linux-kernel - Re: [PATCH v2 4/7] linux/signal.h: Ignore SIGINFO by default in new tasks

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 29 Jul 2019 13:55:21 +0300
From:   Arseny Maslennikov <ar@...msu.ru>
To:     Theodore Ts'o <tytso@....edu>
Cc:     Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Jiri Slaby <jslaby@...e.com>, Ingo Molnar <mingo@...hat.com>,
        Peter Zijlstra <peterz@...radead.org>,
        linux-serial@...r.kernel.org, linux-kernel@...r.kernel.org,
        "Vladimir D. Seleznev" <vseleznv@...linux.org>,
        Rob Landley <rob@...dley.net>,
        "Eric W. Biederman" <ebiederm@...ssion.com>,
        Pavel Machek <pavel@....cz>
Subject: Re: [PATCH v2 4/7] linux/signal.h: Ignore SIGINFO by default in new
 tasks

On Tue, Jun 25, 2019 at 05:32:15PM -0400, Theodore Ts'o wrote:
> <...>
> 
> (In particular it might be worth checking Linux ports of Oracle and
> DB2.)

A couple of weeks of asking around got me to have a look at a DB2 9.7
instance.

sn1:~ # uname -a
Linux sn1 2.6.16.60-0.54.5-ppc64 #1 SMP Fri Sep 4 01:28:03 UTC 2009 ppc64 ppc64 ppc64 GNU/Linux

That particular deployment used the following set of processes:

sn1:~ # ps -eo user,ppid,pid,pgid,sid,tty,state,cmd | grep db2 | grep -v java
bgpsysdb  8505   672  8504  7146 ?        S db2fmp ( ,0,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,171fc0,0x110000000,0x110000000,1600000,18002,2,384f00b7
bgpsysdb  8505   687  8504  7146 ?        S db2fmp ( ,0,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,191fc0,0x110000000,0x110000000,1600000,18002,2,384f80b9
bgpsysdb  8505  4309  8504  7146 ?        S db2fmp ( ,0,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,151fc0,0x110000000,0x110000000,1600000,18002,2,6f5080c5
root         1  8505  8504  7146 ?        S db2wdog 0
bgpsysdb  8505  8511  8511  7146 ?        S db2sysc 0
root      8511  8512  8511  7146 ?        S db2ckpwd 0
root      8511  8513  8511  7146 ?        S db2ckpwd 0
root      8511  8514  8511  7146 ?        S db2ckpwd 0
bgpsysdb  8505  8630  8504  7146 ?        S db2acd 0 ,0,0,0,1,0,0,0,1,0,8a6678,14,1e014,2,0,1,11fc0,0x110000000,0x110000000,1600000,18002,2,398072
bgpsysdb  8505 15280  8504  7146 ?        S db2fmp ( ,1,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,31fc0,0x110000000,0x110000000,1600000,18002,2,bd80bb
bgpsysdb  8505 16032  8504  7146 ?        S db2fmp ( ,0,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,b1fc0,0x110000000,0x110000000,1600000,18002,2,81980ba
bgpsysdb  8505 16073  8504  7146 ?        S db2fmp ( ,0,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,91fc0,0x110000000,0x110000000,1600000,18002,2,12300c0
bgpsysdb  8505 17683  8504  7146 ?        S db2fmp ( ,0,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,131fc0,0x110000000,0x110000000,1600000,18002,2,248600db
bgpsysdb  8505 30731  8504  7146 ?        S db2fmp ( ,0,0,0,0,0,0,0,1,0,8a6678,14,1e014,2,0,1,f1fc0,0x110000000,0x110000000,1600000,18002,2,370080d1

None of them has a controlling tty, so ^T pressed at any tty won't send
them a signal.

To get an idea on how do they handle signals, we can look at
/proc/*/status:

sn1:~ # cat /proc/{8512,8505,687,8511}/status | egrep '^(Pid|Sig|Shd)'
Pid:    687
SigQ:   0/128000
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: 0000000000000000
SigIgn: fffffffe2bbaf007
SigCgt: 00000001c44004f8
Pid:    8505
SigQ:   0/128000
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: fffffffe7ffbfeff
SigIgn: fffffffe2fbaf007
SigCgt: 00000001c0410ef8
Pid:    8511
SigQ:   0/128000
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: fffffffe7ffbfeff
SigIgn: fffffffe23b3c005
SigCgt: 00000001dc483efa
Pid:    8512
SigQ:   0/128000
SigPnd: 0000000000000000
ShdPnd: 0000000000000000
SigBlk: 0000000000000000
SigIgn: fffffffe2fbbf007
SigCgt: 00000001c0400ef8

It can be seen from the above that SIGPWR in particular is always
ignored and sometimes blocked, which means its default disposition has
no effect.

This leads me to think that DB2 in particular would be unaffected by the
patch set.

p.s.: I do not have shell access to the machine, and never did; the commands
cited in this email were executed by a person who does, and the output
was handed back to me.

Download attachment "signature.asc" of type "application/pgp-signature" (834 bytes)