[PATCH] Handle NW DETACH and ME DETACH

[PATCH] Handle NW DETACH and ME DETACH

[Marko Sulejic]

[-- Attachment #1: Type: text/plain, Size: 1503 bytes --]

From: August Mayer <august.mayer@hale.at>

 * According to the Telit AT reference, NW DETACH and ME DETACH
   imply network deactivation.
 * Handle in drivers/atmodem/gprs-context.c, function cgev_notify.
---
 drivers/atmodem/gprs-context.c | 22 +++++++++++++---------
 1 file changed, 13 insertions(+), 9 deletions(-)

diff --git a/drivers/atmodem/gprs-context.c b/drivers/atmodem/gprs-context.c
index 7e9a43c..db32b17 100644
--- a/drivers/atmodem/gprs-context.c
+++ b/drivers/atmodem/gprs-context.c
@@ -360,20 +360,24 @@ static void cgev_notify(GAtResult *result, gpointer user_data)
 	if (!g_at_result_iter_next_unquoted_string(&iter, &event))
 		return;
 
-	if (g_str_has_prefix(event, "NW DEACT") == FALSE &&
+	if (g_str_has_prefix(event, "NW DETACH") == FALSE &&
+			g_str_has_prefix(event, "ME DETACH") == FALSE) {
+
+		if (g_str_has_prefix(event, "NW DEACT") == FALSE &&
 				g_str_has_prefix(event, "ME DEACT") == FALSE)
-		return;
+			return;
 
-	if (!g_at_result_iter_skip_next(&iter))
-		return;
+		if (!g_at_result_iter_skip_next(&iter))
+			return;
 
-	if (!g_at_result_iter_next_number(&iter, &cid))
-		return;
+		if (!g_at_result_iter_next_number(&iter, &cid))
+			return;
 
-	DBG("cid %d", cid);
+		DBG("cid %d", cid);
 
-	if ((unsigned int) cid != gcd->active_context)
-		return;
+		if ((unsigned int) cid != gcd->active_context)
+			return;
+	}
 
 	if (gcd->state != STATE_IDLE && gcd->ppp)
 		g_at_ppp_shutdown(gcd->ppp);
-- 
2.1.0

Re: [PATCH] Handle NW DETACH and ME DETACH

[Denis Kenzior]

[-- Attachment #1: Type: text/plain, Size: 638 bytes --]

Hi Marko,

On 07/03/2015 03:43 AM, Marko Sulejic wrote:
> From: August Mayer <august.mayer@hale.at>
>
>   * According to the Telit AT reference, NW DETACH and ME DETACH
>     imply network deactivation.

Can you please include some description of why this patch is needed? 
E.g. what exceptional conditions is this supposed to be taking care of. 
  If it is Telit specific, then it should be behind logic testing for 
the TELIT vendor quirk.

Also, do realize that PPP session will be terminated by via LCP or by 
using the NO CARRIER final result code.  This logic is (in theory at 
least) redundant.

Regards,
-Denis