linux: add SYS_futex for riscv32 GNU

[jadhavgaurav]

Description

Add SYS_futex for riscv32gc-unknown-linux-gnu as an alias of SYS_futex_time64, matching the riscv32 musl module. #5455 removed the time32 SYS_futex = 98, which does not exist on riscv32; this does not bring that number back. parking_lot_core uses libc::SYS_futex and no longer builds on this target. time_t is 64-bit on riscv32, so the time64 syscall is the one a SYS_futex caller needs.

Fixes #5627

Checklist

• Relevant tests in libc-test/semver have been updated (SYS_futex is already listed in linux.txt, no change needed)
• Commit messages permalink to headers for added or changed API
• Placeholder or unstable values like *LAST or *MAX have the standard doc comment (not applicable)
• Tested locally (cargo test -p libc-test --target mytarget): not run, only cargo check -Zbuild-std=core --target riscv32gc-unknown-linux-gnu and cargo fmt --check on nightly

@rustbot label +stable-nominated

[tgross35]

The links you posted don't back this change up. https://github.com/torvalds/linux/blob/8cd9520d35a6c38db6567e97dd93b1f11f185dc6/include/uapi/asm-generic/unistd.h#L269-L272 is gated behind __ARCH_WANT_TIME32_SYSCALLS, which isn't set for rv32, and there is no __NR_futex at https://github.com/sailfishos-mirror/glibc/blob/decb8b56cc99ff1326a577b17bca96fd9cecc955/sysdeps/unix/sysv/linux/riscv/rv32/arch-syscall.h.

[tgross35]

@rustbot author for feedback

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.

[jadhavgaurav]

Thanks, you are right. The glibc rv32 arch-syscall.h I cited only defines __NR_futex_time64 (422) and has no __NR_futex, and in Linux's asm-generic unistd.h __NR_futex is gated behind __ARCH_WANT_TIME32_SYSCALLS, which rv32 does not set. The SYS_futex alias exists in musl's own headers, but the gnu module follows glibc, so the alias is not backed there. Closing this one. Sorry for the noise.