[U-Boot] migration from Micron to ISSI RAM

[Mauro Marinho]

Hello,
     Continuing with this subject. I didn't mention the registers values
posted last email was already changed for ISSI RAM. Having said that, I have
3 different *write errors msgs and contexts*. Althought they look the same,
they are:

   - *Memory (date line) error at 00000000, wrote cccccccccccccccc, read
   89abcdef89abcdef ! **...*
   - Memory error at 0000018c, wrote 00000000, read 0000ff77 !


   - Memory (address line) error at 000007f8<->000007fc, XOR value 00000004
   !

      Is there differences between these errors? where can I find
information about it?

Regards,
Mauro César

2010/11/24 Mauro Marinho <listasmaneco at gmail.com>

> Hi there,
>      It's my first contact to the group. I just started working with U-boot
> from my Job here in Brazil. Does anyone has already worked with any of the
> RAM's below? both 2 Meg x 16 x 4 banks:
>
>    - Micron - MT48LC8M16A2;
>    - ISSI - IS42S16800E -75ETL;
>
>  I got the mission to make the  necessary software  changes to make a
> MPC5200B board to work with our new RAM from ISSI provider, model
> IS42S16800E which is a 128Mb (16MB), 133Mhz, CAS = 2. So it means the board
> was already working fine with the OLD micron RAM. Offcourse I tried to load
> the same u-boot version to experiment. It failed with the same error
> below(bolded). I have the "*MPC5200B SDRAM Initialization and
> Configuration"* document I used as reference.
>
>      There are 2 CI's on CS0 and CS1 is disabled. So we have 32MB(2x16MB)
> available onBoard. I noticed from the study on each datasheet that Micron
> was configured with CAS Latency = 3 the only clear difference from
> ISSI(CAS=2), and some small timing differences I used on calculations. Then
> i've changed the *control, mode, config1 and config2* REGISTERS to
> corroborate ISSI's datasheet.
>      After that, I compiled and loaded U-boot on the Board. Then, I got a
> WRITE error on screen BELOW :
>
> *U-Boot 1.1.6.Autotrac-R2a-g6b4fc061-dirty (Nov 23 2010 - 10:53:30)*
> *
> *
> *CPU:   MPC5200 v1.2, Core v1.1 at 396 MHz*
> *       Bus 132 MHz, IPB 66 MHz, PCI 33 MHz*
> *Board: Autotrac UCC*
> *DRAM:  32 MB*
> *Memory (date line) error at 00000000, wrote aaaaaaaaaaaaaaaa, read
> 89abcdef89abcdef !*
> *Memory (date line) error at 00000000, wrote cccccccccccccccc, read
> 89abcdef89abcdef !*
> *...*
>
> What could it be? where should I pay more attention? doesn't it look like a
> timing problem?
>
>
>
> The header portion extracted;
> /*
>  * SDRAM configuration
>  */
> /* 2x IS42S16800E - 7.5 ns SDRAMS = 32 MegaBytes Total */
> *#define CFG_DRAM_DDR           0*           /* 0 = SDR mode;   1 = DDR
> mode*/
> #*define CFG_DRAM_EMODE   0 *          /* DDR usage*/
> *#define CFG_DRAM_MODE   0x008D*          //* pg.250 0.1100.11[01] (CAS=2
> burst=8) */
> *#define CFG_DRAM_CONTROL   0x504F0000*   /* pg.251 refresh=0xF half the
> max(64ms) */
> *#define CFG_DRAM_CONFIG1   0xC2222A00*  // 15/7.5    oldValue=0xD2333A00
> /* pg.254 */
> *#define CFG_DRAM_CONFIG2   0x8AC70000*  //* 0x8AD70000 pg.255 */
> *#define CFG_DRAM_TAP_DEL   0x01*     /* pg.146  */
> *#define CFG_DRAM_RAM_SIZE 0x18  *           /* Bitset 11000 -> 32M pg.126
> */
>
>
> If there's any missing info, please let me know.
> My Best Regards,
> --
> Mauro César
>



-- 
--
Mauro César
email/MSN/Gtalk:maneko at gmail.com <Gtalk%3Amaneko at gmail.com>