--- Begin Message ---
- To: Debian Bug Tracking System <submit@bugs.debian.org>
- Subject: cupsd daemon freeze when called with SIGUSR1 not unblocked before execve()
- From: Nobody <origil@orange.fr>
- Date: Sun, 04 Aug 2013 20:18:56 +0200
- Message-id: <20130804181856.5064.90346.reportbug@Bing>
Package: cups
Version: 1.5.3-5
Severity: minor
Tags: upstream
Hello.
My cups version is 1.5.3-5, running on a 3.2 linux kernel.
I wrote a wrapper which execute /etc/init.d/ scripts. The cups init
script appeared to freeze when the wrapper called it : the strace program told
"/usr/sbin/cupsd" was simply waiting for SIGUSR1 to be delivered - but my
wrapper, before calling execve(), dit not unblock SIGUSR1 ... (The workaround
was simply to modify the script to unblock all signals before calling execve(),
but ... )
I presume a daemon should not rely on the initial state of the blocked
signal mask when it is called by another program, but I did not find any clear
statement about the convention in that case.
Should'nt the cupsd daemon be made more independant from its startup context ?
-- System Information:
Debian Release: wheezy/sid
APT prefers unstable
APT policy: (700, 'unstable'), (700, 'testing'), (700, 'stable')
Architecture: amd64 (x86_64)
Foreign Architectures: i386
Kernel: Linux 3.2.41-mien (SMP w/4 CPU cores)
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8) (ignored: LC_ALL set to fr_FR.UTF-8)
Shell: /bin/sh linked to /bin/dash
Versions of packages cups depends on:
ii adduser 3.113+nmu3
ii bc 1.06.95-4
ii cups-client 1.5.3-5
ii cups-common 1.5.3-5
ii cups-filters 1.0.18-2.1
ii cups-ppdc 1.5.3-5
ii debconf [debconf-2.0] 1.5.43
ii dpkg 1.16.10
ii ghostscript 9.05~dfsg-6.3
ii libavahi-client3 0.6.31-2
ii libavahi-common3 0.6.31-2
ii libc-bin 2.13-38
ii libc6 2.17-3
ii libcups2 1.5.3-5
ii libcupscgi1 1.5.3-5
ii libcupsimage2 1.5.3-5
ii libcupsmime1 1.5.3-5
ii libcupsppdc1 1.5.3-5
ii libdbus-1-3 1.6.8-1
ii libgcc1 1:4.7.2-5
ii libgnutls26 2.12.20-5
ii libgssapi-krb5-2 1.10.1+dfsg-4+nmu1
ii libkrb5-3 1.10.1+dfsg-4+nmu1
ii libldap-2.4-2 2.4.28-1.3
ii libpam0g 1.1.3-7.1
ii libpaper1 1.1.24+nmu2
ii libslp1 1.2.1-9
ii libstdc++6 4.7.2-5
ii libusb-1.0-0 2:1.0.12-2
ii lsb-base 4.1+Debian7
ii poppler-utils 0.18.4-6
ii procps 1:3.3.6-1
ii ssl-cert 1.0.32
Versions of packages cups recommends:
ii avahi-daemon 0.6.31-2
ii colord 0.1.21-1
ii foomatic-filters 4.0.17-1
ii ghostscript-cups 9.05~dfsg-6.3
ii printer-driver-gutenprint 5.2.9-1
Versions of packages cups suggests:
ii cups-bsd 1.5.3-5
ii cups-pdf 2.6.1-8
ii foomatic-db 20120523-1
pn hplip <none>
pn printer-driver-hpcups <none>
ii smbclient 2:3.6.15-1
ii udev 175-7
-- Configuration Files:
/etc/init.d/cups changed [not included]
/etc/logrotate.d/cups changed [not included]
-- debconf information:
* cupsys/raw-print: true
* cupsys/backend: ipp, ipp14, lpd, socket, usb, snmp, dnssd
--- End Message ---
--- Begin Message ---
- To: 718717-done@bugs.debian.org
- Subject: Re: Bug#718717: cupsd daemon freeze when called with SIGUSR1 not unblocked before execve()
- From: Brian Potkin <claremont102@gmail.com>
- Date: Tue, 26 Apr 2016 18:16:00 +0100
- Message-id: <26042016180303.3891cf27fd32@desktop.copernicus.demon.co.uk>
- In-reply-to: <20130804181856.5064.90346.reportbug@Bing>
- References: <20130804181856.5064.90346.reportbug@Bing>
On Sun 04 Aug 2013 at 20:18:56 +0200, Nobody wrote:
> Hello.
> My cups version is 1.5.3-5, running on a 3.2 linux kernel.
> I wrote a wrapper which execute /etc/init.d/ scripts. The cups init
> script appeared to freeze when the wrapper called it : the strace program told
> "/usr/sbin/cupsd" was simply waiting for SIGUSR1 to be delivered - but my
> wrapper, before calling execve(), dit not unblock SIGUSR1 ... (The workaround
> was simply to modify the script to unblock all signals before calling execve(),
> but ... )
> I presume a daemon should not rely on the initial state of the blocked
> signal mask when it is called by another program, but I did not find any clear
> statement about the convention in that case.
>
> Should'nt the cupsd daemon be made more independant from its startup context ?
With all the relatively recent changes in the Debian init system and
CUPS and the ending of support for Wheezy it is possible you may have
to rethink this issue. Please feel free to submit a new bug report
based on the present testing/unstable if you think it desirable.
Regards,
Brian.
--- End Message ---