updating u-boot_2013.07.bb recipe

DISCUSSION Cédric Le Goater 2 participants 3 messages May 27, 2016

Share & tools

Cédric Le Goater

author

2016-05-27 09:15 UTC #0 lore raw

Hello.

The uboot recipe in openbmc is still at 4b44678c11. Without them, I see memory corruption under qemu, so I suppose this is the same for real HW.

commit 6f7818cdb8d8 "net: aspeednic: Do not start hardware in initialize" seems to solve the problem.

Is there a reason for staying at that level in uboot ? Could we take all Milton's patches in ?

Thanks,

Joel Stanley

2016-05-27 10:13 UTC #1 lore raw

Hey Cedric,

Quoted 8 lines · Cédric Le Goater

The uboot recipe in openbmc is still at 4b44678c11. Without them, I see memory corruption under qemu, so I suppose this is the same for real HW.
commit 6f7818cdb8d8 "net: aspeednic: Do not start hardware in initialize" seems to solve the problem.
Is there a reason for staying at that level in uboot ? Could we take all Milton's patches in ?

We updated to 44f1262bdf39ad93032d39f17a298165372be82e just yesterday:

https://github.com/openbmc/openbmc/commit/265b0d55113d364554102d532177916741c8c379

The hold up related to some discussion around not going back to the default compiler:

https://github.com/openbmc/openbmc/pull/227

I see Patrick closed this yesterday. I'm not sure what the plans are there; I would strongly recommend we go back to the more modern compiler.

Cheers,

Joel

Cédric Le Goater

author

2016-05-27 11:50 UTC #2 lore raw

Hello Joel,

Quoted 9 lines · Joel Stanley

Hey Cedric,
Quoted 6 lines · Cédric Le Goater
The uboot recipe in openbmc is still at 4b44678c11. Without them, I see memory corruption under qemu, so I suppose this is the same for real HW.
commit 6f7818cdb8d8 "net: aspeednic: Do not start hardware in initialize" seems to solve the problem.

Quoted 6 lines

Quoted 2 lines
Is there a reason for staying at that level in uboot ? Could we take all Milton's patches in ?
We updated to 44f1262bdf39ad93032d39f17a298165372be82e just yesterday:
https://github.com/openbmc/openbmc/commit/265b0d55113d364554102d532177916741c8c379

Ah. I missed that.

So, I baked a new flash and all is fine now. The generated flash boots the qemu guest without modification.

Quoted 7 lines

The hold up related to some discussion around not going back to the default compiler:
https://github.com/openbmc/openbmc/pull/227
I see Patrick closed this yesterday. I'm not sure what the plans are there; I would strongly recommend we go back to the more modern compiler.

u-boot.bin boots a palmetto bmc when compiled with :

	gcc version 5.2.0 (crosstool-NG crosstool-ng-1.22.0)

I didn't try the kernel though.

Cheers,