Re: kern/45361: Consistent kernel panic when building qemu in amd64

[Julio Merino <jmmv%NetBSD.org@localhost>]

The following reply was made to PR kern/45361; it has been noted by GNATS.

From: Julio Merino <jmmv%NetBSD.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: kern-bug-people%netbsd.org@localhost, Matt Thomas 
<matt%3am-software.com@localhost>, 
 Masao Uebayashi <uebayasi%gmail.com@localhost>
Subject: Re: kern/45361: Consistent kernel panic when building qemu in amd64
Date: Tue, 27 Sep 2011 16:07:49 -0400

 OK, so I finally had some time to attempt to pinpoint the problem.
 
 After quite a few attempts, the offending change seems to be this:
 
 http://mail-index.netbsd.org/source-changes/2011/09/01/msg026757.html
 
 Applying this change to sources from 2011-09-01 makes my machine hang 
 (not panic) while building qemu.  Reverting the change to pristine 
 2011-09-01 changes lets the machine survive.
 
 With a kernel from 2011-09-02, I could even get an uncorrupted panic 
 that seems to highlight the same commit.  (Note that this is with 
 Masao's patch applied, but I don't think it triggered):
 
 panic: kernel diagnostic assertion "(opte & (PG_V | PG_U)) != (PG_V |
 PG_U)" failed: file "/home/jmmv/os/netbsd/src/sys/arch/x86/x86/pmap.c",
 line 1002
 fatal breakpoint trap in supervisor mode
 trap type 1 core 0 rip ffffffff80252375 cs 8 rflags 246 cr2 7f7c86f3000
 cpl 0 rsp ffff8000489449c0
 Stopped in pid 0.51 (system) at netbsd:breakpoint+0x5: leave
 db{3}> bt
 breakpoint() at netbsd:breakpoint+0x5
 panic() at netbsd:panic+0x221
 kern_assert() at netbsd:kern_assert+0x2d
 uvm_pagermapin() at netbsd:uvm_pagermapin+0x179
 uvm_swap_io() at netbsd:uvm_swap_io+0x98
 swapcluster_flush() at netbsd:swapcluster_flush+0x4f
 uvm_pageout() at netbsd:uvm_pageout+0x35f