Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
1.3.9-5 seems to be working for me!!! thank you very much! On Mon, 2008-11-17 at 11:44 +0100, Martin Pitt wrote: Martin Pitt [2008-11-16 16:59 +0100]: Upstream committed a different patch: http://www.cups.org/strfiles/3001/str3001.patch This patch is now included in 1.3.9-5, which just got uploaded to experimental. Can you guys please test this version? If it works, I need to push that to unstable and lenny, too, but the window for that gets smaller every day. Thank you! Martin -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Martin Pitt [2008-11-16 16:59 +0100]: Upstream committed a different patch: http://www.cups.org/strfiles/3001/str3001.patch This patch is now included in 1.3.9-5, which just got uploaded to experimental. Can you guys please test this version? If it works, I need to push that to unstable and lenny, too, but the window for that gets smaller every day. Thank you! Martin -- Martin Pitt| http://www.piware.de Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org) signature.asc Description: Digital signature
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Hi, On Mon, Nov 17, 2008 at 11:44:04AM +0100, Martin Pitt wrote: Martin Pitt [2008-11-16 16:59 +0100]: Upstream committed a different patch: http://www.cups.org/strfiles/3001/str3001.patch This patch is now included in 1.3.9-5, which just got uploaded to experimental. Can you guys please test this version? If it works, I need to push that to unstable and lenny, too, but the window for that gets smaller every day. OK, I don't know what to make of this: I downgraded cups to 1.3.8-1lenny2 (the original, hanging version), tried printing two of the complex formerly problematic documents, didn't get any socket backend hangs this time (with printer test which remained configured exactly as last time when it did produce hangs). Verified timestamp of socket backend binary, was ok (October, matching distribution package, _NOT_ my custom-patched package version), and of course I had stopped and started cups, repeatedly even. Then I said so what and upgraded to incoming.debian.org 1.3.9-5 (plus _all_ cups helper packages), first submitted job got stuck at first page without producing any printer traffic, second job worked fine. (first job could possibly have been confused by an earlier Job Abort button action at the printer, though) Puzzled. Thanks a lot for your work, Andreas Mohr -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Hello all, Upstream committed a different patch: http://www.cups.org/strfiles/3001/str3001.patch It's not quite doing the same as Samuel's, but according to upstream it is more correct. Any chance you can apply this instead of Samuel's, and checking if it still works? Many thanks! Martin -- Martin Pitt| http://www.piware.de Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org) signature.asc Description: Digital signature
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Martin Pitt, le Sun 16 Nov 2008 16:59:45 +0100, a écrit : Upstream committed a different patch: http://www.cups.org/strfiles/3001/str3001.patch Ok. I was actually tempted to do something similar but I didn't know the consequences. I guess it's ok then. Samuel -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Hi Samuel, Samuel Thibault [2008-11-11 14:46 +0100]: Right but here the issue is on the input side: device_fd got to EOF, thus select() returning it and read() on it returning 0. The attached patch at least prevents select from returning, avoiding 100% CPU usage. Thanks for your analysis and the patch! It looks correct to me, and I applied it to the package bzr. I forwarded it to upstream (http://www.cups.org/str.php?L3001) and will wait a day or two for upstream's comment before uploading it. Martin -- Martin Pitt| http://www.piware.de Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org) signature.asc Description: Digital signature
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Andreas Mohr, le Tue 11 Nov 2008 19:33:02 +0100, a écrit : Or, simply stated, how to disable the test suite to get a successful .deb package build? Usually you just need to prepend DEB_BUILD_OPTIONS=nocheck Samuel -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Hi, On Tue, Nov 11, 2008 at 07:40:40PM +0100, Samuel Thibault wrote: Andreas Mohr, le Tue 11 Nov 2008 19:33:02 +0100, a écrit : Or, simply stated, how to disable the test suite to get a successful .deb package build? Usually you just need to prepend DEB_BUILD_OPTIONS=nocheck That did it. And re-attempting a 40-page duplexed job on cups Debian unstable 1.3.9-2 did hang just like before, and then downgrading to the patched 1.3.8-1lenny2 _does_ print immediately upon cupsd restart, without any socket backend lockup any more. Very nice work, thanks! Samuel Andreas Mohr -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Hi, On Tue, Nov 11, 2008 at 02:46:55PM +0100, Samuel Thibault wrote: Right but here the issue is on the input side: device_fd got to EOF, thus select() returning it and read() on it returning 0. The attached patch at least prevents select from returning, avoiding 100% CPU usage. Very nice, and fast reaction, too! I wanted to verify this patch to succeed versus the unpatched version to still fail with the very same test setup each, but I failed during package build (via fakeroot dpkg-buildpackage in the apt-get source'd cups-1.3.8/ directory): . . . lsb/usr/cups-included/Zebra/zebra.ppd Zebra ZPL Label Printer, 1.3 zebra.ppd Zebra ZPL Label Printer, 1.3 PASSED Test Summary PASS: Printer 'Test1' correctly produced 55 page(s). PASS: Printer 'Test2' correctly produced 23 page(s). PASS: 154 requests processed. PASS: 0 emergency messages. PASS: 0 alert messages. PASS: 0 critical messages. FAIL: 0 error messages, expected 9. PASS: 0 warning messages. PASS: 0 notice messages. PASS: 1 info messages. FAIL: 0 debug messages, expected more than 0. PASS: 0 debug2 messages. 2 tests failed. Log files can be found in /tmp/cups-root/log. A HTML report was created in /tmp/cups-root/cups-str-1.3-2008-11-11-root.html. Copies of the error_log and cups-str-1.3-2008-11-11-root.html files are in /usr/src/system/cups-1.3.8/test. make[1]: *** [check] Error 1 make[1]: Leaving directory `/usr/src/system/cups-1.3.8' make: *** [debian/stamp-makefile-check] Error 2 dpkg-buildpackage: failure: debian/rules build gave error exit status 2 birgit:/usr/src/system/cups-1.3.8# Or, simply stated, how to disable the test suite to get a successful .deb package build? (maybe I shouldn't ignore a failing test run though; then what?) Or could you tell me what the usual way is to do this cups .deb package build? Thanks a lot, Andreas Mohr -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Andreas Mohr, le Sun 02 Nov 2008 17:17:12 +0100, a écrit :
The CUPS backend mechanism had issues before already
(STR #2664, http://www.cups.org/str.php?L2664+P0+S-2+C0+I0+E0+M20+Q ):
r7204 | mike | 2008-01-09 19:59:55 +0100 (Mi, 09 Jan 2008) | 3 lines
Don't select() on the output side of the device if we have a
side-channel
callback - this causes 100% CPU usage (STR #2664)
Right but here the issue is on the input side: device_fd got to EOF,
thus select() returning it and read() on it returning 0. The attached
patch at least prevents select from returning, avoiding 100% CPU usage.
Samuel
--- backend/runloop.c.orig 2008-11-11 14:38:02.0 +0100
+++ backend/runloop.c 2008-11-11 14:44:42.0 +0100
@@ -161,6 +161,7 @@
#if defined(HAVE_SIGACTION) !defined(HAVE_SIGSET)
struct sigaction action; /* Actions for POSIX signals */
#endif /* HAVE_SIGACTION !HAVE_SIGSET */
+ int device_input = 1;
fprintf(stderr,
@@ -210,7 +211,7 @@
FD_ZERO(input);
if (!print_bytes)
FD_SET(print_fd, input);
-if (use_bc)
+if (use_bc device_input)
FD_SET(device_fd, input);
if (!print_bytes side_cb)
FD_SET(CUPS_SC_FD, input);
@@ -273,6 +274,13 @@
CUPS_LLCAST bc_bytes);
cupsBackChannelWrite(bc_buffer, bc_bytes, 1.0);
}
+ else if (bc_bytes == 0)
+ {
+ /*
+* End of file, stop trying to read from device_fd
+ */
+device_input = 0;
+ }
}
/*
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
OK, using HPLIP (hp backend, which appears to be recommended) submitting larger jobs seems to work without a backend lockup, however it's AWFULLY, almost unusably slow (1 page per 2 minutes or so, adding up to maybe 4 minutes per quadruple-duplexed PDF/PS page). (HPLIP is actually known to be very slow in certain configurations, too, and developers even already said that they're working on improving it) I don't know, but whichever thing I try to configure in CUPS, I'm _ALWAYS_ (yes, that's an always, since chances are about 60%) hitting a brick wall of some more or less severe sort (and I'm far from being the only one, judging from Internet discussions). IOW, I found a crude if acceptable workaround, thus severity should be left at grave, since the normal socket backend should at least be semi-usable, too. Thanks, Andreas Mohr -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Environment clarification: HPLJ4000TN JetDirect J3111A Firmware G.08.49 (newest), on a BNC(!) connection. This being a 10Mbps BNC connection here could be another indication that this 100% CPU lockup issue possibly happens on slower connections only (this issue does not seem to be too wide-spread, thus maybe it affects slow-net users only) The CUPS backend mechanism had issues before already (STR #2664, http://www.cups.org/str.php?L2664+P0+S-2+C0+I0+E0+M20+Q ): r7204 | mike | 2008-01-09 19:59:55 +0100 (Mi, 09 Jan 2008) | 3 lines Don't select() on the output side of the device if we have a side-channel callback - this causes 100% CPU usage (STR #2664) Andreas Mohr -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Processed: Re: [Pkg-cups-devel] Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Processing commands for [EMAIL PROTECTED]: severity 489045 normal Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect Severity set to `normal' from `grave' tag 489045 moreinfo Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect There were no tags set. Tags added: moreinfo thanks Stopping processing here. Please contact me if you need assistance. Debian bug tracking system administrator (administrator, Debian Bugs database) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: [Pkg-cups-devel] Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
severity 489045 normal tag 489045 moreinfo thanks Patricio Rojo [2008-07-02 20:03 -0400]: Nothing happens when trying to print, even the test-page, to a HP 4200 network printer through HPDirect. If I run 'top', I see a process called 'socket' owned by user 'lp' that is using 100% of computer use. Since this does not happen for everyone, just on this particular model, I'm downgrading the severity. Can you please attach your /var/log/cups/error_log right after this happens? Also, by using top, can you find out the process ID of that runaway socket process, and do strace -f -o /tmp/strace.out -p PID (set PID to the process ID of the socket process) for some seconds? Please attach /tmp/strace.out here, too. Thanks, Martin -- Martin Pitt| http://www.piware.de Ubuntu Developer (www.ubuntu.com) | Debian Developer (www.debian.org) -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
Bug#489045: cups: infinite loop, 100%CPU use, while trying to print with HPJetDirect
Package: cups Version: 1.3.7-7 Severity: grave Justification: renders package unusable Hi, Nothing happens when trying to print, even the test-page, to a HP 4200 network printer through HPDirect. If I run 'top', I see a process called 'socket' owned by user 'lp' that is using 100% of computer use. I first saw this problem several months ago when I was using 'etch' on my amd64 machine, but I did not reported a bug because my laptop (running i386 'sid') could print without problems in the same printer. So, I just thought it was an incompatibility of the etch version of cups with the printer. However, I'm now runnning 'lenny' on my amd64 machine (and thus I have the same version of cups in both computers) and I still have the same problem. Since no one else is reporting this problem, I guess is something very specific to amd64. I could not find errors messages anywhere to help with the debugging... so please let me know how else can I help. Thank you very much, Patricio -- System Information: Debian Release: lenny/sid APT prefers testing APT policy: (500, 'testing'), (500, 'stable'), (50, 'unstable') Architecture: amd64 (x86_64) Kernel: Linux 2.6.24-1-amd64 (SMP w/2 CPU cores) Locale: LANG=en_US, LC_CTYPE=en_US (charmap=ISO-8859-1) Shell: /bin/sh linked to /bin/bash Versions of packages cups depends on: ii adduser 3.108 add and remove users and groups ii cups-common 1.3.7-7 Common UNIX Printing System(tm) - ii debconf [debconf-2.0]1.5.22 Debian configuration management sy pi ghostscript 8.62.dfsg.1-2.1 The GPL Ghostscript PostScript/PDF ii libavahi-compat-libdnssd 0.6.22-3Avahi Apple Bonjour compatibility ii libc62.7-10 GNU C Library: Shared libraries ii libcups2 1.3.7-7 Common UNIX Printing System(tm) - pi libcupsimage21.3.7-7 Common UNIX Printing System(tm) - ii libdbus-1-3 1.2.1-2 simple interprocess messaging syst ii libgnutls26 2.4.0-2 the GNU TLS library - runtime libr ii libkrb53 1.6.dfsg.3-2MIT Kerberos runtime libraries ii libldap-2.4-22.4.9-1 OpenLDAP libraries ii libpam0g 0.99.7.1-6 Pluggable Authentication Modules l ii libpaper11.1.23 library for handling paper charact ii libslp1 1.2.1-7.3 OpenSLP libraries ii lsb-base 3.2-12 Linux Standard Base 3.2 init scrip ii perl-modules 5.10.0-11 Core Perl modules ii procps 1:3.2.7-8 /proc file system utilities ii ssl-cert 1.0.20 simple debconf wrapper for OpenSSL ii xpdf-utils [poppler-util 3.02-1.3Portable Document Format (PDF) sui Versions of packages cups recommends: ii avahi-utils 0.6.22-3 Avahi browsing, publishing and dis ii cups-client 1.3.7-7Common UNIX Printing System(tm) - ii foomatic-filters 3.0.2-20080211-3.1 OpenPrinting printer support - fil ii smbclient 2:3.0.30-3 a LanManager-like simple client fo -- debconf information: cupsys/raw-print: true cupsys/backend: ipp, lpd, parallel, scsi, serial, socket, usb, snmp, dnssd -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of unsubscribe. Trouble? Contact [EMAIL PROTECTED]
