Re: [Xen-devel] [xen-4.4-testing test] 65556: regressions - trouble: blocked/broken/fail/pass

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 09.12.15 at 11:13, <JBeulich@xxxxxxxx> wrote:
>>>> On 09.12.15 at 10:47, <JBeulich@xxxxxxxx> wrote:
>>>>> On 09.12.15 at 10:28, <osstest-admin@xxxxxxxxxxxxxx> wrote:
>>> flight 65556 xen-4.4-testing real [real]
>>> http://logs.test-lab.xenproject.org/osstest/logs/65556/ 
>>> 
>>> Regressions :-(
>>> 
>>> Tests which did not succeed and are blocking,
>>> including tests which could not be run:
>>>  test-amd64-amd64-xl-qcow2     3 host-install(3)         broken REGR. vs. 
>>> 64766
>>>  build-armhf                   5 xen-build                 fail REGR. vs. 
> 64766
>> 
>> I'm looking into this (also affecting 4.3, but oddly enough not 4.5
>> and newer).
> 
> I guess it is related to ptdom_max_order being an alias of
> domu_max_order in the pre-4.5 case (due to HAS_PASSTHROUGH
> not being defined there). Looks like a compiler issue though, since
> the return value is getting checked - just that the value used is the
> same in both cases.

And indeed 5.3.0 on this code

int __attribute__((warn_unused_result)) check(void);

unsigned test1(void) {
        return check() ? 0 : 0;
}

unsigned test2(void) {
        if(check())
                return 0;
        return 0;
}

unsigned test3(void) {
        unsigned r = 0;
        if(check())
                r = 0;
        return r;
}

issues a warning in test1() but not in test2() or test3(). I'll craft a
workaround to that effect, after entering a compiler bug report.

Jan


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