Page MenuHomeFreeBSD

Remove a bogus assertion from vm_page_launder().
ClosedPublic

Authored by markj on Feb 23 2018, 5:12 PM.
Tags
None
Referenced Files
Unknown Object (File)
Sep 16 2024, 4:29 PM
Unknown Object (File)
Sep 8 2024, 3:39 PM
Unknown Object (File)
Sep 8 2024, 9:07 AM
Unknown Object (File)
Sep 8 2024, 5:34 AM
Unknown Object (File)
Sep 2 2024, 2:08 AM
Unknown Object (File)
Aug 23 2024, 2:58 AM
Unknown Object (File)
Aug 22 2024, 10:02 PM
Unknown Object (File)
Aug 13 2024, 1:13 AM
Subscribers

Details

Summary

After r328977, a page m may have m->queue != PQ_NONE and
m->wire_count > 0.

Diff Detail

Repository
rS FreeBSD src repository - subversion
Lint
Lint Not Applicable
Unit
Tests Not Applicable

Event Timeline

Hmm, should we queue wired page into laundry at all ?

In D14485#303755, @kib wrote:

Hmm, should we queue wired page into laundry at all ?

Sorry, I don't follow. Here we only queue the page if it isn't wired. If it's wired, there's no point in enqueuing it since the unwiring thread will either enqueue it or ensure that it is not enqueued anywhere.

Yes, it was the bad question. I mean, should we leave the page queued if it is wired and vm_page_launder() is called on it.

In D14485#303890, @kib wrote:

Yes, it was the bad question. I mean, should we leave the page queued if it is wired and vm_page_launder() is called on it.

I don't see any reason to bother dequeuing: we'd need to acquire the page queue lock, and one of the motivations for r328977 was to avoid that lock when it is not needed.

This revision is now accepted and ready to land.Feb 23 2018, 10:21 PM
This revision was automatically updated to reflect the committed changes.