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: Tue, 13 Jan 2026 15:23:14 +0100

 On Mon, Jan 12, 2026 at 11:20:34PM +0100, Manuel Bouyer wrote:
 > On Sun, Jan 11, 2026 at 08:35:10PM -0800, Konrad Schroder wrote:
 > > On 1/10/2026 2:33 PM, Manuel Bouyer wrote:
 > > > And this assembly work may have allowed me to find the problem.
 > > > Basically, the CLI and STI macros are not atomic on Xen PV, and if preemption
 > > > happens at a bad time we may end up updating the upcall_mask of our previous
 > > > CPU. The attached patch should close this race condition; hopefully it's
 > > > the last one.
 > > 
 > > I can confirm that, at least in my case, this allows me to build
 > > continuously all day without a crash.
 > 
 > Good, it points that we're on the right track.
 > Can you try the attached patch instead ? It tries to limit the extra
 > instructions to only places where it's needed, and add DIAGNOSTIC code
 > to check it's really not needed in other places.
 
 Sorry this patch has a bug; working on an updated version
 stay tuned.
 
 -- 
 Manuel Bouyer <bouyer%antioche.eu.org@localhost>
      NetBSD: 26 ans d'experience feront toujours la difference
 --