Ubuntu
ghostscript package

  1. Bug #917148
  2. Comment #0

Comment 0 for bug 917148

Revision history for this message
Till Kamppeter (till-kamppeter) wrote : Kernel problems when accessing USB printers

From Oneiric on, the access to USB printers is exclusively done via libusb and not any more by the "usblp" kernel module. The module is blacklisted by the CUPS package.

The problems occur independently of the CUPS backend (component of CUPS to communicate with printer hardware) used: The original "usb" backend of CUPS (/usr/lib/cups/backend/usb, CUPS package in Oneiric and Oneiric SRUs issued so far, based on deprecated, unmaintained libusb 0.1.x), a port of the "usb" CUPS backend to libusb 1.0.x (also /usr/lib/cups/backend/usb, but of current Precise CUPS package, version 1.5.0-14, or my PPA, https://launchpad.net/~till-kamppeter/+archive/ppa, version 1.5.0-8ubuntu7~ppa2), or HPLIP's "hp" backend (only HP printers, /usr/lib/cups/backend/hp, part of hplip-cups package).

For me the data transfer of print jobs simply stops at a certain point, and always at the same point. It happens on the HP OfficeJet H470 and on the HP PhotoSmart C5280. According to the error_log of CUPS the filters and the driver do their jobs correctly, they do not show errors or crashes. The backend seems to freeze. The job does not complete according to the error_log, the backend process keeps running and the printer tells on its display that it is printing the job. Only after several minutes the display of the printer tells "Job canceled", probably due to a printer-internal timeout on not getting data any more. The CUPS backend keeps running infinitely then and the CUPS job does not terminate. This occurs the same way with all the three backends.

Sending the same job to the printers via network or Bluetooth does not show the problem. The jobs complete correctly in a reasonable time.

There are also other printers where the problem does not occur with all three backends: HP Color LaserJet CM3530, HP LaserJet 1020, HP LaserJet 3390, HP PhotoSmart C8180.

The problem occurs on both Oneiric and Precise. The systems are both up-to-date. Especially all SRUs for CUPS and the kernel are installed, so it is not bug 872711 (or this bug did not get completely fixed).

bug 879933 can be the same problem.

There is also a case of a printer not being detected ate all: bug 916652.