Getting slot ID from slotEventHook

[arrowd]

Reading the __pkcs11h_slotevent_provider function's code, it seems that the slot value returned from C_WaitForSlotEvent is discarded. This makes it impossible to figure what slot exactly has generated an event.

[alonbl]

Hi, In pkcs11-helper the slots are abstracted, as one of the most difficult flows is managing the slots. The notification allows you to refresh your certificate store. Alon

[arrowd]

Do you mean calling pkcs11h_plugAndPlay?

[alonbl]

No, pkcs11h_plugAndPlay() should be called only has hack for providers that do not support plug and play, it destroys the context and created it. Plug and play is defined better these days (20 years later), most providers will allocate a new slot id per each reader unplug/plug.

Probably the pkcs11h_certificate_enumCertificateIds() is to be called when slot event is asserted.

[arrowd]

Sorry, I still don't quite get it.

I want to track which tokens are inserted at the moment. How do I achieve this without getting a slot ID in the hook?

[alonbl]

you can use pkcs11h_token_enumTokenIds() if you wish you can use PKCS#11 directly, this library is fairly old, it designed to abstract the complexity of PKCS#11 while since then the popularity of PKCS#11 (or any standard) became lower.

[arrowd]

Yes, calling pkcs11h_token_enumTokenIds() to figure what changed since last event works. But I still don't see how would it hurt to passthrough the slot value from C_WaitForSlotEvent to the hook.

[alonbl]

slot is meaningless as it is opaque, for plug&play it changes each unplug/plug for usb token it is totally meaningless. it also means nothing to the user. in what case do you need it?

[arrowd]

You said

you can use pkcs11h_token_enumTokenIds() if you wish you can use PKCS#11 directly

but using them requires specifying slot. How do I go from helper token IDs to slots?

[alonbl]

oh, I probably wrote it badly... if you want control over slots, keys, and any other low level feature you can use PKCS#11 directly, avoid using this library.

[arrowd]

Ok, I see. Thank you for your explanations.

[arrowd]

Is is bugfix important enough to make a new release? Otherwise I'll just patch the pkcs11-helper downstream (FreeBSD).

[alonbl]

if you are the maintainer then it will be nice if you apply it, if not I will release.

[arrowd]

I'm not, but I do have authorization to do that. I'll patch it locally then, thanks.