NetBSD Problem Report #55462

From www@netbsd.org  Sun Jul  5 22:15:49 2020
Return-Path: <www@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits))
	(Client CN "mail.NetBSD.org", Issuer "mail.NetBSD.org CA" (not verified))
	by mollari.NetBSD.org (Postfix) with ESMTPS id 6707B1A9213
	for <gnats-bugs@gnats.NetBSD.org>; Sun,  5 Jul 2020 22:15:49 +0000 (UTC)
Message-Id: <20200705221548.72F001A9218@mollari.NetBSD.org>
Date: Sun,  5 Jul 2020 22:15:48 +0000 (UTC)
From: ky0ko@disroot.org
Reply-To: ky0ko@disroot.org
To: gnats-bugs@NetBSD.org
Subject: 9.0 MP kernel does not boot on dual-cpu PowerMac G4
X-Send-Pr-Version: www-1.0

>Number:         55462
>Category:       port-macppc
>Synopsis:       9.0 MP kernel does not boot on dual-cpu PowerMac G4
>Confidential:   no
>Severity:       critical
>Priority:       medium
>Responsible:    port-macppc-maintainer
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Jul 05 22:20:00 +0000 2020
>Closed-Date:    Fri Apr 05 03:02:57 +0000 2024
>Last-Modified:  Fri Apr 05 03:02:57 +0000 2024
>Originator:     ky0ko
>Release:        NetBSD 9.0
>Organization:
>Environment:
NetBSD ringoen 9.0 NetBSD 9.0 (GENERIC) #0: Fri Feb 14 00:06:28 UTC 2020  mkrepro@mkrepro.NetBSD.org:/usr/src/sys/arch/macppc/compile/GENERIC macppc

>Description:
after upgrading to 9.0, i found my machine no longer able to boot correctly. it hangs during the kernel initialization, after the display and usb devices and such have been initialized (but immediately before software watchdog has been initialized - disabling swwdog in userconf does not change this). the kernel is still responsive to removing and re-adding usb devices, and to pressing the power button, but never gets to the point of mounting the root filesystem, even if left for hours. booting the non-MP kernel works around this issue, but leaves me with only one of my two processors usable, which is non-ideal.

the computer in question is a PowerMac G4 (FW800) 1.25DP. Under NetBSD 8.1 and 8.2 i am able to boot the machine with the MP kernel without any issue. 

>How-To-Repeat:
Boot an MP-capable PowerMac with the NetBSD 9.0 MP kernel. (unknown if specific to this model, or if others will show similar issues)
>Fix:

>Release-Note:

>Audit-Trail:
From: "David H. Gutteridge" <david@gutteridge.ca>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: port-macppc/55462: 9.0 MP kernel does not boot on dual-cpu
 PowerMac G4
Date: Sun, 05 Jul 2020 22:06:59 -0400

 This is likely the same issue as I reported as port-macppc/54331. I'd
 closed that issue, as it seemed to be resolved for me in -current, but
 that doesn't address the 9 branch. Which means we'd need to isolate the
 changes and pull them up.

 (You may find port-macppc/55326 is also relevant to you. It's an issue
 for me with -current.)

 Dave


From: ky0ko@disroot.org
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: port-macppc/55462: 9.0 MP kernel does not boot on dual-cpu
 PowerMac G4
Date: Mon, 06 Jul 2020 04:16:23 +0000

 I'm not seeing the same symptoms as described in that PR - my second CPU =
 seems to initialize normally and everything about the boot seems to be go=
 ing correctly until the kernel simply stops progressing (but is clearly s=
 till alive). I'll build a -current kernel and test it, though. =0A=0AJuly=
  5, 2020 7:10 PM, "David H. Gutteridge" <david@gutteridge.ca> wrote:=0A=
 =0A> The following reply was made to PR port-macppc/55462; it has been no=
 ted by GNATS.=0A> =0A> From: "David H. Gutteridge" <david@gutteridge.ca>=
 =0A> To: gnats-bugs@netbsd.org=0A> Cc: =0A> Subject: Re: port-macppc/5546=
 2: 9.0 MP kernel does not boot on dual-cpu=0A> PowerMac G4=0A> Date: Sun,=
  05 Jul 2020 22:06:59 -0400=0A> =0A> This is likely the same issue as I r=
 eported as port-macppc/54331. I'd=0A> closed that issue, as it seemed to =
 be resolved for me in -current, but=0A> that doesn't address the 9 branch=
 . Which means we'd need to isolate the=0A> changes and pull them up.=0A> =
 =0A> (You may find port-macppc/55326 is also relevant to you. It's an iss=
 ue=0A> for me with -current.)=0A> =0A> Dave

From: "David H. Gutteridge" <david@gutteridge.ca>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: port-macppc/55462: 9.0 MP kernel does not boot on dual-cpu
 PowerMac G4
Date: Mon, 06 Jul 2020 06:13:39 -0400

 On Mon, 06 Jul 2020 at 04:16:23 +0000, ky0ko wrote:
 >I'm not seeing the same symptoms as described in that PR - my second CPU =
 >seems to initialize normally and everything about the boot seems to be go=
 >ing correctly until the kernel simply stops progressing (but is clearly s=
 >till alive). I'll build a -current kernel and test it, though.

 Sorry, to clarify, that PR covered more than one issue. What I'm
 referring to was noted in my email of 07 Jul 2019, where I stated:
 "a non-debug MP kernel fails to complete booting. It hangs before
 running /sbin/init. It continues to process hardware-related events
 (e.g., I can unplug and plug in USB devices, and it continues to
 handle these correctly), but otherwise just stalls, without any
 output to indicate why." That no longer occurred for me the last time
 I tested -current, but I encountered PR 55326 instead.

 Dave


From: Michael <macallan@netbsd.org>
To: ky0ko@disroot.org
Cc: gnats-bugs@netbsd.org
Subject: Re: port-macppc/55462: 9.0 MP kernel does not boot on dual-cpu
 PowerMac G4
Date: Tue, 7 Jul 2020 01:35:16 -0400

 I've seen similar issues on my 2x 800MHz Quicksilver.
 My questions:
 - does a -current kernel work? There has been some possibly related
   work since 9.0
 - does a customized, smaller kernel boot?

 ( context: the hangs on my Quicksilver disappeared with a smaller
 kernel. I haven't tried a GENERIC in a while though )

 have fun
 Michael

From: ky0ko@disroot.org
To: gnats-bugs@netbsd.org, port-macppc-maintainer@netbsd.org,
 gnats-admin@netbsd.org, netbsd-bugs@netbsd.org
Cc: 
Subject: Re: port-macppc/55462: 9.0 MP kernel does not boot on dual-cpu
 PowerMac G4
Date: Thu, 09 Jul 2020 03:07:37 +0000

 i can now confirm that a -current kernel does fix the issue (and a 9.0_ST=
 ABLE kernel does not)=0A=0AJuly 6, 2020 3:15 AM, "David H. Gutteridge" <d=
 avid@gutteridge.ca> wrote:=0A=0A> The following reply was made to PR port=
 -macppc/55462; it has been noted by GNATS.=0A> =0A> From: "David H. Gutte=
 ridge" <david@gutteridge.ca>=0A> To: gnats-bugs@netbsd.org=0A> Cc: =0A> S=
 ubject: Re: port-macppc/55462: 9.0 MP kernel does not boot on dual-cpu=0A=
 > PowerMac G4=0A> Date: Mon, 06 Jul 2020 06:13:39 -0400=0A> =0A> On Mon, =
 06 Jul 2020 at 04:16:23 +0000, ky0ko wrote:=0A>> I'm not seeing the same =
 symptoms as described in that PR - my second CPU =3D=0A>> seems to initia=
 lize normally and everything about the boot seems to be go=3D=0A>> ing co=
 rrectly until the kernel simply stops progressing (but is clearly s=3D=0A=
 >> till alive). I'll build a -current kernel and test it, though.=0A> =0A=
 > Sorry, to clarify, that PR covered more than one issue. What I'm=0A> re=
 ferring to was noted in my email of 07 Jul 2019, where I stated:=0A> "a n=
 on-debug MP kernel fails to complete booting. It hangs before=0A> running=
  /sbin/init. It continues to process hardware-related events=0A> (e.g., I=
  can unplug and plug in USB devices, and it continues to=0A> handle these=
  correctly), but otherwise just stalls, without any=0A> output to indicat=
 e why." That no longer occurred for me the last time=0A> I tested -curren=
 t, but I encountered PR 55326 instead.=0A> =0A> Dave

State-Changed-From-To: open->needs-pullups
State-Changed-By: dholland@NetBSD.org
State-Changed-When: Sat, 01 Apr 2023 23:08:16 +0000
State-Changed-Why:
HEAD works as of 2020, but -9 does not, so in principle the fixes out
er, ought to be tracked down and pulled up to -9.

Realistically this is not likely to happen, especially since the fixes
are in -10, but especially since -10 is still in beta technically we
should do it.


State-Changed-From-To: needs-pullups->closed
State-Changed-By: riastradh@NetBSD.org
State-Changed-When: Fri, 05 Apr 2024 03:02:57 +0000
State-Changed-Why:
probably not worth the effort to track down and pull up needed changes
especially now that 10.0 is out


>Unformatted:

Home	PR Database Search