Red Hat Bugzilla – Bug 1478953
[Solved] cups-filters: _IO_file_doallocate(): pdftopdf killed by signal 11
Last modified: 2017-09-01 04:44:43 EDT
Description of problem:
During brother printer installation mfc-7420
Version-Release number of selected component:
cmdline: MFC7420 1 root testprint 1 finishings=3 number-up=1 job-uuid=urn:uuid:56f39d54-1931-31ed-70c0-d1b3d1ffee35 job-originating-host-name=localhost date-time-at-creation= date-time-at-processing= time-at-creation=1502106851 time-at-processing=1502106851 document-name-supplied=testprint
runlevel: N 5
Thread no. 1 (7 frames)
#2 _IO_file_doallocate at filedoalloc.c:101
#3 _IO_doallocbuf at genops.c:398
#4 _IO_new_file_seekoff at fileops.c:1031
#5 fseeko at fseeko.c:36
#6 FileInputSource::seek at libqpdf/FileInputSource.cc:108
#7 QPDF::processFile at libqpdf/QPDF.cc:141
#8 QPDF_PDFTOPDF_Processor::loadFile at filter/pdftopdf/qpdf_pdftopdf_processor.cc:407
Created attachment 1310142 [details]
Created attachment 1310143 [details]
Created attachment 1310144 [details]
Created attachment 1310145 [details]
Created attachment 1310146 [details]
Created attachment 1310147 [details]
Created attachment 1310148 [details]
Created attachment 1310149 [details]
Created attachment 1310150 [details]
Created attachment 1310151 [details]
Created attachment 1310152 [details]
Created attachment 1310153 [details]
I'm also experiencing a crash with pdftopdf when attempting to print different types of files (plain text, PDF, etc.).
Printer: Brother DCP-7020
Driver: Brother DCP-7020 Foomatic/hl1250 (recommended)
I apologize but I'm not familiar with how to create a backtrace, although I would very happily learn. You can see my CUPS error log here: https://pastebin.com/vmG61jVe
I did find this additional information:
Process 7283 (pdftopdf) of user 4 dumped core.
Stack trace of thread 7283:
#0 0x00007f3cf8e5419f _int_malloc (libc.so.6)
#1 0x00007f3cf8e579b8 malloc (libc.so.6)
#2 0x00007f3cf8e3c52c _IO_file_doallocate (libc.so.6)
#3 0x00007f3cf8e4c399 _IO_doallocbuf (libc.so.6)
#4 0x00007f3cf8e48a73 _IO_file_seekoff@@GLIBC_2.2.5 (libc.so.6)
#5 0x00007f3cf8e46f29 fseeko (libc.so.6)
#6 0x00007f3cfb5cc965 _ZN15FileInputSource4seekExi (libqpdf.so.17)
#7 0x00007f3cfb5e3103 _ZN4QPDF11processFileEPKcP8_IO_FILEbS1_ (libqpdf.so.17)
#8 0x000055b388226b0a _ZN23QPDF_PDFTOPDF_Processor8loadFileEP8_IO_FILE12ArgOwnership (pdftopdf)
#9 0x000055b38821d59c main (pdftopdf)
#10 0x00007f3cf8deb4da __libc_start_main (libc.so.6)
#11 0x000055b38821d82a _start (pdftopdf)
I realize this most likely isn't very helpful, but I thought I would share regardless.
Similar problem has been detected:
1. Plugged in printer (Brother DCP-7202). Printer was detected and proper driver was installed: (Brother DCP-7020 Foomatic/hl1250 (recommended)
2. Try to print any file (plain text, PDF, etc.) and you print job will start and then immeditately stop.
3. If you enable CUPS debugging you will see that the job failing is a result of filter errors, specifically /usr/lib/cups/filter/pdftopdf
cmdline: DCP-7020 8 robert Test Page 1 job-uuid=urn:uuid:09d9482a-1589-3525-4c28-3dde751dbad3 job-originating-host-name=localhost date-time-at-creation= date-time-at-processing= time-at-creation=1502114420 time-at-processing=1502114420
reason: pdftopdf killed by signal 11
runlevel: N 5
I got hit by this bug this morning. While I was generating an abrt report a message popped up about an update to qpdf. Trying to track down which version that might be, dnf tells me I am up to date on qpdf-libs, and just installed the qpdf package.
thank you for reporting this issue, it is cause by change in qpdf, would you mind downgrading qpdf-libs package for now? Update reverting changes which caused this issue is in update-testing repository.
I can confirm that downgrading to qpdf-libs.x86_64 6.0.0-5.fc26 fixed this issue for me. Thanks for the suggestion!
I apologize for my low attendance but it was late here in italy ..XD
I confirm that a qpdrf-libs downgrade fixes the problem ..
After removing the repo version, I installed 6.0.0 v3 because the 4 failed to download https://www.rpmfind.net/linux/rpm2html/search.php?query=qpdf-libs and then I reinstalled Drivers .. Now prints work ..
Thanks guys: D
Ah .. Sorry my english I'm doing with the google translator
Yes, I wake up with this error due to an update I did yesterday.
abrt-notification: Process 5806 (pdftopdf) crashed in malloc()
Aug 8 15:36:26 samba cupsd: Job stopped due to filter errors; please consult the error_log file for details.
After I put the LogLevel to debug I always see the error in the subscription file of cups.
*** Bug 1479581 has been marked as a duplicate of this bug. ***
FYI Same here. qpdf-libs-6.0.0-6.fc26.x86_64 makes pdftopdf crash and downgrading to qpdf-libs-6.0.0-5.fc26.x86_64 makes it work again.
I see that qpdf-libs-6.0.0-7.fc26.x86_64 is now available. I can confirm after updating to this version the crash with pdftopdf is resolved. Not sure if this bug can be closed now or if it needs to remain open to track the original issue?
I can confirm after this update the problem is resolved!
I would like to leave this bugzilla open for now, because today I will submit the update, which should have CVE patches (which caused binary incompatibility with cups-filters, which caused this issue - solution was rebuild cups-filters with qpdf-libs) and rebuilt cups-filters package. I would like to ask you after that for testing this update, if it works correctly.
would you mind testing this update https://bodhi.fedoraproject.org/updates/cups-filters-1.13.4-3.fc26%20qpdf-6.0.0-8.fc26 ? This update puts back patches for CVEs in qpdf, which were in qpdf-libs-6.0.0-6.fc26 version, but it created binary incompatibility, which caused segfaults in cups-filters. It should be solved by rebuild of cups-filters against new qpdf-libs release (cups-filters rebuild is part of update). I didn't set auto-push to stable on this update, because I want to push it manually after I get positive feedback about its functionality.
Yes, printing is working for me with the new cups-filters and qpdf-libs packages.
Tested with a Brother HL-2280DW and Brother printer drivers.
Sorry for the late reply.
I did test out qpdf-libs.x86_64 6.0.0-8.fc26 and can confirm that it's working great. No segfaults with pdftopdf anymore, and the CVE patches are in that package.
I would say you can safely close out this bug at this point.
Thanks for the help!
Thank you for testing! Closing as CURRENTRELEASE.