libspl: declare aok extern in header

[dankamongmen]

Motivation and Context

aok is declared and defined in assert.h. This causes anyone who ends up including libspl's assert.h to pick up a definition of aok, resulting in a situation like:

/usr/bin/ld: src/aggregate.o and src/growlight.o: warning: multiple common of `aok'
/usr/bin/ld: src/crypt.o and src/growlight.o: warning: multiple common of `aok'
/usr/bin/ld: src/nvme.o and src/growlight.o: warning: multiple common of `aok'
/usr/bin/ld: src/stats.o and src/growlight.o: warning: multiple common of `aok'

Maybe this is intended, so that independent files can set their assert settings up independently, but I don't see any kind of use like that within ZoL (and would be questionable IMHO in any case). Instead, I create a single definition in zone.c.

Description

How Has This Been Tested?

Recompiled and reinstalled ZoL on my Debian Unstable machine, and reboooted. All zpools were discovered and filesystems mounted. I then ran make check in a checkout, and it succeeded.

[schwarzgerat](0) $ zpool --version
zfs-0.8.0-1
zfs-kmod-0.8.0-1
[schwarzgerat](0) $ 

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Performance enhancement (non-breaking change which improves efficiency)
• Code cleanup (non-breaking change which makes code smaller or more readable)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation (a change to man pages or other documentation)

Checklist:

• My code follows the ZFS on Linux code style requirements.
• I have updated the documentation accordingly.
• I have read the contributing document.
• I have added tests to cover my changes.
• All new and existing tests passed.
• All commit messages are properly formatted and contain Signed-off-by.

[PrivatePuffin]

@dankamongmen It would be awesome if you would restore formatting on your PR description.

If you haven't run ztest and zloop yourself, it might be prefered if you submitted it as a draft PR next time. No biggy, but it keeps clear which PR is expected to work and what isn't. :)

Good luck with getting this working 👍

[dankamongmen]

@dankamongmen It would be awesome if you would restore formatting on your PR description.

I believe this is done, thanks.

[dankamongmen]

Having looked at some of the build failures above, it appears that some of the testing objects need their own definitions of aok, suggesting that they're not getting it from libzfs. I'm looking into this, but would love to hear from someone more informed, thanks!

[PrivatePuffin]

I believe this is done, thanks.

I see, and believe. Thank you!

[PrivatePuffin]

Having looked at some of the build failures above, it appears that some of the testing objects need their own definitions of aok, suggesting that they're not getting it from libzfs. I'm looking into this, but would love to hear from someone more informed, thanks!

I think @jwk404 should be suited to answer this...

[behlendorf]

For additional background information see #9610, but the intended purpose of this variable it so allow zdb to conditionally disable asserts.

I agree, it really should be extern in the header and then declared globally in libspl as you've done. The tricky bit would be that not everything in user space links against libspl, for example libicp only includes some of the headers (hence the build failure). It may be that just lib/libicp/ needs it, in which case adding the full library would be the easiest way to handle this.

[PaulZ-98]

In addition to the changes above, I also added

diff --git a/tests/zfs-tests/tests/functional/checksum/edonr_test.c b/tests/zfs-tests/tests/functional/checksum/edonr_test.c
index a887560..5e82a85 100644
--- a/tests/zfs-tests/tests/functional/checksum/edonr_test.c
+++ b/tests/zfs-tests/tests/functional/checksum/edonr_test.c
@@ -40,6 +40,7 @@
 #include <sys/time.h>
 #include <sys/stdtypes.h>
 
+int aok;
 /*
  * Test messages from:
  * http://csrc.nist.gov/groups/ST/toolkit/documents/Examples/SHA_All.pdf

to get it built, and the new -A feature still worked in zdb.

[dankamongmen]

Looks like most of the tests are passing with this latest, thanks @PaulZ-98 . PTAL.

[codecov]

Codecov Report

Merging #9752 into master will decrease coverage by <1%.
The diff coverage is n/a.

@@           Coverage Diff            @@
##           master    #9752    +/-   ##
========================================
- Coverage      80%      79%   -<1%     
========================================
  Files         385      385            
  Lines      121470   121470            
========================================
- Hits        96787    96472   -315     
- Misses      24683    24998   +315

Flag	Coverage Δ
#kernel	80% <ø> (ø)	⬇️
#user	67% <ø> (ø)	⬇️

---

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 9fb2771...8071501. Read the comment docs.

[behlendorf]

Since it looks like we only have a single exception with the edonr test code I'm good with this. Let's just explicitly initialize the global.