Error in obtaining time on ‘arm-unknown-linux-musleabihf’ platform

[vnt-dev]

fn main() {
    let now = std::time::SystemTime::now();
    println!("now={:?}",now);
}

rustc --version --verbose:

rustc 1.72.0-nightly (f4b80cacf 2023-06-30)
binary: rustc
commit-hash: f4b80cacf93ca216c75f6ae12f4b9dec19eba42f
commit-date: 2023-06-30
host: x86_64-unknown-linux-gnu
release: 1.72.0-nightly
LLVM version: 16.0.5

build

cargo build --target arm-unknown-linux-musleabihf
[target.arm-unknown-linux-musleabihf]
linker = "arm-linux-musleabihf-gcc" # https://musl.cc/arm-linux-musleabihf-cross.tgz

os

uname -a
================================================================
Linux LTBK 3.2.40 #15284 Sat May 19 04:36:51 CST 2018 armv7l GNU/Linux synology_armada370_213j

cat /proc/cpuinfo
================================================================
Processor	: Marvell PJ4Bv7 Processor rev 1 (v7l)
BogoMIPS	: 1196.85
Features	: swp half thumb fastmult vfp edsp vfpv3 vfpv3d16 tls 
CPU implementer	: 0x56
CPU architecture: 7
CPU variant	: 0x1
CPU part	: 0x581
CPU revision	: 1

Hardware	: Marvell Armada-370
Revision	: 0000
Serial		: 0000000000000000

[saethlin]

Is there any chance you can run the executable with strace ./arm_time_test?

[vnt-dev]

strace ./arm_time_test
=====================================
execve("./arm_time_test", ["./arm_time_test"], [/* 19 vars */]) = 0
mmap2(NULL, 328, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x4009b000
set_tls(0x4009b070, 0xaf814, 0x37, 0xbebcfe1e, 0x4009b000) = 0
set_tid_address(0xb3904)                = 14999
poll([{fd=0, events=0}, {fd=1, events=0}, {fd=2, events=0}], 3, 0) = 0 (Timeout)
rt_sigaction(SIGPIPE, {SIG_IGN, [], SA_RESTORER|SA_RESTART, 0x89490}, {SIG_DFL, [], 0}, 8) = 0
rt_sigaction(SIGSEGV, NULL, {SIG_DFL, [], 0}, 8) = 0
rt_sigprocmask(SIG_UNBLOCK, [RT_1 RT_2], NULL, 8) = 0
rt_sigaction(SIGSEGV, {0x65f9c, [], SA_RESTORER|SA_STACK|SA_SIGINFO, 0x89498}, NULL, 8) = 0
rt_sigaction(SIGBUS, NULL, {SIG_DFL, [], 0}, 8) = 0
rt_sigaction(SIGBUS, {0x65f9c, [], SA_RESTORER|SA_STACK|SA_SIGINFO, 0x89498}, NULL, 8) = 0
sigaltstack(NULL, {ss_sp=0, ss_flags=SS_DISABLE, ss_size=0}) = 0
mmap2(NULL, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS|MAP_STACK, -1, 0) = 0x4016e000
mprotect(0x4016e000, 4096, PROT_NONE)   = 0
sigaltstack({ss_sp=0x4016f000, ss_flags=0, ss_size=8192}, NULL) = 0
brk(0)                                  = 0x1d8000
brk(0x1da000)                           = 0x1da000
mmap2(0x1d8000, 4096, PROT_NONE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x1d8000
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40204000
rt_sigprocmask(SIG_BLOCK, ~[RTMIN RT_1 RT_2], [], 8) = 0
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
syscall_403(0, 0xbebcfa78, 0, 0, 0xbebcfaa8, 0xb3134) = -1 (errno 22)
write(2, "thread '", 8thread ')                 = 8
write(2, "main", 4main)                     = 4
write(2, "' panicked at '", 15' panicked at ')         = 15
write(2, "called `Result::unwrap()` on an "..., 109called `Result::unwrap()` on an `Err` value: Os { code: 22, kind: InvalidInput, message: "Invalid argument" }) = 109
write(2, "', ", 3', )                      = 3
write(2, "library/std/src/sys/unix/time.rs", 32library/std/src/sys/unix/time.rs) = 32
write(2, ":", 1:)                        = 1
write(2, "420", 3420)                      = 3
write(2, ":", 1:)                        = 1
write(2, "72", 272)                       = 2
write(2, "\n", 1
)                       = 1
write(2, "note: run with `RUST_BACKTRACE=1"..., 78note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
) = 78
sigaltstack({ss_sp=0, ss_flags=SS_DISABLE, ss_size=8192}, NULL) = 0
munmap(0x4016e000, 12288)               = 0
exit_group(101)                         = ?
+++ exited with 101 +++

[vnt-dev]

rust/library/std/src/sys/unix/time.rs

Line 420 in df99bc1

cvt(unsafe { libc::clock_gettime(clock, t.as_mut_ptr()) }).unwrap();

'unwrap' should not be used

[Fulgen301]

unwrap is what's causing the panic, not what's causing clock_gettime to fail with EINVAL, which means that the specified clock is invalid or not supported on that system. SystemTime::now() is using CLOCK_REALTIME, and the man page for that system call says

All implementations support the system-wide real-time clock, which is identified by CLOCK_REALTIME.

[vnt-dev]

This is how I solve this problem.If there were no panic, it wouldn't be so troublesome.This is also a problem with rust robustness

    #[cfg(all(target_arch = "arm", target_os = "linux",target_pointer_width = "32"))]
    {
        let mut t = std::mem::MaybeUninit::uninit();
        let time = unsafe { libc::clock_gettime(libc::CLOCK_REALTIME, t.as_mut_ptr()) };
        if time <= 0 {
            let output = std::process::Command::new("date")
                .arg("+%s%3N")
                .output()
                .expect("'date'  not found");
            if output.status.success() {
                let date_output = String::from_utf8(output.stdout).unwrap();
                let milliseconds = date_output.trim().parse::<u64>().unwrap();
                return milliseconds;
            }
        }
    }

[Fulgen301]

syscall_403

This is the syscall number for clock_gettime64 on ARM, which was added with Linux 5.1 (as opposed to clock_gettime, which is syscall number 263), and since you're running kernel 3.2.40, that syscall doesn't exist (and that old a version of strace doesn't know it either).

Looks like musl switched over to 64bit times a while ago, so you'd have to build musl without the redirection to the 64bit syscalls.

[vnt-dev]

Sorry, I didn't understand what you meant. Can you explain how to do it more specifically

[vnt-dev]

The 'date' command can obtain the time

root@LTBK:/# strace date

execve("/bin/date", ["date"], [/* 18 vars */]) = 0
brk(0)                                  = 0x1c6f000
uname({sys="Linux", node="LTBK", ...})  = 0
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40004000
access("/etc/ld.so.preload", R_OK)      = -1 ENOENT (No such file or directory)
open("/etc/ld.so.cache", O_RDONLY|O_CLOEXEC) = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=20860, ...}) = 0
mmap2(NULL, 20860, PROT_READ, MAP_PRIVATE, 3, 0) = 0x4008c000
close(3)                                = 0
access("/etc/ld.so.nohwcap", F_OK)      = -1 ENOENT (No such file or directory)
open("/lib/arm-linux-gnueabi/libc.so.6", O_RDONLY|O_CLOEXEC) = 3
read(3, "\177ELF\1\1\1\0\0\0\0\0\0\0\0\0\3\0(\0\1\0\0\0\230\204\1\0004\0\0\0"..., 512) = 512
lseek(3, 1272700, SEEK_SET)             = 1272700
read(3, "\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0"..., 2840) = 2840
lseek(3, 1269268, SEEK_SET)             = 1269268
read(3, "A)\0\0\0aeabi\0\1\37\0\0\0\0054T\0\6\2\10\1\t\1\22\4\23\1\24\1"..., 42) = 42
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40167000
fstat64(3, {st_mode=S_IFREG|0755, st_size=1275540, ...}) = 0
mmap2(NULL, 1344920, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_DENYWRITE, 3, 0) = 0x4025c000
mprotect(0x40390000, 61440, PROT_NONE)  = 0
mmap2(0x4039f000, 12288, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_DENYWRITE, 3, 0x133000) = 0x4039f000
mmap2(0x403a2000, 9624, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_FIXED|MAP_ANONYMOUS, -1, 0) = 0x403a2000
close(3)                                = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x400b7000
set_tls(0x400b74c0, 0x400b7ba8, 0x4011a050, 0x400b74c0, 0x4011a050) = 0
mprotect(0x4039f000, 8192, PROT_READ)   = 0
mprotect(0x2b000, 4096, PROT_READ)      = 0
mprotect(0x40119000, 4096, PROT_READ)   = 0
munmap(0x4008c000, 20860)               = 0
brk(0)                                  = 0x1c6f000
brk(0x1c90000)                          = 0x1c90000
open("/usr/lib/locale/locale-archive", O_RDONLY|O_LARGEFILE|O_CLOEXEC) = 3
fstat64(3, {st_mode=S_IFREG|0644, st_size=3324208, ...}) = 0
mmap2(NULL, 2097152, PROT_READ, MAP_PRIVATE, 3, 0) = 0x403a5000
close(3)                                = 0
clock_gettime(CLOCK_REALTIME, {1695031207, 548680087}) = 0
open("/etc/localtime", O_RDONLY|O_CLOEXEC) = 3
fcntl64(3, F_GETFD)                     = 0x1 (flags FD_CLOEXEC)
fstat64(3, {st_mode=S_IFREG|0644, st_size=118, ...}) = 0
fstat64(3, {st_mode=S_IFREG|0644, st_size=118, ...}) = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x40090000
read(3, "TZif2\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\0\1\0\0\0\1\0\0\0\0"..., 4096) = 118
_llseek(3, -6, [112], SEEK_CUR)         = 0
read(3, "\nUTC0\n", 4096)               = 6
close(3)                                = 0
munmap(0x40090000, 4096)                = 0
fstat64(1, {st_mode=S_IFCHR|0600, st_rdev=makedev(136, 5), ...}) = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) = 0x4010a000
write(1, "Mon Sep 18 10:00:07 UTC 2023\n", 29Mon Sep 18 10:00:07 UTC 2023
) = 29
close(1)                                = 0
munmap(0x4010a000, 4096)                = 0
close(2)                                = 0
exit_group(0)                           = ?
+++ exited with 0 +++

[saethlin]

Sorry, I didn't understand what you meant. Can you explain how to do it more specifically

I think what you're replying to here is an abstract suggestion, not something anyone actually expects you to try.

Personally I'm not even sure if it's correct to say that musl's 64-bit time change has dropped support for old Linux kernels (you are using an old kernel). Musl documents on https://wiki.musl-libc.org/supported-platforms:

musl is built on the Linux syscall layer. Linux kernel >=2.6.39 is necessary for POSIX conformant behaviour,

So if the 64-bit time switch made that statement wrong, the musl docs need to be changed, and we also need to document that our musl targets don't support Linux kernels before clock_gettime64 (because they still statically link musl by default).