Luiz Capitulino <lcapitul...@redhat.com> writes:
> On Tue, 5 Feb 2013 17:22:07 +0100
> Markus Armbruster <arm...@redhat.com> wrote:
>
>>
>> Signed-off-by: Markus Armbruster <arm...@redhat.com>
>> ---
>> qemu-char.c | 5 +++++
>> 1 file changed, 5 insertions(+)
>>
>> diff --git a/qemu-char.c b/qemu-char.c
>> index 892d5bb..9bf53e0 100644
>> --- a/qemu-char.c
>> +++ b/qemu-char.c
>> @@ -2782,6 +2782,10 @@ void qmp_memchar_write(const char *device, const char
>> *data,
>>
>> ret = cirmem_chr_write(chr, write_data, write_count);
>>
>> + if (write_data != (uint8_t *)data) {
>
> Minor, but wouldn't it be clearer to check against has_format and format?
I actually like it this way, because it makes it blatantly obvious that
write_data is either freed or owned by the caller.
>> + g_free((void *)write_data);
>> + }
>> +
>> if (ret < 0) {
>> error_setg(errp, "Failed to write to device %s", device);
>> return;
>> @@ -2825,6 +2829,7 @@ char *qmp_memchar_read(const char *device, int64_t
>> size,
>>
>> if (has_format && (format == DATA_FORMAT_BASE64)) {
>> data = g_base64_encode(read_data, count);
>> + g_free(read_data);
>> } else {
>> data = (char *)read_data;
>> }
