Use qemuMonitorTestNewFromFileFull which allows to test commands used
along with providing replies. This has two advantages:
1) It's easier to see which command was used when looking at the files
2) We check that the used commands are actually in the correct order
Signed-off-by: Peter Krempa <pkrempa@redhat.com>
---
tests/qemucapabilitiestest.c | 13 +++++--------
1 file changed, 5 insertions(+), 8 deletions(-)
diff --git a/tests/qemucapabilitiestest.c b/tests/qemucapabilitiestest.c
index 6981e973a6..633389f263 100644
--- a/tests/qemucapabilitiestest.c
+++ b/tests/qemucapabilitiestest.c
@@ -32,7 +32,7 @@
typedef struct _testQemuData testQemuData;
typedef testQemuData *testQemuDataPtr;
struct _testQemuData {
- virDomainXMLOptionPtr xmlopt;
+ virQEMUDriver driver;
const char *archName;
const char *base;
};
@@ -42,7 +42,7 @@ static int
testQemuCaps(const void *opaque)
{
int ret = -1;
- const testQemuData *data = opaque;
+ testQemuData *data = (void *) opaque;
char *repliesFile = NULL;
char *capsFile = NULL;
qemuMonitorTestPtr mon = NULL;
@@ -55,7 +55,7 @@ testQemuCaps(const void *opaque)
abs_srcdir, data->base, data->archName) < 0)
goto cleanup;
- if (!(mon = qemuMonitorTestNewFromFile(repliesFile, data->xmlopt, false)))
+ if (!(mon = qemuMonitorTestNewFromFileFull(repliesFile, &data->driver, NULL)))
goto cleanup;
if (!(capsActual = virQEMUCapsNew()) ||
@@ -139,7 +139,6 @@ static int
mymain(void)
{
int ret = 0;
- virQEMUDriver driver;
testQemuData data;
#if !WITH_YAJL
@@ -148,13 +147,11 @@ mymain(void)
#endif
if (virThreadInitialize() < 0 ||
- qemuTestDriverInit(&driver) < 0)
+ qemuTestDriverInit(&data.driver) < 0)
return EXIT_FAILURE;
virEventRegisterDefaultImpl();
- data.xmlopt = driver.xmlopt;
-
#define DO_TEST(arch, name) \
do { \
data.archName = arch; \
@@ -200,7 +197,7 @@ mymain(void)
* "tests/qemucapsfixreplies foo.replies" to fix the replies ids.
*/
- qemuTestDriverFree(&driver);
+ qemuTestDriverFree(&data.driver);
return (ret == 0) ? EXIT_SUCCESS : EXIT_FAILURE;
}
--
2.16.2
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
On 06/04/2018 09:58 AM, Peter Krempa wrote:
> Use qemuMonitorTestNewFromFileFull which allows to test commands used
> along with providing replies. This has two advantages:
>
> 1) It's easier to see which command was used when looking at the files
> 2) We check that the used commands are actually in the correct order
>
> Signed-off-by: Peter Krempa <pkrempa@redhat.com>
> ---
> tests/qemucapabilitiestest.c | 13 +++++--------
> 1 file changed, 5 insertions(+), 8 deletions(-)
>
With top of tree this one fails make check, but I assume that has to do
with changes in files since you posted... perhaps it's a combo of this
and the previous patch.
So I'll stop here and wait to see how things work themselves out.
John
> diff --git a/tests/qemucapabilitiestest.c b/tests/qemucapabilitiestest.c
> index 6981e973a6..633389f263 100644
> --- a/tests/qemucapabilitiestest.c
> +++ b/tests/qemucapabilitiestest.c
> @@ -32,7 +32,7 @@
> typedef struct _testQemuData testQemuData;
> typedef testQemuData *testQemuDataPtr;
> struct _testQemuData {
> - virDomainXMLOptionPtr xmlopt;
> + virQEMUDriver driver;
> const char *archName;
> const char *base;
> };
> @@ -42,7 +42,7 @@ static int
> testQemuCaps(const void *opaque)
> {
> int ret = -1;
> - const testQemuData *data = opaque;
> + testQemuData *data = (void *) opaque;
> char *repliesFile = NULL;
> char *capsFile = NULL;
> qemuMonitorTestPtr mon = NULL;
> @@ -55,7 +55,7 @@ testQemuCaps(const void *opaque)
> abs_srcdir, data->base, data->archName) < 0)
> goto cleanup;
>
> - if (!(mon = qemuMonitorTestNewFromFile(repliesFile, data->xmlopt, false)))
> + if (!(mon = qemuMonitorTestNewFromFileFull(repliesFile, &data->driver, NULL)))
> goto cleanup;
>
> if (!(capsActual = virQEMUCapsNew()) ||
> @@ -139,7 +139,6 @@ static int
> mymain(void)
> {
> int ret = 0;
> - virQEMUDriver driver;
> testQemuData data;
>
> #if !WITH_YAJL
> @@ -148,13 +147,11 @@ mymain(void)
> #endif
>
> if (virThreadInitialize() < 0 ||
> - qemuTestDriverInit(&driver) < 0)
> + qemuTestDriverInit(&data.driver) < 0)
> return EXIT_FAILURE;
>
> virEventRegisterDefaultImpl();
>
> - data.xmlopt = driver.xmlopt;
> -
> #define DO_TEST(arch, name) \
> do { \
> data.archName = arch; \
> @@ -200,7 +197,7 @@ mymain(void)
> * "tests/qemucapsfixreplies foo.replies" to fix the replies ids.
> */
>
> - qemuTestDriverFree(&driver);
> + qemuTestDriverFree(&data.driver);
>
> return (ret == 0) ? EXIT_SUCCESS : EXIT_FAILURE;
> }
>
--
libvir-list mailing list
libvir-list@redhat.com
https://www.redhat.com/mailman/listinfo/libvir-list
On Thu, Jun 07, 2018 at 22:02:06 -0400, John Ferlan wrote: > > > On 06/04/2018 09:58 AM, Peter Krempa wrote: > > Use qemuMonitorTestNewFromFileFull which allows to test commands used > > along with providing replies. This has two advantages: > > > > 1) It's easier to see which command was used when looking at the files > > 2) We check that the used commands are actually in the correct order > > > > Signed-off-by: Peter Krempa <pkrempa@redhat.com> > > --- > > tests/qemucapabilitiestest.c | 13 +++++-------- > > 1 file changed, 5 insertions(+), 8 deletions(-) > > > > With top of tree this one fails make check, but I assume that has to do > with changes in files since you posted... perhaps it's a combo of this > and the previous patch. Well, it's probably because you also need the next patch (which is supposed to be squashed into this one) for the tests to pass. I've separatet it because all the changes were generated by the regeneration tool. -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
On 06/08/2018 03:45 AM, Peter Krempa wrote: > On Thu, Jun 07, 2018 at 22:02:06 -0400, John Ferlan wrote: >> >> >> On 06/04/2018 09:58 AM, Peter Krempa wrote: >>> Use qemuMonitorTestNewFromFileFull which allows to test commands used >>> along with providing replies. This has two advantages: >>> >>> 1) It's easier to see which command was used when looking at the files >>> 2) We check that the used commands are actually in the correct order >>> >>> Signed-off-by: Peter Krempa <pkrempa@redhat.com> >>> --- >>> tests/qemucapabilitiestest.c | 13 +++++-------- >>> 1 file changed, 5 insertions(+), 8 deletions(-) >>> >> >> With top of tree this one fails make check, but I assume that has to do >> with changes in files since you posted... perhaps it's a combo of this >> and the previous patch. > > Well, it's probably because you also need the next patch (which is > supposed to be squashed into this one) for the tests to pass. I've > separatet it because all the changes were generated by the regeneration > tool. > Ah, right, I see. Anyway, I figured Martin's patches would change the results here. There's also the SEV patches that have capability changes which bring up a few ordering concerns in which it may be "easier" if these last couple of patches waited for at least the SEV ones to be merged. From recent reviews, I'm aware of 2 other series that are perhaps impacted (Erik's "Enable vfio-pci 'property' for mediated devices" which perhaps just needs a v2 and Jie Wang's "Introduce align for hostmem-file", but that one is not close to being ready yet). While I understand ordering and merges are someone else's problem - still need to play nice though especially if there's any back port concerns 0-). BTW: Should we take this "opportunity" to finally update the 2.12 files to the released version rather than the rc0 version? (only s390x got a final update). As for the changes themselves, consider both patch 8 and 9, Reviewed-by: John Ferlan <jferlan@redhat.com> as long as the commit message is cleaned up a bit. John -- libvir-list mailing list libvir-list@redhat.com https://www.redhat.com/mailman/listinfo/libvir-list
© 2016 - 2025 Red Hat, Inc.