Read card synchronously after reader disconnect/reconnect

[didier....@gmail.com]

Hi all,

We are building a fat client which has to read beID's when the user clicks on a button. For this, we instantiate a BeIDCards, call getAllBeIDCards() and then read the data.

This works pretty well, except after a reader disconnection and reconnection. In that case, we cannot read cards any more and have to restart the application. When trying to read a card we get the following exception:

java.lang.IllegalStateException: Card has been disconnected
      at sun.security.smartcardio.CardImpl.checkState(Unknown Source)
      at sun.security.smartcardio.CardImpl.beginExclusive(Unknown Source)
      at be.fedict.commons.eid.client.BeIDCard.beginExclusive(BeIDCard.java:915)
      at be.fedict.commons.eid.client.BeIDCard.readFile(BeIDCard.java:1034)

(if no card is present we get a “No card found” instead)

The attached file demonstrates the issue. Notice we don't use getOneBeIDCard() because it makes several dialogs flicker before crashing so this is even worse.

The problem seems to be that BeIDCards.close() does not stop the CardAndTerminalManager. However re-using the BeIDCards instance does not work either (getAllBeIDCards() returns an empty set). The only thing that seems to work is accessing the CardAndTerminalManager by reflexion and stop it…

Are we doing something wrong or is this a bug in the synchronous api ?

Best regards,

Didier Loiseau

[fr4...@gmail.com]

FYI: I'll have a look at this afternoon.
BeIDCards.close() is supposed to stop the CardAndTerminalManager if it's a private instance that it created itself, as
in your example.

[Frank Cornelis]

Hi Didier,


I've added an integration test for this case based on your sample code:
http://code.google.com/p/commons-eid/source/browse/commons-eid-tests/src/test/java/test/integ/be/fedict/commons/eid/client/DisconnectConnectTest.java

Under Linux I can disconnect and reconnect my card reader without any problem. The second identity read out works just fine.
Which operating system are you using?


Kind Regards,
Frank.

--
You received this message because you are subscribed to the Google Groups "eID Applet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eid-applet+...@googlegroups.com.
To post to this group, send email to eid-a...@googlegroups.com.
Visit this group at http://groups.google.com/group/eid-applet.
For more options, visit https://groups.google.com/groups/opt_out.

[didier....@gmail.com]

Hi Frank,

I am having this issue on Windows 7 x64, with an ACS ACR38U-CFC-BLi USB card reader.

[Frank Cornelis]

Hi Didier,


I can confirm that there is an issue with the card reader reconnection under Windows 8, 64 bit, Java 7u45. But the exception that I'm getting is totally different.


Kind Regards,
Frank.

On 10/22/2013 10:05 AM, didier....@gmail.com wrote:

Hi Frank,

I am having this issue on Windows 7 x64, with an ACS ACR38U-CFC-BLi USB card reader.

Le mardi 22 octobre 2013 09:48:11 UTC+2, Frank Cornelis a �crit�:

Hi Didier,


I've added an integration test for this case based on your sample code:
http://code.google.com/p/commons-eid/source/browse/commons-eid-tests/src/test/java/test/integ/be/fedict/commons/eid/client/DisconnectConnectTest.java

Under Linux I can disconnect and reconnect my card reader without any problem. The second identity read out works just fine.
Which operating system are you using?


Kind Regards,
Frank.


On 10/21/2013 06:46 PM, didier....@gmail.com wrote:

Hi all,

We are building a fat client which has to read beID's when the user clicks on a button. For this, we instantiate a BeIDCards, call getAllBeIDCards() and then read the data.

This works pretty well, except after a reader disconnection and reconnection. In that case, we cannot read cards any more and have to restart the application. When trying to read a card we get the following exception:

java.lang.IllegalStateException: Card has been disconnected
� � � at sun.security.smartcardio.CardImpl.checkState(Unknown Source)
� � � at sun.security.smartcardio.CardImpl.beginExclusive(Unknown Source)
� � � at be.fedict.commons.eid.client.BeIDCard.beginExclusive(BeIDCard.java:915)
� � � at be.fedict.commons.eid.client.BeIDCard.readFile(BeIDCard.java:1034)

(if no card is present we get a �No card found� instead)

The attached file demonstrates the issue. Notice we don't use getOneBeIDCard() because it makes several dialogs flicker before crashing so this is even worse.

The problem seems to be that BeIDCards.close() does not stop the CardAndTerminalManager. However re-using the BeIDCards instance does not work either (getAllBeIDCards() returns an empty set). The only thing that seems to work is accessing the CardAndTerminalManager by reflexion and stop it�

Are we doing something wrong or is this a bug in the synchronous api ?

Best regards,

Didier Loiseau

--
You received this message because you are subscribed to the Google Groups "eID Applet" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eid-applet+...@googlegroups.com.
To post to this group, send email to eid-a...@googlegroups.com.
Visit this group at http://groups.google.com/group/eid-applet.
For more options, visit https://groups.google.com/groups/opt_out.

[didier....@gmail.com]

It seems indeed that depending on how the API is used, the errors might differ.

I'm currently trying with the asynchronous API, but after reconnecting the reader, I'm getting continuous card insertion/removals.
If I enable debug logging:

• without card reader, I get this continuously (this comes from the this.cardTerminals.waitForChange(this.delay); in CardAndTerminalManager)
  
  2013-10-22 12:59:56,573 [DEBUG] BeidServiceImpl$BeidLogger - subsystem not initialized
  2013-10-22 12:59:56,574 [DEBUG] BeidServiceImpl$BeidLogger - Cannot enumerate card terminals [1] (No Card Readers Connected?): list() failed
  2013-10-22 12:59:56,574 [DEBUG] BeidServiceImpl$BeidLogger - no card readers connected?
  2013-10-22 12:59:56,574 [DEBUG] BeidServiceImpl$BeidLogger - cause: SCARD_E_NO_READERS_AVAILABLE
  2013-10-22 12:59:56,574 [DEBUG] BeidServiceImpl$BeidLogger - cause type: sun.security.smartcardio.PCSCException
  2013-10-22 12:59:56,574 [DEBUG] BeidServiceImpl$BeidLogger - cleared
  
• once I connect a card reader, this hopefully stops and I can read cards
  
• if I disconnect the card reader, this output restarts (as expected)
• if I reconnect the card reader, it does not stop
• if I insert a card afterwards, it is properly read, but my listener is continuously notified of the card insertion and removal:
  
  2013-10-22 13:06:48,223 [DEBUG] BeidServiceImpl$BeidLogger - subsystem not initialized
  2013-10-22 13:06:48,227 [DEBUG] BeidServiceImpl$BeidLogger - Cannot wait for card terminal events [2] (No Card Readers Connected?): waitForChange() failed
  2013-10-22 13:06:48,227 [DEBUG] BeidServiceImpl$BeidLogger - no card readers connected?
  2013-10-22 13:06:48,227 [DEBUG] BeidServiceImpl$BeidLogger - cause: SCARD_E_NO_READERS_AVAILABLE
  2013-10-22 13:06:48,227 [DEBUG] BeidServiceImpl$BeidLogger - cause type: sun.security.smartcardio.PCSCException
  2013-10-22 13:06:48,227 [DEBUG] BeidServiceImpl$BeidLogger - cleared
  2013-10-22 13:06:48,477 [DEBUG] BeidServiceImpl$BeidLogger - subsystem not initialized
  ----- card inserted now
  2013-10-22 13:06:49,082 [INFO ] BeidServiceImpl - Card inserted in ACS ACR38U 0
  2013-10-22 13:06:49,083 [DEBUG] BeidServiceImpl$BeidLogger - ---begin exclusive---
  2013-10-22 13:06:49,200 [DEBUG] BeidServiceImpl$BeidLogger - selecting file
  2013-10-22 13:06:49,316 [DEBUG] BeidServiceImpl$BeidLogger - read binary
  2013-10-22 13:06:49,471 [DEBUG] BeidServiceImpl$BeidLogger - ---end exclusive---
  2013-10-22 13:06:49,473 [DEBUG] DateOfBirthDataConvertor - "XXXXXXXX"
  2013-10-22 13:06:49,473 [DEBUG] DateOfBirthDataConvertor - day: "XX"
  2013-10-22 13:06:49,473 [DEBUG] DateOfBirthDataConvertor - month: "XXXX"
  2013-10-22 13:06:49,474 [DEBUG] DateOfBirthDataConvertor - year: "XXXX"
  2013-10-22 13:06:49,474 [DEBUG] DocumentTypeConvertor - # bytes for document type field: 2
  2013-10-22 13:06:49,474 [DEBUG] BeidServiceImpl$BeidLogger - ---begin exclusive---
  2013-10-22 13:06:49,474 [DEBUG] BeidServiceImpl$BeidLogger - selecting file
  2013-10-22 13:06:49,592 [DEBUG] BeidServiceImpl$BeidLogger - read binary
  2013-10-22 13:06:49,718 [DEBUG] BeidServiceImpl$BeidLogger - ---end exclusive---
  2013-10-22 13:06:49,720 [DEBUG] BeidServiceImpl$BeidLogger - ---begin exclusive---
  2013-10-22 13:06:49,863 [DEBUG] BeidServiceImpl$BeidLogger - selecting file
  2013-10-22 13:06:49,985 [DEBUG] BeidServiceImpl$BeidLogger - read binary
  2013-10-22 13:06:51,506 [DEBUG] BeidServiceImpl$BeidLogger - ---end exclusive---
  2013-10-22 13:06:51,507 [INFO ] BeidServiceImpl - Card read in 2424ms
  2013-10-22 13:06:51,526 [DEBUG] BeidServiceImpl$BeidLogger - Cannot wait for card terminal events [2] (No Card Readers Connected?): waitForChange() failed
  2013-10-22 13:06:51,526 [DEBUG] BeidServiceImpl$BeidLogger - no card readers connected?
  2013-10-22 13:06:51,526 [DEBUG] BeidServiceImpl$BeidLogger - cause: SCARD_E_NO_READERS_AVAILABLE
  2013-10-22 13:06:51,526 [DEBUG] BeidServiceImpl$BeidLogger - cause type: sun.security.smartcardio.PCSCException
  2013-10-22 13:06:51,526 [DEBUG] BeidServiceImpl$BeidLogger - closing eID card
  2013-10-22 13:06:51,527 [INFO ] BeidServiceImpl - Card removed from ACS ACR38U 0
  2013-10-22 13:06:51,527 [DEBUG] BeidServiceImpl$BeidLogger - cleared
  2013-10-22 13:06:51,777 [DEBUG] BeidServiceImpl$BeidLogger - subsystem not initialized
  2013-10-22 13:06:51,778 [INFO ] BeidServiceImpl - Card inserted in ACS ACR38U 0
  2013-10-22 13:06:51,778 [DEBUG] BeidServiceImpl$BeidLogger - ---begin exclusive---
  2013-10-22 13:06:51,779 [DEBUG] BeidServiceImpl$BeidLogger - selecting file
  2013-10-22 13:06:51,892 [DEBUG] BeidServiceImpl$BeidLogger - read binary
  2013-10-22 13:06:52,043 [DEBUG] BeidServiceImpl$BeidLogger - ---end exclusive---
  …
  

So it seems that the reader is detected by the JVM but is not properly handled.

BTW, there are issues with the Logger handling:

• the BeIDCardManager does not transmit the provided Logger when it instantiates the CardAndTerminalManager so I had to instantiate it myself
• the DateOfBirthDataConvertor and DocumentTypeConvertor are using ACL instead (it would actually be great if the entire API relied only on JCL, ACL or another standard logging framework btw - SLF4J maybe?)
  

[Didier Loiseau]

I think the problem is that the CardTerminals instance gets somehow corrupted after a terminal gets disconnected. I have tested this with 2 readers: you can disconnect 1 without issue, but even if you reconnect it, after disconnecting the second you begin to have the repeating SCARD_E_NO_READERS_AVAILABLE.

As a workaround, I added a CardTerminalEventsListener which, on terminalDetached, spawns a new thread that stops and recreates the CardAndTerminalManager and BeIDCardManager (this must be done on a separate thread, otherwise it interrupts itself when calling stop()). This seems to work without issues.

[Frank Cornelis]

Hi Didier,


Interesting. Under Windows 8, even the following fails for me:

��� @Test

��� public void testPCSC() throws Exception {

������� JOptionPane.showMessageDialog(null, "Connect reader.");

������� TerminalFactory terminalFactory = TerminalFactory.getDefault();

������� LOG.debug("number of readers: " 

��������������� + terminalFactory.terminals().list().size());

������� JOptionPane.showMessageDialog(null, "Disconnect and reconnect reader.");

������� terminalFactory = TerminalFactory.getDefault();

������� LOG.debug("number of readers: " 

��������������� + terminalFactory.terminals().list().size());

��� }

From the second terminals() call I receive SCARD_E_SERVICE_STOPPED. So under Windows 8 it seems that it's the JVM itself that does not properly recover from the PC/SC going (temporary) down.


BTW: about the design choices behind the used logging framework: commons-eid-client is using its own logging interface as the client library cannot have any 3rd dependency (since we want to be able to construct very lightweight applets), while for commons-eid-consumer it doesn't really matter.


Kind Regards,
Frank.

On 10/22/2013 03:54 PM, Didier Loiseau wrote:

I think the problem is that the CardTerminals instance gets somehow corrupted after a terminal gets disconnected. I have tested this with 2 readers: you can disconnect 1 without issue, but even if you reconnect it, after disconnecting the second you begin to have the repeating SCARD_E_NO_READERS_AVAILABLE.

As a workaround, I added a CardTerminalEventsListener which, on terminalDetached, spawns a new thread that stops and recreates the CardAndTerminalManager and BeIDCardManager (this must be done on a separate thread, otherwise it interrupts itself when calling stop()). This seems to work without issues.

Le mardi 22 octobre 2013 13:20:19 UTC+2, Didier Loiseau a �crit�:

• �
  

So it seems that the reader is detected by the JVM but is not properly handled.

BTW, there are issues with the Logger handling:

• the BeIDCardManager does not transmit the provided Logger when it instantiates the CardAndTerminalManager so I had to instantiate it myself
• the DateOfBirthDataConvertor and DocumentTypeConvertor are using ACL instead (it would actually be great if the entire API relied only on JCL, ACL or another standard logging framework btw - SLF4J maybe?)
  

Le mardi 22 octobre 2013 11:03:54 UTC+2, Frank Cornelis a �crit�:

Hi Didier,


I can confirm that there is an issue with the card reader reconnection under Windows 8, 64 bit, Java 7u45. But the exception that I'm getting is totally different.


Kind Regards,
Frank.

On 10/22/2013 10:05 AM, didier....@gmail.com wrote:

Hi Frank,

I am having this issue on Windows 7 x64, with an ACS ACR38U-CFC-BLi USB card reader.

--

[Frank Cornelis]

FYI... it's already reported at:
https://bugs.openjdk.java.net/browse/JDK-8026326

[Didier Loiseau]

I don't have this issue with your test on W7. However if I don't reconnect the reader it throws a PCSCException: SCARD_E_NO_READERS_AVAILABLE

I have tested this with JDK 1.6.0_51 (x86) and JDK 1.7.0_21 (x64).

2013/10/22 Frank Cornelis <frank.c...@gmail.com>

FYI... it's already reported at:
https://bugs.openjdk.java.net/browse/JDK-8026326

On 10/22/2013 05:14 PM, Frank Cornelis wrote:

Hi Didier,


Interesting. Under Windows 8, even the following fails for me:

    @Test

    public void testPCSC() throws Exception {

        JOptionPane.showMessageDialog(null, "Connect reader.");

        TerminalFactory terminalFactory = TerminalFactory.getDefault();

        LOG.debug("number of readers: " 

                + terminalFactory.terminals().list().size());

        JOptionPane.showMessageDialog(null, "Disconnect and reconnect reader.");

        terminalFactory = TerminalFactory.getDefault();

        LOG.debug("number of readers: " 

                + terminalFactory.terminals().list().size());

    }

From the second terminals() call I receive SCARD_E_SERVICE_STOPPED. So under Windows 8 it seems that it's the JVM itself that does not properly recover from the PC/SC going (temporary) down.


BTW: about the design choices behind the used logging framework: commons-eid-client is using its own logging interface as the client library cannot have any 3rd dependency (since we want to be able to construct very lightweight applets), while for commons-eid-consumer it doesn't really matter.


Kind Regards,
Frank.

On 10/22/2013 03:54 PM, Didier Loiseau wrote:

I think the problem is that the CardTerminals instance gets somehow corrupted after a terminal gets disconnected. I have tested this with 2 readers: you can disconnect 1 without issue, but even if you reconnect it, after disconnecting the second you begin to have the repeating SCARD_E_NO_READERS_AVAILABLE.

As a workaround, I added a CardTerminalEventsListener which, on terminalDetached, spawns a new thread that stops and recreates the CardAndTerminalManager and BeIDCardManager (this must be done on a separate thread, otherwise it interrupts itself when calling stop()). This seems to work without issues.

• …
  

--
You received this message because you are subscribed to a topic in the Google Groups "eID Applet" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/eid-applet/N1mVFnYJ3VM/unsubscribe.
To unsubscribe from this group and all its topics, send an email to eid-applet+...@googlegroups.com.

To post to this group, send email to eid-a...@googlegroups.com.
Visit this group at http://groups.google.com/group/eid-applet.
For more options, visit https://groups.google.com/groups/opt_out.

--
Boulevard Auguste Reyers 201
1030 Schaerbeek (Belgique)
0479 746 279