Authorized write request with VLOC_USER

RE: Authorized write request with VLOC_USER

Petter Myhre — Mon, 20 Jun 2016 09:46:34 GMT

Anyways, both auth read and auth write should be safe to use with VLOC_USER. If you experience something else please tell me how to reproduce it with the latest SoftDevice, or supply me with a project that exhibits the behavior.

RE: Authorized write request with VLOC_USER

Petter Myhre — Fri, 17 Jun 2016 11:26:56 GMT

I did some more testing, and actually auth read with update flag set seems to be working fine. Could you please elaborate on how you get this hard fault?

RE: Authorized write request with VLOC_USER

Petter Myhre — Wed, 15 Jun 2016 13:57:45 GMT

I haven't been able to reproduce the hard fault with auth read, but there is something strange with VLOC_USER and the update flag set. I will investigate this, and the auth write and get back to you soon.

RE: Authorized write request with VLOC_USER

blungreen — Tue, 14 Jun 2016 15:30:48 GMT

I did open a ticket while i was working with alpha 1.0.3. When i moved to 2.0 i also included the fix suggested by the support team and never got the fault again. I remember i tried 2.0 once without fix and still got hardfault. I instantly reverted to working version and started to be happy again

RE: Authorized write request with VLOC_USER

Petter Myhre — Tue, 14 Jun 2016 08:43:40 GMT

I see. I need to look into this. What SoftDevice version did you experience this with?

RE: Authorized write request with VLOC_USER

blungreen — Mon, 13 Jun 2016 14:46:31 GMT

In the past i was getting Hardfault because i was setting update flag with attribute kept in VLOC_USER when doing a reply to auth_read.

In a ticket reply i've been told that you can't set update flag unless attribute is in VLOCK_STACK, so this is why i did ask. So, to recap and if understood right, if attribute is in VLOC_USER you MUST NOT set update flag if replying to authorized read, and you MUST set it if replying to authorized write request regardless its location in memory.

RE: Authorized write request with VLOC_USER

Petter Myhre — Mon, 13 Jun 2016 14:12:32 GMT

Yes, it is safe. I am not sure what made you ask this question. It could be that the documentation is misinterpreted. When you receive a BLE_GATTS_OP_WRITE_REQ , then the update flag SHALL be set to '1'. Else you will get an error as you did. So please go ahead and set the update flag to '1'.

RE: Authorized write request with VLOC_USER

blungreen — Mon, 13 Jun 2016 13:03:03 GMT

I mean correct, no stack corruption, no umpredictable behaviour

RE: Authorized write request with VLOC_USER

Petter Myhre — Mon, 13 Jun 2016 12:20:45 GMT

What do you mean by is it safe?