Added package v86d which provides a real-mode helper for uvesafb driver.
PATCH 1/2 Added package v86d which provides a real-mode helper for uvesafb driver.
--- package/Config.in | 1 + package/v86d/Config.in | 9 +++++++++ package/v86d/v86d.mk | 26 ++++++++++++++++++++++++++ 3 files changed, 36 insertions(+), 0 deletions(-) create mode 100644 package/v86d/Config.in create mode 100644 package/v86d/v86d.mk
package/Config.in +1 -0
@@ -230,6 +230,7 @@ source "package/unionfs/Config.in" source "package/usb_modeswitch/Config.in" source "package/usbmount/Config.in" source "package/usbutils/Config.in" +source "package/v86d/Config.in" source "package/wipe/Config.in" source "package/xfsprogs/Config.in" endmenu
package/v86d/Config.in new file +9 -0
@@ -0,0 +1,9 @@ +config BR2_PACKAGE_V86D + bool "v86d" + depends on (BR2_i386 || BR2_x86_64) + help + v86d is the userspace helper that runs x86 code in an emulated + environment. uvesafb will not work without v86d. v86d currently + supports the x86 and amd64 (x86-64) architectures. + + http://dev.gentoo.org/~spock/projects/uvesafb/
package/v86d/v86d.mk new file +26 -0
@@ -0,0 +1,26 @@ +############################################ +# +# v86d +# +############################################ + +V86D_VERSION = v86d-0.1.10 +V86D_SITE = git://repo.or.cz/v86d.git + +# It is necessary to define __i386__ explicitly for successful compilation. + +define V86D_BUILD_CMDS + $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) all +endef + +V86D_POST_CONFIGURE_HOOKS += V86D_REALLY_RUN_CONFIGURE + +# v86d's configure script is not autoconf-based: parameters +# supplied by Buildroot cause it to produce no result; explicit +# invocation is needed. + +define V86D_REALLY_RUN_CONFIGURE + (cd $(@D) ; ./configure --default) +endef + +$(eval $(call AUTOTARGETS))
Signature · 1 line
1.7.2.5
PATCH 2/2 Adjust v86d.mk per better source URL and build method.
* git source URL -> http source URL
* AUTOTARGETS -> GENTARGETS
--- package/v86d/v86d.mk | 23 +++++++++++------------ 1 files changed, 11 insertions(+), 12 deletions(-)
package/v86d/v86d.mk +11 -12
@@ -4,23 +4,22 @@ # ############################################ -V86D_VERSION = v86d-0.1.10 -V86D_SITE = git://repo.or.cz/v86d.git - -# It is necessary to define __i386__ explicitly for successful compilation. - -define V86D_BUILD_CMDS - $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) all -endef - -V86D_POST_CONFIGURE_HOOKS += V86D_REALLY_RUN_CONFIGURE +V86D_VERSION = 0.1.10 +V86D_SOURCE = v86d-$(V86D_VERSION).tar.bz2 +V86D_SITE = http://dev.gentoo.org/~spock/projects/uvesafb/archive # v86d's configure script is not autoconf-based: parameters # supplied by Buildroot cause it to produce no result; explicit # invocation is needed. -define V86D_REALLY_RUN_CONFIGURE +define V86D_CONFIGURE_CMDS (cd $(@D) ; ./configure --default) endef -$(eval $(call AUTOTARGETS)) +# It is necessary to define __i386__ explicitly for successful compilation. + +define V86D_BUILD_CMDS + $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) all +endef + +$(eval $(call GENTARGETS))
Signature · 1 line
1.7.2.5
PATCH 2/2 Adjust v86d.mk per better source URL and build method.
Hi Dmitry,
Please squash this patch into the previous one, and send a single fixed patch.
A couple more comments below.
Quoted 23 lines · Dmitry
* git source URL -> http source URL
* AUTOTARGETS -> GENTARGETSSigned-off-by Dmitry <[email protected]>--- package/v86d/v86d.mk | 23 +++++++++++------------ 1 files changed, 11 insertions(+), 12 deletions(-)package/v86d/v86d.mk +0
-7@@ -4,23 +4,22 @@ # ############################################ -V86D_VERSION = v86d-0.1.10 -V86D_SITE = git://repo.or.cz/v86d.git - -# It is necessary to define __i386__ explicitly for successful compilation. - -define V86D_BUILD_CMDS - $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) all
package/v86d/v86d.mk line 13
Quoted 10 lines
package/v86d/v86d.mk +3
-3-endef - -V86D_POST_CONFIGURE_HOOKS += V86D_REALLY_RUN_CONFIGURE +V86D_VERSION = 0.1.10 +V86D_SOURCE = v86d-$(V86D_VERSION).tar.bz2 +V86D_SITE = http://dev.gentoo.org/~spock/projects/uvesafb/archive # v86d's configure script is not autoconf-based: parameters # supplied by Buildroot cause it to produce no result; explicit # invocation is needed.
Since you don't use Buildroot's autoconf support, this comment is no longer relevant.
package/v86d/v86d.mk line 22
Quoted 11 lines
package/v86d/v86d.mk +5
-2-define V86D_REALLY_RUN_CONFIGURE +define V86D_CONFIGURE_CMDS (cd $(@D) ; ./configure --default) endef -$(eval $(call AUTOTARGETS)) +# It is necessary to define __i386__ explicitly for successful compilation. + +define V86D_BUILD_CMDS + $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) all
You can used $(TARGET_CONFIGURE_OPTS) here as well.
baruch
Quoted 3 lines
package/v86d/v86d.mk +3
-0+endef + +$(eval $(call GENTARGETS))
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 -PATCH 2/2 Adjust v86d.mk per better source URL and build method.
Baruch,
Quoted 6 lines · Baruch Siach
Quoted 3 lines
?# v86d's configure script is not autoconf-based: parameters ?# supplied by Buildroot cause it to produce no result; explicit ?# invocation is needed.
Since you don't use Buildroot's autoconf support, this comment is no longer relevant.
Well, I thought this would be the explanation why there is a configure script, yet it is a GENTARGET.
Quoted 5 lines
package/v86d/v86d.mk
Quoted 3 lines
package/v86d/v86d.mk +3
-0+ +define V86D_BUILD_CMDS + ? ? $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) allYou can used $(TARGET_CONFIGURE_OPTS) here as well.
I tried that, and it did not compile. -D__i386__ should be given explicitly to cc.
Thanks.
Signature · 3 lines
Dimitry Golubovsky Anywhere on the Web
PATCH 2/2 Adjust v86d.mk per better source URL and build method.
Hi Dimitry,
Quoted 11 lines · Dimitry Golubovsky
Quoted 6 lines · Baruch Siach
Quoted 3 lines
?# v86d's configure script is not autoconf-based: parameters ?# supplied by Buildroot cause it to produce no result; explicit ?# invocation is needed.
Since you don't use Buildroot's autoconf support, this comment is no longer relevant.
Well, I thought this would be the explanation why there is a configure script, yet it is a GENTARGET.
The comment needs to be reworded, then. Something like: "v86d's configure script is not autoconf-based, so we use GENTARGET".
Quoted 8 lines
Quoted 5 lines
package/v86d/v86d.mk
Quoted 3 lines
package/v86d/v86d.mk +3
-0+ +define V86D_BUILD_CMDS + ? ? $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) allYou can used $(TARGET_CONFIGURE_OPTS) here as well.
I tried that, and it did not compile. -D__i386__ should be given explicitly to cc.
Ah, you're right. I haven't noticed the quotes delimiting the CC variable.
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 -PATCH 2/2 Adjust v86d.mk per better source URL and build method.
In reply to Dimitry GolubovskyLe Mon, 7 May 2012 08:47:10 -0400,
Dimitry Golubovsky <[email protected]> a ?crit :
Quoted 8 lines
Quoted 5 lines
Quoted 3 lines
+define
V86D_BUILD_CMDS
+ ? ? $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) allYou can used $(TARGET_CONFIGURE_OPTS) here as well.
I tried that, and it did not compile. -D__i386__ should be given explicitly to cc.
This -D is rather a CFLAGS, so:
$(MAKE) $(TARGET_CONFIGURE_OPTS) CFLAGS="$(TARGET_CFLAGS) -D__i386__" -C $(@D) all
Thomas
Signature · 4 lines
Thomas Petazzoni, Free Electrons Kernel, drivers, real-time and embedded Linux development, consulting, training and support. http://free-electrons.com
PATCH 2/2 Adjust v86d.mk per better source URL and build method.
OK Thanks, I'll try this way.
Quoted 6 lines · Thomas Petazzoni
Quoted 2 lines
I tried that, and it did not compile. -D__i386__ should be given explicitly to cc.
This -D is rather a CFLAGS, so:
? ? ? ?$(MAKE) $(TARGET_CONFIGURE_OPTS) CFLAGS="$(TARGET_CFLAGS) -D__i386__" -C $(@D) all
Signature · 3 lines
Dimitry Golubovsky Anywhere on the Web
PATCH 2/2 Adjust v86d.mk per better source URL and build method.
In reply to Thomas PetazzoniHi Thomas,
Quoted 7 lines · Thomas Petazzoni
Le Mon, 7 May 2012 08:47:10 -0400,
Dimitry Golubovsky <[email protected]> a ?crit :Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
+define
V86D_BUILD_CMDS
+ ? ? $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) all
Quoted 8 lines
Quoted 4 lines
Quoted 1 line
You can used $(TARGET_CONFIGURE_OPTS) here as well.
I tried that, and it did not compile. -D__i386__ should be given explicitly to cc.
This -D is rather a CFLAGS, so:
$(MAKE) $(TARGET_CONFIGURE_OPTS) CFLAGS="$(TARGET_CFLAGS) -D__i386__" -C $(@D) all
But then you'll have CFLAGS defined twice. Is this OK?
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 -PATCH 2/2 Adjust v86d.mk per better source URL and build method.
Le Mon, 7 May 2012 16:55:19 +0300,
Baruch Siach <[email protected]> a ?crit :
Quoted 4 lines
Quoted 2 lines
$(MAKE) $(TARGET_CONFIGURE_OPTS) CFLAGS="$(TARGET_CFLAGS)-D__i386__" -C $(@D) all
But then you'll have CFLAGS defined twice. Is this OK?
You'll have a first definition of CFLAGS in TARGET_CONFIGURE_OPTS, overriden by the local redefinition. We do this in numerous places in Buildroot already.
Thomas
Signature · 4 lines
Thomas Petazzoni, Free Electrons Kernel, drivers, real-time and embedded Linux development, consulting, training and support. http://free-electrons.com
PATCH 2/2 Adjust v86d.mk per better source URL and build method.
Hi Thomas,
Quoted 11 lines · Thomas Petazzoni
Le Mon, 7 May 2012 16:55:19 +0300,
Baruch Siach <[email protected]> a ?crit :Quoted 4 lines
Quoted 2 lines
$(MAKE) $(TARGET_CONFIGURE_OPTS) CFLAGS="$(TARGET_CFLAGS)-D__i386__" -C $(@D) all
But then you'll have CFLAGS defined twice. Is this OK?
You'll have a first definition of CFLAGS in
TARGET_CONFIGURE_OPTS, overriden by the local redefinition. We do this in numerous places in Buildroot already.
Is this make behaviour documented? I couldn't find any mention of this in the manual.
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 -PATCH 2/2 Adjust v86d.mk per better source URL and build method.
In reply to Thomas PetazzoniHi,
Quoted 7 lines · Thomas Petazzoni
Le Mon, 7 May 2012 08:47:10 -0400,
Dimitry Golubovsky <[email protected]> a ?crit :Quoted 4 lines
Quoted 4 lines
Quoted 3 lines
+define
V86D_BUILD_CMDS
+ ? ? $(MAKE) CC="$(TARGET_CC) -D__i386__" LD="$(TARGET_LD)" -C $(@D) all
Quoted 9 lines
Quoted 4 lines
Quoted 1 line
You can used $(TARGET_CONFIGURE_OPTS) here as well.
I tried that, and it did not compile. -D__i386__ should be given explicitly to cc.
This -D is rather a CFLAGS, so:
? ? ? ?$(MAKE) $(TARGET_CONFIGURE_OPTS) CFLAGS="$(TARGET_CFLAGS) -D__i386__" -C $(@D) all
No, it does not work this way. Value of $(CC) with -D__i386__ should be passed deeper down to the lowest level Makefile in order for v86d to compile.
/40G2/golubovsky/src/buildroot-patching/buildroot/output/host/usr/bin/i586-unknown-linux-uclibc-gcc -D__i386__ -c -pipe -Os -o fpu.o fpu.c
In file included from decode.c:41:0:x86emu/x86emui.h:64:20: fatal error: x86emu.h: No such file or directory
(the way you proposed) does not work as local flags are lost.
But:
/40G2/golubovsky/src/buildroot-patching/buildroot/output/host/usr/bin/i586-unknown-linux-uclibc-gcc -D__i386__ -c -I. -I../../include -I../../include/x86emu -o decode.odecode.cv86_x86emu.c: In function 'v86_init':
v86_x86emu.c:50:3: warning: initialization from incompatible pointer type v86_x86emu.c:53:3: warning: initialization from incompatible pointer type v86_x86emu.c:57:3: warning: initialization from incompatible pointer type v86_x86emu.c:58:3: warning: initialization from incompatible pointer type v86_x86emu.c:59:3: warning: initialization from incompatible pointer type v86_x86emu.c:60:3: warning: initialization from incompatible pointer type v86_x86emu.c:61:3: warning: initialization from incompatible pointer type v86_x86emu.c:62:3: warning: initialization from incompatible pointer type
(my way) works as it is only $CC that is needed by the lower level Makefile.
I'll make other changes as recommended.
Thanks.
Signature · 3 lines
Dmitry Golubovsky Anywhere on the Web