Merge cardmod reader driver with pcsc reader driver

[frankmorgner]

Expected behaviour

PC/SC reader driver should be identical to the reader driver that's used in the minidriver ("cardmod").

Actual behaviour

• cardmod reader driver searches for the almost undocumented "cardmod" block in opensc.conf.
• Since e95b515 the user needs to configure max_send_size and max_recv_size for both, the pcsc reader and the cardmod reader (while the latter is undocumented, as said above) for a reader that doesn't announce dwMaxAPDUDataSize.
• cardmod reader configuration and implementation misses features of what is already implemented in PC/SC (e.g. PACE)
• there is a lot of code duplication in reader-pcsc.c
• reader-pcsc.c contains comments about removing cardmod driver somewhere in the future.

How to proceed?

Merge cardmod and PC/SC driver! Remove some code and use the same configuration blocks.

Since the cardmod driver has been there for years, I'd like to know if the early maintainers are fine with this... @martinpaljak @dengert @viktorTarasov

[martinpaljak]

Absolutely. I practically gave up on OpenSC after the merge of this one.

[dengert]

Maybe I am missing something here. The cardmod driver is in ./reader-pcsc.c.

The main difference is reader-pcsc.c looks at all readers PC/SC provides and presents all the readers to the calling application so the application can select a reader.

Cardmod driver is given a pcsc_context_handle and pcsc_card_handle Microsoft code has already selected the reader and opened a PC/SC connection and passed these to the minidriver. And while the minidriver is still running the Microsoft code can presents a new pcsc_context_handle and pcsc_card_handle. The Microsoft code also handles the PCSC locking as best I can tell.

The main difference is thus:
cardmod_ops.detect_readers = NULL;
cardmod_ops.use_reader = cardmod_use_reader;

"Merge cardmod and PC/SC driver! Remove some code and use the same configuration blocks." is a start.

[frankmorgner]

OK, great! Thanks for the feedback.