linux-backports (new package) and misc related fixes (branch yem/linux-backports)

Hello All!

This series introduces linux-backports, which contains drivers from recent kernels, backported unto older kernels (down to 3.0).

The series starts by fixing a shortcoming in the pkg-kconfig infra, then does a bit of eye-candy in the Linux package, and finally introduces linux-backports.

The thirs patch, linux-backports itself, uses a pure generic package, and does not make use of the kernel-module infra, while the fourth and latest patch converts it over to the kernel-module infra.

I'm not sure using the kernel-module infra is worth it, because it does not gain us much in this case. Yet, it makes for a slightly more beautiful .mk (IMHO!), even if a bit longer (+4 lines).

In the end, using it or not is the same to me (with a very slight preference for using it, of course, but I won't mind either way).

Thanks go to Petr for doing the initial packaging of linux-backports, and Arnout for suggesting the fix in patch 1. :-)

Regards,
Yann E. MORIN.

The following changes since commit 1f640a4e7a42667a88101f6ceb787fbb118a6acf:

zedboard: Switch to U-Boot 2015.07 (2015-07-22 14:31:55 +0200)

are available in the git repository at:

git://git.busybox.net/~ymorin/git/buildroot yem/linux-backports

for you to fetch changes up to 200f17ef13ac12610b33139a1b1f8820fc0c85bd:

package/linux-backports: switch to using the kernel-module infra (2015-07-22 16:44:57 +0200)

----------------------------------------------------------------
Petr Vorel (1):
package/linux-backports: new package

Yann E. MORIN (3):
core/pkg-kconfig: do not override @D
package/linux: trash stderr in LINUX_VERSION_PROBED package/linux-backports: switch to using the kernel-module infra

 linux/linux.mk                               |  2 +-
 package/Config.in                            |  1 +
 package/linux-backports/Config.in            | 52 ++++++++++++++++++
 package/linux-backports/linux-backports.hash |  2 +
 package/linux-backports/linux-backports.mk   | 81 ++++++++++++++++++++++++++++
 package/pkg-kconfig.mk                       | 16 +++---
 6 files changed, 146 insertions(+), 8 deletions(-)
 create mode 100644 package/linux-backports/Config.in
 create mode 100644 package/linux-backports/linux-backports.hash
 create mode 100644 package/linux-backports/linux-backports.mk

.-----------------.--------------------.------------------.--------------------.
|  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.  |
'------------------------------^-------^------------------^--------------------'

Currently, we override @D for the kconfig configurators, so the fixup commands can use $(@D); otherwise @D would be simply '.' because it is not a real file in the package build dir.

However, this breaks the soon-to-be-introduced linux-backports package, which needs to have a dependency on the linux package to be configured.

The underlying reason is that @D is an automatic variable that is always set by make to the directory part of the target of the rule. However, automatic variables loose their "automatic" property when they are manually set.

Furthermore, a variable that is defined for a rule is inherited by all dependencies of that rule, so our manually-set @D is inherited all the way down the dependency chain of linux-backports, down to the linux' own .config rule, which is thus run with @D pointing to linux-backports' build dir, not linux'.

Fix that by using a "static pattern rule", redirecting the configurators to an intermediate stamp-like file which path is in the package build dir, so we get a valid @D from the onset, without having to manually fiddle with it.

Thanks to Arnout for suggesting that in the first place. Sorry I did reject it as "too complex" when it was in fact the best solution.

---
 package/pkg-kconfig.mk | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

index 3e17a70..375607f 100644
@@ -104,15 +104,16 @@ endif
 # that is clean wrt. our requirements.
 #
 # Because commands in $(1)_FIXUP_KCONFIG are probably using $(@D), we
-# fake it for the configurators (otherwise it is set to just '.', i.e.
-# the current directory where make is run, which happens to be in
-# $(TOPDIR), because the target of the rule is not an actual file, so
-# does not have any path component).
+# need to have a valid @D set. But, because the configurators rules are
+# not real files and do not contain the path to the package build dir,
+# @D would be just '.' in this case. So, we use an intermediate rule
+# with a stamp-like file which path is in the package build dir, so we
+# end up having a valid @D.
 #
-$$(addprefix $(1)-,$$($(2)_KCONFIG_EDITORS)): @D=$$($(2)_DIR)
-$$(addprefix $(1)-,$$($(2)_KCONFIG_EDITORS)): $$($(2)_DIR)/.stamp_kconfig_fixup_done
+$$(addprefix $(1)-,$$($(2)_KCONFIG_EDITORS)): $(1)-%: $$($(2)_DIR)/.kconfig_editor_%
+$$($(2)_DIR)/.kconfig_editor_%: $$($(2)_DIR)/.stamp_kconfig_fixup_done
 	$$($(2)_MAKE_ENV) $$(MAKE) -C $$($(2)_DIR) \
-		$$($(2)_KCONFIG_OPTS) $$(subst $(1)-,,$$@)
+		$$($(2)_KCONFIG_OPTS) $$(*)
 	rm -f $$($(2)_DIR)/.stamp_{kconfig_fixup_done,configured,built}
 	rm -f $$($(2)_DIR)/.stamp_{target,staging,images}_installed
 	$$(call $(2)_FIXUP_DOT_CONFIG)

@@ -169,6 +170,7 @@ endif # package enabled
 	$(1)-update-defconfig \
 	$(1)-savedefconfig \
 	$(1)-check-configuration-done \
+	$$($(2)_DIR)/.kconfig_editor_% \
 	$$(addprefix $(1)-,$$($(2)_KCONFIG_EDITORS))
 
 endef # inner-kconfig-package

1.9.1

The Linux buildsystem tries to run the compiler even just for 'kernelrelease' (which we store in LINUX_VERSION_PROBED) and we sometimes need to use it before the toolchain is available; thus we get spurious errors on stderr.

Consign stderr to oblivion when computing the 'kernelrelease'.

---
 linux/linux.mk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 375a9e1..d91dbb2 100644
@@ -76,7 +76,7 @@ LINUX_MAKE_ENV = \
 
 # Get the real Linux version, which tells us where kernel modules are
 # going to be installed in the target filesystem.
-LINUX_VERSION_PROBED = `$(MAKE) $(LINUX_MAKE_FLAGS) -C $(LINUX_DIR) --no-print-directory -s kernelrelease`
+LINUX_VERSION_PROBED = `$(MAKE) $(LINUX_MAKE_FLAGS) -C $(LINUX_DIR) --no-print-directory -s kernelrelease 2>/dev/null`
 
 ifeq ($(BR2_LINUX_KERNEL_USE_INTREE_DTS),y)
 KERNEL_DTS_NAME = $(call qstrip,$(BR2_LINUX_KERNEL_INTREE_DTS_NAME))

1.9.1

From: Petr Vorel <[email protected]>

https://backports.wiki.kernel.org

---
Changes v7 -> v8: (Yann)
- fix option ordering and prompts in Config.in
- drop editors, only menuconfig is available
- add licensing information
- bump to 4.1.1-1, drop 4.0.1-1 patch
- move custom error message before call to kconfig-package
- use $(BR_BUILDING) instead of custom, wrong test on 'make source'
- drop intermediate variable to set LINUX_BACKPORTS_KCONFIG_FILE
- qstrip all variables from Config.in
- fix dependencies (tricky, needs kernel to be configured)
- check kernel is >= 3.0

Changes v6->v7:
* Bump version to lastest stable 4.0.1-1.
* Update to current version of kconfig-package.
* Add option to specify config fragments.
* Include upstream patch to fix build error.

Changes v5->v6:
Added changes by Arnout Vandecappelle in Mon Apr 20 21:46:34 UTC 2015:
* linux-backports.mk: $(LINUX_BACKPORTS_KCONFIG_FILE): linux
this doesn't help: we'd need to use:
LINUX_BACKPORTS_KCONFIG_FILE := $(LINUX_BACKPORTS_DIR)/defconfigs/$(call qstrip,$(BR2_PACKAGE_LINUX_BACKPORTS_DEFCONFIG)) but then we don't get value for $(LINUX_BACKPORTS_DIR) nor $(@D).
* linux-backports.mk: LINUX_BACKPORTS_PATCH_DEPENDENCIES = linux
This also does not help, it does linux extracting, patching, but we
need also configuring. We'd need something like
LINUX_BACKPORTS_CONFIGURE_DEPENDENCIES to be implemented.
* Config.in: info about support from kernel 3.0.
* Config.in: help text wrapped to 72 columns.

Changes v4->v5:
* Added changes made by Thomas Petazzoni in Sun, 19 Apr 2015 11:03:26 +0200.

Changes v3->v4:
* Define LINUX_BACKPORTS_MAKE_ENV instead of LINUX_BACKPORTS_MAKE_OPTS, remove TARGET_MAKE_ENV from it.
* Add nconfig to LINUX_BACKPORTS_KCONFIG_EDITORS.
* Clean formating of build commands.

---
 package/Config.in                            |  1 +
 package/linux-backports/Config.in            | 52 +++++++++++++++++
 package/linux-backports/linux-backports.hash |  2 +
 package/linux-backports/linux-backports.mk   | 85 ++++++++++++++++++++++++++++
 4 files changed, 140 insertions(+)
 create mode 100644 package/linux-backports/Config.in
 create mode 100644 package/linux-backports/linux-backports.hash
 create mode 100644 package/linux-backports/linux-backports.mk

index 9942e3a..db84cd7 100644
@@ -372,6 +372,7 @@ endif
 	source "package/kbd/Config.in"
 	source "package/lcdproc/Config.in"
 	source "package/libump/Config.in"
+	source "package/linux-backports/Config.in"
 	source "package/lirc-tools/Config.in"
 	source "package/lm-sensors/Config.in"
 	source "package/lshw/Config.in"

new file mode 100644
index 0000000..0d41067
@@ -0,0 +1,2 @@
+# From: https://www.kernel.org/pub/linux/kernel/projects/backports/stable/v4.1.1/sha256sums.asc
+sha256 7fca160665b801796ce50def18f2fd6def1c4452290e93ec5332444fb2021bd6  backports-4.1.1-1.tar.xz

1.9.1

Dear Yann E. MORIN,

Doesn't this fail really badly if:

LINUX_OVERRIDE_SRCDIR = <some value>

?

In this case, $(LINUX_DIR) is $(BUILD_DIR)/linux-custom/ and not $(BUILD_DIR)/linux-$(call qstrip,$(BR2_LINUX_KERNEL_VERSION)).

Argh :-/

Thomas

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

---
 package/linux-backports/linux-backports.mk | 32 +++++++++++++-----------------
 1 file changed, 14 insertions(+), 18 deletions(-)

index 790d7e4..7c10270 100644
@@ -11,33 +11,28 @@ LINUX_BACKPORTS_SITE = $(BR2_KERNEL_MIRROR)/linux/kernel/projects/backports/stab
 LINUX_BACKPORTS_LICENSE = GPLv2
 LINUX_BACKPORTS_LICENSE_FILES = COPYING
 
-LINUX_BACKPORTS_DEPENDENCIES = linux
-
-LINUX_BACKPORTS_MAKE_OPTS = \
-	$(LINUX_MAKE_FLAGS) \
-	KLIB_BUILD=$(LINUX_DIR) \
-	KLIB=$(TARGET_DIR)
-
 ifeq ($(BR2_PACKAGE_LINUX_BACKPORTS_USE_DEFCONFIG),y)
 LINUX_BACKPORTS_KCONFIG_FILE = $(LINUX_BACKPORTS_DIR)/defconfigs/$(call qstrip,$(BR2_PACKAGE_LINUX_BACKPORTS_DEFCONFIG))
 else ifeq ($(BR2_PACKAGE_LINUX_BACKPORTS_USE_CUSTOM_CONFIG),y)
 LINUX_BACKPORTS_KCONFIG_FILE = $(call qstrip,$(BR2_PACKAGE_LINUX_BACKPORTS_CUSTOM_CONFIG_FILE))
 endif
 
-define LINUX_BACKPORTS_BUILD_CMDS
-	$(TARGET_MAKE_ENV) $(MAKE) $(LINUX_BACKPORTS_MAKE_OPTS) -C $(@D)
-endef
-
-define LINUX_BACKPORTS_INSTALL_TARGET_CMDS
-	$(TARGET_MAKE_ENV) $(MAKE) $(LINUX_BACKPORTS_MAKE_OPTS) \
-		-C $(LINUX_DIR) M=$(@D) \
-		INSTALL_MOD_DIR=backports \
-		modules_install
-endef
-
 LINUX_BACKPORTS_KCONFIG_FRAGMENT_FILES = $(call qstrip,$(BR2_PACKAGE_LINUX_BACKPORTS_CONFIG_FRAGMENT_FILES))
 LINUX_BACKPORTS_KCONFIG_OPTS = $(LINUX_BACKPORTS_MAKE_OPTS)
 
+LINUX_BACKPORTS_MAKE_OPTS = \
+	BACKPORT_DIR=$(@D) \
+	KLIB_BUILD=$(LINUX_DIR) \
+	KLIB=$(TARGET_DIR)/lib/modules/$(LINUX_VERSION_PROBED) \
+	`sed -r -e '/^\#/d;' $(@D)/.config`
+
+LINUX_BACKPORTS_MODULE_MAKE_OPTS = $(LINUX_BACKPORTS_MAKE_OPTS)
+
+define LINUX_BACKPORTS_AUTOCONF
+	$(MAKE) -C $(@D) $(LINUX_BACKPORTS_MAKE_OPTS) backport-include/backport/autoconf.h
+endef
+LINUX_BACKPORTS_POST_CONFIGURE_HOOKS += LINUX_BACKPORTS_AUTOCONF
+
 # Checks to give errors that the user can understand
 ifeq ($(BR_BUILDING),y)
 

@@ -55,6 +50,7 @@ endif
 
 endif # BR_BUILDING
 
+$(eval $(kernel-module))
 $(eval $(kconfig-package))
 
 # linux-backports' own .config file needs options from the kernel's own

1.9.1

Dear Yann E. MORIN,

What is this useful for?

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

Dear Yann E. MORIN,

I've applied those two preparations patches.

Thanks!

Thomas

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