Skip to contentSkip to conversation

Pull request for branch yem/host-packages

GIT PULL New Yann E. MORIN 3 participants 12 messages Dec 2, 2013 – Dec 10, 2013

lore
Share & tools
lore
#0 lore raw

From: "Yann E. MORIN" <[email protected]>

Hello All!

This series adds parted as a host package. This will allow (eg.) post-image scripts to use parted.

Regards,
Yann E. MORIN.

The following changes since commit edce303b8da95a9b19939ab7a4199a39c0483c6f:

libtasn1: fix botched makeinfo/missing logic (2013-12-02 13:33:02 +0100)

are available in the git repository at:

git://gitorious.org/buildroot/buildroot.git yem/host-packages

for you to fetch changes up to 1891fdbc7ad11ce084f2be315871eb10573953b7:

package/parted: add a host variant (2013-12-02 23:27:45 +0100)

----------------------------------------------------------------
Yann E. MORIN (3):
package/lvm2: remove uninstall commands
package/lvm2: add a host variant
package/parted: add a host variant

 package/Config.in.host        |  1 +
 package/lvm2/lvm2.mk          | 13 ++++---------
 package/parted/Config.in.host |  6 ++++++
 package/parted/parted.mk      |  4 ++++
 4 files changed, 15 insertions(+), 9 deletions(-)
 create mode 100644 package/parted/Config.in.host
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.  |
'------------------------------^-------^------------------^--------------------'
#1 lore raw

PATCH 1/3 package/lvm2: remove uninstall commands

From: "Yann E. MORIN" <[email protected]>

We do not properly support uninstalling packages, so just remove the uninstall commands.

---
 package/lvm2/lvm2.mk | 10 ----------
 1 file changed, 10 deletions(-)
package/lvm2/lvm2.mk +0 -10
index 4099c21..2761100 100644
@@ -50,14 +50,4 @@ else
 LVM2_CONF_OPT += --disable-applib
 endif
 
-define LVM2_UNINSTALL_STAGING_CMDS
-	rm -f $(addprefix $(STAGING_DIR)/usr/sbin/,$(LVM2_BINS))
-	rm -f $(addprefix $(STAGING_DIR)/usr/lib/,libdevmapper.so*)
-endef
-
-define LVM2_UNINSTALL_TARGET_CMDS
-	rm -f $(addprefix $(TARGET_DIR)/usr/sbin/,$(LVM2_BINS))
-	rm -f $(addprefix $(TARGET_DIR)/usr/lib/,libdevmapper.so*)
-endef
-
 $(eval $(autotools-package))
Signature · 1 line
1.8.1.2
#2 lore raw

PATCH 1/3 package/lvm2: remove uninstall commands

Dear Yann E. MORIN,

Quoted 9 lines · Yann E. MORIN

From: "Yann E. MORIN" <[email protected]>

We do not properly support uninstalling packages, so just remove the uninstall commands.

---
 package/lvm2/lvm2.mk | 10 ----------
 1 file changed, 10 deletions(-)

I've marked this patch as superseded in patchwork, since all uninstall commands have been deleted by Thomas DS patch, including the lvm2 one.

Best regards,

Thomas

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

PATCH 1/3 package/lvm2: remove uninstall commands

Thomas, All,

On 2013-12-06 10:38 +0100, Thomas Petazzoni spake thusly:

Quoted 6 lines · Yann E. MORIN
Quoted 5 lines

From: "Yann E. MORIN" <[email protected]>

We do not properly support uninstalling packages, so just remove the uninstall commands.

Quoted 7 lines
Quoted 4 lines
---
 package/lvm2/lvm2.mk | 10 ----------
 1 file changed, 10 deletions(-)

I've marked this patch as superseded in patchwork, since all uninstall commands have been deleted by Thomas DS patch, including the lvm2 one.

Yes, thank you! :-)

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.  |
'------------------------------^-------^------------------^--------------------'
#4 lore raw

PATCH 2/3 package/lvm2: add a host variant

In reply to Yann E. MORIN

From: "Yann E. MORIN" <[email protected]>

Will be needed later for host-parted.

We need to force --with-confdir and --with-default-system-dir, otherwise lvm2 tries to install /etc/lvm/profile which we can't do, and don't want to do.

---
 package/lvm2/lvm2.mk | 5 +++++
 1 file changed, 5 insertions(+)
package/lvm2/lvm2.mk +5 -0
index 2761100..eada3ec 100644
@@ -50,4 +50,9 @@ else
 LVM2_CONF_OPT += --disable-applib
 endif
 
+HOST_LVM2_CONF_OPT += \
+	--with-confdir=$(HOST_DIR)/etc \
+	--with-default-system-dir=$(HOST_DIR)/etc/lvm
+
 $(eval $(autotools-package))
+$(eval $(host-autotools-package))
Signature · 1 line
1.8.1.2
#5 lore raw

PATCH 3/3 package/parted: add a host variant

In reply to Yann E. MORIN

From: "Yann E. MORIN" <[email protected]>

Usefull for for-build scripts to call parted, eg. to generate partition tables and such automatically.

Since the primary gaool is to use parted within scripts, we do not need readline, so it is forcibly disabled.

---
 package/Config.in.host        | 1 +
 package/parted/Config.in.host | 6 ++++++
 package/parted/parted.mk      | 4 ++++
 3 files changed, 11 insertions(+)
 create mode 100644 package/parted/Config.in.host
package/Config.in.host +1 -0
index 5fd2570..34e84bf 100644
@@ -10,6 +10,7 @@ source "package/lpc3250loader/Config.in.host"
 source "package/mtools/Config.in.host"
 source "package/omap-u-boot-utils/Config.in.host"
 source "package/openocd/Config.in.host"
+source "package/parted/Config.in.host"
 source "package/sam-ba/Config.in.host"
 source "package/sunxi-tools/Config.in.host"
 source "package/uboot-tools/Config.in.host"
package/parted/Config.in.host new file +6 -0
new file mode 100644
index 0000000..f5d7acc
@@ -0,0 +1,6 @@
+config BR2_PACKAGE_HOST_PARTED
+	bool "host parted"
+	help
+	  parted, the GNU partition resizing program
+
+	  http://www.gnu.org/software/parted/
package/parted/parted.mk +4 -0
index 773e332..a1c0e9e 100644
@@ -8,8 +8,12 @@ PARTED_VERSION = 3.1
 PARTED_SOURCE = parted-$(PARTED_VERSION).tar.xz
 PARTED_SITE = $(BR2_GNU_MIRROR)/parted
 PARTED_DEPENDENCIES = readline util-linux lvm2
+HOST_PARTED_DEPENDENCIES = host-util-linux host-lvm2
 PARTED_INSTALL_STAGING = YES
 PARTED_LICENSE = GPLv3+
 PARTED_LICENSE_FILES = COPYING
 
+HOST_PARTED_CONF_OPT += --without-readline
+
 $(eval $(autotools-package))
+$(eval $(host-autotools-package))
Signature · 1 line
1.8.1.2
#6 lore raw

PATCH 3/3 package/parted: add a host variant

Dear Yann E. MORIN,

Quoted 3 lines · Yann E. MORIN

From: "Yann E. MORIN" <[email protected]>

Usefull for for-build scripts to call parted, eg. to generate

Useful. See http://en.wiktionary.org/wiki/usefull.

Quoted 3 lines

partition tables and such automatically.

Since the primary gaool is to use parted within scripts, we

goal

Quoted 5 lines

do not need readline, so it is forcibly disabled.

PARTED_DEPENDENCIES = readline util-linux lvm2
+HOST_PARTED_DEPENDENCIES = host-util-linux host-lvm2

Do we really want the lvm2 dependency? It can be avoided by passing --disable-device-mapper.

package/parted/parted.mk PARTED_VERSION = 3.1 · line 19

Quoted 8 lines

PARTED_INSTALL_STAGING = YES
PARTED_LICENSE = GPLv3+
PARTED_LICENSE_FILES = COPYING

package/parted/parted.mk +3 -0
+HOST_PARTED_CONF_OPT += --without-readline
+
 $(eval $(autotools-package))
+$(eval $(host-autotools-package))

So maybe something like:

PARTED_DEPENDENCIES = util-linux

ifeq ($(BR2_PACKAGE_READLINE),y)
PARTED_DEPENDENCIES += readline
PARTED_CONF_OPT += --with-readline
else
PARTED_CONF_OPT += --without-readline
endif

ifeq ($(BR2_PACKAGE_LVM2),y)
PARTED_DEPENDENCIES += lvm2
PARTED_CONF_OPT += --enable-device-mapper
else
PARTED_CONF_OPT += --disable-device-mapper
endif

HOST_PARTED_DEPENDENCIES = host-util-linux

HOST_PARTED_CONF_OPT = \
	--without-readline \
	--disable-device-mapper

In the host variant of parted, lvm2 support is never enabled, as I believe it's pretty unlikely that lvm2 support will be needed to generate disk images. This would make PATCH 2/3 of your series unnecessary, of course.

In the target variant, we automatically enable lvm2 and readline support when available, which allows to build parted without lvm2 support when not needed.

Thomas

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

PATCH 3/3 package/parted: add a host variant

Thomas, All,

On 2013-12-06 10:48 +0100, Thomas Petazzoni spake thusly:

Quoted 13 lines · Yann E. MORIN
Quoted 3 lines

From: "Yann E. MORIN" <[email protected]>

Usefull for for-build scripts to call parted, eg. to generate

Useful. See http://en.wiktionary.org/wiki/usefull.

Quoted 3 lines

partition tables and such automatically.

Since the primary gaool is to use parted within scripts, we

goal

Quoted 8 lines
Quoted 5 lines

do not need readline, so it is forcibly disabled.

PARTED_DEPENDENCIES = readline util-linux lvm2
+HOST_PARTED_DEPENDENCIES = host-util-linux host-lvm2

Do we really want the lvm2 dependency? It can be avoided by passing --disable-device-mapper.

I think we want the same dependency the target parted has. If the target parted can handle lvm2 volumes, then we may have to generate them in the first place.

So, what about:

# If target-parted can handle lvm volumes, then host-parted

# should be, too, so as to be able to generate them.
# If target-parted can't handle lvm volumes, there is no reason
# for host-aprted to handle them.
ifeq ($(BR2_PACKAGE_LVM2),y)
PARTED_DEPENDENCIES += lvm2
HOST_PARTED_DEPENDENCIES += lvm2
PARTED_CONF_OPT += --enable-device-mapper
HOST_PARTED_CONF_OPT += --enable-device-mapper
else
PARTED_CONF_OPT += --disable-device-mapper
HOST_PARTED_CONF_OPT += --disable-device-mapper
endif

But see further thoughts below...

Quoted 28 lines

package/parted/parted.mk PARTED_VERSION = 3.1 · line 19

Quoted 8 lines

PARTED_INSTALL_STAGING = YES
PARTED_LICENSE = GPLv3+
PARTED_LICENSE_FILES = COPYING

package/parted/parted.mk +3 -0
+HOST_PARTED_CONF_OPT += --without-readline
+
 $(eval $(autotools-package))
+$(eval $(host-autotools-package))

So maybe something like:

PARTED_DEPENDENCIES = util-linux

ifeq ($(BR2_PACKAGE_READLINE),y)
PARTED_DEPENDENCIES += readline
PARTED_CONF_OPT += --with-readline
else
PARTED_CONF_OPT += --without-readline
endif

ifeq ($(BR2_PACKAGE_LVM2),y)
PARTED_DEPENDENCIES += lvm2
PARTED_CONF_OPT += --enable-device-mapper
else
PARTED_CONF_OPT += --disable-device-mapper
endif

HOST_PARTED_DEPENDENCIES = host-util-linux

Quoted 8 lines
HOST_PARTED_CONF_OPT = \
	--without-readline \
	--disable-device-mapper

In the host variant of parted, lvm2 support is never enabled, as I believe it's pretty unlikely that lvm2 support will be needed to generate disk images. This would make PATCH 2/3 of your series unnecessary, of course.

I don't think we should not be able to generate lvm volumes. I can see at least a few cases where it would be needed.

Of course, we may just add host-parted without lvm support right now, and add it later when the need really arises.

Thanks for the review; v2 on its way later tonight.

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.  |
'------------------------------^-------^------------------^--------------------'
#8 lore raw

PATCH 3/3 package/parted: add a host variant

Dear Yann E. MORIN,

Quoted 3 lines · Yann E. MORIN

I think we want the same dependency the target parted has. If the target parted can handle lvm2 volumes, then we may have to generate them in the first place.

Yes, we may. But that seems unlikely. I believe the most common usage of LVM on embedded platforms is for NAS/storage type devices, and for those systems, the root filesystem image is generally not on the LVM/RAID storage I believe, no?

Quoted 15 lines

So, what about:

# If target-parted can handle lvm volumes, then host-parted

# should be, too, so as to be able to generate them.
# If target-parted can't handle lvm volumes, there is no reason
# for host-aprted to handle them.
ifeq ($(BR2_PACKAGE_LVM2),y)
PARTED_DEPENDENCIES += lvm2
HOST_PARTED_DEPENDENCIES += lvm2
PARTED_CONF_OPT += --enable-device-mapper
HOST_PARTED_CONF_OPT += --enable-device-mapper
else
PARTED_CONF_OPT += --disable-device-mapper
HOST_PARTED_CONF_OPT += --disable-device-mapper
endif

While I do understand the logic behind what you're proposing, I'm not really comfortable with having the configuration of tools built for the host changed depending on the target configuration. It seems to be creating a bad precedent.

Quoted 10 lines
Quoted 4 lines

In the host variant of parted, lvm2 support is never enabled, as I believe it's pretty unlikely that lvm2 support will be needed to generate disk images. This would make PATCH 2/3 of your series unnecessary, of course.

I don't think we should not be able to generate lvm volumes. I can see at least a few cases where it would be needed.

Of course, we may just add host-parted without lvm support right now, and add it later when the need really arises.

Yes, that would be my proposal.

Best regards,

Thomas

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

PATCH 3/3 package/parted: add a host variant

Quoted 4 lines · Thomas Petazzoni

Dear Yann E. MORIN,

On Fri, 6 Dec 2013 17:56:14 +0100, Yann E. MORIN wrote:

Quoted 8 lines
Quoted 3 lines

I think we want the same dependency the target parted has. If the target parted can handle lvm2 volumes, then we may have to generate them in the first place.

Yes, we may. But that seems unlikely. I believe the most common usage of LVM on embedded platforms is for NAS/storage type devices, and for those systems, the root filesystem image is generally not on the LVM/RAID storage I believe, no?

It's also used for encrypted volumes, no?

Quoted 17 lines
Quoted 4 lines

So, what about:

# If target-parted can handle lvm volumes, then host-parted

Quoted 12 lines
# should be, too, so as to be able to generate them.
# If target-parted can't handle lvm volumes, there is no reason
# for host-aprted to handle them.
ifeq ($(BR2_PACKAGE_LVM2),y)
PARTED_DEPENDENCIES += lvm2
HOST_PARTED_DEPENDENCIES += lvm2
PARTED_CONF_OPT += --enable-device-mapper
HOST_PARTED_CONF_OPT += --enable-device-mapper
else
PARTED_CONF_OPT += --disable-device-mapper
HOST_PARTED_CONF_OPT += --disable-device-mapper
endif
Quoted 4 lines

While I do understand the logic behind what you're proposing, I'm not really comfortable with having the configuration of tools built for the host changed depending on the target configuration. It seems to be creating a bad precedent.

We already have a precedent: libxml2.

What is so bad about one package's configuration depending on another
package configuration?

Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

In the host variant of parted, lvm2 support is never enabled, as I believe it's pretty unlikely that lvm2 support will be needed to generate disk images. This would make PATCH 2/3 of your series unnecessary, of course.

Quoted 7 lines
Quoted 5 lines

I don't think we should not be able to generate lvm volumes. I can see at least a few cases where it would be needed.

Of course, we may just add host-parted without lvm support right now, and add it later when the need really arises.

Yes, that would be my proposal.

That's of course also fine - only implement it when it's actually used.

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 3/3 package/parted: add a host variant

Dear Arnout Vandecappelle,

Quoted 14 lines · Arnout Vandecappelle
Quoted 14 lines
Quoted 13 lines

# If target-parted can handle lvm volumes, then host-parted

# should be, too, so as to be able to generate them.
# If target-parted can't handle lvm volumes, there is no reason
# for host-aprted to handle them.
ifeq ($(BR2_PACKAGE_LVM2),y)
PARTED_DEPENDENCIES += lvm2
HOST_PARTED_DEPENDENCIES += lvm2
PARTED_CONF_OPT += --enable-device-mapper
HOST_PARTED_CONF_OPT += --enable-device-mapper
else
PARTED_CONF_OPT += --disable-device-mapper
HOST_PARTED_CONF_OPT += --disable-device-mapper
endif
Quoted 6 lines
Quoted 4 lines

While I do understand the logic behind what you're proposing, I'm not really comfortable with having the configuration of tools built for the host changed depending on the target configuration. It seems to be creating a bad precedent.

We already have a precedent: libxml2.

No, libxml2 is not the same case. libxml2 host configuration is adjusted according to the hidden BR2_PACKAGE_HOST_LIBXML2_PYTHON configuration option, which can be selected by another package.

Therefore, it is not directly the fact of changing the selection of target packages that modifies the configuration of host-libxml2.

Quoted 2 lines

What is so bad about one package's configuration depending on another
package configuration?

What I found bad here is that we're trying the configuration of the host package to the configuration of the corresponding target package, even though there is no real connection between the two things. But maybe I'm over-zealous I don't know.

Best regards,

Thomas

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

PATCH 3/3 package/parted: add a host variant

Quoted 23 lines · Thomas Petazzoni

Dear Arnout Vandecappelle,

Quoted 19 lines · Arnout Vandecappelle
Quoted 18 lines
Quoted 13 lines

# If target-parted can handle lvm volumes, then host-parted

# should be, too, so as to be able to generate them.
# If target-parted can't handle lvm volumes, there is no reason
# for host-aprted to handle them.
ifeq ($(BR2_PACKAGE_LVM2),y)
PARTED_DEPENDENCIES += lvm2
HOST_PARTED_DEPENDENCIES += lvm2
PARTED_CONF_OPT += --enable-device-mapper
HOST_PARTED_CONF_OPT += --enable-device-mapper
else
PARTED_CONF_OPT += --disable-device-mapper
HOST_PARTED_CONF_OPT += --disable-device-mapper
endif

While I do understand the logic behind what you're proposing, I'm not really comfortable with having the configuration of tools built for the host changed depending on the target configuration. It seems to be creating a bad precedent.

Quoted 5 lines
Quoted 1 line

We already have a precedent: libxml2.

No, libxml2 is not the same case. libxml2 host configuration is adjusted according to the hidden BR2_PACKAGE_HOST_LIBXML2_PYTHON configuration option, which can be selected by another package.

Yes, it is selected by the _target_ package mesa3d.

Quoted 10 lines

Therefore, it is not directly the fact of changing the selection of target packages that modifies the configuration of host-libxml2.

Quoted 2 lines

What is so bad about one package's configuration depending on another
package configuration?

What I found bad here is that we're trying the configuration of the host package to the configuration of the corresponding target package, even though there is no real connection between the two things. But

There you have a point: it's really an assumed connection, and one can
probably construct scenarios where you would want the lvm support on the host but not on the target or vice versa.

Well, let's solve that problem when it is relevant, i.e. there's an
actual use case.

Regards,
Arnout

Quoted 6 lines

maybe I'm over-zealous I don't know.

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