carl9170fw.git
6 weeks agocarl9170 toolchain: update gcc to 11.1 master
Christian Lamparter [Sat, 1 May 2021 19:43:13 +0000 (21:43 +0200)]
carl9170 toolchain: update gcc to 11.1

All the way to 11!

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170 firmware build: remove byproducts carl9170's dsc bin files
Christian Lamparter [Sat, 13 Feb 2021 11:42:09 +0000 (12:42 +0100)]
carl9170 firmware build: remove byproducts carl9170's dsc bin files

These files are produced as part of the firmware built. These
should be removed by "make clean".

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170 firmware miniboot: remove miniboot.fw on make clean
Christian Lamparter [Sat, 13 Feb 2021 11:44:59 +0000 (12:44 +0100)]
carl9170 firmware miniboot: remove miniboot.fw on make clean

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170 toolchain: update binutils, mprf, gmp, newlib, mpc
Christian Lamparter [Sat, 13 Feb 2021 09:49:18 +0000 (10:49 +0100)]
carl9170 toolchain: update binutils, mprf, gmp, newlib, mpc

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170 firmware build: explicitly list carl9170.elf in dependency
Christian Lamparter [Sat, 13 Feb 2021 10:05:07 +0000 (11:05 +0100)]
carl9170 firmware build: explicitly list carl9170.elf in dependency

This should fix build errors like:

sh-elf-objcopy: 'carl9170.elf': No such file
make[2]: *** [carlfw/CMakeFiles/firmware.dir/build.make:76: firmware] Error 1
make[1]: *** [CMakeFiles/Makefile2:204: carlfw/CMakeFiles/firmware.dir/all] Error 2
make: *** [Makefile:103: all] Error 2

no idea what broke in cmake there.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: fix return value of do_error_if()
Masahiro Yamada [Sat, 19 Dec 2020 18:18:42 +0000 (03:18 +0900)]
kconfig: fix return value of do_error_if()

$(error-if,...) is expanded to an empty string. Currently, it relies on
eval_clause() returning xstrdup("") when all attempts for expansion fail,
but the correct implementation is to make do_error_if() return xstrdup("").

Fixes: 1d6272e6fe43 ("kconfig: add 'info', 'warning-if', and 'error-if' built-in functions")
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: clean up header inclusion
Boris Kolpackov [Mon, 23 Nov 2020 09:38:18 +0000 (11:38 +0200)]
kconfig: clean up header inclusion

- Add missing includes.
- Remove no longer necessary includes.

Signed-off-by: Boris Kolpackov <boris@codesynthesis.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: make lkc.h self-sufficient #include-wise
Boris Kolpackov [Thu, 29 Oct 2020 15:51:51 +0000 (17:51 +0200)]
kconfig: make lkc.h self-sufficient #include-wise

Signed-off-by: Boris Kolpackov <boris@codesynthesis.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: qconf: fix incomplete type 'struct gstr' warning
Masahiro Yamada [Tue, 8 Sep 2020 22:16:38 +0000 (07:16 +0900)]
kconfig: qconf: fix incomplete type 'struct gstr' warning

"make HOSTCXX=clang++ xconfig" reports the following:

  HOSTCXX scripts/kconfig/qconf.o
In file included from scripts/kconfig/qconf.cc:23:
In file included from scripts/kconfig/lkc.h:15:
scripts/kconfig/lkc_proto.h:26:13: warning: 'get_relations_str' has C-linkage specified, but returns incomplete type 'struct gstr' which could be incompatible with C [-Wreturn-type-c-linkage]
struct gstr get_relations_str(struct symbol **sym_arr, struct list_head *head);
            ^

Currently, get_relations_str() is declared before the struct gstr
definition.

Move all declarations of menu.c functions below.

BTW, some are declared in lkc.h and some in lkc_proto.h, but the
difference is unclear. I guess some refactoring is needed.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Acked-by: Boris Kolpackov <boris@codesynthesis.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: add 'static' to some file-local data
Masahiro Yamada [Wed, 29 Jul 2020 03:18:37 +0000 (12:18 +0900)]
kconfig: add 'static' to some file-local data

Fix some warnings from sparce like follows:

  warning: symbol '...' was not declared. Should it be static?

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: remove '---help---' support
Masahiro Yamada [Wed, 17 Jun 2020 03:02:19 +0000 (12:02 +0900)]
kconfig: remove '---help---' support

The conversion is done. No more user of '---help---'.

Cc: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170 firmware: replace ---help--- with help syntax
Christian Lamparter [Fri, 5 Feb 2021 11:03:09 +0000 (12:03 +0100)]
carl9170 firmware: replace ---help--- with help syntax

upstream kconfig is replacing the ---help--- command syntax with help.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: allow only 'config', 'comment', and 'if' inside 'choice'
Masahiro Yamada [Fri, 24 Apr 2020 05:49:29 +0000 (14:49 +0900)]
kconfig: allow only 'config', 'comment', and 'if' inside 'choice'

The code block surrounded by 'if' ... 'endif' is reduced into if_stmt,
which is accepted in the 'choice' context. Therefore, you can write any
statements within a choice block by wrapping 'if y' ... 'end'.

For example, you can create a menu inside a choice, like follows:

---------------->8----------------
  choice
          prompt "choice"

  config A
          bool "A"

  config B
          bool "B"

  if y

  menu "strange menu"

  config C
          bool "C"

  endmenu

  endif

  endchoice
---------------->8----------------

I want to change such a weird structure into a syntax error.

In fact, the USB gadget Kconfig had used nested 'choice' for no good
reason until commit df8df5e4bc37 ("usb: get rid of 'choice' for
legacy gadget drivers") killed it.

I think the 'source' inside 'choice' is on the fence. It is at least
gramatically sensible as long as the included file contains only
bool/tristate configs. However, it makes the code unreadable, and people
tend to forget the fact that the file is included from the choice
block. Commit 10e5e6c24963 ("usb: gadget: move choice ... endchoice to
legacy/Kconfig") got rid of the only usecase.

Going forward, you can only use 'config', 'comment', and 'if' inside
'choice'. This also recursively applies to 'if' blocks inside 'choice'.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: do not assign a variable in the return statement
Masahiro Yamada [Mon, 13 Apr 2020 15:35:42 +0000 (00:35 +0900)]
kconfig: do not assign a variable in the return statement

I am not a big fan of doing assignment in a return statement.
Split it into two lines.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: do not use OR-assignment for zero-cleared structure
Masahiro Yamada [Mon, 13 Apr 2020 15:33:20 +0000 (00:33 +0900)]
kconfig: do not use OR-assignment for zero-cleared structure

The simple assignment is enough because memset() three lines above
has zero-cleared the structure.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months ago.gitignore: add SPDX License Identifier
Masahiro Yamada [Tue, 3 Mar 2020 13:35:59 +0000 (22:35 +0900)]
.gitignore: add SPDX License Identifier

Add SPDX License Identifier to all .gitignore files.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: make 'imply' obey the direct dependency
Masahiro Yamada [Mon, 2 Mar 2020 06:23:40 +0000 (15:23 +0900)]
kconfig: make 'imply' obey the direct dependency

The 'imply' statement may create unmet direct dependency when the
implied symbol depends on m.

[Test Code]

  config FOO
          tristate "foo"
          imply BAZ

  config BAZ
          tristate "baz"
          depends on BAR

  config BAR
          def_tristate m

  config MODULES
          def_bool y
          option modules

If you set FOO=y, BAZ is also promoted to y, which results in the
following .config file:

  CONFIG_FOO=y
  CONFIG_BAZ=y
  CONFIG_BAR=m
  CONFIG_MODULES=y

This does not meet the dependency 'BAZ depends on BAR'.

Unlike 'select', what is worse, Kconfig never shows the
'WARNING: unmet direct dependencies detected for ...' for this case.

Because 'imply' is considered to be weaker than 'depends on', Kconfig
should take the direct dependency into account.

For clarification, describe this case in kconfig-language.rst too.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Tested-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: allow symbols implied by y to become m
Masahiro Yamada [Mon, 2 Mar 2020 06:23:39 +0000 (15:23 +0900)]
kconfig: allow symbols implied by y to become m

The 'imply' keyword restricts a symbol to y or n, excluding m
when it is implied by y. This is the original behavior since
commit 237e3ad0f195 ("Kconfig: Introduce the "imply" keyword").

However, the author of this feature, Nicolas Pitre, stated that
the 'imply' keyword should not impose any restrictions.
(https://lkml.org/lkml/2020/2/19/714)

I agree, and want to get rid of this tricky behavior.

Suggested-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Acked-by: Nicolas Pitre <nico@fluxnic.net>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: Invalidate all symbols after changing to y or m.
Tetsuo Handa [Tue, 4 Feb 2020 04:08:44 +0000 (13:08 +0900)]
kconfig: Invalidate all symbols after changing to y or m.

Since commit 89b9060987d9 ("kconfig: Add yes2modconfig and
mod2yesconfig targets.") forgot to clear SYMBOL_VALID bit after
changing to y or m, these targets did not save the changes.
Call sym_clear_all_valid() so that all symbols are revalidated.

Fixes: 89b9060987d9 ("kconfig: Add yes2modconfig and mod2yesconfig targets.")
Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: fix broken dependency in randconfig-generated .config
Masahiro Yamada [Sat, 1 Feb 2020 05:03:11 +0000 (14:03 +0900)]
kconfig: fix broken dependency in randconfig-generated .config

Running randconfig on arm64 using KCONFIG_SEED=0x40C5E904 (e.g. on v5.5)
produces the .config with CONFIG_EFI=y and CONFIG_CPU_BIG_ENDIAN=y,
which does not meet the !CONFIG_CPU_BIG_ENDIAN dependency.

This is because the user choice for CONFIG_CPU_LITTLE_ENDIAN vs
CONFIG_CPU_BIG_ENDIAN is set by randomize_choice_values() after the
value of CONFIG_EFI is calculated.

When this happens, the has_changed flag should be set.

Currently, it takes the result from the last iteration. It should
accumulate all the results of the loop.

Fixes: 3b9a19e08960 ("kconfig: loop as long as we changed some symbols in randconfig")
Reported-by: Vincenzo Frascino <vincenzo.frascino@arm.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: fix nesting of symbol help text
Thomas Hebb [Tue, 17 Dec 2019 16:15:45 +0000 (08:15 -0800)]
kconfig: fix nesting of symbol help text

When we generate the help text of a symbol (e.g. when a user presses '?'
in menuconfig), we do two things:

 1. We iterate through every prompt that belongs to that symbol,
    printing its text and its location in the menu tree.
 2. We print symbol-wide information that's not linked to a particular
    prompt, such as what it selects/is selected by and what it
    implies/is implied by.

Each prompt we print for 1 starts with a line that's not indented
indicating where the prompt is defined, then continues with indented
lines that describe properties of that particular definition.

Once we get to 2, however, we print all the global data indented as
well! Visually, this makes it look like the symbol-wide data is
associated with the last prompt we happened to print rather than
the symbol as a whole.

Fix this by removing the indentation for symbol-wide information.

Before:

  Symbol: CPU_FREQ [=n]
  Type  : bool
  Defined at drivers/cpufreq/Kconfig:4
    Prompt: CPU Frequency scaling
    Location:
      -> CPU Power Management
        -> CPU Frequency scaling
    Selects: SRCU [=n]
    Selected by [n]:
    - ARCH_SA1100 [=n] && <choice>

After:

  Symbol: CPU_FREQ [=n]
  Type  : bool
  Defined at drivers/cpufreq/Kconfig:4
    Prompt: CPU Frequency scaling
    Location:
      -> CPU Power Management
        -> CPU Frequency scaling
  Selects: SRCU [=n]
  Selected by [n]:
    - ARCH_SA1100 [=n] && <choice>

Signed-off-by: Thomas Hebb <tommyhebb@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: distinguish between dependencies and visibility in help text
Thomas Hebb [Tue, 17 Dec 2019 16:15:44 +0000 (08:15 -0800)]
kconfig: distinguish between dependencies and visibility in help text

Kconfig makes a distinction between dependencies (defined by "depends
on" expressions and enclosing "if" blocks) and visibility (which
includes all dependencies, but also includes inline "if" expressions of
individual properties as well as, for prompts, "visible if" expressions
of enclosing menus).

Before commit bcdedcc1afd6 ("menuconfig: print more info for symbol
without prompts"), the "Depends on" lines of a symbol's help text
indicated the visibility of the prompt property they appeared under.
After bcdedcc1afd, there was always only a single "Depends on" line,
which indicated the visibility of the first P_SYMBOL property of the
symbol. Since P_SYMBOLs never have inline if expressions, this was in
effect the same as the dependencies of the menu item that the P_SYMBOL
was attached to.

Neither of these situations accurately conveyed the dependencies of a
symbol--the first because it was actually the visibility, and the second
because it only showed the dependencies from a single definition.

With this series, we are back to printing separate dependencies for each
definition, but we print the actual dependencies (rather than the
visibility) in the "Depends on" line. However, it can still be useful to
know the visibility of a prompt, so this patch adds a "Visible if" line
that shows the visibility only if the visibility is different from the
dependencies (which it isn't for most prompts in Linux).

Before:

  Symbol: THUMB2_KERNEL [=n]
  Type  : bool
  Defined at arch/arm/Kconfig:1417
    Prompt: Compile the kernel in Thumb-2 mode
    Depends on: (CPU_V7 [=y] || CPU_V7M [=n]) && !CPU_V6 [=n] && !CPU_V6K [=n]
    Location:
      -> Kernel Features
    Selects: ARM_UNWIND [=n]

After:

   Symbol: THUMB2_KERNEL [=n]
   Type  : bool
   Defined at arch/arm/Kconfig:1417
     Prompt: Compile the kernel in Thumb-2 mode
     Depends on: (CPU_V7 [=y] || CPU_V7M [=n]) && !CPU_V6 [=n] && !CPU_V6K [=n]
     Visible if: (CPU_V7 [=y] || CPU_V7M [=n]) && !CPU_V6 [=n] && !CPU_V6K [=n] && !CPU_THUMBONLY [=n]
     Location:
       -> Kernel Features
     Selects: ARM_UNWIND [=n]

Signed-off-by: Thomas Hebb <tommyhebb@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: list all definitions of a symbol in help text
Thomas Hebb [Tue, 17 Dec 2019 16:15:43 +0000 (08:15 -0800)]
kconfig: list all definitions of a symbol in help text

In Kconfig, each symbol (representing a config option) can be defined in
multiple places. Each definition may or may not have a prompt, which
allows the option to be set via an interface like menuconfig. Each
definition has a set of dependencies, which determine whether its prompt
is visible and whether other pieces of the definition, like a default
value, take effect.

Historically, a symbol's help text (i.e. what's shown when a user
presses '?' in menuconfig) contained some symbol-wide information not
tied to any particular definition (e.g. what other symbols it selects)
as well as the location (file name and line number) and dependencies of
each prompt. Notably, the help text did not show the location or
dependencies of definitions without prompts.

Because this made it hard to reason about symbols that had no prompts,
commit bcdedcc1afd6 ("menuconfig: print more info for symbol without
prompts") changed the help text so that, instead of containing the
location and dependencies of each prompt, it contained the location and
dependencies of the symbol's first definition, regardless of whether or
not that definition had a prompt.

For symbols with only one definition, that change makes sense. However,
it breaks down for symbols with multiple definitions: each definition
has its own set of dependencies (the `dep` field of `struct menu`), and
those dependencies are ORed together to get the symbol's dependency list
(the `dir_dep` field of `struct symbol`). By printing only the
dependencies of the first definition, the help text misleads users into
believing that an option is more narrowly-applicable than it actually
is.

For an extreme example of this, we can look at the SYS_TEXT_BASE symbol
in the Das U-Boot project (version 2019.10), which also uses Kconfig. (I
unfortunately could not find an illustrative example in Linux.) This
config option specifies the load address of the built binary and, as
such, is applicable to basically every configuration possible. And yet,
without this patch, its help text is as follows:

  Symbol: SYS_TEXT_BASE [=]
  Type  : hex
  Prompt: U-Boot base address
    Location:
      -> ARM architecture
  Prompt: Text Base
    Location:
      -> Boot images
    Defined at arch/arm/mach-aspeed/Kconfig:9
    Depends on: ARM [=n] && ARCH_ASPEED [=n]

The help text indicates that the option is applicable only for a
specific unselected architecture (aspeed), because that architecture's
promptless definition (which just sets a default value), happens to be
the first one seen. No definition or dependency information is printed
for either of the two prompts listed.

Because source locations and dependencies are fundamentally properties
of definitions and not of symbols, we should treat them as such. This
patch brings back the pre-bcdedcc1afd6 behavior for definitions with
prompts but also separately prints the location and dependencies of
those without prompts, solving the original problem in a different way.
With this change, our SYS_TEXT_BASE example becomes

   Symbol: SYS_TEXT_BASE [=]
   Type  : hex
   Defined at arch/arm/mach-stm32mp/Kconfig:83
     Prompt: U-Boot base address
     Depends on: ARM [=n] && ARCH_STM32MP [=n]
     Location:
       -> ARM architecture
   Defined at Kconfig:532
     Prompt: Text Base
     Depends on: !NIOS2 [=n] && !XTENSA [=n] && !EFI_APP [=n]
     Location:
       -> Boot images
   Defined at arch/arm/mach-aspeed/Kconfig:9
     Depends on: ARM [=n] && ARCH_ASPEED [=n]
   Defined  at arch/arm/mach-socfpga/Kconfig:25
     Depends on: ARM [=n] && ARCH_SOCFPGA [=n]
   <snip>
   Defined at board/sifive/fu540/Kconfig:15
     Depends on: RISCV [=n] && TARGET_SIFIVE_FU540 [=n]

which is a much more accurate representation.

Note that there is one notable difference between what gets printed for
prompts after this change and what got printed before bcdedcc1afd6: the
"Depends on" line now accurately represents the prompt's dependencies
instead of conflating those with the prompt's visibility (which can
include extra conditions). See the patch later in this series titled
"kconfig: distinguish between dependencies and visibility in help text"
for more details and better handling of that nuance.

Signed-off-by: Thomas Hebb <tommyhebb@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: Add yes2modconfig and mod2yesconfig targets.
Tetsuo Handa [Tue, 17 Dec 2019 09:42:06 +0000 (18:42 +0900)]
kconfig: Add yes2modconfig and mod2yesconfig targets.

Since kernel configs provided by syzbot are close to "make allyesconfig",
it takes long time to rebuild. This is especially waste of time when we
need to rebuild for many times (e.g. doing manual printk() inspection,
bisect operations).

We can save time if we can exclude modules which are irrelevant to each
problem. But "make localmodconfig" cannot exclude modules which are built
into vmlinux because /sbin/lsmod output is used as the source of modules.

Therefore, this patch adds "make yes2modconfig" which converts from =y
to =m if possible. After confirming that the interested problem is still
reproducible, we can try "make localmodconfig" (and/or manually tune
based on "Modules linked in:" line) in order to exclude modules which are
irrelevant to the interested problem. While we are at it, this patch also
adds "make mod2yesconfig" which converts from =m to =y in case someone
wants to convert from =m to =y after "make localmodconfig".

Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp>
Cc: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: squash prop_alloc() into menu_add_prop()
Masahiro Yamada [Tue, 17 Dec 2019 04:14:23 +0000 (13:14 +0900)]
kconfig: squash prop_alloc() into menu_add_prop()

prop_alloc() is only called from menu_add_prop(). Squash it.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: remove sym from struct property
Masahiro Yamada [Tue, 17 Dec 2019 04:14:22 +0000 (13:14 +0900)]
kconfig: remove sym from struct property

struct property can reference to the symbol that it is associated with
by prop->menu->sym.

Fix up the one usage of prop->sym, and remove sym from struct property.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: remove 'prompt' argument from menu_add_prop()
Masahiro Yamada [Tue, 17 Dec 2019 04:14:21 +0000 (13:14 +0900)]
kconfig: remove 'prompt' argument from menu_add_prop()

This function no longer uses the 'prompt' argument.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: move prompt handling to menu_add_prompt() from menu_add_prop()
Masahiro Yamada [Tue, 17 Dec 2019 04:14:20 +0000 (13:14 +0900)]
kconfig: move prompt handling to menu_add_prompt() from menu_add_prop()

menu_add_prompt() is the only function that calls menu_add_prop() with
non-NULL prompt.

So, the code inside the if-conditional block of menu_add_prop() can be
moved to menu_add_prompt().

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: remove 'prompt' symbol
Masahiro Yamada [Tue, 17 Dec 2019 04:14:19 +0000 (13:14 +0900)]
kconfig: remove 'prompt' symbol

Now that 'prompt' is only reduced from T_WORD_QUOTE without any action,
use T_WORD_QUOTE directly.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: drop T_WORD from the RHS of 'prompt' symbol
Masahiro Yamada [Tue, 17 Dec 2019 04:14:18 +0000 (13:14 +0900)]
kconfig: drop T_WORD from the RHS of 'prompt' symbol

Commit 8636a1f9677d ("treewide: surround Kconfig file paths with double
quotes") killed use-cases to reduce an unquoted string into the 'prompt'
symbol.

Kconfig still allows to use an unquoted string in the context of menu,
source, or prompt.

So, you can omit quoting if the prompt is a single word:

    bool foo

..., but I do not think this is so useful.

Let's require quoting:

    bool "foo"

All the Kconfig files in the kernel are written in this way.

Remove the T_WORD from the right-hand side of the symbol 'prompt'.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: use parent->dep as the parentdep of 'menu'
Masahiro Yamada [Tue, 17 Dec 2019 04:14:17 +0000 (13:14 +0900)]
kconfig: use parent->dep as the parentdep of 'menu'

In menu_finalize(), the dependency of a menu entry is propagated
downwards.

For the 'menu', parent->dep and parent->prompt->visible.expr have
the same expression. Both accumulate the 'depends on' of itself and
upper menu entries.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokconfig: remove the rootmenu check in menu_add_prop()
Masahiro Yamada [Tue, 17 Dec 2019 04:14:16 +0000 (13:14 +0900)]
kconfig: remove the rootmenu check in menu_add_prop()

This reverts commit ba6ff60d5eb4 ("kconfig: don't emit warning upon
rootmenu's prompt redefinition").

At that time, rootmenu.prompt was always set first, then it was set
again if a "mainmenu" statement was specified in the Kconfig file.

This is no longer the case since commit 0724a7c32a54 ("kconfig: Don't
leak main menus during parsing"). Remove the unneeded check.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agokbuild: create modules.builtin without Makefile.modbuiltin or tristate.conf
Masahiro Yamada [Thu, 19 Dec 2019 08:33:29 +0000 (17:33 +0900)]
kbuild: create modules.builtin without Makefile.modbuiltin or tristate.conf

Commit bc081dd6e9f6 ("kbuild: generate modules.builtin") added
infrastructure to generate modules.builtin, the list of all
builtin modules.

Basically, it works like this:

  - Kconfig generates include/config/tristate.conf, the list of
    tristate CONFIG options with a value in a capital letter.

  - scripts/Makefile.modbuiltin makes Kbuild descend into
    directories to collect the information of builtin modules.

I am not a big fan of it because Kbuild ends up with traversing
the source tree twice.

I am not sure how perfectly it should work, but this approach cannot
avoid false positives; even if the relevant CONFIG option is tristate,
some Makefiles forces obj-m to obj-y.

Some examples are:

  arch/powerpc/platforms/powermac/Makefile:
    obj-$(CONFIG_NVRAM:m=y)         += nvram.o

  net/ipv6/Makefile:
    obj-$(subst m,y,$(CONFIG_IPV6)) += inet6_hashtables.o

  net/netlabel/Makefile:
    obj-$(subst m,y,$(CONFIG_IPV6)) += netlabel_calipso.o

Nobody has complained about (or noticed) it, so it is probably fine to
have false positives in modules.builtin.

This commit simplifies the implementation. Let's exploit the fact
that every module has MODULE_LICENSE(). (modpost shows a warning if
MODULE_LICENSE is missing. If so, 0-day bot would already have blocked
such a module.)

I added MODULE_FILE to <linux/module.h>. When the code is being compiled
as builtin, it will be filled with the file path of the module, and
collected into modules.builtin.info. Then, scripts/link-vmlinux.sh
extracts the list of builtin modules out of it.

This new approach fixes the false-positives above, but adds another
type of false-positives; non-modular code may have MODULE_LICENSE()
by mistake. This is not a big deal, it is just the code is always
orphan. We can clean it up if we like. You can see cleanup examples by:

  $ git log --grep='make.* explicitly non-modular'

To sum up, this commits deletes lots of code, but still produces almost
equivalent results. Please note it does not increase the vmlinux size at
all. As you can see in include/asm-generic/vmlinux.lds.h, the .modinfo
section is discarded in the link stage.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: update reduced neighbor report TBTT info length
Avraham Stern [Sun, 29 Nov 2020 15:30:52 +0000 (17:30 +0200)]
ieee80211: update reduced neighbor report TBTT info length

A new field (20MHz PSD - 1 byte) was added to the RNR TBTT info field.
Adjust the expected TBTT info length accordingly.

Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20201129172929.b503adccce6a.Ie684e1d3039c111bf2d521bf762aaec3f7a24d2e@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: Parse SAE H2E only membership selector
Ilan Peer [Sun, 29 Nov 2020 15:30:46 +0000 (17:30 +0200)]
cfg80211: Parse SAE H2E only membership selector

This extends the support for drivers that rebuild IEs in the
FW (same as with HT/VHT/HE).

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20201129172929.4012647275f3.I1a93ae71c57ef0b6f58f99d47fce919d19d65ff0@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: Add definition for WFA DPP
Kurt Lee [Mon, 12 Oct 2020 08:43:46 +0000 (03:43 -0500)]
ieee80211: Add definition for WFA DPP

Add Wi-Fi Alliance definition for DPP (Device Provisioning Protocol).

Signed-off-by: Kurt Lee <kurt.lee@cypress.com>
Signed-off-by: Wright Feng <wright.feng@cypress.com>
Link: https://lore.kernel.org/r/20201012084347.121557-2-wright.feng@cypress.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agonl80211: extend support to config spatial reuse parameter set
Rajkumar Manoharan [Mon, 28 Sep 2020 07:28:11 +0000 (00:28 -0700)]
nl80211: extend support to config spatial reuse parameter set

Allow the user to configure below Spatial Reuse Parameter Set element.
  * Non-SRG OBSS PD Max Offset
  * SRG BSS Color Bitmap
  * SRG Partial BSSID Bitmap

Signed-off-by: Rajkumar Manoharan <rmanohar@codeaurora.org>
Link: https://lore.kernel.org/r/1601278091-20313-2-git-send-email-rmanohar@codeaurora.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: support S1G association
Thomas Pedersen [Tue, 22 Sep 2020 02:28:15 +0000 (19:28 -0700)]
mac80211: support S1G association

The changes required for associating in S1G are:

- apply S1G BSS channel info before assoc
- mark all S1G STAs as QoS STAs
- include and parse AID request element
- handle new Association Response format
- don't fail assoc if supported rates element is missing

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200922022818.15855-15-thomas@adapt-ip.com
[pass skb to ieee80211_add_aid_request_ie(), remove unused variable 'bss']
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: encode listen interval for S1G
Thomas Pedersen [Tue, 22 Sep 2020 02:28:10 +0000 (19:28 -0700)]
mac80211: encode listen interval for S1G

S1G allows listen interval up to 2^14 * 10000 beacon
intervals. In order to do this listen interval needs a
scaling factor applied to the lower 14 bits. Calculate
this and properly encode the listen interval for S1G STAs.

See IEEE802.11ah-2016 Table 9-44a for reference.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200922022818.15855-10-thomas@adapt-ip.com
[move listen_int_usf into function using it]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: convert S1G beacon to scan results
Thomas Pedersen [Tue, 22 Sep 2020 02:28:06 +0000 (19:28 -0700)]
cfg80211: convert S1G beacon to scan results

The S1G beacon is an extension frame as opposed to
management frame for the regular beacon. This means we may
have to occasionally cast the frame buffer to a different
header type. Luckily this isn't too bad as scan results
mostly only care about the IEs.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200922022818.15855-6-thomas@adapt-ip.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agonl80211: support S1G capability overrides in assoc
Thomas Pedersen [Tue, 22 Sep 2020 02:28:04 +0000 (19:28 -0700)]
nl80211: support S1G capability overrides in assoc

NL80211_ATTR_S1G_CAPABILITY can be passed along with
NL80211_ATTR_S1G_CAPABILITY_MASK to NL80211_CMD_ASSOCIATE
to indicate S1G capabilities which should override the
hardware capabilities in eg. the association request.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200922022818.15855-4-thomas@adapt-ip.com
[johannes: always require both attributes together, commit message]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: redefine S1G bits with GENMASK
Thomas Pedersen [Tue, 8 Sep 2020 19:03:02 +0000 (12:03 -0700)]
ieee80211: redefine S1G bits with GENMASK

The S1G capability fields were defined by ORing BITS()
together, and expecting a custom macro to use the _SHIFT
definitions. Use the Linux kernel GENMASK for the
definitions now, and FIELD_{GET,PREP} to access the fields
in the future.

Take the chance to rename eg. S1G_CAPAB_B0 to the more
compact S1G_CAP0.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200908190323.15814-2-thomas@adapt-ip.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: Handle special status codes in SAE commit
Jouni Malinen [Fri, 31 Jul 2020 18:38:30 +0000 (21:38 +0300)]
mac80211: Handle special status codes in SAE commit

SAE authentication has been extended with H2E (IEEE 802.11 REVmd) and PK
(WFA) options. Those extensions use special status code values in the
SAE commit messages (Authentication frame with transaction sequence
number 1) to identify which extension is in use. mac80211 was
interpreting those new values as the AP denying authentication and that
resulted in failure to complete SAE authentication in some cases.

Fix this by adding exceptions for the new status code values 126 and
127.

Signed-off-by: Jouni Malinen <jouni@codeaurora.org>
Link: https://lore.kernel.org/r/20200731183830.18735-1-jouni@codeaurora.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: Add missing and new AKM suite selector definitions
Veerendranath Jakkam [Wed, 17 Jun 2020 11:31:32 +0000 (17:01 +0530)]
ieee80211: Add missing and new AKM suite selector definitions

Add the definitions for missing AKM selectors defined in
IEEE P802.11-REVmd/D3.0, table 9-151. These definitions will
be used by various drivers that support these new AKM suites.

Signed-off-by: Veerendranath Jakkam <vjakkam@codeaurora.org>
Link: https://lore.kernel.org/r/20200617113132.13477-1-vjakkam@codeaurora.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: check the correct bit for EMA AP
Shaul Triebitz [Thu, 28 May 2020 19:34:37 +0000 (21:34 +0200)]
mac80211: check the correct bit for EMA AP

An AP supporting EMA (Enhanced Multi-BSSID advertisement) should set
bit 83 in the extended capabilities IE (9.4.2.26 in the 802.11ax D5 spec).
So the *3rd* bit of the 10th byte should be checked.
Also, in one place, the wrong byte was checked.
(cfg80211_find_ie returns a pointer to the beginning of the IE,
 so the data really starts at ie[2], so the 10th byte
 should be ie[12]. To avoid this confusion, use cfg80211_find_elem
 instead).

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Link: https://lore.kernel.org/r/20200528213443.4316121fa2a3.I9745582f8d41ad8e689dac0fefcd70b276d7c1ea@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: add HE ext EIDs and 6 GHz capability defines
Johannes Berg [Thu, 28 May 2020 19:34:28 +0000 (21:34 +0200)]
ieee80211: add HE ext EIDs and 6 GHz capability defines

Add the HE extended element IDs and the definitions for the
HE 6 GHz band capabilities element, from Draft 5.0.

Link: https://lore.kernel.org/r/20200528213443.1a6689fe093f.Ifdc5400fb01779351354daf38663ebeea03c9ad9@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: add code to obtain and parse 6 GHz operation field
Johannes Berg [Thu, 28 May 2020 19:34:27 +0000 (21:34 +0200)]
ieee80211: add code to obtain and parse 6 GHz operation field

Add some code to obtain and parse the 6 GHz operation field
inside the HE operation element.

While at it, fix the required length using sizeof() the new
struct, which is 5 instead of 4 now.

Link: https://lore.kernel.org/r/20200528213443.42ca72c45ca9.Id74bc1b03da9ea6574f9bc70deeb60dfc1634359@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: definitions for reduced neighbor reports
Tova Mussai [Thu, 28 May 2020 19:34:26 +0000 (21:34 +0200)]
ieee80211: definitions for reduced neighbor reports

Add the necessary definitions to parse reduced neighbor
report elements.

Signed-off-by: Tova Mussai <tova.mussai@intel.com>
[change struct name, remove IEEE80211_MIN_AP_NEIGHBOR_INFO_SIZE]
Link: https://lore.kernel.org/r/20200528213443.4f9154461c06.I518d9898ad982f838112ea9ca14a20d6bbb16394@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: Replace zero-length array with flexible-array
Gustavo A. R. Silva [Thu, 7 May 2020 18:39:09 +0000 (13:39 -0500)]
cfg80211: Replace zero-length array with flexible-array

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Link: https://lore.kernel.org/r/20200507183909.GA12993@embeddedor
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: Add new AMPDU factor macro for HE peer caps
Tamizh Chelvam [Mon, 4 May 2020 17:04:59 +0000 (22:34 +0530)]
mac80211: Add new AMPDU factor macro for HE peer caps

Add IEEE80211_HE_VHT_MAX_AMPDU_FACTOR and IEEE80211_HE_HT_MAX_AMPDU_FACTOR
as per spec to use for peer max ampdu factor.

Signed-off-by: Tamizh Chelvam <tamizhr@codeaurora.org>
Link: https://lore.kernel.org/r/1588611900-21185-1-git-send-email-tamizhr@codeaurora.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: S1G defines
Thomas Pedersen [Thu, 30 Apr 2020 17:25:53 +0000 (10:25 -0700)]
ieee80211: S1G defines

These are found in IEEE-802.11ah-2016.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200430172554.18383-5-thomas@adapt-ip.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: add KHz variants of frame RX API
Thomas Pedersen [Thu, 30 Apr 2020 17:25:50 +0000 (10:25 -0700)]
cfg80211: add KHz variants of frame RX API

Drivers may wish to report the RX frequency in units of
KHz. Provide cfg80211_rx_mgmt_khz() and wrap it with
cfg80211_rx_mgmt() so exisiting drivers which can't report
KHz anyway don't need to change. Add a similar wrapper for
cfg80211_report_obss_beacon() so the frequency units stay
somewhat consistent.

This doesn't actually change the nl80211 API yet.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200430172554.18383-2-thomas@adapt-ip.com
[fix mac80211 calling the non-khz version of obss beacon report,
 drop trace point name changes]
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: Fix incorrect mask for default PE duration
Pradeep Kumar Chitrapu [Wed, 6 May 2020 10:24:30 +0000 (03:24 -0700)]
ieee80211: Fix incorrect mask for default PE duration

Fixes bitmask for HE opration's default PE duration.

Fixes: daa5b83513a7 ("mac80211: update HE operation fields to D3.0")
Signed-off-by: Pradeep Kumar Chitrapu <pradeepc@codeaurora.org>
Link: https://lore.kernel.org/r/20200506102430.5153-1-pradeepc@codeaurora.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: share 802.11 unit conversion helpers
Thomas Pedersen [Thu, 2 Apr 2020 01:18:02 +0000 (18:18 -0700)]
ieee80211: share 802.11 unit conversion helpers

MHZ_TO_KHZ, and KHZ_TO_MHZ are useful to drivers and
elsewhere so export these in the common ieee80211 header.
Move the power helpers also because we might as well.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200402011810.22947-2-thomas@adapt-ip.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: Parse HE membership selector
Ilan Peer [Thu, 26 Mar 2020 13:09:35 +0000 (15:09 +0200)]
cfg80211: Parse HE membership selector

This extends the support for drivers that rebuilds IEs in the
FW (same as with HT/VHT).

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200326150855.20feaabfb484.I886252639604c8e3e84b8ef97962f1b0e4beec81@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: implement Operating Mode Notification extended NSS support
Johannes Berg [Thu, 26 Mar 2020 13:09:32 +0000 (15:09 +0200)]
mac80211: implement Operating Mode Notification extended NSS support

Somehow we missed this for a long time, but similar to the extended
NSS support in VHT capabilities, we need to have this in Operating
Mode notification.

Implement it by
 * parsing the 160/80+80 bit there and setting the bandwidth
   appropriately
 * having callers of ieee80211_get_vht_max_nss() pass in the current
   max NSS value as received in the operating mode notification in
   order to modify it appropriately depending on the extended NSS
   bits.

This updates all drivers that use it, i.e. only iwlwifi/mvm.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/iwlwifi.20200326150855.098483728cfa.I4e8c25d3288441759c2793247197229f0696a37d@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: fix HE SPR size calculation
Johannes Berg [Wed, 25 Mar 2020 08:09:19 +0000 (09:09 +0100)]
ieee80211: fix HE SPR size calculation

The he_sr_control field is just a u8, so le32_to_cpu()
shouldn't be applied to it; this was evidently copied
from ieee80211_he_oper_size(). Fix it, and also adjust
the type of the local variable.

Fixes: ef11a931bd1c ("mac80211: HE: add Spatial Reuse element parsing support")
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Link: https://lore.kernel.org/r/20200325090918.dfe483b49e06.Ia53622f23b2610a2ae6ea39a199866196fe946c1@changeid
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: add WPA3 OWE AKM suite selector
Sergey Matyukevich [Thu, 13 Feb 2020 13:16:17 +0000 (13:16 +0000)]
ieee80211: add WPA3 OWE AKM suite selector

Add the definition for Opportunistic Wireless Encryption AKM selector.

Signed-off-by: Sergey Matyukevich <sergey.matyukevich.os@quantenna.com>
Link: https://lore.kernel.org/r/20200213131608.10541-3-sergey.matyukevich.os@quantenna.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: parse also the RSNXE IE
Shaul Triebitz [Fri, 31 Jan 2020 11:12:57 +0000 (13:12 +0200)]
mac80211: parse also the RSNXE IE

Parse also the RSN Extension IE when parsing the rest of the IEs.
It will be used in a later patch.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Link: https://lore.kernel.org/r/20200131111300.891737-21-luca@coelho.fi
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: fix 'the' doubling in comments
Zvika Yehudai [Mon, 3 Feb 2020 08:08:23 +0000 (10:08 +0200)]
ieee80211: fix 'the' doubling in comments

Remove redundant 'the' where 'the the' was written.

Signed-off-by: Zvika Yehudai <zvikayeh@gmail.com>
Link: https://lore.kernel.org/r/20200203080823.24949-1-zvikayeh@gmail.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: add ieee80211_is_any_nullfunc()
Thomas Pedersen [Tue, 14 Jan 2020 05:59:40 +0000 (21:59 -0800)]
mac80211: add ieee80211_is_any_nullfunc()

commit 08a5bdde3812 ("mac80211: consider QoS Null frames for STA_NULLFUNC_ACKED")
Fixed a bug where we failed to take into account a
nullfunc frame can be either non-QoS or QoS. It turns out
there is at least one more bug in
ieee80211_sta_tx_notify(), introduced in
commit 7b6ddeaf27ec ("mac80211: use QoS NDP for AP probing"),
where we forgot to check for the QoS variant and so
assumed the QoS nullfunc frame never went out

Fix this by adding a helper ieee80211_is_any_nullfunc()
which consolidates the check for non-QoS and QoS nullfunc
frames. Replace existing compound conditionals and add a
couple more missing checks for QoS variant.

Signed-off-by: Thomas Pedersen <thomas@adapt-ip.com>
Link: https://lore.kernel.org/r/20200114055940.18502-3-thomas@adapt-ip.com
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agotrivial: mac80211: fix indentation
John Crispin [Fri, 6 Dec 2019 14:31:03 +0000 (15:31 +0100)]
trivial: mac80211: fix indentation

Signed-off-by: John Crispin <john@phrozen.org>
Link: https://lore.kernel.org/r/20191206143103.3645-1-john@phrozen.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: add support for the ADDBA extension element
John Crispin [Mon, 29 Jul 2019 10:45:12 +0000 (12:45 +0200)]
mac80211: add support for the ADDBA extension element

HE allows peers to negotiate the aggregation fragmentation level to be used
during transmission. The level can be 1-3. The Ext element is added behind
the ADDBA request inside the action frame. The responder will then reply
with the same level or a lower one if the requested one is not supported.
This patch only handles the negotiation part as the ADDBA frames get passed
to the ATH11k firmware, which does the rest of the magic for us aswell as
generating the requests.

Signed-off-by: Shashidhar Lakkavalli <slakkavalli@datto.com>
Signed-off-by: John Crispin <john@phrozen.org>
Link: https://lore.kernel.org/r/20190729104512.27615-1-john@phrozen.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: fix ieee80211_he_oper_size() comment
John Crispin [Mon, 29 Jul 2019 10:23:41 +0000 (12:23 +0200)]
mac80211: fix ieee80211_he_oper_size() comment

Johannes mentioned that the comment should not reference mac80211 as other
subsystems might call the helper.

Signed-off-by: John Crispin <john@phrozen.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20190729102342.8659-1-john@phrozen.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: HE: add Spatial Reuse element parsing support
John Crispin [Tue, 18 Jun 2019 06:19:14 +0000 (08:19 +0200)]
mac80211: HE: add Spatial Reuse element parsing support

Add support to mac80211 for parsing SPR elements as per
P802.11ax_D4.0 section 9.4.2.241.

Signed-off-by: Shashidhar Lakkavalli <slakkavalli@datto.com>
Signed-off-by: John Crispin <john@phrozen.org>
Link: https://lore.kernel.org/r/20190618061915.7102-2-john@phrozen.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: add support for parsing ADDBA_EXT IEs
John Crispin [Sat, 13 Jul 2019 16:36:41 +0000 (18:36 +0200)]
mac80211: add support for parsing ADDBA_EXT IEs

ADDBA_EXT IEs can be used to negotiate the BA fragmentation level.

Signed-off-by: John Crispin <john@phrozen.org>
Link: https://lore.kernel.org/r/20190713163642.18491-2-john@phrozen.org
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agotreewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500
Thomas Gleixner [Tue, 4 Jun 2019 08:11:33 +0000 (10:11 +0200)]
treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 500

Based on 2 normalized pattern(s):

  this program is free software you can redistribute it and or modify
  it under the terms of the gnu general public license version 2 as
  published by the free software foundation

  this program is free software you can redistribute it and or modify
  it under the terms of the gnu general public license version 2 as
  published by the free software foundation #

extracted by the scancode license scanner the SPDX license identifier

  GPL-2.0-only

has been chosen to replace the boilerplate/reference in 4122 file(s).

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Enrico Weigelt <info@metux.net>
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190604081206.933168790@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: Add a missing extended capability flag definition
Ilan Peer [Wed, 29 May 2019 12:25:33 +0000 (15:25 +0300)]
ieee80211: Add a missing extended capability flag definition

Add the "OBSS Narrow Bandwidth RU In OFDMA Tolerance Support" flag
definition to the definitions of the flags covered by the Extended
Capability IE.

Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agonl80211: add support for SAE authentication offload
Chung-Hsien Hsu [Thu, 9 May 2019 09:49:06 +0000 (09:49 +0000)]
nl80211: add support for SAE authentication offload

Let drivers advertise support for station-mode SAE authentication
offload with a new NL80211_EXT_FEATURE_SAE_OFFLOAD flag.

Signed-off-by: Chung-Hsien Hsu <stanley.hsu@cypress.com>
Signed-off-by: Chi-Hsien Lin <chi-hsien.lin@cypress.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: update HE IEs to D4.0 spec
Liad Kaufman [Fri, 15 Mar 2019 15:39:07 +0000 (17:39 +0200)]
ieee80211: update HE IEs to D4.0 spec

Update the out-dated comments as well, and have them point to
the correct sections in the D4.0 spec.

Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: support non-inheritance element
Sara Sharon [Fri, 15 Mar 2019 15:39:03 +0000 (17:39 +0200)]
cfg80211: support non-inheritance element

Subelement profile may specify element IDs it doesn't inherit
from the management frame. Support it.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: update HE IEs to D3.3
Liad Kaufman [Wed, 6 Feb 2019 11:17:20 +0000 (13:17 +0200)]
mac80211: update HE IEs to D3.3

Update element names and new fields according to D3.3 of
the HE spec.

Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: support max channel switch time element
Sara Sharon [Wed, 6 Feb 2019 11:17:09 +0000 (13:17 +0200)]
mac80211: support max channel switch time element

2018 REVmd of the spec introduces the max channel switch time
element which is optionally included in beacons/probes when there
is a channel switch / extended channel switch element.
The value represents the maximum delay between the time the AP
transmitted the last beacon in current channel and the expected
time of the first beacon in the new channel, in TU.

Parse the value and pass it to the driver.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoiwlwifi: mvm: support FTM initiator
Johannes Berg [Wed, 5 Dec 2018 10:33:34 +0000 (11:33 +0100)]
iwlwifi: mvm: support FTM initiator

Add support for FTM initiator, i.e. peer measurements with FTM
if the firmware supports FTM.

Additionally, add two defines we depend on in
include/linux/ieee80211.h.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Avraham Stern <avraham.stern@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: Use const more consistently in for_each_element macros
Jouni Malinen [Mon, 11 Feb 2019 14:29:04 +0000 (16:29 +0200)]
cfg80211: Use const more consistently in for_each_element macros

Enforce the first argument to be a correct type of a pointer to struct
element and avoid unnecessary typecasts from const to non-const pointers
(the change in validate_ie_attr() is needed to make this part work). In
addition, avoid signed/unsigned comparison within for_each_element() and
mark struct element packed just in case.

Signed-off-by: Jouni Malinen <j@w1.fi>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: fix for_each_element_extid()
Johannes Berg [Fri, 8 Feb 2019 16:56:33 +0000 (17:56 +0100)]
ieee80211: fix for_each_element_extid()

The data/datalen argument names cannot be used as those
are also the struct element names, fix that.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: indicate support for multiple BSSID
Sara Sharon [Wed, 16 Jan 2019 21:03:25 +0000 (23:03 +0200)]
mac80211: indicate support for multiple BSSID

Set multi-bssid support flags according to driver support.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: support multi-bssid
Sara Sharon [Wed, 16 Jan 2019 16:22:56 +0000 (18:22 +0200)]
mac80211: support multi-bssid

Add support for multi-bssid.

This includes:
- Parsing multi-bssid element
- Overriding DTIM values
- Taking into account in various places the inner BSSID instead of
  transmitter BSSID
- Save aside some multi-bssid properties needed by drivers

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: add and use strongly typed element iteration macros
Johannes Berg [Thu, 7 Feb 2019 20:44:41 +0000 (21:44 +0100)]
cfg80211: add and use strongly typed element iteration macros

Rather than always iterating elements from frames with pure
u8 pointers, add a type "struct element" that encapsulates
the id/datalen/data format of them.

Then, add the element iteration macros
 * for_each_element
 * for_each_element_id
 * for_each_element_extid

which take, as their first 'argument', such a structure and
iterate through a given u8 array interpreting it as elements.

While at it and since we'll need it, also add
 * for_each_subelement
 * for_each_subelement_id
 * for_each_subelement_extid

which instead of taking data/length just take an outer element
and use its data/datalen.

Also add for_each_element_completed() to determine if any of
the loops above completed, i.e. it was able to parse all of
the elements successfully and no data remained.

Use for_each_element_id() in cfg80211_find_ie_match() as the
first user of this.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: update HE operation fields to D3.0
Shaul Triebitz [Sat, 15 Dec 2018 09:03:05 +0000 (11:03 +0200)]
mac80211: update HE operation fields to D3.0

HE Operation element has changed in 11ax D3.0.  Update the fields
accordingly.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: add bits for TWT in Extended Capabilities IE
Emmanuel Grumbach [Sat, 15 Dec 2018 09:03:03 +0000 (11:03 +0200)]
ieee80211: add bits for TWT in Extended Capabilities IE

These bits are defined in ieee802.11ax to advertise support
for TWT in addition to the bits in the HE IE.

Signed-off-by: Emmanuel Grumbach <emmanuel.grumbach@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months ago{nl,mac}80211: report gate connectivity in station info
Bob Copeland [Thu, 25 Oct 2018 19:48:53 +0000 (15:48 -0400)]
{nl,mac}80211: report gate connectivity in station info

Capture the current state of gate connectivity from the mesh
formation field in mesh config whenever we receive a beacon,
and report that via GET_STATION.  This allows applications
doing mesh peering in userspace to make peering decisions
based on peers' current upstream connectivity.

Signed-off-by: Bob Copeland <bobcopeland@fb.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: minstrel: Enable STBC and LDPC for VHT Rates
Chaitanya T K [Sat, 6 Oct 2018 17:34:59 +0000 (19:34 +0200)]
mac80211: minstrel: Enable STBC and LDPC for VHT Rates

If peer support reception of STBC and LDPC, enable them for better
performance.

Signed-off-by: Chaitanya TK <chaitanya.mgit@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agowireless: align to draft 11ax D3.0
Shaul Triebitz [Wed, 5 Sep 2018 05:06:08 +0000 (08:06 +0300)]
wireless: align to draft 11ax D3.0

Align to new 11ax draft D3.0.  Change/add new MAC and PHY capabilities
and update drivers' 11ax capabilities and mac80211's debugfs
accordingly.

Signed-off-by: Shaul Triebitz <shaul.triebitz@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: add new VHT capability fields/parsing
Johannes Berg [Fri, 31 Aug 2018 08:31:16 +0000 (11:31 +0300)]
ieee80211: add new VHT capability fields/parsing

IEEE 802.11-2016 extended the VHT capability fields to allow
indicating the number of spatial streams depending on the
actually used bandwidth, add support for decoding this.

Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: remove redundant leading zeroes
Sara Sharon [Fri, 31 Aug 2018 08:31:09 +0000 (11:31 +0300)]
ieee80211: remove redundant leading zeroes

The defines of IEEE80211_HE_OPERATION_VHT_OPER_INFO and
IEEE80211_HE_OPERATION_MULTI_BSSID_AP have leading zeroes
that makes the number look like it is bigger than 32 bit.
This is misleading, remove it.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: bump IEEE80211_MAX_AMPDU_BUF to support HE
Johannes Berg [Mon, 18 Jun 2018 20:39:29 +0000 (22:39 +0200)]
ieee80211: bump IEEE80211_MAX_AMPDU_BUF to support HE

Bump the IEEE80211_MAX_AMPDU_BUF size to 0x100 for HE support
and - for now - use IEEE80211_MAX_AMPDU_BUF_HT everywhere.

This is derived from my internal patch, parts of which Luca
had sent upstream.

Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocfg80211: Add support for HE
Luca Coelho [Sat, 9 Jun 2018 06:14:42 +0000 (09:14 +0300)]
cfg80211: Add support for HE

Add support for the HE in cfg80211 and also add userspace API to
nl80211 to send rate information out, conforming with P802.11ax_D2.0.

Signed-off-by: Liad Kaufman <liad.kaufman@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Ilan Peer <ilan.peer@intel.com>
Signed-off-by: Ido Yariv <idox.yariv@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes@sipsolutions.net>
[added missing DIV_ROUND_UP + hweight from the linux kernel]
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agomac80211: add get TID helper
Sara Sharon [Mon, 19 Feb 2018 12:48:40 +0000 (14:48 +0200)]
mac80211: add get TID helper

Extracting the TID from the QOS header is common enough
to justify helper.

Signed-off-by: Sara Sharon <sara.sharon@intel.com>
Signed-off-by: Luca Coelho <luciano.coelho@intel.com>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agoieee80211: Increase PMK maximum length to 64 bytes
Srinivas Dasari [Tue, 6 Feb 2018 14:19:35 +0000 (19:49 +0530)]
ieee80211: Increase PMK maximum length to 64 bytes

Increase the PMK maximum length to 64 bytes to accommodate
the key length used in DPP with the NIST P-521 and
Brainpool 512 curves.

Signed-off-by: Srinivas Dasari <dasaris@codeaurora.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170 firmware: refresh ch9.h
Christian Lamparter [Fri, 5 Feb 2021 10:38:29 +0000 (11:38 +0100)]
carl9170 firmware: refresh ch9.h

This patch refreshes the upstream linux ch9.h header definitions.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170 firmware: prepare header update
Christian Lamparter [Fri, 5 Feb 2021 09:24:38 +0000 (10:24 +0100)]
carl9170 firmware: prepare header update

This patch brings the shared firmware headers in line with the carl9170
linux kernel driver.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agotreewide: Replace GPLv2 boilerplate/reference with SPDX - rule 325
Thomas Gleixner [Wed, 29 May 2019 23:57:40 +0000 (16:57 -0700)]
treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 325

Based on 1 normalized pattern(s):

  this program is free software you can redistribute it and or modify
  it under the terms of the gnu general public license as published by
  the free software foundation either version 2 of the license this
  program is distributed in the hope that it will be useful but
  without any warranty without even the implied warranty of
  merchantability or fitness for a particular purpose see the gnu
  general public license for more details you should have received a
  copy of the gnu general public license along with this program see
  the file copying if not see http www gnu org licenses

extracted by the scancode license scanner the SPDX license identifier

  GPL-2.0-only

has been chosen to replace the boilerplate/reference in 1 file(s).

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Alexios Zavras <alexios.zavras@intel.com>
Reviewed-by: Armijn Hemel <armijn@tjaldur.nl>
Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org>
Reviewed-by: Allison Randal <allison@lohutok.net>
Cc: linux-spdx@vger.kernel.org
Link: https://lkml.kernel.org/r/20190530000435.742096485@linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170: fix struct alignment conflict
Arnd Bergmann [Thu, 4 Feb 2021 16:29:17 +0000 (17:29 +0100)]
carl9170: fix struct alignment conflict

Multiple structures in the carl9170 driver have alignment
impossible alignment constraints that gcc warns about when
building with 'make W=1':

include/shared/fwcmd.h:243:2: warning: alignment 1 of 'union <anonymous>' is less than 4 [-Wpacked-not-aligned]
include/shared/wlan.h:373:1: warning: alignment 1 of 'struct ar9170_rx_frame_single' is less than 2 [-Wpacked-not-aligned]

In the carl9170_cmd structure, multiple members that have an explicit
alignment requirement of four bytes are added into a union with explicit
byte alignment, but this in turn is part of a structure that also has
four-byte alignment.

In the wlan.h header, multiple structures contain a ieee80211_hdr member
that is required to be two-byte aligned to avoid alignmnet faults when
processing network headers, but all members are forced to be byte-aligned
using the __packed tag at the end of the struct definition.

In both cases, leaving out the packing does not change the internal
layout of the structure but changes the alignment constraint of the
structure itself.

Change all affected structures to only apply packing where it does
not violate the alignment requirement of the contained structure.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
4 months agocarl9170: Replace zero-length array with flexible-array
Gustavo A. R. Silva [Thu, 7 May 2020 15:19:21 +0000 (10:19 -0500)]
carl9170: Replace zero-length array with flexible-array

The current codebase makes use of the zero-length array language
extension to the C90 standard, but the preferred mechanism to declare
variable-length types such as these ones is a flexible array member[1][2],
introduced in C99:

struct foo {
        int stuff;
        struct boo array[];
};

By making use of the mechanism above, we will get a compiler warning
in case the flexible array does not occur last in the structure, which
will help us prevent some kind of undefined behavior bugs from being
inadvertently introduced[3] to the codebase from now on.

Also, notice that, dynamic memory allocations won't be affected by
this change:

"Flexible array members have incomplete type, and so the sizeof operator
may not be applied. As a quirk of the original implementation of
zero-length arrays, sizeof evaluates to zero."[1]

sizeof(flexible-array-member) triggers a warning because flexible array
members have incomplete type[1]. There are some instances of code in
which the sizeof operator is being incorrectly/erroneously applied to
zero-length arrays and the result is zero. Such instances may be hiding
some bugs. So, this work (flexible-array member conversions) will also
help to get completely rid of those sorts of issues.

This issue was found with the help of Coccinelle.

[1] https://gcc.gnu.org/onlinedocs/gcc/Zero-Length.html
[2] https://github.com/KSPP/linux/issues/21
[3] commit 76497732932f ("cxgb3/l2t: Fix undefined behaviour")

Signed-off-by: Gustavo A. R. Silva <gustavoars@kernel.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Link: https://lore.kernel.org/r/20200507151921.GA5083@embeddedor
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
10 months agocarl9170 toolchain: update to gcc 10.2.0
Christian Lamparter [Fri, 24 Jul 2020 22:47:12 +0000 (00:47 +0200)]
carl9170 toolchain: update to gcc 10.2.0

This patch also removes some of SHA256SUMS' old entries.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
13 months agocarl9170 toolchain: update to gcc 10.1.0
Christian Lamparter [Sat, 9 May 2020 20:33:20 +0000 (22:33 +0200)]
carl9170 toolchain: update to gcc 10.1.0

This patch also removes some of SHA256SUMS' old entries.

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
14 months agocarl9170 toolchain: update to gcc 9.3.0
Christian Lamparter [Fri, 20 Mar 2020 19:27:34 +0000 (20:27 +0100)]
carl9170 toolchain: update to gcc 9.3.0

Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
16 months agokconfig: don't crash on NULL expressions in expr_eq()
Thomas Hebb [Mon, 9 Dec 2019 08:19:17 +0000 (00:19 -0800)]
kconfig: don't crash on NULL expressions in expr_eq()

NULL expressions are taken to always be true, as implemented by the
expr_is_yes() macro and by several other functions in expr.c. As such,
they ought to be valid inputs to expr_eq(), which compares two
expressions.

Signed-off-by: Thomas Hebb <tommyhebb@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
16 months agokconfig: Add option to get the full help text with listnewconfig
Laura Abbott [Mon, 4 Nov 2019 22:10:08 +0000 (17:10 -0500)]
kconfig: Add option to get the full help text with listnewconfig

make listnewconfig will list the individual options that need to be set.
This is useful but there's no easy way to get the help text associated
with the options at the same time. Introduce a new targe
'make helpnewconfig' which lists the full help text of all the
new options as well. This makes it easier to automatically generate
changes that are easy for humans to review. This command also adds
markers between each option for easier parsing.

Signed-off-by: Laura Abbott <labbott@redhat.com>
Acked-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>