Enable the machine function split optimization for AutoFDO in Clang.

Machine function split (MFS) is a pass in the Clang compiler that
splits a function into hot and cold parts. The linker groups all
cold blocks across functions together. This decreases hot code
fragmentation and improves iCache and iTLB utilization.

MFS requires a profile so this is enabled only for the AutoFDO builds.

Co-developed-by: Han Shen <shenhan@google.com>
Signed-off-by: Han Shen <shenhan@google.com>
Signed-off-by: Rong Xu <xur@google.com>
Suggested-by: Sriraman Tallam <tmsriram@google.com>
Suggested-by: Krzysztof Pszeniczny <kpszeniczny@google.com>
Tested-by: Yonghong Song <yonghong.song@linux.dev>
Tested-by: Yabin Cui <yabinc@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Enable -ffunction-sections by default for the AutoFDO build.

With -ffunction-sections, the compiler places each function in its own
section named .text.function_name instead of placing all functions in
the .text section. In the AutoFDO build, this allows the linker to
utilize profile information to reorganize functions for improved
utilization of iCache and iTLB.

Co-developed-by: Han Shen <shenhan@google.com>
Signed-off-by: Han Shen <shenhan@google.com>
Signed-off-by: Rong Xu <xur@google.com>
Suggested-by: Sriraman Tallam <tmsriram@google.com>
Tested-by: Yonghong Song <yonghong.song@linux.dev>
Tested-by: Yabin Cui <yabinc@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Add markers like __hot_text_start, __hot_text_end, __unlikely_text_start,
and __unlikely_text_end which will be included in System.map. These markers
indicate how the compiler groups functions, providing valuable information
to developers about the layout and optimization of the code.

Co-developed-by: Han Shen <shenhan@google.com>
Signed-off-by: Han Shen <shenhan@google.com>
Signed-off-by: Rong Xu <xur@google.com>
Suggested-by: Sriraman Tallam <tmsriram@google.com>
Tested-by: Yonghong Song <yonghong.song@linux.dev>
Tested-by: Yabin Cui <yabinc@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

When the -ffunction-sections compiler option is enabled, each function
is placed in a separate section named .text.function_name rather than
putting all functions in a single .text section.

However, using -function-sections can cause problems with the
linker script. The comments included in include/asm-generic/vmlinux.lds.h
note these issues.:
  “TEXT_MAIN here will match .text.fixup and .text.unlikely if dead
   code elimination is enabled, so these sections should be converted
   to use ".." first.”

It is unclear whether there is a straightforward method for converting
a suffix to "..".

This patch modifies the order of subsections within the text output
section. Specifically, it changes current order:
  .text.hot, .text, .text_unlikely, .text.unknown, .text.asan
to the new order:
  .text.asan, .text.unknown, .text_unlikely, .text.hot, .text

Here is the rationale behind the new layout:

The majority of the code resides in three sections: .text.hot, .text,
and .text.unlikely, with .text.unknown containing a negligible amount.
.text.asan is only generated in ASAN builds.

The primary goal is to group code segments based on their execution
frequency (hotness).

First, we want to place .text.hot adjacent to .text. Since we cannot put
.text.hot after .text (Due to constraints with -ffunction-sections,
placing .text.hot after .text is problematic), we need to put
.text.hot before .text.

Then it comes to .text.unlikely, we cannot put it after .text (same
-ffunction-sections issue) . Therefore, we position .text.unlikely
before .text.hot.

.text.unknown and .tex.asan follow the same logic.

This revised ordering effectively reverses the original arrangement (for
.text.unlikely, .text.unknown, and .tex.asan), maintaining a similar level
of affinity between sections.

It also places .text.hot section at the beginning of a page to better
utilize the TLB entry.

Note that the limitation arises because the linker script employs glob
patterns instead of regular expressions for string matching. While there
is a method to maintain the current order using complex patterns, this
significantly complicates the pattern and increases the likelihood of
errors.

This patch also changes vmlinux.lds.S for the sparc64 architecture to
accommodate specific symbol placement requirements.

Co-developed-by: Han Shen <shenhan@google.com>
Signed-off-by: Han Shen <shenhan@google.com>
Signed-off-by: Rong Xu <xur@google.com>
Suggested-by: Sriraman Tallam <tmsriram@google.com>
Suggested-by: Krzysztof Pszeniczny <kpszeniczny@google.com>
Tested-by: Yonghong Song <yonghong.song@linux.dev>
Tested-by: Yabin Cui <yabinc@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Mark __kernel_entry as ".head.text" and place HEAD_TEXT before
TEXT_TEXT in the linker script. This ensures that __kernel_entry
will be placed at the beginning of text section.

Drop mips from scripts/head-object-list.txt.

Signed-off-by: Rong Xu <xur@google.com>
Reported-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Closes: https://lore.kernel.org/lkml/c6719149-8531-4174-824e-a3caf4bc6d0e@alliedtelesis.co.nz/T/


Tested-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

In the presence of both weak and strong function definitions, the
linker drops the weak symbol in favor of a strong symbol, but
leaves the code in place. Code in ignore_unreachable_insn() has
some heuristics to suppress the warning, but it does not work when
-ffunction-sections is enabled.

Suppose function foo has both strong and weak definitions.
Case 1: The strong definition has an annotated section name,
like .init.text. Only the weak definition will be placed into
.text.foo. But since the section has no symbols, there will be no
"hole" in the section.

Case 2: Both sections are without an annotated section name.
Both will be placed into .text.foo section, but there will be only one
symbol (the strong one). If the weak code is before the strong code,
there is no "hole" as it fails to find the right-most symbol before
the offset.

The fix is to use the first node to compute the hole if hole.sym
is empty. If there is no symbol in the section, the first node
will be NULL, in which case, -1 is returned to skip the whole
section.

Co-developed-by: Han Shen <shenhan@google.com>
Signed-off-by: Han Shen <shenhan@google.com>
Signed-off-by: Rong Xu <xur@google.com>
Suggested-by: Sriraman Tallam <tmsriram@google.com>
Suggested-by: Krzysztof Pszeniczny <kpszeniczny@google.com>
Tested-by: Yonghong Song <yonghong.song@linux.dev>
Tested-by: Yabin Cui <yabinc@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Acked-by: Josh Poimboeuf <jpoimboe@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Add the build support for using Clang's AutoFDO. Building the kernel
with AutoFDO does not reduce the optimization level from the
compiler. AutoFDO uses hardware sampling to gather information about
the frequency of execution of different code paths within a binary.
This information is then used to guide the compiler's optimization
decisions, resulting in a more efficient binary. Experiments
showed that the kernel can improve up to 10% in latency.

The support requires a Clang compiler after LLVM 17. This submission
is limited to x86 platforms that support PMU features like LBR on
Intel machines and AMD Zen3 BRS. Support for SPE on ARM 1,
 and BRBE on ARM 1 is part of planned future work.

Here is an example workflow for AutoFDO kernel:

1) Build the kernel on the host machine with LLVM enabled, for example,
       $ make menuconfig LLVM=1
    Turn on AutoFDO build config:
      CONFIG_AUTOFDO_CLANG=y
    With a configuration that has LLVM enabled, use the following
    command:
       scripts/config -e AUTOFDO_CLANG
    After getting the config, build with
      $ make LLVM=1

2) Install the kernel on the test machine.

3) Run the load tests. The '-c' option in perf specifies the sample
   event period. We suggest     using a suitable prime number,
   like 500009, for this purpose.
   For Intel platforms:
      $ perf record -e BR_INST_RETIRED.NEAR_TAKEN:k -a -N -b -c <count> \
        -o <perf_file> -- <loadtest>
   For AMD platforms:
      The supported system are: Zen3 with BRS, or Zen4 with amd_lbr_v2
     For Zen3:
      $ cat proc/cpuinfo | grep " brs"
      For Zen4:
      $ cat proc/cpuinfo | grep amd_lbr_v2
      $ perf record --pfm-events RETIRED_TAKEN_BRANCH_INSTRUCTIONS:k -a \
        -N -b -c <count> -o <perf_file> -- <loadtest>

4) (Optional) Download the raw perf file to the host machine.

5) To generate an AutoFDO profile, two offline tools are available:
   create_llvm_prof and llvm_profgen. The create_llvm_prof tool is part
   of the AutoFDO project and can be found on GitHub
   (https://github.com/google/autofdo

), version v0.30.1 or later. The
   llvm_profgen tool is included in the LLVM compiler itself. It's
   important to note that the version of llvm_profgen doesn't need to
   match the version of Clang. It needs to be the LLVM 19 release or
   later, or from the LLVM trunk.
      $ llvm-profgen --kernel --binary=<vmlinux> --perfdata=<perf_file> \
        -o <profile_file>
   or
      $ create_llvm_prof --binary=<vmlinux> --profile=<perf_file> \
        --format=extbinary --out=<profile_file>

   Note that multiple AutoFDO profile files can be merged into one via:
      $ llvm-profdata merge -o <profile_file>  <profile_1> ... <profile_n>

6) Rebuild the kernel using the AutoFDO profile file with the same config
   as step 1, (Note CONFIG_AUTOFDO_CLANG needs to be enabled):
      $ make LLVM=1 CLANG_AUTOFDO_PROFILE=<profile_file>

Co-developed-by: Han Shen <shenhan@google.com>
Signed-off-by: Han Shen <shenhan@google.com>
Signed-off-by: Rong Xu <xur@google.com>
Suggested-by: Sriraman Tallam <tmsriram@google.com>
Suggested-by: Krzysztof Pszeniczny <kpszeniczny@google.com>
Suggested-by: Nick Desaulniers <ndesaulniers@google.com>
Suggested-by: Stephane Eranian <eranian@google.com>
Tested-by: Yonghong Song <yonghong.song@linux.dev>
Tested-by: Yabin Cui <yabinc@google.com>
Tested-by: Nathan Chancellor <nathan@kernel.org>
Reviewed-by: Kees Cook <kees@kernel.org>
Tested-by: Peter Jung <ptr1337@cachyos.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

There is no need to prune the rust/alloc directory because it was
removed by commit 9d0441ba ("rust: alloc: remove our fork of the
`alloc` crate").

There is no need to prune the rust/test directory because no '*.rs'
files are generated within it.

To avoid forking the 'grep -Fv generated' process, filter out generated
files using the option, ! -name '*generated*'.

Now that the '-path ... -prune' option is no longer used, there is no
need to use the absolute path. Searching in $(srctree), which can be
a relative path, is sufficient.

The comment mentions the use case where $(srctree) is '..', that is,
$(objtree) is a sub-directory of $(srctree). In this scenario, all
'*.rs' files under $(objtree) are generated files and filters out by
the '*generated*' pattern.

Add $(RCS_FIND_IGNORE) as a shortcut. Although I do not believe '*.rs'
files would exist under the .git directory, there is no need for the
'find' command to traverse it.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicolas Schier <n.schier@avm.de>
Acked-by: Miguel Ojeda <ojeda@kernel.org>

The positional argument specifies the top-level Kconfig. Include this
information in the help message.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The P_MENU entries ("menu" and "menuconfig") are never displayed in
symbolMode.

The condition, list->mode == symbolMode, is never met here.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Most of the code in ConfigInfoView::clicked() is unnecessary.
There is no need to use the regular expression to search for a symbol.
Calling sym_find() is simpler and faster.

The hyperlink always begins with the "s" tag, and there is no other
tag used. Remove it.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Split out the code that retrieves the menu entry with a prompt, so it
can be reused in other functions.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The only functional tag is href="s<symbol_name>".

Commit c4f7398b ("kconfig: qconf: make debug links work again")
changed prop->name to sym->name for this reference, but it missed to
change the tag "m" to "s".

This tag is not functional at all.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The same check is performed in the configList->setParentMenu() call.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

When the ESC key is pressed, the parentSelected() signal is currently
emitted for singleMode, menuMode, and symbolMode.

However, parentSelected() signal is functional only for singleMode.

In menuMode, the signal is connected to the goBack() slot, but nothing
occurs because configList->rootEntry is always &rootmenu.

In symbolMode (in the right pane), the parentSelected() signal is not
connected to any slot.

This commit prevents the unnecessary emission of the parentSelected()
signal.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The " (NEW)" string should be displayed regardless of the visibility
of the associated menu.

The ConfigItem::visible member is not used for any other purpose.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

When a menu is selected in the split view, the right pane displays the
goParent button, but it is never functional.

This is unnecessary, as you can select a menu from the menu tree in the
left pane.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Commit a2574c12 ("kconfig: qconf: convert to Qt5 new signal/slot
connection syntax") converted most of the old string-based connections,
but one more instance still remains. Convert it to the new style.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The default value of the quitOnLastWindowClosed property is true.

Hence, the application implicitly quits when the last window is closed.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The default value of the rootIsDecorated property is true.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Since commit fde19251 ("kconfig: remove tristate choice support"),
choice members are always boolean. The type check is redundant.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

These functions simply passes the event to the parent.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Replace the hand crafted lookup table with a QHash. This has the nice benefit
that the added offsets can not get out of sync with the length of the
replacement strings.

Signed-off-by: Rolf Eike Beer <eb@emlix.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

This renames "Load" to "Open" and switches Ctrl-L to Ctrl-O for the default
platforms. This may break the workflow for those used to it, but will make it
actually work for everyone else like me who would just expect the default
behavior. Add some more standard shortcuts where available. If they replace
the existing shortcuts they would have the same value in my case.

Signed-off-by: Rolf Eike Beer <eb@emlix.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

This is the native type used by the file dialogs and avoids any hassle with
filename encoding when converting this back and forth to a character array.

Signed-off-by: Rolf Eike Beer <eb@emlix.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

s/handles/handled/

Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Ignore process select warnings for config entries that have a default
option. Some config entries have no prompt, and nothing selects them, but
these config options are okay because they have a default option.

Signed-off-by: David Hunter <david.hunter.linux@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Track default options on the second line. On the second line of some
config entries, default and dependency options sometimes appear. In those
instances, the state will be "NEW" and not "DEP".

Signed-off-by: David Hunter <david.hunter.linux@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Put in the dollar sign for the variable '$config'. That way, the debug
message has more meaning.

Signed-off-by: David Hunter <david.hunter.linux@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Add the ability to cycle through dialog buttons with the TAB key.

Signed-off-by: Thorsten Blum <thorsten.blum@linux.dev>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The ->addWidget() method re-parents the widget. The parent QWidget can
be specified directly in the constructor.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Rearrange the code to make the upcoming refactoring easier to understand.

No functional changes intended.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

The orientation of the QSplitter can be specified directly in its
constructor.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Turn all warnings during parsing into hard errors.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Since commit fde19251 ("kconfig: remove tristate choice support"),
all choice blocks are now boolean. There is no longer a need to specify
the choice type explicitly.

All "bool" prompts in choice entries have been converted to "prompt".

This commit removes support for the "bool" syntax in choice entries.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Since commit fde19251 ("kconfig: remove tristate choice support"),
all choice blocks are now boolean. There is no longer a need to specify
the choice type explicitly.

Most choice blocks already use "prompt". Before the next commit removes
support for the "bool" syntax in choice entries, this commit converts
the remaining "bool" occurences under drivers/usb/.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Some architectures embed boot DTBs in vmlinux. A potential issue for
these architectures is a race condition during parallel builds because
Kbuild descends into arch/*/boot/dts/ twice.

One build thread is initiated by the 'dtbs' target, which is a
prerequisite of the 'all' target in the top-level Makefile:

  ifdef CONFIG_OF_EARLY_FLATTREE
  all: dtbs
  endif

For architectures that support the built-in boot dtb, arch/*/boot/dts/
is visited also during the ordinary directory traversal in order to
build obj-y objects that wrap DTBs.

Since these build threads are unaware of each other, they can run
simultaneously during parallel builds.

This commit introduces a generic build rule to scripts/Makefile.vmlinux
to support embedded boot DTBs in a race-free way. Architectures that
want to use this rule need to select CONFIG_GENERIC_BUILTIN_DTB.

After the migration, Makefiles under arch/*/boot/dts/ will be visited
only once to build only *.dtb files.

This change also aims to unify the CONFIG options used for built-in DTBs
support. Currently, different architectures use different CONFIG options
for the same purposes.

With this commit, the CONFIG options will be unified as follows:

 - CONFIG_GENERIC_BUILTIN_DTB

   This enables the generic rule for built-in boot DTBs. This will be
   renamed to CONFIG_BUILTIN_DTB after all architectures migrate to the
   generic rule.

 - CONFIG_BUILTIN_DTB_NAME

   This specifies the path to the embedded DTB.
   (relative to arch/*/boot/dts/)

 - CONFIG_BUILTIN_DTB_ALL

   If this is enabled, all DTB files compiled under arch/*/boot/dts/ are
   embedded into vmlinux. Only used by MIPS.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>

Since commit 2f7ab126 ("Kbuild: add Rust support"), Kconfig
generates include/generated/rustc_cfg, but its presence is not checked
in the top-level Makefile. It should be checked similarly to the C
header counterpart, include/generated/autoconf.h.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicolas Schier <n.schier@avm.de>
Acked-by: Miguel Ojeda <ojeda@kernel.org>
Tested-by: Miguel Ojeda <ojeda@kernel.org>

This commit refactors the check for missing configuration files, making
it easier to add more files to the list.

The format of the error message has been slightly changed, as follows:

[Before]

    ERROR: Kernel configuration is invalid.
           include/generated/autoconf.h or include/config/auto.conf are missing.
           Run 'make oldconfig && make prepare' on kernel src to fix it.

[After]

  ***
  ***  ERROR: Kernel configuration is invalid. The following files are missing:
  ***    - include/generated/autoconf.h
  ***    - include/config/auto.conf
  ***  Run "make oldconfig && make prepare" on kernel source to fix it.
  ***

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicolas Schier <n.schier@avm.de>
Tested-by: Miguel Ojeda <ojeda@kernel.org>

Commit e68a558f ("speakup: Fix building as extmod") was intended
to support building this as an external module.

Since commit b1992c37 ("kbuild: use $(src) instead of
$(srctree)/$(src) for source directory"), $(src) consistently points
to the source directory, regardless of whether it is compiled as an
external module or not.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>