Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-26 Thread Mikko.Rapeli 
On Fri, Jun 23, 2017 at 04:20:41PM -0700, Khem Raj wrote:
> On Fri, Jun 23, 2017 at 7:17 AM,   wrote:
> > Hi,
> >
> > I'm chipping in since I've been messing with these things a bit in upstream
> > Linux kernel.
> >
> > On Fri, Jun 23, 2017 at 06:37:52AM -0700, Khem Raj wrote:
> >> On Fri, Jun 23, 2017 at 3:46 AM, André Draszik  wrote:
> >> > connman is not doing anything wrong here.
> >> >
> >>
> >> yes I am aware of this
> >>
> >> > The kernel is redefining IFF_LOWER_UP, because it thinks the libc doesn't
> >> > define it yet (and glibc doesn't).
> >> >
> >> > libc-compat.h is the way to solve these kind of issues. There also is 
> >> > https:
> >> > //lkml.org/lkml/2017/3/12/238 which is very similar. I'll pick that 
> >> > instead.
> >> >
> >> see the comment https://lkml.org/lkml/2017/3/16/121
> >> that worries me for this patch
> >
> > I'm aware of those review comments but I have not seen any patches posted 
> > which
> > fix the problem in some other way. Thus I would propose to apply these 
> > patches
> > as a workaround until upstream fixes the issues.
> >
> > These header files do not change that often either.
> 
> problem is you become incompatible ABI forever that worries me.
> However if bruce is fine to carry this patch as part of linux-yocto
> I might relent. It still will be hassle where folks will have to apply
> this patch to there kernels if they are building musl based systems.

API or ABI? But agreed this is a problem.

> >
> >> I am not questioning the correctness of patch too. But
> >> it would be better to get this patch accepted into kernel
> >> before applying to OE since these are kind of patches which
> >> you can get stuck with for life if upstream is not accepting it.
> >
> > Upstream-Status: Denied
> >
> > would be a correct marker for now I guess.
> 
> I would rather see some progress made to get it resolved :)
> we need to actually remove glibc'ness completely from kernel.
> and this will fix itself.

Yes. With the glibc'ness in the kernel headers I was just following existing
examples and adding some more. Some of those changes got accepted and only
later came the resistance to remove glibc'ness completely. No-one else
has proposed patches. Maybe I should invent some hobby project time again
for this...

-Mikko
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-23 Thread Andre McCurdy 
On Fri, Jun 23, 2017 at 4:41 PM, Denys Dmytriyenko  wrote:
> On Fri, Jun 23, 2017 at 04:20:41PM -0700, Khem Raj wrote:
>> On Fri, Jun 23, 2017 at 7:17 AM,   wrote:
>> > Hi,
>> >
>> > I'm chipping in since I've been messing with these things a bit in upstream
>> > Linux kernel.
>> >
>> > On Fri, Jun 23, 2017 at 06:37:52AM -0700, Khem Raj wrote:
>> >> On Fri, Jun 23, 2017 at 3:46 AM, André Draszik  wrote:
>> >> > connman is not doing anything wrong here.
>> >> >
>> >>
>> >> yes I am aware of this
>> >>
>> >> > The kernel is redefining IFF_LOWER_UP, because it thinks the libc 
>> >> > doesn't
>> >> > define it yet (and glibc doesn't).
>> >> >
>> >> > libc-compat.h is the way to solve these kind of issues. There also is 
>> >> > https:
>> >> > //lkml.org/lkml/2017/3/12/238 which is very similar. I'll pick that 
>> >> > instead.
>> >> >
>> >> see the comment https://lkml.org/lkml/2017/3/16/121
>> >> that worries me for this patch
>> >
>> > I'm aware of those review comments but I have not seen any patches posted 
>> > which
>> > fix the problem in some other way. Thus I would propose to apply these 
>> > patches
>> > as a workaround until upstream fixes the issues.
>> >
>> > These header files do not change that often either.
>>
>> problem is you become incompatible ABI forever that worries me.
>
> +1
>
>> However if bruce is fine to carry this patch as part of linux-yocto
>> I might relent. It still will be hassle where folks will have to apply
>> this patch to there kernels if they are building musl based systems.
>
> Don't forget that not everyone is using linux-yocto kernel!

Neither is linux-libc-headers, so getting these musl patches into the
linux-yocto kernel isn't going to change much...

>> >> I am not questioning the correctness of patch too. But
>> >> it would be better to get this patch accepted into kernel
>> >> before applying to OE since these are kind of patches which
>> >> you can get stuck with for life if upstream is not accepting it.
>> >
>> > Upstream-Status: Denied
>> >
>> > would be a correct marker for now I guess.
>>
>> I would rather see some progress made to get it resolved :)
>> we need to actually remove glibc'ness completely from kernel.
>> and this will fix itself.
> --
> ___
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-23 Thread Denys Dmytriyenko 
On Fri, Jun 23, 2017 at 04:20:41PM -0700, Khem Raj wrote:
> On Fri, Jun 23, 2017 at 7:17 AM,   wrote:
> > Hi,
> >
> > I'm chipping in since I've been messing with these things a bit in upstream
> > Linux kernel.
> >
> > On Fri, Jun 23, 2017 at 06:37:52AM -0700, Khem Raj wrote:
> >> On Fri, Jun 23, 2017 at 3:46 AM, André Draszik  wrote:
> >> > connman is not doing anything wrong here.
> >> >
> >>
> >> yes I am aware of this
> >>
> >> > The kernel is redefining IFF_LOWER_UP, because it thinks the libc doesn't
> >> > define it yet (and glibc doesn't).
> >> >
> >> > libc-compat.h is the way to solve these kind of issues. There also is 
> >> > https:
> >> > //lkml.org/lkml/2017/3/12/238 which is very similar. I'll pick that 
> >> > instead.
> >> >
> >> see the comment https://lkml.org/lkml/2017/3/16/121
> >> that worries me for this patch
> >
> > I'm aware of those review comments but I have not seen any patches posted 
> > which
> > fix the problem in some other way. Thus I would propose to apply these 
> > patches
> > as a workaround until upstream fixes the issues.
> >
> > These header files do not change that often either.
> 
> problem is you become incompatible ABI forever that worries me.

+1


> However if bruce is fine to carry this patch as part of linux-yocto
> I might relent. It still will be hassle where folks will have to apply
> this patch to there kernels if they are building musl based systems.

Don't forget that not everyone is using linux-yocto kernel!


> >> I am not questioning the correctness of patch too. But
> >> it would be better to get this patch accepted into kernel
> >> before applying to OE since these are kind of patches which
> >> you can get stuck with for life if upstream is not accepting it.
> >
> > Upstream-Status: Denied
> >
> > would be a correct marker for now I guess.
> 
> I would rather see some progress made to get it resolved :)
> we need to actually remove glibc'ness completely from kernel.
> and this will fix itself.
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-23 Thread Khem Raj 
On Fri, Jun 23, 2017 at 7:17 AM,   wrote:
> Hi,
>
> I'm chipping in since I've been messing with these things a bit in upstream
> Linux kernel.
>
> On Fri, Jun 23, 2017 at 06:37:52AM -0700, Khem Raj wrote:
>> On Fri, Jun 23, 2017 at 3:46 AM, André Draszik  wrote:
>> > connman is not doing anything wrong here.
>> >
>>
>> yes I am aware of this
>>
>> > The kernel is redefining IFF_LOWER_UP, because it thinks the libc doesn't
>> > define it yet (and glibc doesn't).
>> >
>> > libc-compat.h is the way to solve these kind of issues. There also is 
>> > https:
>> > //lkml.org/lkml/2017/3/12/238 which is very similar. I'll pick that 
>> > instead.
>> >
>> see the comment https://lkml.org/lkml/2017/3/16/121
>> that worries me for this patch
>
> I'm aware of those review comments but I have not seen any patches posted 
> which
> fix the problem in some other way. Thus I would propose to apply these patches
> as a workaround until upstream fixes the issues.
>
> These header files do not change that often either.

problem is you become incompatible ABI forever that worries me.
However if bruce is fine to carry this patch as part of linux-yocto
I might relent. It still will be hassle where folks will have to apply
this patch to there kernels if they are building musl based systems.

>
>> I am not questioning the correctness of patch too. But
>> it would be better to get this patch accepted into kernel
>> before applying to OE since these are kind of patches which
>> you can get stuck with for life if upstream is not accepting it.
>
> Upstream-Status: Denied
>
> would be a correct marker for now I guess.

I would rather see some progress made to get it resolved :)
we need to actually remove glibc'ness completely from kernel.
and this will fix itself.
>
> Hope this helps,
>
> -Mikko
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-23 Thread Mikko.Rapeli 
Hi,

I'm chipping in since I've been messing with these things a bit in upstream
Linux kernel.

On Fri, Jun 23, 2017 at 06:37:52AM -0700, Khem Raj wrote:
> On Fri, Jun 23, 2017 at 3:46 AM, André Draszik  wrote:
> > connman is not doing anything wrong here.
> >
> 
> yes I am aware of this
> 
> > The kernel is redefining IFF_LOWER_UP, because it thinks the libc doesn't
> > define it yet (and glibc doesn't).
> >
> > libc-compat.h is the way to solve these kind of issues. There also is https:
> > //lkml.org/lkml/2017/3/12/238 which is very similar. I'll pick that instead.
> >
> see the comment https://lkml.org/lkml/2017/3/16/121
> that worries me for this patch

I'm aware of those review comments but I have not seen any patches posted which
fix the problem in some other way. Thus I would propose to apply these patches
as a workaround until upstream fixes the issues.

These header files do not change that often either.

> I am not questioning the correctness of patch too. But
> it would be better to get this patch accepted into kernel
> before applying to OE since these are kind of patches which
> you can get stuck with for life if upstream is not accepting it.

Upstream-Status: Denied

would be a correct marker for now I guess.

Hope this helps,

-Mikko
-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core

Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-23 Thread Khem Raj 
On Fri, Jun 23, 2017 at 3:46 AM, André Draszik  wrote:
> connman is not doing anything wrong here.
>

yes I am aware of this

> The kernel is redefining IFF_LOWER_UP, because it thinks the libc doesn't
> define it yet (and glibc doesn't).
>
> libc-compat.h is the way to solve these kind of issues. There also is https:
> //lkml.org/lkml/2017/3/12/238 which is very similar. I'll pick that instead.
>
see the comment https://lkml.org/lkml/2017/3/16/121
that worries me for this patch

I am not questioning the correctness of patch too. But
it would be better to get this patch accepted into kernel
before applying to OE since these are kind of patches which
you can get stuck with for life if upstream is not accepting it.

> Cheers,
> Andre'
>
> On Thu, 2017-06-22 at 22:26 +, Khem Raj wrote:
>> I would rather fix the package instead of kernel headers unless kernel
>> community accepts this patch
>>
>> On Thu, Jun 22, 2017 at 11:14 AM André Draszik  wrote:
>>
>> > From: André Draszik 
>> >
>> > musl _does_ define IFF_LOWER_UP DORMANT ECHO so we should
>> > prevent redefinition of these when on musl.
>> >
>> > As per the included patch, this can be triggered by
>> > (from connman 6to4.c):
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> > include 
>> >
>> > In file included from ../git/src/6to4.c:34:0:
>> > .../usr/include/linux/if.h:97:2: error: expected identifier before
>> > numeric
>> > constant
>> >   IFF_LOWER_UP   = 1<<16, /* __volatile__ */
>> >   ^
>> >
>> > This is because at that time, IFF_LOWER_UP has been converted
>> > to 0x1 already:
>> > enum net_device_flags {
>> >  0x1 = 1<<16,
>> >  0x2 = 1<<17,
>> >  0x4 = 1<<18,
>> >
>> > };
>> >
>> > By defining __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
>> > to 0, we avoid the duplicated definition.
>> >
>> > Signed-off-by: André Draszik 
>> > ---
>> >  h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch | 62
>> > ++
>> >  .../linux-libc-headers/linux-libc-headers_4.10.bb  |  1 +
>> >  2 files changed, 63 insertions(+)
>> >  create mode 100644
>> > meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
>> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
>> >
>> > diff --git
>> > a/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
>> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
>> > b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
>> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
>> > new file mode 100644
>> > index 00..fd1bb9edb8
>> > --- /dev/null
>> > +++
>> > b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
>> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
>> > @@ -0,0 +1,62 @@
>> > +From 08a04c025395b0580913285045f2af36ee0985db Mon Sep 17 00:00:00 2001
>> > +From: =?UTF-8?q?Andr=C3=A9=20Draszik?= 
>> > +Date: Thu, 15 Jun 2017 16:55:33 +0100
>> > +Subject: [PATCH] libc-compat.h: musl _does_ define IFF_LOWER_UP DORMANT
>> > ECHO
>> > +MIME-Version: 1.0
>> > +Content-Type: text/plain; charset=UTF-8
>> > +Content-Transfer-Encoding: 8bit
>> > +
>> > +To trigger this (from connman 6to4.c):
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +include 
>> > +
>> > +In file included from ../git/src/6to4.c:34:0:
>> > +/scratch/yocto/build-tgm-r3-poky-fpp-tgm.pyro/tmp/work/tgm_r3-poky-
>> > linux-musl/connman/1.33-git4+AUTOINC+aa5b3dc12b-r0/recipe-
>> > sysroot/usr/include/linux/if.h:97:2:
>> > error: expected identifier before numeric constant
>> > +  IFF_LOWER_UP   = 1<<16, /* __volatile__ */
>> > +  ^
>> > +
>> > +This is because at that time, IFF_LOWER_UP has been converted
>> > +to 0x1 already:
>> > +enum net_device_flags {
>> > + 0x1 = 1<<16,
>> > + 0x2 = 1<<17,
>> > + 0x4 = 1<<18,
>> > +
>> > +};
>> > +
>> > +Upstream-Status: Pending
>> > +Signed-off-by: André Draszik 
>> > +Acked-by: Stephane Ayotte 
>> > +---
>> > + include/uapi/linux/libc-compat.h | 3 ++-
>> > + 1 file changed, 2 insertions(+), 1 deletion(-)
>> > +
>> > +diff --git a/include/uapi/linux/libc-compat.h
>> > b/include/uapi/linux/libc-compat.h
>> > +index f11ec0e..2003ae4 100644
>> > +--- a/include/uapi/linux/libc-compat.h
>> >  b/include/uapi/linux/libc-compat.h
>> > +@@ -70,7 +70,8 @@
>> > + #define __UAPI_DEF_IF_NET_DEVICE_FLAGS 0
>> > + /* For the future if glibc adds IFF_LOWER_UP, IFF_DORMANT and IFF_ECHO
>> > */
>> > + #ifndef

Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-23 Thread André Draszik 
connman is not doing anything wrong here.

The kernel is redefining IFF_LOWER_UP, because it thinks the libc doesn't
define it yet (and glibc doesn't).

libc-compat.h is the way to solve these kind of issues. There also is https:
//lkml.org/lkml/2017/3/12/238 which is very similar. I'll pick that instead.

Cheers,
Andre'

On Thu, 2017-06-22 at 22:26 +, Khem Raj wrote:
> I would rather fix the package instead of kernel headers unless kernel
> community accepts this patch
> 
> On Thu, Jun 22, 2017 at 11:14 AM André Draszik  wrote:
> 
> > From: André Draszik 
> > 
> > musl _does_ define IFF_LOWER_UP DORMANT ECHO so we should
> > prevent redefinition of these when on musl.
> > 
> > As per the included patch, this can be triggered by
> > (from connman 6to4.c):
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > include 
> > 
> > In file included from ../git/src/6to4.c:34:0:
> > .../usr/include/linux/if.h:97:2: error: expected identifier before
> > numeric
> > constant
> >   IFF_LOWER_UP   = 1<<16, /* __volatile__ */
> >   ^
> > 
> > This is because at that time, IFF_LOWER_UP has been converted
> > to 0x1 already:
> > enum net_device_flags {
> >  0x1 = 1<<16,
> >  0x2 = 1<<17,
> >  0x4 = 1<<18,
> > 
> > };
> > 
> > By defining __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
> > to 0, we avoid the duplicated definition.
> > 
> > Signed-off-by: André Draszik 
> > ---
> >  h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch | 62
> > ++
> >  .../linux-libc-headers/linux-libc-headers_4.10.bb  |  1 +
> >  2 files changed, 63 insertions(+)
> >  create mode 100644
> > meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
> > 
> > diff --git
> > a/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
> > b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
> > new file mode 100644
> > index 00..fd1bb9edb8
> > --- /dev/null
> > +++
> > b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-
> > compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
> > @@ -0,0 +1,62 @@
> > +From 08a04c025395b0580913285045f2af36ee0985db Mon Sep 17 00:00:00 2001
> > +From: =?UTF-8?q?Andr=C3=A9=20Draszik?= 
> > +Date: Thu, 15 Jun 2017 16:55:33 +0100
> > +Subject: [PATCH] libc-compat.h: musl _does_ define IFF_LOWER_UP DORMANT
> > ECHO
> > +MIME-Version: 1.0
> > +Content-Type: text/plain; charset=UTF-8
> > +Content-Transfer-Encoding: 8bit
> > +
> > +To trigger this (from connman 6to4.c):
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +include 
> > +
> > +In file included from ../git/src/6to4.c:34:0:
> > +/scratch/yocto/build-tgm-r3-poky-fpp-tgm.pyro/tmp/work/tgm_r3-poky-
> > linux-musl/connman/1.33-git4+AUTOINC+aa5b3dc12b-r0/recipe-
> > sysroot/usr/include/linux/if.h:97:2:
> > error: expected identifier before numeric constant
> > +  IFF_LOWER_UP   = 1<<16, /* __volatile__ */
> > +  ^
> > +
> > +This is because at that time, IFF_LOWER_UP has been converted
> > +to 0x1 already:
> > +enum net_device_flags {
> > + 0x1 = 1<<16,
> > + 0x2 = 1<<17,
> > + 0x4 = 1<<18,
> > +
> > +};
> > +
> > +Upstream-Status: Pending
> > +Signed-off-by: André Draszik 
> > +Acked-by: Stephane Ayotte 
> > +---
> > + include/uapi/linux/libc-compat.h | 3 ++-
> > + 1 file changed, 2 insertions(+), 1 deletion(-)
> > +
> > +diff --git a/include/uapi/linux/libc-compat.h
> > b/include/uapi/linux/libc-compat.h
> > +index f11ec0e..2003ae4 100644
> > +--- a/include/uapi/linux/libc-compat.h
> >  b/include/uapi/linux/libc-compat.h
> > +@@ -70,7 +70,8 @@
> > + #define __UAPI_DEF_IF_NET_DEVICE_FLAGS 0
> > + /* For the future if glibc adds IFF_LOWER_UP, IFF_DORMANT and IFF_ECHO
> > */
> > + #ifndef __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
> > +-#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 1
> > ++/* musl has these defined */
> > ++#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 0
> > + #endif /* __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO */
> > +
> > + #else /* _NET_IF_H */
> > +--
> > +2.11.0
> > +
> > diff --git a/meta/recipes-kernel/linux-libc-headers/
> > linux-libc-headers_4.10.bb b/meta/recipes-kernel/linux-libc-headers/
> > linux-libc-headers_4.10.bb
> > index 108446aa34..29262789a8 100644
> > ---

Re: [OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-22 Thread Khem Raj 
I would rather fix the package instead of kernel headers unless kernel
community accepts this patch

On Thu, Jun 22, 2017 at 11:14 AM André Draszik  wrote:

> From: André Draszik 
>
> musl _does_ define IFF_LOWER_UP DORMANT ECHO so we should
> prevent redefinition of these when on musl.
>
> As per the included patch, this can be triggered by
> (from connman 6to4.c):
> include 
> include 
> include 
> include 
> include 
> include 
> include 
> include 
> include 
> include 
> include 
> include 
> include 
> include 
>
> In file included from ../git/src/6to4.c:34:0:
> .../usr/include/linux/if.h:97:2: error: expected identifier before numeric
> constant
>   IFF_LOWER_UP   = 1<<16, /* __volatile__ */
>   ^
>
> This is because at that time, IFF_LOWER_UP has been converted
> to 0x1 already:
> enum net_device_flags {
>  0x1 = 1<<16,
>  0x2 = 1<<17,
>  0x4 = 1<<18,
>
> };
>
> By defining __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
> to 0, we avoid the duplicated definition.
>
> Signed-off-by: André Draszik 
> ---
>  h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch | 62
> ++
>  .../linux-libc-headers/linux-libc-headers_4.10.bb  |  1 +
>  2 files changed, 63 insertions(+)
>  create mode 100644
> meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
>
> diff --git
> a/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
> b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
> new file mode 100644
> index 00..fd1bb9edb8
> --- /dev/null
> +++
> b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
> @@ -0,0 +1,62 @@
> +From 08a04c025395b0580913285045f2af36ee0985db Mon Sep 17 00:00:00 2001
> +From: =?UTF-8?q?Andr=C3=A9=20Draszik?= 
> +Date: Thu, 15 Jun 2017 16:55:33 +0100
> +Subject: [PATCH] libc-compat.h: musl _does_ define IFF_LOWER_UP DORMANT
> ECHO
> +MIME-Version: 1.0
> +Content-Type: text/plain; charset=UTF-8
> +Content-Transfer-Encoding: 8bit
> +
> +To trigger this (from connman 6to4.c):
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +include 
> +
> +In file included from ../git/src/6to4.c:34:0:
> +/scratch/yocto/build-tgm-r3-poky-fpp-tgm.pyro/tmp/work/tgm_r3-poky-linux-musl/connman/1.33-git4+AUTOINC+aa5b3dc12b-r0/recipe-sysroot/usr/include/linux/if.h:97:2:
> error: expected identifier before numeric constant
> +  IFF_LOWER_UP   = 1<<16, /* __volatile__ */
> +  ^
> +
> +This is because at that time, IFF_LOWER_UP has been converted
> +to 0x1 already:
> +enum net_device_flags {
> + 0x1 = 1<<16,
> + 0x2 = 1<<17,
> + 0x4 = 1<<18,
> +
> +};
> +
> +Upstream-Status: Pending
> +Signed-off-by: André Draszik 
> +Acked-by: Stephane Ayotte 
> +---
> + include/uapi/linux/libc-compat.h | 3 ++-
> + 1 file changed, 2 insertions(+), 1 deletion(-)
> +
> +diff --git a/include/uapi/linux/libc-compat.h
> b/include/uapi/linux/libc-compat.h
> +index f11ec0e..2003ae4 100644
> +--- a/include/uapi/linux/libc-compat.h
>  b/include/uapi/linux/libc-compat.h
> +@@ -70,7 +70,8 @@
> + #define __UAPI_DEF_IF_NET_DEVICE_FLAGS 0
> + /* For the future if glibc adds IFF_LOWER_UP, IFF_DORMANT and IFF_ECHO */
> + #ifndef __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
> +-#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 1
> ++/* musl has these defined */
> ++#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 0
> + #endif /* __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO */
> +
> + #else /* _NET_IF_H */
> +--
> +2.11.0
> +
> diff --git a/meta/recipes-kernel/linux-libc-headers/
> linux-libc-headers_4.10.bb b/meta/recipes-kernel/linux-libc-headers/
> linux-libc-headers_4.10.bb
> index 108446aa34..29262789a8 100644
> --- a/meta/recipes-kernel/linux-libc-headers/linux-libc-headers_4.10.bb
> +++ b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers_4.10.bb
> @@ -4,6 +4,7 @@ SRC_URI_append_libc-musl = "\
>  file://0001-libc-compat.h-fix-some-issues-arising-from-in6.h.patch \
>  file://0002-libc-compat.h-prevent-redefinition-of-struct-ethhdr.patch
> \
>  file://0003-remove-inclusion-of-sysinfo.h-in-kernel.h.patch \
> +
> file://0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch \
> "
>
>  SRC_URI[md5sum] = "b5e7f6b9b2fe1b6cc7bc56a3a0bfc090"
> --
> 2.11.0
>
> --
> ___
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
>

[OE-core] [PATCH] linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

2017-06-22 Thread André Draszik 
From: André Draszik 

musl _does_ define IFF_LOWER_UP DORMANT ECHO so we should
prevent redefinition of these when on musl.

As per the included patch, this can be triggered by
(from connman 6to4.c):
include 
include 
include 
include 
include 
include 
include 
include 
include 
include 
include 
include 
include 
include 

In file included from ../git/src/6to4.c:34:0:
.../usr/include/linux/if.h:97:2: error: expected identifier before numeric 
constant
  IFF_LOWER_UP   = 1<<16, /* __volatile__ */
  ^

This is because at that time, IFF_LOWER_UP has been converted
to 0x1 already:
enum net_device_flags {
 0x1 = 1<<16,
 0x2 = 1<<17,
 0x4 = 1<<18,

};

By defining __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
to 0, we avoid the duplicated definition.

Signed-off-by: André Draszik 
---
 h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch | 62 ++
 .../linux-libc-headers/linux-libc-headers_4.10.bb  |  1 +
 2 files changed, 63 insertions(+)
 create mode 100644 
meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch

diff --git 
a/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
 
b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
new file mode 100644
index 00..fd1bb9edb8
--- /dev/null
+++ 
b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers/0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch
@@ -0,0 +1,62 @@
+From 08a04c025395b0580913285045f2af36ee0985db Mon Sep 17 00:00:00 2001
+From: =?UTF-8?q?Andr=C3=A9=20Draszik?= 
+Date: Thu, 15 Jun 2017 16:55:33 +0100
+Subject: [PATCH] libc-compat.h: musl _does_ define IFF_LOWER_UP DORMANT ECHO
+MIME-Version: 1.0
+Content-Type: text/plain; charset=UTF-8
+Content-Transfer-Encoding: 8bit
+
+To trigger this (from connman 6to4.c):
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+include 
+
+In file included from ../git/src/6to4.c:34:0:
+/scratch/yocto/build-tgm-r3-poky-fpp-tgm.pyro/tmp/work/tgm_r3-poky-linux-musl/connman/1.33-git4+AUTOINC+aa5b3dc12b-r0/recipe-sysroot/usr/include/linux/if.h:97:2:
 error: expected identifier before numeric constant
+  IFF_LOWER_UP   = 1<<16, /* __volatile__ */
+  ^
+
+This is because at that time, IFF_LOWER_UP has been converted
+to 0x1 already:
+enum net_device_flags {
+ 0x1 = 1<<16,
+ 0x2 = 1<<17,
+ 0x4 = 1<<18,
+
+};
+
+Upstream-Status: Pending
+Signed-off-by: André Draszik 
+Acked-by: Stephane Ayotte 
+---
+ include/uapi/linux/libc-compat.h | 3 ++-
+ 1 file changed, 2 insertions(+), 1 deletion(-)
+
+diff --git a/include/uapi/linux/libc-compat.h 
b/include/uapi/linux/libc-compat.h
+index f11ec0e..2003ae4 100644
+--- a/include/uapi/linux/libc-compat.h
 b/include/uapi/linux/libc-compat.h
+@@ -70,7 +70,8 @@
+ #define __UAPI_DEF_IF_NET_DEVICE_FLAGS 0
+ /* For the future if glibc adds IFF_LOWER_UP, IFF_DORMANT and IFF_ECHO */
+ #ifndef __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO
+-#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 1
++/* musl has these defined */
++#define __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO 0
+ #endif /* __UAPI_DEF_IF_NET_DEVICE_FLAGS_LOWER_UP_DORMANT_ECHO */
+ 
+ #else /* _NET_IF_H */
+-- 
+2.11.0
+
diff --git a/meta/recipes-kernel/linux-libc-headers/linux-libc-headers_4.10.bb 
b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers_4.10.bb
index 108446aa34..29262789a8 100644
--- a/meta/recipes-kernel/linux-libc-headers/linux-libc-headers_4.10.bb
+++ b/meta/recipes-kernel/linux-libc-headers/linux-libc-headers_4.10.bb
@@ -4,6 +4,7 @@ SRC_URI_append_libc-musl = "\
 file://0001-libc-compat.h-fix-some-issues-arising-from-in6.h.patch \
 file://0002-libc-compat.h-prevent-redefinition-of-struct-ethhdr.patch \
 file://0003-remove-inclusion-of-sysinfo.h-in-kernel.h.patch \
+file://0001-libc-compat.h-musl-_does_-define-IFF_LOWER_UP-DORMAN.patch \
"
 
 SRC_URI[md5sum] = "b5e7f6b9b2fe1b6cc7bc56a3a0bfc090"
-- 
2.11.0

-- 
___
Openembedded-core mailing list
Openembedded-core@lists.openembedded.org
http://lists.openembedded.org/mailman/listinfo/openembedded-core