[PATCH BlueZ 1/2] android/system-emulator: Pass --track-origin=yes to valgrind

[PATCH BlueZ 1/2] android/system-emulator: Pass --track-origin=yes to valgrind

[Luiz Augusto von Dentz]

From: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>

This adds --track-origin=yes option:

  "Controls whether Memcheck tracks the origin of uninitialised values.
   By default, it does not, which means that although it can tell you
   that an uninitialised value is being used in a dangerous way, it
   cannot tell you where the uninitialised value came from. This often
   makes it difficult to track down the root problem."
---
 android/system-emulator.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/android/system-emulator.c b/android/system-emulator.c
index dc46814..c1b1b25 100644
--- a/android/system-emulator.c
+++ b/android/system-emulator.c
@@ -50,7 +50,7 @@ static pid_t snoop_pid = -1;
 static void ctl_start(void)
 {
 	char prg_name[PATH_MAX + 1];
-	char *prg_argv[5];
+	char *prg_argv[6];
 	char *prg_envp[3];
 	pid_t pid;
 
@@ -58,9 +58,10 @@ static void ctl_start(void)
 
 	prg_argv[0] = "/usr/bin/valgrind";
 	prg_argv[1] = "--leak-check=full";
-	prg_argv[2] = prg_name;
-	prg_argv[3] = "-d";
-	prg_argv[4] = NULL;
+	prg_argv[2] = "--track-origins=yes";
+	prg_argv[3] = prg_name;
+	prg_argv[4] = "-d";
+	prg_argv[5] = NULL;
 
 	prg_envp[0] = "G_SLICE=always-malloc";
 	prg_envp[1] = "G_DEBUG=gc-friendly";
-- 
1.9.0

[PATCH BlueZ 2/2] android/hal-gatt: Fix sending uninitialised byte

[Luiz Augusto von Dentz]

From: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>

This fixes sending uninitialised byte on init:

Syscall param sendmsg(msg.msg_iov[1]) points to uninitialised byte(s)
   at 0x534ADDD: ??? (in /usr/lib64/libpthread-2.18.so)
   by 0x6B5628C: hal_ipc_cmd (hal-ipc.c:357)
   by 0x6B54F6D: init (hal-gatt.c:1268)
   by 0x4092DB: init_p (if-gatt.c:779)
   by 0x401CCD: main (haltest.c:413)
 Address 0xffefffda1 is on thread 1's stack
 Uninitialised value was created by a stack allocation
   at 0x6B54EC0: init (hal-gatt.c:1252)
---
 android/hal-gatt.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/android/hal-gatt.c b/android/hal-gatt.c
index d8c0d54..6fde143 100644
--- a/android/hal-gatt.c
+++ b/android/hal-gatt.c
@@ -1264,6 +1264,7 @@ static bt_status_t init(const btgatt_callbacks_t *callbacks)
 				sizeof(ev_handlers)/sizeof(ev_handlers[0]));
 
 	cmd.service_id = HAL_SERVICE_ID_GATT;
+	cmd.mode = HAL_MODE_DEFAULT;
 
 	ret = hal_ipc_cmd(HAL_SERVICE_ID_CORE, HAL_OP_REGISTER_MODULE,
 					sizeof(cmd), &cmd, 0, NULL, NULL);
-- 
1.9.0

Re: [PATCH BlueZ 1/2] android/system-emulator: Pass --track-origin=yes to valgrind

[Szymon Janc]

Hi Luiz,

On Thursday 10 of April 2014 15:17:08 Luiz Augusto von Dentz wrote:
> From: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
> 
> This adds --track-origin=yes option:
> 
>   "Controls whether Memcheck tracks the origin of uninitialised values.
>    By default, it does not, which means that although it can tell you
>    that an uninitialised value is being used in a dangerous way, it
>    cannot tell you where the uninitialised value came from. This often
>    makes it difficult to track down the root problem."
> ---
>  android/system-emulator.c | 9 +++++----
>  1 file changed, 5 insertions(+), 4 deletions(-)
> 
> diff --git a/android/system-emulator.c b/android/system-emulator.c
> index dc46814..c1b1b25 100644
> --- a/android/system-emulator.c
> +++ b/android/system-emulator.c
> @@ -50,7 +50,7 @@ static pid_t snoop_pid = -1;
>  static void ctl_start(void)
>  {
>  	char prg_name[PATH_MAX + 1];
> -	char *prg_argv[5];
> +	char *prg_argv[6];
>  	char *prg_envp[3];
>  	pid_t pid;
>  
> @@ -58,9 +58,10 @@ static void ctl_start(void)
>  
>  	prg_argv[0] = "/usr/bin/valgrind";
>  	prg_argv[1] = "--leak-check=full";
> -	prg_argv[2] = prg_name;
> -	prg_argv[3] = "-d";
> -	prg_argv[4] = NULL;
> +	prg_argv[2] = "--track-origins=yes";
> +	prg_argv[3] = prg_name;
> +	prg_argv[4] = "-d";
> +	prg_argv[5] = NULL;
>  
>  	prg_envp[0] = "G_SLICE=always-malloc";
>  	prg_envp[1] = "G_DEBUG=gc-friendly";
> 

Both patches applied, thanks.

-- 
Best regards, 
Szymon Janc