linux-block.git
8 months agotools/nolibc: fix return type of getpagesize()
Thomas Weißschuh [Thu, 3 Aug 2023 07:28:46 +0000 (09:28 +0200)]
tools/nolibc: fix return type of getpagesize()

It's documented as returning int which is also implemented by glibc and
musl, so adopt that return type.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: drop unused variables
Thomas Weißschuh [Thu, 3 Aug 2023 07:28:45 +0000 (09:28 +0200)]
tools/nolibc: drop unused variables

Nobody needs it, get rid of it.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add testcase for pipe
Yuan Tan [Tue, 1 Aug 2023 15:40:23 +0000 (23:40 +0800)]
selftests/nolibc: add testcase for pipe

Add a test case of pipe that sends and receives message in a single
process.

Suggested-by: Thomas Weißschuh <thomas@t-8ch.de>
Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/all/c5de2d13-3752-4e1b-90d9-f58cca99c702@t-8ch.de/
Signed-off-by: Yuan Tan <tanyuan@tinylab.org>
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
[wt: fixed the "len" type to size_t to address a sign-compare warning
 with upcoming patches]
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: add pipe() and pipe2() support
Yuan Tan [Tue, 1 Aug 2023 15:39:54 +0000 (23:39 +0800)]
tools/nolibc: add pipe() and pipe2() support

According to manual page [1], posix spec [2] and source code like
arch/mips/kernel/syscall.c, for historic reasons, the sys_pipe() syscall
on some architectures has an unusual calling convention.  It returns
results in two registers which means there is no need for it to do
verify the validity of a userspace pointer argument.  Historically that
used to be expensive in Linux.  These days the performance advantage is
negligible.

Nolibc doesn't support the unusual calling convention above, luckily
Linux provides a generic sys_pipe2() with an additional flags argument
from 2.6.27. If flags is 0, then pipe2() is the same as pipe(). So here
we use sys_pipe2() to implement the pipe().

pipe2() is also provided to allow users to use flags argument on demand.

[1]: https://man7.org/linux/man-pages/man2/pipe.2.html
[2]: https://pubs.opengroup.org/onlinepubs/9699919799/functions/pipe.html

Suggested-by: Zhangjin Wu <falcon@tinylab.org>
Link: https://lore.kernel.org/all/20230729100401.GA4577@1wt.eu/
Signed-off-by: Yuan Tan <tanyuan@tinylab.org>
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: mmap_munmap_good: fix up return value
Zhangjin Wu [Tue, 25 Jul 2023 16:43:36 +0000 (00:43 +0800)]
selftests/nolibc: mmap_munmap_good: fix up return value

The other tests use 1 as failure, mmap_munmap_good uses -1 as failure,
let's fix up this.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: avoid buffer underrun in space printing
Thomas Weißschuh [Wed, 26 Jul 2023 06:08:13 +0000 (08:08 +0200)]
selftests/nolibc: avoid buffer underrun in space printing

If the test description is longer than the status alignment the
parameter 'n' to putcharn() would lead to a signed underflow that then
gets converted to a very large unsigned value.
This in turn leads out-of-bound writes in memset() crashing the
application.

The failure case of EXPECT_PTRER() used in "mmap_bad" exhibits this
exact behavior.

Fixes: 29f5540be392 ("selftests/nolibc: add EXPECT_PTREQ, EXPECT_PTRNE and EXPECT_PTRER")
Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc/stdio: add setvbuf() to set buffering mode
Ryan Roberts [Wed, 26 Jul 2023 07:06:55 +0000 (08:06 +0100)]
tools/nolibc/stdio: add setvbuf() to set buffering mode

Add a minimal implementation of setvbuf(), which error checks the mode
argument (as required by spec) and returns. Since nolibc never buffers
output, nothing needs to be done.

The kselftest framework recently added a call to setvbuf(). As a result,
any tests that use the kselftest framework and nolibc cause a compiler
error due to missing function. This provides an urgent fix for the
problem which is preventing arm64 testing on linux-next.

Example:

clang --target=aarch64-linux-gnu -fintegrated-as
-Werror=unknown-warning-option -Werror=ignored-optimization-argument
-Werror=option-ignored -Werror=unused-command-line-argument
--target=aarch64-linux-gnu -fintegrated-as
-fno-asynchronous-unwind-tables -fno-ident -s -Os -nostdlib \
-include ../../../../include/nolibc/nolibc.h -I../..\
-static -ffreestanding -Wall za-fork.c
build/kselftest/arm64/fp/za-fork-asm.o
-o build/kselftest/arm64/fp/za-fork
In file included from <built-in>:1:
In file included from ./../../../../include/nolibc/nolibc.h:97:
In file included from ./../../../../include/nolibc/arch.h:25:
./../../../../include/nolibc/arch-aarch64.h:178:35: warning: unknown
attribute 'optimize' ignored [-Wunknown-attributes]
void __attribute__((weak,noreturn,optimize("omit-frame-pointer")))
__no_stack_protector _start(void)
                                  ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from za-fork.c:12:
../../kselftest.h:123:2: error: call to undeclared function 'setvbuf';
ISO C99 and later do not support implicit function declarations
[-Wimplicit-function-declaration]
        setvbuf(stdout, NULL, _IOLBF, 0);
        ^
../../kselftest.h:123:24: error: use of undeclared identifier '_IOLBF'
        setvbuf(stdout, NULL, _IOLBF, 0);
                              ^
1 warning and 2 errors generated.

Signed-off-by: Ryan Roberts <ryan.roberts@arm.com>
Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>
Link: https://lore.kernel.org/linux-kselftest/CA+G9fYus3Z8r2cg3zLv8uH8MRrzLFVWdnor02SNr=rCz+_WGVg@mail.gmail.com/
Reviewed-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: allow test -include /path/to/nolibc.h
Zhangjin Wu [Sat, 15 Jul 2023 18:36:46 +0000 (02:36 +0800)]
selftests/nolibc: allow test -include /path/to/nolibc.h

As the head comment of nolibc-test.c shows, it can be built in 3 ways:

    $(CC) -nostdlib -include /path/to/nolibc.h => NOLIBC already defined
    $(CC) -nostdlib -I/path/to/nolibc/sysroot  => _NOLIBC_* guards are present
    $(CC) with default libc                    => NOLIBC* never defined

Only last two of them are tested currently, let's allow test the first one too.

This may help to find issues about using nolibc.h to build programs. it
derives from this change:

    commit 3a8039e289a3 ("tools/nolibc: Fix build of stdio.h due to header ordering")

Usage:

    // test with sysroot by default
    $ make run-user

    // test without sysroot, using nolibc.h directly
    $ make run-user NOLIBC_SYSROOT=0

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: allow run nolibc-test locally
Zhangjin Wu [Sat, 15 Jul 2023 18:35:39 +0000 (02:35 +0800)]
selftests/nolibc: allow run nolibc-test locally

It is able to run nolibc-test directly without qemu-user when the target
machine is the same as the host machine.

Sometimes, the result running locally may help a lot when the qemu-user
package is too old.

When the target machine differs from the host machine, it is also able
to run nolibc-test directly with qemu-user-static + binfmt_misc.

Link: https://lore.kernel.org/lkml/ZKutZwIOfy5MqedG@1wt.eu/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add testcases for startup code
Zhangjin Wu [Sat, 15 Jul 2023 18:34:32 +0000 (02:34 +0800)]
selftests/nolibc: add testcases for startup code

The startup code is critical to get the right argc, argv, envp/environ
and _auxv, let's add a startup test group and the corresponding
testcases.

The "environ" test case is also moved from the stdlib test group to this
new startup test group and it is renamed to "environ_envp".

Since argv0 has been used by many other test cases, let's add testcases
to gurantee it too.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add EXPECT_PTRGE, EXPECT_PTRGT, EXPECT_PTRLE, EXPECT_PTRLT
Zhangjin Wu [Sat, 15 Jul 2023 18:33:26 +0000 (02:33 +0800)]
selftests/nolibc: add EXPECT_PTRGE, EXPECT_PTRGT, EXPECT_PTRLE, EXPECT_PTRLT

4 new pointer compare macros are added, they are similar to the integer
compare macros.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: s390: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:32:19 +0000 (02:32 +0800)]
tools/nolibc: s390: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: riscv: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:31:12 +0000 (02:31 +0800)]
tools/nolibc: riscv: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: loongarch: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:30:02 +0000 (02:30 +0800)]
tools/nolibc: loongarch: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: mips: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:28:55 +0000 (02:28 +0800)]
tools/nolibc: mips: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Also clean up the instructions in delayed slots.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: x86_64: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:27:49 +0000 (02:27 +0800)]
tools/nolibc: x86_64: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: i386: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:26:42 +0000 (02:26 +0800)]
tools/nolibc: i386: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: aarch64: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:25:35 +0000 (02:25 +0800)]
tools/nolibc: aarch64: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: arm: shrink _start with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:24:29 +0000 (02:24 +0800)]
tools/nolibc: arm: shrink _start with _start_c

move most of the _start operations to _start_c(), include the
stackprotector initialization.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: crt.h: initialize stack protector
Zhangjin Wu [Sat, 15 Jul 2023 18:23:21 +0000 (02:23 +0800)]
tools/nolibc: crt.h: initialize stack protector

As suggested by Thomas, It is able to move the stackprotector
initialization from the assembly _start to the beginning of the new
_start_c(). Let's call __stack_chk_init() in _start_c() as a
preparation.

Suggested-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/lkml/a00284a6-54b1-498c-92aa-44997fa78403@t-8ch.de/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: stackprotector.h: add empty __stack_chk_init for !_NOLIBC_STACKPROTECTOR
Zhangjin Wu [Sat, 15 Jul 2023 18:22:15 +0000 (02:22 +0800)]
tools/nolibc: stackprotector.h: add empty __stack_chk_init for !_NOLIBC_STACKPROTECTOR

Let's define an empty __stack_chk_init for the !_NOLIBC_STACKPROTECTOR
branch.

This allows to remove #ifdef around every call of __stack_chk_init().

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: add new crt.h with _start_c
Zhangjin Wu [Sat, 15 Jul 2023 18:21:08 +0000 (02:21 +0800)]
tools/nolibc: add new crt.h with _start_c

As the environ and _auxv support added for nolibc, the assembly _start
function becomes more and more complex and therefore makes the porting
of nolibc to new architectures harder and harder.

To simplify portability, this C version of _start_c() is added to do
most of the assembly start operations in C, which reduces the complexity
a lot and will eventually simplify the porting of nolibc to the new
architectures.

The new _start_c() only requires a stack pointer argument, it will find
argc, argv, envp/environ and _auxv for us, and then call main(),
finally, it exit() with main's return status. With this new _start_c(),
the future new architectures only require to add very few assembly
instructions.

As suggested by Thomas, users may use a different signature of main
(e.g. void main(void)), a _nolibc_main alias is added for main to
silence the warning about potential conflicting types.

As suggested by Willy, the code is carefully polished for both smaller
size and better readability with local variables and the right types.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/20230715095729.GC24086@1wt.eu/
Suggested-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/lkml/90fdd255-32f4-4caf-90ff-06456b53dac3@t-8ch.de/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: remove the old sys_stat support
Zhangjin Wu [Sat, 15 Jul 2023 18:20:01 +0000 (02:20 +0800)]
tools/nolibc: remove the old sys_stat support

The statx manpage [1] shows that it has been supported from Linux 4.11
and glibc 2.28, the Linux support can be checked for all of the
architectures with this command:

    $ git grep -r statx v4.11 arch/ include/uapi/asm-generic/unistd.h \
      | grep -E "aarch64|arm|mips|s390|x86|:include/uapi"

Besides riscv and loongarch, all of the nolibc supported architectures
have added sys_statx from Linux v4.11. riscv is mainlined to v4.15,
loongarch is mainlined to v5.19, both of them use the generic unistd.h,
so, they have added sys_statx from their first mainline versions.

The current oldest stable branch is v4.14, only reserving sys_statx
still preserves compatibility with all of the supported stable branches,
So, let's remove the old arch related and dependent sys_stat support
completely.

This is friendly to the future new architecture porting.

[1]: https://man7.org/linux/man-pages/man2/statx.2.html

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: fix up startup failures for -O0 under gcc < 11.1.0
Zhangjin Wu [Sat, 15 Jul 2023 18:18:54 +0000 (02:18 +0800)]
tools/nolibc: fix up startup failures for -O0 under gcc < 11.1.0

As gcc doc [1] shows:

  Most optimizations are completely disabled at -O0 or if an -O level is
  not set on the command line, even if individual optimization flags are
  specified.

Test result [2] shows, gcc>=11.1.0 deviates from the above description,
but before gcc 11.1.0, "-O0" still forcely uses frame pointer in the
_start function even if the individual optimize("omit-frame-pointer")
flag is specified.

The frame pointer related operations will change the stack pointer (e.g.
In x86_64, an extra "push %rbp" will be inserted at the beginning of
_start) and make it differs from the one we expected, as a result, break
the whole startup function.

To fix up this issue, as suggested by Thomas, the individual "Os" and
"omit-frame-pointer" optimize flags are used together on _start function
to disable frame pointer completely even if the -O0 is set on the
command line.

[1]: https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html
[2]: https://lore.kernel.org/lkml/20230714094723.140603-1-falcon@tinylab.org/

Suggested-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/lkml/34b21ba5-7b59-4b3b-9ed6-ef9a3a5e06f7@t-8ch.de/
Fixes: 7f8548589661 ("tools/nolibc: make compiler and assembler agree on the section around _start")
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: arch-*.h: add missing space after ','
Zhangjin Wu [Sat, 15 Jul 2023 18:17:43 +0000 (02:17 +0800)]
tools/nolibc: arch-*.h: add missing space after ','

Fix up such errors reported by scripts/checkpatch.pl:

    ERROR: space required after that ',' (ctx:VxV)
    #148: FILE: tools/include/nolibc/arch-aarch64.h:148:
    +void __attribute__((weak,noreturn,optimize("omit-frame-pointer"))) __no_stack_protector _start(void)
                             ^

    ERROR: space required after that ',' (ctx:VxV)
    #148: FILE: tools/include/nolibc/arch-aarch64.h:148:
    +void __attribute__((weak,noreturn,optimize("omit-frame-pointer"))) __no_stack_protector _start(void)
                                      ^

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: avoid gaps in test numbers
Thomas Weißschuh [Tue, 11 Jul 2023 09:48:42 +0000 (11:48 +0200)]
selftests/nolibc: avoid gaps in test numbers

As the test numbers are based on line numbers gaps without testcases are
to be avoided.
Instead use the already existing test condition logic to implement
conditional execution.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: simplify status printing
Thomas Weißschuh [Tue, 11 Jul 2023 09:48:40 +0000 (11:48 +0200)]
selftests/nolibc: simplify status printing

pad_spc() is only ever used to print the status message of testcases.
The line size is always constant, the return value is never used and the
format string is never used as such.

Remove all the unneeded logic and simplify the API and its users.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: make evaluation of test conditions
Thomas Weißschuh [Tue, 11 Jul 2023 09:48:39 +0000 (11:48 +0200)]
selftests/nolibc: make evaluation of test conditions

If "cond" is a multi-token statement the behavior of the preprocessor
will lead to the negation "!" to be only applied to the first token.
Although currently no test uses such multi-token conditions but it can
happen at any time.

Put braces around "cond" to ensure the negation works as expected.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: completely remove optional environ support
Thomas Weißschuh [Mon, 10 Jul 2023 18:01:34 +0000 (20:01 +0200)]
tools/nolibc: completely remove optional environ support

In commit 52e423f5b93e ("tools/nolibc: export environ as a weak symbol on i386")
and friends the asm startup logic was extended to directly populate the
"environ" array.

This makes it impossible for "environ" to be dropped by the linker.
Therefore also drop the other logic to handle non-present "environ".

Also add a testcase to validate the initialization of environ.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: report: add newline before test failures
Zhangjin Wu [Thu, 6 Jul 2023 09:12:28 +0000 (17:12 +0800)]
selftests/nolibc: report: add newline before test failures

a newline is inserted just before the test failures to avoid mixing the
test failures with the raw test log.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: report: extrude the test status line
Zhangjin Wu [Thu, 6 Jul 2023 09:11:17 +0000 (17:11 +0800)]
selftests/nolibc: report: extrude the test status line

two newlines are added around the test summary line to extrude the test
status.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: report: align passed, skipped and failed
Zhangjin Wu [Thu, 6 Jul 2023 09:10:08 +0000 (17:10 +0800)]
selftests/nolibc: report: align passed, skipped and failed

align the test values for different runs and different architectures.

Since the total number of tests is not bigger than 1000 currently, let's
align them with "%3d".

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
[wt: s/%03d/%3d/ as discussed with Zhangjin]
Link: https://lore.kernel.org/lkml/20230709185112.97236-1-falcon@tinylab.org/
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: report: print total tests
Zhangjin Wu [Thu, 6 Jul 2023 09:08:16 +0000 (17:08 +0800)]
selftests/nolibc: report: print total tests

Let's count and print the total number of tests, now, the data of
passed, skipped and failed have the same format.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: report: print a summarized test status
Zhangjin Wu [Thu, 6 Jul 2023 09:03:34 +0000 (17:03 +0800)]
selftests/nolibc: report: print a summarized test status

one of the test status: success, warning and failure is printed to
summarize the passed, skipped and failed values.

- "success" means no skipped and no failed.
- "warning" means has at least one skipped and no failed.
- "failure" means all tests are failed.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/20230702164358.GB16233@1wt.eu/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add chmod_argv0 test
Zhangjin Wu [Fri, 7 Jul 2023 18:46:20 +0000 (02:46 +0800)]
selftests/nolibc: add chmod_argv0 test

argv0 is readable and chmodable, let's use it for chmod test, but a safe
umask should be used, the readable and executable modes should be
reserved.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: chroot_exe: remove procfs dependency
Zhangjin Wu [Fri, 7 Jul 2023 18:45:08 +0000 (02:45 +0800)]
selftests/nolibc: chroot_exe: remove procfs dependency

Since argv0 also works for CONFIG_PROC_FS=n, let's use it instead of
'/proc/self/exe'.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: stat_timestamps: remove procfs dependency
Zhangjin Wu [Fri, 7 Jul 2023 18:43:44 +0000 (02:43 +0800)]
selftests/nolibc: stat_timestamps: remove procfs dependency

'/proc/self/' is a good path which doesn't have stale time info but it
is only available for CONFIG_PROC_FS=y.

When CONFIG_PROC_FS=n, use argv0 instead of '/proc/self', use '/' for the
worst case.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: chdir_root: restore current path after test
Zhangjin Wu [Fri, 7 Jul 2023 18:42:01 +0000 (02:42 +0800)]
selftests/nolibc: chdir_root: restore current path after test

The PWD environment variable has the path of the nolibc-test program,
the current path must be the same as it, otherwise, the test cases will
fail with relative path (e.g. ./nolibc-test).

Since only chdir_root really changes the current path, let's restore it
with the PWD environment variable.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: vfprintf: remove MEMFD_CREATE dependency
Zhangjin Wu [Fri, 7 Jul 2023 18:40:08 +0000 (02:40 +0800)]
selftests/nolibc: vfprintf: remove MEMFD_CREATE dependency

The vfprintf test case require to open a temporary file to write, the
old memfd_create() method is perfect but has strong dependency on
MEMFD_CREATE and also TMPFS or HUGETLBFS (see fs/Kconfig):

    config MEMFD_CREATE
def_bool TMPFS || HUGETLBFS

And from v6.2, MFD_NOEXEC_SEAL must be passed for the non-executable
memfd, otherwise, The kernel warning will be output to the test result
like this:

        Running test 'vfprintf'
        0 emptymemfd_create() without MFD_EXEC nor MFD_NOEXEC_SEAL, pid=1 'init'
         "" = ""                                                  [OK]

To avoid such warning and also to remove the MEMFD_CREATE dependency,
let's open a file from tmpfs directly.

The /tmp directory is used to detect the existing of tmpfs, if not
there, skip instead of fail.

And further, for pid == 1, the initramfs is loaded as ramfs, which can
be used as tmpfs, so, it is able to further remove TMPFS dependency too.

Suggested-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/lkml/9ad51430-b7c0-47dc-80af-20c86539498d@t-8ch.de
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: prepare /tmp for tests that need to write
Zhangjin Wu [Fri, 7 Jul 2023 18:38:57 +0000 (02:38 +0800)]
selftests/nolibc: prepare /tmp for tests that need to write

create a /tmp directory. If it succeeds, the directory is writable,
which is normally the case when booted from an initramfs anyway.

This will be used instead of procfs for some tests.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Link: https://lore.kernel.org/lkml/20230710050600.9697-1-falcon@tinylab.org/
[wt: removed the unneeded mount() call]
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: fix up failures when CONFIG_PROC_FS=n
Zhangjin Wu [Fri, 7 Jul 2023 18:37:45 +0000 (02:37 +0800)]
selftests/nolibc: fix up failures when CONFIG_PROC_FS=n

For CONFIG_PROC_FS=n, the /proc is not mountable, but the /proc
directory has been created in the prepare() stage whenever /proc is
there or not.

so, the checking of /proc in the run_syscall() stage will be always true
and at last it will fail all of the procfs dependent test cases, which
deviates from the 'cond' check design of the EXPECT_xx macros, without
procfs, these test cases should be skipped instead of failed.

To solve this issue, one method is checking /proc/self instead of /proc,
another method is removing the /proc directory completely for
CONFIG_PROC_FS=n, we apply the second method to avoid misleading the
users.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add a new rmdir() test case
Zhangjin Wu [Fri, 7 Jul 2023 18:35:38 +0000 (02:35 +0800)]
selftests/nolibc: add a new rmdir() test case

A new rmdir_blah test case is added to remove a non-existing /blah,
which expects failure with ENOENT errno.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: add rmdir() support
Zhangjin Wu [Fri, 7 Jul 2023 18:34:21 +0000 (02:34 +0800)]
tools/nolibc: add rmdir() support

a reverse operation of mkdir() is meaningful, add rmdir() here.

required by nolibc-test to remove /proc while CONFIG_PROC_FS is not
enabled.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: link_cross: use /proc/self/cmdline
Zhangjin Wu [Fri, 7 Jul 2023 18:33:13 +0000 (02:33 +0800)]
selftests/nolibc: link_cross: use /proc/self/cmdline

For CONFIG_NET=n, there would be no /proc/self/net, so, use
/proc/self/cmdline instead.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: fix up kernel parameters support
Zhangjin Wu [Fri, 7 Jul 2023 18:32:05 +0000 (02:32 +0800)]
selftests/nolibc: fix up kernel parameters support

kernel parameters allow pass two types of strings, one type is like
'noapic', another type is like 'panic=5', the first type is passed as
arguments of the init program, the second type is passed as environment
variables of the init program.

when users pass kernel parameters like this:

    noapic NOLIBC_TEST=syscall

our nolibc-test program will use the test setting from argv[1] and
ignore the one from NOLIBC_TEST environment variable, and at last, it
will print the following line and ignore the whole test setting.

    Ignoring unknown test name 'noapic'

reversing the parsing order does solve the above issue:

    test = getenv("NOLIBC_TEST");
    if (test)
        test = argv[1];

but it still doesn't work with such kernel parameters (without
NOLIBC_TEST environment variable):

    noapic FOO=bar

To support all of the potential kernel parameters, let's verify the test
setting from both of argv[1] and NOLIBC_TEST environment variable.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: prefer <sys/reboot.h> to <linux/reboot.h>
Zhangjin Wu [Fri, 7 Jul 2023 18:30:58 +0000 (02:30 +0800)]
selftests/nolibc: prefer <sys/reboot.h> to <linux/reboot.h>

Since both glibc and musl provide RB_ flags via <sys/reboot.h>, and we
just add RB_ flags for nolibc, let's use RB_ flags instead of
LINUX_REBOOT_ flags and only reserve the required <sys/reboot.h> header.

This allows compile libc-test for musl libc without the linux headers.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: types.h: add RB_ flags for reboot()
Zhangjin Wu [Fri, 7 Jul 2023 18:29:30 +0000 (02:29 +0800)]
tools/nolibc: types.h: add RB_ flags for reboot()

Both glibc and musl provide RB_ flags via <sys/reboot.h> for reboot(),
they don't need to include <linux/reboot.h>, let nolibc provide RB_
flags too.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: fix up int_fast16/32_t test cases for musl
Zhangjin Wu [Fri, 7 Jul 2023 18:27:40 +0000 (02:27 +0800)]
selftests/nolibc: fix up int_fast16/32_t test cases for musl

musl limits the fast signed int in 32bit, but glibc and nolibc don't, to
let such test cases work on musl, let's provide the type based
SINT_MAX_OF_TYPE(type) and SINT_MIN_OF_TYPE(type).

Suggested-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/lkml/bc635c4f-67fe-4e86-bfdf-bcb4879b928d@t-8ch.de/
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add _LARGEFILE64_SOURCE for musl
Zhangjin Wu [Fri, 7 Jul 2023 18:26:24 +0000 (02:26 +0800)]
selftests/nolibc: add _LARGEFILE64_SOURCE for musl

_GNU_SOURCE Implies _LARGEFILE64_SOURCE in glibc, but in musl, the
default configuration doesn't enable _LARGEFILE64_SOURCE.

>From include/dirent.h of musl, getdents64 is provided as getdents when
_LARGEFILE64_SOURCE is defined.

    #if defined(_LARGEFILE64_SOURCE)
    ...
    #define getdents64 getdents
    #endif

Let's define _LARGEFILE64_SOURCE to fix up this compile error:

    tools/testing/selftests/nolibc/nolibc-test.c: In function ‘test_getdents64’:
    tools/testing/selftests/nolibc/nolibc-test.c:453:8: warning: implicit declaration of function ‘getdents64’; did you mean ‘getdents’? [-Wimplicit-function-declaration]
      453 |  ret = getdents64(fd, (void *)buffer, sizeof(buffer));
          |        ^~~~~~~~~~
          |        getdents
    /usr/bin/ld: /tmp/ccKILm5u.o: in function `test_getdents64':
    nolibc-test.c:(.text+0xe3e): undefined reference to `getdents64'
    collect2: error: ld returned 1 exit status

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: gettid: restore for glibc and musl
Zhangjin Wu [Fri, 7 Jul 2023 18:25:11 +0000 (02:25 +0800)]
selftests/nolibc: gettid: restore for glibc and musl

As the gettid manpage [1] shows, glibc 2.30 has gettid support, so,
let's enable the test for glibc >= 2.30.

gettid works on musl too.

[1]: https://man7.org/linux/man-pages/man2/gettid.2.html

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: stat_fault: silence NULL argument warning with glibc
Zhangjin Wu [Fri, 7 Jul 2023 18:23:45 +0000 (02:23 +0800)]
selftests/nolibc: stat_fault: silence NULL argument warning with glibc

Use another invalid address (void *)1 instead of NULL to silence this
compile warning with glibc:

    $ make libc-test
      CC      libc-test
    nolibc-test.c: In function ‘run_syscall’:
    nolibc-test.c:622:49: warning: null argument where non-null required (argument 1) [-Wnonnull]
      622 |   CASE_TEST(stat_fault);        EXPECT_SYSER(1, stat(NULL, &stat_buf), -1, EFAULT); break;
          |                                                 ^~~~
    nolibc-test.c:304:79: note: in definition of macro ‘EXPECT_SYSER2’
      304 |  do { if (!cond) pad_spc(llen, 64, "[SKIPPED]\n"); else ret += expect_syserr2(expr, expret, experr1, experr2, llen); } while (0)
          |                                                                               ^~~~
    nolibc-test.c:622:33: note: in expansion of macro ‘EXPECT_SYSER’
      622 |   CASE_TEST(stat_fault);        EXPECT_SYSER(1, stat(NULL, &stat_buf), -1, EFAULT); break;

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add run-libc-test target
Zhangjin Wu [Fri, 7 Jul 2023 18:22:39 +0000 (02:22 +0800)]
selftests/nolibc: add run-libc-test target

allow run and report glibc or musl based libc-test.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add mmap_munmap_good test case
Zhangjin Wu [Fri, 7 Jul 2023 15:08:20 +0000 (23:08 +0800)]
selftests/nolibc: add mmap_munmap_good test case

mmap() a file with a good offset and then munmap() it. a non-zero offset
is passed to test the 6th argument of my_syscall6().

Note, it is not easy to find a unique file for mmap() in different
scenes, so, a file list is used to search the right one:

- /dev/zero: is commonly used to allocate anonymous memory and is likely
  present and readable

- /proc/1/exe: for 'run' and 'run-user' target, 'run-user' can not find
  '/proc/self/exe'

- /proc/self/exe: for 'libc-test' target, normal program 'libc-test' has
  no permission to access '/proc/1/exe'

- argv0: the path of the program itself, let it pass even with worst
  case scene: no procfs and no /dev/zero

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/20230702193306.GK16233@1wt.eu/
Suggested-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/lkml/bff82ea6-610b-4471-a28b-6c76c28604a6@t-8ch.de/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add munmap_bad test case
Zhangjin Wu [Fri, 7 Jul 2023 15:07:13 +0000 (23:07 +0800)]
selftests/nolibc: add munmap_bad test case

The addr argument of munmap() must be a multiple of the page size,
passing invalid (void *)1 addr expects failure with -EINVAL.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add mmap_bad test case
Zhangjin Wu [Fri, 7 Jul 2023 15:05:49 +0000 (23:05 +0800)]
selftests/nolibc: add mmap_bad test case

The length argument of mmap() must be greater than 0, passing a zero
length argument expects failure with -EINVAL.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add sbrk_0 to test current brk getting
Zhangjin Wu [Fri, 7 Jul 2023 15:04:38 +0000 (23:04 +0800)]
selftests/nolibc: add sbrk_0 to test current brk getting

>From musl 0.9.14 (to the latest version 1.2.3), both sbrk() and brk()
have almost been disabled for they conflict with malloc, only sbrk(0) is
still permitted as a way to get the current location of the program
break, let's support such case.

EXPECT_PTRNE() is used to expect sbrk() always successfully getting the
current break.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: add EXPECT_PTREQ, EXPECT_PTRNE and EXPECT_PTRER
Zhangjin Wu [Fri, 7 Jul 2023 15:03:26 +0000 (23:03 +0800)]
selftests/nolibc: add EXPECT_PTREQ, EXPECT_PTRNE and EXPECT_PTRER

The syscalls like sbrk() and mmap() return pointers, to test them, more
pointer compare test macros are required, add them:

- EXPECT_PTREQ() expects two equal pointers.
- EXPECT_PTRNE() expects two non-equal pointers.
- EXPECT_PTRER() expects failure with a specified errno.
- EXPECT_PTRER2() expects failure with one of two specified errnos.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: prepare: create /dev/zero
Zhangjin Wu [Fri, 7 Jul 2023 15:02:16 +0000 (23:02 +0800)]
selftests/nolibc: prepare: create /dev/zero

/dev/zero is commonly used to allocate anonymous memory, it is a very
good file for tests, let's prepare it.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/20230702193306.GK16233@1wt.eu/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: export argv0 for some tests
Zhangjin Wu [Fri, 7 Jul 2023 15:01:09 +0000 (23:01 +0800)]
selftests/nolibc: export argv0 for some tests

argv0 is the path to nolibc-test program itself, which is a very good
always existing readable file for some tests, let's export it.

Note, the path may be absolute or relative, please make sure the tests
work with both of them. If it is relative, we must make sure the current
path is the one specified by the PWD environment variable.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/ZKKbS3cwKcHgnGwu@1wt.eu/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: clean up sbrk() routine
Zhangjin Wu [Fri, 7 Jul 2023 14:59:53 +0000 (22:59 +0800)]
tools/nolibc: clean up sbrk() routine

Fix up the error reported by scripts/checkpatch.pl:

    ERROR: do not use assignment in if condition
    #95: FILE: tools/include/nolibc/sys.h:95:
    + if ((ret = sys_brk(0)) && (sys_brk(ret + inc) == ret + inc))

Apply the new generic __sysret() to merge the SET_ERRNO() and return
lines.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: clean up mmap() routine
Zhangjin Wu [Fri, 7 Jul 2023 14:58:20 +0000 (22:58 +0800)]
tools/nolibc: clean up mmap() routine

Do several cleanups together:

- Since all supported architectures have my_syscall6() now, remove the
  #ifdef check.

- Move the mmap() related macros to tools/include/nolibc/types.h and
  reuse most of them from <linux/mman.h>

- Apply the new generic __sysret() to convert the calling of sys_map()
  to oneline code

Note, since MAP_FAILED is -1 on Linux, so we can use the generic
__sysret() which returns -1 upon error and still satisfy user land that
checks for MAP_FAILED.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/20230702192347.GJ16233@1wt.eu/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: __sysret: support syscalls who return a pointer
Zhangjin Wu [Fri, 7 Jul 2023 14:56:59 +0000 (22:56 +0800)]
tools/nolibc: __sysret: support syscalls who return a pointer

No official reference states the errno range, here aligns with musl and
glibc and uses [-MAX_ERRNO, -1] instead of all negative ones.

- musl: src/internal/syscall_ret.c
- glibc: sysdeps/unix/sysv/linux/sysdep.h

The MAX_ERRNO used by musl and glibc is 4095, just like the one nolibc
defined in tools/include/nolibc/errno.h.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/ZKKdD%2Fp4UkEavru6@1wt.eu/
Suggested-by: David Laight <David.Laight@ACULAB.COM>
Link: https://lore.kernel.org/linux-riscv/94dd5170929f454fbc0a10a2eb3b108d@AcuMS.aculab.com/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: add missing my_syscall6() for mips
Zhangjin Wu [Fri, 7 Jul 2023 14:55:35 +0000 (22:55 +0800)]
tools/nolibc: add missing my_syscall6() for mips

It is able to pass the 6th argument like the 5th argument via the stack
for mips, let's add a new my_syscall6() now, see [1] for details:

  The mips/o32 system call convention passes arguments 5 through 8 on
  the user stack.

Both mmap() and pselect6() require my_syscall6().

[1]: https://man7.org/linux/man-pages/man2/syscall.2.html

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: arch-mips.h: shrink with _NOLIBC_SYSCALL_CLOBBERLIST
Zhangjin Wu [Fri, 7 Jul 2023 14:54:26 +0000 (22:54 +0800)]
tools/nolibc: arch-mips.h: shrink with _NOLIBC_SYSCALL_CLOBBERLIST

my_syscall<N> share the same long clobber list, define a macro for them.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: arch-loongarch.h: shrink with _NOLIBC_SYSCALL_CLOBBERLIST
Zhangjin Wu [Fri, 7 Jul 2023 14:53:20 +0000 (22:53 +0800)]
tools/nolibc: arch-loongarch.h: shrink with _NOLIBC_SYSCALL_CLOBBERLIST

my_syscall<N> share the same long clobber list, define a macro for them.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotoolc/nolibc: arch-*.h: clean up whitespaces after __asm__
Zhangjin Wu [Fri, 7 Jul 2023 14:52:09 +0000 (22:52 +0800)]
toolc/nolibc: arch-*.h: clean up whitespaces after __asm__

replace "__asm__  volatile" with "__asm__ volatile" and insert necessary
whitespace before "\" to make sure the lines are aligned.

    $ sed -i -e 's/__asm__  volatile ( /__asm__ volatile (  /g' tools/include/nolibc/*.h

Note, arch-s390.h uses post-tab instead of post-whitespaces, must avoid
insert whitespace just before the tabs:

    $ sed -i -e 's/__asm__  volatile (\t/__asm__ volatile (\t/g' tools/include/nolibc/arch-*.h

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agotools/nolibc: arch-*.h: fix up code indent errors
Zhangjin Wu [Fri, 7 Jul 2023 14:50:34 +0000 (22:50 +0800)]
tools/nolibc: arch-*.h: fix up code indent errors

More than 8 whitespaces of the code indent are replaced with "tab +
whitespaces" to fix up such errors reported by scripts/checkpatch.pl:

    ERROR: code indent should use tabs where possible
    #64: FILE: tools/include/nolibc/arch-mips.h:64:
    +^I                                                                      \$

    ERROR: code indent should use tabs where possible
    #72: FILE: tools/include/nolibc/arch-mips.h:72:
    +^I          "t0", "t1", "t2", "t3", "t4", "t5", "t6", "t7", "t8", "t9"  \$

This command is used:

    $ sed -i -e '/^\t*        /{s/        /\t/g}' tools/include/nolibc/arch-*.h

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
8 months agoselftests/nolibc: simplify call to ioperm
Thomas Weißschuh [Mon, 3 Jul 2023 08:11:08 +0000 (10:11 +0200)]
selftests/nolibc: simplify call to ioperm

Since commit 53fcfafa8c5c ("tools/nolibc/unistd: add syscall()") nolibc
has support for syscall(2).
Use it to get rid of some ifdef-ery.

Signed-off-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agotools/nolibc: unistd.h: reorder the syscall macros
Zhangjin Wu [Wed, 28 Jun 2023 13:14:53 +0000 (21:14 +0800)]
tools/nolibc: unistd.h: reorder the syscall macros

Tune the macros in the using order and align most of them.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agotools/nolibc: sys.h: apply __sysret() helper
Zhangjin Wu [Wed, 28 Jun 2023 13:13:25 +0000 (21:13 +0800)]
tools/nolibc: sys.h: apply __sysret() helper

Use __sysret() to shrink most of the library routines to oneline code.

Removed 266 lines of duplicated code.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agotools/nolibc: unistd.h: apply __sysret() helper
Zhangjin Wu [Wed, 28 Jun 2023 13:11:44 +0000 (21:11 +0800)]
tools/nolibc: unistd.h: apply __sysret() helper

Use __sysret() to shrink the whole _syscall() to oneline code.

Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agotools/nolibc: sys.h: add a syscall return helper
Zhangjin Wu [Wed, 28 Jun 2023 13:08:32 +0000 (21:08 +0800)]
tools/nolibc: sys.h: add a syscall return helper

Most of the library routines share the same syscall return logic:

  In general, a 0 return value indicates success.  A -1 return value
  indicates an error, and an error number is stored in errno. [1]

Let's add a __sysret() helper for the above logic to simplify the coding
and shrink the code lines too.

Thomas suggested to use inline function instead of macro for __sysret().

Willy suggested to make __sysret() be always inline.

[1]: https://man7.org/linux/man-pages/man2/syscall.2.html

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/linux-riscv/ZH1+hkhiA2+ItSvX@1wt.eu/
Suggested-by: Thomas Weißschuh <linux@weissschuh.net>
Link: https://lore.kernel.org/linux-riscv/ea4e7442-7223-4211-ba29-70821e907888@t-8ch.de/
Reviewed-by: Thomas Weißschuh <linux@weissschuh.net>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agotools/nolibc: fix up undeclared syscall macros with #ifdef and -ENOSYS
Zhangjin Wu [Mon, 19 Jun 2023 12:25:47 +0000 (20:25 +0800)]
tools/nolibc: fix up undeclared syscall macros with #ifdef and -ENOSYS

Compiling nolibc for rv32 got such errors:

    nolibc/sysroot/riscv/include/sys.h: In function ‘sys_gettimeofday’:
    nolibc/sysroot/riscv/include/sys.h:557:21: error: ‘__NR_gettimeofday’ undeclared (first use in this function); did you mean ‘sys_gettimeofday’?
      557 |  return my_syscall2(__NR_gettimeofday, tv, tz);
          |                     ^~~~~~~~~~~~~~~~~
    nolibc/sysroot/riscv/include/sys.h: In function ‘sys_lseek’:
    nolibc/sysroot/riscv/include/sys.h:675:21: error: ‘__NR_lseek’ undeclared (first use in this function)
      675 |  return my_syscall3(__NR_lseek, fd, offset, whence);
          |                     ^~~~~~~~~~
    nolibc/sysroot/riscv/include/sys.h: In function ‘sys_wait4’:
    nolibc/sysroot/riscv/include/sys.h:1341:21: error: ‘__NR_wait4’ undeclared (first use in this function)
     1341 |  return my_syscall4(__NR_wait4, pid, status, options, rusage);

If a syscall macro is not supported by a target platform, wrap it with
'#ifdef' and 'return -ENOSYS' for the '#else' branch, which lets the
other syscalls work as-is and allows developers to fix up the test
failures reported by nolibc-test one by one later.

This wraps all of the failed syscall macros with '#ifdef' and 'return
-ENOSYS' for the '#else' branch, so, all of the undeclared failures are
fixed.

Suggested-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/linux-riscv/5e7d2adf-e96f-41ca-a4c6-5c87a25d4c9c@app.fastmail.com/
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agotools/nolibc: fix up #error compile failures with -ENOSYS
Zhangjin Wu [Mon, 19 Jun 2023 12:24:15 +0000 (20:24 +0800)]
tools/nolibc: fix up #error compile failures with -ENOSYS

Compiling nolibc for rv32 got such errors:

    In file included from nolibc/sysroot/riscv/include/nolibc.h:99,
                     from nolibc/sysroot/riscv/include/errno.h:26,
                     from nolibc/sysroot/riscv/include/stdio.h:14,
                     from tools/testing/selftests/nolibc/nolibc-test.c:12:
    nolibc/sysroot/riscv/include/sys.h:946:2: error: #error Neither __NR_ppoll nor __NR_poll defined, cannot implement sys_poll()
      946 | #error Neither __NR_ppoll nor __NR_poll defined, cannot implement sys_poll()
          |  ^~~~~
    nolibc/sysroot/riscv/include/sys.h:1062:2: error: #error None of __NR_select, __NR_pselect6, nor __NR__newselect defined, cannot implement sys_select()
     1062 | #error None of __NR_select, __NR_pselect6, nor __NR__newselect defined, cannot implement sys_select()

If a syscall is not supported by a target platform, 'return -ENOSYS' is
better than '#error', which lets the other syscalls work as-is and
allows developers to fix up the test failures reported by nolibc-test
one by one later.

This converts all of the '#error' to 'return -ENOSYS', so, all of the
'#error' failures are fixed.

Suggested-by: Arnd Bergmann <arnd@arndb.de>
Link: https://lore.kernel.org/linux-riscv/5e7d2adf-e96f-41ca-a4c6-5c87a25d4c9c@app.fastmail.com/
Reviewed-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agoselftests/nolibc: restore the failed tests print
Zhangjin Wu [Mon, 19 Jun 2023 07:01:43 +0000 (15:01 +0800)]
selftests/nolibc: restore the failed tests print

The commit fa0df56a804b ("selftests/nolibc: also count skipped and
failed tests in output") added counting for the skipped and failed
tests, but also removed the 'FAIL' results print, let's restore it for
it really allow users to learn the failed details without opening the
log file.

Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agoselftests/nolibc: always print the path to test log file
Zhangjin Wu [Mon, 19 Jun 2023 06:57:34 +0000 (14:57 +0800)]
selftests/nolibc: always print the path to test log file

Even when there is no failure, developers may be still interested in the
test log file, especially, string alignment, duplicated print, kernel
message and so forth, so, always print the path to test log file.

A new line is added for such a print to avoid annoying people who don't
care about it when the test pass completely.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/ZIB792FtG6ibOudp@1wt.eu/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
9 months agoselftests/nolibc: add a standalone test report macro
Zhangjin Wu [Mon, 19 Jun 2023 06:56:22 +0000 (14:56 +0800)]
selftests/nolibc: add a standalone test report macro

The run-user, run and rerun targets use the same test report script,
let's add a standalone test report macro for them.

This shrinks code lines and simplify the future maintainability.

Suggested-by: Willy Tarreau <w@1wt.eu>
Link: https://lore.kernel.org/lkml/ZIB792FtG6ibOudp@1wt.eu/
Signed-off-by: Zhangjin Wu <falcon@tinylab.org>
Signed-off-by: Willy Tarreau <w@1wt.eu>
10 months agoLinux 6.5-rc1 v6.5-rc1
Linus Torvalds [Sun, 9 Jul 2023 20:53:13 +0000 (13:53 -0700)]
Linux 6.5-rc1

10 months agoMAINTAINERS 2: Electric Boogaloo
Linus Torvalds [Sun, 9 Jul 2023 17:29:53 +0000 (10:29 -0700)]
MAINTAINERS 2: Electric Boogaloo

We just sorted the entries and fields last release, so just out of a
perverse sense of curiosity, I decided to see if we can keep things
ordered for even just one release.

The answer is "No. No we cannot".

I suggest that all kernel developers will need weekly training sessions,
involving a lot of Big Bird and Sesame Street.  And at the yearly
maintainer summit, we will all sing the alphabet song together.

I doubt I will keep doing this.  At some point "perverse sense of
curiosity" turns into just a cold dark place filled with sadness and
despair.

Repeats: 80e62bc8487b ("MAINTAINERS: re-sort all entries and fields")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agoMerge tag 'dma-mapping-6.5-2023-07-09' of git://git.infradead.org/users/hch/dma-mapping
Linus Torvalds [Sun, 9 Jul 2023 17:24:22 +0000 (10:24 -0700)]
Merge tag 'dma-mapping-6.5-2023-07-09' of git://git.infradead.org/users/hch/dma-mapping

Pull dma-mapping fixes from Christoph Hellwig:

 - swiotlb area sizing fixes (Petr Tesarik)

* tag 'dma-mapping-6.5-2023-07-09' of git://git.infradead.org/users/hch/dma-mapping:
  swiotlb: reduce the number of areas to match actual memory pool size
  swiotlb: always set the number of areas before allocating the pool

10 months agoMerge tag 'irq_urgent_for_v6.5_rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 9 Jul 2023 17:16:04 +0000 (10:16 -0700)]
Merge tag 'irq_urgent_for_v6.5_rc1' of git://git./linux/kernel/git/tip/tip

Pull irq update from Borislav Petkov:

 - Optimize IRQ domain's name assignment

* tag 'irq_urgent_for_v6.5_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  irqdomain: Use return value of strreplace()

10 months agoMerge tag 'x86_urgent_for_v6.5_rc1' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 9 Jul 2023 17:13:32 +0000 (10:13 -0700)]
Merge tag 'x86_urgent_for_v6.5_rc1' of git://git./linux/kernel/git/tip/tip

Pull x86 fpu fix from Borislav Petkov:

 - Do FPU AP initialization on Xen PV too which got missed by the recent
   boot reordering work

* tag 'x86_urgent_for_v6.5_rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/xen: Fix secondary processors' FPU initialization

10 months agoMerge tag 'x86-core-2023-07-09' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 9 Jul 2023 17:08:38 +0000 (10:08 -0700)]
Merge tag 'x86-core-2023-07-09' of git://git./linux/kernel/git/tip/tip

Pull x86 fix from Thomas Gleixner:
 "A single fix for the mechanism to park CPUs with an INIT IPI.

  On shutdown or kexec, the kernel tries to park the non-boot CPUs with
  an INIT IPI. But the same code path is also used by the crash utility.
  If the CPU which panics is not the boot CPU then it sends an INIT IPI
  to the boot CPU which resets the machine.

  Prevent this by validating that the CPU which runs the stop mechanism
  is the boot CPU. If not, leave the other CPUs in HLT"

* tag 'x86-core-2023-07-09' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/smp: Don't send INIT to boot CPU

10 months agoMerge tag 'mips_6.5_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux
Linus Torvalds [Sun, 9 Jul 2023 17:02:49 +0000 (10:02 -0700)]
Merge tag 'mips_6.5_1' of git://git./linux/kernel/git/mips/linux

Pull MIPS fixes from Thomas Bogendoerfer:

 - fixes for KVM

 - fix for loongson build and cpu probing

 - DT fixes

* tag 'mips_6.5_1' of git://git.kernel.org/pub/scm/linux/kernel/git/mips/linux:
  MIPS: kvm: Fix build error with KVM_MIPS_DEBUG_COP0_COUNTERS enabled
  MIPS: dts: add missing space before {
  MIPS: Loongson: Fix build error when make modules_install
  MIPS: KVM: Fix NULL pointer dereference
  MIPS: Loongson: Fix cpu_probe_loongson() again

10 months agoMerge tag 'xfs-6.5-merge-6' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux
Linus Torvalds [Sun, 9 Jul 2023 16:50:42 +0000 (09:50 -0700)]
Merge tag 'xfs-6.5-merge-6' of git://git./fs/xfs/xfs-linux

Pull xfs fix from Darrick Wong:
 "Nothing exciting here, just getting rid of a gcc warning that I got
  tired of seeing when I turn on gcov"

* tag 'xfs-6.5-merge-6' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: fix uninit warning in xfs_growfs_data

10 months agoMerge tag '6.5-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sun, 9 Jul 2023 16:45:32 +0000 (09:45 -0700)]
Merge tag '6.5-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6

Pull more smb client updates from Steve French:

 - fix potential use after free in unmount

 - minor cleanup

 - add worker to cleanup stale directory leases

* tag '6.5-rc-smb3-client-fixes-part2' of git://git.samba.org/sfrench/cifs-2.6:
  cifs: Add a laundromat thread for cached directories
  smb: client: remove redundant pointer 'server'
  cifs: fix session state transition to avoid use-after-free issue

10 months agoMerge tag 'ntb-6.5' of https://github.com/jonmason/ntb
Linus Torvalds [Sun, 9 Jul 2023 16:35:51 +0000 (09:35 -0700)]
Merge tag 'ntb-6.5' of https://github.com/jonmason/ntb

Pull NTB updates from Jon Mason:
 "Fixes for pci_clean_master, error handling in driver inits, and
  various other issues/bugs"

* tag 'ntb-6.5' of https://github.com/jonmason/ntb:
  ntb: hw: amd: Fix debugfs_create_dir error checking
  ntb.rst: Fix copy and paste error
  ntb_netdev: Fix module_init problem
  ntb: intel: Remove redundant pci_clear_master
  ntb: epf: Remove redundant pci_clear_master
  ntb_hw_amd: Remove redundant pci_clear_master
  ntb: idt: drop redundant pci_enable_pcie_error_reporting()
  MAINTAINERS: git://github -> https://github.com for jonmason
  NTB: EPF: fix possible memory leak in pci_vntb_probe()
  NTB: ntb_tool: Add check for devm_kcalloc
  NTB: ntb_transport: fix possible memory leak while device_register() fails
  ntb: intel: Fix error handling in intel_ntb_pci_driver_init()
  NTB: amd: Fix error handling in amd_ntb_pci_driver_init()
  ntb: idt: Fix error handling in idt_pci_driver_init()

10 months agomm: lock newly mapped VMA with corrected ordering
Hugh Dickins [Sat, 8 Jul 2023 23:04:00 +0000 (16:04 -0700)]
mm: lock newly mapped VMA with corrected ordering

Lockdep is certainly right to complain about

  (&vma->vm_lock->lock){++++}-{3:3}, at: vma_start_write+0x2d/0x3f
                 but task is already holding lock:
  (&mapping->i_mmap_rwsem){+.+.}-{3:3}, at: mmap_region+0x4dc/0x6db

Invert those to the usual ordering.

Fixes: 33313a747e81 ("mm: lock newly mapped VMA which can be modified after it becomes visible")
Cc: stable@vger.kernel.org
Signed-off-by: Hugh Dickins <hughd@google.com>
Tested-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agoMerge tag 'mm-hotfixes-stable-2023-07-08-10-43' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sat, 8 Jul 2023 21:30:25 +0000 (14:30 -0700)]
Merge tag 'mm-hotfixes-stable-2023-07-08-10-43' of git://git./linux/kernel/git/akpm/mm

Pull hotfixes from Andrew Morton:
 "16 hotfixes. Six are cc:stable and the remainder address post-6.4
  issues"

The merge undoes the disabling of the CONFIG_PER_VMA_LOCK feature, since
it was all hopefully fixed in mainline.

* tag 'mm-hotfixes-stable-2023-07-08-10-43' of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm:
  lib: dhry: fix sleeping allocations inside non-preemptable section
  kasan, slub: fix HW_TAGS zeroing with slub_debug
  kasan: fix type cast in memory_is_poisoned_n
  mailmap: add entries for Heiko Stuebner
  mailmap: update manpage link
  bootmem: remove the vmemmap pages from kmemleak in free_bootmem_page
  MAINTAINERS: add linux-next info
  mailmap: add Markus Schneider-Pargmann
  writeback: account the number of pages written back
  mm: call arch_swap_restore() from do_swap_page()
  squashfs: fix cache race with migration
  mm/hugetlb.c: fix a bug within a BUG(): inconsistent pte comparison
  docs: update ocfs2-devel mailing list address
  MAINTAINERS: update ocfs2-devel mailing list address
  mm: disable CONFIG_PER_VMA_LOCK until its fixed
  fork: lock VMAs of the parent process when forking

10 months agofork: lock VMAs of the parent process when forking
Suren Baghdasaryan [Sat, 8 Jul 2023 19:12:12 +0000 (12:12 -0700)]
fork: lock VMAs of the parent process when forking

When forking a child process, the parent write-protects anonymous pages
and COW-shares them with the child being forked using copy_present_pte().

We must not take any concurrent page faults on the source vma's as they
are being processed, as we expect both the vma and the pte's behind it
to be stable.  For example, the anon_vma_fork() expects the parents
vma->anon_vma to not change during the vma copy.

A concurrent page fault on a page newly marked read-only by the page
copy might trigger wp_page_copy() and a anon_vma_prepare(vma) on the
source vma, defeating the anon_vma_clone() that wasn't done because the
parent vma originally didn't have an anon_vma, but we now might end up
copying a pte entry for a page that has one.

Before the per-vma lock based changes, the mmap_lock guaranteed
exclusion with concurrent page faults.  But now we need to do a
vma_start_write() to make sure no concurrent faults happen on this vma
while it is being processed.

This fix can potentially regress some fork-heavy workloads.  Kernel
build time did not show noticeable regression on a 56-core machine while
a stress test mapping 10000 VMAs and forking 5000 times in a tight loop
shows ~5% regression.  If such fork time regression is unacceptable,
disabling CONFIG_PER_VMA_LOCK should restore its performance.  Further
optimizations are possible if this regression proves to be problematic.

Suggested-by: David Hildenbrand <david@redhat.com>
Reported-by: Jiri Slaby <jirislaby@kernel.org>
Closes: https://lore.kernel.org/all/dbdef34c-3a07-5951-e1ae-e9c6e3cdf51b@kernel.org/
Reported-by: Holger Hoffstätte <holger@applied-asynchrony.com>
Closes: https://lore.kernel.org/all/b198d649-f4bf-b971-31d0-e8433ec2a34c@applied-asynchrony.com/
Reported-by: Jacob Young <jacobly.alt@gmail.com>
Closes: https://bugzilla.kernel.org/show_bug.cgi?id=217624
Fixes: 0bff0aaea03e ("x86/mm: try VMA lock-based page fault handling first")
Cc: stable@vger.kernel.org
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm: lock newly mapped VMA which can be modified after it becomes visible
Suren Baghdasaryan [Sat, 8 Jul 2023 19:12:11 +0000 (12:12 -0700)]
mm: lock newly mapped VMA which can be modified after it becomes visible

mmap_region adds a newly created VMA into VMA tree and might modify it
afterwards before dropping the mmap_lock.  This poses a problem for page
faults handled under per-VMA locks because they don't take the mmap_lock
and can stumble on this VMA while it's still being modified.  Currently
this does not pose a problem since post-addition modifications are done
only for file-backed VMAs, which are not handled under per-VMA lock.
However, once support for handling file-backed page faults with per-VMA
locks is added, this will become a race.

Fix this by write-locking the VMA before inserting it into the VMA tree.
Other places where a new VMA is added into VMA tree do not modify it
after the insertion, so do not need the same locking.

Cc: stable@vger.kernel.org
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agomm: lock a vma before stack expansion
Suren Baghdasaryan [Sat, 8 Jul 2023 19:12:10 +0000 (12:12 -0700)]
mm: lock a vma before stack expansion

With recent changes necessitating mmap_lock to be held for write while
expanding a stack, per-VMA locks should follow the same rules and be
write-locked to prevent page faults into the VMA being expanded. Add
the necessary locking.

Cc: stable@vger.kernel.org
Signed-off-by: Suren Baghdasaryan <surenb@google.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
10 months agoMerge tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Linus Torvalds [Sat, 8 Jul 2023 19:35:18 +0000 (12:35 -0700)]
Merge tag 'scsi-misc' of git://git./linux/kernel/git/jejb/scsi

Pull more SCSI updates from James Bottomley:
 "A few late arriving patches that missed the initial pull request. It's
  mostly bug fixes (the dt-bindings is a fix for the initial pull)"

* tag 'scsi-misc' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
  scsi: ufs: core: Remove unused function declaration
  scsi: target: docs: Remove tcm_mod_builder.py
  scsi: target: iblock: Quiet bool conversion warning with pr_preempt use
  scsi: dt-bindings: ufs: qcom: Fix ICE phandle
  scsi: core: Simplify scsi_cdl_check_cmd()
  scsi: isci: Fix comment typo
  scsi: smartpqi: Replace one-element arrays with flexible-array members
  scsi: target: tcmu: Replace strlcpy() with strscpy()
  scsi: ncr53c8xx: Replace strlcpy() with strscpy()
  scsi: lpfc: Fix lpfc_name struct packing

10 months agoMerge tag 'i2c-for-6.5-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 8 Jul 2023 19:28:00 +0000 (12:28 -0700)]
Merge tag 'i2c-for-6.5-rc1-part2' of git://git./linux/kernel/git/wsa/linux

Pull more i2c updates from Wolfram Sang:

 - xiic patch should have been in the original pull but slipped through

 - mpc patch fixes a build regression

 - nomadik cleanup

* tag 'i2c-for-6.5-rc1-part2' of git://git.kernel.org/pub/scm/linux/kernel/git/wsa/linux:
  i2c: mpc: Drop unused variable
  i2c: nomadik: Remove a useless call in the remove function
  i2c: xiic: Don't try to handle more interrupt events after error

10 months agoMerge tag 'hardening-v6.5-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sat, 8 Jul 2023 19:08:39 +0000 (12:08 -0700)]
Merge tag 'hardening-v6.5-rc1-fixes' of git://git./linux/kernel/git/kees/linux

Pull hardening fixes from Kees Cook:

 - Check for NULL bdev in LoadPin (Matthias Kaehlcke)

 - Revert unwanted KUnit FORTIFY build default

 - Fix 1-element array causing boot warnings with xhci-hub

* tag 'hardening-v6.5-rc1-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux:
  usb: ch9: Replace bmSublinkSpeedAttr 1-element array with flexible array
  Revert "fortify: Allow KUnit test to build without FORTIFY"
  dm: verity-loadpin: Add NULL pointer check for 'bdev' parameter

10 months agontb: hw: amd: Fix debugfs_create_dir error checking
Anup Sharma [Fri, 12 May 2023 20:24:34 +0000 (01:54 +0530)]
ntb: hw: amd: Fix debugfs_create_dir error checking

The debugfs_create_dir function returns ERR_PTR in case of error, and the
only correct way to check if an error occurred is 'IS_ERR' inline function.
This patch will replace the null-comparison with IS_ERR.

Signed-off-by: Anup Sharma <anupnewsmail@gmail.com>
Suggested-by: Ivan Orlov <ivan.orlov0322@gmail.com>
Signed-off-by: Jon Mason <jdmason@kudzu.us>
10 months agoMerge tag 'perf-tools-for-v6.5-2-2023-07-06' of git://git.kernel.org/pub/scm/linux...
Linus Torvalds [Sat, 8 Jul 2023 17:21:51 +0000 (10:21 -0700)]
Merge tag 'perf-tools-for-v6.5-2-2023-07-06' of git://git./linux/kernel/git/perf/perf-tools-next

Pull more perf tools updates from Namhyung Kim:
 "These are remaining changes and fixes for this cycle.

  Build:

   - Allow generating vmlinux.h from BTF using `make GEN_VMLINUX_H=1`
     and skip if the vmlinux has no BTF.

   - Replace deprecated clang -target xxx option by --target=xxx.

  perf record:

   - Print event attributes with well known type and config symbols in
     the debug output like below:

       # perf record -e cycles,cpu-clock -C0 -vv true
       <SNIP>
       ------------------------------------------------------------
       perf_event_attr:
         type                             0 (PERF_TYPE_HARDWARE)
         size                             136
         config                           0 (PERF_COUNT_HW_CPU_CYCLES)
         { sample_period, sample_freq }   4000
         sample_type                      IP|TID|TIME|CPU|PERIOD|IDENTIFIER
         read_format                      ID
         disabled                         1
         inherit                          1
         freq                             1
         sample_id_all                    1
         exclude_guest                    1
       ------------------------------------------------------------
       sys_perf_event_open: pid -1  cpu 0  group_fd -1  flags 0x8 = 5
       ------------------------------------------------------------
       perf_event_attr:
         type                             1 (PERF_TYPE_SOFTWARE)
         size                             136
         config                           0 (PERF_COUNT_SW_CPU_CLOCK)
         { sample_period, sample_freq }   4000
         sample_type                      IP|TID|TIME|CPU|PERIOD|IDENTIFIER
         read_format                      ID
         disabled                         1
         inherit                          1
         freq                             1
         sample_id_all                    1
         exclude_guest                    1

   - Update AMD IBS event error message since it now support per-process
     profiling but no priviledge filters.

       $ sudo perf record -e ibs_op//k -C 0
       Error:
       AMD IBS doesn't support privilege filtering. Try again without
       the privilege modifiers (like 'k') at the end.

  perf lock contention:

   - Support CSV style output using -x option

       $ sudo perf lock con -ab -x, sleep 1
       # output: contended, total wait, max wait, avg wait, type, caller
       19, 194232, 21415, 10222, spinlock, process_one_work+0x1f0
       15, 162748, 23843, 10849, rwsem:R, do_user_addr_fault+0x40e
       4, 86740, 23415, 21685, rwlock:R, ep_poll_callback+0x2d
       1, 84281, 84281, 84281, mutex, iwl_mvm_async_handlers_wk+0x135
       8, 67608, 27404, 8451, spinlock, __queue_work+0x174
       3, 58616, 31125, 19538, rwsem:W, do_mprotect_pkey+0xff
       3, 52953, 21172, 17651, rwlock:W, do_epoll_wait+0x248
       2, 30324, 19704, 15162, rwsem:R, do_madvise+0x3ad
       1, 24619, 24619, 24619, spinlock, rcu_core+0xd4

   - Add --output option to save the data to a file not to be interfered
     by other debug messages.

  Test:

   - Fix event parsing test on ARM where there's no raw PMU nor supports
     PERF_PMU_CAP_EXTENDED_HW_TYPE.

   - Update the lock contention test case for CSV output.

   - Fix a segfault in the daemon command test.

  Vendor events (JSON):

   - Add has_event() to check if the given event is available on system
     at runtime. On Intel machines, some transaction events may not be
     present when TSC extensions are disabled.

   - Update Intel event metrics.

  Misc:

   - Sort symbols by name using an external array of pointers instead of
     a rbtree node in the symbol. This will save 16-bytes or 24-bytes
     per symbol whether the sorting is actually requested or not.

   - Fix unwinding DWARF callstacks using libdw when --symfs option is
     used"

* tag 'perf-tools-for-v6.5-2-2023-07-06' of git://git.kernel.org/pub/scm/linux/kernel/git/perf/perf-tools-next: (38 commits)
  perf test: Fix event parsing test when PERF_PMU_CAP_EXTENDED_HW_TYPE isn't supported.
  perf test: Fix event parsing test on Arm
  perf evsel amd: Fix IBS error message
  perf: unwind: Fix symfs with libdw
  perf symbol: Fix uninitialized return value in symbols__find_by_name()
  perf test: Test perf lock contention CSV output
  perf lock contention: Add --output option
  perf lock contention: Add -x option for CSV style output
  perf lock: Remove stale comments
  perf vendor events intel: Update tigerlake to 1.13
  perf vendor events intel: Update skylakex to 1.31
  perf vendor events intel: Update skylake to 57
  perf vendor events intel: Update sapphirerapids to 1.14
  perf vendor events intel: Update icelakex to 1.21
  perf vendor events intel: Update icelake to 1.19
  perf vendor events intel: Update cascadelakex to 1.19
  perf vendor events intel: Update meteorlake to 1.03
  perf vendor events intel: Add rocketlake events/metrics
  perf vendor metrics intel: Make transaction metrics conditional
  perf jevents: Support for has_event function
  ...

10 months agoMerge tag 'bitmap-6.5-rc1' of https://github.com/norov/linux
Linus Torvalds [Sat, 8 Jul 2023 17:02:24 +0000 (10:02 -0700)]
Merge tag 'bitmap-6.5-rc1' of https://github.com/norov/linux

Pull bitmap updates from Yury Norov:
 "Fixes for different bitmap pieces:

   - lib/test_bitmap: increment failure counter properly

     The tests that don't use expect_eq() macro to determine that a test
     is failured must increment failed_tests explicitly.

   - lib/bitmap: drop optimization of bitmap_{from,to}_arr64

     bitmap_{from,to}_arr64() optimization is overly optimistic
     on 32-bit LE architectures when it's wired to
     bitmap_copy_clear_tail().

   - nodemask: Drop duplicate check in for_each_node_mask()

     As the return value type of first_node() became unsigned, the node
     >= 0 became unnecessary.

   - cpumask: fix function description kernel-doc notation

   - MAINTAINERS: Add bits.h and bitfield.h to the BITMAP API record

     Add linux/bits.h and linux/bitfield.h for visibility"

* tag 'bitmap-6.5-rc1' of https://github.com/norov/linux:
  MAINTAINERS: Add bitfield.h to the BITMAP API record
  MAINTAINERS: Add bits.h to the BITMAP API record
  cpumask: fix function description kernel-doc notation
  nodemask: Drop duplicate check in for_each_node_mask()
  lib/bitmap: drop optimization of bitmap_{from,to}_arr64
  lib/test_bitmap: increment failure counter properly

10 months agolib: dhry: fix sleeping allocations inside non-preemptable section
Geert Uytterhoeven [Wed, 5 Jul 2023 14:54:04 +0000 (16:54 +0200)]
lib: dhry: fix sleeping allocations inside non-preemptable section

The Smatch static checker reports the following warnings:

    lib/dhry_run.c:38 dhry_benchmark() warn: sleeping in atomic context
    lib/dhry_run.c:43 dhry_benchmark() warn: sleeping in atomic context

Indeed, dhry() does sleeping allocations inside the non-preemptable
section delimited by get_cpu()/put_cpu().

Fix this by using atomic allocations instead.
Add error handling, as atomic these allocations may fail.

Link: https://lkml.kernel.org/r/bac6d517818a7cd8efe217c1ad649fffab9cc371.1688568764.git.geert+renesas@glider.be
Fixes: 13684e966d46283e ("lib: dhry: fix unstable smp_processor_id(_) usage")
Reported-by: Dan Carpenter <dan.carpenter@linaro.org>
Closes: https://lore.kernel.org/r/0469eb3a-02eb-4b41-b189-de20b931fa56@moroto.mountain
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
10 months agokasan, slub: fix HW_TAGS zeroing with slub_debug
Andrey Konovalov [Wed, 5 Jul 2023 12:44:02 +0000 (14:44 +0200)]
kasan, slub: fix HW_TAGS zeroing with slub_debug

Commit 946fa0dbf2d8 ("mm/slub: extend redzone check to extra allocated
kmalloc space than requested") added precise kmalloc redzone poisoning to
the slub_debug functionality.

However, this commit didn't account for HW_TAGS KASAN fully initializing
the object via its built-in memory initialization feature.  Even though
HW_TAGS KASAN memory initialization contains special memory initialization
handling for when slub_debug is enabled, it does not account for in-object
slub_debug redzones.  As a result, HW_TAGS KASAN can overwrite these
redzones and cause false-positive slub_debug reports.

To fix the issue, avoid HW_TAGS KASAN memory initialization when
slub_debug is enabled altogether.  Implement this by moving the
__slub_debug_enabled check to slab_post_alloc_hook.  Common slab code
seems like a more appropriate place for a slub_debug check anyway.

Link: https://lkml.kernel.org/r/678ac92ab790dba9198f9ca14f405651b97c8502.1688561016.git.andreyknvl@google.com
Fixes: 946fa0dbf2d8 ("mm/slub: extend redzone check to extra allocated kmalloc space than requested")
Signed-off-by: Andrey Konovalov <andreyknvl@google.com>
Reported-by: Will Deacon <will@kernel.org>
Acked-by: Marco Elver <elver@google.com>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrey Ryabinin <ryabinin.a.a@gmail.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: David Rientjes <rientjes@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Feng Tang <feng.tang@intel.com>
Cc: Hyeonggon Yoo <42.hyeyoo@gmail.com>
Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: kasan-dev@googlegroups.com
Cc: Pekka Enberg <penberg@kernel.org>
Cc: Peter Collingbourne <pcc@google.com>
Cc: Roman Gushchin <roman.gushchin@linux.dev>
Cc: Vincenzo Frascino <vincenzo.frascino@arm.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>