spin_delay.hpp: fix for Darwin PPC

[barracuda156]

Current assembler is wrong for Darwin PPC, it won’t work and even build. In addition to accommodate ppc64 on Darwin, __ppc64__ should be added (or otherwise __POWERPC__).

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated
seqan3	✅ Ready (Inspect)	Visit Preview	💬 Add your feedback	Jan 30, 2023 at 11:06AM (UTC)

[codecov]

Codecov Report

Base: 98.18% // Head: 98.18% // No change to project coverage 👍

Coverage data is based on head (2fb5449) compared to base (329e5a3).
Patch has no changes to coverable lines.

❗ Current head 2fb5449 differs from pull request most recent head 619db34. Consider uploading reports for the commit 619db34 to get more accurate results

Additional details and impacted files

@@           Coverage Diff           @@
##           master    #3129   +/-   ##
=======================================
  Coverage   98.18%   98.18%           
=======================================
  Files         276      276           
  Lines       12292    12292           
=======================================
  Hits        12069    12069           
  Misses        223      223           

Impacted Files	Coverage Δ
...lude/seqan3/utility/parallel/detail/spin_delay.hpp	100.00% <ø> (ø)

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report at Codecov.
📢 Do you have feedback about the report comment? Let us know in this issue.

[eseiler]

Thank you for the PR!

[eseiler]

@barracuda156 do you by chance have a source/reference for the Darwin PPC assembly instruction?

I only found documentation on the "normal" or 27,27,27 instruction, but not really much on the or r27,r27,r27. Though I found source code that shows that Darwin PPC usually does, e.g., r1,r1,r1 instead of 1,1,1.

[barracuda156]

I only found documentation on the "normal" or 27,27,27 instruction, but not really much on the or r27,r27,r27. Though I found source code that shows that Darwin PPC usually does, e.g., r1,r1,r1 instead of 1,1,1.

@eseiler As long as insns are supported by ISA (and for Darwin that means <= v2.03), they should do what they are supposed to, regardless of the OS (well, provided they do not violate ABI). However, Darwin ABI uses prefixed register names, so 27,27,27 just looks nonsense to the as, since it expects to get registers as arguments of or :)