tools/tests/cpu-policy/Makefile | 14 +++++++++++++- tools/tests/x86_emulator/Makefile | 2 +- 2 files changed, 14 insertions(+), 2 deletions(-)
GCC 4.4 (as included in CentOS 6) is too old to handle designated initialisers
in anonymous unions. As this is just a developer tool, skip the test in this
case, rather than sacraficing the legibility/expresibility of the test cases.
This fixes the Gitlab CI tests.
While adding this logic to cpu-polcy, adjust the equivelent logic from
x86_emulator on which this was based. Printing:
Test harness not built, use newer compiler than "gcc"
isn't helpful for anyone unexpectedly encountering the error.
Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
---
CC: Jan Beulich <JBeulich@suse.com>
CC: Wei Liu <wei.liu2@citrix.com>
CC: Roger Pau Monné <roger.pau@citrix.com>
CC: Ian Jackson <Ian.Jackson@citrix.com>
---
tools/tests/cpu-policy/Makefile | 14 +++++++++++++-
tools/tests/x86_emulator/Makefile | 2 +-
2 files changed, 14 insertions(+), 2 deletions(-)
diff --git a/tools/tests/cpu-policy/Makefile b/tools/tests/cpu-policy/Makefile
index eeed7f3..4b6caec 100644
--- a/tools/tests/cpu-policy/Makefile
+++ b/tools/tests/cpu-policy/Makefile
@@ -1,8 +1,20 @@
XEN_ROOT = $(CURDIR)/../../..
include $(XEN_ROOT)/tools/Rules.mk
+TARGET-y := test-cpu-policy
+
+# For brevity, these tests make extensive use of designated initialisers, but
+# GCCs older than 4.6 can't cope. Ignore the test in this case.
+ifneq ($(clang),y)
+TARGET-$(call cc-ver,$(CC),lt,0x040600) :=
+endif
+
+ifeq ($(TARGET-y),)
+$(warning Test harness not built, use newer compiler than $(CC) $(shell $(CC) -dumpversion))
+endif
+
.PHONY: all
-all: test-cpu-policy
+all: $(TARGET-y)
.PHONY: clean
clean:
diff --git a/tools/tests/x86_emulator/Makefile b/tools/tests/x86_emulator/Makefile
index 4f4c0f6..970ec3e 100644
--- a/tools/tests/x86_emulator/Makefile
+++ b/tools/tests/x86_emulator/Makefile
@@ -97,7 +97,7 @@ $(foreach flavor,$(SIMD) $(FMA),$(eval $(call simd-check-cc,$(flavor))))
TARGET-$(shell echo 'asm("{evex} vzeroall");' | $(CC) -x c -c -o /dev/null - || echo y) :=
ifeq ($(TARGET-y),)
-$(warning Test harness not built, use newer compiler than "$(CC)")
+$(warning Test harness not built, use newer compiler than $(CC) $(shell $(CC) -dumpversion) and an "{evex}" capable assembler)
endif
all: $(TARGET-y)
--
2.1.4
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
On Fri, May 24, 2019 at 02:29:46PM +0100, Andrew Cooper wrote: > GCC 4.4 (as included in CentOS 6) is too old to handle designated initialisers > in anonymous unions. As this is just a developer tool, skip the test in this > case, rather than sacraficing the legibility/expresibility of the test cases. > > This fixes the Gitlab CI tests. > > While adding this logic to cpu-polcy, adjust the equivelent logic from > x86_emulator on which this was based. Printing: > > Test harness not built, use newer compiler than "gcc" > > isn't helpful for anyone unexpectedly encountering the error. > > Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com> Reviewed-by: Wei Liu <wei.liu2@citrix.com> _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
>>> On 24.05.19 at 15:29, <andrew.cooper3@citrix.com> wrote:
> GCC 4.4 (as included in CentOS 6) is too old to handle designated initialisers
> in anonymous unions. As this is just a developer tool, skip the test in this
> case, rather than sacraficing the legibility/expresibility of the test cases.
>
> This fixes the Gitlab CI tests.
>
> While adding this logic to cpu-polcy, adjust the equivelent logic from
> x86_emulator on which this was based. Printing:
>
> Test harness not built, use newer compiler than "gcc"
>
> isn't helpful for anyone unexpectedly encountering the error.
>
> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
Fundamentally
Reviewed-by: Jan Beulich <jbeulich@suse.com>
But there are remarks:
> --- a/tools/tests/cpu-policy/Makefile
> +++ b/tools/tests/cpu-policy/Makefile
> @@ -1,8 +1,20 @@
> XEN_ROOT = $(CURDIR)/../../..
> include $(XEN_ROOT)/tools/Rules.mk
>
> +TARGET-y := test-cpu-policy
> +
> +# For brevity, these tests make extensive use of designated initialisers, but
> +# GCCs older than 4.6 can't cope. Ignore the test in this case.
Designated initializers alone are no problem for old gcc. The issue is
with ones used for sub-structures/-unions without field name.
Perhaps worth slightly extending the comment to this effect?
> --- a/tools/tests/x86_emulator/Makefile
> +++ b/tools/tests/x86_emulator/Makefile
> @@ -97,7 +97,7 @@ $(foreach flavor,$(SIMD) $(FMA),$(eval $(call simd-check-cc,$(flavor))))
> TARGET-$(shell echo 'asm("{evex} vzeroall");' | $(CC) -x c -c -o /dev/null
> - || echo y) :=
>
> ifeq ($(TARGET-y),)
> -$(warning Test harness not built, use newer compiler than "$(CC)")
> +$(warning Test harness not built, use newer compiler than $(CC) $(shell $(CC) -dumpversion) and an "{evex}" capable assembler)
> endif
I appreciate the idea of providing mode information, but I'm afraid
this is going to be clumsy in the other direction now:
"Test harness not built, use newer compiler than gcc-4.8 4.8 and ..."
Naming the compiler binary, I thought, allows the user to figure
out the version easily enough. Therefore, please consider
dropping that part again.
I'm unconditionally fine with the {evex} addition.
Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
On 24/05/2019 15:19, Jan Beulich wrote:
>>>> On 24.05.19 at 15:29, <andrew.cooper3@citrix.com> wrote:
>> GCC 4.4 (as included in CentOS 6) is too old to handle designated initialisers
>> in anonymous unions. As this is just a developer tool, skip the test in this
>> case, rather than sacraficing the legibility/expresibility of the test cases.
>>
>> This fixes the Gitlab CI tests.
>>
>> While adding this logic to cpu-polcy, adjust the equivelent logic from
>> x86_emulator on which this was based. Printing:
>>
>> Test harness not built, use newer compiler than "gcc"
>>
>> isn't helpful for anyone unexpectedly encountering the error.
>>
>> Signed-off-by: Andrew Cooper <andrew.cooper3@citrix.com>
> Fundamentally
> Reviewed-by: Jan Beulich <jbeulich@suse.com>
> But there are remarks:
>
>> --- a/tools/tests/cpu-policy/Makefile
>> +++ b/tools/tests/cpu-policy/Makefile
>> @@ -1,8 +1,20 @@
>> XEN_ROOT = $(CURDIR)/../../..
>> include $(XEN_ROOT)/tools/Rules.mk
>>
>> +TARGET-y := test-cpu-policy
>> +
>> +# For brevity, these tests make extensive use of designated initialisers, but
>> +# GCCs older than 4.6 can't cope. Ignore the test in this case.
> Designated initializers alone are no problem for old gcc. The issue is
> with ones used for sub-structures/-unions without field name.
> Perhaps worth slightly extending the comment to this effect?
" in anonymous unions" ? I can never remember exactly which bit it
chokes on, but I think there are two different ones in practice which
interfere.
>
>> --- a/tools/tests/x86_emulator/Makefile
>> +++ b/tools/tests/x86_emulator/Makefile
>> @@ -97,7 +97,7 @@ $(foreach flavor,$(SIMD) $(FMA),$(eval $(call simd-check-cc,$(flavor))))
>> TARGET-$(shell echo 'asm("{evex} vzeroall");' | $(CC) -x c -c -o /dev/null
>> - || echo y) :=
>>
>> ifeq ($(TARGET-y),)
>> -$(warning Test harness not built, use newer compiler than "$(CC)")
>> +$(warning Test harness not built, use newer compiler than $(CC) $(shell $(CC) -dumpversion) and an "{evex}" capable assembler)
>> endif
> I appreciate the idea of providing mode information, but I'm afraid
> this is going to be clumsy in the other direction now:
>
> "Test harness not built, use newer compiler than gcc-4.8 4.8 and ..."
>
> Naming the compiler binary, I thought, allows the user to figure
> out the version easily enough. Therefore, please consider
> dropping that part again.
I'm afraid you have a selection bias here. Your compiler binaries may
have a version suffix, but the overwhelming majority of people who are
going to hit that error and need to figure out what to do will be using
their system-provided binaries, as per the commit message.
What about:
... than "$(CC)" (version $(shell $(CC) -dumpversion)) and ...
which should (in your example) render as:
... than "gcc-4.8" (version 4.8) and ...
?
~Andrew
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
>>> On 24.05.19 at 16:43, <andrew.cooper3@citrix.com> wrote:
> On 24/05/2019 15:19, Jan Beulich wrote:
>>>>> On 24.05.19 at 15:29, <andrew.cooper3@citrix.com> wrote:
>>> --- a/tools/tests/cpu-policy/Makefile
>>> +++ b/tools/tests/cpu-policy/Makefile
>>> @@ -1,8 +1,20 @@
>>> XEN_ROOT = $(CURDIR)/../../..
>>> include $(XEN_ROOT)/tools/Rules.mk
>>>
>>> +TARGET-y := test-cpu-policy
>>> +
>>> +# For brevity, these tests make extensive use of designated initialisers, but
>>> +# GCCs older than 4.6 can't cope. Ignore the test in this case.
>> Designated initializers alone are no problem for old gcc. The issue is
>> with ones used for sub-structures/-unions without field name.
>> Perhaps worth slightly extending the comment to this effect?
>
> " in anonymous unions" ? I can never remember exactly which bit it
> chokes on, but I think there are two different ones in practice which
> interfere.
" in anonymous unions" is fine with me.
>>> --- a/tools/tests/x86_emulator/Makefile
>>> +++ b/tools/tests/x86_emulator/Makefile
>>> @@ -97,7 +97,7 @@ $(foreach flavor,$(SIMD) $(FMA),$(eval $(call simd-check-cc,$(flavor))))
>>> TARGET-$(shell echo 'asm("{evex} vzeroall");' | $(CC) -x c -c -o /dev/null > - || echo y) :=
>>>
>>> ifeq ($(TARGET-y),)
>>> -$(warning Test harness not built, use newer compiler than "$(CC)")
>>> +$(warning Test harness not built, use newer compiler than $(CC) $(shell $(CC) -dumpversion) and an "{evex}" capable assembler)
>>> endif
>> I appreciate the idea of providing mode information, but I'm afraid
>> this is going to be clumsy in the other direction now:
>>
>> "Test harness not built, use newer compiler than gcc-4.8 4.8 and ..."
>>
>> Naming the compiler binary, I thought, allows the user to figure
>> out the version easily enough. Therefore, please consider
>> dropping that part again.
>
> I'm afraid you have a selection bias here. Your compiler binaries may
> have a version suffix, but the overwhelming majority of people who are
> going to hit that error and need to figure out what to do will be using
> their system-provided binaries, as per the commit message.
Well, I can only judge by what the distro does that I use; I wasn't
aware they do something non-standard. I've already avoided
mentioning my own compiler naming scheme.
> What about:
>
> ... than "$(CC)" (version $(shell $(CC) -dumpversion)) and ...
>
> which should (in your example) render as:
>
> ... than "gcc-4.8" (version 4.8) and ...
>
> ?
Better, so let's go with this.
Jan
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xenproject.org
https://lists.xenproject.org/mailman/listinfo/xen-devel
© 2016 - 2026 Red Hat, Inc.