Calibre stopped working with Sony PRS-650 - reports write protection

[phees]

Originally Posted by kovidgoyal

I use 4.9.11 and havent seen any problems, although I no longer have a SONY device to test with, my other devices continue to work fine with it.

I have a PRS-600 and running into the same issue here, BUT:

When I am inserting a SD-Card, Calibre accepts it as "Main" memory and the device works fine.

I would be happy to help debugging this, as I would love to run my Sony reader without SD-card.

It looks as calibre skips my internel memory partition (/dev/sde) and only checks the read-only LAUNCH/SETTINGS (/dev/sdh) and SD-card (/dev/sdg) partitions. Somehow it looks as it only starts looking from /dev/sdf on (which is the reader's MemoryStick driver, which has no media in it on my device.)

How can I help, what other information do you need? (I am on Linux Kernel 4.12 by the way)


Here is a debug-log:

calibre Debug log
calibre 3.7  embedded-python: True is64bit: True
Linux-4.12.8-2-MANJARO-x86_64-with-glibc2.2.5 Linux ('64bit', 'ELF')
('Linux', '4.12.8-2-MANJARO', '#1 SMP PREEMPT Sun Aug 20 08:24:42 UTC 2017')
Python 2.7.12
Linux: ('', '', '')
Interface language: None
calibre 3.7  embedded-python: True is64bit: True
Linux-4.12.8-2-MANJARO-x86_64-with-glibc2.2.5 Linux ('64bit', 'ELF')
('Linux', '4.12.8-2-MANJARO', '#1 SMP PREEMPT Sun Aug 20 08:24:42 UTC 2017')
Python 2.7.12
Linux: ('', '', '')
Interface language: None
Turning on automatic hidpi scaling
QGtkStyle could not resolve GTK. Make sure you have installed the proper libraries.
devicePixelRatio: 1.0
logicalDpi: 95.8921348315 x 95.9352517986
physicalDpi: 90.0253164557 x 90.1013513514
Using calibre Qt style: True
QGtkStyle could not resolve GTK. Make sure you have installed the proper libraries.
Starting up...
Failed to read metadata from font file: /usr/share/fonts/TTF/TSCu_Times.ttf tuple index out of range
Failed to read metadata from font file: /usr/share/fonts/TTF/TSCu_Comic.ttf tuple index out of range

Ignoring the node: sdf as could not read size from: /sys/devices/pci0000:00/0000:00:01.3/0000:03:00.0/usb1/1-9/1-9:1.0/host12/target12:0:0/12:0:0:1/block/sdf/size

Found device nodes: /dev/sdg1 /dev/sdh None

The mountpoint /run/media/user/LAUNCHER/ is readonly, ignoring it
Job: 1 Get device information started
DeviceJob: 1 Get device information done, calling callback
DeviceJob: 1 Get device information callback returned
Job: 2 Set library information started
DeviceJob: 2 Set library information done, calling callback
DeviceJob: 2 Set library information callback returned
Job: 3 Get list of books on device started
DEBUG:    0.0 PRS505: starting fetching books for card None
DEBUG:    0.0 USBMS: Fetching list of books from device. Device= PRS505 oncard= None
DEBUG:    0.0 USBMS: dirs are: /run/media/user/C5AC-1301/ database/media/books
DEBUG:    0.0 USBMS: scan from root True /run/media/user/C5AC-1301/
DEBUG:    0.0 USBMS: count found in cache: 11, count of files in metadata: 11, need_sync: False
DEBUG:    0.0 USBMS: Finished fetching list of books from device. oncard= None
DEBUG:    0.0 Building XMLCache... {0: '/run/media/user/C5AC-1301/database/cache/media.xml'}
DEBUG:    0.0 Done building XMLCache...
DEBUG:    0.0 Updating JSON cache: 0
DEBUG:    0.0 Start build_id_playlist_map
DEBUG:    0.0 after cleaning playlists
DEBUG:    0.0 Finish build_id_playlist_map. Found 9
DEBUG:    0.0 Finished updating JSON cache: 0
DEBUG:    0.0 PRS505: finished fetching books for card None
DEBUG:    0.0 PRS505: starting fetching books for card carda
DEBUG:    0.0 USBMS: Fetching list of books from device. Device= PRS505 oncard= carda
DEBUG:    0.0 Building XMLCache... {0: '/run/media/user/C5AC-1301/database/cache/media.xml'}
DEBUG:    0.0 Done building XMLCache...
DEBUG:    0.0 PRS505: finished fetching books for card carda
DEBUG:    0.0 PRS505: starting fetching books for card cardb
DEBUG:    0.0 USBMS: Fetching list of books from device. Device= PRS505 oncard= cardb
DEBUG:    0.0 Building XMLCache... {0: '/run/media/user/C5AC-1301/database/cache/media.xml'}
DEBUG:    0.0 Done building XMLCache...
DEBUG:    0.0 PRS505: finished fetching books for card cardb
DeviceJob: 3 Get list of books on device done, calling callback
DeviceJob: metadata_downloaded: Starting set_books_in_library
DeviceJob: set_books_in_library: books to process= 11
DeviceJob: set_books_in_library finished: time= 0.00396800041199
DeviceJob: metadata_downloaded: updating views
DeviceJob: metadata_downloaded: syncing
DeviceJob: metadata_downloaded: refreshing ondevice
DeviceJob: metadata_downloaded: sending metadata_available signal
DeviceJob: 3 Get list of books on device callback returned
Job: 1 Get device information finished
No details available.
Job: 2 Set library information finished
No details available.
Job: 3 Get list of books on device finished
No details available.
Started up in 10.20 seconds with 292 books

Just to make sure, I looked at the three partitions. Linux lables the Model as:
"Sony PRS-600" for my /dev/sde, which is the Main memory
"Sony PRS-600 SD" for my /dev/sdg, which is the SD card.
"Sony PRS-600 Launcher" for my /dev/sdh, which is the system partition (LAUNCHER/SETTINGS)

===============================

There might also be a typo in the driver, but I am not sure:

The device driver.py for the PRS505 class might have an error in line 39:
Current line:

    WINDOWS_MAIN_MEM   = re.compile(
            r'(PRS-(505|500|300))|'
            r'(PRS-((700[#/])|((6|9|3)(0|5)0&)))'
            )

But it should be:

    WINDOWS_MAIN_MEM   = re.compile(
            r'(PRS-(505|500|300))|'
            r'(PRS-((700[#/])|((6|9|3)(0|5)0)))'
            )

The "&" might prevent calibre from detecting the MAIN partition. I guess "re.fullmatch" is being used and would not match "PRS-600" with WINDOWS_MAIN_MEM. But sadly that is not the solution. Also, from the object name, I guess this is only used on Windows machines?
Anyway, I think calibre is not checking all devices. Maybe because there is a gap in my device list? (/dev/sdf is not being assigned!)

As I said, I can deliver you all information you need, just let me know.

One last thing: When I use "connect to folder" and point to my device's internal/Main memory, that also works without a hitch since it is in no way mounted as read-only.

[kovidgoyal]

Given that this device line is discontinued, I'm not particualrly interested in spending time on it myself. But I'm happy to provide guidance on how to fix it yourself. The problem is likely the read-only launcher partition being detected as the main memory. On windows this is filtered by windows_filter_pnp_id

Linux has no equivalent mechanism, that I am aware of, anyway.

The relevant function you will need to play with is find_device_nodes() in usbms/device.py

WINDOWS_MAIN_MEM is a red herring, it is not used for anything in modern calibre.

[selenetic]

Hello, I have the same issue with my Sony PRS-650. I took a look at the function suggested find_device_nodes() in usbms/device.py. I'm pretty much a novice at any of this sort of thing, and relatively unfamiliar with python (but familiar enough with code to work my way through it).

But it seems the problems is in that function there's the block starting on 511

        devnodes += list(repeat(None, 3))
        ans = ['/dev/'+x if ok.get(x, False) else None for x in devnodes[:3]]
        ans.sort(key=lambda x: x[5:] if x else 'zzzzz')
        return self.linux_swap_drives(ans)

The function assumes that devnodes will be at most 3 entries long, but the sony reader has four devices associated with it. The READER drive, the main memory, is being cut off by devnodes[:3].

So what i did for the moment is edit usbms/device.py to this effect. Seems to be working. Don't know if its optimal or not, but it works for sony reader.

        devnodes += list(repeat(None, 3))
        ans = ['/dev/'+x if ok.get(x, False) else None for x in devnodes]
        ans.sort(key=lambda x: x[5:] if x else 'zzzzz')
        return self.linux_swap_drives(ans[:3])

[kovidgoyal]

That seems correct: https://github.com/kovidgoyal/calibr...824773605f8559

[phees]

Confirmed on two machines, fix works!
Thanks @selenetic for the quick work!