Re: [PATCH] qga: fix a memory leak in qmp_guest_exec_status()

Re: [PATCH] qga: fix a memory leak in qmp_guest_exec_status()

[Marc-André Lureau]

[-- Attachment #1: Type: text/plain, Size: 1883 bytes --]

Hi

On Wed, Jun 30, 2021 at 11:42 AM Yi Wang <wang.yi59@zte.com.cn> wrote:

> From: Wang Yechao <wang.yechao255@zte.com.cn>
>
> The $GuestExecInfo.out.data is alloced in guest_exec_output_watch(),
> and the buffer size is $GuestExecInfo.out.size. We should free the
> $GuestExecInfo.out.data judge by the size, not length. Because the
> $GuestExecInfo.out.length maybe zero in some case.
>
> $GuestExecInfo.err.data has the same problem.
>
> Signed-off-by: Yechao Wang <wang.yechao255@zte.com.cn>
> Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
> ---
>  qga/commands.c | 8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
>
> diff --git a/qga/commands.c b/qga/commands.c
> index d3fec80..58d4da9 100644
> --- a/qga/commands.c
> +++ b/qga/commands.c
>
> @@ -214,17 +214,21 @@ GuestExecStatus *qmp_guest_exec_status(int64_t pid, Error **errp)
>          if (gei->out.length > 0) {
>              ges->has_out_data = true;
>
>              ges->out_data = g_base64_encode(gei->out.data, gei->out.length);
> -            g_free(gei->out.data);
>              ges->has_out_truncated = gei->out.truncated;
>          }
>
>          if (gei->err.length > 0) {
>              ges->has_err_data = true;
>
>              ges->err_data = g_base64_encode(gei->err.data, gei->err.length);
> -            g_free(gei->err.data);
>              ges->has_err_truncated = gei->err.truncated;
>          }
>
> +        if (gei->out.size > 0)
> +            g_free(gei->out.data);
> +
> +        if (gei->err.size > 0)
> +             g_free(gei->err.data);
>

You shouldn't need to check if size > 0. You can just call g_free() on the
fields (NULL is ignored).

Looks good otherwise

+
>          QTAILQ_REMOVE(&guest_exec_state.processes, gei, next);
>          g_free(gei);
>      }
> --
> 1.8.3.1
>


-- 
Marc-André Lureau

[-- Attachment #2: Type: text/html, Size: 3041 bytes --]

[PATCH] qga: fix a memory leak in qmp_guest_exec_status()

[Yi Wang]

[-- Attachment #1.1.1: Type: text/plain, Size: 1547 bytes --]

From: Wang Yechao <wang.yechao255@zte.com.cn>

The $GuestExecInfo.out.data is alloced in guest_exec_output_watch(),
and the buffer size is $GuestExecInfo.out.size. We should free the
$GuestExecInfo.out.data judge by the size, not length. Because the
$GuestExecInfo.out.length maybe zero in some case.

$GuestExecInfo.err.data has the same problem.

Signed-off-by: Yechao Wang <wang.yechao255@zte.com.cn>
Signed-off-by: Yi Wang <wang.yi59@zte.com.cn>
---
 qga/commands.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/qga/commands.c b/qga/commands.c
index d3fec80..58d4da9 100644
--- a/qga/commands.c
+++ b/qga/commands.c
@@ -214,17 +214,21 @@ GuestExecStatus *qmp_guest_exec_status(int64_t pid, Error **errp)
         if (gei->out.length > 0) {
             ges->has_out_data = true;
             ges->out_data = g_base64_encode(gei->out.data, gei->out.length);
-            g_free(gei->out.data);
             ges->has_out_truncated = gei->out.truncated;
         }
 
         if (gei->err.length > 0) {
             ges->has_err_data = true;
             ges->err_data = g_base64_encode(gei->err.data, gei->err.length);
-            g_free(gei->err.data);
             ges->has_err_truncated = gei->err.truncated;
         }
 
+        if (gei->out.size > 0)
+            g_free(gei->out.data);
+
+        if (gei->err.size > 0)
+             g_free(gei->err.data);
+
         QTAILQ_REMOVE(&guest_exec_state.processes, gei, next);
         g_free(gei);
     }
-- 
1.8.3.1

[-- Attachment #1.1.2: Type: text/html , Size: 3338 bytes --]