Laundry size is always 0

[Konstantin Belousov]

On Sat, May 09, 2020 at 08:43:08AM +0400, Igor Pokrovsky wrote:
> Hello!
> 
> I’ve got strange values for vm laundry size on -STABLE.
> And these values do not change even under load.
> 
> vm.domain.0.stats.laundry: 0
> vm.stats.vm.v_laundry_count: 0
> 
> I cannot complete buildworld on this box, compiler throughs segmentation core at some random points.
> Can it be connected to empty laundry?
> 
> I tested RAM and hard disk for errors - none found.
> Has anyone seen this? Relevant boot messages follows.
> 
> ---<<BOOT>>---
> FreeBSD 12.1-STABLE r360207 GENERIC amd64
> FreeBSD clang version 9.0.1 (git at github.com:llvm/llvm-project.git c1a0a213378a458fbea1a5c77b315c7dce08fd05) (based on LLVM 9.0.1)
> VT(vga): resolution 640x480
> CPU: AMD Athlon(tm) 64 Processor 3000+ (1890.05-MHz K8-class CPU)
>  Origin="AuthenticAMD"  Id=0x10ff0  Family=0xf  Model=0x1f  Stepping=0
>  Features=0x78bfbff<FPU,VME,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,CLFLUSH,MMX,FXSR,SSE,SSE2>
>  AMD Features=0xe2500800<SYSCALL,NX,MMX+,FFXSR,LM,3DNow!+,3DNow!>
>  AMD Features2=0x1<LAHF>
> real memory  = 2147155968 (2047 MB)
> avail memory = 2039795712 (1945 MB)

This machine is quite interesting, it is the first generation of long mode
capable CPUs without cmpxchg16b support.  For me it is interesting that
we boot at all, because I was not able to test some important implementation
of pmap features on such CPU.

Can you provide:
- full verbose dmesg
- output of "sysctl vm".  I am mostly interested in the value of
	vm.disable_swapspace_pageouts.
- output of ps -axlHww

Do you ever have some processes hang on the machine, as opposed to faulting
with SIGSEGV/SIGBUS ?