u-boot/doc/api/linker_lists.rst

160 lines
5.6 KiB
ReStructuredText
Raw Normal View History

.. SPDX-License-Identifier: GPL-2.0+
Linker-Generated Arrays
=======================
A linker list is constructed by grouping together linker input
sections, each containing one entry of the list. Each input section
contains a constant initialized variable which holds the entry's
content. Linker list input sections are constructed from the list
and entry names, plus a prefix which allows grouping all lists
together. Assuming _list and _entry are the list and entry names,
then the corresponding input section name is
::
.u_boot_list_ + 2_ + @_list + _2_ + @_entry
and the C variable name is
::
_u_boot_list + _2_ + @_list + _2_ + @_entry
This ensures uniqueness for both input section and C variable name.
Note that the names differ only in the first character, "." for the
section and "_" for the variable, so that the linker cannot confuse
section and symbol names. From now on, both names will be referred
to as
::
%u_boot_list_ + 2_ + @_list + _2_ + @_entry
Entry variables need never be referred to directly.
The naming scheme for input sections allows grouping all linker lists
into a single linker output section and grouping all entries for a
single list.
Note the two '_2_' constant components in the names: their presence
allows putting a start and end symbols around a list, by mapping
these symbols to sections names with components "1" (before) and
"3" (after) instead of "2" (within).
Start and end symbols for a list can generally be defined as
::
%u_boot_list_2_ + @_list + _1_...
%u_boot_list_2_ + @_list + _3_...
Start and end symbols for the whole of the linker lists area can be
defined as
::
%u_boot_list_1_...
%u_boot_list_3_...
Here is an example of the sorted sections which result from a list
"array" made up of three entries : "first", "second" and "third",
iterated at least once.
::
.u_boot_list_2_array_1
.u_boot_list_2_array_2_first
.u_boot_list_2_array_2_second
.u_boot_list_2_array_2_third
.u_boot_list_2_array_3
If lists must be divided into sublists (e.g. for iterating only on
part of a list), one can simply give the list a name of the form
'outer_2_inner', where 'outer' is the global list name and 'inner'
is the sub-list name. Iterators for the whole list should use the
global list name ("outer"); iterators for only a sub-list should use
the full sub-list name ("outer_2_inner").
Here is an example of the sections generated from a global list
named "drivers", two sub-lists named "i2c" and "pci", and iterators
defined for the whole list and each sub-list:
::
%u_boot_list_2_drivers_1
%u_boot_list_2_drivers_2_i2c_1
%u_boot_list_2_drivers_2_i2c_2_first
%u_boot_list_2_drivers_2_i2c_2_first
%u_boot_list_2_drivers_2_i2c_2_second
%u_boot_list_2_drivers_2_i2c_2_third
%u_boot_list_2_drivers_2_i2c_3
%u_boot_list_2_drivers_2_pci_1
%u_boot_list_2_drivers_2_pci_2_first
%u_boot_list_2_drivers_2_pci_2_second
%u_boot_list_2_drivers_2_pci_2_third
%u_boot_list_2_drivers_2_pci_3
%u_boot_list_2_drivers_3
linker_lists: Fix alignment issue The linker script uses alphabetic sorting to group the different linker lists together. Each group has its own struct and potentially its own alignment. But when the linker packs the structs together it cannot ensure that a linker list starts on the expected alignment boundary. For example, if the first list has a struct size of 8 and we place 3 of them in the image, that means that the next struct will start at offset 0x18 from the start of the linker_list section. If the next struct has a size of 16 then it will start at an 8-byte aligned offset, but not a 16-byte aligned offset. With sandbox on x86_64, a reference to a linker list item using ll_entry_get() can force alignment of that particular linker_list item, if it is in the same file as the linker_list item is declared. Consider this example, where struct driver is 0x80 bytes: ll_entry_declare(struct driver, fred, driver) ... void *p = ll_entry_get(struct driver, fred, driver) If these two lines of code are in the same file, then the entry is forced to be aligned at the 'struct driver' alignment, which is 16 bytes. If the second line of code is in a different file, then no action is taken, since the compiler cannot update the alignment of the linker_list item. In the first case, an 8-byte 'fill' region is added: .u_boot_list_2_driver_2_testbus_drv 0x0000000000270018 0x80 test/built-in.o 0x0000000000270018 _u_boot_list_2_driver_2_testbus_drv .u_boot_list_2_driver_2_testfdt1_drv 0x0000000000270098 0x80 test/built-in.o 0x0000000000270098 _u_boot_list_2_driver_2_testfdt1_drv *fill* 0x0000000000270118 0x8 .u_boot_list_2_driver_2_testfdt_drv 0x0000000000270120 0x80 test/built-in.o 0x0000000000270120 _u_boot_list_2_driver_2_testfdt_drv .u_boot_list_2_driver_2_testprobe_drv 0x00000000002701a0 0x80 test/built-in.o 0x00000000002701a0 _u_boot_list_2_driver_2_testprobe_drv With this, the linker_list no-longer works since items after testfdt1_drv are not at the expected address. Ideally we would have a way to tell gcc not to align structs in this way. It is not clear how we could do this, and in any case it would require us to adjust every struct used by the linker_list feature. One possible fix is to force each separate linker_list to start on the largest possible boundary that can be required by the compiler. However that does not seem to work on x86_64, which uses 16-byte alignment in this case but needs 32-byte alignment. So add a Kconfig option to handle this. Set the default value to 4 so as to avoid changing platforms that don't need it. Update the ll_entry_start() accordingly. Signed-off-by: Simon Glass <sjg@chromium.org>
2020-12-17 04:20:06 +00:00
Alignment issues
----------------
The linker script uses alphabetic sorting to group the different linker
lists together. Each group has its own struct and potentially its own
alignment. But when the linker packs the structs together it cannot ensure
that a linker list starts on the expected alignment boundary.
For example, if the first list has a struct size of 8 and we place 3 of
them in the image, that means that the next struct will start at offset
0x18 from the start of the linker_list section. If the next struct has
a size of 16 then it will start at an 8-byte aligned offset, but not a
16-byte aligned offset.
With sandbox on x86_64, a reference to a linker list item using
ll_entry_get() can force alignment of that particular linker_list item,
if it is in the same file as the linker_list item is declared.
Consider this example, where struct driver is 0x80 bytes::
ll_entry_declare(struct driver, fred, driver)
...
void *p = ll_entry_get(struct driver, fred, driver)
If these two lines of code are in the same file, then the entry is forced
to be aligned at the 'struct driver' alignment, which is 16 bytes. If the
second line of code is in a different file, then no action is taken, since
the compiler cannot update the alignment of the linker_list item.
In the first case, an 8-byte 'fill' region is added::
.u_boot_list_2_driver_2_testbus_drv
0x0000000000270018 0x80 test/built-in.o
0x0000000000270018 _u_boot_list_2_driver_2_testbus_drv
.u_boot_list_2_driver_2_testfdt1_drv
0x0000000000270098 0x80 test/built-in.o
0x0000000000270098 _u_boot_list_2_driver_2_testfdt1_drv
*fill* 0x0000000000270118 0x8
.u_boot_list_2_driver_2_testfdt_drv
0x0000000000270120 0x80 test/built-in.o
0x0000000000270120 _u_boot_list_2_driver_2_testfdt_drv
.u_boot_list_2_driver_2_testprobe_drv
0x00000000002701a0 0x80 test/built-in.o
0x00000000002701a0 _u_boot_list_2_driver_2_testprobe_drv
With this, the linker_list no-longer works since items after testfdt1_drv
are not at the expected address.
Ideally we would have a way to tell gcc not to align structs in this way.
It is not clear how we could do this, and in any case it would require us
to adjust every struct used by the linker_list feature.
The simplest fix seems to be to force each separate linker_list to start
on the largest possible boundary that can be required by the compiler. This
is the purpose of CONFIG_LINKER_LIST_ALIGN
.. kernel-doc:: include/linker_lists.h
:internal: