Ludovic Rousseau's blog: 2022

Friday, December 30, 2022

New version of pcsc-tools: 1.6.1

I just released a new version of pcsc-tools, a suite of tools for PC/SC.

Changes:

1.6.1 - 30 December 2022, Ludovic ROUSSEAU

196 new ATRs
use colors on NetBSD ("wsvt25" terminal)
pcsc_scan: always print the version in verbose mode

Tuesday, December 27, 2022

PySCard (smartcard Python wrapper) API documentation update

The documentation for the PySCard API was not updated since 2017 (5 years ago).

epydoc → pydoctor

The tool used to generate the HTML documentation is epydoc. The latest epydoc version was released in 2008 (14 years ago). This tool is no more available in Debian so I had to move to something else.

The (new) tool to replace epydoc is pydoctor. It is actively maintained and generates nicer HTML pages.

For example, nice changes are:

use of HTML list
use syntax colorization for sample source code
provide indexes for names, classes and modules
no use of HTML frame so the URL is always correct and can be shared

Previous documentation

The old (epydoc) documentation is still available online at https://pyscard.sourceforge.io/epydoc/

New documentation

Conclusion

The new API documentation is available at https://pyscard.sourceforge.io/apidocs/index.html or from the project main documentation page https://pyscard.sourceforge.io/index.html

I guess I could improve the documentation. Feel free to report issues or suggest improvements.

Saturday, December 17, 2022

AlcorMicro AU9560 reader and fast smart cards: help from crowd needed

The AlcorMicro AU9560 smart card reader has problems with high speed smart cards.

The reported problem

The same problem is already reported in different places:

Example of cards in reported problems

The reader is always the same but used with different cards.

The card must have a TA₁ > 0x95. To know the TA₁ of your smart card you need to get and parse the card's ATR, for example using pcsc_scan program from pcsc-tools.

The problem

The reader declares it can support card/reader communication speed up to 688 172 bps.
See dwMaxDataRate field in AlcorMicro_AU9560.txt.

It is fast but I have 4.5 % of readers in my list that declare to be faster (not including contactless readers).

But speeds above ~200 000 bps are problematic with the AlcorMicro AU9560. It works fine, until an APDU exchange fails with a "Hardware error" (for example) message from the reader:

00000009 APDU: 00 A4 00 00 02 3F 00 00
00000007 ifdhandler.c:1333:IFDHTransmitToICC() usb:058f/9540:libudev:0:/dev/bus/usb/001/002 (lun: 0)
00000006 commands.c:1670:CmdXfrBlockAPDU_extended() T=0 (extended): 8 bytes
00000017 -> 000000 6F 08 00 00 00 00 0E 00 00 00 00 A4 00 00 02 3F 00 00
02734396 <- 000000 80 00 00 00 00 00 0E 41 FB 00
00000025 commands.c:1563:CCID_Receive Hardware error
00000009 APDU: 00 A4 00 00 02 3F 00 00

Where are the bogus readers?

The AlcorMicro AU9560 is not a stand alone USB reader. It must be integrated in a laptop. For example it is present in these laptop models:

Lenovo Thinkpad P17
Lenovo Thinkpad L15
Lenovo X1 Extreme Gen 2
HP Zbook
HP EliteBook.

I guess the same AlcorMicro AU9560 reader is present in a lot of other Lenovo or HP laptops and also laptops from other brands.

Patch

I worked on a patch to remove the highest speeds so that the CCID driver will negotiate a lower speed that is still supported by the reader.

Problem with the patch

My patch works fine with one card I have (NXP JCOP 4). But it generates problems with another card (Acos-ID).

The error occurs when the driver set the communication speed. The Set Parameters commands fails and the driver gets a "Card absent or mute" error.

00000004 [140396399142464] commands.c:2324:SetParameters() length: 7 bytes
00000006 [140396399142464] -> 000000 61 07 00 00 00 00 07 01 00 00 95 10 FF 46 00 FE 00
01646726 [140396415927872] ccid_usb.c:1532:InterruptRead() after (0) (2)
00621370 [140396399142464] <- 000000 82 00 00 00 00 00 07 41 FE 00
00000025 [140396399142464] commands.c:2351:SetParameters Card absent or mute
00000005 [140396399142464] prothandler.c:141:PHSetProtocol() Set PTS failed (612)

I suspect the problem to be specific to this card. But I am not sure. That is why I need your help to test with as much possible combinations as possible.

Your help is welcome

If your have:

a laptop with the AlcorMicro AU9560 smart card reader
a fast enough smart card (i.e. TA₁ > 0x95)

then you can help me.

The AlcorMicro AU9560 and the AlcorMicro AU9540 both use the same USB idProduct value of 0x9540 even if the two readers are a bit different. So even if you have a AU9560 the PC/SC name will be "Alcor Micro AU9540 xx yy". If you do not know what reader you have just suppose you have a AU9560.

You can check the 2 conditions above (AlcorMicro and TA1 value) using the pcsc_scan tools. See

$ pcsc_scan 
Using reader plug'n play mechanism
Scanning present readers...
0: Alcor Micro AU9540 00 00
 
Sat Dec 17 11:02:51 2022
 Reader 0: Alcor Micro AU9540 00 00
  Event number: 0
  Card state: Card inserted, Shared Mode, 
  ATR: 3B DC 96 FF 81 11 FE 80 31 C8 54 43 56 35 07 73 FF A1 C0 3C

ATR: 3B DC 96 FF 81 11 FE 80 31 C8 54 43 56 35 07 73 FF A1 C0 3C
+ TS = 3B --> Direct Convention
+ T0 = DC, Y(1): 1101, K: 12 (historical bytes)
  TA(1) = 96 --> Fi=512, Di=32, 16 cycles/ETU
    250000 bits/s at 4 MHz, fMax for Fi = 5 MHz => 312500 bits/s
  TC(1) = FF --> Extra guard time: 255 (special value)
  TD(1) = 81 --> Y(i+1) = 1000, Protocol T = 1 
-----
  TD(2) = 11 --> Y(i+1) = 0001, Protocol T = 1 
-----
  TA(3) = FE --> IFSC: 254
+ Historical bytes: 80 31 C8 54 43 56 35 07 73 FF A1 C0
[...]
+ TCK = 3C (correct checksum)

Possibly identified card (using /home/rousseau/.cache/smartcard_list.txt):
3B DC 96 FF 81 11 FE 80 31 C8 54 43 56 35 07 73 FF A1 C0 3C
	NXP JCOP 4, J3R200P0X3U/0ZA16CP NXD6.2 (JavaCard)

Please do:

download, build and install the patched version of the CCID driver from https://ccid.apdu.fr/files/ccid-1.5.1-3ac3a1a.tar.bz2
test it fixes the problems you had with the previous CCID driver
test it does not create new problems that were not present with the previous CCID driver
if a new problem appears I would like you to report it including:

the exact computer model you use
the name and ATR of the smart card you use
a complete pcscd trace as documented in How to get support

You can report your results using different channels:

on the MUSCLE mailing list
by email to me
by creating a new issue on the CCID project

Conclusion

Your help will greatly improve support of this reader commonly found in laptops.

Thank you.

[Update: Jan, 20th 2023]

The patch is now included in the CCID driver and will be present in the CCID release 1.5.2 (to be released later).

[Update: Jan, 31th 2023]

The CCID driver version 1.5.2 is now available.

Support of the AlcorMicro AU9560 should be better now.

Tuesday, November 29, 2022

PC/SC sample in TypeScript (Deno)

To continue the list of PC/SC wrappers initiated in 2010 with "PC/SC sample in different languages" I now present a new sample code in Deno a modern runtime for JavaScript and TypeScript.

pcsc-deno

The wrapper is available at https://github.com/cryptographix/pcsc-deno and https://deno.land/x/pcsc

The author is Sean Michael Wykes.

The license is MIT.

I used version 0.4.
This version includes the fixes I proposed for GNU/Linux.

Deno

From Wikipedia Deno article:

Deno is a runtime for JavaScript, TypeScript, and WebAssembly that is based on the V8 JavaScript engine and the Rust programming language. Deno was co-created by Ryan Dahl, who also created Node.js.

Deno explicitly takes on the role of both runtime and package manager within a single executable, rather than requiring a separate package-management program.

Installation

Installation is very easy. First install Deno as documented in https://deno.land/#installation

The PC/SC wrapper will be downloaded and installed automatically at run time.

Source code

import {
  CommandAPDU,
  ContextProvider,
  ISO7816,
  PCSC,
} from 'https://deno.land/x/pcsc/mod.ts';

try {
  // establish a PC/SC context
  const context = ContextProvider.establishContext();

  // get all available readers
  const readers = context.listReaders();

  for (const reader of readers) {
    console.log(`Using reader: ${reader.name}`);
    if (reader.isMute) {
      console.log(`Reader ${reader.name}: MUTE`);
    } else if (reader.isPresent) {
      // connect
      const card = await reader.connect();

      // send Select Applet APDU
      const selectApplet = CommandAPDU
        .from([ISO7816.CLA.ISO, ISO7816.INS.SelectFile, 0x04, 0x00]) // ISO SELECT
        .setData([0xA0, 0x00, 0x00, 0x00, 0x62, 0x03, 0x01, 0x0C, 0x06, 0x01]);

      const resp = await card.transmitAPDU(selectApplet);

      // check for 0x90 0x00
      if (resp.SW == ISO7816.SW.SUCCESS) {
        // success ..
        console.log(`Reader ${reader.name}: applet successfully selected`);

        // send Test APDU
        const command = CommandAPDU
          .from([ISO7816.CLA.ISO, 0, 0, 0]);

        const resp = await card.transmitAPDU(command);
        if (resp.SW == ISO7816.SW.SUCCESS) {
          // success ..
          console.log(`Reader ${reader.name}: Test command successful`);

          // convert from bytes to string and display
          console.log(String.fromCharCode.apply(null, resp.data));
        } else {
          // something went wrong ..
          console.error(
            `Reader ${reader.name}: error SW=${resp.SW.toString(16)}`,
          );
        }
      } else {
        // something went wrong ..
        console.error(
          `Reader ${reader.name}: error SW=${resp.SW.toString(16)}`,
        );
      }

      // unpower and disconnect
      await card.disconnect(PCSC.Disposition.UnpowerCard);
    } else {
      console.log(`Reader ${reader.name}: NO CARD`);
    }
  }
  // release the PC/SC context
  context.shutdown();
} catch (e: PCSCException) {
  console.log(e, "error");
}

Output

$ deno run --unstable --allow-ffi blog.ts
Using reader: Gemalto PC Twin Reader (F8345B4A) 00 00
Reader Gemalto PC Twin Reader (F8345B4A) 00 00: applet successfully selected
Reader Gemalto PC Twin Reader (F8345B4A) 00 00: command successful
Hello world!

Conclusion

Nothing special to say. Thanks Sean for the wrapper.

If you work on a Free Software PC/SC wrapper that is not yet in my list please let me know.

Wednesday, November 16, 2022

Share a smart card reader between a host and its guest VM(s)

As I wrote in "One smart card reader accessible from many computers" it is possible to share a smart card reader between 2 or more systems.

Problem

I recently received a bug report about a problem between pcsc-lite and VirtualBox. When the smart card reader is connected to the VM guest then the kernel on the host reports errors like:

2022-11-11T14:25:01.186983-08:00 track pcscd[2474]: 00000001 eventhandler.c:336:EHStatusHandlerThread() Error communicating to: SCM Microsystems Inc. SCR 3310 [CCID Interface] (53311514247933) 00 00

2022-11-11T14:25:01.186993-08:00 track pcscd[2474]: 00000005 ccid_usb.c:1356:InterruptRead() libusb_submit_transfer failed: LIBUSB_ERROR_IO

2022-11-11T14:25:01.188050-08:00 track kernel: [ 1247.705353][ T2521] usb 1-2: usbfs: process 2521 (pcscd) did not claim interface 0 before use

2022-11-11T14:25:01.188053-08:00 track kernel: [ 1247.705386][ T2521] usb 1-2: usbfs: process 2521 (pcscd) did not claim interface 0 before use

2022-11-11T14:25:01.587034-08:00 track pcscd[2474]: 00400173 ccid_usb.c:865:WriteUSB() write failed (1/2): -1 LIBUSB_ERROR_IO

2022-11-11T14:25:01.587076-08:00 track pcscd[2474]: 00000008 ifdwrapper.c:364:IFDStatusICC() Card not transacted: 612

And after some times (in days) the host kernel crashes.

A Linux kernel crash is never a good thing. pcsc-lite may be very powerful but it can't crash the Linux kernel. Only a bug in the kernel itself can generate a crash. Here I suspect the VirtualBox Linux kernel module to do something bad.

Solution

Instead of connecting the USB smart card reader in the guest VM (and disconnecting it from the host) it is possible to share the smart card reader(s) between the host and guest with some help from pcsc-lite.

Setup

My demo setup:

Host: Debian testing system
Guest: NetBSD 9.3 running inside KVM (Kernel Virtual Machine).

I use 2 very different operating systems, GNU/Linux and NetBSD, on purpose. It is to show it is possible to mix systems.

Host

In the host, no change to the configuration. But we will redirect (inject) /run/pcscd/pcscd.comm in the virtual machine.

On the Debian host I run:

$ ssh -N -R/tmp/pcscd.comm:/run/pcscd/pcscd.comm VMNetBSD

Guest

On the NetBSD VM I use:

$ export PCSCLITE_CSOCK_NAME=/tmp/pcscd.comm

Then I can run any application using pcsc-lite and get access to the smart card(s) and reader(s) from the host. For example:

$ pcsc_scan -c

Wed Nov 16 17:26:55 2022
 Reader 0: Alcor Micro AU9540 00 00
  Event number: 0
  Card state: Card inserted, 
  ATR: 3B A7 00 40 18 80 65 A2 08 01 01 52

With a screenshot:

Limitations

pcsc-lite to pcsc-lite

As I demonstrated the solution is not limited to GNU/Linux. Any Unix system using pcsc-lite can be used. But you must use the same pcsc-lite protocol on both sides.

For example the current protocol version used by pcsc-lite 1.9.9 (current version) is 4.4. It is the same protocol version since pcsc-lite 1.8.24 released in Oct 2018.

macOS or Windows host

It should be technically possible to use Windows or macOS as the host OS. That would involve a new development. Contact me if you need something like that.

Conclusion

No need to disconnect/reconnect the USB reader in the VM. Just share it with the host.

You will be able to use the same smart card at the same time on the two sides. Isn't it nice?

Tuesday, November 15, 2022

macOS Ventura and smart cards status

Ventura (macOS 13.0) is now available since October, 2022.

I will compare this version to the previous one in Monterey I presented in macOS Monterey and smart cards status.

CCID

$ grep -A 1 CFBundleShortVersionString /usr/libexec/SmartCardServices/drivers/ifd-ccid.bundle/Contents/Info.plist
	<key>CFBundleShortVersionString</key>
	<string>1.5.0</string>

The CCID driver has been upgraded from version 1.4.34 as in Monterey to version 1.5.0.

Apple Open Source

The Open Source components included in macOS are listed at https://opensource.apple.com/releases/

In addition to a .tar.gz archive, the source code is also available in a github (acquired by Microsoft in 2018) repository at https://github.com/apple-oss-distributions/SmartcardCCID.

It is then easy to see the patches applied by Apple to the CCID driver:

But the patches have no documentation on the why the patches are needed.

The only obvious patch is ccid-info-plist.patch that changes the value of ifdLogLevel from Info.plist configuration file from 3 (CRITICAL + INFO) to 1 (CRITICAL) in order to generate less logs.

It is also easy to compare two versions. For example the differences between the version for Monterey and the version for Ventura is available as a github diff between tags SmartcardCCID-55028 and SmartcardCCID-55031.

Crypto Token Kit

Nothing special to say. The source code of this part is not available.

My Objective-C sample "PC/SC" sample in Objective-C (synchronous) still builds and works fine.

Security message on first connection

On the first connection of my USB smart card reader I got this dialogue box:

It is nice to see the security improvements.

As expected, I do not get the dialogue box again after I selected "Allow".

Conclusion

No big changes in Ventura for the smart card world.

Monday, November 14, 2022

New version of libccid: 1.5.1

I just released version 1.5.1 of libccid the Free Software CCID class smart card reader driver.

Changes:

1.5.1 - 14 November 2022, Ludovic Rousseau

Add support of

Access IS ATR220 with idProduct: 0x0184
Alcor Link AK9567
Alcor Link AK9572
BLUTRONICS TAURUS NFC
CHERRY SmartTerminal ST-1144
CREATOR CRT-603(CZ1) CCR
Dexon Tecnologias Digitais LTDA DXToken
ESMART Reader ER433x ICC
ESMART Reader ER773x Dual & 1S
Flight system consulting Incredist
Ledger Nano S
Ledger Nano S Plus
Ledger Nano SP
Ledger Nano X
SafeNet eToken Fusion
Sensyl SSC-NFC Reader

Adjust USB drivers path at run-time via environment variable PCSCLITE_HP_DROPDIR
configure.ac: add --enable-strict option
Fix a problem with AUTO PPS readers and ATR convention inverse cards
examples/scardcontrol:

add support of 6A xx error codes
check WinSCard error early
parse wLcdLayout & bEntryValidationCondition

macOS: log non sensitive strings as "%{public}s"
Some other minor improvements

Monday, November 7, 2022

Updated CCID driver for UEFI

In 2015 (7 years ago) I ported my CCID driver to UEFI (Unified Extensible Firmware Interface). For example read "UEFI Smart Card Reader Protocol implementation" and "PCSC sample in C for UEFI".

New version

I now updated the driver to use:

source code from CCID driver 1.5.0
source code from pcsc-lite 1.9.9
build using edk2 stable version edk2-stable202208

The driver is no more a patch for edk2 but an independent UEFI driver in its own repository UEFI-SmartCardReader. It should now be easier to build.

I also updated my samples applications in UEFI-SmartCardReader-Samples.

Conclusion

This driver will not be used by a lot of people. The driver is for applications that are run in the UEFI (i.e. before the main operating system is started) and with a need to access smart cards.

If you use it and want to have new features, or just want to discuss, you can contact me. I am curious to know what people can do with it.

Friday, October 28, 2022

Ubuntu 22.04 and pcscd auto start failure

Problem

Since Ubuntu 22.04 LTS there is a problem with PC/SC daemon automatic start. pcscd is supposed to start when an application makes the first PC/SC call. See "pcscd auto start using systemd". In some cases pcscd will not start and SCardEstablishContext() will receive the error SCARD_E_NO_SERVICE.

A ticket is opened at Ubuntu with bug #1971984: pcscd 1.9.5-3 do not start automatically, only manual

The problem is present on some configurations only. I have no idea why on some systems it works fine and on some others we have the problem.

For an unknown reason pcscd.socket is inactive.

$ systemctl status pcscd.socket
○ pcscd.socket - PC/SC Smart Card Daemon Activation Socket
     Loaded: loaded (/lib/systemd/system/pcscd.socket; disabled; vendor preset: enabled)
     Active: inactive (dead)
   Triggers: ● pcscd.service
     Listen: /run/pcscd/pcscd.comm (Stream)

Fix

The fix is easy:

sudo systemctl enable pcscd.socket

And reboot.

No problem on Debian

The same pcscd package has no problem on Debian.

I compared the configuration scripts between the Ubutun and Debian packages and they are the same. I guess the problem comes from systemd or dpkg on Ubuntu and the systemd configuration files are not always installed correctly. Again, no idea why.

Conclusion

The bug is opened since 2022-05-06. I don't know if someone at Ubuntu is working on it.

The pcscd package is in the section Universe (Community-Maintained, Open-Source Software) at Ubuntu. It is not in the section Main (Officially Supported, Open-Source Software) so maybe no engineer from Ubuntu is looking at this issue.

The problem may not be fixed until Ubuntu 24.04, the next LTS version.

Wednesday, September 28, 2022

PySCard 2.0.5 released

I just released a new version 2.0.5 of pyscard. PySCard is a python module adding smart cards support (PC/SC) to Python.

The PySCard project is available at:

This version is a not even a bug fix release. No code has changed.

The problem is that for the previous version, 2.0.4, I uploaded a incorrect source archive to Pypi. I inadvertently included some generated files in the .tar.gz archive. I discovered the problem while creating the Debian package. I removed the incorrect file from Pypi. But then it is not possible to upload a new file with a name that was already present on Pypi.

The source .tar.gz is already present in the pyscard project on sourceforge.net so I thought it was OK.

But then I received bug reports like Missing source release for 2.0.4? or pyscard 2.0.4 not available on linux from pypi so I had to do something.

Changes:

New version just because I failed the previous upload to https://pypi.org/

Sunday, September 11, 2022

New version of pcsc-lite: 1.9.9

I just released a new version of pcsc-lite 1.9.9.
pcsc-lite is a Free Software implementation of the PC/SC (or WinSCard) API for Unix systems.

Changes:

1.9.9: Ludovic Rousseau
11 September 2022

SCardEstablishContext() may return SCARD_W_SECURITY_VIOLATION if refused by Polkit
Fix SCardReleaseContext() failure on orphan handles
Fix SCardDisconnect() on orphan handle
pcsc-spy: log the pioSendPci & pioRecvPci SCardTransmit() parameters
Improve the log from pcscd: log the return code in text instead of hex
Some other minor improvements

Sunday, September 4, 2022

New PyKCS11 1.5.11 available

I just released a new version of PyKCS11, a Python wrapper above the PKCS#11 API.
See "PyKCS11 introduction" or "PyKCS11’s documentation".

The project is registered at Pypi: https://pypi.org/project/PyKCS11/

Changes:

1.5.11 - September 2022, Ludovic Rousseau

add deriveKey() with CKM_ECDH1_DERIVE and CK_ECDH1_DERIVE_PARAMS
support pSourceData in OAEP params
remove use of (deprecated) distutils
samples: port to Python 3
fix code coverage generation

Sunday, August 28, 2022

PySCard 2.0.4 released

I just released a new version 2.0.4 of pyscard. PySCard is a python module adding smart cards support (PC/SC) to Python.

The PySCard project is available at:

This version is a bug fix release.

Changes:

2.0.4 (August 2022)

Fix a problem on PCSCCardConnection.disconnect
Add support of BaseSCardException(hresult) format
Do not use deprecated distutils anymore

Tuesday, July 5, 2022

CCID Windows 10 driver?

Just a quick article to talk about something funny.

Google Search Console Team sent me an email with the top most Google searches used to arrive on my website https://ccid.apdu.fr/.

For the month of June 2022 the second top search is: "ccid driver windows 10".

Google "think" someone will find a Windows 10 driver on https://ccid.apdu.fr/?

The Windows CCID is so hard to find or to install or is so buggy that users are searching for an alternative?

On the last 12 months the result is less worrisome. The "ccid driver windows 10" search is only 7th in the list:

Conclusion

Please propose your own explanation of this result.

Maybe I should add "Windows 10" in the web site description metadata to confuse Google search a bit more? :-)

Sunday, June 12, 2022

PCSC API spy, update

10 years ago I documented in "PCSC API spy, third try" a way to generate PC/SC API traces when using pcsc-lite.

Since then the ecosystem has changed. This article is an update of the previous blog article with more up-to-date information.

Changes

The pcsc-spy.py command has been renamed pcsc-spy (in 2012)
The libpcscspy.so library has been moved from /usr/lib/ to /usr/lib/x86_64-linux-gnu/ (for Intel 64-bits CPU systems)
opensc-tool can't be used with LD_PRELOAD= any more

Demo

As before we have two cases for the use of libpcsclite.so.1.

Applications linked with libpcsclite.so.1

This is the case of the pcsc_scan command for example.

You can use the ldd command to know what library has been dynamically linked at build time:

$ ldd /usr/bin/pcsc_scan 
	linux-vdso.so.1 (0x00007fffac11b000)
	libpcsclite.so.1 => /lib/x86_64-linux-gnu/libpcsclite.so.1 (0x00007f85a5f24000)
	libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f85a5f03000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f85a5d2a000)
	/lib64/ld-linux-x86-64.so.2 (0x00007f85a5f4f000)

You can use the LD_PRELOAD solution by doing:

In on terminal you run the pcsc-spy program. In another terminal you run:

$ LD_PRELOAD=/usr/lib/x86_64-linux-gnu/libpcscspy.so.0 pcsc_scan -r
libpcsclite_nospy.so.1: cannot open shared object file: No such file or directory
No reader found.

In the first terminal you get the trace:

$ pcsc-spy 
SCardEstablishContext
 i dwScope: SCARD_SCOPE_SYSTEM (0x00000002)
 o hContext: 0x0E4C693B
 => Command successful. (SCARD_S_SUCCESS [0x00000000])  [0.002456]
SCardGetStatusChange
 i hContext: 0x0E4C693B
 i dwTimeout: 0x00000000 (0)
 i cReaders: 1
 i szReader: \\?PnP?\Notification
 i  dwCurrentState:  (0x00000000)
 i  dwEventState: SCARD_STATE_IGNORE, SCARD_STATE_UNKNOWN, SCARD_STATE_UNAVAILABLE, SCARD_STATE_EMPTY, SCARD_STATE_INUSE, SCARD_STATE_MUTE (0x55EDE352031D)
 i  Atr length: 0x55EDE352032C (94480209412908)
 i  Atr: NULL
 o szReader: \\?PnP?\Notification
 o  dwCurrentState:  (0x00000000)
 o  dwEventState:  (0x00000000)
 o  Atr length: 0x55EDE352032C (94480209412908)
 o  Atr: NULL
 => Command timeout. (SCARD_E_TIMEOUT [0x8010000A])  [0.007774]
SCardListReaders
 i hContext: 0x0E4C693B
 i mszGroups: (null)
 o pcchReaders: 0x00000001
 o mszReaders: NULL
 => Cannot find a smart card reader. (SCARD_E_NO_READERS_AVAILABLE [0x8010002E])  [0.000908]
SCardListReaders
 i hContext: 0x0E4C693B
 i mszGroups: (null)
 o pcchReaders: 0x00000001
 o mszReaders: NULL
 => Cannot find a smart card reader. (SCARD_E_NO_READERS_AVAILABLE [0x8010002E])  [0.000531]

Thread 1/1
Results sorted by total execution time
total time: 0.011769 sec
0.007774 sec (  1 calls) 66.06% SCardGetStatusChange
0.002456 sec (  1 calls) 20.87% SCardEstablishContext
0.001439 sec (  2 calls) 12.23% SCardListReaders

Application loading libpcsclite.so.1

In this case you need to modify the system configuration to replace the libpcsclite.so.1 library. This is done by the install_spy.sh script. You only need to run the script once.

$ sudo bash /usr/share/doc/libpcsclite-dev/install_spy.sh
Using directory: /lib/x86_64-linux-gnu
Spying library is: /lib/x86_64-linux-gnu/libpcscspy.so.0

On Debian (and derivatives like Ubuntu) and with pcsc-lite version 1.9.8 and more the script is provided by the libpcsclite-dev package.

In on terminal you run the pcsc-spy program. In another terminal you run the program you want to spy. For example:

$ opensc-tool -a
No smart card readers found.
Failed to connect to reader: No readers found

In the first terminal you get the trace:

SCardEstablishContext
 i dwScope: SCARD_SCOPE_USER (0x00000000)
 o hContext: 0x2667F6DA
 => Command successful. (SCARD_S_SUCCESS [0x00000000])  [0.005316]
SCardListReaders
 i hContext: 0x2667F6DA
 i mszGroups: (null)
 o pcchReaders: 0x00000001
 o mszReaders: NULL
 => Cannot find a smart card reader. (SCARD_E_NO_READERS_AVAILABLE [0x8010002E])  [0.000079]
SCardReleaseContext
 i hContext: 0x2667F6DA
 => Command successful. (SCARD_S_SUCCESS [0x00000000])  [0.000074]

Thread 1/1
Results sorted by total execution time
total time: 0.007195 sec
0.005316 sec (  1 calls) 73.88% SCardEstablishContext
0.000079 sec (  1 calls)  1.10% SCardListReaders
0.000074 sec (  1 calls)  1.03% SCardReleaseContext

Do not forget to restore the system configuration using the uninstall_spy.sh script.

$ sudo bash /usr/share/doc/libpcsclite-dev/uninstall_spy.sh
Using directory: /lib/x86_64-linux-gnu

Redirection in a file

It is still possible to redirect the traces in a file. Instead of running pcsc-spy you do:

$ mkfifo ~/pcsc-spy
$ cat ~/pcsc-spy > logfile

And in another terminal you start the application as indicated above (i.e. using LD_PRELOAD= or after running install_spy.sh)

You can then analyse the logs later using:

$ pcsc-spy logfile

Remarks

Bugs found

I note that SCardReleaseContext() is not always called by pcsc_scan before exit. I just fixed this problem in pcsc-tools.

Install/uninstall

It is important to run the uninstall_spy.sh script to undo the changes made by the install_spy.sh script.

It is important you undo the changes before any execution of the ldconfig (configure dynamic linker run-time bindings) administration command. ldconfig is used, for example, during the installation of a package.

If you run uninstall_spy.sh after an execution of ldconfig you may get a broken libpcsclite installation with an error like:

$ pcsc_scan 
pcsc_scan: error while loading shared libraries:
libpcsclite.so.1: cannot open shared object file: No such file or directory

To fix te problem you can force reinstall the libpcsclite1 (or equivalent) package.

Order of execution

It is important to start pcsc-spy before the application you want to spy. If you start pcsc-spy after the application you have 2 cases:

if the fifo file ~/pcsc-spy does not yet exist then pcsc-spy will display nothing
if the fifo file ~/pcsc-spy already exists then libpcscspy.so will use it to send logs and will be blocked until something reads the file (pcsc-spy or the cat command to redirect the content)

Conclusion

I hope this update is useful.

if you have ideas to improve the logs please contact me.

Saturday, June 11, 2022

New version of pcsc-lite: 1.9.8

I just released a new version of pcsc-lite 1.9.8.
pcsc-lite is a Free Software implementation of the PC/SC (or WinSCard) API for Unix systems.

Changes:

1.9.8: Ludovic Rousseau
11 June 2022

Install install_spy.sh & uninstall_spy.sh scripts in docdir
SCardTransmit(): do not fail if receive buffer is "too large"
SCardControl(): do not fail if receive buffer is "too large"
fix some memory leaks on shutdown
use a better random number generator
Some other minor improvements

Friday, May 20, 2022

ATR parsing in JSON

I updated my ATR parsing web site at https://smartcard-atr.apdu.fr/ to add the option to get the result not in an HTML page but as a JSON document.

If you do not yet know this service you can read "Parsing an ATR: new web site URL" to get some history.

JSON

The URL for the JSON service is https://smartcard-atr.apdu.fr/V1/parse2json?ATR=3BFF9700008131FE4380318065B0846160FB120FFD8290000D.

It is the same as for the HTML output https://smartcard-atr.apdu.fr/parse?ATR=3BFF9700008131FE4380318065B0846160FB120FFD8290000D except you replace "parse" by "V1/parse2json" in the URL.

Of course you update the ATR parameter value to use the ATR value you want to parse.

Examples

You can use it to get the description of a particular field by piping the result in the jq command.

jq is a lightweight and flexible command-line JSON processor.

For example to get the meaning of the TA₁ byte you can use:

$ curl -s https://smartcard-atr.apdu.fr/V1/parse2json?ATR=3B12953606 | jq '.TA."1"'
{
  "description": "Fi=512, Di=16, 32 cycles/ETU (125000 bits/s at 4.00 MHz, 156250 bits/s for fMax=5 MHz)",
  "value": 149
}

Or for just the TA₁ description:

$ curl -s https://smartcard-atr.apdu.fr/V1/parse2json?ATR=3B90160187 | jq '.TA."1".description'
"Fi=372, Di=32, 11.625 cycles/ETU (344086 bits/s at 4.00 MHz, 430107 bits/s for fMax=5 MHz)"

Card matching

I also added the possibility to get the cards matching a given ATR.

For example: https://smartcard-atr.apdu.fr/V1/match?ATR=3B8F800180318065B0850300EF120FFE82900072

$ curl -s https://smartcard-atr.apdu.fr/V1/match?ATR=3B8F800180318065B0850300EF120FFE82900072 | jq
{
  "3B 8F 80 01 80 31 80 65 B0 .. .. .. .. 12 0F FE 82 90 00 ..": [
    "IDPrime MD 3810 T=Contactless (Prox DU)"
  ],
  "3B 8F 80 01 80 31 80 65 B0 85 03 00 EF 12 0F FE 82 90 00 72": [
    "Gemalto IDPrime MD 3840",
    "http://www.gemalto.com/dwnld/6891_IDPrimeMD3840_Product_Datasheet_May14.pdf"
  ]
}

Because of the joker mechanism using the character '.' it is possible that one ATR matches more than one "card".

Conclusion

If you have ideas how to improve the service just tell me.

Saturday, May 14, 2022

SCardListReaders() and non-initialized pcchReaders parameter

Sample C sample

We start with the following sample.c program:

01: #include <stdio.h>
02: #include <stdlib.h>
03: #include <winscard.h>
04: 
05: #define CHECK(f, rv) \
06:  if (SCARD_S_SUCCESS != rv) \
07:  { \
08:   printf(f ": %s\n", pcsc_stringify_error(rv)); \
09:   return -1; \
10:  }
11: 
12: int main(void)
13: {
14:     LONG rv;
15: 
16:     SCARDCONTEXT hContext;
17:     LPTSTR mszReaders;
18:     DWORD dwReaders;
19: 
20:     rv = SCardEstablishContext(SCARD_SCOPE_SYSTEM, NULL, NULL, &hContext);
21:     CHECK("SCardEstablishContext", rv)
22: 
23:     rv = SCardListReaders(hContext, NULL, NULL, &dwReaders);
24:     CHECK("SCardListReaders", rv)
25: 
26:     mszReaders = calloc(dwReaders, sizeof(char));
27:     rv = SCardListReaders(hContext, NULL, mszReaders, &dwReaders);
28:     CHECK("SCardListReaders", rv)
29:     printf("1st reader name: %s\n", mszReaders);
30: 
31:     free(mszReaders);
32: 
33:     rv = SCardReleaseContext(hContext);
34:     CHECK("SCardReleaseContext", rv)
35: 
36:     return 0;
37: }
38:

Makefile

I used this Makefile

# Linux
PCSC_CFLAGS := $(shell pkg-config --cflags libpcsclite)
LDLIBS := $(shell pkg-config --libs libpcsclite)

CFLAGS = $(PCSC_CFLAGS) -g

sample: sample.c

clean:
	rm -f sample

Output

Compilation and execution give:

$ make
cc -pthread -I/usr/include/PCSC -g   sample.c  -lpcsclite -o sample
$ ./sample 
1st reader name: Alcor Micro AU9540 00 00

It looks like everything is correct.

But can you find the problem?

Valgrind is your friend

valgrind is a very nice tool to debug issues.

Valgrind is an instrumentation framework for building dynamic analysis tools. There are Valgrind tools that can automatically detect many memory management and threading bugs, and profile your programs in detail. You can also use Valgrind to build new tools.

valgrind sees a problem:

$ valgrind --track-origins=yes ./sample

==99878== Memcheck, a memory error detector
==99878== Copyright (C) 2002-2017, and GNU GPL'd, by Julian Seward et al.
==99878== Using Valgrind-3.18.1 and LibVEX; rerun with -h for copyright info
==99878== Command: ./sample
==99878== 
==99878== Conditional jump or move depends on uninitialised value(s)
==99878==    at 0x486E02E: SCardListReaders (in /usr/lib/x86_64-linux-gnu/libpcsclite.so.1.0.0)
==99878==    by 0x109210: main (sample.c:23)
==99878==  Uninitialised value was created by a stack allocation
==99878==    at 0x109199: main (sample.c:13)
==99878== 
1st reader name: Alcor Micro AU9540 00 00
==99878== 
==99878== HEAP SUMMARY:
==99878==     in use at exit: 112 bytes in 4 blocks
==99878==   total heap usage: 10 allocs, 6 frees, 1,434 bytes allocated
==99878== 
==99878== LEAK SUMMARY:
==99878==    definitely lost: 0 bytes in 0 blocks
==99878==    indirectly lost: 0 bytes in 0 blocks
==99878==      possibly lost: 0 bytes in 0 blocks
==99878==    still reachable: 112 bytes in 4 blocks
==99878==         suppressed: 0 bytes in 0 blocks
==99878== Rerun with --leak-check=full to see details of leaked memory
==99878== 
==99878== For lists of detected and suppressed errors, rerun with: -s
==99878== ERROR SUMMARY: 1 errors from 1 contexts (suppressed: 0 from 0)

Using uninitialized variables is bad. It creates bugs that declare at random time and are not easy to find.

SCardListReaders

The problem is at line 23 which is:

23:     rv = SCardListReaders(hContext, NULL, NULL, &dwReaders);

If we read the documentation for SCardListReaders() we have:

Returns a list of currently available readers on the system.

mszReaders is a pointer to a character string that is allocated by the application. If the application sends mszGroups and mszReaders as NULL then this function will return the size of the buffer needed to allocate in pcchReaders.

If *pcchReaders is equal to SCARD_AUTOALLOCATE then the function will allocate itself the needed memory. Use SCardFreeMemory() to release it.

Parameters

[in] hContext Connection context to the PC/SC Resource Manager.

[in] mszGroups List of groups to list readers (not used).

[out] mszReaders Multi-string with list of readers.

[in,out] pcchReaders Size of multi-string buffer including NULL's.

You can note that the parameter pcchReaders is in and out. This is because the value is compared to SCARD_AUTOALLOCATE. So the value of pcchReaders shall be set before calling SCardListReaders().

The fix is simple. Just update line 18 like this:

18: DWORD dwReaders = 0;

Looking for problems

What happens if we initialize dwReaders to the special value SCARD_AUTOALLOCATE instead of 0?

18: DWORD dwReaders = SCARD_AUTOALLOCATE;

We build and run the sample:

$ ./sample 
SCardListReaders: Invalid parameter given.

The execution of SCardListReaders() fails.

23:     rv = SCardListReaders(hContext, NULL, NULL, &dwReaders);

This is because we are asking SCardListReaders() to allocate and store the result in the parameter mszReaders. But in our sample this parameter is NULL. pcsc-lite has a check for that and returns an error code.

Instead of using NULL we could use something else like 0x1234.

23:     rv = SCardListReaders(hContext, NULL, 0x1234, &dwReaders);

This time we have a nice crash:

$ ./sample 
Segmentation fault

This is because we asked SCardListReaders() to write at the address 0x1234. This address is, in general, not valid.

This problem could happen if you use something like:

23:     rv = SCardListReaders(hContext, NULL, mszReaders, &dwReaders);

And both variables mszReaders and dwReaders are uninitialized, and by a total lack of luck dwReaders has the value SCARD_AUTOALLOCATE (i.e. -1).

GnuPG

The problem was found in by oddlama and reported to PCSC at "pcscd fails to read future yubikeys after removing a yubikey, until restarted #125".

The problem is not in pcsc-lite but in GnugPG. So I reported the problem at "SCardListReaders: Conditional jump or move depends on uninitialised value(s)".

WinSCard API

macOS does not support (yet) the value SCARD_AUTOALLOCATE. But pcsc-lite and Windows WinSCard do.

I agree the API to use SCARD_AUTOALLOCATE is ugly. We pass the address of a buffer pointer in a parameter that is a buffer pointer. We have to cast the variable like I did in the C sample like:

  dwReaders = SCARD_AUTOALLOCATE;
  rv = SCardListReaders(hContext, NULL, (LPTSTR)&mszReaders, &dwReaders);

This feature is a Microsoft extension that is not present in the PCSC workgroup specification.

Conclusion

I think the problem is not very known and should be better documented. That was my motivation for writting this blog article.

Morality: in C language, always initialize your variables to a known/safe value (like 0).

Friday, May 13, 2022

New version of pcsc-lite: 1.9.7

I just released a new version of pcsc-lite 1.9.7.
pcsc-lite is a Free Software implementation of the PC/SC (or WinSCard) API for Unix systems.

Changes

1.9.7: Ludovic Rousseau
13 May 2022

disable strict compilation by default
fix 3 warnings

Issues with strict compilation

In the previous version (1.9.6) I enabled a strict compilation option by default. The compiler was configured to use the -Werror argument. The effect is to transform warnings into errors. The compilation would fail instead of just displaying a warning.

On my side I had no warnings so I was confident.

But some build systems use a different set of default compilation arguments. For example Arch Linux or openSUSE Tumbleweed use the argument -flto=auto to perform some link-time optimizer.

This option generated 3 warnings. The warnings are minor but since the compilation used -Werror they were treated as errors and the compilation failed.

Sorry for generating the problem. Thanks to Axel Braun and frederik for reporting the failure.

Wednesday, May 11, 2022

New version of pcsc-lite: 1.9.6

I just released a new version of pcsc-lite 1.9.6.
pcsc-lite is a Free Software implementation of the PC/SC (or WinSCard) API for Unix systems.

Changes:

1.9.6: Ludovic Rousseau
11 May 2022

do not fail reader removal in some specific cases (USB/Thunderbolt port)
improve documentation regarding /etc/reader.conf.d/
SCardGetStatusChange: speedup the case DISABLE_AUTO_POWER_ON
configure:

add --disable-strict option
By default the compiler arguments are now:
-Wall -Wextra -Wno-unused-parameter -Werror ${CFLAGS}
fail if flex is not found

fix different data races
pcscdaemon: -v displays internal constants values: MAX_READERNAME & PCSCLITE_MAX_READERS_CONTEXTS
Some other minor improvements

Reader removal issue

The problem with reader removal is that, with some hardware configurations, the USB bus is removed when the USB device is disconnected. You can see that using the libusb command.

Before connecting the device:

Bus 002 Device 002: ID 0bda:0328 Realtek Semiconductor Corp. USB3.0-CRW
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 003: ID 1bcf:2c6b Sunplus Innovation Technology Inc. HD WebCam
Bus 001 Device 005: ID 1044:7a39 Chu Yuen Enterprise Co., Ltd USB-HID Keyboard
Bus 001 Device 004: ID 8087:0a2b Intel Corp. Bluetooth wireless interface
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

After connecting the Yubikey token

Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 003 Device 002: ID 1050:0407 Yubico.com Yubikey 4/5 OTP+U2F+CCID
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
Bus 002 Device 002: ID 0bda:0328 Realtek Semiconductor Corp. USB3.0-CRW
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 003: ID 1bcf:2c6b Sunplus Innovation Technology Inc. HD WebCam
Bus 001 Device 005: ID 1044:7a39 Chu Yuen Enterprise Co., Ltd USB-HID Keyboard
Bus 001 Device 004: ID 8087:0a2b Intel Corp. Bluetooth wireless interface
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

The new lines are:

Bus 004 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 003 Device 002: ID 1050:0407 Yubico.com Yubikey 4/5 OTP+U2F+CCID
Bus 003 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

2 new USB buses are created 003 & 004. The Yubikey token is connected on the bus 003.

After removing the Yubikey

Bus 002 Device 002: ID 0bda:0328 Realtek Semiconductor Corp. USB3.0-CRW
Bus 002 Device 001: ID 1d6b:0003 Linux Foundation 3.0 root hub
Bus 001 Device 003: ID 1bcf:2c6b Sunplus Innovation Technology Inc. HD WebCam
Bus 001 Device 005: ID 1044:7a39 Chu Yuen Enterprise Co., Ltd USB-HID Keyboard
Bus 001 Device 004: ID 8087:0a2b Intel Corp. Bluetooth wireless interface
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub

The USB buses 3 & 4 have disappeared. This confused pcsc-lite. The token was not removed for pcsc-lite and caused the generation of log errors. This is now fixed.

It looks like the problem occurred with a USB-C + Thunderbolt port and was reported in "pcscd fails to read future yubikeys after removing a yubikey, until restarted #125" and "Yubikey 5c Reader "stuck" #57".

Wednesday, April 27, 2022

30 shades of PC/SC

In 2010 I started to write about the different programming languages you can use to talk to a smart card and a smart card reader.

12 years later I have 30 languages in the list.

They are, in alphabetical order:

Top Computer Languages

Different lists exist for the top computer languages popularity. I will use the TIOBE Index for April 2022.

The "support" column indicates if the language has a smart card API available.

#	Language	support
1	Python	Yes
2	C	Yes
3	Java	Yes
4	C++	Yes
5	C#	Yes
6	Visual Basic	?
7	JavaScript	Yes
8	Assembly	?
9	SQL	?
10	PHP	Yes
11	R	?
12	Pascal	Yes
13	Go	Yes
14	Swift	Yes
15	Ruby	Yes
16	Classic Visual Basic	?
17	Objective-C	Yes
18	Perl	Yes
19	Lua	Yes
20	Mathlab	?
21	SAS	?
22	Prolog	Yes
23	(Visual) FoxPro	?
24	Scratch	?
25	COBOL	?
26	Julia	?
27	Kotlin	Yes
28	Rust	Yes
29	Ada	Yes
30	Lisp	Yes
31	Fortran	?
32	Groovy	?
33	VBScript	?
34	PL/SQL	?
35	D	?
36	Scala	Yes
37	Haskell	?
38	Dart	Yes

Comments

It should be easy to call the C WinSCard API from assembly language. So the answer could be "yes". I learned assembly language with a Motorola 6809 and then a Motorola 68000. I am not a fan of the Intel x86 assembly language with its very strange register names (to be backward compatible). Yes, I could also use ARM or RISC-V CPUs. I let the exercise to write an example in assembly language to the readers.

Not listed

Some programming languages are not listed in the TIOBE index but do provide a PC/SC wrapper:

Elixir
Erlang
OCaml

Windows technologies

Some languages are specific to the Microsoft Windows system. I can't use them since I do not use Windows. They are:

Visual Basic
Classic Visual Basic
VBScript
Visual FoxPro.

It looks like Visual Basic is also available for GNU/Linux. But I don't want to invest time in these technologies.

Also note that all the computer languages with PC/SC support exist in a Free Software implementation. I am not sure for the Microsoft languages.

Language I would be surprised to get PC/SC support

R and SAS are languages used for statistics.

Scratch is used for education.

Mathlab is for numeric computing.

COBOL is very old. I do not expect any COBOL program needing access to smart cards. Same remark for Fortran. COBOL and Fortran are still used and it looks like it is possible to call C functions from both of them. So there is hope ☺.

Languages that could have a PC/SC wrapper

Julia
it should be possible to use PyCall and call the Python PC/SC wrapper.
D
Haskell

What I had not yet found

SQL: (SQL Server) PC/SC List Smart Card Readers (and USB Tokens)

This is for SQL Server, so may be limited to the Microsoft product and not the SQL language.

Conclusion

It was a lot of fun to discover and write a sample codes for all the 30 languages.

For some (4) of the PC/SC wrappers I am the maintainer so writing a sample code was easy.

The list is not finished and may never be as new programming languages are created from time to time. If you want me to add a new language please contact me.

[in]	hContext	Connection context to the PC/SC Resource Manager.
[in]	mszGroups	List of groups to list readers (not used).
[out]	mszReaders	Multi-string with list of readers.
[in,out]	pcchReaders	Size of multi-string buffer including NULL's.

Important!

Friday, December 30, 2022

Changes:

Tuesday, December 27, 2022

epydoc → pydoctor

Previous documentation

New documentation

Conclusion

Saturday, December 17, 2022

The reported problem

Example of cards in reported problems

The problem

Where are the bogus readers?

Patch

Problem with the patch

Your help is welcome

Conclusion

[Update: Jan, 20th 2023]

[Update: Jan, 31th 2023]

Tuesday, November 29, 2022

pcsc-deno

Deno

Installation

Source code

Output

Conclusion

Wednesday, November 16, 2022

Problem

Solution

Setup

Host

Guest

Limitations

pcsc-lite to pcsc-lite

macOS or Windows host

Conclusion

Tuesday, November 15, 2022

CCID

Apple Open Source

Crypto Token Kit

Security message on first connection

Conclusion

Monday, November 14, 2022

Changes:

Monday, November 7, 2022

New version

Conclusion

Friday, October 28, 2022

Problem

Fix

No problem on Debian

Conclusion

Wednesday, September 28, 2022

Changes:

Sunday, September 11, 2022

Changes:

Sunday, September 4, 2022

Changes:

Sunday, August 28, 2022

Changes:

Tuesday, July 5, 2022

Conclusion

Sunday, June 12, 2022

Changes

Demo

Applications linked with libpcsclite.so.1

Application loading libpcsclite.so.1

Redirection in a file

Remarks

Bugs found

Install/uninstall

Order of execution

Conclusion

Saturday, June 11, 2022

Changes:

Friday, May 20, 2022

JSON

Examples

Card matching

Conclusion