[Buildroot] [PATCH 1/3] mtd: fix build issues on older hosts

[Buildroot] [PATCH 1/3] mtd: fix build issues on older hosts

[Thomas Petazzoni]

Building the MTD test programs requires the MS_DIRSYNC, which is not
necessarily available on old build machines. But obviously, MTD test
programs are not needed, so we can simply disable them, as they were
prior to the migration to mtd 2.0.

Fixes:

  http://autobuild.buildroot.net/results/21e1ad2a4560b6d3ba6490d20ae064246e66d5c1/

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/mtd/mtd.mk | 1 +
 1 file changed, 1 insertion(+)

diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
index 7c3a2b6..043f1c7 100644
--- a/package/mtd/mtd.mk
+++ b/package/mtd/mtd.mk
@@ -45,6 +45,7 @@ MTD_CONF_OPTS += --without-xattr
 endif
 
 HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-e2fsprogs
+HOST_MTD_CONF_OPTS = --disable-tests
 
 MKFS_JFFS2 = $(HOST_DIR)/sbin/mkfs.jffs2
 SUMTOOL = $(HOST_DIR)/sbin/sumtool
-- 
2.9.4

[Buildroot] [PATCH 2/3] mtd: host variant needs host-util-linux, host-e2fsprogs

[Thomas Petazzoni]

Since mtd was converted to the package infrastructure in commit
de4cf4e9135e198d4c3beefc8ad63c03870eb78a ("mtd: convert to gentargets,
add host package"), its host variant depended on host-e2fsprogs. At
the time, only a host variant of the mtd package was available.

When a target variant of mtd was introduced in commit
b50e0fa113bf641a3764ae99b94bb7ba4e1e8f85 ("mtd: add option to build
mkfs.ubifs for target"), it depended on util-linux.

So today, the target variant continues to depend on util-linux, while
the target variant depends on e2fsprogs. What mkfs.ubifs really needs
is libuuid, which is provided by util-linux. It was in fact provided
by the fact that host-e2fsprogs depends on host-util-linux.

But really, host-e2fsprogs is not needed as a dependency, so use
host-util-linux to be consistent with the target variant.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/mtd/mtd.mk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
index 043f1c7..3bac2d5 100644
--- a/package/mtd/mtd.mk
+++ b/package/mtd/mtd.mk
@@ -44,7 +44,7 @@ else
 MTD_CONF_OPTS += --without-xattr
 endif
 
-HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-e2fsprogs
+HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-util-linux
 HOST_MTD_CONF_OPTS = --disable-tests
 
 MKFS_JFFS2 = $(HOST_DIR)/sbin/mkfs.jffs2
-- 
2.9.4

[Buildroot] [PATCH 3/3] mtd: be explicit about config options for the host

[Thomas Petazzoni]

Instead of letting auto-detection do its job, be explicit about the
fact that we want the JFFS2 and UBIFS utilities when building the host
variant of mtd.

Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
---
 package/mtd/mtd.mk | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
index 3bac2d5..807f559 100644
--- a/package/mtd/mtd.mk
+++ b/package/mtd/mtd.mk
@@ -45,7 +45,10 @@ MTD_CONF_OPTS += --without-xattr
 endif
 
 HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-util-linux
-HOST_MTD_CONF_OPTS = --disable-tests
+HOST_MTD_CONF_OPTS = \
+	--with-jffs \
+	--with-ubifs \
+	--disable-tests
 
 MKFS_JFFS2 = $(HOST_DIR)/sbin/mkfs.jffs2
 SUMTOOL = $(HOST_DIR)/sbin/sumtool
-- 
2.9.4

[Buildroot] [PATCH 1/3] mtd: fix build issues on older hosts

[Arnout Vandecappelle]

On 06-07-17 22:58, Thomas Petazzoni wrote:
> Building the MTD test programs requires the MS_DIRSYNC, which is not
> necessarily available on old build machines. But obviously, MTD test
> programs are not needed, so we can simply disable them, as they were
> prior to the migration to mtd 2.0.
> 
> Fixes:
> 
>   http://autobuild.buildroot.net/results/21e1ad2a4560b6d3ba6490d20ae064246e66d5c1/
> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>


 Regards,
 Arnout


> ---
>  package/mtd/mtd.mk | 1 +
>  1 file changed, 1 insertion(+)
> 
> diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
> index 7c3a2b6..043f1c7 100644
> --- a/package/mtd/mtd.mk
> +++ b/package/mtd/mtd.mk
> @@ -45,6 +45,7 @@ MTD_CONF_OPTS += --without-xattr
>  endif
>  
>  HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-e2fsprogs
> +HOST_MTD_CONF_OPTS = --disable-tests
>  
>  MKFS_JFFS2 = $(HOST_DIR)/sbin/mkfs.jffs2
>  SUMTOOL = $(HOST_DIR)/sbin/sumtool
> 

-- 
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:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

[Buildroot] [PATCH 2/3] mtd: host variant needs host-util-linux, host-e2fsprogs

[Arnout Vandecappelle]

On 06-07-17 22:58, Thomas Petazzoni wrote:
> Since mtd was converted to the package infrastructure in commit
> de4cf4e9135e198d4c3beefc8ad63c03870eb78a ("mtd: convert to gentargets,
> add host package"), its host variant depended on host-e2fsprogs. At
> the time, only a host variant of the mtd package was available.
> 
> When a target variant of mtd was introduced in commit
> b50e0fa113bf641a3764ae99b94bb7ba4e1e8f85 ("mtd: add option to build
> mkfs.ubifs for target"), it depended on util-linux.
> 
> So today, the target variant continues to depend on util-linux, while
> the target variant depends on e2fsprogs. What mkfs.ubifs really needs
> is libuuid, which is provided by util-linux. It was in fact provided
> by the fact that host-e2fsprogs depends on host-util-linux.
> 
> But really, host-e2fsprogs is not needed as a dependency, so use
> host-util-linux to be consistent with the target variant.
> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>

Reviewed-by: Arnout Vandecappelle (Essensium/Mind) <arnout@mind.be>

 Excellent commit message!

 Regards,
 Arnout

> ---
>  package/mtd/mtd.mk | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
> index 043f1c7..3bac2d5 100644
> --- a/package/mtd/mtd.mk
> +++ b/package/mtd/mtd.mk
> @@ -44,7 +44,7 @@ else
>  MTD_CONF_OPTS += --without-xattr
>  endif
>  
> -HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-e2fsprogs
> +HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-util-linux
>  HOST_MTD_CONF_OPTS = --disable-tests
>  
>  MKFS_JFFS2 = $(HOST_DIR)/sbin/mkfs.jffs2
> 

-- 
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:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

[Buildroot] [PATCH 3/3] mtd: be explicit about config options for the host

[Arnout Vandecappelle]

On 06-07-17 22:58, Thomas Petazzoni wrote:
> Instead of letting auto-detection do its job, be explicit about the
> fact that we want the JFFS2 and UBIFS utilities when building the host
> variant of mtd.
> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> ---
>  package/mtd/mtd.mk | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
> index 3bac2d5..807f559 100644
> --- a/package/mtd/mtd.mk
> +++ b/package/mtd/mtd.mk
> @@ -45,7 +45,10 @@ MTD_CONF_OPTS += --without-xattr
>  endif
>  
>  HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-util-linux
> -HOST_MTD_CONF_OPTS = --disable-tests
> +HOST_MTD_CONF_OPTS = \
> +	--with-jffs \
> +	--with-ubifs \

 If you do this, it would also seem natural to add

--without-xattr --with-lzo

(although we probably want xattr support in ubifs images, at least if
BR2_ROOTFS_DEVICE_TABLE_SUPPORTS_EXTENDED_ATTRIBUTES is set, but that's for
another day :-).

 Regards,
 Arnout

> +	--disable-tests
>  
>  MKFS_JFFS2 = $(HOST_DIR)/sbin/mkfs.jffs2
>  SUMTOOL = $(HOST_DIR)/sbin/sumtool
> 

-- 
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:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

[Buildroot] [PATCH 2/3] mtd: host variant needs host-util-linux, host-e2fsprogs

[Baruch Siach]

Hi Thomas,

On Thu, Jul 06, 2017 at 10:58:48PM +0200, Thomas Petazzoni wrote:
> Since mtd was converted to the package infrastructure in commit
> de4cf4e9135e198d4c3beefc8ad63c03870eb78a ("mtd: convert to gentargets,
> add host package"), its host variant depended on host-e2fsprogs. At
> the time, only a host variant of the mtd package was available.
> 
> When a target variant of mtd was introduced in commit
> b50e0fa113bf641a3764ae99b94bb7ba4e1e8f85 ("mtd: add option to build
> mkfs.ubifs for target"), it depended on util-linux.
> 
> So today, the target variant continues to depend on util-linux, while
> the target variant depends on e2fsprogs. What mkfs.ubifs really needs
> is libuuid, which is provided by util-linux. It was in fact provided
> by the fact that host-e2fsprogs depends on host-util-linux.
>
> But really, host-e2fsprogs is not needed as a dependency, so use
> host-util-linux to be consistent with the target variant.

Missing 'not' in the subject line.

baruch

> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> ---
>  package/mtd/mtd.mk | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
> index 043f1c7..3bac2d5 100644
> --- a/package/mtd/mtd.mk
> +++ b/package/mtd/mtd.mk
> @@ -44,7 +44,7 @@ else
>  MTD_CONF_OPTS += --without-xattr
>  endif
>  
> -HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-e2fsprogs
> +HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-util-linux
>  HOST_MTD_CONF_OPTS = --disable-tests
>  
>  MKFS_JFFS2 = $(HOST_DIR)/sbin/mkfs.jffs2

-- 
     http://baruch.siach.name/blog/                  ~. .~   Tk Open Systems
=}------------------------------------------------ooO--U--Ooo------------{=
   - baruch at tkos.co.il - tel: +972.52.368.4656, http://www.tkos.co.il -

[Buildroot] [PATCH 3/3] mtd: be explicit about config options for the host

[Thomas Petazzoni]

Hello,

On Fri, 7 Jul 2017 00:57:59 +0200, Arnout Vandecappelle wrote:
> On 06-07-17 22:58, Thomas Petazzoni wrote:
> > Instead of letting auto-detection do its job, be explicit about the
> > fact that we want the JFFS2 and UBIFS utilities when building the host
> > variant of mtd.
> > 
> > Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> > ---
> >  package/mtd/mtd.mk | 5 ++++-
> >  1 file changed, 4 insertions(+), 1 deletion(-)
> > 
> > diff --git a/package/mtd/mtd.mk b/package/mtd/mtd.mk
> > index 3bac2d5..807f559 100644
> > --- a/package/mtd/mtd.mk
> > +++ b/package/mtd/mtd.mk
> > @@ -45,7 +45,10 @@ MTD_CONF_OPTS += --without-xattr
> >  endif
> >  
> >  HOST_MTD_DEPENDENCIES = host-zlib host-lzo host-util-linux
> > -HOST_MTD_CONF_OPTS = --disable-tests
> > +HOST_MTD_CONF_OPTS = \
> > +	--with-jffs \
> > +	--with-ubifs \  
> 
>  If you do this, it would also seem natural to add
> 
> --without-xattr --with-lzo

Indeed for --without-xattr. However, we don't pass
--with-lzo/--without-lzo for the target variant. Perhaps we should as
well.

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com

[Buildroot] [PATCH 1/3] mtd: fix build issues on older hosts

[Thomas Petazzoni]

Hello,

On Thu,  6 Jul 2017 22:58:47 +0200, Thomas Petazzoni wrote:
> Building the MTD test programs requires the MS_DIRSYNC, which is not
> necessarily available on old build machines. But obviously, MTD test
> programs are not needed, so we can simply disable them, as they were
> prior to the migration to mtd 2.0.
> 
> Fixes:
> 
>   http://autobuild.buildroot.net/results/21e1ad2a4560b6d3ba6490d20ae064246e66d5c1/
> 
> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
> ---
>  package/mtd/mtd.mk | 1 +
>  1 file changed, 1 insertion(+)

Patches 1 and 2 applied. Thanks Arnout for the review.

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux, Kernel and Android engineering
http://free-electrons.com

[Buildroot] [PATCH 1/3] mtd: fix build issues on older hosts

[Arnout Vandecappelle]

On 07-07-17 10:02, Thomas Petazzoni wrote:
> Hello,
> 
> On Thu,  6 Jul 2017 22:58:47 +0200, Thomas Petazzoni wrote:
>> Building the MTD test programs requires the MS_DIRSYNC, which is not
>> necessarily available on old build machines. But obviously, MTD test
>> programs are not needed, so we can simply disable them, as they were
>> prior to the migration to mtd 2.0.
>>
>> Fixes:
>>
>>   http://autobuild.buildroot.net/results/21e1ad2a4560b6d3ba6490d20ae064246e66d5c1/
>>
>> Signed-off-by: Thomas Petazzoni <thomas.petazzoni@free-electrons.com>
>> ---
>>  package/mtd/mtd.mk | 1 +
>>  1 file changed, 1 insertion(+)
> 
> Patches 1 and 2 applied. Thanks Arnout for the review.

 Unfortunately you missed Baruch's comment...

 Regards,
 Arnout

-- 
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:  7493 020B C7E3 8618 8DEC 222C 82EB F404 F9AC 0DDF

[Buildroot] [PATCH 1/3] mtd: fix build issues on older hosts

[Thomas Petazzoni]

Hello,

On Fri, 7 Jul 2017 10:21:36 +0200, Arnout Vandecappelle wrote:

> > Patches 1 and 2 applied. Thanks Arnout for the review.  
> 
>  Unfortunately you missed Baruch's comment...

Dammit :-/

And still, I did notice and fixed a mistake that neither you nor Baruch
spotted:

"""
So today, the target variant continues to depend on util-linux, while
the target variant depends on e2fsprogs. 
"""

Obviously the second "target variant" should have been "host variant",
and I fixed that before pushing.

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com