Reverse comparison in mmap test

[dougm]

Summary

Peter Holm reports that the kernel is now unbootable. This should fix the error.

[pho]

Yes, this fixes the boot issue for me.

I ran some mmap() tests on i386. I must confess to have lost track of what's fixed and what is in the pipeline.

20190611 01:35:26 all (28/36): mmap34.sh
panic: caller failed to provide space -545882112 at address 0x20a00000
cpuid = 3
time = 1560209731
KDB: stack backtrace:
db_trace_self_wrapper(e633ad,1bcec40,0,262a498c,bd2061,...) at db_trace_self_wrapper+0x2a/frame 0x262a4960
kdb_backtrace(7,3,3,fd2644,25345530,...) at kdb_backtrace+0x2e/frame 0x262a49c0
vpanic(160c9c2,262a4a04,262a4a04,262a4a58,12bc154,...) at vpanic+0x121/frame 0x262a49e4
panic(160c9c2,df768000,20a00000,3,ffffffff,...) at panic+0x14/frame 0x262a49f8
vm_map_find(25345530,22388d68,0,0,262a4ba0,...) at vm_map_find+0x5c4/frame 0x262a4a58
vm_map_find_min(25345530,22388d68,0,0,262a4ba0,df768000,20402000,0,2,3,7,411) at vm_map_find_min+0x37/frame 0x262a4a9c
vm_mmap_object(25345530,262a4ba0,df768000,3,7,1,22388d68,0,0,1,25488700) at vm_mmap_object+0x2f6/frame 0x262a4aec
vn_mmap(210d0e38,25345530,262a4ba0,df768000,3,7,1,0,0,25488700) at vn_mmap+0x144/frame 0x262a4b60
kern_mmap(25488700,0,df767846,3,1,3,0,0) at kern_mmap+0x3e0/frame 0x262a4be0
sys_mmap(25488700,25488988) at sys_mmap+0x22/frame 0x262a4c08
syscall(262a4ce8,3b,3b,3b,20624604,...) at syscall+0x2d9/frame 0x262a4cdc
Xint0x80_syscall() at 0xffc033b7/frame 0x262a4cdc
--- syscall (477, FreeBSD ELF32, sys_mmap), eip = 0x205df0ab, esp = 0xfbafcf2c, ebp = 0xfbafcfa4 ---

https://people.freebsd.org/~pho/stress/log/dougm031.txt