Re: Raspberry Pi 3B USB Printing Issue

From: Archimedes Gaviola <archimedes.gaviola_at_gmail.com>
Date: Sun, 27 Mar 2022 14:28:07 UTC
On Sun, Mar 27, 2022 at 5:05 PM Hans Petter Selasky <hps@selasky.org> wrote:

> On 3/27/22 07:55, Archimedes Gaviola wrote:
> > On Sat, Mar 12, 2022 at 4:41 PM Hans Petter Selasky <hps@selasky.org>
> wrote:
> >
> >> On 3/12/22 08:07, Archimedes Gaviola wrote:
> >>> ugen1.5: <EPSON EPSON UB-U03II> at usbus1
> >>> ulpt1 on uhub1
> >>> ulpt1: <EPSON EPSON UB-U03II, class 0/0, rev 1.10/2.00, addr 5> on
> usbus1
> >>> device_attach: ulpt1 attach returned 12
> >>
> >> 12 : man errno :
> >>        12 ENOMEM Cannot allocate memory.
> >>
> >> I guess the EPSON printer you've got is not compatible with ulpt<n>
> >>
> >> When printing, can you make sure that the length transferred is never a
> >> multiple of 64 bytes?
> >>
> >> Also, there might be a bug lurking in the USB host controller driver,
> >> like already mentioned.
> >>
> >
> >
> > Hi Hans,
> >
> > I just figured-out the ulpt(4) driver in my Epson printer while comparing
> > with my Xprinter printer's USB device info. My Epson printer is providing
> > vendor specific values of 255 in the bInterfaceClass and
> bInterfaceSubClass
> > respectively.
> >
> >        bInterfaceClass = 0x00ff  <Vendor specific>
> >        bInterfaceSubClass = 0x00ff
> >
> > It should be a value of 7 for bInterfaceClass and a value of 1 in
> > bInterfaceSubClass.
> >
> >        bInterfaceClass = 0x0007  <Printer device>
> >        bInterfaceSubClass = 0x0001
> >
> > So, the ulpt_attach() routine below will break upon validation for
> > mismatched values in UICLASS_PRINTER and  UISUBCLASS_PRINTER.
> >
> >                          } else {
> >                                  alt_index++;
> >                                  if ((id->bInterfaceClass ==
> > UICLASS_PRINTER) &&
> >                                      (id->bInterfaceSubClass ==
> > UISUBCLASS_PRINTER) &&
> >                                      (id->bInterfaceProtocol ==
> > UIPROTO_PRINTER_BI)) {
> >                                          goto found;
> >                                  }
> >                          }
> >
> > What I did is temporarily replace these values in the USB definition. In
> > this case, how should the project handle this non-compliance USB devices?
> > Though I will raise this to Epson if they could provide an updated
> firmware.
> >
> > freebsd@generic:~ % diff -Nur /usr/src/sys/dev/usb/usb.h.orig
> > /usr/src/sys/dev/usb/usb.h
> > --- /usr/src/sys/dev/usb/usb.h.orig     2022-03-27 02:55:01.319235000
> +0800
> > +++ /usr/src/sys/dev/usb/usb.h  2022-03-27 02:57:10.608518000 +0800
> > @@ -459,8 +459,10 @@
> >   #define        UICLASS_PHYSICAL        0x05
> >   #define        UICLASS_IMAGE           0x06
> >   #define        UISUBCLASS_SIC          1       /* still image class */
> > -#define        UICLASS_PRINTER         0x07
> > -#define        UISUBCLASS_PRINTER      1
> > +/* #define     UICLASS_PRINTER         0x07 */
> > +/* #define     UISUBCLASS_PRINTER      1 */
> > +#define        UICLASS_PRINTER         0xff
> > +#define        UISUBCLASS_PRINTER      0xff
> >
> > I can print now with ulpt(4) driver but need further testing for any
> issues.
> >
> > ugen1.5: <EPSON EPSON UB-U03II> at usbus1
> > ulpt0 on uhub1
> > ulpt0: <EPSON EPSON UB-U03II, class 0/0, rev 1.10/2.00, addr 5> on usbus1
> > ulpt_attach: setting alternate config number: 0
> > ulpt0: using bi-directional mode
> >
>
> Hi,
>
> I think you can just extend that piece of code to accept either value
> using a boolean OR, ||.
>


Hi Hans,

Ah okay, this is noted. I just thought it's not allowed. I already extended
the code and on the way re-building the kernel.

Thanks,
Archimedes