Skip to contentSkip to conversation

Build results for 2013-11-22

Buildroot New Thomas Petazzoni 6 participants 24 messages Nov 23, 2013 – Nov 28, 2013

lore
Share & tools
lore
#0 lore raw

Build statistics for 2013-11-22
===============================

success : 68
failures : 34
timeouts : 0
TOTAL : 102

Classification of failures by reason
====================================

              e2fsprogs-1.42.8 | 2 
               libglib2-2.36.3 | 2 
                     gmp-5.1.3 | 2 
                 libnspr-4.9.6 | 2 
                      icu-51.2 | 2 
                  libv4l-0.8.9 | 2 
               alsa-lib-1.0.26 | 2 
ne10-88c18f02199947b2c8b577... | 1 
                      qt-4.8.5 | 1 
                 mplayer-1.1.1 | 1 
           imagemagick-6.8.7-5 | 1 
                     lxc-0.9.0 | 1 
                 schifra-0.0.1 | 1 
libwebsockets-v1.22-chrome2... | 1 
                     rpm-5.2.0 | 1 
                directfb-1.6.3 | 1 
            libvncserver-0.9.9 | 1 
                    systemd-44 | 1 
                     nut-2.6.5 | 1 
              fontconfig-2.6.0 | 1 
                  poco-1.4.6p1 | 1 
                valgrind-3.8.1 | 1 
                openssl-1.0.1e | 1 
                      gpsd-3.9 | 1 
                  udisks-1.0.4 | 1 
           argp-standalone-1.3 | 1 
                libiscsi-1.6.0 | 1 

Detail of failures
===================

      bfin |                alsa-lib-1.0.26 | NOK | http://autobuild.buildroot.net/results/65b2966853297b16cdc472149dae6b7ed4603f03/
    mipsel |                alsa-lib-1.0.26 | NOK | http://autobuild.buildroot.net/results/51ad394266c07f386349a5df94df1c1f9fab2722/
    mipsel |            argp-standalone-1.3 | NOK | http://autobuild.buildroot.net/results/db81f4cf5424a1c125dd94e7ae1f7baef9d08c30/
    x86_64 |                 directfb-1.6.3 | NOK | http://autobuild.buildroot.net/results/2d6573eb6e93e3659422658c247ce64745b51221/
      bfin |               e2fsprogs-1.42.8 | NOK | http://autobuild.buildroot.net/results/5fec58000e36062da4a618935c72620a65214ed3/
      bfin |               e2fsprogs-1.42.8 | NOK | http://autobuild.buildroot.net/results/2d819fc65a5bfc698be4050203a95f6bb64c4e9c/
      bfin |               fontconfig-2.6.0 | NOK | http://autobuild.buildroot.net/results/41d6cecf13aa23ffa6ef6eac8a8a5d91cb78eaf6/
    mipsel |                      gmp-5.1.3 | NOK | http://autobuild.buildroot.net/results/36d354880acd04a3e5e3d366a654d005be92f397/
    mipsel |                      gmp-5.1.3 | NOK | http://autobuild.buildroot.net/results/c8c5f58e8c0b7629599a979ed7777b94f5d28eee/
       arc |                       gpsd-3.9 | NOK | http://autobuild.buildroot.net/results/d3dd298b9ada76e9a2f15fe7803312c57e8efec1/
       arm |                       icu-51.2 | NOK | http://autobuild.buildroot.net/results/5ce9baf6085036eed16429a1eead6e723effe23f/
      bfin |                       icu-51.2 | NOK | http://autobuild.buildroot.net/results/e720e911fc2d989db4799e5cbe6809995dab25b2/
   powerpc |            imagemagick-6.8.7-5 | NOK | http://autobuild.buildroot.net/results/d4665b1e16df8f26b573311cac77d43aca97a315/
      bfin |                libglib2-2.36.3 | NOK | http://autobuild.buildroot.net/results/e7439cd27e28129292bfabfdc48b0749d288a777/
      bfin |                libglib2-2.36.3 | NOK | http://autobuild.buildroot.net/results/74eb1cec76ee44d3fd1d656183dafa4c35840ae4/
  mips64el |                 libiscsi-1.6.0 | NOK | http://autobuild.buildroot.net/results/d0889cd7b908b5a9c7558a63b0ee0ca079bee29d/
     nios2 |                  libnspr-4.9.6 | NOK | http://autobuild.buildroot.net/results/690d25402b0715bde1edd2feade1a8b9187d3acd/
    xtensa |                  libnspr-4.9.6 | NOK | http://autobuild.buildroot.net/results/14ba6426edbffa100de924aa69157b3f59368ff2/
       arc |                   libv4l-0.8.9 | NOK | http://autobuild.buildroot.net/results/a177d552d36a8127b90b30086b260ef232ce7c55/
   aarch64 |                   libv4l-0.8.9 | NOK | http://autobuild.buildroot.net/results/c31d99de5441133ddcd77a5f0e2350c228cf5b96/
       arm |             libvncserver-0.9.9 | NOK | http://autobuild.buildroot.net/results/94192fe9e795286afc053c65e3bad2993ba2da26/
      bfin | libwebsockets-v1.22-chrome2... | NOK | http://autobuild.buildroot.net/results/63331f0fe53bc77320d31c3d3921ecc73b2f7a9d/
     nios2 |                      lxc-0.9.0 | NOK | http://autobuild.buildroot.net/results/942c1dd275ce153b7f94d2359fb0e71a9a0f75dd/
       arm |                  mplayer-1.1.1 | NOK | http://autobuild.buildroot.net/results/f38611e49ced60519f3c0b1177426f344430ca6e/
       arm | ne10-88c18f02199947b2c8b577... | NOK | http://autobuild.buildroot.net/results/947871ee21f2baa86b9f8809e6e1789b48c55871/
   powerpc |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/b6f619f9c249fc653dc65f0f516cb846683700bb/
microblaze |                 openssl-1.0.1e | NOK | http://autobuild.buildroot.net/results/3ab8ad3ea288a3b8f4e3a4ddf60764432880b287/
      sh4a |                   poco-1.4.6p1 | NOK | http://autobuild.buildroot.net/results/8dc9bca5969d455cc30206fc1ae6e1d3ba6eb6ca/
       arm |                       qt-4.8.5 | NOK | http://autobuild.buildroot.net/results/c9cca5e31b9eac01ec212dca9d486a6c639aab14/
      mips |                      rpm-5.2.0 | NOK | http://autobuild.buildroot.net/results/343579619909559b9b2bc6eafe4fb082a63fa048/
      i686 |                  schifra-0.0.1 | NOK | http://autobuild.buildroot.net/results/424fc76a7e439e274e1cc66b09486fad8da05f8c/
       arm |                     systemd-44 | NOK | http://autobuild.buildroot.net/results/e5484bcdcca7f2cb48913d65271de4b957ea3450/
       arm |                   udisks-1.0.4 | NOK | http://autobuild.buildroot.net/results/ea4627cae45e972ebba5b33a2b2871ce7f46fedc/
       arm |                 valgrind-3.8.1 | NOK | http://autobuild.buildroot.net/results/32dc58dd4219b5b74c7e623d53a1fe8034fdf31b/
Signature · 1 line
http://autobuild.buildroot.net
#1 lore raw

Hi,

Quoted 8 lines · Thomas Petazzoni

Build statistics for 2013-11-22
===============================

success : 68
failures : 34
timeouts : 0
TOTAL : 102

[..]

Quoted 1 line
      sh4a |                   poco-1.4.6p1 | NOK | http://autobuild.buildroot.net/results/8dc9bca5969d455cc30206fc1ae6e1d3ba6eb6ca/

Error is:

In file included from src/FPEnvironment_C99.cpp:37:0,
from src/FPEnvironment.cpp:48:
include/Poco/FPEnvironment_C99.h:56:30: error: 'FE_DOWNWARD' was not
declared in this scope
include/Poco/FPEnvironment_C99.h:57:30: error: 'FE_UPWARD' was not
declared in this scope
make[2]: *** [/home/test/test/2/output/build/poco-1.4.6p1/Foundation/obj/Linux/sh4a/release_shared/FPEnvironment.o] Error 1
make[2]: Leaving directory
`/home/test/test/2/output/build/poco-1.4.6p1/Foundation'
make[1]: *** [Foundation-libexec] Error 2
make[1]: Leaving directory `/home/test/test/2/output/build/poco-1.4.6p1'
make: *** [/home/test/test/2/output/build/poco-1.4.6p1/.stamp_built] Error 2

The same build failure has been discussed a long time ago here:
http://lists.busybox.net/pipermail/buildroot/2011-September/046013.html
when the poco package was introduced. Then, the fpenvironment was disabled when the C library is uclibc.
In the current failure, the C library seems to be glibc. I'm including Baruch on this one, who was involved in the above thread. Baruch: any idea?

Quoted 2 lines
       arm |                     systemd-44 | NOK | http://autobuild.buildroot.net/results/e5484bcdcca7f2cb48913d65271de4b957ea3450/
       arm |                   udisks-1.0.4 | NOK | http://autobuild.buildroot.net/results/ea4627cae45e972ebba5b33a2b2871ce7f46fedc/

I'm looking at the above two ones...

[..]

Best regards,
Thomas

#2 lore raw
Quoted 41 lines · Thomas De Schampheleire

Hi,

Quoted 8 lines · Thomas Petazzoni

Build statistics for 2013-11-22
===============================

success : 68
failures : 34
timeouts : 0
TOTAL : 102

[..]

Quoted 1 line
      sh4a |                   poco-1.4.6p1 | NOK | http://autobuild.buildroot.net/results/8dc9bca5969d455cc30206fc1ae6e1d3ba6eb6ca/

Error is:

In file included from src/FPEnvironment_C99.cpp:37:0,
from src/FPEnvironment.cpp:48:
include/Poco/FPEnvironment_C99.h:56:30: error: 'FE_DOWNWARD' was not
declared in this scope
include/Poco/FPEnvironment_C99.h:57:30: error: 'FE_UPWARD' was not
declared in this scope
make[2]: *** [/home/test/test/2/output/build/poco-1.4.6p1/Foundation/obj/Linux/sh4a/release_shared/FPEnvironment.o] Error 1
make[2]: Leaving directory
`/home/test/test/2/output/build/poco-1.4.6p1/Foundation'
make[1]: *** [Foundation-libexec] Error 2
make[1]: Leaving directory `/home/test/test/2/output/build/poco-1.4.6p1'
make: *** [/home/test/test/2/output/build/poco-1.4.6p1/.stamp_built] Error 2

The same build failure has been discussed a long time ago here:
http://lists.busybox.net/pipermail/buildroot/2011-September/046013.html
when the poco package was introduced. Then, the fpenvironment was disabled when the C library is uclibc.
In the current failure, the C library seems to be glibc. I'm including Baruch on this one, who was involved in the above thread. Baruch: any idea?

Quoted 4 lines
Quoted 2 lines
       arm |                     systemd-44 | NOK | http://autobuild.buildroot.net/results/e5484bcdcca7f2cb48913d65271de4b957ea3450/
       arm |                   udisks-1.0.4 | NOK | http://autobuild.buildroot.net/results/ea4627cae45e972ebba5b33a2b2871ce7f46fedc/

I'm looking at the above two ones...

The udisks problem is the following: the configuration requests lvm2 support in udisks, which causes:
select BR2_PACKAGE_LVM2_APP_LIBRARY
The configuration also has BR2_PACKAGE_LVM2_DMSETUP_ONLY selected. The problem is that both are not supposed to be configured together:

config BR2_PACKAGE_LVM2_DMSETUP_ONLY
bool "install dmsetup only"
depends on BR2_PACKAGE_LVM2
help
Install dmsetup only and skip the LVM2 suite.

config BR2_PACKAGE_LVM2_APP_LIBRARY
bool "install application library" depends on BR2_PACKAGE_LVM2 depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY help
Install application library (liblvm2app)

In the configuration menu, it seems you can indeed create such an invalid configuration: start with LVM2_DMSETUP_ONLY selected, then enable udisks with lvm2 support. If you save, you get a .config that has:

$ grep LVM .config
BR2_PACKAGE_LVM2=y
BR2_PACKAGE_LVM2_DMSETUP_ONLY=y
BR2_PACKAGE_LVM2_APP_LIBRARY=y
BR2_PACKAGE_UDISKS_LVM2=y

Because DMSETUP_ONLY is setup, the install command does not install the lvm2app, which is needed for udisks.

This could be solved by adding 'depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY' to the option BR2_PACKAGE_UDISKS_LVM2, so that you cannot enable lvm2 support in udisks if DMSETUP_ONLY is selected.
However, is there a better solution? I tried 'select !BR2_PACKAGE_LVM2_DMSETUP_ONLY' to explicitly turn off DMSETUP_ONLY when lvm2 support is requested, but this seems invalid.

Thanks for your input,
Thomas

#3 lore raw

Dear Thomas De Schampheleire,

Quoted 33 lines · Thomas De Schampheleire

The udisks problem is the following: the configuration requests lvm2 support in udisks, which causes:
select BR2_PACKAGE_LVM2_APP_LIBRARY
The configuration also has BR2_PACKAGE_LVM2_DMSETUP_ONLY selected. The problem is that both are not supposed to be configured together:

config BR2_PACKAGE_LVM2_DMSETUP_ONLY
bool "install dmsetup only"
depends on BR2_PACKAGE_LVM2
help
Install dmsetup only and skip the LVM2 suite.

config BR2_PACKAGE_LVM2_APP_LIBRARY
bool "install application library" depends on BR2_PACKAGE_LVM2 depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY help
Install application library (liblvm2app)

In the configuration menu, it seems you can indeed create such an invalid configuration: start with LVM2_DMSETUP_ONLY selected, then enable udisks with lvm2 support. If you save, you get a .config that has:

$ grep LVM .config
BR2_PACKAGE_LVM2=y
BR2_PACKAGE_LVM2_DMSETUP_ONLY=y
BR2_PACKAGE_LVM2_APP_LIBRARY=y
BR2_PACKAGE_UDISKS_LVM2=y

Because DMSETUP_ONLY is setup, the install command does not install the lvm2app, which is needed for udisks.

Quoted 7 lines

This could be solved by adding 'depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY' to the option BR2_PACKAGE_UDISKS_LVM2, so that you cannot enable lvm2 support in udisks if DMSETUP_ONLY is selected.
However, is there a better solution? I tried 'select !BR2_PACKAGE_LVM2_DMSETUP_ONLY' to explicitly turn off DMSETUP_ONLY when lvm2 support is requested, but this seems invalid.

Thanks for the investigation!

I unfortunately don't have a good knowledge of the various components inside LVM2. However, what surprises me here is to have an option called 'install this thing *only*'. Options should only install something *more* but not exclude other options. So I believe the right way of solving this problem is to remove the "only" aspect of this option, and rethink completely the set of sub-options offered for the lvm2 package. People wanting to install only dmsetup will just have to enable only one specific sub-option, and that's it. But it shouldn't prevent another package from requesting the library or other applications to be installed.

Best regards,

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#4 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

BR2_PACKAGE_LVM2_DMSETUP_ONLY is a "negative" option that disables build and install of part of the suite. This option cannot be unselected by other config options, which gives a problem for BR2_PACKAGE_UDISKS_LVM2:
it needs BR2_PACKAGE_LVM2_APP_LIBRARY, which requires the full suite.

Therefore, replace the negative BR2_PACKAGE_LVM2_DMSETUP_ONLY by a positive BR2_PACKAGE_LVM2_FULL_SUITE. To make sure that existing defconfigs keep working, the new option defaults to y unless the legacy BR2_PACKAGE_LVM2_DMSETUP_ONLY was selected.

Fixes http://autobuild.buildroot.net/results/ea4627cae45e972ebba5b33a2b2871ce7f46fedc/

---
Note that in the legacy handling I've assumed that this will still be taken in 2013.11.

---
 Config.in.legacy       |  9 +++++++++
 package/lvm2/Config.in | 10 ++++++----
 package/lvm2/lvm2.mk   |  2 +-
 3 files changed, 16 insertions(+), 5 deletions(-)
Config.in.legacy +9 -0
index 587afe1..6f9050a 100644
@@ -101,6 +101,15 @@ endif
 ###############################################################################
 comment "Legacy options removed in 2013.11"
 
+config BR2_PACKAGE_LVM2_DMSETUP_ONLY
+	bool "lvm2's 'dmsetup only' option removed"
+	select BR2_LEGACY
+	help
+	  The BR2_PACKAGE_LVM2_DMSETUP_ONLY was a negative option, which
+	  led to problems with other packages that need the full lvm2
+	  suite. Therefore, the option has been replaced with the positive
+	  BR2_PACKAGE_LVM2_FULL_SUITE option.
+
 config BR2_PACKAGE_QT_JAVASCRIPTCORE
 	bool "qt javascriptcore option removed"
 	select BR2_LEGACY
package/lvm2/Config.in +6 -4
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
 	  user-space. This makes it useful for not only LVM, but EVMS,
 	  software raid, and other drivers that create "virtual" block devices.
 
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-	bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+	bool "install the full LVM2 suite"
+	default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
 	depends on BR2_PACKAGE_LVM2
 	help
-	  Install dmsetup only and skip the LVM2 suite.
+	  Install the full LVM2 suite. When this option is not set, only
+	  dmsetup is installed.
 
 config BR2_PACKAGE_LVM2_APP_LIBRARY
 	bool "install application library"
 	depends on BR2_PACKAGE_LVM2
-	depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+	select BR2_PACKAGE_LVM2_FULL_SUITE
 	help
 	  Install application library (liblvm2app).
 
package/lvm2/lvm2.mk +1 -1
index 79a0d05..698a212 100644
@@ -38,7 +38,7 @@ else
 LVM2_CONF_OPT += --disable-readline
 endif
 
-ifeq ($(BR2_PACKAGE_LVM2_DMSETUP_ONLY),y)
+ifeq ($(BR2_PACKAGE_LVM2_FULL_SUITE),)
 LVM2_MAKE_OPT = device-mapper
 LVM2_INSTALL_STAGING_OPT = DESTDIR=$(STAGING_DIR) install_device-mapper
 LVM2_INSTALL_TARGET_OPT = DESTDIR=$(TARGET_DIR) install_device-mapper
Signature · 1 line
1.8.4.3
#5 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Hi Arnout,

Quoted 31 lines · Arnout Vandecappelle (Essensium/Mind)

BR2_PACKAGE_LVM2_DMSETUP_ONLY is a "negative" option that disables build and install of part of the suite. This option cannot be unselected by other config options, which gives a problem for BR2_PACKAGE_UDISKS_LVM2:
it needs BR2_PACKAGE_LVM2_APP_LIBRARY, which requires the full suite.

Therefore, replace the negative BR2_PACKAGE_LVM2_DMSETUP_ONLY by a positive BR2_PACKAGE_LVM2_FULL_SUITE. To make sure that existing defconfigs keep working, the new option defaults to y unless the legacy BR2_PACKAGE_LVM2_DMSETUP_ONLY was selected.

Fixes http://autobuild.buildroot.net/results/ea4627cae45e972ebba5b33a2b2871ce7f46fedc/

---
Note that in the legacy handling I've assumed that this will still be taken in 2013.11.

---
 Config.in.legacy       |  9 +++++++++
 package/lvm2/Config.in | 10 ++++++----
 package/lvm2/lvm2.mk   |  2 +-
 3 files changed, 16 insertions(+), 5 deletions(-)
Config.in.legacy +0 -0
index 587afe1..6f9050a 100644
@@ -101,6 +101,15 @@ endif
 ###############################################################################
 comment "Legacy options removed in 2013.11"

Config.in.legacy endif · line 112

Quoted 10 lines
Config.in.legacy +9 -0
 
+config BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       bool "lvm2's 'dmsetup only' option removed"
+       select BR2_LEGACY
+       help
+         The BR2_PACKAGE_LVM2_DMSETUP_ONLY was a negative option, which
+         led to problems with other packages that need the full lvm2
+         suite. Therefore, the option has been replaced with the positive
+         BR2_PACKAGE_LVM2_FULL_SUITE option.
+

There should be a comment here that refers to the remaining legacy reference in package/lvm2/Config.in, see
comment at BR2_TARGET_UBOOT_CUSTOM_GIT_REPO_URL_WRAP.

Quoted 19 lines
Config.in.legacy +0 -0
 config BR2_PACKAGE_QT_JAVASCRIPTCORE
        bool "qt javascriptcore option removed"
        select BR2_LEGACY
package/lvm2/Config.in +3 -3
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
          user-space. This makes it useful for not only LVM, but EVMS,
          software raid, and other drivers that create "virtual" block devices.
 
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-       bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+       bool "install the full LVM2 suite"
+       default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
        depends on BR2_PACKAGE_LVM2
        help
-         Install dmsetup only and skip the LVM2 suite.

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 32

Quoted 10 lines
package/lvm2/Config.in +3 -1
+         Install the full LVM2 suite. When this option is not set, only
+         dmsetup is installed.
 
 config BR2_PACKAGE_LVM2_APP_LIBRARY
        bool "install application library"
        depends on BR2_PACKAGE_LVM2
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_FULL_SUITE
        help
          Install application library (liblvm2app).

It seems odd to me that the 'full suite' does not include the app
library: it is not 'full'.
It would have been nice if we could have options for the different components, but at first sight there seem to be quite a number of components in lvm2: (rough estimate below)

 $ find -name Makefile.in
./lib/thin/Makefile.in
./lib/format_pool/Makefile.in
./lib/replicator/Makefile.in
./lib/snapshot/Makefile.in
./lib/mirror/Makefile.in
./lib/format1/Makefile.in
./lib/Makefile.in
./lib/locking/Makefile.in
./lib/raid/Makefile.in
./python/Makefile.in
./libdaemon/client/Makefile.in
./libdaemon/Makefile.in
./libdaemon/server/Makefile.in
./test/unit/Makefile.in
./test/api/Makefile.in
./test/Makefile.in
./daemons/cmirrord/Makefile.in
./daemons/Makefile.in
./daemons/dmeventd/plugins/thin/Makefile.in
./daemons/dmeventd/plugins/snapshot/Makefile.in
./daemons/dmeventd/plugins/mirror/Makefile.in
./daemons/dmeventd/plugins/lvm2/Makefile.in
./daemons/dmeventd/plugins/Makefile.in
./daemons/dmeventd/plugins/raid/Makefile.in
./daemons/dmeventd/Makefile.in
./daemons/clvmd/Makefile.in
./daemons/lvmetad/Makefile.in
./tools/Makefile.in
./include/Makefile.in
./man/Makefile.in
./scripts/Makefile.in
./po/Makefile.in
./unit-tests/mm/Makefile.in
./unit-tests/datastruct/Makefile.in
./unit-tests/regex/Makefile.in
./Makefile.in
./conf/Makefile.in
./libdm/Makefile.in
./udev/Makefile.in
./liblvm/Makefile.in

Best regards,
Thomas

#6 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Quoted 4 lines · Thomas De Schampheleire

Hi Arnout,

On Sun, Nov 24, 2013 at 12:50 AM, Arnout Vandecappelle (Essensium/Mind) <[email protected]> wrote:

[snip]

Quoted 5 lines
Quoted 5 lines
Config.in.legacy +5 -0
+config BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       bool "lvm2's 'dmsetup only' option removed"
+       select BR2_LEGACY
+       help
+         The BR2_PACKAGE_LVM2_DMSETUP_ONLY was a negative option, which
Quoted 8 lines

Config.in.legacy endif · line 112

Quoted 4 lines
Config.in.legacy +4 -0
+         led to problems with other packages that need the full lvm2
+         suite. Therefore, the option has been replaced with the positive
+         BR2_PACKAGE_LVM2_FULL_SUITE option.
+

There should be a comment here that refers to the remaining legacy reference in package/lvm2/Config.in, see
comment at BR2_TARGET_UBOOT_CUSTOM_GIT_REPO_URL_WRAP.

True, I forgot about that. Thanks.

Quoted 23 lines
Quoted 23 lines

config BR2_PACKAGE_QT_JAVASCRIPTCORE
bool "qt javascriptcore option removed"
select BR2_LEGACY

package/lvm2/Config.in +5 -3
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
           user-space. This makes it useful for not only LVM, but EVMS,
           software raid, and other drivers that create "virtual" block devices.
 
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-       bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+       bool "install the full LVM2 suite"
+       default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
         depends on BR2_PACKAGE_LVM2
         help
-         Install dmsetup only and skip the LVM2 suite.
+         Install the full LVM2 suite. When this option is not set, only
+         dmsetup is installed.
 
Quoted 10 lines

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 32

Quoted 7 lines
package/lvm2/Config.in +1 -1
  config BR2_PACKAGE_LVM2_APP_LIBRARY
         bool "install application library"
         depends on BR2_PACKAGE_LVM2
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_FULL_SUITE
         help
           Install application library (liblvm2app).

It seems odd to me that the 'full suite' does not include the app
library: it is not 'full'.

Granted, but I can't think of a better name. Anyone?

Quoted 3 lines

It would have been nice if we could have options for the different components, but at first sight there seem to be quite a number of components in lvm2: (rough estimate below)

Yeah, and that would anyway be a separate patch.

Regards,
Arnout

[snip]

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#7 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Quoted 66 lines · Arnout Vandecappelle
Quoted 5 lines · Thomas De Schampheleire

Hi Arnout,

On Sun, Nov 24, 2013 at 12:50 AM, Arnout Vandecappelle (Essensium/Mind) <[email protected]> wrote:

[snip]

Quoted 7 lines
Quoted 7 lines
Config.in.legacy +7 -0
+config BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       bool "lvm2's 'dmsetup only' option removed"
+       select BR2_LEGACY
+       help
+         The BR2_PACKAGE_LVM2_DMSETUP_ONLY was a negative option, which
+         led to problems with other packages that need the full lvm2
+         suite. Therefore, the option has been replaced with the
Quoted 8 lines

Config.in.legacy endif · line 112

Quoted 3 lines

positive

Config.in.legacy +2 -0
+         BR2_PACKAGE_LVM2_FULL_SUITE option.
+

There should be a comment here that refers to the remaining legacy reference in package/lvm2/Config.in, see
comment at BR2_TARGET_UBOOT_CUSTOM_GIT_REPO_URL_WRAP.

True, I forgot about that. Thanks.

Quoted 24 lines
Quoted 24 lines

config BR2_PACKAGE_QT_JAVASCRIPTCORE
bool "qt javascriptcore option removed"
select BR2_LEGACY

package/lvm2/Config.in +0 -0
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
           user-space. This makes it useful for not only LVM, but EVMS,
           software raid, and other drivers that create "virtual" block

devices.

package/lvm2/Config.in +5 -3
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-       bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+       bool "install the full LVM2 suite"
+       default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
         depends on BR2_PACKAGE_LVM2
         help
-         Install dmsetup only and skip the LVM2 suite.
+         Install the full LVM2 suite. When this option is not set, only
+         dmsetup is installed.
Quoted 11 lines

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 32

Quoted 7 lines

config BR2_PACKAGE_LVM2_APP_LIBRARY
bool "install application library"
depends on BR2_PACKAGE_LVM2

package/lvm2/Config.in +1 -1
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_FULL_SUITE
         help
           Install application library (liblvm2app).

It seems odd to me that the 'full suite' does not include the app
library: it is not 'full'.

Quoted 1 line

Granted, but I can't think of a better name. Anyone?

How 'large' is this application library? In this particular configuration it is 810K (liblvm2app.so). Is it acceptable to include it always in the 'full suite' (and thus remove the option BR2_PACKAGE_LVM2_APP_LIBRARY)?

Best regards,
Thomas

#8 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Quoted 67 lines · Thomas De Schampheleire
Quoted 66 lines · Arnout Vandecappelle
Quoted 5 lines · Thomas De Schampheleire

Hi Arnout,

On Sun, Nov 24, 2013 at 12:50 AM, Arnout Vandecappelle (Essensium/Mind) <[email protected]> wrote:

[snip]

Quoted 7 lines
Quoted 7 lines
Config.in.legacy +7 -0
+config BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       bool "lvm2's 'dmsetup only' option removed"
+       select BR2_LEGACY
+       help
+         The BR2_PACKAGE_LVM2_DMSETUP_ONLY was a negative option, which
+         led to problems with other packages that need the full lvm2
+         suite. Therefore, the option has been replaced with the
Quoted 8 lines

Config.in.legacy endif · line 112

Quoted 3 lines

positive

Config.in.legacy +2 -0
+         BR2_PACKAGE_LVM2_FULL_SUITE option.
+

There should be a comment here that refers to the remaining legacy reference in package/lvm2/Config.in, see
comment at BR2_TARGET_UBOOT_CUSTOM_GIT_REPO_URL_WRAP.

True, I forgot about that. Thanks.

Quoted 35 lines
Quoted 21 lines

config BR2_PACKAGE_QT_JAVASCRIPTCORE
bool "qt javascriptcore option removed"
select BR2_LEGACY

package/lvm2/Config.in +0 -0
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
            user-space. This makes it useful for not only LVM, but EVMS,
            software raid, and other drivers that create "virtual" block

devices.

package/lvm2/Config.in +3 -3
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-       bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+       bool "install the full LVM2 suite"
+       default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
          depends on BR2_PACKAGE_LVM2
          help
-         Install dmsetup only and skip the LVM2 suite.

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 32

Quoted 10 lines
package/lvm2/Config.in +3 -1
+         Install the full LVM2 suite. When this option is not set, only
+         dmsetup is installed.
 
   config BR2_PACKAGE_LVM2_APP_LIBRARY
          bool "install application library"
          depends on BR2_PACKAGE_LVM2
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_FULL_SUITE
          help
            Install application library (liblvm2app).

It seems odd to me that the 'full suite' does not include the app
library: it is not 'full'.

Quoted 6 lines
Quoted 1 line

Granted, but I can't think of a better name. Anyone?

How 'large' is this application library? In this particular configuration it is 810K (liblvm2app.so). Is it acceptable to include it always in the 'full suite' (and thus remove the option BR2_PACKAGE_LVM2_APP_LIBRARY)?

To help answer that question, on my x86_64 glibc build:

3.0M    /tmp/lvm2-app-library
2.1M    /tmp/lvm2-full-suite
472K    /tmp/lvm2-minimal

So the app library adds another 40% to the package size. I guess that's
significant enough not to remove the existing option. Especially because the only reason is that we can't find a better name for it :-)

My original name was "install all binaries", but I found that less
descriptive than "full suite". Maybe "install all LVM2 programs"?

I could also copy the full list (which is in the .mk file) to the help
text.

Regards,
Arnout

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#9 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Quoted 3 lines · Arnout Vandecappelle
Quoted 2 lines · Thomas De Schampheleire

On Mon, Nov 25, 2013 at 12:02 PM, Arnout Vandecappelle <[email protected]> wrote:

[snip]

Quoted 11 lines
Quoted 5 lines

How 'large' is this application library? In this particular configuration it is 810K (liblvm2app.so). Is it acceptable to include it always in the 'full suite' (and thus remove the option BR2_PACKAGE_LVM2_APP_LIBRARY)?

To help answer that question, on my x86_64 glibc build:

3.0M    /tmp/lvm2-app-library
2.1M    /tmp/lvm2-full-suite
472K    /tmp/lvm2-minimal

Sorry, I forgot to strip and remove the redundant files. It should be:

2.5M    /tmp/lvm2-app-library
1.6M    /tmp/lvm2-full-suite
396K    /tmp/lvm2-minimal

Conclusion remains the same, though.

Quoted 17 lines

So the app library adds another 40% to the package size. I guess that's
significant enough not to remove the existing option. Especially because the only reason is that we can't find a better name for it :-)

My original name was "install all binaries", but I found that less
descriptive than "full suite". Maybe "install all LVM2 programs"?

I could also copy the full list (which is in the .mk file) to the help
text.

Regards,
Arnout

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#10 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

In reply to Arnout Vandecappelle
Quoted 1 line · Arnout Vandecappelle

On 25/11/13 16:10, Thomas De Schampheleire wrote:

[..]

Quoted 15 lines
Quoted 6 lines

How 'large' is this application library? In this particular configuration it is 810K (liblvm2app.so). Is it acceptable to include it always in the 'full suite' (and thus remove the option BR2_PACKAGE_LVM2_APP_LIBRARY)?

To help answer that question, on my x86_64 glibc build:

3.0M    /tmp/lvm2-app-library
2.1M    /tmp/lvm2-full-suite
472K    /tmp/lvm2-minimal
Quoted 6 lines

So the app library adds another 40% to the package size. I guess that's
significant enough not to remove the existing option. Especially because the only reason is that we can't find a better name for it :-)

My original name was "install all binaries", but I found that less
descriptive than "full suite". Maybe "install all LVM2 programs"?

It's more than just programs, right? The default SUBDIRS includes some libraries too:
SUBDIRS = conf daemons include lib libdaemon libdm man scripts tools

Maybe we should add 'default' in the name somewhere, as what it installed is the set of programs/libraries/headers that are installed by default by 'make install'. The APPLIB option then additionally passes an extra request to configure. So for example:
BR2_PACKAGE_LVM2_DEFAULT_INSTALL BR2_PACKAGE_LVM2_DEFAULT_SUITE BR2_PACKAGE_LVM2_NOT_ONLY_DEVICEMAPPER ;-)
...

Quoted 3 lines

I could also copy the full list (which is in the .mk file) to the help
text.

That list is only used for the uninstall, so are we sure that it is up-to-date?

Best regards,
Thomas

#11 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Quoted 22 lines · Thomas De Schampheleire
Quoted 1 line · Arnout Vandecappelle

On 25/11/13 16:10, Thomas De Schampheleire wrote:

[..]

Quoted 19 lines
Quoted 6 lines

How 'large' is this application library? In this particular configuration it is 810K (liblvm2app.so). Is it acceptable to include it always in the 'full suite' (and thus remove the option BR2_PACKAGE_LVM2_APP_LIBRARY)?

To help answer that question, on my x86_64 glibc build:

3.0M    /tmp/lvm2-app-library
2.1M    /tmp/lvm2-full-suite
472K    /tmp/lvm2-minimal

So the app library adds another 40% to the package size. I guess that's
significant enough not to remove the existing option. Especially because the only reason is that we can't find a better name for it :-)

Quoted 6 lines
Quoted 2 lines

My original name was "install all binaries", but I found that less
descriptive than "full suite". Maybe "install all LVM2 programs"?

It's more than just programs, right? The default SUBDIRS includes some libraries too:
SUBDIRS = conf daemons include lib libdaemon libdm man scripts tools

I looked at what is actually installed: the only library is
libdevmapper. So I guess libdaemon is linked statically.

Quoted 11 lines

Maybe we should add 'default' in the name somewhere, as what it installed is the set of programs/libraries/headers that are installed by default by 'make install'. The APPLIB option then additionally passes an extra request to configure. So for example:
BR2_PACKAGE_LVM2_DEFAULT_INSTALL BR2_PACKAGE_LVM2_DEFAULT_SUITE BR2_PACKAGE_LVM2_NOT_ONLY_DEVICEMAPPER ;-)
...

Quoted 1 line
Quoted 4 lines
Quoted 2 lines

I could also copy the full list (which is in the .mk file) to the help
text.

That list is only used for the uninstall, so are we sure that it is up-to-date?

I can put the up-to-date list there for the moment, but of course we
will not be sure in the future.

Regards,
Arnout

Quoted 5 lines

Best regards,
Thomas

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#12 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

In reply to Arnout Vandecappelle
Quoted 1 line · Arnout Vandecappelle

On 25/11/13 10:46, Thomas De Schampheleire wrote:

[..]

Quoted 33 lines
Quoted 20 lines
Quoted 20 lines
package/lvm2/Config.in +0 -0
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
           user-space. This makes it useful for not only LVM, but EVMS,
           software raid, and other drivers that create "virtual" block

devices.

package/lvm2/Config.in +5 -3
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-       bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+       bool "install the full LVM2 suite"
+       default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
         depends on BR2_PACKAGE_LVM2
         help
-         Install dmsetup only and skip the LVM2 suite.
+         Install the full LVM2 suite. When this option is not set, only
+         dmsetup is installed.
Quoted 11 lines

package/lvm2/Config.in

Quoted 7 lines

config BR2_PACKAGE_LVM2_APP_LIBRARY
bool "install application library"
depends on BR2_PACKAGE_LVM2

package/lvm2/Config.in +1 -1
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_FULL_SUITE
         help
           Install application library (liblvm2app).

It seems odd to me that the 'full suite' does not include the app
library: it is not 'full'.

Quoted 1 line

Granted, but I can't think of a better name. Anyone?

I was trying an alternative approach with 'choice', but failed. Here is what I tried:

$ hg diff package/
package/lvm2/Config.in +13 -3
@@ -16,19 +16,29 @@ config BR2_PACKAGE_LVM2
          user-space. This makes it useful for not only LVM, but EVMS,
          software raid, and other drivers that create "virtual" block devices.
 
+if BR2_PACKAGE_LVM2
+
+choice
+       prompt "Installation type"
+
+config BR2_PACKAGE_LVM2_STANDARD_INSTALL
+       bool "standard installation"
+
 config BR2_PACKAGE_LVM2_DMSETUP_ONLY
        bool "install dmsetup only"
-       depends on BR2_PACKAGE_LVM2
        help
          Install dmsetup only and skip the LVM2 suite.
 
+endchoice
+
 config BR2_PACKAGE_LVM2_APP_LIBRARY
        bool "install application library"
-       depends on BR2_PACKAGE_LVM2
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_STANDARD_INSTALL
        help
          Install application library (liblvm2app).
 
+endif
+
 comment "lvm2 needs a toolchain w/ largefile"
        depends on BR2_USE_MMU
        depends on !BR2_LARGEFILE
package/udisks/Config.in +1 -0
@@ -30,6 +30,7 @@ if BR2_PACKAGE_UDISKS
 config BR2_PACKAGE_UDISKS_LVM2
        bool "lvm2 support"
        select BR2_PACKAGE_LVM2_APP_LIBRARY
+       select BR2_PACKAGE_LVM2_STANDARD_INSTALL
        help
          Enable LVM2 support
 

If anyone can make this work, I'd prefer it over the 'magic' if-not-full-suite-you-get-dmonly approach.

If not, what about the naming 'standard installation' to refer to what is called 'full suite' in your patch. If it is not standard install, it is dmonly.

Best regards,
Thomas

#13 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Quoted 65 lines · Thomas De Schampheleire
Quoted 1 line · Arnout Vandecappelle

On 25/11/13 10:46, Thomas De Schampheleire wrote:

[..]

Quoted 33 lines
Quoted 20 lines
Quoted 20 lines
package/lvm2/Config.in +0 -0
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
            user-space. This makes it useful for not only LVM, but EVMS,
            software raid, and other drivers that create "virtual" block

devices.

package/lvm2/Config.in +5 -3
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-       bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+       bool "install the full LVM2 suite"
+       default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
          depends on BR2_PACKAGE_LVM2
          help
-         Install dmsetup only and skip the LVM2 suite.
+         Install the full LVM2 suite. When this option is not set, only
+         dmsetup is installed.
Quoted 11 lines

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 38

Quoted 7 lines

config BR2_PACKAGE_LVM2_APP_LIBRARY
bool "install application library"
depends on BR2_PACKAGE_LVM2

package/lvm2/Config.in +1 -1
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_FULL_SUITE
          help
            Install application library (liblvm2app).

It seems odd to me that the 'full suite' does not include the app
library: it is not 'full'.

Quoted 1 line

Granted, but I can't think of a better name. Anyone?

I was trying an alternative approach with 'choice', but failed. Here is what I tried:

$ hg diff package/
package/lvm2/Config.in +10 -1
@@ -16,19 +16,29 @@ config BR2_PACKAGE_LVM2
           user-space. This makes it useful for not only LVM, but EVMS,
           software raid, and other drivers that create "virtual" block devices.
 
+if BR2_PACKAGE_LVM2
+
+choice
+       prompt "Installation type"
+
+config BR2_PACKAGE_LVM2_STANDARD_INSTALL
+       bool "standard installation"
+
  config BR2_PACKAGE_LVM2_DMSETUP_ONLY
         bool "install dmsetup only"
-       depends on BR2_PACKAGE_LVM2
         help
           Install dmsetup only and skip the LVM2 suite.
 
+endchoice
+

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 41

Quoted 10 lines
package/lvm2/Config.in +3 -2
  config BR2_PACKAGE_LVM2_APP_LIBRARY
         bool "install application library"
-       depends on BR2_PACKAGE_LVM2
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_STANDARD_INSTALL
         help
           Install application library (liblvm2app).
 
+endif
+

So, what didn't work with this?

Quoted 15 lines
package/lvm2/Config.in +0 -0
  comment "lvm2 needs a toolchain w/ largefile"
         depends on BR2_USE_MMU
         depends on !BR2_LARGEFILE
package/udisks/Config.in +1 -0
@@ -30,6 +30,7 @@ if BR2_PACKAGE_UDISKS
  config BR2_PACKAGE_UDISKS_LVM2
         bool "lvm2 support"
         select BR2_PACKAGE_LVM2_APP_LIBRARY
+       select BR2_PACKAGE_LVM2_STANDARD_INSTALL
         help
           Enable LVM2 support
 
Quoted 2 lines

If anyone can make this work, I'd prefer it over the 'magic' if-not-full-suite-you-get-dmonly approach.

I don't.

* Choice is not covered by randconfig, so not tried by the autobuilders.

* Choice is more cumbersome for the user

* I don't see what is wrong with a boolean option that installs more stuff.

Quoted 4 lines

If not, what about the naming 'standard installation' to refer to what is called 'full suite' in your patch. If it is not standard install, it is dmonly.

I called it:

config BR2_PACKAGE_LVM2_STANDARD_INSTALL
bool "standard install instead of only dmsetup" default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11 help
Install the standard suite of lvm2 programs. When this option is not set, only dmsetup is installed.

Regards,
Arnout

Signature · 6 lines
Arnout Vandecappelle                          arnout at mind be
Senior Embedded Software Architect            +32-16-286500
Essensium/Mind                                http://www.mind.be
G.Geenslaan 9, 3001 Leuven, Belgium           BE 872 984 063 RPR Leuven
LinkedIn profile: http://www.linkedin.com/in/arnoutvandecappelle
GPG fingerprint:  7CB5 E4CC 6C2E EFD4 6E3D A754 F963 ECAB 2450 2F1F
#14 lore raw

PATCH 1/1 lvm2: replace !BR2_PACKAGE_LVM2_DMSETUP_ONLY by BR2_PACKAGE_LVM2_FULL_SUITE

Quoted 74 lines · Arnout Vandecappelle
Quoted 73 lines · Thomas De Schampheleire
Quoted 1 line · Arnout Vandecappelle

On 25/11/13 10:46, Thomas De Schampheleire wrote:

[..]

Quoted 33 lines
Quoted 20 lines
Quoted 20 lines
package/lvm2/Config.in +0 -0
index b3988ea..849aab3 100644
@@ -16,16 +16,18 @@ config BR2_PACKAGE_LVM2
            user-space. This makes it useful for not only LVM, but EVMS,
            software raid, and other drivers that create "virtual" block

devices.

package/lvm2/Config.in +5 -3
-config BR2_PACKAGE_LVM2_DMSETUP_ONLY
-       bool "install dmsetup only"
+config BR2_PACKAGE_LVM2_FULL_SUITE
+       bool "install the full LVM2 suite"
+       default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11
          depends on BR2_PACKAGE_LVM2
          help
-         Install dmsetup only and skip the LVM2 suite.
+         Install the full LVM2 suite. When this option is not set, only
+         dmsetup is installed.
Quoted 11 lines

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 38

Quoted 7 lines

config BR2_PACKAGE_LVM2_APP_LIBRARY
bool "install application library"
depends on BR2_PACKAGE_LVM2

package/lvm2/Config.in +1 -1
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_FULL_SUITE
          help
            Install application library (liblvm2app).

It seems odd to me that the 'full suite' does not include the app
library: it is not 'full'.

Quoted 1 line

Granted, but I can't think of a better name. Anyone?

I was trying an alternative approach with 'choice', but failed. Here is what I tried:

$ hg diff package/
package/lvm2/Config.in +11 -3
@@ -16,19 +16,29 @@ config BR2_PACKAGE_LVM2
           user-space. This makes it useful for not only LVM, but EVMS,
           software raid, and other drivers that create "virtual" block devices.
 
+if BR2_PACKAGE_LVM2
+
+choice
+       prompt "Installation type"
+
+config BR2_PACKAGE_LVM2_STANDARD_INSTALL
+       bool "standard installation"
+
  config BR2_PACKAGE_LVM2_DMSETUP_ONLY
         bool "install dmsetup only"
-       depends on BR2_PACKAGE_LVM2
         help
           Install dmsetup only and skip the LVM2 suite.
 
+endchoice
+
  config BR2_PACKAGE_LVM2_APP_LIBRARY
         bool "install application library"
-       depends on BR2_PACKAGE_LVM2
-       depends on !BR2_PACKAGE_LVM2_DMSETUP_ONLY
+       select BR2_PACKAGE_LVM2_STANDARD_INSTALL
         help
           Install application library (liblvm2app).
 
Quoted 4 lines

package/lvm2/Config.in config BR2_PACKAGE_LVM2 · line 41

Quoted 2 lines
package/lvm2/Config.in +2 -0
+endif
+

So, what didn't work with this?

It behaved as before, udisks lvm support would not enable the standard install and keep the dmonly option...

Quoted 17 lines
Quoted 17 lines
package/lvm2/Config.in +0 -0
  comment "lvm2 needs a toolchain w/ largefile"
         depends on BR2_USE_MMU
         depends on !BR2_LARGEFILE
package/udisks/Config.in +1 -0
@@ -30,6 +30,7 @@ if BR2_PACKAGE_UDISKS
  config BR2_PACKAGE_UDISKS_LVM2
         bool "lvm2 support"
         select BR2_PACKAGE_LVM2_APP_LIBRARY
+       select BR2_PACKAGE_LVM2_STANDARD_INSTALL
         help
           Enable LVM2 support
 
Quoted 6 lines
Quoted 2 lines

If anyone can make this work, I'd prefer it over the 'magic' if-not-full-suite-you-get-dmonly approach.

I don't.

* Choice is not covered by randconfig, so not tried by the autobuilders.

Didn't know that.

Quoted 10 lines

* Choice is more cumbersome for the user

* I don't see what is wrong with a boolean option that installs more stuff.

Quoted 4 lines

If not, what about the naming 'standard installation' to refer to what is called 'full suite' in your patch. If it is not standard install, it is dmonly.

Quoted 10 lines

I called it:

config BR2_PACKAGE_LVM2_STANDARD_INSTALL
bool "standard install instead of only dmsetup" default y if !BR2_PACKAGE_LVM2_DMSETUP_ONLY # legacy 2013.11 help
Install the standard suite of lvm2 programs. When this option is not set, only dmsetup is installed.

Works for me!

Best regards,
Thomas

#15 lore raw
In reply to Thomas De Schampheleire

Dear Thomas De Schampheleire,

Quoted 1 line · Thomas De Schampheleire
Quoted 1 line
       arm |                     systemd-44 | NOK | http://autobuild.buildroot.net/results/e5484bcdcca7f2cb48913d65271de4b957ea3450/

It's a fairly old toolchain that I've built with Crosstool-NG. It uses uClibc 0.9.33.2, but there might be some missing uClibc configuration options, or the kernel headers might be old.

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#16 lore raw

Hi Thomas,

Quoted 9 lines · Thomas Petazzoni

Dear Thomas De Schampheleire,

Quoted 1 line · Thomas De Schampheleire
Quoted 1 line
       arm |                     systemd-44 | NOK | http://autobuild.buildroot.net/results/e5484bcdcca7f2cb48913d65271de4b957ea3450/

It's a fairly old toolchain that I've built with Crosstool-NG. It uses uClibc 0.9.33.2, but there might be some missing uClibc configuration options, or the kernel headers might be old.

Then what do we do? Does it make sense to use this old toolchain, or should we update it?

Thanks,
Thomas

#17 lore raw

Dear Thomas De Schampheleire,

Quoted 6 lines · Thomas De Schampheleire
Quoted 3 lines

It's a fairly old toolchain that I've built with Crosstool-NG. It uses uClibc 0.9.33.2, but there might be some missing uClibc configuration options, or the kernel headers might be old.

Then what do we do? Does it make sense to use this old toolchain, or should we update it?

It's on my TODO-list to update all the toolchains used in the autobuilders. I've rebuilt all the Buildroot toolchains already, and prepared the configs for the Crosstool-NG toolchains as well. Just need a bit of time to finalize this (I have been working on SELinux patches yesterday evening with Ryan and Clayton).

Thomas

Signature · 3 lines
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com
#18 lore raw
In reply to Thomas De Schampheleire

Hi Thomas,

Quoted 30 lines · Thomas De Schampheleire
Quoted 8 lines · Thomas Petazzoni

Build statistics for 2013-11-22
===============================

success : 68
failures : 34
timeouts : 0
TOTAL : 102

[..]

Quoted 1 line
      sh4a |                   poco-1.4.6p1 | NOK | http://autobuild.buildroot.net/results/8dc9bca5969d455cc30206fc1ae6e1d3ba6eb6ca/

Error is:

In file included from src/FPEnvironment_C99.cpp:37:0,
from src/FPEnvironment.cpp:48:
include/Poco/FPEnvironment_C99.h:56:30: error: 'FE_DOWNWARD' was not
declared in this scope
include/Poco/FPEnvironment_C99.h:57:30: error: 'FE_UPWARD' was not
declared in this scope
make[2]: *** [/home/test/test/2/output/build/poco-1.4.6p1/Foundation/obj/Linux/sh4a/release_shared/FPEnvironment.o] Error 1
make[2]: Leaving directory
`/home/test/test/2/output/build/poco-1.4.6p1/Foundation'
make[1]: *** [Foundation-libexec] Error 2
make[1]: Leaving directory `/home/test/test/2/output/build/poco-1.4.6p1'
make: *** [/home/test/test/2/output/build/poco-1.4.6p1/.stamp_built] Error 2

Quoted 7 lines

The same build failure has been discussed a long time ago here:
http://lists.busybox.net/pipermail/buildroot/2011-September/046013.html
when the poco package was introduced. Then, the fpenvironment was disabled when the C library is uclibc.
In the current failure, the C library seems to be glibc. I'm including Baruch on this one, who was involved in the above thread. Baruch: any idea?

I see the following comment in glibc sysdeps/sh/sh4/fpu/bits/fenv.h:

/* The SH FPU supports two of the four defined rounding modes: round to nearest
and round to zero. We use again the bit positions in the FPU control word as the values for the appropriate macros. */

I don't know enough about fenv to know what the right solution here would be. The easiest one is to add --no-fpenvironment for sh4a. Is this acceptable?

baruch

Signature · 3 lines
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.2.679.5364, http://www.tkos.co.il -
#19 lore raw
Quoted 41 lines · Baruch Siach

Hi Thomas,

Quoted 30 lines · Thomas De Schampheleire
Quoted 8 lines · Thomas Petazzoni

Build statistics for 2013-11-22
===============================

success : 68
failures : 34
timeouts : 0
TOTAL : 102

[..]

Quoted 1 line
      sh4a |                   poco-1.4.6p1 | NOK | http://autobuild.buildroot.net/results/8dc9bca5969d455cc30206fc1ae6e1d3ba6eb6ca/

Error is:

In file included from src/FPEnvironment_C99.cpp:37:0,
from src/FPEnvironment.cpp:48:
include/Poco/FPEnvironment_C99.h:56:30: error: 'FE_DOWNWARD' was not
declared in this scope
include/Poco/FPEnvironment_C99.h:57:30: error: 'FE_UPWARD' was not
declared in this scope
make[2]: *** [/home/test/test/2/output/build/poco-1.4.6p1/Foundation/obj/Linux/sh4a/release_shared/FPEnvironment.o] Error 1
make[2]: Leaving directory
`/home/test/test/2/output/build/poco-1.4.6p1/Foundation'
make[1]: *** [Foundation-libexec] Error 2
make[1]: Leaving directory `/home/test/test/2/output/build/poco-1.4.6p1'
make: *** [/home/test/test/2/output/build/poco-1.4.6p1/.stamp_built] Error 2

Quoted 7 lines

The same build failure has been discussed a long time ago here:
http://lists.busybox.net/pipermail/buildroot/2011-September/046013.html
when the poco package was introduced. Then, the fpenvironment was disabled when the C library is uclibc.
In the current failure, the C library seems to be glibc. I'm including Baruch on this one, who was involved in the above thread. Baruch: any idea?

Quoted 8 lines

I see the following comment in glibc sysdeps/sh/sh4/fpu/bits/fenv.h:

/* The SH FPU supports two of the four defined rounding modes: round to nearest
and round to zero. We use again the bit positions in the FPU control word as the values for the appropriate macros. */

I don't know enough about fenv to know what the right solution here would be. The easiest one is to add --no-fpenvironment for sh4a. Is this acceptable?

I couldn't tell you. We already pass this option in one situation, at first sight it would be acceptable to pass it in another.

Best regards,
Thomas

#20 lore raw
In reply to Thomas Petazzoni

Hi all,

I tried to reproduce the following build failures:

powerpc |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/b6f619f9c249fc653dc65f0f516cb846683700bb/
arm     |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/7e4/7e4e369dc5640619e95983ad0b4571f3a3f0d84b/

(also on i686 last week...)

I have not been able to reproduce this error...
I simply found a missing dependency on libltdl

Any Idea ?

Best regards,

Signature · 8 lines
Romain Naour

OPEN WIDE Ing?nierie - Paris
23/25, rue Daviel 75013 PARIS
http://ingenierie.openwide.fr

Le blog des technologies libres et embarqu?es :
http://www.linuxembedded.fr
#21 lore raw

Romain, All,

On 2013-11-23 21:38 +0100, Romain Naour spake thusly:

Quoted 10 lines

I tried to reproduce the following build failures:

powerpc |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/b6f619f9c249fc653dc65f0f516cb846683700bb/
arm     |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/7e4/7e4e369dc5640619e95983ad0b4571f3a3f0d84b/

(also on i686 last week...)

I have not been able to reproduce this error...
I simply found a missing dependency on libltdl

Any Idea ?

Hmmm. Maybe an issue with libusb vs. libusv-compat? Both config above have libusb-compat enabled.

I'll try a build here with only libusb.

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#22 lore raw

Romain, All,

On 2013-11-23 22:22 +0100, Yann E. MORIN spake thusly:

Quoted 10 lines

On 2013-11-23 21:38 +0100, Romain Naour spake thusly:

Quoted 9 lines

I tried to reproduce the following build failures:

powerpc |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/b6f619f9c249fc653dc65f0f516cb846683700bb/
arm     |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/7e4/7e4e369dc5640619e95983ad0b4571f3a3f0d84b/

(also on i686 last week...)

I have not been able to reproduce this error...
I simply found a missing dependency on libltdl

Quoted 6 lines
Quoted 1 line

Any Idea ?

Hmmm. Maybe an issue with libusb vs. libusv-compat? Both config above have libusb-compat enabled.

I'll try a build here with only libusb.

Well, no, it really needs libusb-compat.

OK, I'll try to rebuild those failing configs...

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'
#23 lore raw
In reply to Romain Naour

Romain, All,

On 2013-11-23 21:38 +0100, Romain Naour spake thusly:

Quoted 7 lines

I tried to reproduce the following build failures:

powerpc |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/b6f619f9c249fc653dc65f0f516cb846683700bb/
arm     |                      nut-2.6.5 | NOK | http://autobuild.buildroot.net/results/7e4/7e4e369dc5640619e95983ad0b4571f3a3f0d84b/

(also on i686 last week...)

I have not been able to reproduce this error...

Neither was I.

The errors happen on Peter's build machine, which is a PPC. Maybe something from the host is leaking in the build, and does leak on neither your machine nor mine.

Regards,
Yann E. MORIN.

Signature · 6 lines
.-----------------.--------------------.------------------.--------------------.
|  Yann E. MORIN  | Real-Time Embedded | /"\ ASCII RIBBON | Erics' conspiracy: |
| +33 662 376 056 | Software  Designer | \ / CAMPAIGN     |  ___               |
| +33 223 225 172 `------------.-------:  X  AGAINST      |  \e/  There is no  |
| http://ymorin.is-a-geek.org/ | _/*\_ | / \ HTML MAIL    |   v   conspiracy.  |
'------------------------------^-------^------------------^--------------------'