[Spice-devel] [PATCH spice-server 1/8] test-stream-device: Better Qemu emulation for data reading

[Christophe Fergeau]

Acked-by: Christophe Fergeau <cfergeau at redhat.com>

couple of comments on the comments below

On Sun, Feb 18, 2018 at 06:58:07PM +0000, Frediano Ziglio wrote:
> Qemu does not trigger a new data read if we don't read all data in
> the buffer.
> 
> Signed-off-by: Frediano Ziglio <fziglio at redhat.com>
> ---
>  server/tests/test-stream-device.c | 43 +++++++++++++++++++++++++++++++++++++--
>  1 file changed, 41 insertions(+), 2 deletions(-)
> 
> diff --git a/server/tests/test-stream-device.c b/server/tests/test-stream-device.c
> index 656bf56b..edd618e2 100644
> --- a/server/tests/test-stream-device.c
> +++ b/server/tests/test-stream-device.c
> @@ -23,6 +23,7 @@
>  #include <string.h>
>  #include <stdlib.h>
>  #include <stdio.h>
> +#include <unistd.h>
>  
>  #include <spice/protocol.h>
>  #include <spice/stream-device.h>
> @@ -68,8 +69,12 @@ static int vmc_read(SPICE_GNUC_UNUSED SpiceCharDeviceInstance *sin,
>      pos += ret;
>      // kick off next message read
>      // currently Qemu kicks the device so we need to do it manually
> -    // here
> -    spice_server_char_device_wakeup(&vmc_instance);
> +    // here. If not all data are read the device goes into blocking

comma after "all data are read"?

> +    // state and we get the wake only when we read from the device
> +    // again
> +    if (pos >= *message_sizes_curr) {
> +        spice_server_char_device_wakeup(&vmc_instance);
> +    }
>      return ret;
>  }
>  
> @@ -177,11 +182,45 @@ static void test_stream_device(void)
>      basic_event_loop_destroy();
>  }
>  
> +// check if sending a partial message cause issues

"causes"

> +static void test_stream_device_unfinished(void)
> +{
> +    uint8_t *p = message;
> +    SpiceCoreInterface *core = basic_event_loop_init();
> +    Test *test = test_new(core);
> +
> +    pos = 0;
> +    message_sizes_curr = message_sizes;
> +    message_sizes_end = message_sizes;
> +
> +    // this long and not finished message should not cause an infinite loop
> +    p = add_stream_hdr(p, STREAM_TYPE_DATA, 100000);
> +    *message_sizes_end = p - message;
> +    ++message_sizes_end;
> +
> +    vmc_instance.base.sif = &vmc_interface.base;
> +    spice_server_add_interface(test->server, &vmc_instance.base);
> +
> +    // we need to open the device and kick the start
> +    // the alarm is to avoid program to stuck

"to prevent the program from getting stuck" ? (same comment on the next
patch).

> +    alarm(5);
> +    spice_server_port_event(&vmc_instance, SPICE_PORT_EVENT_OPENED);
> +    spice_server_char_device_wakeup(&vmc_instance);
> +    alarm(0);
> +
> +    // we should read all data

"we should have read all data"?


Christophe
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/spice-devel/attachments/20180223/5fa145ee/attachment.sig>