[Pcsclite-muscle] question about locking and reconnect

[Ricardo Martin]

Hi Ludovic,
I have tested a similar sample.c (attached) in a windows vista (sorry, it's the only windows I have) with Windows SDK 6.1 and it is clear that the SCardRconnect clears the connections.? Compilation and output is presented below:

C:\Users\ricky\Desktop>cl sample.c Winscard.lib
Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 15.00.21022.08 for 80x86
Copyright (C) Microsoft Corporation.? All rights reserved.

sample.c
Microsoft (R) Incremental Linker Version 9.00.21022.08
Copyright (C) Microsoft Corporation.? All rights reserved.

/out:sample.exe
sample.obj
Winscard.lib
C:\Users\ricky\Desktop>sample.exe
reader name: Gemplus USB SmartCard Reader 0
response: 86 CC 40 80 7A 68 C9 90 00
response: 86 CC 40 80 7A 68 C9 90 00
SCardEndTransaction: 0x80100016 (0x80100016)
So the SCardEndTransaction returns SCARD_E_NOT_TRANSACTED after the reconnect. Besides the documentation says the following:
"The application must immediately call the SCardDisconnect, SCardReconnect, or SCardReleaseContext function to avoid an existing transaction blocking other threads and processes from communicating with the smart card."

See here: https://msdn.microsoft.com/es-es/library/windows/desktop/aa379477%28v=vs.85%29.aspx

Regards!
     On Monday, June 1, 2015 2:52 PM, Ludovic Rousseau <ludovic.rousseau at gmail.com> wrote:
   

 2015-06-01 13:57 GMT+02:00 Ricardo Martin <rickyepoderi at yahoo.es>:
> Hi everybody,

Hello,

> Playing with opensc and the dnie driver I found a weird behavior. After some
> tests I opened the following issue in opensc github:
>
> https://github.com/OpenSC/OpenSC/issues/475
>
> It seems that opensc in the method sc_reset finally sends a SCardReconnect
> with SCARD_UNPOWER_CARD or SCARD_RESET_CARD (depending the do_cold_reset
> parameter). But at the end of the sc_reset function opensc performs the
> following:
>
>? ? ? ? /* pcsc_reconnect unlocks card... try to lock it again if it was
> locked */
>? ? ? ? if(old_locked)
>? ? ? ? ? ? ? ? r = pcsc_lock(reader);
>
> If the connection was previously locked, after the reconnect, it adds a lock
> again (SCardBeginTransaction is called again inside pcsc_lock). My
> impression is that this extra call adds one more lock in the pcsc
> (rContext->LockCount) which is never released, and this ends in the complete
> lock of the card until the current process finishes (no other process can
> access the card in shared mode). Every call to pcsc_reset in opensc seems to
> add a lost lock in pcsc (and the dnie driver calls to this reset).
>
> I have done some little test with pcsc (with the little sample.c copied
> below) and it seems that SCardReconnect does not lose the locks (in any
> case: SCARD_SHARE_SHARED or SCARD_SHARE_EXCLUSIVE, SCARD_UNPOWER_CARD or
> SCARD_RESET_CARD). But, just to be sure, can anybody confirm that
> SCardReconnect maintains the locks (transactions)? Is the extra call done in
> opensc to SCardBeginTransaction needed?

It looks like you are right.

I checked the source code of pcsc-lite and SCardReconnect() does not
release or take a lock. It only checks if a lock (from another PC/SC
context) is not yet present.

To know if this is a bug or not in pcsc-lite we should know how it
works on Windows.
I could not find any reference to transactions or locks in MSDN
documentation [1].

Can you do some tests on Windows and tell us how SCardReconnect()
behaves on Windows?

Thanks

[1] https://msdn.microsoft.com/en-us/library/windows/desktop/aa379797%28v=vs.85%29.aspx

-- 
 Dr. Ludovic Rousseau


  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.alioth.debian.org/pipermail/pcsclite-muscle/attachments/20150601/720506bd/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: sample.c
Type: application/octet-stream
Size: 2992 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/pcsclite-muscle/attachments/20150601/720506bd/attachment.obj>