Patchwork [U-Boot] fdt: Ensure that libfdt_env.h comes from U-Boot

login
register
mail settings
Submitter Simon Glass
Date March 11, 2013, 4:30 p.m.
Message ID <1363019427-31520-1-git-send-email-sjg@chromium.org>
Download mbox | patch
Permalink /patch/226608/
State Accepted, archived
Delegated to: Tom Rini
Headers show

Comments

Simon Glass - March 11, 2013, 4:30 p.m.
When building host utilities, we include libfdt.h from the host, not from
U-Boot. This in turn brings in libfdt_env.h from the host, which can mess
up the types and cause a build failure, depending on the host environment.
To fix this, force inclusion of U-Boot's libfdt_env.h so that the types
are correct.

Another way to fix this is to use -nostdinc and -idirafter to ensure that
system includes are included after U-Boot ones. Unfortunately this means
that U-Boot's errno.h gets included instead of the system one. This in
turn requires a hack to errno.h to redirect things, so all in all the
solution in this patch is probably cleaner.

Signed-off-by: Simon Glass <sjg@chromium.org>
---
 tools/Makefile | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
Simon Glass - April 12, 2013, 8:24 p.m.
Hi Tom / Jerry,

On Mon, Mar 11, 2013 at 9:30 AM, Simon Glass <sjg@chromium.org> wrote:
> When building host utilities, we include libfdt.h from the host, not from
> U-Boot. This in turn brings in libfdt_env.h from the host, which can mess
> up the types and cause a build failure, depending on the host environment.
> To fix this, force inclusion of U-Boot's libfdt_env.h so that the types
> are correct.
>
> Another way to fix this is to use -nostdinc and -idirafter to ensure that
> system includes are included after U-Boot ones. Unfortunately this means
> that U-Boot's errno.h gets included instead of the system one. This in
> turn requires a hack to errno.h to redirect things, so all in all the
> solution in this patch is probably cleaner.
>
> Signed-off-by: Simon Glass <sjg@chromium.org>
> ---
>  tools/Makefile | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/tools/Makefile b/tools/Makefile
> index c5952fc..889c897 100644
> --- a/tools/Makefile
> +++ b/tools/Makefile
> @@ -164,7 +164,8 @@ NOPEDOBJS := $(addprefix $(obj),$(NOPED_OBJ_FILES-y))
>  # Use native tools and options
>  # Define __KERNEL_STRICT_NAMES to prevent typedef overlaps
>  #
> -HOSTCPPFLAGS = -idirafter $(SRCTREE)/include \
> +HOSTCPPFLAGS = -include $(SRCTREE)/include/libfdt_env.h \
> +               -idirafter $(SRCTREE)/include \
>                 -idirafter $(OBJTREE)/include2 \
>                 -idirafter $(OBJTREE)/include \
>                 -I $(SRCTREE)/lib/libfdt \
> --
> 1.8.1.3
>

This is a very late patch, and I'm not 100% sure this is the best way
to fix it. But with the upstream dtc package we get these errors
without it:

Configuring for smdk5250 board...
In file included from fdt_host.h:24:0,
                 from mkimage.h:35,
                 from aisimage.c:27:
../include/libfdt.h: In function ‘fdt_setprop_inplace_u32’:
../include/libfdt.h:885:2: error: unknown type name ‘fdt32_t’
../include/libfdt.h: In function ‘fdt_setprop_inplace_u64’:
../include/libfdt.h:920:2: error: unknown type name ‘fdt64_t’
../include/libfdt.h: In function ‘fdt_property_u32’:
../include/libfdt.h:996:2: error: unknown type name ‘fdt32_t’
../include/libfdt.h: In function ‘fdt_property_u64’:
../include/libfdt.h:1001:2: error: unknown type name ‘fdt64_t’
../include/libfdt.h: In function ‘fdt_setprop_u32’:
../include/libfdt.h:1157:2: error: unknown type name ‘fdt32_t’
../include/libfdt.h: In function ‘fdt_setprop_u64’:
../include/libfdt.h:1192:2: error: unknown type name ‘fdt64_t’
../include/libfdt.h: In function ‘fdt_appendprop_u32’:
../include/libfdt.h:1299:2: error: unknown type name ‘fdt32_t’
../include/libfdt.h: In function ‘fdt_appendprop_u64’:
../include/libfdt.h:1334:2: error: unknown type name ‘fdt64_t’

Can this be considered for the release please?

Regards,
Simon
Jerry Van Baren - April 17, 2013, 2:25 a.m.
Hi Simon,

On 04/12/2013 04:24 PM, Simon Glass wrote:
> Hi Tom / Jerry,

[snip]

>> Another way to fix this is to use -nostdinc and -idirafter to ensure that
>> system includes are included after U-Boot ones. Unfortunately this means
>> that U-Boot's errno.h gets included instead of the system one. This in
>> turn requires a hack to errno.h to redirect things, so all in all the
>> solution in this patch is probably cleaner.

[snip]

> Can this be considered for the release please?

This appears to be a reasonable request to fix an outstanding bug.  I 
don't have time today or tomorrow, but I'll pull in your patch and move 
it forward, Tom willing.

> Regards,
> Simon

Thanks,
gvb
Tom Rini - April 17, 2013, 3:35 p.m.
On Mon, Mar 11, 2013 at 09:30:27AM -0700, Simon Glass wrote:

> When building host utilities, we include libfdt.h from the host, not from
> U-Boot. This in turn brings in libfdt_env.h from the host, which can mess
> up the types and cause a build failure, depending on the host environment.
> To fix this, force inclusion of U-Boot's libfdt_env.h so that the types
> are correct.
> 
> Another way to fix this is to use -nostdinc and -idirafter to ensure that
> system includes are included after U-Boot ones. Unfortunately this means
> that U-Boot's errno.h gets included instead of the system one. This in
> turn requires a hack to errno.h to redirect things, so all in all the
> solution in this patch is probably cleaner.
> 
> Signed-off-by: Simon Glass <sjg@chromium.org>

Applied to u-boot/master, thanks!
Joe Hershberger - June 14, 2013, 2:56 p.m.
Hi Simon,

On Wed, Apr 17, 2013 at 10:35 AM, Tom Rini <trini@ti.com> wrote:
> On Mon, Mar 11, 2013 at 09:30:27AM -0700, Simon Glass wrote:
>
>> When building host utilities, we include libfdt.h from the host, not from
>> U-Boot. This in turn brings in libfdt_env.h from the host, which can mess
>> up the types and cause a build failure, depending on the host environment.
>> To fix this, force inclusion of U-Boot's libfdt_env.h so that the types
>> are correct.
>>
>> Another way to fix this is to use -nostdinc and -idirafter to ensure that
>> system includes are included after U-Boot ones. Unfortunately this means
>> that U-Boot's errno.h gets included instead of the system one. This in
>> turn requires a hack to errno.h to redirect things, so all in all the
>> solution in this patch is probably cleaner.
>>
>> Signed-off-by: Simon Glass <sjg@chromium.org>
>
> Applied to u-boot/master, thanks!
>
> --
> Tom

For me this patch causes a regression that looks like this:

Configuring for sandbox board...
aisimage.c: In function âaisimage_generateâ:
aisimage.c:314: warning: implicit declaration of function âgetlineâ
imximage.c: In function âparse_cfg_fileâ:
imximage.c:424: warning: implicit declaration of function âgetlineâ
kwbimage.c: In function âkwdimage_set_ext_headerâ:
kwbimage.c:236: warning: implicit declaration of function âgetlineâ
pblimage.c: In function âpbl_parserâ:
pblimage.c:119: warning: implicit declaration of function âgetlineâ
ublimage.c: In function âparse_cfg_fileâ:
ublimage.c:179: warning: implicit declaration of function âgetlineâ

Did you not see this?  I see it on ARM, PPC, and sandbox.

-Joe
Tom Rini - June 14, 2013, 3:07 p.m.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 06/14/2013 10:56 AM, Joe Hershberger wrote:
> Hi Simon,
> 
> On Wed, Apr 17, 2013 at 10:35 AM, Tom Rini <trini@ti.com> wrote:
>> On Mon, Mar 11, 2013 at 09:30:27AM -0700, Simon Glass wrote:
>> 
>>> When building host utilities, we include libfdt.h from the
>>> host, not from U-Boot. This in turn brings in libfdt_env.h from
>>> the host, which can mess up the types and cause a build
>>> failure, depending on the host environment. To fix this, force
>>> inclusion of U-Boot's libfdt_env.h so that the types are
>>> correct.
>>> 
>>> Another way to fix this is to use -nostdinc and -idirafter to
>>> ensure that system includes are included after U-Boot ones.
>>> Unfortunately this means that U-Boot's errno.h gets included
>>> instead of the system one. This in turn requires a hack to
>>> errno.h to redirect things, so all in all the solution in this
>>> patch is probably cleaner.
>>> 
>>> Signed-off-by: Simon Glass <sjg@chromium.org>
>> 
>> Applied to u-boot/master, thanks!
>> 
>> -- Tom
> 
> For me this patch causes a regression that looks like this:
> 
> Configuring for sandbox board... aisimage.c: In function
> âaisimage_generateâ: aisimage.c:314: warning: implicit declaration
> of function âgetlineâ imximage.c: In function âparse_cfg_fileâ: 
> imximage.c:424: warning: implicit declaration of function
> âgetlineâ kwbimage.c: In function âkwdimage_set_ext_headerâ: 
> kwbimage.c:236: warning: implicit declaration of function
> âgetlineâ pblimage.c: In function âpbl_parserâ: pblimage.c:119:
> warning: implicit declaration of function âgetlineâ ublimage.c: In
> function âparse_cfg_fileâ: ublimage.c:179: warning: implicit
> declaration of function âgetlineâ
> 
> Did you not see this?  I see it on ARM, PPC, and sandbox.

Nope.  What toolchains?

- -- 
Tom
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Thunderbird - http://www.enigmail.net/

iQIcBAEBAgAGBQJRuzGcAAoJENk4IS6UOR1W2KEP/Roaf4ewQxnpag59QdHHy7GJ
FaMHOA36GbjLt8ay4KaEfFazxpmJiVPWApEt5RkbKICa8sYot6X+nmcOhLMau46t
/mSx5ig6NMQeCvVmFvppRmBdJvew1dfxGiIAXd8GEAQQgDQsJ18KS0qOOyr9UlUI
o+z6unVXybQQgsijiQaph2uhIkV2Fo5wUY2JrNzbHDepjHgKkfS+6CeIRxJE7+yr
sCyn5hAAWDBIUR3Xlqw2pfrWPCNJNORSQhWtfMNVQbf1PFUYEv4U008QldsaR1qF
AcXZv7tzXahJYYqQTKuEGIdO9Ed9Ij3eJ+oPg5guZ5+6PYoX8ZLAyOKEtp/hdFs6
bHXdj5URUzv1bW81p4Hsuk+O+h1izIgqeUqfzTQfnZucgCPl87HVsevdA2N22deC
2PuM8vvEdl0UgJmwd9fFyrUYXzE+xzOLf3XRXVwOWtKrQUl8WE6FMmkGfAWLQNNx
LbXwmyQKXOPM1YAM+P7mxEk4L7gpeRF/r+uARBJ36Oi1RtY4zM16FYT25cOoppuL
aFCssOuxL02WiX4q9EWZ8sGtJ9H6c5Dz/oFMwWmse9i+QNMnLGpoS4m6imnDJw26
yju4fIztjajAJN+dFV5J9CJHp2ddon3sa+6Sdn7aEBrCUgjU0Zi9ne+Ge+UghoRG
3fHhQNqmvtq0OgrCmLPY
=+4Fc
-----END PGP SIGNATURE-----

Patch

diff --git a/tools/Makefile b/tools/Makefile
index c5952fc..889c897 100644
--- a/tools/Makefile
+++ b/tools/Makefile
@@ -164,7 +164,8 @@  NOPEDOBJS := $(addprefix $(obj),$(NOPED_OBJ_FILES-y))
 # Use native tools and options
 # Define __KERNEL_STRICT_NAMES to prevent typedef overlaps
 #
-HOSTCPPFLAGS =	-idirafter $(SRCTREE)/include \
+HOSTCPPFLAGS =	-include $(SRCTREE)/include/libfdt_env.h \
+		-idirafter $(SRCTREE)/include \
 		-idirafter $(OBJTREE)/include2 \
 		-idirafter $(OBJTREE)/include \
 	        -I $(SRCTREE)/lib/libfdt \