[alsa-devel] Still have problems/issues using SGTL5000 with i.MX28 CPU

[gianluca]

On 05/29/2015 03:50 PM, Fabio Estevam wrote:
>> Now we need to record some sounds as soon as the SGTL5000 has the SAIF1 line
>> used for caputure (line in / mic in).
>
> Does recording work on mx28evk with 4.0.4?
>

Well, is that linux kernel version so important? I can surely say it 
works with 3.12.1.

Porting to 4.0.4 means a (partial) rewrite of some device-tree files. I 
can do in a couple of days if you want. But I think I can check out what 
is really happened in our board. Something related to dma / irq??

In IMX28EVK it works out-of-the-box, in EK340 it does not.

Regarding at the last mail I can clearly see:

> When recording from IMX28EVK boards here is the SAIF1_STAT register (0x80046010):
> 0x80000001 (BIT31 PRESENT and BIT0 BUSY)
> ...
> 0x80010001 (BIT31 PRESENT, BIT16 DMA_PREQ and BIT0 BUSY)
> ...
> 0x80000001 (BIT31 PRESENT and BIT0 BUSY)
> ...
> 0x80010001 (BIT31 PRESENT, BIT16 DMA_PREQ and BIT0 BUSY)
>
> and when finishing recording:
>
> 0x80010010 (BIT31 PRESENT, BIT16 DMA_PREQ and BIT4 FIFO_SERVICE_IRQ)
>
> in the EK340 while recording (0x80046010):
> 0x80000000 (BIT31 PRESENT)
> ...
> ...
> ...
> 0x80000000 (BIT31 PRESENT)
>
> and finally:
>> # arecord: pcm_read:1801: read error: Input/output error
>
>> # devmem2 0x80046010
>> /dev/mem opened.
>> Memory mapped at address 0xb6f07000.
>> Value at address 0x80046010 (0xb6f07010): 0x80000000
>
> It seems no IRQ or DMA actions are involved in EK340 for recording.
> May be I have some DMA Clash or IRQ bad handling? In the device tree I cannot see anything wrong... Some other driver is using the same DMA Channel for SAIF1 ?????
>
> Anyway here is the cat /proc/interrupts for IMX28EVK:
>>
>> # cat /proc/interrupts
>>            CPU0
>>  16:      32790         -  48  MXS Timer Tick
>>  17:       5327         -  82  mxs-dma
>>  18:          0         -  83  mxs-dma
>>  25:          6         -  96  mxs-mmc
>>  26:          0         -  97  mxs-mmc
>> 195:          0         -  80  mxs-dma
>> 196:        141         -  81  mxs-dma
>> 197:         73         -  68  mxs-dma
>> 213:          0         -  59  mxs-saif
>> 214:          2         -  58  mxs-saif
>> 215:          0         -  10  mxs-lradc-touchscreen
>> 216:          0         -  14  mxs-lradc-thresh0
>> 217:          0         -  15  mxs-lradc-thresh1
>> 218:          0         -  16  mxs-lradc-channel0
>> 219:          0         -  17  mxs-lradc-channel1
>> 220:          0         -  18  mxs-lradc-channel2
>> 221:          0         -  19  mxs-lradc-channel3
>> 222:          0         -  20  mxs-lradc-channel4
>> 223:          0         -  21  mxs-lradc-channel5
>> 224:          0         -  22  mxs-lradc-channel6
>> 225:          0         -  23  mxs-lradc-channel7
>> 226:          0         -  24  mxs-lradc-button0
>> 227:          0         -  25  mxs-lradc-button1
>> 228:          0         -  29  RTC alarm
>> 229:         79         - 111  80058000.i2c
>> 233:          0         - 112  8006a000.serial
>> 234:          0         - 115  80070000.serial
>> 235:       1759         -  47  uart-pl011
>> 236:          0         -  93  ci_hdrc_imx
>> 237:          0         -  92  ci_hdrc_imx
>> 238:         64         - 101  800f0000.ethernet
>> 239:          0         - 102  800f4000.ethernet
>> Err:          0
>
> I suppose IRQ 196 and 197 are used for SAIF core because I can see them changing when recording...
>
>
> and here the same for EK340:
>
>> # cat /proc/interrupts
>>            CPU0
>>  16:      28189         -  48  MXS Timer Tick
>>  17:       6735         -  82  mxs-dma
>>  18:          0         -  83  mxs-dma
>>  20:        989         -  85  mxs-dma
>>  25:         11         -  96  mxs-mmc
>>  26:          0         -  97  mxs-mmc
>>  27:         24         -  99  mxs-sdio
>>  48:          0  gpio-mxs  19  80064000.pwm
>> 196:          0         -  80  mxs-dma
>> 197:          0         -  81  mxs-dma
>> 198:         76         -  68  mxs-dma
>> 199:          0         -  69  mxs-dma
>> 213:          0         -  59  mxs-saif
>> 214:          3         -  58  mxs-saif
>> 215:          1         -  10  mxs-lradc-touchscreen
>> 216:          0         -  14  mxs-lradc-thresh0
>> 217:          0         -  15  mxs-lradc-thresh1
>> 218:          0         -  16  mxs-lradc-channel0
>> 219:          0         -  17  mxs-lradc-channel1
>> 220:          0         -  18  mxs-lradc-channel2
>> 221:          0         -  19  mxs-lradc-channel3
>> 222:          0         -  20  mxs-lradc-channel4
>> 223:          0         -  21  mxs-lradc-channel5
>> 224:          0         -  22  mxs-lradc-channel6
>> 225:          0         -  23  mxs-lradc-channel7
>> 226:          0         -  24  mxs-lradc-button0
>> 227:          0         -  25  mxs-lradc-button1
>> 228:          0         -  29  RTC alarm
>> 229:         84         - 111  80058000.i2c
>> 230:          0         - 110  8005a000.i2c
>> 234:          0         - 113  8006c000.serial
>> 235:          0         - 114  8006e000.serial
>> 236:       2062         -  47  uart-pl011
>> 237:          0         -  93  ci_hdrc_imx
>> 238:          0         -  92  ci_hdrc_imx
>> 239:         64         - 101  800f0000.ethernet
>> Err:          0
>
> When recording in EK340, the only IRQ is changing between recording sessions is the 198, but it is not present in the IMX28EVK...
>
> Some device-tree clash???

On the EK340 I have 3 sdconnector and I use the SDIO3 for wifi. On the 
IMX28EVK I still have only two sd-card connector.
Is it something related to it?

the IRQs number 198 and 199 what do they mean?? How can I check where 
they are claimed??

Regards,
-- 
Eurek s.r.l.                          |
Electronic Engineering                | http://www.eurek.it
via Celletta 8/B, 40026 Imola, Italy  | Phone: +39-(0)542-609120
p.iva 00690621206 - c.f. 04020030377  | Fax:   +39-(0)542-609212