Linux 5.2 compat: Define wait_on_page_writeback() if GPL-only exported

[kusumi]

Motivation and Context

Fixes #8775.

Description

wait_on_page_writeback() was made GPL only in torvalds/linux@19343b5bdd.
Define one if it's GPL only.

How Has This Been Tested?

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.

[codecov]

Codecov Report

Merging #8794 into master will increase coverage by 0.01%.
The diff coverage is 0%.

@@            Coverage Diff             @@
##           master    #8794      +/-   ##
==========================================
+ Coverage   78.72%   78.74%   +0.01%     
==========================================
  Files         382      382              
  Lines      117809   117810       +1     
==========================================
+ Hits        92747    92771      +24     
+ Misses      25062    25039      -23

Flag	Coverage Δ
#kernel	79.29% <0%> (-0.05%)	⬇️
#user	67.28% <ø> (-0.22%)	⬇️

---

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 e55db32...9f83bf6. Read the comment docs.

[behlendorf]

We're going to need to add an configure check for this, see the ZFS_AC_KERNEL_BIO_SET_DEV_GPL_ONLY check in config/kernel-bio_set_dev.m4 for an example.

Although, what might be simpler in this particular case is to update zfs_putpage to directly call wait_on_page_bit() since it's the only caller. Then we don't need any compatibility code.

[kusumi]

We're going to need to add an configure check for this, see the ZFS_AC_KERNEL_BIO_SET_DEV_GPL_ONLY check in config/kernel-bio_set_dev.m4 for an example.

Added.

Although, what might be simpler in this particular case is to update zfs_putpage to directly call wait_on_page_bit() since it's the only caller. Then we don't need any compatibility code.

It now only has wait_on_page_writeback() in module/zfs/zfs_vnops.c for GPL export case.

[behlendorf]

Actually, that's not quite what I was suggesting, but after re-reading my comment I see why it was confusing. We could absolutely go with your updated patch. However, since there's only this one caller I was thinking we could do away with the configure check entirely and do something much simpler, like this:

@@ -4526,8 +4526,10 @@ zfs_putpage(struct inode *ip, struct page *pp, struct wri
                unlock_page(pp);
                rangelock_exit(lr);
 
-               if (wbc->sync_mode != WB_SYNC_NONE)
-                       wait_on_page_writeback(pp);
+               if (wbc->sync_mode != WB_SYNC_NONE) {
+                       if (PageWriteback(pp))
+                               wait_on_page_bit(pp, PG_writeback);
+               }
 
                ZFS_EXIT(zfsvfs);
                return (0);

[kusumi]

Oh.
I could change to do like that, but it would start to be more complex if there are >1 consumers of it.
Well, maybe not, it's just two lines of code in the end...

[behlendorf]

This code actually hasn't changed in quite a while, and offhand I can't think of any pending work where we'd need to add another caller. So why don't we keep it simple for now and avoid the wrapper entirely. As you said, it's only two lines after all.

[kusumi]

Updated.

[behlendorf]

Should be dropped when refreshed to include the configure check

[kusumi]

Dropped accordingly.