Re: install/49470: NetBSD 9.0 RC1 reboots after bootloader (HP Compaq 6005 Pro / AMD Phenom II)

[Andrew Doran <ad%netbsd.org@localhost>]

The following reply was made to PR install/49470; it has been noted by GNATS.

From: Andrew Doran <ad%netbsd.org@localhost>
To: Matthias Petermann <mp%petermann-it.de@localhost>
Cc: Andrius V <vezhlys%gmail.com@localhost>, Paul Goyette <paul%whooppee.com@localhost>,
	install-manager%netbsd.org@localhost, gnats-admin%netbsd.org@localhost,
	netbsd-bugs%netbsd.org@localhost, gnats-bugs%netbsd.org@localhost
Subject: Re: install/49470: NetBSD 9.0 RC1 reboots after bootloader (HP
 Compaq 6005 Pro / AMD Phenom II)
Date: Wed, 11 Dec 2019 10:59:09 +0000

 Hi,
 
 I understand it's quite difficult to do if you don't have a working build
 environment, and getting the kernels in place is tough too, but assuming
 that your machine actually gets into the kernel and crashes there rather
 than the boot loader, and you are determined to find the problem...
 
 If I wanted to debug this I would try sprinkling printf()s in the kernel's
 main() in src/sys/kern/init_main.c, after the call to consinit().
 
 If a machine does not even get that far, or it reboots so quickly that the
 messages cannot be read, I keep this assembly function around to dump into
 src/sys/arch/amd64/amd64/locore.S (or i386, it should assemble correctly on
 both) and then call it from C code.  It produces long beeps from the speaker
 and by counting the beeps and following a process of approximate bisection I
 can figure out where it's going wrong.
 
 	http://www.netbsd.org/~ad/keep/beep.S
 
 It is an awful lot of work though...
 
 Andrew
 
 
 On Wed, Dec 11, 2019 at 09:48:58AM +0100, Matthias Petermann wrote:
 
 > thanks for your response. You are right, the issue maps not exactly to the
 > reported problem as a different CPU is installed.
 > 
 > Anyway, the following actions did not succeed so far:
 > 
 > - booting with boot -12
 > - booting from SATA hard drive with pre-installed NetBSD
 > - USB options not configurable in BIOS
 > - older Release NetBSD 6.0 did also not boot (hangs at the second spinner
 > 123456\ but does not reboot)
 > - switching from AHCI to legacy/native IDE mode
 > - Booting FreeBSD 11.2 also hangs in primary bootstrap
 > 
 > Especially the fact that FreeBSD doesn't boot either led me to [1]
 > 
 > I will try to find out if I get to the point where I could debug into the
 > kernel from my serial console...
 > 
 > Kind regards
 > Matthias
 > 
 > 
 > [1] http://freebsd.1045724.x6.nabble.com/boot-hang-with-certain-Phenom-II-cpu-models-td6300528.html
 > 
 > Am 09.12.2019 um 22:08 schrieb Andrius V:
 > > Hi,
 > > 
 > > It is quite unlikely that your issue is related to the original bug
 > > report since hardware was VIA based (VIA C7-M+VN896) HP 2133 netbook.
 > > The typical actions I would try to do: 1) boot without ACPI/SMP
 > > enabled (boot -12) 2) if you're booting using USB media, you can try
 > > to change some relevant USB options (speed let's say, depends on BIOS)
 > > or try to boot from non USB media (CD, netboot or install it to HDD on
 > > the compatible machine and attach it to your system). If any of those
 > > boot, you can check dmesg output for potentially unsupported hardware.
 > > Finally you can try to boot older releases to see if it worked before.
 > > Otherwise you are probably out of luck (playing with some BIOS options
 > > like switching between AHCI/IDE modes may help but not necessarily).
 > > 
 > > P.S. gnats-bugs%netbsd.org@localhost should be added as one of email addresses.
 > > 
 > > 
 > > 
 > > On Mon, Dec 9, 2019 at 9:18 PM Matthias Petermann <mp%petermann-it.de@localhost> wrote:
 > > > 
 > > > resent mail, removed tuotila%gmx.com@localhost from cc (address seams to be not
 > > > existing anymore)
 > > > 
 > > > ---
 > > > 
 > > > Hi,
 > > > 
 > > > Am 09.12.2019 um 14:38 schrieb Paul Goyette:
 > > > > > WARNING: 1 module failed to load
 > > > > > 
 > > > > > How do I find out which module failed to load here?
 > > > > 
 > > > > Almost certainly, this is a failure to load the ffs filesystem,
 > > > > since the filesystem module is already built-in to the kernel.
 > > > > 
 > > > > The message can safely be ignored.
 > > > 
 > > > That's bad news ;-) So there is no known way to debug this further. I
 > > > wish I had more knowledge on possible causes. The box boots without any
 > > > issue into Slackware Linux. I should try a BIOS upgrade and/or try a
 > > > NetBSD version older than the one where the issue was first reported.
 > > > 
 > > > Kind regards
 > > > Matthias
 > > > 
 > > > --
 > > > Matthias Petermann <mp%petermann-it.de@localhost> | www.petermann-it.de
 > > > Innovative IT-L?sungen, Systemintegration, Linux/FreeBSD/Unix-Support
 > > > Wildparkring 13, 01458 Ottendorf-Okrilla | Tel.: +49 (0)35205 597 991
 > > > 
 > 
 > -- 
 > Matthias Petermann <mp%petermann-it.de@localhost> | www.petermann-it.de
 > Innovative IT-L?sungen, Systemintegration, Linux/FreeBSD/Unix-Support
 > Wildparkring 13, 01458 Ottendorf-Okrilla | Tel.: +49 (0)35205 597 991
 >