NetBSD Problem Report #55922

From www@netbsd.org  Tue Jan 12 14:46:17 2021
Return-Path: <www@netbsd.org>
Received: from mail.netbsd.org (mail.netbsd.org [199.233.217.200])
	(using TLSv1.3 with cipher TLS_AES_256_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 98DB91A9217
	for <gnats-bugs@gnats.NetBSD.org>; Tue, 12 Jan 2021 14:46:17 +0000 (UTC)
Message-Id: <20210112144616.BFA521A9248@mollari.NetBSD.org>
Date: Tue, 12 Jan 2021 14:46:16 +0000 (UTC)
From: bsiegert@gmail.com
Reply-To: bsiegert@gmail.com
To: gnats-bugs@NetBSD.org
Subject: Kernel panics with nvme+gpt on Pinebook Pro
X-Send-Pr-Version: www-1.0

>Number:         55922
>Category:       port-arm
>Synopsis:       Kernel panics with nvme+gpt on Pinebook Pro
>Confidential:   no
>Severity:       critical
>Priority:       high
>Responsible:    port-arm-maintainer
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Jan 12 14:50:00 +0000 2021
>Last-Modified:  Wed Jan 13 11:05:01 +0000 2021
>Originator:     Benny Siegert
>Release:        NetBSD 9.99.76
>Organization:
The NetBSD Foundation
>Environment:
NetBSD 9.99.76 kernel (since update to 9.99.77) + NetBSD 9 userland, Pinebook Pro, evbarm-aarch64
>Description:
I added an NVMe drive to the Pinebook Pro and created a GPT with some partitions on it (root, EFI, swap, home). The kernel is loaded from eMMC but the root FS is on the NVMe.

Trying to build a Rust program using MAKE_JOBS=6 (thus, highly parallel, using all cores), I can make the machine crash after several minutes of such activity.

gdb is responsive but "sync" hangs for obvious reasons.

Here are two photos I snapped of the backtraces:

https://photos.app.goo.gl/f58UfkoURGxTjF3j8
https://photos.app.goo.gl/JF25NDTEDYK4Uvt26

The backtraces are a little different but both of them are in writes to the file system (below dofilewrite).
>How-To-Repeat:
- aarch64 system (Pinebook Pro)
- root FS on NVME, FFS on GPT
- build some Rust program (in my case, wip/sccache, which is not committed yet).
>Fix:
?

>Audit-Trail:
From: matthew green <mrg@eterna.com.au>
To: gnats-bugs@netbsd.org
Cc: port-arm-maintainer@netbsd.org, gnats-admin@netbsd.org,
    netbsd-bugs@netbsd.org
Subject: re: port-arm/55922: Kernel panics with nvme+gpt on Pinebook Pro
Date: Wed, 13 Jan 2021 22:00:52 +1100

 not that i have any particular insight to the crash (the images
 don't show the panic message -- can you get a copy of those?
 from ddb, 'dmesg' should show it all with a pager), and we
 should figure out what is crashing here, but i do not recommend
 using MAKE_JOBS=6 or -j6 on the PBP.

 there are two serious issues:

 - 3.875GiB of ram vs 6 jobs is simply not enough for many
   modern packges.  some use >1GiB each.

 - the PBP battery controller setup is not the best.  it can only
   charge at 15W.  however, the SoC and devices are quite capable
   of using more than 15W, so the battery will drain, even while
   plugged into barrel or usb-c power.


 i strongly recommend using no more than 3 concurrent jobs.


 .mrg.

NetBSD Home
NetBSD PR Database Search

(Contact us) $NetBSD: query-full-pr,v 1.46 2020/01/03 16:35:01 leot Exp $
$NetBSD: gnats_config.sh,v 1.9 2014/08/02 14:16:04 spz Exp $
Copyright © 1994-2020 The NetBSD Foundation, Inc. ALL RIGHTS RESERVED.