package/makedumpfile: fix build for powerpc target

PATCH Superseded Alexander Egorenkov 2 participants 3 messages Sep 9, 2020 – Sep 10, 2020

lore

Share & tools

lore

Alexander Egorenkov

author

2020-09-09 16:41 UTC #0 lore raw

PATCH 1/1 package/makedumpfile: fix build for powerpc target

Signed-off-by Alexander Egorenkov <[email protected]>

---
 ...Makefile-fix-ARCH-for-TARGET-powerpc.patch | 28 +++++++++++++++++++
 1 file changed, 28 insertions(+)
 create mode 100644 package/makedumpfile/0003-Makefile-fix-ARCH-for-TARGET-powerpc.patch

package/makedumpfile/0003-Makefile-fix-ARCH-for-TARGET-powerpc.patch new file +28 -0

new file mode 100644
index 0000000000..ba75463ec4
@@ -0,0 +1,28 @@
+From 18acea4df1d02f0275606fa4e73292d7759bdf15 Mon Sep 17 00:00:00 2001
+From: Alexander Egorenkov <[email protected]>
+Date: Wed, 9 Sep 2020 18:35:05 +0200
+Subject: [PATCH] Makefile: fix ARCH for TARGET powerpc
+
+TARGET 'ppc' and 'powerpc' shall result in ARCH 'powerpc32'.
+
+Signed-off-by: Alexander Egorenkov <[email protected]>
+---
+ Makefile | 2 +-
+ 1 file changed, 1 insertion(+), 1 deletion(-)
+
+diff --git a/Makefile b/Makefile
+index ef20672..b55821c 100644
+--- a/Makefile
++++ b/Makefile
+@@ -24,7 +24,7 @@ endif
+ ARCH := $(shell echo ${TARGET}  | sed -e s/i.86/x86/ -e s/sun4u/sparc64/ \
+ 			       -e s/arm.*/arm/ -e s/sa110/arm/ \
+ 			       -e s/s390x/s390/ -e s/parisc64/parisc/ \
+-			       -e s/ppc64/powerpc64/ -e s/ppc/powerpc32/)
++			       -e s/ppc64/powerpc64/ -e s/ppc/powerpc32/ -e s/powerpc/powerpc32/)
+ 
+ CROSS :=
+ ifneq ($(TARGET), $(HOST_ARCH))
+-- 
+2.28.0
+

Signature · 1 line

2.28.0

Thomas Petazzoni

2020-09-09 21:20 UTC #1 lore raw

PATCH 1/1 package/makedumpfile: fix build for powerpc target

Hello Alexander,

Quoted 1 line · On Wed, 9 Sep 2020 18:41:04 +0200 Alexander Egorenkov…

Signed-off-by Alexander Egorenkov <[email protected]>

Is there an autobuilder failure for this? If so, there should be a link to it in your commit log. See for example this commit log:

https://git.buildroot.org/buildroot/commit/?id=a85ae8b297fd0526f57fe99d73087dd840ea6219

Quoted 6 lines

package/makedumpfile/0003-Makefile-fix-ARCH-for-TARGET-powerpc.patch new file +0 -0
new file mode 100644
index 0000000000..ba75463ec4

package/makedumpfile/0003-Makefile-fix-ARCH-for-TARGET-powerpc.patch line 8

Quoted 8 lines

package/makedumpfile/0003-Makefile-fix-ARCH-for-TARGET-powerpc.patch new file +8 -0

@@ -0,0 +1,28 @@
+From 18acea4df1d02f0275606fa4e73292d7759bdf15 Mon Sep 17 00:00:00 2001
+From: Alexander Egorenkov <[email protected]>
+Date: Wed, 9 Sep 2020 18:35:05 +0200
+Subject: [PATCH] Makefile: fix ARCH for TARGET powerpc
+
+TARGET 'ppc' and 'powerpc' shall result in ARCH 'powerpc32'.
+
+Signed-off-by: Alexander Egorenkov <[email protected]>

I'm not sure the best solution here is to patch makedumpfile. Indeed, Buildroot's makedumpfile.mk is passing TARGET=$(BR2_ARCH), but there is nothing that "forces" makedumpfile to understand the BR2_ARCH values.

So instead, I think we should have some logic in makedumpfile.mk that calculates the correct TARGET values depending on the CPU architecture. Something along the lines of:

ifeq ($(BR2_aarch64)$(BR2_aarch64_be),y)
MAKEDUMPFILE_TARGET = ...
else ifeq ($(BR2_arm)$(BR2_armeb),y)
MAKEDUMPFILE_TARGET = ...
else ifeq ...
...
endif

Perhaps if for some architectures the BR2_ARCH value is OK, you could finish this if sequence by:

else
MAKEDUMPFILE_TARGET = $(BR2_ARCH)
endif

and then of course, pass TARGET=$(MAKEDUMPFILE_TARGET)

Thanks!

Thomas

Signature · 3 lines

Thomas Petazzoni, CTO, Bootlin
Embedded Linux and Kernel engineering
https://bootlin.com

Alexander Egorenkov

author

2020-09-10 07:08 UTC #2 lore raw

PATCH 1/1 package/makedumpfile: fix build for powerpc target

Quoted 6 lines

Is there an autobuilder failure for this? If so, there should be a link to it in your commit log. See for example this commit log:
https://git.buildroot.org/buildroot/commit/?id=a85ae8b297fd0526f57fe99d73087dd840ea6219

Fixed.

Quoted 17 lines

I'm not sure the best solution here is to patch makedumpfile. Indeed, Buildroot's makedumpfile.mk is passing TARGET=$(BR2_ARCH), but there is nothing that "forces" makedumpfile to understand the BR2_ARCH values.
So instead, I think we should have some logic in makedumpfile.mk that calculates the correct TARGET values depending on the CPU architecture. Something along the lines of:
ifeq ($(BR2_aarch64)$(BR2_aarch64_be),y)
MAKEDUMPFILE_TARGET = ...
else ifeq ($(BR2_arm)$(BR2_armeb),y)
MAKEDUMPFILE_TARGET = ...
else ifeq ...
...
endif

Quoted 8 lines

Perhaps if for some architectures the BR2_ARCH value is OK, you could finish this if sequence by:
else
MAKEDUMPFILE_TARGET = $(BR2_ARCH)
endif
and then of course, pass TARGET=$(MAKEDUMPFILE_TARGET)

Fixed. I find your suggestion a lot cleaner and like it better.

The new version of patch is on the way.

Thanks for your constructive criticism
Regards
Alex