Cp 1.15

[tallclair]

What type of PR is this?
/kind bug

What this PR does / why we need it:

Fix the kubectl cp untarr symlink logic. Unpacking the symlinks after all the regular files have been unpacked guarantees that a file can't be written through a symlink. Furthermore, we still call mkDirAll for the symlink names, so symlink chaining is prevented as well.

This greatly simplifies the symlink resolution logic, and is robust to corner cases we have missed in the past.

Special notes for your reviewer:

Please review carefully. This is changing a bunch of logic that we've had a lot of issues with in the past.

Does this PR introduce a user-facing change?:

kubectl cp now safely allows unpacking of symlinks that may point outside the destination directory

/sig cli
/milestone v1.15
/assign @soltysh

[tallclair]

/assign @M00nF1sh

[soltysh]

One nit.
/lgtm
/approve

[k8s-ci-robot]

[APPROVALNOTIFIER] This PR is APPROVED

This pull-request has been approved by: soltysh, tallclair

The full list of commands accepted by this bot can be found here.

The pull request process is described here

Needs approval from an approver in each of these files:

• pkg/kubectl/OWNERS [soltysh]

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[tallclair]

Fixed the gofmt/vet issues

[soltysh]

/lgtm
/priority important-longterm

[M00nF1sh]

/lgtm
wondering whether it's fine to allow create symbol-links points to outside of working directory..
e.g. if some one use kubectl cp twice to same working directory, which first one create symbol link to outside of working directory and second one try to write the file, is this considered a security issue?

[tallclair]

if some one use kubectl cp twice to same working directory, which first one create symbol link to outside of working directory and second one try to write the file, is this considered a security issue?

Valid question, but IMO we shouldn't worry about this case.

[k8s-ci-robot]

New changes are detected. LGTM label has been removed.

[tallclair]

Fixed comment typo. Reapplying lgtm.

[tallclair]

If we want to minimize the changes, I can also add the symlink restrictions back in, but still create the symlinks at the end.

[tallclair]

/assign @Bubblemelon
For cherry-pick approval

[tallclair]

/retest

[seans3]

/retest