JIT suggestion: optimize comparisons of uint against int.MaxValue

[GrabYourPitchforks]

We have a few places in the code where we compare locals of type uint against int.MaxValue.

runtime/src/libraries/Common/src/System/Buffers/ArrayBufferWriter.cs

Lines 180 to 183 in a0c3324

	if ((uint)newSize > int.MaxValue)
	{
	newSize = currentLength + sizeHint;
	if ((uint)newSize > int.MaxValue)

runtime/src/libraries/Common/src/System/Security/Cryptography/ECDiffieHellmanOpenSsl.Derive.cs

Lines 149 to 151 in a0c3324

	using (SafeEvpPKeyCtxHandle ctx = Interop.Crypto.EvpPKeyCtxCreate(ourKey, theirKey, out uint secretLengthU))
	{
	if (ctx == null || ctx.IsInvalid || secretLengthU == 0 || secretLengthU > int.MaxValue)

runtime/src/libraries/Common/src/System/Security/Cryptography/Asn1Reader/AsnReader.cs

Lines 417 to 419 in a0c3324

	if (Utf8Parser.TryParse(data, out uint value, out int consumed) &&
	value <= int.MaxValue &&
	consumed == data.Length)

runtime/src/libraries/System.Reflection.Metadata/src/System/Reflection/Metadata/TypeSystem/FieldDefinition.cs

Lines 119 to 120 in a0c3324

	uint offset = _reader.FieldLayoutTable.GetOffset(layoutRowId);
	if (offset > int.MaxValue)

There's an opportunity for a few bytes of codegen savings across these methods if the JIT treat these calls as (int)value < 0 or (int)value >= 0 instead of performing a direct comparison against the constant value 0x7FFFFFFF.

category:cq
theme:codegen
skill-level:intermediate
cost:small