Re: 5.99.3 amd64 kernel panic when loading modules

[Andrew Doran <ad%netbsd.org@localhost>]

On Sat, Nov 22, 2008 at 06:42:52PM +0100, Christoph Egger wrote:

> Christoph Egger wrote:
> > Christoph Egger wrote:
> >> Hi!
> >>
> >> Machine paniced at boot when it wanted to load a module
> >> for launch postfix.
> >>
> >> My machine: NetBSD/amd64 GENERIC kernel 5.99.3.
> >> The kernel has DEBUG+DIAGNOSTIC enabled.
> >>
> >> panic: kernel debugging assertion "end <= VM_MAX_KERNEL_ADDRESS" failed:
> >> file "sys/uvm/uvm_km.c", line 495
> >> fatal breakpoint trap in supervisor modetrap type 1 code 0 rip
> >> ffffffff802250b5 cs 8 rflags 246 cr2 ffff80004da880b0 cpl 0 rsp
> >> ffff80004e4a2390
> >> Stopped in pid 276.1 (postfix) at netbsd:breakpoint+0x5:     leave
> >> db{0}> bt
> >> breakpoint() at netbsd:breakpoint+0x5
> >> panic() at netbsd:panic+0x260
> >> __kernassert() at netbsd:__kernassert+0x2d
> >> uvm_km_check_empty() at netbsd:uvm_km_check_empty+0x13c
> >> uvm_map() at netbsd:uvm_map+0x14f
> >> uvm_km_alloc() at netbsd:uvm_km_alloc+0xa8kobj_load() at
> >> netbsd:kobj_load+0xabak
> >> obj_load_file() at netbsd:kobj_load_file+0x113
> >> module_do_load() at netbsd:module_do_load+0x31f
> >> module_autoload() at netbsd:module_autoload+0xb2
> >> execve1() at netbsd:execve1+0x1ca
> >> syscall() at netbsd:syscall+0xb6
> >> db{0}>
> >>
> > 
> > The value of
> > VM_MAX_KERNEL_ADDRESS is 0xffffff8000000000
> > 
> > The value of   end    is 0xffffffff80eeb000
> 
> I found and solved this.
> The issue was simply userland too old.
> Updating the hole userland (including /boot) to match the kernel version
> fixed it.

That's unlikely! See PR 32816. I have a fix for it.

Andrew