Xen devel - May 2012 - [ANNOUNCE] First release candidates for Xen 4.0.4 and 4.1.3

Folks,

I have just tagged first release candidates for 4.0.4 and 4.1.3:

http://xenbits.xen.org/staging/xen-4.0-testing.hg (tag 4.0.4-rc1)
http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)

Please test!

 -- Keir

On Mon 7. of May 2012 14:58:44 Keir Fraser wrote:
> Folks,
> 
> I have just tagged first release candidates for 4.0.4 and 4.1.3:
> 
> http://xenbits.xen.org/staging/xen-4.0-testing.hg (tag 4.0.4-rc1)
> http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)
> 
> Please test!
> 
>  -- Keir
And which kernel should we use please?
-- 
Pavel Mateja

On 07/05/2012 18:18, "Pavel Matěja" <pavel@netsafe.cz> wrote:
> On Mon 7. of May 2012 14:58:44 Keir Fraser wrote:
>> Folks,
>> 
>> I have just tagged first release candidates for 4.0.4 and 4.1.3:
>> 
>> http://xenbits.xen.org/staging/xen-4.0-testing.hg (tag 4.0.4-rc1)
>> http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)
>> 
>> Please test!
>> 
>>  -- Keir
> 
> And which kernel should we use please?
The kernel of your choice. We don''t do kernels and hypervisors as
matched
pairs these days. A recent pv_ops kernel ought to work with either 4.0 or
4.1.

 -- Keir

On 07/05/2012 13:58, Keir Fraser wrote:
> Folks,
>
> I have just tagged first release candidates for 4.0.4 and 4.1.3:
>
> http://xenbits.xen.org/staging/xen-4.0-testing.hg (tag 4.0.4-rc1)
> http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)
>
> Please test!
>
>  -- Keir
XenServer trunk is currently running on xen-4.1-testing tip (minus the 3
changesets today) (plus a patch queue).

I am not aware of any outstanding hypervisor issues.

~Andrew
>
>
>
> _______________________________________________
> Xen-devel mailing list
> Xen-devel@lists.xen.org
> http://lists.xen.org/xen-devel

On Mon, 7 May 2012, Keir Fraser wrote:
> Folks,
>
> I have just tagged first release candidates for 4.0.4 and 4.1.3:
>
> http://xenbits.xen.org/staging/xen-4.0-testing.hg (tag 4.0.4-rc1)
> http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)
>
> Please test!
I am seeing problems building 4.1.3-rc1 on Fedora 17, though it might be 
changes in Fedora 17 rather than xen changes that have triggered it. When 
trying to compile objects such as scheduler.c in tools/blktap2/drivers I 
get the error
/usr/include/features.h:329:3: error: #warning _FORTIFY_SOURCE requested but
disabled [-Werror=cpp]

On checking that file I see that this is because -Wp,-D_FORTIFY_SOURCE=2 
is requested with -O0 . Fedora 17 decides to warn about this and it 
becomes an error as -Werror is specified.
This conflict occurs because of the line
CFLAGS += -Werror -g -O0
in tools/blktap2/drivers/Makefile
I can of course work around the problem in my build, but I was wondering 
if optimization level 0 is still necessary for this code. I think the 
only other place -O0 is used now is tools/security/Makefile which I 
assume will also trigger the problem.

 	Michael Young

On 07/05/2012 21:52, "M A Young" <m.a.young@durham.ac.uk> wrote:
> On Mon, 7 May 2012, Keir Fraser wrote:
> 
>> Folks,
>> 
>> I have just tagged first release candidates for 4.0.4 and 4.1.3:
>> 
>> http://xenbits.xen.org/staging/xen-4.0-testing.hg (tag 4.0.4-rc1)
>> http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)
>> 
>> Please test!
> 
> I am seeing problems building 4.1.3-rc1 on Fedora 17, though it might be
> changes in Fedora 17 rather than xen changes that have triggered it. When
> trying to compile objects such as scheduler.c in tools/blktap2/drivers I
> get the error
> /usr/include/features.h:329:3: error: #warning _FORTIFY_SOURCE requested
but
> disabled [-Werror=cpp]
> 
> On checking that file I see that this is because -Wp,-D_FORTIFY_SOURCE=2
> is requested with -O0 . Fedora 17 decides to warn about this and it
> becomes an error as -Werror is specified.
> This conflict occurs because of the line
> CFLAGS += -Werror -g -O0
> in tools/blktap2/drivers/Makefile
> I can of course work around the problem in my build, but I was wondering
> if optimization level 0 is still necessary for this code. I think the
> only other place -O0 is used now is tools/security/Makefile which I
> assume will also trigger the problem.
Thanks. The same appears to be true in xen-unstable also. Is the solution to
simply remove -O0 from the command line? We can test that out in
xen-unstable if so, and backport if it causes no problems.

 -- Keir

> Michael Young

On Tue, 8 May 2012, Keir Fraser wrote:
> Thanks. The same appears to be true in xen-unstable also. Is the solution
to
> simply remove -O0 from the command line? We can test that out in
> xen-unstable if so, and backport if it causes no problems.
I got it to build with the attached patch -
http://koji.fedoraproject.org/koji/taskinfo?taskID=4063767 .

It seems the -O0 option was masking some other compile issues so I had to 
add -Wno-error=unused-result -Wno-error=array-bounds as well though I 
imagine the right solution would be to fix the code so it doesn''t give 
warnings (if they haven''t already been fixed in xen-unstable).

 	Michael Young

_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel

On 09/05/12 20:03, M A Young wrote:
> On Tue, 8 May 2012, Keir Fraser wrote:
>
>> Thanks. The same appears to be true in xen-unstable also. Is the
solution to
>> simply remove -O0 from the command line? We can test that out in
>> xen-unstable if so, and backport if it causes no problems.
> I got it to build with the attached patch -
> http://koji.fedoraproject.org/koji/taskinfo?taskID=4063767 .
>
> It seems the -O0 option was masking some other compile issues so I had to 
> add -Wno-error=unused-result -Wno-error=array-bounds as well though I 
> imagine the right solution would be to fix the code so it doesn''t
give
> warnings (if they haven''t already been fixed in xen-unstable).
>
>  	Michael Young
If it is failing -Warray-bounds then the source code probably needs fixing.

Do you have a log of the failures caused by -Warray-bounds ?

-- 
Andrew Cooper - Dom0 Kernel Engineer, Citrix XenServer
T: +44 (0)1223 225 900, http://www.citrix.com

On Wed, May 09, Andrew Cooper wrote:
> If it is failing -Warray-bounds then the source code probably needs fixing.
> 
> Do you have a log of the failures caused by -Warray-bounds ?
http://lists.xen.org/archives/html/xen-devel/2012-03/msg02583.html

Olaf

On Mon, May 07, Keir Fraser wrote:
> http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)
Keir,

the changes to unmodified_drivers//linux-2.6/ should be backported to
the 4.1 branch:

changeset:   25069:46bf3ab42baf
changeset:   25067:05768bd498f2

optional:
changeset:   24045:4ed766d70396
changeset:   25068:e4460795ee66

Also the asm/system.h patch I sent out today is a candidate.


Olaf

>>> On 11.05.12 at 17:51, Olaf Hering <olaf@aepfle.de> wrote:
> On Mon, May 07, Keir Fraser wrote:
> 
>> http://xenbits.xen.org/staging/xen-4.1-testing.hg (tag 4.1.3-rc1)
> 
> Keir,
> 
> the changes to unmodified_drivers//linux-2.6/ should be backported to
> the 4.1 branch:
> 
> changeset:   25069:46bf3ab42baf
> changeset:   25067:05768bd498f2
> 
> optional:
> changeset:   24045:4ed766d70396
> changeset:   25068:e4460795ee66
> 
> Also the asm/system.h patch I sent out today is a candidate.
25327:cc7a054a5a27

Jan