Re: kern/54818: 9.0_RC1 pagedaemon spins

[Havard Eidnes <he%uninett.no@localhost>]

The following reply was made to PR kern/54818; it has been noted by GNATS.

From: Havard Eidnes <he%uninett.no@localhost>
To: ad%netbsd.org@localhost, tsutsui%ceres.dti.ne.jp@localhost
Cc: gnats-bugs%netbsd.org@localhost
Subject: Re: kern/54818: 9.0_RC1 pagedaemon spins
Date: Mon, 02 Mar 2020 10:42:34 +0100 (CET)

 >  > > Thank you for the vmstat output, that was very helpful.  I
 >  > > think it trying to balance the queues but your other
 >  > > processes are competing with it, and because of
 >  > > uvm_pageqlock nobody can make progress.  This is not a
 >  > > problem on -current any more because uvm_pageqlock is
 >  > > gone..  I have updated the patch to try and help this
 >  > > situation.  I have not compiled it because I don't have a
 >  > > 9.0 system at the moment.  Could you please try it again?
 >  > >
 >  > > 	http://www.netbsd.org/~ad/2020/uvm-9.diff =
 
 >  >
 >  > Thanks, but unfortunately it fails because this change includes
 >  > "TAILQ_ENTRY(vm_page) pdqueue" not available in netbsd-9:
 >
 >  Oops.  Updated patch is here:
 >
 >  	http://www.netbsd.org/~ad/2020/uvm-9.diff
 
 I've manually applied this diff to my 8.1_RC1 src tree, and the
 build succeeded.  Any predictions whether this has any chance of
 improving the situation?  It's completely "cargo cult" to me...
 I see that Izumi says the pdaemon looping still happens though.
 Despite this, I'll test this kernel myself as well, and if it
 re-occurs I can try to collect information as best I can, similar
 to what Izumi has done here.
 
 I noticed in my latest wedge, the X11 server was stuck in "vmem",
 the pdaemon was spinning, and "top" showed 117MB free memory.
 The end result was that the X11 session was stuck, and I had to
 login remotely to reboot the machine (and that even eventually
 required a push of the power butteon).
 
 Regards,
 
 - H=E5vard