Xen devel - Nov 2012 - ARM build with v3.7-rc3 with CONFIG_XEN_GNTDEV=m

Hey Stefano,

I setup a cross ARM build system so that the generic changes won''t
cause build issues
on ARM (and actually caught a bug with fallback.o being enabled there -
I''ve a fix
for that).

But more interestingly the build system I do also makes a variations of config
entries.
So there is the allmodules, allyes, allyes, but all Xen drivers are modules,
etc.

I found out that with:
CONFIG_XEN_GNTDEV=m
I get this

ERROR: "HYPERVISOR_event_channel_op" [drivers/xen/xen-gntdev.ko]
undefined!

If I make it CONFIG_XEN_GNTDEV=y the issue disappears. Any thoughts?

(FYI, if you want to do this on v3.7-rc4 please apply this patch to your tree)


commit e1a982642262820d68779c06de64f8857a040f5a
Author: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Date:   Tue Nov 6 15:49:27 2012 -0500

    xen/generic: Disable fallback build on ARM.
    
    As there is no need for it (the fallback code is for older
    hypervisors and they won''t run under ARM), and also b/c
    we get:
    
    drivers/xen/fallback.c: In function
''xen_event_channel_op_compat'':
    drivers/xen/fallback.c:10:19: error: storage size of ''op''
isn''t known
    drivers/xen/fallback.c:15:2: error: implicit declaration of function
''_hypercall1'' [-Werror=implicit-function-declaration]
    drivers/xen/fallback.c:15:19: error: expected expression before
''int''
    drivers/xen/fallback.c:18:7: error: ''EVTCHNOP_close''
undeclared (first use in this function)
    drivers/xen/fallback.c:18:7: note: each undeclared identifier is reported
only once for each function it appears in
    .. and more
    
    Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>

diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
index 46de6cd..273d2b9 100644
--- a/drivers/xen/Makefile
+++ b/drivers/xen/Makefile
@@ -1,8 +1,8 @@
 ifneq ($(CONFIG_ARM),y)
-obj-y	+= manage.o balloon.o
+obj-y	+= manage.o balloon.o fallback.o
 obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
 endif
-obj-y	+= grant-table.o features.o events.o fallback.o
+obj-y	+= grant-table.o features.o events.o
 obj-y	+= xenbus/
 
 nostackp := $(call cc-option, -fno-stack-protector)

On Tue, 2012-11-06 at 21:19 +0000, Konrad Rzeszutek Wilk
wrote:
> Hey Stefano,
> 
> I setup a cross ARM build system so that the generic changes won''t
cause build issues
> on ARM (and actually caught a bug with fallback.o being enabled there -
I''ve a fix
> for that).
> 
> But more interestingly the build system I do also makes a variations of
config entries.
> So there is the allmodules, allyes, allyes, but all Xen drivers are
modules, etc.
> 
> I found out that with:
> CONFIG_XEN_GNTDEV=m
> I get this
> 
> ERROR: "HYPERVISOR_event_channel_op" [drivers/xen/xen-gntdev.ko]
undefined!
> 
> If I make it CONFIG_XEN_GNTDEV=y the issue disappears. Any thoughts?
I wonder if this is to do with missing EXPORT_SYMBOL for the various
things in hypercall.S ? Quite possibly this is the same thing as Russell
King just reported r.e. privmcd?
> 
> (FYI, if you want to do this on v3.7-rc4 please apply this patch to your
tree)
> 
> 
> commit e1a982642262820d68779c06de64f8857a040f5a
> Author: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
> Date:   Tue Nov 6 15:49:27 2012 -0500
> 
>     xen/generic: Disable fallback build on ARM.
>     
>     As there is no need for it (the fallback code is for older
>     hypervisors and they won''t run under ARM), and also b/c
>     we get:
>     
>     drivers/xen/fallback.c: In function
''xen_event_channel_op_compat'':
>     drivers/xen/fallback.c:10:19: error: storage size of
''op'' isn''t known
>     drivers/xen/fallback.c:15:2: error: implicit declaration of function
''_hypercall1'' [-Werror=implicit-function-declaration]
>     drivers/xen/fallback.c:15:19: error: expected expression before
''int''
>     drivers/xen/fallback.c:18:7: error: ''EVTCHNOP_close''
undeclared (first use in this function)
>     drivers/xen/fallback.c:18:7: note: each undeclared identifier is
reported only once for each function it appears in
>     .. and more
>     
>     Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
> 
> diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> index 46de6cd..273d2b9 100644
> --- a/drivers/xen/Makefile
> +++ b/drivers/xen/Makefile
> @@ -1,8 +1,8 @@
>  ifneq ($(CONFIG_ARM),y)
> -obj-y	+= manage.o balloon.o
> +obj-y	+= manage.o balloon.o fallback.o
>  obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
>  endif
> -obj-y	+= grant-table.o features.o events.o fallback.o
> +obj-y	+= grant-table.o features.o events.o
>  obj-y	+= xenbus/
>  
>  nostackp := $(call cc-option, -fno-stack-protector)

On Tue, Nov 06, 2012 at 10:39:59PM +0100, Ian Campbell
wrote:
> On Tue, 2012-11-06 at 21:19 +0000, Konrad Rzeszutek Wilk wrote:
> > Hey Stefano,
> > 
> > I setup a cross ARM build system so that the generic changes
won''t cause build issues
> > on ARM (and actually caught a bug with fallback.o being enabled there
- I''ve a fix
> > for that).
> > 
> > But more interestingly the build system I do also makes a variations
of config entries.
> > So there is the allmodules, allyes, allyes, but all Xen drivers are
modules, etc.
> > 
> > I found out that with:
> > CONFIG_XEN_GNTDEV=m
> > I get this
> > 
> > ERROR: "HYPERVISOR_event_channel_op"
[drivers/xen/xen-gntdev.ko] undefined!
> > 
> > If I make it CONFIG_XEN_GNTDEV=y the issue disappears. Any thoughts?
> 
> I wonder if this is to do with missing EXPORT_SYMBOL for the various
> things in hypercall.S ? Quite possibly this is the same thing as Russell
> King just reported r.e. privmcd?
I think so, I was unable to find the code for HYPERVISOR_event_channel_op
until I did a grep for ''hypercall'' and found the hypercall.S
file :-(

Will add an EXPORT_SYMBOL_GPL for said issue and see if that fixes it.


> 
> > 
> > (FYI, if you want to do this on v3.7-rc4 please apply this patch to
your tree)
> > 
> > 
> > commit e1a982642262820d68779c06de64f8857a040f5a
> > Author: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
> > Date:   Tue Nov 6 15:49:27 2012 -0500
> > 
> >     xen/generic: Disable fallback build on ARM.
> >     
> >     As there is no need for it (the fallback code is for older
> >     hypervisors and they won''t run under ARM), and also b/c
> >     we get:
> >     
> >     drivers/xen/fallback.c: In function
''xen_event_channel_op_compat'':
> >     drivers/xen/fallback.c:10:19: error: storage size of
''op'' isn''t known
> >     drivers/xen/fallback.c:15:2: error: implicit declaration of
function ''_hypercall1'' [-Werror=implicit-function-declaration]
> >     drivers/xen/fallback.c:15:19: error: expected expression before
''int''
> >     drivers/xen/fallback.c:18:7: error:
''EVTCHNOP_close'' undeclared (first use in this function)
> >     drivers/xen/fallback.c:18:7: note: each undeclared identifier is
reported only once for each function it appears in
> >     .. and more
> >     
> >     Signed-off-by: Konrad Rzeszutek Wilk
<konrad.wilk@oracle.com>
> > 
> > diff --git a/drivers/xen/Makefile b/drivers/xen/Makefile
> > index 46de6cd..273d2b9 100644
> > --- a/drivers/xen/Makefile
> > +++ b/drivers/xen/Makefile
> > @@ -1,8 +1,8 @@
> >  ifneq ($(CONFIG_ARM),y)
> > -obj-y	+= manage.o balloon.o
> > +obj-y	+= manage.o balloon.o fallback.o
> >  obj-$(CONFIG_HOTPLUG_CPU)		+= cpu_hotplug.o
> >  endif
> > -obj-y	+= grant-table.o features.o events.o fallback.o
> > +obj-y	+= grant-table.o features.o events.o
> >  obj-y	+= xenbus/
> >  
> >  nostackp := $(call cc-option, -fno-stack-protector)
>