[llvm] [AMDGPU][GFX12] Restrict scalar subword loads to PAL (PR #117576)

[Matt Arsenault via llvm-commits]

Juan Manuel Martinez =?utf-8?q?Caamaño?= <juamarti at amd.com>,
Juan Manuel Martinez =?utf-8?q?Caamaño?= <juamarti at amd.com>,
Juan Manuel Martinez =?utf-8?q?Caamaño?= <juamarti at amd.com>
Message-ID:
In-Reply-To: <llvm.org/llvm/llvm-project/pull/117576 at github.com>


arsenm wrote:

> `llvm.amdgcn.s.buffer.load` is part of the family of intrinsics that is _specifically_ designed to expose ISA instructions to whatever sits on top of LLVM. It is the _user_ of the intrinsic who has to know what they're doing. This has always been the case in the past, e.g. the user must be aware that the intrinsic cannot be used with swizzled buffer descriptors.
> 

We should better document this then. The bugs don't always make it into the public isa docs, and are not mentioned in the (limited) intrinsic documentation either

https://github.com/llvm/llvm-project/pull/117576