kernel/linux.git/tools/include/nolibc, branch v6.18.21

tools/nolibc/dirent: avoid errno in readdir_r

2025-12-18T13:02:43+00:00

[ Upstream commit 4ada5679f18dbbe92d87c37a842c3368e6ab5e4a ] Using errno is not possible when NOLIBC_IGNORE_ERRNO is set. Use sys_lseek instead of lseek as that avoids using errno. Fixes: 665fa8dea90d ("tools/nolibc: add support for directory access") Signed-off-by: Benjamin Berg Signed-off-by: Thomas Weißschuh Signed-off-by: Sasha Levin

tools/nolibc/stdio: let perror work when NOLIBC_IGNORE_ERRNO is set

2025-12-18T13:02:43+00:00

[ Upstream commit c485ca3aff2442adea4c08ceb5183e671ebed22a ] There is no errno variable when NOLIBC_IGNORE_ERRNO is defined. As such, simply print the message with "unknown error" rather than the integer value of errno. Fixes: acab7bcdb1bc ("tools/nolibc/stdio: add perror() to report the errno value") Signed-off-by: Benjamin Berg Signed-off-by: Thomas Weißschuh Signed-off-by: Sasha Levin

tools/nolibc: x86: fix section mismatch caused by asm "mem*" functions

2025-12-18T13:02:39+00:00

[ Upstream commit 2602949b22330f1275138e2b5aea5d49126b9757 ] I recently got occasional build failures at -Os or -Oz that would always involve waitpid(), where the assembler would complain about this: init.s: Error: .size expression for waitpid.constprop.0 does not evaluate to a constant And without -fno-asynchronous-unwind-tables it could also spit such errors: init.s:836: Error: CFI instruction used without previous .cfi_startproc init.s:838: Error: .cfi_endproc without corresponding .cfi_startproc init.s: Error: open CFI at the end of file; missing .cfi_endproc directive A trimmed down reproducer is as simple as this: int main(int argc, char **argv) { int ret, status; if (argc == 0) ret = waitpid(-1, &status, 0); else ret = waitpid(-1, &status, 0); return status; } It produces the following asm code on x86_64: .text .section .text.nolibc_memmove_memcpy .weak memmove .weak memcpy memmove: memcpy: movq %rdx, %rcx (...) retq .section .text.nolibc_memset .weak memset memset: xchgl %eax, %esi movq %rdx, %rcx pushq %rdi rep stosb popq %rax retq .type waitpid.constprop.0.isra.0, @function waitpid.constprop.0.isra.0: subq $8, %rsp (...) jmp *.L5(,%rax,8) .section .rodata .align 8 .align 4 .L5: .quad .L10 (...) .quad .L4 .text .L10: (...) .cfi_def_cfa_offset 8 ret .cfi_endproc .LFE273: .size waitpid.constprop.0.isra.0, .-waitpid.constprop.0.isra.0 It's a bit dense, but here's the explanation: the compiler has emitted a ".text" statement because it knows it's working in the .text section. Then, our hand-written asm code for the mem* functions forced the section to .text.something without the compiler knowing about it, so it thinks the code is still being emitted for .text. As such, without any .section statement, the waitpid.constprop.0.isra.0 label is in fact placed in the previously created section, here .text.nolibc_memset. The waitpid() function involves a switch/case statement that can be turned to a jump table, which is what the compiler does with the .rodata section, and after that it restores .text, which is no longer the previous .text.nolibc_memset section. Then the CFI statements cross a section, so does the .size calculation, which explains the error. While a first approach consisting in placing an explicit ".text" at the end of these functions was verified to work, it's still unreliable as it depends on what the compiler remembers having emitted previously. A better approach is to replace the ".section" with ".pushsection", and place a ".popsection" at the end, so that these code blocks are agnostic to where they're placed relative to other blocks. Fixes: 553845eebd60 ("tools/nolibc: x86-64: Use `rep movsb` for `memcpy()` and `memmove()`") Fixes: 12108aa8c1a1 ("tools/nolibc: x86-64: Use `rep stosb` for `memset()`") Signed-off-by: Willy Tarreau Signed-off-by: Thomas Weißschuh Signed-off-by: Sasha Levin

tools/nolibc: handle NULL wstatus argument to waitpid()

2025-12-18T13:02:27+00:00

[ Upstream commit 812f223fe9be03dc22abb85240b6f075135d2386 ] wstatus is allowed to be NULL. Avoid a segmentation fault in this case. Fixes: 0c89abf5ab3f ("tools/nolibc: implement waitpid() in terms of waitid()") Signed-off-by: Thomas Weißschuh Acked-by: Willy Tarreau Signed-off-by: Thomas Weißschuh Signed-off-by: Sasha Levin

tools/nolibc: add stdbool.h to nolibc includes

2025-09-25T16:47:08+00:00

Otherwise tests compiled with only "-include nolibc.h" will fail with "error: unknown type name 'bool'", even though a stdbool.h is available from nolibc. Fixes: ae1f550efc11 ("tools/nolibc: add stdbool.h header") Fixes: f2662ec26b26 ("selftests: kselftest: Create ksft_print_dbg_msg()") Reported-by: Mark Brown Closes: https://lore.kernel.org/lkml/833f5ae5-190e-47ec-9ad9-127ad166c80c@sirena.org.uk/ Signed-off-by: André Almeida [Thomas: add Fixes tags and massage commit message a bit] Signed-off-by: Thomas Weißschuh

tools/nolibc: make time_t robust if __kernel_old_time_t is missing in host headers

2025-09-20T09:06:37+00:00

Commit d5094bcb5bfd ("tools/nolibc: define time_t in terms of __kernel_old_time_t") made nolibc use the kernel's time type so that `time_t` matches `timespec::tv_sec` on all ABIs (notably x32). But since __kernel_old_time_t is fairly new, notably from 2020 in commit 94c467ddb273 ("y2038: add __kernel_old_timespec and __kernel_old_time_t"), nolibc builds that rely on host headers may fail. Switch to __kernel_time_t, which is the same as __kernel_old_time_t and has existed for longer. Tested in PPC VM of Open Source Lab of Oregon State University (./tools/testing/selftests/rcutorture/bin/mkinitrd.sh) Fixes: d5094bcb5bfd ("tools/nolibc: define time_t in terms of __kernel_old_time_t") Signed-off-by: Zhouyi Zhou [Thomas: Reformat commit and its message a bit] Signed-off-by: Thomas Weißschuh

tools/nolibc: drop wait4() support

2025-09-01T18:48:40+00:00

Not all architectures implement the wait4() syscall. It can be implemented in terms of the waitid() syscall, but that would require some rework of the other wait-related functions in wait.h. As wait4() is non-standard and deprecated, remove it. Signed-off-by: Thomas Weißschuh Acked-by: Willy Tarreau Link: https://lore.kernel.org/r/20250821-nolibc-enosys-v1-7-4b63f2caaa89@weissschuh.net

tools/nolibc: fold llseek fallback into lseek()

2025-09-01T18:47:54+00:00

Align the implementation of the fallback handling inside sys_lseek() with the rest of nolibc. Signed-off-by: Thomas Weißschuh Acked-by: Willy Tarreau Link: https://lore.kernel.org/r/20250821-nolibc-enosys-v1-5-4b63f2caaa89@weissschuh.net

tools/nolibc: remove __nolibc_enosys() fallback from fork functions

2025-09-01T18:47:54+00:00

All architectures have one of the real functions available. The additional fallback to __nolibc_enosys() is superfluous. Signed-off-by: Thomas Weißschuh Acked-by: Willy Tarreau Link: https://lore.kernel.org/r/20250821-nolibc-enosys-v1-4-4b63f2caaa89@weissschuh.net

tools/nolibc: remove __nolibc_enosys() fallback from dup2()

2025-09-01T18:47:53+00:00