Skip to contentSkip to conversation

Using yocto/standard by default

DISCUSSION Darren Hart 2 participants 3 messages Dec 13, 2011 – Dec 14, 2011

lore
Share & tools
lore
#0 lore raw

We hit another lock-step SRCREV bug earlier on the FRI2 BSP. This was due mostly to my pushing the efi changes to meta-intel too early - but, it highlights a maintenance step that I believe could be eliminated for most boards.

We have a yocto/standard/fri2 branch, but it doesn't contain any additional changes over yocto/standard/base. If we were to make yocto/standard/base the default for KBRANCH, shouldn't we be able to eliminate all the BSP branches that are identical to yocto/standard/base? This would significantly reduce the number of SRCREV updates that are required and likely reduce the number of Autobuilder failures we experience as a result. Seems like it would also help make the git tree easier to deal with.

Any opinions here?

Signature · 3 lines
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel
#1 lore raw
Quoted 14 lines · Darren Hart

We hit another lock-step SRCREV bug earlier on the FRI2 BSP. This was due mostly to my pushing the efi changes to meta-intel too early - but, it highlights a maintenance step that I believe could be eliminated for most boards.

We have a yocto/standard/fri2 branch, but it doesn't contain any additional changes over yocto/standard/base. If we were to make yocto/standard/base the default for KBRANCH, shouldn't we be able to eliminate all the BSP branches that are identical to yocto/standard/base? This would significantly reduce the number of SRCREV updates that are required and likely reduce the number of Autobuilder failures we experience as a result. Seems like it would also help make the git tree easier to deal with.

Quoted 1 line

Any opinions here?

It's a valid config, and something that works now. So there's no reason to not use it. New branches can be created IF a board really does need to merge conflicting patches. The emgd stuff was a problem and required branches, but if we have nothing like that, squashing the branches is a nice simplification.

Cheers,

Bruce

Quoted 1 line
#2 lore raw
Quoted 15 lines · Bruce Ashfield
Quoted 14 lines · Darren Hart

We hit another lock-step SRCREV bug earlier on the FRI2 BSP. This was due mostly to my pushing the efi changes to meta-intel too early - but, it highlights a maintenance step that I believe could be eliminated for most boards.

We have a yocto/standard/fri2 branch, but it doesn't contain any additional changes over yocto/standard/base. If we were to make yocto/standard/base the default for KBRANCH, shouldn't we be able to eliminate all the BSP branches that are identical to yocto/standard/base? This would significantly reduce the number of SRCREV updates that are required and likely reduce the number of Autobuilder failures we experience as a result. Seems like it would also help make the git tree easier to deal with.

Quoted 8 lines
Quoted 1 line

Any opinions here?

It's a valid config, and something that works now. So there's no reason to not use it. New branches can be created IF a board really does need to merge conflicting patches. The emgd stuff was a problem and required branches, but if we have nothing like that, squashing the branches is a nice simplification.

Hrm maybe I missed that in the fri2 branches. It does need emgd, so I'll double check that.

Quoted 6 lines

Cheers,

Bruce

Quoted 1 line
Signature · 3 lines
Darren Hart
Intel Open Source Technology Center
Yocto Project - Linux Kernel