Skip to contentSkip to conversation

Various Misc tweaks

PATCH In review Richard Purdie 7 participants 18 messages Jul 27, 2011 – Aug 2, 2011

lore
Share & tools
lore
#0 lore raw

cover Various Misc tweaks

The following series is for a few things I've noticed needed fixing as various patches have gone by:

git://git.openembedded.org/openembedded-core-contrib rpurdie/misc2
http://cgit.openembedded.org/cgit.cgi/openembedded-core-contrib/log/?h=rpurdie/misc2

Richard Purdie (4):
base.bbclass: Add compatibility package name mapping handler
gcc: Drop part of the 64bithack patch which is no longer used
module-init-tools-cross: Drop static binaries patch as a better fix
has been merged
base.bbclass: Fix PACKAGE_ARCH typo

 meta/classes/base.bbclass                          |   10 +++++++-
 .../recipes-devtools/gcc/gcc-4.5.1/64bithack.patch |   13 ----------
 meta/recipes-devtools/gcc/gcc-4.6/64bithack.patch  |   13 ----------
 .../files/no-static-binaries.patch                 |   25 --------------------
 .../module-init-tools-cross_3.16.bb                |    2 -
 5 files changed, 9 insertions(+), 54 deletions(-)
 delete mode 100644 meta/recipes-kernel/module-init-tools/files/no-static-binaries.patch
Signature · 1 line
1.7.4.1
#1 lore raw

PATCH 1/4 base.bbclass: Add compatibility package name mapping handler

This means if PKGARCHCOMPAT_ARMV7A is set, "armv7a-vfp-neon" is renamed to be "armv7a". Other compatibility mappings can be added as needed.

---
 meta/classes/base.bbclass |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)
meta/classes/base.bbclass +8 -0
index f12b3cb..3ed1bb8 100644
@@ -133,6 +133,13 @@ def generate_git_config(e):
                 f.write(proxy_command)
                 f.close
 
+def pkgarch_mapping(d):
+    # Compatibility mappings of TUNE_PKGARCH (opt in)
+    if d.getVar("PKGARCHCOMPAT_ARMV7A", True):
+        if d.getVar("TUNE_PKGARCH", True) == "armv7a-vfp-neon":
+            d.setVar("TUNE_PKGARCH", "armv7a")
+
+
 addhandler base_eventhandler
 python base_eventhandler() {
 	from bb import note, error, data
@@ -203,6 +210,7 @@ python base_eventhandler() {
 
         if name == "ConfigParsed":
                 generate_git_config(e)
+                pkgarch_mapping(e.data)
 
 	if not data in e.__dict__:
 		return
Signature · 1 line
1.7.4.1
#2 lore raw

PATCH 2/4 gcc: Drop part of the 64bithack patch which is no longer used

Since we now handle GLIBC_DYNAMIC_LINKER in gcc-configure-common.inc:

's#\(GLIBC_DYNAMIC_LINKER[^ ]*\)\( *"/lib.*\)#\1 SYSTEMLIBS_DIR\2#'

we can drop the patch which changes a hardcoded value for this.

No PR bump since there is no code change resulting from this, its just a cleanup.

There is still a valid question over the remainder of this patch and its interaction with multilib configurations.

---
 .../recipes-devtools/gcc/gcc-4.5.1/64bithack.patch |   13 -------------
 meta/recipes-devtools/gcc/gcc-4.6/64bithack.patch  |   13 -------------
 2 files changed, 0 insertions(+), 26 deletions(-)
meta/recipes-devtools/gcc/gcc-4.5.1/64bithack.patch +0 -13
index 890cc2f..05eafe3 100644
@@ -20,16 +20,3 @@ Index: gcc-4.5.0/gcc/config/i386/t-linux64
  
  LIBGCC = stmp-multilib
  INSTALL_LIBGCC = install-multilib
-Index: gcc-4.5.0/gcc/config/i386/linux64.h
-===================================================================
---- gcc-4.5.0.orig/gcc/config/i386/linux64.h	2010-08-10 15:49:06.000000000 +0100
-+++ gcc-4.5.0/gcc/config/i386/linux64.h	2010-08-10 15:49:16.000000000 +0100
-@@ -59,7 +59,7 @@
-    done.  */
- 
- #define GLIBC_DYNAMIC_LINKER32 "/lib/ld-linux.so.2"
--#define GLIBC_DYNAMIC_LINKER64 "/lib64/ld-linux-x86-64.so.2"
-+#define GLIBC_DYNAMIC_LINKER64 "/lib/ld-linux-x86-64.so.2"
- 
- #if TARGET_64BIT_DEFAULT
- #define SPEC_32 "m32"
meta/recipes-devtools/gcc/gcc-4.6/64bithack.patch +0 -13
index 84962e6..4b846bb 100644
@@ -20,16 +20,3 @@ Index: gcc-4.6.0/gcc/config/i386/t-linux64
  
  LIBGCC = stmp-multilib
  INSTALL_LIBGCC = install-multilib
-Index: gcc-4.6.0/gcc/config/i386/linux64.h
-===================================================================
---- gcc-4.6.0.orig/gcc/config/i386/linux64.h
-+++ gcc-4.6.0/gcc/config/i386/linux64.h
-@@ -63,7 +63,7 @@ see the files COPYING3 and COPYING.RUNTI
-    done.  */
- 
- #define GLIBC_DYNAMIC_LINKER32 "/lib/ld-linux.so.2"
--#define GLIBC_DYNAMIC_LINKER64 "/lib64/ld-linux-x86-64.so.2"
-+#define GLIBC_DYNAMIC_LINKER64 "/lib/ld-linux-x86-64.so.2"
- 
- #if TARGET_64BIT_DEFAULT
- #define SPEC_32 "m32"
Signature · 1 line
1.7.4.1
#3 lore raw

Re: [PATCH 2/4] gcc: Drop part of the 64bithack patch which is no longer used

Richard Purdie wrote on 2011-07-27:

Quoted 4 lines

Since we now handle GLIBC_DYNAMIC_LINKER in gcc-configure-common.inc:

's#\(GLIBC_DYNAMIC_LINKER[^ ]*\)\( *"/lib.*\)#\1 SYSTEMLIBS_DIR\2#'

It seems to me that this patch may not be removed. The above code in gcc-configure-common.inc said it only useful for SDK. For gcc-cross, it doesn't change /lib64/ld-linux-x86-64.so.2 to /lib/ld-linux-x86-64.so.2 as what the 64bithack.patch does.

-Lianhao

Quoted 13 lines

we can drop the patch which changes a hardcoded value for this.

No PR bump since there is no code change resulting from this, its just a cleanup.

There is still a valid question over the remainder of this patch and its interaction with multilib configurations.

---
 .../recipes-devtools/gcc/gcc-4.5.1/64bithack.patch |   13 -------------
 meta/recipes-devtools/gcc/gcc-4.6/64bithack.patch  |   13 -------------
 2 files changed, 0 insertions(+), 26 deletions(-)
#4 lore raw

Re: [PATCH 2/4] gcc: Drop part of the 64bithack patch which is no longer used

Quoted 9 lines · Lianhao

Richard Purdie wrote on 2011-07-27:

Quoted 4 lines

Since we now handle GLIBC_DYNAMIC_LINKER in gcc-configure-common.inc:

's#\(GLIBC_DYNAMIC_LINKER[^ ]*\)\( *"/lib.*\)#\1 SYSTEMLIBS_DIR\2#'

It seems to me that this patch may not be removed. The above code in gcc-configure-common.inc said it only useful for SDK. For gcc-cross, it doesn't change /lib64/ld-linux-x86-64.so.2 to /lib/ld-linux-x86-64.so.2 as what the 64bithack.patch does.

-Lianhao

Agreed, seeing similar issues on ppc64. I think the sed line might not be doing with Richard intended.

We get:

./i386/linux.h:#define GLIBC_DYNAMIC_LINKER SYSTEMLIBS_DIR "/lib/ld-linux.so.2" ./i386/linux64.h:#define GLIBC_DYNAMIC_LINKER32 SYSTEMLIBS_DIR "/lib/ld-linux.so.2" ./i386/linux64.h:#define GLIBC_DYNAMIC_LINKER64 SYSTEMLIBS_DIR "/lib64/ld-linux-x86-64.so.2"

- k

Quoted 21 lines
Quoted 6 lines

we can drop the patch which changes a hardcoded value for this.

No PR bump since there is no code change resulting from this, its just a cleanup.

Quoted 8 lines

There is still a valid question over the remainder of this patch and its interaction with multilib configurations.

---

.../recipes-devtools/gcc/gcc-4.5.1/64bithack.patch |   13 -------------
meta/recipes-devtools/gcc/gcc-4.6/64bithack.patch  |   13 -------------
2 files changed, 0 insertions(+), 26 deletions(-)

_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

#5 lore raw

Re: [PATCH 2/4] gcc: Drop part of the 64bithack patch which is no longer used

In reply to Richard Purdie
Quoted 26 lines · Richard Purdie

Since we now handle GLIBC_DYNAMIC_LINKER in gcc-configure-common.inc:

's#\(GLIBC_DYNAMIC_LINKER[^ ]*\)\( *"/lib.*\)#\1 SYSTEMLIBS_DIR\2#'

we can drop the patch which changes a hardcoded value for this.

No PR bump since there is no code change resulting from this, its just a cleanup.

There is still a valid question over the remainder of this patch and its interaction with multilib configurations.

---
 .../recipes-devtools/gcc/gcc-4.5.1/64bithack.patch |   13 -------------
 meta/recipes-devtools/gcc/gcc-4.6/64bithack.patch  |   13 -------------
 2 files changed, 0 insertions(+), 26 deletions(-)
meta/recipes-devtools/gcc/gcc-4.5.1/64bithack.patch +0 -0
index 890cc2f..05eafe3 100644
@@ -20,16 +20,3 @@ Index: gcc-4.5.0/gcc/config/i386/t-linux64
 
  LIBGCC = stmp-multilib
  INSTALL_LIBGCC = install-multilib

meta/recipes-devtools/gcc/gcc-4.5.1/64bithack.patch Index: gcc-4.5.0/gcc/config/i386/t-linux64 · line 32

Quoted 10 lines
meta/recipes-devtools/gcc/gcc-4.5.1/64bithack.patch +0 -10
-Index: gcc-4.5.0/gcc/config/i386/linux64.h
-===================================================================
---- gcc-4.5.0.orig/gcc/config/i386/linux64.h	2010-08-10 15:49:06.000000000 +0100
-+++ gcc-4.5.0/gcc/config/i386/linux64.h	2010-08-10 15:49:16.000000000 +0100
-@@ -59,7 +59,7 @@
-    done.  */
- 
- #define GLIBC_DYNAMIC_LINKER32 "/lib/ld-linux.so.2"
--#define GLIBC_DYNAMIC_LINKER64 "/lib64/ld-linux-x86-64.so.2"
-+#define GLIBC_DYNAMIC_LINKER64 "/lib/ld-linux-x86-64.so.2"

This change while I like it that we are moving to use default will now emit /lib64/ld-linux-x86-64.so.2 into INTERP section but I dont see a corresponding patch to fix root file systems where the defualt libdir for x86-64 is changed to be lib64

So this will end up in userspace not booting at all since init wont find the dynamic linker to start with.

Thanks

-Khem

#6 lore raw

PATCH 3/4 module-init-tools-cross: Drop static binaries patch as a better fix has been merged

In reply to Richard Purdie
---
 .../files/no-static-binaries.patch                 |   25 --------------------
 .../module-init-tools-cross_3.16.bb                |    2 -
 2 files changed, 0 insertions(+), 27 deletions(-)
 delete mode 100644 meta/recipes-kernel/module-init-tools/files/no-static-binaries.patch
meta/recipes-kernel/module-init-tools/files/no-static-binaries.patch deleted +0 -25
deleted file mode 100644
index 33fb6e0..0000000
@@ -1,25 +0,0 @@
-# Building a native, statically-linked insmod can result in ld failing because
-# it can't find a static libc.  A native, static insmod is unnecessary anyway,
-# so this patch just removes it.
-#
-# - [email protected]
-#
-# Change to a simpler method by just disabling static link flag, since even
-# removing those lines don't prevent insmod.static from compile
-#
-# by Kevin Tian <[email protected]>, 2010-07-21
-
-Upstream-Status: Pending
-
-Index: module-init-tools-3.2.2/Makefile.am
-===================================================================
---- module-init-tools-3.2.2.orig/Makefile.am	2010-05-26 15:43:16.000000000 -0400
-+++ module-init-tools-3.2.2/Makefile.am	2010-05-26 16:04:06.232990785 -0400
-@@ -10,7 +10,6 @@
- modindex_SOURCES = modindex.c zlibsupport.c zlibsupport.h testing.h
- 
- insmod_static_SOURCES = insmod.c
--insmod_static_LDFLAGS = -static
- # We don't want the $(zlib_flags) here: that makes a dynamic executable
- insmod_static_LDADD = libmodtools.a
- 
meta/recipes-kernel/module-init-tools/module-init-tools-cross_3.16.bb +0 -2
index 38f0476..a6c5d21 100644
@@ -3,8 +3,6 @@ PR = "r0"
 inherit cross
 PROVIDES += "virtual/${TARGET_PREFIX}depmod"
 
-SRC_URI += "file://no-static-binaries.patch"
-
 SRC_URI[md5sum] = "bc44832c6e41707b8447e2847d2019f5"
 SRC_URI[sha256sum] = "e1f2cdcae64a8effc25e545a5e0bdaf312f816ebbcd0916e4e87450755fab64b"
 
Signature · 1 line
1.7.4.1
#7 lore raw

PATCH 4/4 base.bbclass: Fix PACKAGE_ARCH typo

In reply to Richard Purdie
---
 meta/classes/base.bbclass |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
meta/classes/base.bbclass +1 -1
index 3ed1bb8..a0a189f 100644
@@ -407,7 +407,7 @@ python () {
         # if multiple differences are present?
         # Look through PACKAGE_ARCHS for the priority order?
         if pkgarch and pkgarch == mach_arch:
-            bb.data.setVar('PACAKGE_ARCH', "${MACHINE_ARCH}", d)
+            bb.data.setVar('PACKAGE_ARCH', "${MACHINE_ARCH}", d)
             bb.warn("Recipe %s is marked as only being architecture specific but seems to have machine specific packages?! The recipe may as well mark itself as machine specific directly." % d.getVar("PN", True))
 }
 
Signature · 1 line
1.7.4.1
#8 lore raw

Re: [PATCH 4/4] base.bbclass: Fix PACKAGE_ARCH typo

Quoted 10 lines · Richard Purdie
---
 meta/classes/base.bbclass |    2 +-
 1 files changed, 1 insertions(+), 1 deletions(-)
meta/classes/base.bbclass +0 -0
index 3ed1bb8..a0a189f 100644

meta/classes/base.bbclass python () { · line 410

Quoted 5 lines
meta/classes/base.bbclass +1 -1
@@ -407,7 +407,7 @@ python () {
         # if multiple differences are present?
         # Look through PACKAGE_ARCHS for the priority order?
         if pkgarch and pkgarch == mach_arch:
-            bb.data.setVar('PACAKGE_ARCH', "${MACHINE_ARCH}", d)
+            bb.data.setVar('PACKAGE_ARCH', "${MACHINE_ARCH}", d)

nice I was wondering this warning in my builds

Quoted 3 lines
meta/classes/base.bbclass +0 -0
             bb.warn("Recipe %s is marked as only being architecture

specific but seems to have machine specific packages?! The recipe may as well mark itself as machine specific directly." % d.getVar("PN", True)) }

Signature · 1 line
Khem Raj
#9 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

In reply to Richard Purdie
Quoted 2 lines · Richard Purdie

This means if PKGARCHCOMPAT_ARMV7A is set, "armv7a-vfp-neon" is renamed to be "armv7a". Other compatibility mappings can be added as needed.

There are multiple armv7 cores without neon... I think there might even be one or two custom cores w/o VFP. (Yes I know this violates the core spec from ARM, but people do that all of the time.)

--Mark

Quoted 31 lines
---
 meta/classes/base.bbclass |    8 ++++++++
 1 files changed, 8 insertions(+), 0 deletions(-)
meta/classes/base.bbclass +8 -0
index f12b3cb..3ed1bb8 100644
@@ -133,6 +133,13 @@ def generate_git_config(e):
                 f.write(proxy_command)
                 f.close
 
+def pkgarch_mapping(d):
+    # Compatibility mappings of TUNE_PKGARCH (opt in)
+    if d.getVar("PKGARCHCOMPAT_ARMV7A", True):
+        if d.getVar("TUNE_PKGARCH", True) == "armv7a-vfp-neon":
+            d.setVar("TUNE_PKGARCH", "armv7a")
+
+
 addhandler base_eventhandler
 python base_eventhandler() {
 	from bb import note, error, data
@@ -203,6 +210,7 @@ python base_eventhandler() {
 
         if name == "ConfigParsed":
                 generate_git_config(e)
+                pkgarch_mapping(e.data)
 
 	if not data in e.__dict__:
 		return
#10 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

Quoted 7 lines · Mark Hatle
Quoted 2 lines · Richard Purdie

This means if PKGARCHCOMPAT_ARMV7A is set, "armv7a-vfp-neon" is renamed to be "armv7a". Other compatibility mappings can be added as needed.

There are multiple armv7 cores without neon... I think there might even be one or two custom cores w/o VFP. (Yes I know this violates the core spec from ARM, but people do that all of the time.)

I know there are but for some distros in OE have established "armv7a" with a specific meaning. This gives us some way of merging the tune code yet defer fixing those conflicts which can then happen at a time of their choosing.

Its assumed anyone enabling the option knows what it means (which Angstrom does and is where the request came from).

FWIW I have just merged the tune changes and this workaround flag.

Cheers,

Richard

#11 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

Op 27 jul. 2011, om 16:50 heeft Richard Purdie het volgende geschreven:

Quoted 14 lines · Mark Hatle
Quoted 7 lines · Richard Purdie
Quoted 2 lines

This means if PKGARCHCOMPAT_ARMV7A is set, "armv7a-vfp-neon" is renamed to be "armv7a". Other compatibility mappings can be added as needed.

There are multiple armv7 cores without neon... I think there might even be one or two custom cores w/o VFP. (Yes I know this violates the core spec from ARM, but people do that all of the time.)

I know there are but for some distros in OE have established "armv7a" with a specific meaning. This gives us some way of merging the tune code yet defer fixing those conflicts which can then happen at a time of their choosing.

Quoted 4 lines

Its assumed anyone enabling the option knows what it means (which Angstrom does and is where the request came from).

FWIW I have just merged the tune changes and this workaround flag.

I think we might need a armv5e -> armv5te one and a slightly more tricky armv4 -> armv4t.

The armv4t without thumb case is interesting since we do want to use the EABI bx,lr but not other thumb instructions. My main reason for avoiding t1 is that the toolchain is perpetually broken for it. If gcc fixes a bug binutils will introduce a new one.

regards,

Koen

#12 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

In reply to Mark Hatle
Quoted 7 lines · Mark Hatle
Quoted 2 lines · Richard Purdie

This means if PKGARCHCOMPAT_ARMV7A is set, "armv7a-vfp-neon" is renamed to be "armv7a". Other compatibility mappings can be added as needed.

There are multiple armv7 cores without neon... I think there might even be one or two custom cores w/o VFP. (Yes I know this violates the core spec from ARM, but people do that all of the time.)

Which spec are you thinking of? As far as I know, VFP has always been optional for both ARM11 and Cortex.

p.

#13 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

Quoted 11 lines · Phil Blundell
Quoted 7 lines · Mark Hatle
Quoted 2 lines · Richard Purdie

This means if PKGARCHCOMPAT_ARMV7A is set, "armv7a-vfp-neon" is renamed to be "armv7a". Other compatibility mappings can be added as needed.

There are multiple armv7 cores without neon... I think there might even be one or two custom cores w/o VFP. (Yes I know this violates the core spec from ARM, but people do that all of the time.)

Which spec are you thinking of? As far as I know, VFP has always been optional for both ARM11 and Cortex.

I thought VFP was part of the instruction set as of ARMv7. Neon is recommended but options. As I mentioned though, I know of chips in design that don't have Neon, and I suspect also don't have VFP -- but they're ARMv7.. (they also have stripped out the thumb2 part as well..)

--Mark

Quoted 8 lines

p.

_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

#14 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

In reply to Phil Blundell

Op 27 jul. 2011, om 17:02 heeft Phil Blundell het volgende geschreven:

Quoted 11 lines · Mark Hatle
Quoted 7 lines · Richard Purdie
Quoted 2 lines

This means if PKGARCHCOMPAT_ARMV7A is set, "armv7a-vfp-neon" is renamed to be "armv7a". Other compatibility mappings can be added as needed.

There are multiple armv7 cores without neon... I think there might even be one or two custom cores w/o VFP. (Yes I know this violates the core spec from ARM, but people do that all of the time.)

Which spec are you thinking of? As far as I know, VFP has always been optional for both ARM11 and Cortex.

From the conversations I had with ARM they implied that it isn't optional for cortex a8/9/15, only neon is (e.g. marvell dove, nvidia tegra2). I haven't seen any black on white contracts saying that, though :)

regards,

Koen

#15 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

Quoted 1 line · Koen Kooi

From the conversations I had with ARM they implied that it isn't optional for cortex a8/9/15, only neon is (e.g. marvell dove, nvidia tegra2). I haven't seen any black on white contracts saying that, though :)

The spec on the web seems to say fairly clearly that it's optional, at least in theory. I guess they might have commercial reasons for not wanting to sell it without VFP in any given situation.

See:

http://www.arm.com/products/processors/cortex-a/cortex-a5.php?tab=Specifications

for example

p.

#16 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

Quoted 9 lines · Phil Blundell
Quoted 4 lines · Koen Kooi

From the conversations I had with ARM they implied that it isn't optional for cortex a8/9/15, only neon is (e.g. marvell dove, nvidia tegra2). I haven't seen any black on white contracts saying that, though :)

The spec on the web seems to say fairly clearly that it's optional, at least in theory. I guess they might have commercial reasons for not wanting to sell it without VFP in any given situation.

Quoted 6 lines

See:

http://www.arm.com/products/processors/cortex-a/cortex-a5.php?tab=Specificat
ions

for example

thats only for for a5. For procesors koen mentioned VFP is mandatory

Quoted 9 lines

p.

_______________________________________________
Openembedded-core mailing list
[email protected]
http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-core

Signature · 1 line
Khem Raj
#17 lore raw

Re: [PATCH 1/4] base.bbclass: Add compatibility package name mapping handler

Quoted 10 lines · Khem Raj
Quoted 9 lines · Phil Blundell
Quoted 4 lines · Koen Kooi

From the conversations I had with ARM they implied that it isn't optional for cortex a8/9/15, only neon is (e.g. marvell dove, nvidia tegra2). I haven't seen any black on white contracts saying that, though :)

The spec on the web seems to say fairly clearly that it's optional, at least in theory. I guess they might have commercial reasons for not wanting to sell it without VFP in any given situation.

Quoted 8 lines
Quoted 6 lines

See:

http://www.arm.com/products/processors/cortex-a/cortex-a5.php?tab=Specificat
ions

for example

thats only for for a5. For procesors koen mentioned VFP is mandatory

Oh yeah, you're right. Sorry, I didn't read his email carefully enough.

The original question, though, was whether VFP is mandatory for the ARMv7-A architecture. Irrespective of what the situation is with A8/A9/A15, Cortex-A5 is still an ARMv7-A processor and hence VFP can't be mandatory for the architecture as a whole.

p.