AnsweredAssumed Answered

ANOMALY_05000263 unexpected behavior.

Question asked by josemanuel on Oct 7, 2016
Latest reply on Jan 2, 2018 by Aaronwu

PROBLEM DESCRIPTION

 

Ad1939 does not output any audio neither on cold reset nor power up.

 

ENVIRONMENT.

 

Blackfin BF537, rev 0.3
Ad1939

 

Ad1939 is connected to sport 0. Sport 0 uses dma to transmit and receive data.

 

uClinux-dist.R1.1-RC3. The linux kernel version is 2.6.19.3-ADI-2007R1.1. The output of uname is:

 

Linux version 2.6.19.3-ADI-2007R1.1[KFSBFD-2A]-svn (user@Precision-3510) (gcc version 4.1.1 (ADI 07R1)) #1 Fri Jul 22 15:36:33 CEST 2016

 

Kernel compiled for revisión 2.

 

MODIFICATIONS

 

Reported in http://www.analog.com/media/en/dsp-documentation/integrated-circuit-anomalies/icanom_ADSP_BF534.BF536.BF537_RevC_0208.pdf,
the anomaly 05000263 affects only to cpu's revision 0.2. Currently we are using cpu's revision 0.3, free of that anomaly.

 

We have created a patch to avoid this errata in cpu's revision 0.3. This way we made the kernel compatible with cpu's  revision 0.2. That patch is attached to this mail.

 

TESTS

 

Test1: OK

 

1- Compile kernel without patch applied.
2- Power off machine
3- Wait for 5 seconds.
4- Power on machine.
5- Run the compiled kernel.
6- Writing to SPORT a signal with a frequency of 1khz and a voltage of 2Vrms, we get the same signal at output.

 

Test1:  FAIL

 

1- Compile kernel with patch applied.
2- Power off machine
3- Wait for 5 seconds.
4- Power on machine.
5- Run the compiled kernel.
6- Writing to SPORT a signal with a frequency of 1khz and a voltage of 2Vrms, we don't get any signal at output.

 

At this point,

 

7- Execute reboot.
8- Run compiled kernel,
9- Writing to SPORT a signal with a frequency of 1khz and a voltage of 2Vrms, we get the same signal at output.

 

CHECKS

 

1- We've checked every pin on ad1939. All seems be the same value on both tests
2- We've checked ad1939 spi configuration. It's equal. We checked the spi bus too, it seems to be ok.
3- We've enlarged time before configuration of the ad1939 .
4- We've enlarged time between configuration operations of the ad1939 .
5- We've checked the memory using u-boot extended memory test. Test passed ok.

 

ASSUMPTIONS

 

A) Ad1939 is correctly configured: it works correctly either in test1 or in test2.
B) Memory is coherent and consistent.

 

ASKS

 

1- Does the anomaly 05000263 affect to dma engine?
2- Does we need to wait for an extra time if we do not apply the errata?
3- Could the TDM of sport be corrupt if we do not apply the errata?

 

Thanks in advance.

Outcomes