[PATCH gnupg] scd: Add Nitrokey 3 to pcsc-shared interference detection

[Artemii Bigdan]

Hello,

I gave your suggestion a try. I applied your little patch to the current 
master branch.

Unfortunately, it did not achieve the desired bugfix. The `gpg: OpenPGP 
card not available: Card error` is back after KeePassXC interference, as 
before.

I tried several different combinations of fixes based around yours; none 
yielded satisfactory results.

It seems that, as Robin Krahl <robin at nitrokey.com> noted, there are 
other CARDTYPE_YUBIKEY workarounds scattered around the codebase. 
Unfortunately, my free time is rather limited, so the specific root 
cause is yet to be determined.

Still, it seems plausible to me that some Yubikey-specific workaround 
interferes, given that the original patch with a separate CARDTYPE_NK3 
enum works flawlessly.

While I understand you want a minimal change to reduce the maintenance 
burden, I agree with Robin that it would be wise to create a new enum 
value for Nitrokey 3. Nitrokey is not developed or marketed as a Yubikey 
clone. And while I understand that Nitrokey is a rather niche platform, 
it is one of the more popular fully free-software alternatives to 
Yubikey currently on the market.

So, as of my testing, I unfortunately have to insist on the 
`0025-scdaemon-nitrokey3-pcsc-shared.patch` as it was submitted earlier. 
I am open to perform more tests if you can come up with a more concise 
way to achieve the same result.

I would also like to ask Robin and NIIBE Yutaka: maybe Nitrokey and 
GnuPG could cooperate and deliver an R&D Nitrokey 3 device to one of 
GnuPG's core maintainers for better first-class support?

Thank you for your time and attention to this issue. It means a lot, 
this bug has bothered me constantly in my daily workflow, and I really 
do not want to manually re-patch each update. I hope we can upstream this.

- Artemii


On 6/30/26 6:34 AM, NIIBE Yutaka wrote:
> Hello,
>
> I register your DCO.  Then, I found a mistake in the code and pushed:
>
> 	ca25a7a61bebbe4e27dd5568c5b049675b7aa4ae
>
> Artemii Bigdan<artem at bigdan.in> wrote:
>> I do not have documentation from Nitrokey on the CCID behavior of the
>> device. The observations below are from testing. More information
>> about the hardware can be found athttps://www.nitrokey.com
> Ummm...  So, all might be guess work (unfortunately, I don't have the
> device).
>
> In this situation, simpler changes are better.  If it's mostly the same
> (feature-wise), how about something like the following patch?  If it
> works somehow, I think that it is good for our maintenance reason.
>
> ==========================
> diff --git a/scd/app.c b/scd/app.c
> index 63eb84190..bbd10d6e7 100644
> --- a/scd/app.c
> +++ b/scd/app.c
> @@ -667,6 +667,12 @@ atr_to_cardtype (int slot, const unsigned char *atr, size_t atrlen)
>         if (!atrbuf)
>           return 0;
>         atr = atrbuf;
> +      if (atrlen >= 16 && !memcmp (&atr[5], "Nitrokey", 8))
> +        {
> +          /* It looks like Yubikey variant/clone.  */
> +          xfree (atrbuf);
> +          return CARDTYPE_YUBIKEY;
> +        }
>       }
>   
>     for (i=0; atrlist[i].atrlen; i++)
> ==========================
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 228 bytes
Desc: OpenPGP digital signature
URL: <https://lists.gnupg.org/pipermail/gnupg-devel/attachments/20260630/f1d0c507/attachment.sig>