Message ID | 25db4e01-c391-fd32-af40-70bbedcc2278@canonical.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
>>> On 10.03.17 at 18:22, <stefan.bader@canonical.com> wrote: > On 08.03.2017 13:54, Jan Beulich wrote: >> All, >> >> I am pleased to announce the release of Xen 4.6.5. This is >> available immediately from its git repository >> http://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=refs/heads/stable-4.6 >> (tag RELEASE-4.6.5) or from the XenProject download page >> http://www.xenproject.org/downloads/xen-archives/xen-46-series/xen-465.html >> (where a list of changes can also be found). >> >> We recommend all users of the 4.6 stable series to update to this >> latest point release. > > This does not seem to compile for me (x86_64) without the attached > (admittedly > brutish) change. I guess it's the emulator test code which has a problem here (I did notice this myself), but that doesn't get built by default (and I see no reason why anyone would want to build it when putting together packages for people to consume - this is purely a dev tool). Please clarify. Jan
On 13/03/17 09:24, Jan Beulich wrote: >>>> On 10.03.17 at 18:22, <stefan.bader@canonical.com> wrote: >> On 08.03.2017 13:54, Jan Beulich wrote: >>> All, >>> >>> I am pleased to announce the release of Xen 4.6.5. This is >>> available immediately from its git repository >>> http://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=refs/heads/stable-4.6 >>> (tag RELEASE-4.6.5) or from the XenProject download page >>> http://www.xenproject.org/downloads/xen-archives/xen-46-series/xen-465.html >>> (where a list of changes can also be found). >>> >>> We recommend all users of the 4.6 stable series to update to this >>> latest point release. >> This does not seem to compile for me (x86_64) without the attached >> (admittedly >> brutish) change. > I guess it's the emulator test code which has a problem here (I > did notice this myself), but that doesn't get built by default (and > I see no reason why anyone would want to build it when putting > together packages for people to consume - this is purely a dev > tool). Please clarify. These tools are all built automatically. Therefore, build fixes should be backported. To avoid building them, you need override CONFIG_TESTS := n in the root .config file to override the default in Config.mk ~Andrew
On 13.03.2017 11:29, Andrew Cooper wrote: > On 13/03/17 09:24, Jan Beulich wrote: >>>>> On 10.03.17 at 18:22, <stefan.bader@canonical.com> wrote: >>> On 08.03.2017 13:54, Jan Beulich wrote: >>>> All, >>>> >>>> I am pleased to announce the release of Xen 4.6.5. This is >>>> available immediately from its git repository >>>> http://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=refs/heads/stable-4.6 >>>> (tag RELEASE-4.6.5) or from the XenProject download page >>>> http://www.xenproject.org/downloads/xen-archives/xen-46-series/xen-465.html >>>> (where a list of changes can also be found). >>>> >>>> We recommend all users of the 4.6 stable series to update to this >>>> latest point release. >>> This does not seem to compile for me (x86_64) without the attached >>> (admittedly >>> brutish) change. >> I guess it's the emulator test code which has a problem here (I >> did notice this myself), but that doesn't get built by default (and >> I see no reason why anyone would want to build it when putting >> together packages for people to consume - this is purely a dev >> tool). Please clarify. > > These tools are all built automatically. Therefore, build fixes should > be backported. > > To avoid building them, you need override CONFIG_TESTS := n in the root > .config file to override the default in Config.mk Thanks Andrew, I was not sure but I did not do anything special except replacing the orig tarballs. The rest of the build is as we share it with Debian. So for a minor release / stable release update I would rather not change the environment. For the patch I just copied the definition from lib.h because gcc seems to be called without access to hypervisor includes (probably adapting the Makefile plus adding an include would be the better path but it was late'ish on a Friday and I wanted something compiling quickly). -Stefan
>>> On 13.03.17 at 11:29, <andrew.cooper3@citrix.com> wrote: > On 13/03/17 09:24, Jan Beulich wrote: >>>>> On 10.03.17 at 18:22, <stefan.bader@canonical.com> wrote: >>> On 08.03.2017 13:54, Jan Beulich wrote: >>>> All, >>>> >>>> I am pleased to announce the release of Xen 4.6.5. This is >>>> available immediately from its git repository >>>> http://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=refs/heads/stable-4.6 >>>> (tag RELEASE-4.6.5) or from the XenProject download page >>>> http://www.xenproject.org/downloads/xen-archives/xen-46-series/xen-465.html >>>> (where a list of changes can also be found). >>>> >>>> We recommend all users of the 4.6 stable series to update to this >>>> latest point release. >>> This does not seem to compile for me (x86_64) without the attached >>> (admittedly >>> brutish) change. >> I guess it's the emulator test code which has a problem here (I >> did notice this myself), but that doesn't get built by default (and >> I see no reason why anyone would want to build it when putting >> together packages for people to consume - this is purely a dev >> tool). Please clarify. > > These tools are all built automatically. If so, how come osstest didn't notice the issue (long ago)? Jan
On 13/03/17 10:59, Jan Beulich wrote: >>>> On 13.03.17 at 11:29, <andrew.cooper3@citrix.com> wrote: >> On 13/03/17 09:24, Jan Beulich wrote: >>>>>> On 10.03.17 at 18:22, <stefan.bader@canonical.com> wrote: >>>> On 08.03.2017 13:54, Jan Beulich wrote: >>>>> All, >>>>> >>>>> I am pleased to announce the release of Xen 4.6.5. This is >>>>> available immediately from its git repository >>>>> http://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=refs/heads/stable-4.6 >>>>> (tag RELEASE-4.6.5) or from the XenProject download page >>>>> http://www.xenproject.org/downloads/xen-archives/xen-46-series/xen-465.html >>>>> (where a list of changes can also be found). >>>>> >>>>> We recommend all users of the 4.6 stable series to update to this >>>>> latest point release. >>>> This does not seem to compile for me (x86_64) without the attached >>>> (admittedly >>>> brutish) change. >>> I guess it's the emulator test code which has a problem here (I >>> did notice this myself), but that doesn't get built by default (and >>> I see no reason why anyone would want to build it when putting >>> together packages for people to consume - this is purely a dev >>> tool). Please clarify. >> These tools are all built automatically. > If so, how come osstest didn't notice the issue (long ago)? Because OSSTest highly customises its build. It doesn't test the default. (but really should...) ~Andrew
>>> On 13.03.17 at 12:22, <andrew.cooper3@citrix.com> wrote: > On 13/03/17 10:59, Jan Beulich wrote: >>>>> On 13.03.17 at 11:29, <andrew.cooper3@citrix.com> wrote: >>> On 13/03/17 09:24, Jan Beulich wrote: >>>>>>> On 10.03.17 at 18:22, <stefan.bader@canonical.com> wrote: >>>>> On 08.03.2017 13:54, Jan Beulich wrote: >>>>>> All, >>>>>> >>>>>> I am pleased to announce the release of Xen 4.6.5. This is >>>>>> available immediately from its git repository >>>>>> http://xenbits.xen.org/gitweb/?p=xen.git;a=shortlog;h=refs/heads/stable-4.6 >>>>>> (tag RELEASE-4.6.5) or from the XenProject download page >>>>>> http://www.xenproject.org/downloads/xen-archives/xen-46-series/xen-465.html >>>>>> (where a list of changes can also be found). >>>>>> >>>>>> We recommend all users of the 4.6 stable series to update to this >>>>>> latest point release. >>>>> This does not seem to compile for me (x86_64) without the attached >>>>> (admittedly >>>>> brutish) change. >>>> I guess it's the emulator test code which has a problem here (I >>>> did notice this myself), but that doesn't get built by default (and >>>> I see no reason why anyone would want to build it when putting >>>> together packages for people to consume - this is purely a dev >>>> tool). Please clarify. >>> These tools are all built automatically. >> If so, how come osstest didn't notice the issue (long ago)? > > Because OSSTest highly customises its build. Well, I don't customize the default, and I don't see any build failure either (I simply had the desire to run the harness on one of the backports, which made me notice this issue). I can't even seem to spot where CONFIG_TESTS is being (or supposed to be) set. In fact grep-ing tools/ and config/ for TESTS yields a rather limited set of hits. Jan
Index: xen-4.6.5/xen/arch/x86/x86_emulate/x86_emulate.c =================================================================== --- xen-4.6.5.orig/xen/arch/x86/x86_emulate/x86_emulate.c +++ xen-4.6.5/xen/arch/x86/x86_emulate/x86_emulate.c @@ -20,6 +20,8 @@ * along with this program; If not, see <http://www.gnu.org/licenses/>. */ +#define MASK_EXTR(v, m) (((v) & (m)) / ((m) & -(m))) + /* Operand sizes: 8-bit operands or specified/overridden size. */ #define ByteOp (1<<0) /* 8-bit operands. */ /* Destination operand type. */