kassert MDL_FPU_IN_CPU

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

Hello,
in NetBSD/Xen automated tests, I've occasionally seen this panic on HEAD,
on both amd64 and i386 (with similar stack traces):
b/libc/sys/t_ptrace_waitid (342/924): 569 test cases
    access_fpregs1: [ 1655.3561773] panic: kernel diagnostic assertion "!(l->l_md.md_flags & MDL_FPU_IN_CPU)" failed: file "/usr/src/sys/arch/x86/x86/fpu.c", line 144 
[ 1655.3561773] cpu1: Begin traceback...
[ 1655.3661924] vpanic(c12bbe34,d9e7e944,d9e7e96c,c01253be,c12bbe34,c12bbd9b,c12bd43c,c12bd3ac,90,c2687800) at netbsd:vpanic+0x196
[ 1655.3661924] kern_assert(c12bbe34,c12bbd9b,c12bd43c,c12bd3ac,90,c2687800,d9e7e9f8,da9a0000,6c,d9e7ee14) at netbsd:kern_assert+0x23
[ 1655.3661924] process_read_fpregs_s87(c4b7d3c0,d9e7e9a0,d9e7edbc,c0dc2758,c4b7d3c0,d9e7e9a0,d9e7e99c,6c,c0121a92,c4b7d3c0) at netbsd:process_read_fpregs_s87+0x1d4
[ 1655.3661924] process_read_fpregs(c4b7d3c0,d9e7e9a0,d9e7e99c,6c,c0121a92,c4b7d3c0,6c,d9e7e9c0,c0d12097,d9e7ea34) at netbsd:process_read_fpregs+0x18
[ 1655.3661924] proc_regio(c01219c4,c0121a92,d9e7ee14,c30ba680,23,0,0,0,c1643800,bfb57354) at netbsd:proc_regio+0x76
[ 1655.3661924] do_ptrace(c1643800,c3586780,23,783,bfb57354,0,d9e7ef60,d9e7ef9c,c04ae20b,c3586780) at netbsd:do_ptrace+0xed5
[ 1655.3761856] sys_ptrace(c3586780,d9e7ef68,d9e7ef60,c2798788,1,1a,d9e7ef60,d9e7ef68,0,0) at netbsd:sys_ptrace+0x3e
[ 1655.3761856] syscall() at netbsd:syscall+0x17c
[ 1655.3761856] --- syscall (number 26) ---
[ 1655.3761856] b041ac6e:
[ 1655.3761856] cpu1: End traceback...

This was with build 202208140400Z
Does it ring a bell to someone ?

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--