[U-Boot] [PATCH v9 18/30] nand: mxc: Switch NAND SPL to generic SPL

[Benoît Thébaudeau]

On Tuesday, April 9, 2013 5:04:45 PM, Benoît Thébaudeau wrote:
> On Tuesday, April 9, 2013 4:38:23 PM, Benoît Thébaudeau wrote:
> > Hi Fabio,
> > 
> > On Tuesday, April 9, 2013 2:46:21 PM, Fabio Estevam wrote:
> > > On Tue, Apr 9, 2013 at 9:30 AM, Fabio Estevam <festevam at gmail.com> wrote:
> > > 
> > > > Just tested v10 on mx31pdk.
> > > >
> > > > As reported in v9 I start to get reboots and never reach the prompt.
> > > >
> > > > Then I disabled watchdog options in mx31pdk.h and I get the following:
> > > >
> > > > U-Boot 2013.04-rc1-14135-gab21c72-dirty (Apr 09 2013 - 09:23:01)
> > > >
> > > > CPU:   Freescale i.MX31 rev 2.0 at 532 MHz.
> > > > Reset cause: POR
> > > > Board: MX31PDK
> > > > DRAM:  128 MiB
> > > > NAND:  256 MiB
> > > > In:    serial
> > > > Out:   serial
> > > > Err:   serial
> > > > Net:
> > > >
> > > > (Board hangs here).
> > > >
> > > > When I tested v9 and removed watchdog support I was able to get into
> > > > the U-boot prompt.
> > > 
> > > I just added "#define DEBUG" in arch/arm/lib/board to start debugging
> > > this and for my surprise I can reach the prompt now:
> > > 
> > > U-Boot 2013.04-rc1-14135-gab21c72-dirty (Apr 09 2013 - 09:35:51)
> > > 
> > > U-Boot code: 87F00000 -> 87F2C34C  BSS: -> 87F3137C
> > > CPU:   Freescale i.MX31 rev 2.0 at 532 MHz.
> > > Reset cause: POR
> > > Board: MX31PDK
> > > monitor len: 0003137C
> > > ramsize: 08000000
> > > TLB table from 87ff0000 to 87ff4000
> > > Top of RAM usable for U-Boot at: 87ff0000
> > > Reserving 196k for U-Boot at: 87fbe000
> > > Reserving 640k for malloc() at: 87f1e000
> > > Reserving 32 Bytes for Board Info at: 87f1dfe0
> > > Reserving 160 Bytes for Global Data at: 87f1df40
> > > New Stack Pointer is: 87f1df30
> > > RAM Configuration:
> > > Bank #0: 80000000 128 MiB
> > > relocation Offset is: 000be000
> > > monitor flash len: 000310F4
> > > Now running in RAM - U-Boot at: 87fbe000
> > > NAND:  256 MiB
> > > In:    serial
> > > Out:   serial
> > > Err:   serial
> > > Net:   smc911x-0
> > > Hit any key to stop autoboot:  0
> > > MX31PDK U-Boot >
> > > 
> > > I remember to have added "#define DEBUG" in v9 as well, so looks like
> > > the behavior between v9 and v10 is consistent.
> > > 
> > > Now we need to understand why the prompt is not reached without
> > > ""#define DEBUG".
> > 
> > Can you try different values for the following configs in mx31pdk.h?
> >  - CONFIG_SYS_NAND_U_BOOT_SIZE
> >  - CONFIG_SPL_TEXT_BASE
> >  - CONFIG_SYS_TEXT_BASE
> >  - (CONFIG_ENV_OFFSET)
> >  - (CONFIG_ENV_OFFSET_REDUND)
> > 
> > I would try:
> >  - 0x40000 for CONFIG_SYS_NAND_U_BOOT_SIZE
> >  - 0x86000000 for CONFIG_SPL_TEXT_BASE
> >  - 0x87000000 for CONFIG_SYS_TEXT_BASE
> 
> I have retrieved your previous test trace:
> ---
> U-Boot 2013.01 (Feb 22 2013 - 16:25:48)
> 
> CPU:   Freescale i.MX31 rev 2.0 at 532 MHz.
> Reset cause: POR
> Board: MX31PDK
> DRAM:  128 MiB
> 
> (hangs here).
> ---
> 
> So it does not hang at the same step.
> 
> Can you try to disable CONFIG_CMD_NET or to comment out eth_initialize() in
> arch/arm/lib/board.c? Do you know if something has changed in this function
> recently, or if there is something about it that has not yet been merged in
> the
> arm tree?

gd->bd might also be broken for some reason. Can you print it with and without
DEBUG, right before the call to eth_initialize()?

Also, can you try to apply this patch?
http://patchwork.ozlabs.org/patch/233964/

Best regards,
Benoît