AM625-Q1: DDR4 not initializing- DDR Reset is never Released

Pawan, can you add

#define DEBUG

to the very top (above all #include) of k3_ddrss.c.  This will give us a little more verbose debug statements.  

Which SDK did you start with?  Linux or MCU+?  If linux, can you post the results of 'git diff' in u-boot.  I'd like to see what changes you have done bootloader.  If MCU+, can you post the changes you have made to the code?

Is it possible to perform a warm reset on your board, and does it produce the same behavior?

Regards,

James

Hi James

Sorry for the delay 

JJD said:

Linux or MCU+

We are using Linux.

JJD said:

This will give us a little more verbose debug statements.  

Did not get verbose after adding #define DEBUG in k3_ddrss.c. Don't know why ?

JJD said:

If linux, can you post the results of 'git diff' in u-boot

please find the changes made

diff --git a/board/ti/am62x/evm.c b/board/ti/am62x/evm.c
index 9146361b3a..b786329e41 100644
--- a/board/ti/am62x/evm.c
+++ b/board/ti/am62x/evm.c
@@ -50,7 +50,7 @@ static struct gpio_desc board_det_gpios[AM62X_LPSK_BRD_DET_COUNT];
 
 #define board_is_am62x_skevm()  (board_ti_k3_is("AM62-SKEVM") || \
                 board_ti_k3_is("AM62B-SKEVM"))
-#define board_is_am62b_p1_skevm() board_ti_k3_is("AM62B-SKEVM-P1")
+#define board_is_am62b_p1_skevm() 1
 #define board_is_am62x_lp_skevm()  board_ti_k3_is("AM62-LP-SKEVM")
 #define board_is_am62x_sip_skevm()  board_ti_k3_is("AM62SIP-SKEVM")
 #define board_is_am62x_play()  board_ti_k3_is("BEAGLEPLAY-A0-")
@@ -191,16 +191,18 @@ static void setup_board_eeprom_env(void)
    if (do_board_detect())
        goto invalid_eeprom;
 
-   if (board_is_am62x_skevm())
-       name = "am62x_skevm";
-   else if (board_is_am62b_p1_skevm())
+   if (1)
        name = "am62b_p1_skevm";
-   else if (board_is_am62x_lp_skevm())
-       name = "am62x_lp_skevm";
-   else if (board_is_am62x_sip_skevm())
-       name = "am62x_sip_skevm";
-   else if (board_is_am62x_play())
-       name = "am62x_beagleplay";
+   // if (board_is_am62x_skevm())
+   //  name = "am62x_skevm";
+   // else if (board_is_am62b_p1_skevm())
+   //  name = "am62b_p1_skevm";
+   // else if (board_is_am62x_lp_skevm())
+   //  name = "am62x_lp_skevm";
+   // else if (board_is_am62x_sip_skevm())
+   //  name = "am62x_sip_skevm";
+   // else if (board_is_am62x_play())
+   //  name = "am62x_beagleplay";
    else
        printf("Unidentified board claims %s in eeprom header\n",
               board_ti_get_name());
diff --git a/drivers/ram/k3-ddrss/k3-ddrss.c b/drivers/ram/k3-ddrss/k3-ddrss.c
index 322e8e86de..5bc53f078c 100644
--- a/drivers/ram/k3-ddrss/k3-ddrss.c
+++ b/drivers/ram/k3-ddrss/k3-ddrss.c
@@ -26,6 +26,8 @@
 #include "lpddr4_structs_if.h"
 #include "lpddr4_ctl_regs.h"
 
+#define DEBUG
+
 #define SRAM_MAX 512
 
 #define CTRLMMR_DDR4_FSP_CLKCHNG_REQ_OFFS  0x80

JJD said:

warm reset on your board, and does it produce the same behavior

warm reset has same behaviour (Reset always low)

Sure James, we will make the suggested change and probing and get back to you soon.

One more doubt, what does k3_lpddr4_start: Fail has to do with Board not having LPDDR4? (Our Board has only DDR4 ) ?

Regards

Pawan

Pawan, the k3_lpddr4_start function (as well as other functions) are part of the driver which supports both LPDDR4 and DDR4.  This is just a misnomer in the function name.  You will see certain parts of the driver where the functionality changes depending on whether you are using LPDDR4 or DDR4

Regards,

James

Hi James,

Sorry for the delay.

Thanks for suggesting, we found VDDA_DDR_PLL0 rail not connecting to Core voltage. We fixed it.

now we are getting RAM initialized but board fails at below stage

(Unable to import code with insert code so adding below)

U-Boot SPL 2023.04-ti-gd2612223375a (Feb 01 2024 - 15:06:40 +0000)
SYSFW ABI: 3.1 (firmware rev 0x0009 '9.2.7--v09.02.07 (Kool Koala)')
SPL initial stack usage: 13392 bytes
Trying to boot from MMC2
Authentication passed
Authentication passed
Authentication passed
Authentication passed
Authentication passed
Starting ATF on ARM64 core...

NOTICE: BL31: v2.10.0(release):v2.10.0-367-g00f1ec6b87-dirty
NOTICE: BL31: Built : 15:06:40, Feb 1 2024

U-Boot SPL 2023.04-ti-gd2612223375a (Feb 01 2024 - 15:06:40 +0000)
SYSFW ABI: 3.1 (firmware rev 0x0009 '9.2.7--v09.02.07 (Kool Koala)')
SPL initial stack usage: 1904 bytes
i2c_write: error waiting for data ACK (status=0x116)
pca953x gpio@22: Error reading output register
Trying to boot from MMC2
Authentication passed
Authentication passed

U-Boot 2023.04-ti-gd2612223375a (Feb 01 2024 - 15:06:40 +0000)

SoC: AM62X SR1.0 HS-FS
Model: Texas Instruments AM625 SK
EEPROM not available at 80, trying to read at 81
Reading on-board EEPROM at 0x51 failed -1
DRAM: 2 GiB
Core: 72 devices, 32 uclasses, devicetree: separate
MMC: mmc@fa10000: 0, mmc@fa00000: 1
Loading Environment from nowhere... OK
In: serial
Out: serial
Err: serial
EEPROM not available at 80, trying to read at 81
Reading on-board EEPROM at 0x51 failed -1
Net: eth0: ethernet@8000000port@1
Hit any key to stop autoboot: 0

Looks like you still have code in there that reads an EEPROM to get board info (which is what we have on the EVMs).  You will need to remove that.  Please check https://dev.ti.com/tirex/explore/node?node=A__AS3yAWRDEB3oRrtGCnsJ7A__AM62-ACADEMY__uiYMDcq__LATEST for some guidance on this.  

Regards,

James