riscv64: Implement SIMD shifts, v{all,any}_true and vhigh_bits

[afonso360]

👋 Hey,

This PR implements lowerings for SIMD shifts, and a few boolean operations ( v{all,any}_true/vhigh_bits).

The implementation for shifts is fairly straightforward, but the boolean operations don't have a equivalent instruction, so we need to do some different operations.

[github-actions]

Subscribe to Label Action

cc @cfallin, @fitzgen

This issue or pull request has been labeled: "cranelift", "isle"

Thus the following users have been cc'd because of the following labels:

• cfallin: isle
• fitzgen: isle

To subscribe or unsubscribe from this label, edit the .github/subscribe-to-label.json configuration file.

Learn more.

[alexcrichton]

All looks good to me, thanks!

For CI abrupt interruptions like that sometimes mean that QEMU is OOM-ing perhaps because of something like the pooling allocator reserving memory or something like that (or perhaps a flaky run, but we don't get many of those). Other than that though I'm not sure what would cause such a failure in CI.

[afonso360]

This is somewhat weird, I ran it a few times yesterday and it always got canceled. I tested today in my own branch, and it worked, and I just reran CI here and it worked again.

I also tested running the CI tests on my machine and it only used ~500MB, so I'm not entirely sure what is going on. I thought we didn't run any pooling tests on CI due to QEMU issues?

I'm going to tentatively merge this and if it causes issues again, we should revert it and look at it again.