Re: port-xen/58561 (panic: kernel diagnostic assertion, "x86_read_psl() == 0" failed: file, "/home/netbsd/10/src/sys/arch/x86/x86/pmap.c", line 3581)

["Manuel Bouyer via gnats" <gnats-admin%NetBSD.org@localhost>]

The following reply was made to PR port-xen/58561; it has been noted by GNATS.

From: Manuel Bouyer <bouyer%antioche.eu.org@localhost>
To: Konrad Schroder <perseant%hhhh.org@localhost>
Cc: gnats-bugs%NetBSD.org@localhost, port-xen-maintainer%netbsd.org@localhost,
        netbsd-bugs%netbsd.org@localhost, gnats-admin%netbsd.org@localhost, riastradh%NetBSD.org@localhost,
        campbell+netbsd%mumble.net@localhost, cherry%NetBSD.org@localhost
Subject: Re: port-xen/58561 (panic: kernel diagnostic assertion,
 "x86_read_psl() == 0" failed: file,
 "/home/netbsd/10/src/sys/arch/x86/x86/pmap.c", line 3581)
Date: Sat, 10 Jan 2026 13:55:33 +0100

 --arOAjUDb+7TMRYPD
 Content-Type: text/plain; charset=iso-8859-1
 Content-Disposition: inline
 Content-Transfer-Encoding: 8bit
 
 On Fri, Jan 09, 2026 at 07:24:17AM -0800, Konrad Schroder wrote:
 > I can reliably reproduce this issue, quickly enough that a NetBSD
 > tools+release build usually will not complete before the system panics.  I
 > am using yesterday's -current, with a kernel config as follows:
 > 
 >     netbsd# cat sys/arch/amd64/conf/LFS
 >     include "arch/amd64/conf/XEN3_DOMU"
 > 
 >     options         LOCKDEBUG
 > 
 > The host system is Ubuntu 22.04.5 running xen-hypervisor-4.16-amd64, and the
 > guest config is:
 > 
 >     linux# cat lfs
 >     name = "lfs"
 >     kernel = "/etc/xen/netbsd-lfs.gz"
 >     memory = 16384
 >     vcpus = 2
 >     vif = [ 'bridge=br0' ]
 >     disk = [ '/dev/md3p6,raw,hda,rw' ]
 > 
 >     on_poweroff = 'destroy'
 >     on_reboot   = 'destroy'
 >     on_crash    = 'destroy'
 > 
 > The load running on the system is from
 > 
 >     ./build.sh -j4 -U -u -O obj.amd64 tools release
 > 
 > in a freshly unpacked src.
 > 
 > If I configure the VM with "vcpus = 1", or turn off LOCKDEBUG, it doesn't
 > panic.
 > 
 > (As the names imply, I'm hoping to use this configuration to test LFS, but
 > the problem is unrelated: the system panics without any LFS file system ever
 > having been active since boot.)
 > 
 > The panic traces vary, but are consistent after the call to copyout():
 > 
 >     [ 842.2695658] panic: kernel diagnostic assertion "(psl =
 > x86_read_psl()) == 0" failed: file
 > "/home/src-current/sys/arch/x86/x86/pmap.c", line 3596 psl=0x1
 >     [ 842.2695658] cpu1: Begin traceback...
 >     [ 842.2695658] vpanic() at netbsd:vpanic+0x164
 >     [ 842.2695658] kern_assert() at netbsd:kern_assert+0x4b
 >     [ 842.2695658] pmap_load() at netbsd:pmap_load+0x13d
 >     [ 842.2695658] do_pmap_load() at netbsd:do_pmap_load+0x1d
 >     [ 842.2695658] copyout() at netbsd:copyout+0x48
 >     [ 842.2695658] ubc_uiomove() at netbsd:ubc_uiomove+0x12e
 >     [ 842.2695658] ffs_read() at netbsd:ffs_read+0xf0
 >     [ 842.2695658] VOP_READ() at netbsd:VOP_READ+0x3c
 >     [ 842.2695658] vn_rdwr() at netbsd:vn_rdwr+0x100
 >     [ 842.2695658] vmcmd_readvn() at netbsd:vmcmd_readvn+0x56
 >     [ 842.2695658] execve_runproc() at netbsd:execve_runproc+0x34e
 >     [ 842.2695658] execve1() at netbsd:execve1+0x4c
 >     [ 842.2695658] sys_execve() at netbsd:sys_execve+0x2a
 >     [ 842.2695658] syscall() at netbsd:syscall+0x98
 >     [ 842.2695658] --- syscall (number 59) ---
 >     [ 842.2695658] netbsd:syscall+0x98:
 >     [ 842.2695658] cpu1: End traceback...
 > 
 > Let me know if there is anything you'd like me to do to help test or further
 > diagnose the issue.
 
 Hello,
 can you try with the attached patch ? It won't fix the problem but
 should let us know if syscall() is already called with interrupts disabled,
 or if they're disabled later
 
 -- 
 Manuel Bouyer <bouyer%antioche.eu.org@localhost>
      NetBSD: 26 ans d'experience feront toujours la difference
 --
 
 --arOAjUDb+7TMRYPD
 Content-Type: text/plain; charset=us-ascii
 Content-Disposition: attachment; filename=diff
 
 Index: sys/arch/amd64/amd64/copy.S
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/amd64/amd64/copy.S,v
 retrieving revision 1.36
 diff -u -p -u -r1.36 copy.S
 --- sys/arch/amd64/amd64/copy.S	24 Sep 2024 20:19:30 -0000	1.36
 +++ sys/arch/amd64/amd64/copy.S	10 Jan 2026 12:52:12 -0000
 @@ -176,6 +176,7 @@ ENTRY(kcopy)
  END(kcopy)
  
  ENTRY(copyout)
 +	call _C_LABEL(amd64_check_psl)
  	DEFERRED_SWITCH_CHECK
  
  	xchgq	%rdi,%rsi		/* kernel address to %rsi, user to %rdi */
 Index: sys/arch/amd64/amd64/trap.c
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/amd64/amd64/trap.c,v
 retrieving revision 1.130
 diff -u -p -u -r1.130 trap.c
 --- sys/arch/amd64/amd64/trap.c	20 Jun 2025 17:02:18 -0000	1.130
 +++ sys/arch/amd64/amd64/trap.c	10 Jan 2026 12:52:12 -0000
 @@ -720,3 +722,13 @@ sigdebug(const struct trapframe *tf, con
  	frame_dump(tf, lwp_getpcb(l));
  }
  #endif
 +
 +void amd64_check_psl(void);
 +void
 +amd64_check_psl(void)
 +{
 +#ifdef XENPV
 +		/* Check to see if interrupts are enabled (ie; no events are masked) */
 +		KASSERT(x86_read_psl() == 0);
 +#endif
 +}
 Index: sys/arch/x86/x86/syscall.c
 ===================================================================
 RCS file: /cvsroot/src/sys/arch/x86/x86/syscall.c,v
 retrieving revision 1.22
 diff -u -p -u -r1.22 syscall.c
 --- sys/arch/x86/x86/syscall.c	5 Oct 2023 19:41:06 -0000	1.22
 +++ sys/arch/x86/x86/syscall.c	10 Jan 2026 12:52:12 -0000
 @@ -48,6 +48,7 @@ __KERNEL_RCSID(0, "$NetBSD: syscall.c,v 
  #include <machine/userret.h>
  
  #include "opt_dtrace.h"
 +#include "opt_xen.h"
  
  #ifndef __x86_64__
  int		x86_copyargs(void *, void *, size_t);
 @@ -90,6 +91,10 @@ static
  void
  syscall(struct trapframe *frame)
  {
 +#ifdef XENPV
 +		/* Check to see if interrupts are enabled (ie; no events are masked) */
 +		KASSERT(x86_read_psl() == 0);
 +#endif
  	const struct sysent *callp;
  	struct proc *p;
  	struct lwp *l;
 
 --arOAjUDb+7TMRYPD--