Commit Graph

13 Commits

Author SHA1 Message Date
Masahiro Yamada 87c7ee67de scripts: handle BrokenPipeError for python scripts
In the follow-up of commit fb3041d61f ("kbuild: fix SIGPIPE error
message for AR=gcc-ar and AR=llvm-ar"), Kees Cook pointed out that
tools should _not_ catch their own SIGPIPEs [1] [2].

Based on his feedback, LLVM was fixed [3].

However, Python's default behavior is to show noisy bracktrace when
SIGPIPE is sent. So, scripts written in Python are basically in the
same situation as the buggy llvm tools.

Example:

  $ make -s allnoconfig
  $ make -s allmodconfig
  $ scripts/diffconfig .config.old .config | head -n1
  -ALIX n
  Traceback (most recent call last):
    File "/home/masahiro/linux/scripts/diffconfig", line 132, in <module>
      main()
    File "/home/masahiro/linux/scripts/diffconfig", line 130, in main
      print_config("+", config, None, b[config])
    File "/home/masahiro/linux/scripts/diffconfig", line 64, in print_config
      print("+%s %s" % (config, new_value))
  BrokenPipeError: [Errno 32] Broken pipe

Python documentation [4] notes how to make scripts die immediately and
silently:

  """
  Piping output of your program to tools like head(1) will cause a
  SIGPIPE signal to be sent to your process when the receiver of its
  standard output closes early. This results in an exception like
  BrokenPipeError: [Errno 32] Broken pipe. To handle this case,
  wrap your entry point to catch this exception as follows:

    import os
    import sys

    def main():
        try:
            # simulate large output (your code replaces this loop)
            for x in range(10000):
                print("y")
            # flush output here to force SIGPIPE to be triggered
            # while inside this try block.
            sys.stdout.flush()
        except BrokenPipeError:
            # Python flushes standard streams on exit; redirect remaining output
            # to devnull to avoid another BrokenPipeError at shutdown
            devnull = os.open(os.devnull, os.O_WRONLY)
            os.dup2(devnull, sys.stdout.fileno())
            sys.exit(1)  # Python exits with error code 1 on EPIPE

    if __name__ == '__main__':
        main()

  Do not set SIGPIPE’s disposition to SIG_DFL in order to avoid
  BrokenPipeError. Doing that would cause your program to exit
  unexpectedly whenever any socket connection is interrupted while
  your program is still writing to it.
  """

Currently, tools/perf/scripts/python/intel-pt-events.py seems to be the
only script that fixes the issue that way.

tools/perf/scripts/python/compaction-times.py uses another approach
signal.signal(signal.SIGPIPE, signal.SIG_DFL) but the Python
documentation clearly says "Don't do it".

I cannot fix all Python scripts since there are so many.
I fixed some in the scripts/ directory.

[1]: https://lore.kernel.org/all/202211161056.1B9611A@keescook/
[2]: https://github.com/llvm/llvm-project/issues/59037
[3]: 4787efa380
[4]: https://docs.python.org/3/library/signal.html#note-on-sigpipe

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>
2023-01-26 12:43:33 +09:00
Masahiro Yamada 92215e7a80 kbuild: rename cmd_$@ to savedcmd_$@ in *.cmd files
The cmd-check macro compares $(cmd_$@) and $(cmd_$1), but a pitfall is
that you cannot use cmd_<target> as the variable name for the command.

For example, the following code will not work in the top Makefile
or ./Kbuild.

    quiet_cmd_foo = GEN     $@
          cmd_foo = touch $@

    targets += foo
    foo: FORCE
            $(call if_changed,foo)

In this case, both $@ and $1 are expanded to 'foo', so $(cmd_check)
is always empty.

We do not need to use the same prefix for cmd_$@ and cmd_$1.
Rename the former to savedcmd_$@.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>
2023-01-22 23:43:33 +09:00
Masahiro Yamada f65a486821 kbuild: change module.order to list *.o instead of *.ko
scripts/Makefile.build replaces the suffix .o with .ko, then
scripts/Makefile.modpost calls the sed command to change .ko back
to the original .o suffix.

Instead of converting the suffixes back-and-forth, store the .o paths
in modules.order, and replace it with .ko in 'make modules_install'.

This avoids the unneeded sed command.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Luis Chamberlain <mcgrof@kernel.org>
2022-12-14 15:42:40 +09:00
Guru Das Srinagesh 04518e4c2e scripts/clang-tools: Convert clang-tidy args to list
Convert list of clang-tidy arguments to a list for ease of adding to
them and extending them as required.

Signed-off-by: Guru Das Srinagesh <quic_gurus@quicinc.com>
Suggested-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2022-10-14 22:32:52 +09:00
Masahiro Yamada 3216484550 kbuild: use obj-y instead extra-y for objects placed at the head
The objects placed at the head of vmlinux need special treatments:

 - arch/$(SRCARCH)/Makefile adds them to head-y in order to place
   them before other archives in the linker command line.

 - arch/$(SRCARCH)/kernel/Makefile adds them to extra-y instead of
   obj-y to avoid them going into built-in.a.

This commit gets rid of the latter.

Create vmlinux.a to collect all the objects that are unconditionally
linked to vmlinux. The objects listed in head-y are moved to the head
of vmlinux.a by using 'ar m'.

With this, arch/$(SRCARCH)/kernel/Makefile can consistently use obj-y
for builtin objects.

There is no *.o that is directly linked to vmlinux. Drop unneeded code
in scripts/clang-tools/gen_compile_commands.py.

$(AR) mPi needs 'T' to workaround the llvm-ar bug. The fix was suggested
by Nathan Chancellor [1].

[1]: https://lore.kernel.org/llvm/YyjjT5gQ2hGMH0ni@dev-arch.thelio-3990X/

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Reviewed-by: Nicolas Schier <nicolas@fjasle.eu>
2022-10-02 18:04:05 +09:00
yangxingwu 237fe72749 scripts/clang-tools: remove unused module
Remove unused imported 'os' module.

Signed-off-by: yangxingwu <xingwu.yang@gmail.com>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2022-09-24 04:30:06 +09:00
Guru Das Srinagesh 4be72c1b9f scripts/clang-tools: Remove DeprecatedOrUnsafeBufferHandling check
This `clang-analyzer` check flags the use of memset(), suggesting a more
secure version of the API, such as memset_s(), which does not exist in
the kernel:

  warning: Call to function 'memset' is insecure as it does not provide
  security checks introduced in the C11 standard. Replace with analogous
  functions that support length arguments or provides boundary checks such
  as 'memset_s' in case of C11
  [clang-analyzer-security.insecureAPI.DeprecatedOrUnsafeBufferHandling]

Signed-off-by: Guru Das Srinagesh <quic_gurus@quicinc.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2022-08-21 02:47:49 +09:00
John Hubbard a4ab14e1d8 gen_compile_commands: handle multiple lines per .mod file
scripts/clang-tools/gen_compile_commands.py incorrectly assumes that
each .mod file only contains one line. That assumption was correct when
the script was originally created, but commit 9413e76405 ("kbuild:
split the second line of *.mod into *.usyms") changed the .mod file
format so that there is one entry per line, and potentially many lines.

The problem can be reproduced by using Kbuild to generate
compile_commands.json, like this:

    make CC=clang compile_commands.json

In many cases, the problem might be overlooked because many subsystems
only have one line anyway. However, in some subsystems (Nouveau, with
762 entries, is a notable example) it results in skipping most of the
subsystem.

Fix this by fully processing each .mod file.

Fixes: 9413e76405 ("kbuild: split the second line of *.mod into *.usyms")
Signed-off-by: John Hubbard <jhubbard@nvidia.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2022-06-29 11:43:13 +09:00
Kortan ec783c7cb2 gen_compile_commands: fix missing 'sys' package
We need to import the 'sys' package since the script has called
sys.exit() method.

Fixes: 6ad7cbc015 ("Makefile: Add clang-tidy and static analyzer support to makefile")
Signed-off-by: Kortan <kortanzh@gmail.com>
Reviewed-by: Nathan Chancellor <nathan@kernel.org>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2021-09-19 10:13:03 +09:00
Masahiro Yamada 265264b814 gen_compile_commands: extract compiler command from a series of commands
The current gen_compile_commands.py assumes that objects are always
built by a single command.

It makes sense to support cases where objects are built by a series of
commands:

  cmd_<object> := <command1> ; <command2>

One use-case is that <command1> is a compiler command, and <command2>
an objtool command.

It allows *.cmd files to contain an objtool command so that any change
in it triggers object rebuilds.

If ; appears after the C source file, take the first command.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
2021-09-03 08:17:20 +09:00
Masahiro Yamada 585d32f9b0 gen_compile_commands: prune some directories
If directories are passed to gen_compile_commands.py, os.walk() traverses
all the subdirectories to search for .cmd files, but we know some of them
are not worth traversing.

Use the 'topdown' parameter of os.walk to prune them.

Documentation about the 'topdown' option of os.walk:
  When topdown is True, the caller can modify the dirnames list
  in-place (perhaps using del or slice assignment), and walk() will
  only recurse into the subdirectories whose names remain in dirnames;
  this can be used to prune the search, impose a specific order of
  visiting, or even to inform walk() about directories the caller
  creates or renames before it resumes walk() again. Modifying
  dirnames when topdown is False has no effect on the behavior of
  the walk, because in bottom-up mode the directories in dirnames
  are generated before dirpath itself is generated.

This commit prunes four directories, .git, Documentation, include, and
tools.

The first three do not contain any C files, so skipping them makes this
script work slightly faster. My main motivation is the last one, tools/
directory.

Commit 6ca4c6d259 ("gen_compile_commands: do not support .cmd files
under tools/ directory") stopped supporting the tools/ directory.
The current code no longer picks up .cmd files from the tools/
directory.

If you run:

  ./scripts/clang-tools/gen_compile_commands.py --log_level=INFO

then, you will see several "File ... not found" log messages.

This is expected, and I do not want to support the tools/ directory.
However, without an explicit comment "do not support tools/", somebody
might try to get it back. Clarify this.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Acked-by: Nathan Chancellor <nathan@kernel.org>
2021-02-16 22:23:56 +09:00
Masahiro Yamada 074075aea2 scripts/clang-tools: switch explicitly to Python 3
For the same reason as commit 51839e29cb ("scripts: switch explicitly
to Python 3"), switch some more scripts, which I tested and confirmed
working on Python 3.

Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
Acked-by: Nathan Chancellor <nathan@kernel.org>
2021-02-03 09:01:08 +09:00
Nathan Huckleberry 6ad7cbc015 Makefile: Add clang-tidy and static analyzer support to makefile
This patch adds clang-tidy and the clang static-analyzer as make
targets. The goal of this patch is to make static analysis tools
usable and extendable by any developer or researcher who is familiar
with basic c++.

The current static analysis tools require intimate knowledge of the
internal workings of the static analysis. Clang-tidy and the clang
static analyzers expose an easy to use api and allow users unfamiliar
with clang to write new checks with relative ease.

===Clang-tidy===

Clang-tidy is an easily extendable 'linter' that runs on the AST.
Clang-tidy checks are easy to write and understand. A check consists of
two parts, a matcher and a checker. The matcher is created using a
domain specific language that acts on the AST
(https://clang.llvm.org/docs/LibASTMatchersReference.html).  When AST
nodes are found by the matcher a callback is made to the checker. The
checker can then execute additional checks and issue warnings.

Here is an example clang-tidy check to report functions that have calls
to local_irq_disable without calls to local_irq_enable and vice-versa.
Functions flagged with __attribute((annotation("ignore_irq_balancing")))
are ignored for analysis. (https://reviews.llvm.org/D65828)

===Clang static analyzer===

The clang static analyzer is a more powerful static analysis tool that
uses symbolic execution to find bugs. Currently there is a check that
looks for potential security bugs from invalid uses of kmalloc and
kfree. There are several more general purpose checks that are useful for
the kernel.

The clang static analyzer is well documented and designed to be
extensible.
(https://clang-analyzer.llvm.org/checker_dev_manual.html)
(https://github.com/haoNoQ/clang-analyzer-guide/releases/download/v0.1/clang-analyzer-guide-v0.1.pdf)

The main draw of the clang tools is how accessible they are. The clang
documentation is very nice and these tools are built specifically to be
easily extendable by any developer. They provide an accessible method of
bug-finding and research to people who are not overly familiar with the
kernel codebase.

Signed-off-by: Nathan Huckleberry <nhuck@google.com>
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Signed-off-by: Masahiro Yamada <masahiroy@kernel.org>
2020-08-27 00:44:33 +09:00