[U-Boot] powerpc, tqm5200s: fails with current head

[U-Boot] powerpc, tqm5200s: fails with current head

[Heiko Schocher]

Hello Simon,

I just tried current mainline on the tqm5200s board, and it fails.
git bisect showed:

f05ad9ba4c22b5c5929406394a0889791c2778da is the first bad commit
commit f05ad9ba4c22b5c5929406394a0889791c2778da
Author: Simon Glass <sjg@chromium.org>
Date:   Tue Aug 4 12:33:39 2015 -0600

Add a way to skip relocation

When running U-Boot as an EFI application we cannot relocate since we do not
have relocation information. U-Boot has already been relocated to a suitable
address.

Add a global_data flag to control skipping relocation.

Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Bin Meng <bmeng.cn@gmail.com>

You can look into the hole git bisect session here [1]

I reverted this patch, and then the board boots again [2]

Do you have an idea, what can be wrong here?
I try to find time to look into it.

bye,
Heiko
[1] tbot automated git bisect session
http://xeidos.ddns.net/buildbot/builders/tqm5200s/builds/3/steps/shell/logs/tbotlog

[2] bootlog without commit f05ad9ba
read 1: U-Boot 2016.01-rc1-00241-g71cbed3-dirty (Nov 24 2015 - 17:03:06 +0100)
read 1:
read 1: CPU:   MPC5200B v2.2, Core v1.4 at 396 MHz
read 1:        Bus 132 MHz, IPB 132 MHz, PCI 66 MHz
read 1: Board: TQM5200S (TQ-Components GmbH)
read 1:        on a STK52xx carrier board
read 1: I2C:   85 kHz, ready
read 1: DRAM:  64 MiB
read 1: Flash: 32 MiB
read 1: In:    serial
read 1: Out:   serial
read 1: Err:   serial
read 1: Net:   FEC
read 1: IDE:   Bus 0: OK
read 1:   Device 0: Model: HITACHI_DK23DA-20 Firm: 00J2A0A1 Ser#: 12Y0MN
read 1:             Type: Hard Disk
read 1:             Capacity: 19077.1 MB = 18.6 GB (39070080 x 512)
read 1:   Device 1: not available
read 1: SRAM:  512 kB
read 1:
PS/2:
read 1: No device found
read 1:

read 1: Kbd:   reset failed, no ACK
read 1:
read 1: Type run flash_nfs to mount root filesystem over NFS
read 1:
read 1:
=>


-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany

[U-Boot] powerpc, tqm5200s: fails with current head

[Simon Glass]

Hi Heiko,

On 24 November 2015 at 09:10, Heiko Schocher <hs@denx.de> wrote:
> Hello Simon,
>
> I just tried current mainline on the tqm5200s board, and it fails.
> git bisect showed:
>
> f05ad9ba4c22b5c5929406394a0889791c2778da is the first bad commit
> commit f05ad9ba4c22b5c5929406394a0889791c2778da
> Author: Simon Glass <sjg@chromium.org>
> Date:   Tue Aug 4 12:33:39 2015 -0600
>
> Add a way to skip relocation
>
> When running U-Boot as an EFI application we cannot relocate since we do not
> have relocation information. U-Boot has already been relocated to a suitable
> address.
>
> Add a global_data flag to control skipping relocation.
>
> Signed-off-by: Simon Glass <sjg@chromium.org>
> Reviewed-by: Bin Meng <bmeng.cn@gmail.com>
>
> You can look into the hole git bisect session here [1]
>
> I reverted this patch, and then the board boots again [2]
>
> Do you have an idea, what can be wrong here?
> I try to find time to look into it.
>
> bye,
> Heiko
> [1] tbot automated git bisect session
> http://xeidos.ddns.net/buildbot/builders/tqm5200s/builds/3/steps/shell/logs/tbotlog
>
> [2] bootlog without commit f05ad9ba
> read 1: U-Boot 2016.01-rc1-00241-g71cbed3-dirty (Nov 24 2015 - 17:03:06
> +0100)
> read 1:
> read 1: CPU:   MPC5200B v2.2, Core v1.4 at 396 MHz
> read 1:        Bus 132 MHz, IPB 132 MHz, PCI 66 MHz
> read 1: Board: TQM5200S (TQ-Components GmbH)
> read 1:        on a STK52xx carrier board
> read 1: I2C:   85 kHz, ready
> read 1: DRAM:  64 MiB
> read 1: Flash: 32 MiB
> read 1: In:    serial
> read 1: Out:   serial
> read 1: Err:   serial
> read 1: Net:   FEC
> read 1: IDE:   Bus 0: OK
> read 1:   Device 0: Model: HITACHI_DK23DA-20 Firm: 00J2A0A1 Ser#: 12Y0MN
> read 1:             Type: Hard Disk
> read 1:             Capacity: 19077.1 MB = 18.6 GB (39070080 x 512)
> read 1:   Device 1: not available
> read 1: SRAM:  512 kB
> read 1:
> PS/2:
> read 1: No device found
> read 1:
>
> read 1: Kbd:   reset failed, no ACK
> read 1:
> read 1: Type run flash_nfs to mount root filesystem over NFS
> read 1:
> read 1:
> =>

I wonder if the GD_FLG_SKIP_RELOC flag is somehow being set? I don't
see any local powerpc flags, though. It's hard to imagine any other
reason though.

Does the code in arch/powerpc/cpu/mpc85xx/start.S set the parameter to
0 before calling board_init_f()? I can't see it.

Regards,
Simon