1:59 a.m.
From 278a406ff09f52679ae9a7d4126344ddc7a83d3a Mon Sep 17 00:00:00 2001
From: Paul Sherwood <paul.sherwood(a)codethink.co.uk>
Date: Thu, 12 Sep 2013 16:02:21 +0000
Subject: [PATCH] force make to notice $MAKEFLAGS
We noticed that kernels have not been building on all cores, even
though we're setting MAKE FLAGS. This fixes that.
---
linux.morph | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/linux.morph b/linux.morph
index e11a50a..5278f8e 100644
--- a/linux.morph
+++ b/linux.morph
@@ -55,7 +55,7 @@
"yes '' | make oldconfig"
],
"build-commands": [
- "make"
+ "make $MAKEFLAGS"
],
"install-commands": [
"mkdir -p \"$DESTDIR\"/boot",
--
1.7.9-rc0
6:57 p.m.
On Thu, Sep 12, 2013 at 05:59:47PM +0100, Paul Sherwood wrote:
From 278a406ff09f52679ae9a7d4126344ddc7a83d3a Mon Sep 17 00:00:00
2001
From: Paul Sherwood <paul.sherwood(a)codethink.co.uk>
Date: Thu, 12 Sep 2013 16:02:21 +0000
Subject: [PATCH] force make to notice $MAKEFLAGS
We noticed that kernels have not been building on all cores, even
though we're setting MAKE FLAGS. This fixes that.
---
linux.morph | 2 +-
1 files changed, 1 insertions(+), 1 deletions(-)
diff --git a/linux.morph b/linux.morph
index e11a50a..5278f8e 100644
--- a/linux.morph
+++ b/linux.morph
@@ -55,7 +55,7 @@
"yes '' | make oldconfig"
],
"build-commands": [
- "make"
+ "make $MAKEFLAGS"
],
"install-commands": [
"mkdir -p \"$DESTDIR\"/boot",
That's a +1 from me.
In general I think we should also change morph's build systems to use
`make $MAKEFLAGS`, but there seems to be an oddity with linux that it
ignores MAKEFLAGS from the environment.
9:46 p.m.
On Thu, Sep 12, 2013 at 17:59:47 +0100, Paul Sherwood wrote:
From 278a406ff09f52679ae9a7d4126344ddc7a83d3a Mon Sep 17 00:00:00
2001
The patch itself is okay, but since it doesn't address any of the other BSPs,
I'm tempted to -1 this.
I'd rather sort out a kernel buildsystem and have that solve this once and for
all.
Overall my vote is -0.
D.
--
Daniel Silverstone http://www.codethink.co.uk/
Principal Managing Engineer GPG 4096/R Key Id: 3CCE BABE 206C 3B69
11:41 p.m.
On 16/09/2013 13:46, Daniel Silverstone wrote:
On Thu, Sep 12, 2013 at 17:59:47 +0100, Paul Sherwood wrote:
> From 278a406ff09f52679ae9a7d4126344ddc7a83d3a Mon Sep 17 00:00:00 2001
The patch itself is okay, but since it doesn't address any of the other BSPs,
I'm tempted to -1 this.
OK - but in the meantime I don't see this as a reason to reject this patch?
I'd rather sort out a kernel buildsystem and have that solve this
once and for
all.
Overall my vote is -0.
2:46 a.m.
On Mon, Sep 16, 2013 at 03:41:50PM +0100, Paul Sherwood wrote:
On 16/09/2013 13:46, Daniel Silverstone wrote:
> On Thu, Sep 12, 2013 at 17:59:47 +0100, Paul Sherwood wrote:
>> From 278a406ff09f52679ae9a7d4126344ddc7a83d3a Mon Sep 17 00:00:00 2001
>
> The patch itself is okay, but since it doesn't address any of the other BSPs,
> I'm tempted to -1 this.
OK - but in the meantime I don't see this as a reason to reject this patch?
I think we want to change into Baserock sooner rather than later: it
affects build speeds quite dramatically. However, doing so will
currently require changing potentially many linux.morphs in a number
of git branches, and that's quite tedious. Hence Daniel's wish to do
this properly by having Morph understand the kernel build system.
However, the tedious thing should be doable in a day (perhaps two, but
then most of the time is spent waiting on builds), whereas the kernel
build system requires requirement gathering and design, so is likely
to take longer.
Assuming I can get some other things done and finished, I'd be happy
to do the tedious thing later this week or next week. Daniel, what do
you think?
--
http://www.codethink.co.uk/ http://wiki.baserock.org/ http://www.baserock.com/
3661
days inactive
3665
days old
4 comments
4 participants
participants (4)
-
Daniel Silverstone -
Lars Wirzenius -
Paul Sherwood -
Richard Maw