Re: [PATCH 1/2] copy_process: Handle architectures where sizeof(unsigned long) < sizeof(u64)

[David Hildenbrand <david@redhat.com>]

On 21.08.25 13:27, Simon Schuster via B4 Relay wrote:
> From: Simon Schuster <schuster.simon@siemens-energy.com>
> 
> With the introduction of clone3 in commit 7f192e3cd316 ("fork: add
> clone3") the effective bit width of clone_flags on all architectures was
> increased from 32bit to 64bit. However, the signature of the copy_*
> helper functions (e.g., copy_sighand) used by copy_process was not
> adapted, as such, they potentially truncate the flags on architectures
> such as nios2, where unsigned long is a 32bit unsigned integer type.
> 
> This can, for instance, be observed via failures of kernel selftest
> clone3_clear_sighand, which attempts to trigger the conditional
> 
> 	if (clone_flags & CLONE_CLEAR_SIGHAND)
> 
> in function copy_sighand within fork.c that will always fail given:
> 
> 	unsigned long /* == uint32_t */ clone_flags
> 	#define CLONE_CLEAR_SIGHAND 0x100000000ULL
> 
> This commit fixes the bug by always passing clone_flags via their
> declared u64 type, invariant of architecture-dependent integer sizes.

Sounds reasonable.

But is this actually something that is already exposed before patch#2 on 
other architectures?

(I assume above output is with patch #2 but without patch #1)

If so, we need a Fixes:. If not, we're good.

-- 
Cheers

David / dhildenb