Hi All, Now we have Matthias on board with his MX375.
Please don't use the old patches. I started changing code in pixma.c where we might get problems with other scanners, e.g. with the Imageclass sub-backend. So, we should find a fix which affects pixma_mp150.c only. We should start at the beginning. Please use attached patch with latest code from git. I added some additional debug outputs to show the status of adf scanning. Please create a debug file with 'export SANE_DEBUG_PIXMA=4'. I also added TODO comments as hints for possible fixes. Cheers, Rolf Am 14.03.2014 22:41, schrieb Matthias Peter Walther: > Hallo Rolf, > > thanks for your quick reply. I'm from Germany too, by the way. But to > keep this conformed, I'm gonna answer you in english. > > If I understood your posts correctly, you want me to use patch 3+4. > Patch 5 is only for the feature to initiate a scan from the device. > Which is nice, but I will look into that later. > > So I pulled my git-sane-backend-repo and patched it: > > mpw at Server0:~/sane$ patch -p1 < attachment-0003.bin > patching file sane-backends/backend/pixma.c > mpw at Server0:~/sane$ patch -p1 < attachment-0004.bin > patching file sane-backends/backend/pixma_mp150.c > Hunk #2 succeeded at 1397 (offset -12 lines). > Hunk #3 succeeded at 1439 (offset -12 lines). > Hunk #4 succeeded at 1493 (offset -12 lines). > Hunk #5 succeeded at 1560 (offset -12 lines). > Hunk #6 succeeded at 1568 (offset -12 lines). > Hunk #7 succeeded at 1580 (offset -12 lines). > > I think the offset is related to the age of the patch. > > After installing, I check the ldconfig: > > $ sudo ldconfig -v | grep libsane > /sbin/ldconfig.real: Kann ?stat()? f?r ?/lib/i686-linux-gnu? nicht > aufrufen: Datei oder Verzeichnis nicht gefunden > /sbin/ldconfig.real: Kann ?stat()? f?r ?/usr/lib/i686-linux-gnu? nicht > aufrufen: Datei oder Verzeichnis nicht gefunden > /sbin/ldconfig.real: Pfad ?/usr/local/lib? mehrfach angegeben > /sbin/ldconfig.real: Pfad ?/lib/x86_64-linux-gnu? mehrfach angegeben > /sbin/ldconfig.real: Pfad ?/usr/lib/x86_64-linux-gnu? mehrfach angegeben > libsane.so.1 -> libsane.so.1.0.25 > libsane.so.1 -> libsane.so.1.0.22 > /sbin/ldconfig.real: Es ist nicht m?glich, den Status (stat()) der Datei > /usr/lib/x86_64-linux-gnu/libsoftokn3.so zu lesen: Datei oder > Verzeichnis nicht gefunden > libsane.so.1 -> libsane.so.1.0.22 > libsane-canon_mfp.so.1 -> libsane-canon_mfp.so.1.0.9 > > This looks good, I think I had 1.0.24 from last october before. Okay, I > tried a scan then: > > $ scanadf > scanadf: sane_read: Error during device I/O > Error during device I/O > Scanned 0 pages > $ > > This did not work before, either. Maybe my sane-frontend is too old? I > have the original one from Ubuntu 12.04.4. > $ scanadf --version > scanadf (sane-frontends) 1.0.14 > > Next try, scanimage: > $ scanimage --source "Automatic Document Feeder" -p --batch > Scanning -1 pages, incrementing by 1, numbering from 1 > Scanning page 1 > Scanned page 1. (scanner status = 5) > Scanning page 2 > scanimage: sane_read: Document feeder out of documents > Scanned page 2. (scanner status = 7) > mpw at Server0:~/sane$ ls > attachment-0003.bin attachment-0004.bin out1.pnm sane-backends > > This is a first success. But as you can see in the output of ls, it > stores only the first page, the second is lost. Or is there something > wrong with my scanimage-command? > > BTW: Can you please change "Automatic Document Feeder" to simply "ADF", > that would be much more comfortable to type :) > > So thanks for your patch, I think it's almost perfect. I just need to > store the 2nd page. > > Please let me know, if you need more detailed debugging output. Will you > commit this two patches to the git repository? > > Gr??e aus M?nster > Matthias > > Am 14.03.2014 19:31, schrieb Rolf Bensch: >> Hi Matthias, >> >> Please read this thread: >> http://lists.alioth.debian.org/pipermail/sane-devel/2014-March/032138.html. >> >> Cheers, >> Rolf >> >> >> Am 12.03.2014 21:11, schrieb Matthias Peter Walther: >>> Hello, >>> >>> I'm new to this, but I'm gonna try my luck. I own a Canaon MX375 >>> all-in-one. The PIXMA-backend has this annoying adf-empty-bug. >>> >>> Is there any information I could provice to help to fix this? >>> >>> The binary driver from canon.uk works with their gui-frontend only and >>> I'd like to get it working on a headless system. >>> >>> Thanks for any comments. >>> >>> Bye >>> Matthias Walther >>> >>> > > -------------- next part -------------- --- ./pixma_mp150.c 2014-03-15 19:33:34.000000000 +0100 +++ ../sane-backends/backend/pixma_mp150.c 2014-03-15 19:52:05.000000000 +0100 @@ -1395,7 +1395,9 @@ tmo = 10; if (s->param->adf_pageid == 0 || mp->generation <= 2) - { + { /* single sheet or first sheet from ADF */ + /* TODO: s->param->adf_pageid != 0 @ next scan session! */ + PDBG (pixma_dbg (4, "*mp150_scan***** start scanning *****\n")); error = start_session (s); while (error == PIXMA_EBUSY && --tmo >= 0) { @@ -1436,7 +1438,11 @@ error = send_set_tpu_info (s); } else /* ADF pageid != 0 and gen3 or above */ + { /* next sheet from ADF */ + PDBG (pixma_dbg (4, "*mp150_scan***** scan next sheet from ADF *****\n")); pixma_sleep (1000000); + } + if ((error >= 0) || (mp->generation >= 3)) mp->state = state_warmup; @@ -1485,7 +1491,13 @@ if (s->cancel) return PIXMA_ECANCELED; if ((mp->last_block & 0x28) == 0x28) - { /* end of image */ + { /* end of image */ + /* TODO: reset ADF page counter here! + * there are 2 counters: s->param->adf_pageid == sp->adf_pageid and + * ss->page_count (see pixma.c) + * OR + * define a new parameter in mp150_t for new session detection in mp150_scan() */ + PDBG (pixma_dbg (4, "*mp150_fill_buffer***** end of image *****\n")); mp->state = state_finished; return 0; } @@ -1557,6 +1569,7 @@ * abort_session and start_session between pages (last_block=0x28) */ if (mp->generation <= 2 || !is_scanning_from_adf (s) || mp->last_block == 0x38) { + PDBG (pixma_dbg (4, "*mp150_finish_scan***** abort session *****\n")); error = abort_session (s); /* FIXME: it probably doesn't work in duplex mode! */ if (error < 0) PDBG (pixma_dbg (1, "WARNING:abort_session() failed %d\n", error)); @@ -1568,6 +1581,9 @@ PDBG (pixma_dbg (1, "WARNING:XML_END dialog failed \n")); } } + else + PDBG (pixma_dbg (4, "*mp150_finish_scan***** wait for next page from ADF *****\n")); + mp->state = state_idle; /* fall through */ case state_idle:
