On 07/10/2016 18:38, Greg Kurz wrote:
> On Sat,  1 Oct 2016 20:56:02 +0200
> Laurent Vivier <lviv...@redhat.com> wrote:
> 
>> This patch replaces calls to qtest_start() and qtest_end() by
>> calls to qtest_pc_boot() and qtest_shutdown().
>>
>> This allows to initialize memory allocator and PCI interface
>> functions. This will ease to enable virtio tests on other
>> architectures by only adding a specific qtest_XXX_boot() (like
>> qtest_spapr_boot()).
>>
>> Signed-off-by: Laurent Vivier <lviv...@redhat.com>
>> Reviewed-by: Greg Kurz <gr...@kaod.org>
>> ---
> 
> Oops I had missed it during review but I have a single remark for the 9p
> test.
> 
> My R-b stands anyway.
> 
>>  tests/virtio-9p-test.c   |  51 +++++++---------
>>  tests/virtio-blk-test.c  | 150 
>> ++++++++++++++++++++---------------------------
>>  tests/virtio-net-test.c  |  39 +++++-------
>>  tests/virtio-scsi-test.c |  67 +++++++++------------
>>  4 files changed, 129 insertions(+), 178 deletions(-)
>>
>> diff --git a/tests/virtio-9p-test.c b/tests/virtio-9p-test.c
>> index e8b2196..7698014 100644
>> --- a/tests/virtio-9p-test.c
>> +++ b/tests/virtio-9p-test.c
>> @@ -10,62 +10,56 @@
>>  #include "qemu/osdep.h"
>>  #include "libqtest.h"
>>  #include "qemu-common.h"
>> -#include "libqos/pci-pc.h"
>> +#include "libqos/libqos-pc.h"
>>  #include "libqos/virtio.h"
>>  #include "libqos/virtio-pci.h"
>> -#include "libqos/malloc.h"
>> -#include "libqos/malloc-pc.h"
>>  #include "standard-headers/linux/virtio_ids.h"
>>  #include "standard-headers/linux/virtio_pci.h"
>>  
>>  static const char mount_tag[] = "qtest";
>>  static char *test_share;
>>  
>> -static void qvirtio_9p_start(void)
>> -{
>> -    char *args;
>>  
>> +static QOSState *qvirtio_9p_start(void)
>> +{
>>      test_share = g_strdup("/tmp/qtest.XXXXXX");
>>      g_assert_nonnull(mkdtemp(test_share));
>> +    const char *cmd = "-fsdev local,id=fsdev0,security_model=none,path=%s "
>> +                      "-device virtio-9p-pci,fsdev=fsdev0,mount_tag=%s";
>>  
> 
> Even if C99 supports declarations within the code, I prefer the old school way
> of declaring variables followed by an empty line, at the top.

Yes, you're right, I prefer too...

I will resend the series next week.

Thanks,
Laurent


Reply via email to