* [Buildroot] Atmel specific handling in buildroot
@ 2009-01-30 19:33 Peter Korsgaard
2009-02-01 15:03 ` Peter Korsgaard
0 siblings, 1 reply; 6+ messages in thread
From: Peter Korsgaard @ 2009-01-30 19:33 UTC (permalink / raw)
To: buildroot
Hi,
Why is it again we have special handling of Atmel targets?
git grep -B2 'default.*if BR2_TARGET_ATMEL' ~/source/buildroot
target/Config.in-choice
target/Config.in- prompt "Kernel type"
target/Config.in: default BR2_KERNEL_LINUX_ADVANCED if BR2_TARGET_ATMEL
target/device/Config.in.linux.patches- depends on !BR2_KERNEL_PREPATCHED
target/device/Config.in.linux.patches- default y if BR2_avr32
target/device/Config.in.linux.patches: default y if BR2_TARGET_ATMEL && BR2_arm
--
target/linux/Config.in.advanced-config BR2_LINUX_COPYTO_TFTPBOOT
target/linux/Config.in.advanced- bool "Copy kernel to /tftpboot"
target/linux/Config.in.advanced: default y if BR2_TARGET_ATMEL
--
target/linux/Config.in.advanced-config BR2_LINUX_COPY_CONFIGURATION
target/linux/Config.in.advanced- bool "Copy buildroot configuration to Li
target/linux/Config.in.advanced: default y if BR2_TARGET_ATMEL
Do we really need that? It would imho be a lot more sensible if BR
behaved the same way for all archs.
--
Bye, Peter Korsgaard
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Buildroot] Atmel specific handling in buildroot
2009-01-30 19:33 [Buildroot] Atmel specific handling in buildroot Peter Korsgaard
@ 2009-02-01 15:03 ` Peter Korsgaard
2009-02-01 15:36 ` Ulf Samuelsson
0 siblings, 1 reply; 6+ messages in thread
From: Peter Korsgaard @ 2009-02-01 15:03 UTC (permalink / raw)
To: buildroot
>>>>> "Peter" == Peter Korsgaard <jacmet@uclibc.org> writes:
Peter> Hi,
Peter> Why is it again we have special handling of Atmel targets?
..
Peter> target/linux/Config.in.advanced: default y if BR2_TARGET_ATMEL
Peter> Do we really need that? It would imho be a lot more sensible if BR
Peter> behaved the same way for all archs.
No objections? Good, then I'll remove them.
--
Bye, Peter Korsgaard
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Buildroot] Atmel specific handling in buildroot
2009-02-01 15:03 ` Peter Korsgaard
@ 2009-02-01 15:36 ` Ulf Samuelsson
2009-02-01 16:13 ` Peter Korsgaard
0 siblings, 1 reply; 6+ messages in thread
From: Ulf Samuelsson @ 2009-02-01 15:36 UTC (permalink / raw)
To: buildroot
s?n 2009-02-01 klockan 16:03 +0100 skrev Peter Korsgaard:
> >>>>> "Peter" == Peter Korsgaard <jacmet@uclibc.org> writes:
>
> Peter> Hi,
> Peter> Why is it again we have special handling of Atmel targets?
>
> ..
> Peter> target/linux/Config.in.advanced: default y if BR2_TARGET_ATMEL
>
> Peter> Do we really need that? It would imho be a lot more sensible if BR
> Peter> behaved the same way for all archs.
There are several things here
1) If there is an architecture patch available for an architecture,
then it should have default yes.
I can't think of a situation where you prefer not to apply
the architecture patch for the AT91 or AVR32.
The only reason is that you are building a kernel for where
there is no valid patch.
If there is no architecture patch, then obviously
it should be default no.
2) I think that we should reduce the choice to
only have no kernel or the advanced configuration.
3) For copy to, that should be collected to a single copy to
BR
Ulf Samuelsson
>
> No objections? Good, then I'll remove them.
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Buildroot] Atmel specific handling in buildroot
2009-02-01 15:36 ` Ulf Samuelsson
@ 2009-02-01 16:13 ` Peter Korsgaard
2009-02-01 17:33 ` Ulf Samuelsson
0 siblings, 1 reply; 6+ messages in thread
From: Peter Korsgaard @ 2009-02-01 16:13 UTC (permalink / raw)
To: buildroot
>>>>> "Ulf" == Ulf Samuelsson <ulf.samuelsson@atmel.com> writes:
Ulf> There are several things here
Ulf> 1) If there is an architecture patch available for an architecture,
Ulf> then it should have default yes.
Ulf> I can't think of a situation where you prefer not to apply
Ulf> the architecture patch for the AT91 or AVR32.
Ulf> The only reason is that you are building a kernel for where
Ulf> there is no valid patch.
Ulf> If there is no architecture patch, then obviously
Ulf> it should be default no.
Ok, but that's only one of the settings, so you're fine with me fixing
the COPYTO / kernel type thingies?
--
Bye, Peter Korsgaard
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Buildroot] Atmel specific handling in buildroot
2009-02-01 16:13 ` Peter Korsgaard
@ 2009-02-01 17:33 ` Ulf Samuelsson
2009-02-01 18:57 ` Peter Korsgaard
0 siblings, 1 reply; 6+ messages in thread
From: Ulf Samuelsson @ 2009-02-01 17:33 UTC (permalink / raw)
To: buildroot
s?n 2009-02-01 klockan 17:13 +0100 skrev Peter Korsgaard:
> >>>>> "Ulf" == Ulf Samuelsson <ulf.samuelsson@atmel.com> writes:
>
> Ulf> There are several things here
>
> Ulf> 1) If there is an architecture patch available for an architecture,
> Ulf> then it should have default yes.
> Ulf> I can't think of a situation where you prefer not to apply
> Ulf> the architecture patch for the AT91 or AVR32.
> Ulf> The only reason is that you are building a kernel for where
> Ulf> there is no valid patch.
> Ulf> If there is no architecture patch, then obviously
> Ulf> it should be default no.
>
> Ok, but that's only one of the settings, so you're fine with me fixing
> the COPYTO / kernel type thingies?
I dont think you will build correctly unless you use the advanced kernel
setting for the AVR32 and AT91.
Copying the build configuration to the root file system
is in line with the open source model, since
it helps people to reproduce whatever is on the board
if they buy the product.
If you want to debug something, then it is good to
know how it was configured.
If someone calls in with a board without this, then it is
too late to do anything about it.
Obviously, since I put it there, I want to have it ON
for all my customers.
As for the copyto, this is a convenience, and people
can configure whatever they want.
I find that it saves time to have this on.
BR
Ulf Samuelsson
^ permalink raw reply [flat|nested] 6+ messages in thread
* [Buildroot] Atmel specific handling in buildroot
2009-02-01 17:33 ` Ulf Samuelsson
@ 2009-02-01 18:57 ` Peter Korsgaard
0 siblings, 0 replies; 6+ messages in thread
From: Peter Korsgaard @ 2009-02-01 18:57 UTC (permalink / raw)
To: buildroot
>>>>> "Ulf" == Ulf Samuelsson <ulf.samuelsson@atmel.com> writes:
Hi,
Ulf> I dont think you will build correctly unless you use the
Ulf> advanced kernel setting for the AVR32 and AT91.
Huh, why wouldn't BR2_KERNEL_none work for those targets?
Ulf> Copying the build configuration to the root file system
Ulf> is in line with the open source model, since
Ulf> it helps people to reproduce whatever is on the board
Ulf> if they buy the product.
Not only that, That info is required by the license.
Ulf> If you want to debug something, then it is good to
Ulf> know how it was configured.
Ulf> If someone calls in with a board without this, then it is
Ulf> too late to do anything about it.
Ulf> Obviously, since I put it there, I want to have it ON
Ulf> for all my customers.
Fine, that's why we have the configuration in the first place I
guess - But the question isn't about what you or your customers want,
it's about what the default should be in BR.
It makes sense to default these to off like for everything else, so
I'll remove the atmel special case.
Ulf> As for the copyto, this is a convenience, and people
Ulf> can configure whatever they want.
Ulf> I find that it saves time to have this on.
Again, feel free to use it, but it shouldn't be default (E.G. it
breaks the build when you don't have a /tftpboot and/or it isn't
writable.
--
Bye, Peter Korsgaard
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2009-02-01 18:57 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2009-01-30 19:33 [Buildroot] Atmel specific handling in buildroot Peter Korsgaard
2009-02-01 15:03 ` Peter Korsgaard
2009-02-01 15:36 ` Ulf Samuelsson
2009-02-01 16:13 ` Peter Korsgaard
2009-02-01 17:33 ` Ulf Samuelsson
2009-02-01 18:57 ` Peter Korsgaard
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox