7444 fs/xattr.c should be more transparent (zfs_acl_test)

Review Request #446 — Created April 27, 2017 and discarded

gwr
illumos-gate
7444
general

Oh... I accidently uploaded an unrelated diff into this review,
and I don't know of any way to undo that, so I've opened a new one:
https://www.illumos.org/rb/r/518/

7444 fs/xattr.c should be more transparent (zfs_acl_test)

The function xattr_dir_access() was missing some logic
that was there in the earlier version.

Also presented as a github compare:
https://github.com/illumos/illumos-gate/compare/master...gwr:il-7444e

Reviewed by: Yuri Pankov yuri.pankov@nexenta.com
Reviewed by: Dan Fields dan.fields@nexenta.com
Reviewed by: Roman Strashkin roman.strashkin@nexenta.com

Re-ran the test case Yuri described. OK now.

  • 0
  • 0
  • 1
  • 0
  • 1
Description From Last Updated
danmcd
  1. I'm FINE with this modulo my itching question below.

  2. usr/src/uts/common/fs/xattr.c (Diff revision 1)
     
     

    A cursory look at the xattr_dir_realdir() source doesn't show why EINVAL is morally equivalent to the makes-sense ENOENT. Why does EINVAL mean "no real xattr dir"?

    1. This is one of the possible returns from VOP_LOOKUP with flags LOOKUP_XATTR,
      when extended attribute directories are not supported. (i.e. see ufs_lookup)
      I'm basically just reinstating the special handling of those two errors
      from the code before 7444 went in.

  3. 
      
danmcd
  1. Ship It!
  2. 
      
yuripv
  1. Ship It!
  2. 
      
gwr
gwr
gwr
gwr
gwr
gwr
Review request changed

Status: Discarded

Loading...