NetBSD Problem Report #56560

From gson@gson.org  Mon Dec 20 11:48:34 2021
Return-Path: <gson@gson.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 A09571A9239
	for <gnats-bugs@gnats.NetBSD.org>; Mon, 20 Dec 2021 11:48:34 +0000 (UTC)
Message-Id: <20211220114825.AFA2F254286@guava.gson.org>
Date: Mon, 20 Dec 2021 13:48:25 +0200 (EET)
From: gson@gson.org (Andreas Gustafsson)
Reply-To: gson@gson.org (Andreas Gustafsson)
To: gnats-bugs@NetBSD.org
Subject: execsnoop test now fails
X-Send-Pr-Version: 3.95

>Number:         56560
>Category:       port-amd64
>Synopsis:       execsnoop test now fails
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    port-amd64-maintainer
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon Dec 20 11:50:00 +0000 2021
>Closed-Date:    Mon Dec 20 19:04:47 +0000 2021
>Last-Modified:  Mon Dec 20 19:04:47 +0000 2021
>Originator:     Andreas Gustafsson
>Release:        NetBSD-current, source date 2021.12.19.16.01.29
>Organization:

>Environment:
System: NetBSD
Architecture: x86_64
Machine: amd64
>Description:

Since yesteday's DRM update, the usr.sbin/execsnoop/t_execsnoop:basic
test case is failing on amd64, both under qemu and on real hardware:

  http://releng.netbsd.org/b5reports/amd64/2021/2021.12.19.12.45.50/test.html#usr.sbin_execsnoop_t_execsnoop_basic

  https://www.gson.org/netbsd/bugs/build/amd64-baremetal/2021/2021.12.19.16.01.29/test.html#usr.sbin_execsnoop_t_execsnoop_basic

The 51-line long error message ends with:

  "/usr/lib/dtrace/psinfo.d", line 46: failed to copy type of 'pr_arglen': Type information is in parent and unavailable

I don't see any obvious connection between DRM and DTrace, but perhaps
this is another case of the total number of symbols, types, or similar
in the kernel growing past some fixed limit in DTrace.

>How-To-Repeat:

>Fix:

>Release-Note:

>Audit-Trail:
From: "J. Hannken-Illjes" <hannken@mailbox.org>
To: NetBSD GNATS <gnats-bugs@netbsd.org>
Cc: 
Subject: Re: port-amd64/56560: execsnoop test now fails
Date: Mon, 20 Dec 2021 13:29:25 +0100

 --Apple-Mail=_5D608CFD-A969-43D9-B78F-1582AB2907D6
 Content-Transfer-Encoding: 7bit
 Content-Type: text/plain;
 	charset=us-ascii

 > I don't see any obvious connection between DRM and DTrace, but perhaps
 > this is another case of the total number of symbols, types, or similar
 > in the kernel growing past some fixed limit in DTrace.

 To test please run

 	ctfdump /netbsd | grep "total number of types"

 DTrace will fail if this number is greater than 32767.

 --
 J. Hannken-Illjes - hannken@mailbox.org

 --Apple-Mail=_5D608CFD-A969-43D9-B78F-1582AB2907D6
 Content-Transfer-Encoding: 7bit
 Content-Disposition: attachment;
 	filename=signature.asc
 Content-Type: application/pgp-signature;
 	name=signature.asc
 Content-Description: Message signed with OpenPGP

 -----BEGIN PGP SIGNATURE-----

 iQIzBAEBCAAdFiEEyLVMkhxs8fxixv+2IOocBq6p/bMFAmHAdyUACgkQIOocBq6p
 /bO8sw/9HjbVaAcZPQ/1zisnqKuOXKubz0QaHfv8Q9MNGTb9V6nxMFYuklw35bYl
 VliiUUa3kMnb/XCLKoA3M6WGcq98XC09iJqj9bWohxjz+p8TaMz3JJT9HnqdDRWf
 x5117FpJr0BVhAq1TsoGTrYshAbzBz3TMm5BKERTAtFXhQUKxWkJDmUZ/E1G1Exr
 FrVvzWQL/pJNBhvEFDmG8/TH4/3rP06XGeGZpe7i7pXndENkAZ9jj6PkMvOz963/
 DR+FBvXwh15ZByvTLSum58aigtyqTOUfs74h79AmgB5V+JAtiKFNUrnsRQEYyC1a
 eNkhgMlX1ZcKtD6jkbPa8FpMdRrNx2xmr+lxlvmWu/SPNOACkPl5Wf1jpPdS6RuP
 p8J/k9hdvBSG6VQZObNhTXtBcdXB4xj5AOhi9nn8x7HdVM7tJlLpL9tIqPk89cVb
 mjWXwN8elSS74dmzuHnv9miEht3hqSlJixgUMS8Lq3cj/uhl2CVgAgHbSYSzKoUw
 41FcTpGukXszVRt0Dz/oYpyUsbQp4u6zkOTwxh7yNCVlPtlFMiAd5/Qi2p/1zirw
 a0Lwp3xg4Gnr3jNHeLMtqeNC0TQNYvrcxnNobFtW425WYX6JXeuOdZ0ytQNK2Y1W
 a9QFYv9iiLSyXUCm2tGft0mu9FQ9qPBdz08VMg7JBlpfOytesbM=
 =Aqjx
 -----END PGP SIGNATURE-----

 --Apple-Mail=_5D608CFD-A969-43D9-B78F-1582AB2907D6--

From: Andreas Gustafsson <gson@gson.org>
To: "J. Hannken-Illjes" <hannken@mailbox.org>
Cc: gnats-bugs@netbsd.org
Subject: Re: port-amd64/56560: execsnoop test now fails
Date: Mon, 20 Dec 2021 14:35:03 +0200

 J. Hannken-Illjes wrote:
 >  To test please run
 >  
 >  	ctfdump /netbsd | grep "total number of types"
 >  
 >  DTrace will fail if this number is greater than 32767.

 # ctfdump /netbsd | grep "total number of types"
   total number of types               = 34066

 -- 
 Andreas Gustafsson, gson@gson.org

From: Taylor R Campbell <riastradh@NetBSD.org>
To: gnats-bugs@netbsd.org
Cc: port-amd64-maintainer@netbsd.org, gnats-admin@netbsd.org, netbsd-bugs@netbsd.org, gson@gson.org
Subject: Re: port-amd64/56560: execsnoop test now fails
Date: Mon, 20 Dec 2021 12:38:00 +0000

 > Date: Mon, 20 Dec 2021 11:50:00 +0000 (UTC)
 > From: gson@gson.org (Andreas Gustafsson)
 > 
 > I don't see any obvious connection between DRM and DTrace, but perhaps
 > this is another case of the total number of symbols, types, or similar
 > in the kernel growing past some fixed limit in DTrace.

 This is it -- owing to some namespace shenanigans, a lot of drm
 structs got duplicated many times over in the CTF data, so there's
 well over 40k CTF types in the kernel.  Working on reducing the
 duplication.

From: Taylor R Campbell <riastradh@NetBSD.org>
To: gnats-bugs@netbsd.org
Cc: port-amd64-maintainer@netbsd.org,
	gnats-admin@netbsd.org, netbsd-bugs@netbsd.org, gson@gson.org
Subject: Re: port-amd64/56560: execsnoop test now fails
Date: Mon, 20 Dec 2021 13:09:07 +0000

 cvs up and see if this fixes it?

 https://mail-index.netbsd.org/source-changes/2021/12/20/msg135983.html

 (May still be a problem when we turn on amdgpu, though, and that's
 mostly because amdgpu is gigantic, not because of duplicate types...)

From: Andreas Gustafsson <gson@gson.org>
To: Taylor R Campbell <riastradh@NetBSD.org>
Cc: gnats-bugs@netbsd.org
Subject: Re: port-amd64/56560: execsnoop test now fails
Date: Mon, 20 Dec 2021 20:44:29 +0200

 Taylor R Campbell wrote:
 > cvs up and see if this fixes it?
 > 
 > https://mail-index.netbsd.org/source-changes/2021/12/20/msg135983.html

 Yes, that did fix it - thanks.  The test now passes, and the command
 hannken suggested now shows

 # ctfdump /netbsd | grep "total number of types"
   total number of types               = 29658

 -- 
 Andreas Gustafsson, gson@gson.org

State-Changed-From-To: open->closed
State-Changed-By: gson@NetBSD.org
State-Changed-When: Mon, 20 Dec 2021 19:04:47 +0000
State-Changed-Why:
Fixed, thanks!


>Unformatted:

Home	PR Database Search