* stubs-32.h not found
@ 2016-02-02 20:07 M. Gregory
2016-02-02 21:59 ` M. Gregory
0 siblings, 1 reply; 6+ messages in thread
From: M. Gregory @ 2016-02-02 20:07 UTC (permalink / raw)
To: meta-virtualization
I am working with the Xen recipes and I getting an odd behavior. The
build will sometimes fail compiling Xen. The failure is a result of the
absence of stubs-32.h (see error below). If I run a bitbake -c clean xen
and build xen again everything builds as expected. According to the logs
the do_stubs task runs and stubs-64.h is present in the sysroot, so
stubs-32.h should have been created by the do_stubs task. I checked the
timestamps on the do_populate_sysroot task of glibc and it was run after
do_stubs, which indicates a dependence problem. I thought that adding
glibc to the DEPENDS would fix the problem, but it did not. Any ideas?
-Machon
fatal error: gnu/stubs-32.h: No such file or directory
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: stubs-32.h not found
2016-02-02 20:07 stubs-32.h not found M. Gregory
@ 2016-02-02 21:59 ` M. Gregory
2016-02-04 20:52 ` Chris Patterson
0 siblings, 1 reply; 6+ messages in thread
From: M. Gregory @ 2016-02-02 21:59 UTC (permalink / raw)
To: meta-virtualization
On 02/02/2016 03:07 PM, M. Gregory wrote:
> I am working with the Xen recipes and I getting an odd behavior. The
> build will sometimes fail compiling Xen. The failure is a result of the
> absence of stubs-32.h (see error below). If I run a bitbake -c clean xen
> and build xen again everything builds as expected. According to the logs
> the do_stubs task runs and stubs-64.h is present in the sysroot, so
> stubs-32.h should have been created by the do_stubs task. I checked the
> timestamps on the do_populate_sysroot task of glibc and it was run after
> do_stubs, which indicates a dependence problem. I thought that adding
> glibc to the DEPENDS would fix the problem, but it did not. Any ideas?
> -Machon
>
> fatal error: gnu/stubs-32.h: No such file or directory
>
So I read the Yocto manual and I understand why adding glibc to the
DEPENDS does not work. Specifically, DEPENDS only ensures "all the
dependencies listed have been built and have their contents in the
appropriate sysroots before the recipe's configure task is executed."
The do_stubs tasks running before configure. Adding the the line below
seems to fix the problem. Is this the correct solution?
Signed-off-by: Machon Gregory <mbgrego@tycho.nsa.gov>
diff --git a/recipes-extended/xen/xen.inc b/recipes-extended/xen/xen.inc
index 5555ed3..192bbee 100644
--- a/recipes-extended/xen/xen.inc
+++ b/recipes-extended/xen/xen.inc
@@ -783,6 +783,7 @@ do_stubs() {
}
addtask stubs after do_patch before do_configure
+do_stubs[depends] += "glibc:do_populate_sysroot"
do_configure() {
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: stubs-32.h not found
2016-02-02 21:59 ` M. Gregory
@ 2016-02-04 20:52 ` Chris Patterson
2016-02-04 20:54 ` M. Gregory
0 siblings, 1 reply; 6+ messages in thread
From: Chris Patterson @ 2016-02-04 20:52 UTC (permalink / raw)
To: M. Gregory; +Cc: meta-virtualization@yoctoproject.org
I repro'd this with a fresh tree. Instead of the depends approach, I
simply moved the step between do_configure and do_compile, which
should be OK?
diff --git a/recipes-extended/xen/xen.inc b/recipes-extended/xen/xen.inc
index 5555ed3..d35bc3f 100644
--- a/recipes-extended/xen/xen.inc
+++ b/recipes-extended/xen/xen.inc
@@ -782,7 +782,7 @@ do_stubs() {
fi
}
-addtask stubs after do_patch before do_configure
+addtask stubs after do_configure before do_compile
do_configure() {
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: stubs-32.h not found
2016-02-04 20:52 ` Chris Patterson
@ 2016-02-04 20:54 ` M. Gregory
2016-02-10 14:06 ` M. Gregory
0 siblings, 1 reply; 6+ messages in thread
From: M. Gregory @ 2016-02-04 20:54 UTC (permalink / raw)
To: Chris Patterson; +Cc: meta-virtualization@yoctoproject.org
On 02/04/2016 03:52 PM, Chris Patterson wrote:
> I repro'd this with a fresh tree. Instead of the depends approach, I
> simply moved the step between do_configure and do_compile, which
> should be OK?
>
> diff --git a/recipes-extended/xen/xen.inc b/recipes-extended/xen/xen.inc
> index 5555ed3..d35bc3f 100644
> --- a/recipes-extended/xen/xen.inc
> +++ b/recipes-extended/xen/xen.inc
> @@ -782,7 +782,7 @@ do_stubs() {
> fi
> }
>
> -addtask stubs after do_patch before do_configure
> +addtask stubs after do_configure before do_compile
>
> do_configure() {
yes.
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: stubs-32.h not found
2016-02-04 20:54 ` M. Gregory
@ 2016-02-10 14:06 ` M. Gregory
2016-02-12 15:25 ` Chris Patterson
0 siblings, 1 reply; 6+ messages in thread
From: M. Gregory @ 2016-02-10 14:06 UTC (permalink / raw)
To: meta-virtualization
On 02/04/2016 03:54 PM, M. Gregory wrote:
> On 02/04/2016 03:52 PM, Chris Patterson wrote:
>> I repro'd this with a fresh tree. Instead of the depends approach, I
>> simply moved the step between do_configure and do_compile, which
>> should be OK?
>>
>> diff --git a/recipes-extended/xen/xen.inc b/recipes-extended/xen/xen.inc
>> index 5555ed3..d35bc3f 100644
>> --- a/recipes-extended/xen/xen.inc
>> +++ b/recipes-extended/xen/xen.inc
>> @@ -782,7 +782,7 @@ do_stubs() {
>> fi
>> }
>>
>> -addtask stubs after do_patch before do_configure
>> +addtask stubs after do_configure before do_compile
>>
>> do_configure() {
> yes.
Can you submit this as a patch?
-Machon
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: stubs-32.h not found
2016-02-10 14:06 ` M. Gregory
@ 2016-02-12 15:25 ` Chris Patterson
0 siblings, 0 replies; 6+ messages in thread
From: Chris Patterson @ 2016-02-12 15:25 UTC (permalink / raw)
To: M. Gregory; +Cc: meta-virtualization@yoctoproject.org
Sure thing. :)
On Wed, Feb 10, 2016 at 9:06 AM, M. Gregory <mbgrego@tycho.nsa.gov> wrote:
>
>
> On 02/04/2016 03:54 PM, M. Gregory wrote:
>> On 02/04/2016 03:52 PM, Chris Patterson wrote:
>>> I repro'd this with a fresh tree. Instead of the depends approach, I
>>> simply moved the step between do_configure and do_compile, which
>>> should be OK?
>>>
>>> diff --git a/recipes-extended/xen/xen.inc b/recipes-extended/xen/xen.inc
>>> index 5555ed3..d35bc3f 100644
>>> --- a/recipes-extended/xen/xen.inc
>>> +++ b/recipes-extended/xen/xen.inc
>>> @@ -782,7 +782,7 @@ do_stubs() {
>>> fi
>>> }
>>>
>>> -addtask stubs after do_patch before do_configure
>>> +addtask stubs after do_configure before do_compile
>>>
>>> do_configure() {
>> yes.
> Can you submit this as a patch?
> -Machon
> --
> _______________________________________________
> meta-virtualization mailing list
> meta-virtualization@yoctoproject.org
> https://lists.yoctoproject.org/listinfo/meta-virtualization
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2016-02-12 15:25 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2016-02-02 20:07 stubs-32.h not found M. Gregory
2016-02-02 21:59 ` M. Gregory
2016-02-04 20:52 ` Chris Patterson
2016-02-04 20:54 ` M. Gregory
2016-02-10 14:06 ` M. Gregory
2016-02-12 15:25 ` Chris Patterson
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.