Why is Array.MaxLength in .NET 2147483591 (0x7FFFFFC7) instead of Int32.MaxValue?

[semihartan]

In .NET, the Array.MaxLength property is defined as 2147483591 (0x7FFFFFC7), which is slightly less than Int32.MaxValue (2,147,483,647).

• Arrays are indexed by int, so I would expect the maximum length to be Int32.MaxValue.
• The documentation (https://learn.microsoft.com/en-us/dotnet/api/system.array.maxlength) only states that this is a runtime limitation, but does not explain why this specific number was chosen.
• 0x7FFFFFC7 is Int32.MaxValue - 56. What runtime or memory layout considerations led to this exact value?
• Is this related to array headers, alignment, or internal allocation safety margins?

Can someone with knowledge of the CLR internals explain why the maximum array length is 2147483591 instead of Int32.MaxValue? Is there an official design note or source code reference that documents this choice?

[tannergooding]

This is just to help avoid issues with integer overflow, as per:

runtime/src/coreclr/vm/gchelpers.cpp

Lines 604 to 609 in 02bea82

	inline SIZE_T MaxArrayLength()
	{
	// Impose limits on maximum array length to prevent corner case integer overflow bugs
	// Keep in sync with Array.MaxLength in BCL.
	return 0X7FFFFFC7;
	}

There is various logic that needs to do offset and other calculations using the number of elements and a slightly smaller limit helps avoid potential issues, making typical code more robust without being limiting.

Notably, allocating such large arrays, regardless of what the actual runtime limit is, has other fundamental problems and is typically not recommended.

[colejohnson66]

What kind of arithmetic? The object header? That's only 8 bytes, no?

[tannergooding]

It's just a reasonable limit that simplifies things and helps avoid general edge case bugs.

Often when picking limits you pick something that is a bit more than necessary as well, since it helps avoid breaks over time.