NetBSD Problem Report #55736

From gson@gson.org  Mon Oct 19 18:12:57 2020
Return-Path: <gson@gson.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 715D01A9217
	for <gnats-bugs@gnats.NetBSD.org>; Mon, 19 Oct 2020 18:12:57 +0000 (UTC)
Message-Id: <20201019181252.CACF6253ED2@guava.gson.org>
Date: Mon, 19 Oct 2020 21:12:52 +0300 (EEST)
From: gson@gson.org (Andreas Gustafsson)
Reply-To: gson@gson.org (Andreas Gustafsson)
To: gnats-bugs@NetBSD.org
Subject: ipsec_natt and npf tests have regressed on the -9 branch
X-Send-Pr-Version: 3.95

>Number:         55736
>Category:       kern
>Synopsis:       ipsec_natt and npf tests have regressed on the -9 branch
>Confidential:   no
>Severity:       serious
>Priority:       high
>Responsible:    rmind
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Oct 19 18:15:00 +0000 2020
>Last-Modified:  Mon Apr 26 20:10:01 +0000 2021
>Originator:     Andreas Gustafsson
>Release:        NetBSD-9; soon to be in 9.1
>Organization:
>Environment:
System: NetBSD
Architecture: x86_64
Machine: amd64
>Description:

The following test cases are now failing on the -9 branch, tested on
real amd64 hardware.  They did not fail on 9.0.

  net/if_ipsec/t_ipsec_natt:ipsecif_natt_transport_null
  net/if_ipsec/t_ipsec_natt:ipsecif_natt_transport_rijndaelcbc
  net/ipsec/t_ipsec_natt:ipsec_natt_transport_ipv4_null
  net/ipsec/t_ipsec_natt:ipsec_natt_transport_ipv4_rijndaelcbc
  net/npf/t_npf:npf_gc

Logs can be found at:

  http://www.gson.org/netbsd/bugs/build/amd64-9-baremetal/2020/2020.10.16.08.17.09/test.html#failed-tcs-summary

The problems started with one of the pullups to the -9 branch performed
between source dates 2020.06.20.15.46.47 and 2020.06.20.20.03.27.
Auto-bisecting this down to a single pullup failed because of build
breakage.

>How-To-Repeat:

>Fix:

>Release-Note:

>Audit-Trail:

Responsible-Changed-From-To: kern-bug-people->martin
Responsible-Changed-By: maya@NetBSD.org
Responsible-Changed-When: Mon, 19 Oct 2020 18:27:08 +0000
Responsible-Changed-Why:
Ping


Responsible-Changed-From-To: martin->rmind
Responsible-Changed-By: martin@NetBSD.org
Responsible-Changed-When: Mon, 19 Oct 2020 18:28:57 +0000
Responsible-Changed-Why:
Over to original commiter / requestor of the change


From: Martin Husemann <martin@duskware.de>
To: gnats-bugs@netbsd.org
Cc: rmind@NetBSD.org
Subject: Re: kern/55736: ipsec_natt and npf tests have regressed on the -9
 branch
Date: Mon, 19 Oct 2020 20:27:42 +0200

 On Mon, Oct 19, 2020 at 06:15:01PM +0000, Andreas Gustafsson wrote:
 > Logs can be found at:
 > 
 >   http://www.gson.org/netbsd/bugs/build/amd64-9-baremetal/2020/2020.10.16.08.17.09/test.html#failed-tcs-summary
 > 
 > The problems started with one of the pullups to the -9 branch performed
 > between source dates 2020.06.20.15.46.47 and 2020.06.20.20.03.27.
 > Auto-bisecting this down to a single pullup failed because of build
 > breakage.

 The pullups in that time frame are (unless I mismatched something):

 sys/arch/mac68k/dev/ams.c			1.23

 	Fix inverted Y-axis value passed to wsmouse_input(9).
 	[tsutsui, ticket #955]

 lib/libnpf/libnpf.3				1.12
 lib/libnpf/npf.c				1.49
 lib/libnpf/npf.h				1.39
 sys/net/npf/files.npf				1.23
 sys/net/npf/npf.c				1.43
 sys/net/npf/npf.h				1.63
 sys/net/npf/npf_alg.c				1.22
 sys/net/npf/npf_alg_icmp.c			1.33
 sys/net/npf/npf_conf.c				1.17
 sys/net/npf/npf_conn.c				1.32
 sys/net/npf/npf_conn.h				1.20
 sys/net/npf/npf_conndb.c			1.9
 sys/net/npf/npf_connkey.c			1.2
 sys/net/npf/npf_ctl.c				1.60
 sys/net/npf/npf_ext_log.c			1.17
 sys/net/npf/npf_ext_normalize.c			1.10
 sys/net/npf/npf_ext_rndblock.c			1.9
 sys/net/npf/npf_handler.c			1.49
 sys/net/npf/npf_if.c				1.13
 sys/net/npf/npf_ifaddr.c			1.7
 sys/net/npf/npf_impl.h				1.81
 sys/net/npf/npf_inet.c				1.57
 sys/net/npf/npf_mbuf.c				1.24
 sys/net/npf/npf_nat.c				1.50
 sys/net/npf/npf_os.c				1.18
 sys/net/npf/npf_params.c			1.3
 sys/net/npf/npf_portmap.c			1.5
 sys/net/npf/npf_rproc.c				1.20
 sys/net/npf/npf_ruleset.c			1.51
 sys/net/npf/npf_sendpkt.c			1.22
 sys/net/npf/npf_state.c				1.23
 sys/net/npf/npf_state_tcp.c			1.21
 sys/net/npf/npf_tableset.c			1.35
 sys/net/npf/npf_worker.c			1.8
 sys/net/npf/npf_worker.c			1.9
 sys/net/npf/npfkern.h				1.5
 tests/net/npf/t_npf.sh				1.4
 usr.sbin/npf/npf-params.7			1.4-1.7
 usr.sbin/npf/npfctl/Makefile			1.15
 usr.sbin/npf/npfctl/npf.conf.5			1.91
 usr.sbin/npf/npfctl/npf_bpf_comp.c		1.16
 usr.sbin/npf/npfctl/npf_build.c			1.55
 usr.sbin/npf/npfctl/npf_cmd.c			1.1
 usr.sbin/npf/npfctl/npf_parse.y			1.51
 usr.sbin/npf/npfctl/npf_scan.l			1.32
 usr.sbin/npf/npfctl/npf_show.c			1.32
 usr.sbin/npf/npfctl/npf_var.c			1.13
 usr.sbin/npf/npfctl/npf_var.h			1.10
 usr.sbin/npf/npfctl/npfctl.8			1.25
 usr.sbin/npf/npfctl/npfctl.c			1.64
 usr.sbin/npf/npfctl/npfctl.h			1.53
 usr.sbin/npf/npftest/libnpftest/Makefile	1.11
 usr.sbin/npf/npftest/libnpftest/npf_conn_test.c	delete
 usr.sbin/npf/npftest/libnpftest/npf_gc_test.c	1.1 (adapted)
 usr.sbin/npf/npftest/libnpftest/npf_mbuf_subr.c	1.11
 usr.sbin/npf/npftest/libnpftest/npf_nat_test.c	1.14
 usr.sbin/npf/npftest/libnpftest/npf_state_test.c 1.10
 usr.sbin/npf/npftest/libnpftest/npf_test.h	1.21
 usr.sbin/npf/npftest/libnpftest/npf_test_subr.c	1.17
 usr.sbin/npf/npftest/libnpftest/npf_test_subr.c	1.18
 usr.sbin/npf/npftest/npftest.c			1.25
 usr.sbin/npf/npftest/npftest.conf		1.9
 usr.sbin/npf/npftest/npftest.h			1.17

 	Major NPF improvements:
 	- npfkern: introduce the 'state.key.interface' and
 	  'state.key.direction' settings.
 	  Users can now choose whether the connection state should be
 	  strictly per-interface or global at the configuration level.
 	  Keep NAT logic to be always per-interface, though.
 	- npfctl: lots of fixes for the 'npfctl show' logic;
 	  make 'npfctl list' more informative; misc usability improvements
 	  and more user-friendly error messages.
 	- Amend and improve the manual pages.
 	- Various bug fixes.
 	[rmind, ticket #956]

 sys/dev/acpi/acpi_ec.c				1.84

 	PR 53687: only install space handler and enable interrupt for EC if
 	ACPI claims the device is actually present .
 	[jdolecek, ticket #959]

 sys/arch/x86/x86/cpu_rng.c			1.16

 	Use x86_read_psl/x86_disable_intr/x86_write_psl to defer interrupts.
 	This avoids accidently enabling interrupts too early.
 	[riastradh, ticket #960]

 usr.sbin/hdaudioctl/graph.c			1.5
 usr.sbin/hdaudioctl/hdaudioctl.8		1.7
 usr.sbin/hdaudioctl/hdaudioctl.c		1.3,1.4
 usr.sbin/hdaudioctl/hdaudioctl.h		1.3

 	Add show subcommand that displays the codec configuration in
 	human-readable tabulated form.
 	[sborrill, ticket #961]

 etc/MAKEDEV.awk					1.29
 etc/etc.cobalt/MAKEDEV.conf			1.18,1.19
 sys/arch/cobalt/conf/majors.cobalt		1.34

 	PR 55009: rename /dev/panel to lcdpanel.
 	[tsutsui, ticket #962]

 sys/arch/mips/mips/bus_dma.c			1.41

 	PR 54216: fix inconsistent mips_o32, _mips_o32,
 	and __mips_o32 macro.
 	[tsutsui, ticket #963]

 sys/arch/mips/mips/cache.c			1.67

 	PR 55138: use 32 byte cacheline ops (not 16 byte ones) for
 	R5000 picache.
 	[tsutsui, ticket #964]

 sys/arch/mips/include/cache_r5k.h		1.5
 sys/arch/mips/mips/cache_r5k.c			1.21
 sys/arch/mips/mips/cache_r5k_subr.S		1.4

 	PR 55139: use proper "page" alignments for R5k Page Invalidate(S) op.
 	[tsutsui, ticket #965]

 etc/rc.d/postfix				1.19

 	PR 55396: show "Starting postfix" message at startup.
 	[rin, ticket #966]



 with the obvious most likely culprit:

 	Major NPF improvements:
 	[rmind, ticket #956]


 Mindaugas, could you please have a look?
 Also note there is PR 55586 which might be related.

 Martin

From: Andreas Gustafsson <gson@gson.org>
To: Martin Husemann <martin@duskware.de>
Cc: rmind@netbsd.org, gnats-bugs@netbsd.org
Subject: Re: kern/55736: ipsec_natt and npf tests have regressed on the -9 branch
Date: Mon, 19 Oct 2020 21:58:24 +0300

 Martin Husemann wrote:
 >  The pullups in that time frame are (unless I mismatched something):
 >  
 >  sys/arch/mac68k/dev/ams.c			1.23

 That one is outside the time frame, but the commit that fixed the
 broken build is part of it:

   2020.06.20.20.03.27 martin src/usr.sbin/npf/npftest/libnpftest/npf_gc_test.c 1.1.2.3

   Adapt to netbsd-9 branch

 A list of the commits and test results is here:

   http://www.gson.org/netbsd/bugs/build/amd64-9-baremetal/commits-2020.06.html#2020.06.20.15.46.48

 -- 
 Andreas Gustafsson, gson@gson.org

From: Andreas Gustafsson <gson@gson.org>
To: gnats-bugs@netbsd.org
Cc: 
Subject: Re: kern/55736: ipsec_natt and npf tests have regressed on the -9 branch
Date: Mon, 26 Apr 2021 23:05:31 +0300

 Most of the issues reported in this PR have been fixed by now, but
 the net/npf/t_npf:npf_gc test case is still failing.  It fails only
 on the -9 branch, not in -current.
 -- 
 Andreas Gustafsson, gson@gson.org

>Unformatted:

Home	PR Database Search