Hugepage: Don't check for system setting

[mkroening]

madvise still works if hugepages are disabled system wide. It just does not have an effect then.

[codecov]

Codecov Report

Merging #260 (2c453a3) into master (fd9d2fd) will increase coverage by 0.66%.
The diff coverage is 0.00%.

@@            Coverage Diff             @@
##           master     #260      +/-   ##
==========================================
+ Coverage   62.58%   63.25%   +0.66%     
==========================================
  Files          16       16              
  Lines        2475     2449      -26     
==========================================
  Hits         1549     1549              
+ Misses        926      900      -26     

Impacted Files	Coverage Δ
src/bin/uhyve.rs	0.45% <0.00%> (ø)
src/utils.rs	100.00% <ø> (+49.05%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update fd9d2fd...2c453a3. Read the comment docs.

[stlankes]

Do we not use transparent huge pages (https://access.redhat.com/documentation/en-us/red_hat_enterprise_linux/6/html/performance_tuning_guide/s-memory-transhuge)?

[mkroening]

We do: https://github.com/hermitcore/uhyve/blob/fd9d2fd54c4561f51d5262a43b91f071af643b87/src/linux/uhyve.rs#L422-L427

But madvise only gives an advise to the kernel. We do not have to check weather the user has disabled transparent hugepages. If the user did explicitly disable it on the system, madvise is ignored, as far as I know (at my system at least).

This merely removes the check and defaults to on. Users can still disable huge pages either system wide or via the CLI flag.

[stlankes]

bors r+

[bors]

Build succeeded!

And happy new year! 🎉

• Build (macos-latest)
• Build (ubuntu-latest)
• Cargo tests (macos-latest)
• Cargo tests (ubuntu-latest)
• Clippy
• Format
• KVM Tests
• Rusty-Hermit Kernel based integration tests