NetBSD Problem Report #25529

Received: (qmail 8877 invoked by uid 605); 10 May 2004 21:08:27 -0000
Message-Id: <20040510210818.9EF1910560@colleen.internal.johnrshannon.com>
Date: Mon, 10 May 2004 15:08:18 -0600 (MDT)
From: john@johnrshannon.com
Sender: gnats-bugs-owner@NetBSD.org
Reply-To: john@johnrshannon.com
To: gnats-bugs@gnats.NetBSD.org
Subject: smp and ptrace() - markedly different becavior from single-processor
X-Send-Pr-Version: 3.95

>Number:         25529
>Category:       kern
>Synopsis:       With a MP kernel, ptrace() behaves inconsistently
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    kamil
>State:          closed
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Mon May 10 21:09:00 +0000 2004
>Closed-Date:    Mon Sep 16 21:45:24 +0000 2019
>Last-Modified:  Mon Sep 16 21:45:24 +0000 2019
>Originator:     John R. Shannon
>Release:        NetBSD 2.0E
>Organization:
	johnrshannon.com
>Environment:
System: NetBSD colleen.internal.johnrshannon.com 2.0E NetBSD 2.0E (GENERIC.MP) #1: Mon May 10 13:10:09 MDT 2004 root@colleen.internal.johnrshannon.com:/usr/obj/usr/src/sys/arch/i386/compile/GENERIC.MP i386
Architecture: i386
Machine: i386 - dual Xeon processors
>Description:
	While working on a gdb 6.1 package, I've been running the
	gdb testsuite. The test results vary each time I run the
	testsuite unless I execute the testsuite while running 
	a single-processor kernel. I get the same results with 
	pkgsrc/devel/gdb.

	With a single-processor kernel, I get:

		gdb 5.3 GENERIC KERNEL

	        === gdb Summary ===

		# of expected passes            7848
		# of unexpected failures        69
		# of unexpected successes       1
		# of expected failures          70
		# of unresolved testcases       2
		# of untested testcases         3
		# of unsupported tests          5


		gdb 6.1 GENERIC KERNEL

		=== gdb Summary ===

		# of expected passes            9329
		# of unexpected failures        48
		# of expected failures          50
		# of unknown successes          40
		# of known failures             43
		# of unresolved testcases       1
		# of untested testcases         4
		# of unsupported tests          9

	With GENERIC.MP, the tests in the suite start timing out. 


	With gdb 6.1 and GENERIC.MP I got:

		=== gdb Summary ===

		# of expected passes            9292
		# of unexpected failures        86
		# of expected failures          50
		# of unknown successes          40
		# of known failures             42
		# of unresolved testcases       1
		# of untested testcases         4
		# of unsupported tests          9

	On some other runs, after 24 hours, the testsuite had not finished; 
	execution was bogged down timing out for every real-time signal in test
	gdb.base/sigall.exp.

	Some testing on a dual-processor AMD 64 (x86-64) suggests that 
	possibly this problem may not be universal.

>How-To-Repeat:
	1) cd pkgsrc/devel/gdb
	2) make
	3) cd ${WRKSRC} (e.g., work/gdb-5.3)
	4) gmake check
	Try with single and multi-processor kernels
>Fix:
	<how to correct  if known (multiple lines)>
>Release-Note:
>Audit-Trail:

From: "John R. Shannon" <john@johnrshannon.com>
To: gnats-bugs@netbsd.org
Cc:  
Subject: Re: kern/25529
Date: Tue, 11 May 2004 04:37:44 -0600

 =2D----BEGIN PGP SIGNED MESSAGE-----
 Hash: SHA1

 1) Typo corrections in original pr:

 =2D From and Reply-To should be john@johnrshannon.com

 2) Further test runs on x68_64 demonstrated that problem occurs on this=20
 platform also.

 =2D --=20

 John R. Shannon
 john@johnrshannon.com
 =2D----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.2.4 (NetBSD)

 iEYEARECAAYFAkCgrPgACgkQOKbCxya4HYsOoQCfXiLzOsh99A2w+tbG6+mTj6r2
 aycAn3WCdU53X/vh6vrwRpG5FCf2wHfh
 =3Dv9zn
 =2D----END PGP SIGNATURE-----
Responsible-Changed-From-To: gnats-admin->kern-bug-people 
Responsible-Changed-By: fair 
Responsible-Changed-When: Thu May 13 09:01:32 UTC 2004 
Responsible-Changed-Why:  

This is a kern bug - how did it end up in "pending"? 

Responsible-Changed-From-To: kern-bug-people->kamil
Responsible-Changed-By: kamil@NetBSD.org
Responsible-Changed-When: Fri, 06 Oct 2017 23:11:08 +0200
Responsible-Changed-Why:
Take.


State-Changed-From-To: open->closed
State-Changed-By: kamil@NetBSD.org
State-Changed-When: Mon, 16 Sep 2019 23:45:24 +0200
State-Changed-Why:
Close as old. All recent ptrace(2) and GDB/LLDB work is with MP kernels.
All bugs are respected in the context of SMP and MP kernels now.


>Unformatted:

Home	PR Database Search