Panic (KASSERT) in src/sys/uvm/uvm_map.c", line 2120 (today's HEAD).

[Robert Elz <kre%munnari.OZ.AU@localhost>]

I just got the following panic, from running the ATF tests, after an amd64
build of HEAD (updated sources within the past 24 hours).   Aside from Jason's
driver changes (which can't be relevant, this system has none of those
devices) I don't think I'm missing any meaningful changes from HEAD as
it is now.

The test system is a XEN DomU - currently sitting in ddb (there's no
particular hurry to make it do anything else, if anyone has any info
they'd like me to extract from it).    No MKDEBUG on the build though.


lib/libc/sys/t_mmap (322/930): 11 test cases
    mmap_block: [0.002733s] Skipped: The test case causes a panic (PR kern/38889, PR kern/46592)
    mmap_err: [0.002748s] Passed.
    mmap_hint: [ 991.7219923] panic: kernel diagnostic assertion "!topdown || hint <= orig_hint" failed: file "/release/src/sys/uvm/uvm_map.c", line 2120 hint: 0x1ff000, orig_hint: 0x1000
[ 991.7219923] cpu1: Begin traceback...
[ 991.7219923] vpanic() at netbsd:vpanic+0x177
[ 991.7219923] kern_assert() at netbsd:kern_assert+0x4b
[ 991.7219923] uvm_map_findspace() at netbsd:uvm_map_findspace+0x41e
[ 991.7219923] uvm_map_prepare() at netbsd:uvm_map_prepare+0x1f6
[ 991.7219923] uvm_map() at netbsd:uvm_map+0x70
[ 991.7219923] uvm_mmap.part.0() at netbsd:uvm_mmap.part.0+0x15a
[ 991.7219923] sys_mmap() at netbsd:sys_mmap+0x39f
[ 991.7219923] syscall() at netbsd:syscall+0x9c
[ 991.7219923] --- syscall (number 197) ---
[ 991.7219923] netbsd:syscall+0x9c:
[ 991.7219923] cpu1: End traceback...
[ 991.7219923] fatal breakpoint trap in supervisor mode
[ 991.7219923] trap type 1 code 0 rip 0xffffffff8022e93d cs 0xe030 rflags 0x202 cr2 0x7f7ff7ecb000 ilevel 0 rsp 0xffffb0805d3d0b20
[ 991.7219923] curlwp 0xffffb080028c22c0 pid 1380.1380 lowest kstack 0xffffb0805d3cc2c0
Stopped in pid 1380.1380 (t_mmap) at    netbsd:breakpoint+0x5:  leave


If there's anything you'd like to see, tell me just how to extract
the info that you want, and I shall attempt to do that.

The system was doing nothing except running the ATF tests (full run)
and the normal background work (somewhat limited compared to usual, as
there no networking configured up).

kre