2_array_2_third |
| 70 | .u_boot_list_2_array_3 |
| 71 | |
| 72 | If lists must be divided into sublists (e.g. for iterating only on |
| 73 | part of a list), one can simply give the list a name of the form |
| 74 | 'outer_2_inner', where 'outer' is the global list name and 'inner' |
| 75 | is the sub-list name. Iterators for the whole list should use the |
| 76 | global list name ("outer"); iterators for only a sub-list should use |
| 77 | the full sub-list name ("outer_2_inner"). |
| 78 | |
| 79 | Here is an example of the sections generated from a global list |
| 80 | named "drivers", two sub-lists named "i2c" and "pci", and iterators |
| 81 | defined for the whole list and each sub-list: |
| 82 | |
| 83 | :: |
| 84 | |
| 85 | %u_boot_list_2_drivers_1 |
| 86 | %u_boot_list_2_drivers_2_i2c_1 |
| 87 | %u_boot_list_2_drivers_2_i2c_2_first |
| 88 | %u_boot_list_2_drivers_2_i2c_2_first |
| 89 | %u_boot_list_2_drivers_2_i2c_2_second |
| 90 | %u_boot_list_2_drivers_2_i2c_2_third |
| 91 | %u_boot_list_2_drivers_2_i2c_3 |
| 92 | %u_boot_list_2_drivers_2_pci_1 |
| 93 | %u_boot_list_2_drivers_2_pci_2_first |
| 94 | %u_boot_list_2_drivers_2_pci_2_second |
| 95 | %u_boot_list_2_drivers_2_pci_2_third |
| 96 | %u_boot_list_2_drivers_2_pci_3 |
| 97 | %u_boot_list_2_drivers_3 |
| 98 | |
| 99 | .. kernel-doc:: include/linker_lists.h |
| 100 | :internal: |