Skip to contentSkip to conversation

broken patch "do not fixup NULL ptrs" ?

U-Boot André Schwarz 2 participants 4 messages Oct 22, 2010 – Oct 23, 2010

lore
Share & tools
lore
#0 lore raw

Joakim,

your commit d1e0b10accdbac2e0a8b2cbf7c589645442f87c5 renders my MPC8377 based board unbootable.

Two weeks ago I got this after power on :

U-Boot 2010.09-00486-gfda6278-dirty (Oct 22 2010 - 16:19:20) MPC83XX

Reset Status:

CPU:   e300c4, MPC8377, Rev: 2.1 at 533.333 MHz, CSB: 266.667 MHz
Board: Matrix Vision MergerBox

....

After rebasing my work to current master completely silences the serial line and the CPU doesn't come up. Bisection leads to your patch.

reverting it makes the board work again.

Any ideas ?

BTW:
- I'm using ELDK-4.2 toolchain.
- mvBLM7 (=MPC8343 based) works fine.

Regards,
Andr?

MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

#1 lore raw

Andr? Schwarz <[email protected]> wrote on 2010/10/22 16:30:49:

Quoted 21 lines

Joakim,

your commit d1e0b10accdbac2e0a8b2cbf7c589645442f87c5 renders my MPC8377 based board unbootable.

Two weeks ago I got this after power on :

U-Boot 2010.09-00486-gfda6278-dirty (Oct 22 2010 - 16:19:20) MPC83XX

Reset Status:

CPU:   e300c4, MPC8377, Rev: 2.1 at 533.333 MHz, CSB: 266.667 MHz
Board: Matrix Vision MergerBox

....

Quoted 6 lines

After rebasing my work to current master completely silences the serial line and the CPU doesn't come up. Bisection leads to your patch.

reverting it makes the board work again.

Any ideas ?

Yes, the problem is else where. This code runs after relocation to RAM. I had a similar problem and the reason was the patch:
"powerpc: Remove warm reset entry point"

Try sticking 4 nop's right after _start:
in start.S

#2 lore raw
Quoted 26 lines · Joakim Tjernlund

Andr? Schwarz <[email protected]> wrote on 2010/10/22 16:30:49:

Quoted 24 lines

Joakim,

your commit d1e0b10accdbac2e0a8b2cbf7c589645442f87c5 renders my MPC8377 based board unbootable.

Two weeks ago I got this after power on :

U-Boot 2010.09-00486-gfda6278-dirty (Oct 22 2010 - 16:19:20) MPC83XX

Reset Status:

CPU:   e300c4, MPC8377, Rev: 2.1 at 533.333 MHz, CSB: 266.667 MHz
Board: Matrix Vision MergerBox

....

After rebasing my work to current master completely silences the serial line and the CPU doesn't come up. Bisection leads to your patch.

Quoted 7 lines
Quoted 3 lines

reverting it makes the board work again.

Any ideas ?

Yes, the problem is else where. This code runs after relocation to RAM. I had a similar problem and the reason was the patch:
"powerpc: Remove warm reset entry point"

hmm, ok.
But what's the root cause ?

Quoted 4 lines

Try sticking 4 nop's right after _start:
in start.S

... and why should this work ?
Has this something to do with trial and error ?

Can you shed some light on this ?

Regards,
Andr?

MATRIX VISION GmbH, Talstrasse 16, DE-71570 Oppenweiler
Registergericht: Amtsgericht Stuttgart, HRB 271090
Geschaeftsfuehrer: Gerhard Thullner, Werner Armingeon, Uwe Furtner

#3 lore raw

Andr? Schwarz <[email protected]> wrote on 2010/10/22 21:25:34:

Quoted 30 lines · Joakim Tjernlund
Quoted 28 lines

Andr? Schwarz <[email protected]> wrote on 2010/10/22 16:30:49:

Quoted 26 lines

Joakim,

your commit d1e0b10accdbac2e0a8b2cbf7c589645442f87c5 renders my MPC8377 based board unbootable.

Two weeks ago I got this after power on :

U-Boot 2010.09-00486-gfda6278-dirty (Oct 22 2010 - 16:19:20) MPC83XX

Reset Status:

CPU:   e300c4, MPC8377, Rev: 2.1 at 533.333 MHz, CSB: 266.667 MHz
Board: Matrix Vision MergerBox

....

After rebasing my work to current master completely silences the serial line and the CPU doesn't come up. Bisection leads to your patch.

reverting it makes the board work again.

Quoted 8 lines
Quoted 5 lines
Quoted 1 line

Any ideas ?

Yes, the problem is else where. This code runs after relocation to RAM. I had a similar problem and the reason was the patch:
"powerpc: Remove warm reset entry point"

hmm, ok.
But what's the root cause ?

Not sure, I think my board still use 0x110 as boot entry

Quoted 8 lines
Quoted 5 lines

Try sticking 4 nop's right after _start:
in start.S

... and why should this work ?
Has this something to do with trial and error ?

yes, see above.

Quoted 2 lines

Can you shed some light on this ?

not much, can you? Did the extra nop's make any
difference?

Jocke