| Summary: | [Beaglebone] Xserver segmentation fault when reboot the system | ||||||
|---|---|---|---|---|---|---|---|
| Product: | [Build System, Metadata & Runtime] OE-Core | Reporter: | Yi Zhao <yi.zhao> | ||||
| Component: | graphics | Assignee: | Ross Burton <ross.burton> | ||||
| Status: | VERIFIED FIXED | QA Contact: | Yi Zhao <yi.zhao> | ||||
| Severity: | normal | ||||||
| Priority: | Medium | CC: | alexandru.c.georgescu, ke.zou, markus.boos, meta.mr.watcher, meta.watcher, richard.purdie, yi.zhao | ||||
| Version: | 1.7 | ||||||
| Target Milestone: | 1.8 M3 | ||||||
| Hardware: | BeagleBone | ||||||
| OS: | arm | ||||||
| Whiteboard: | |||||||
| OS type for building Yocto: | --- | Type of Regression: | Regression (Found by QA test) | ||||
| Verified: | Documentation change: | No (bug/feature does not impact docs) | |||||
| Attachments: |
|
||||||
|
Description
Yi Zhao
2014-08-28 07:46:44 UTC
This is not a BSP issue. The should be an issue in xserver-xorg. It seems that we try to close the connections repetitively. Something like the following would address it. But I am not sure this is the right fix.
diff --git a/os/connection.c b/os/connection.c
index 162e1d93e483..6dff9ad7e2ea 100644
--- a/os/connection.c
+++ b/os/connection.c
@@ -510,9 +510,14 @@ void
CloseWellKnownConnections(void)
{
int i;
+ XtransConnInfo *p = ListenTransConns;
- for (i = 0; i < ListenTransCount; i++)
- _XSERVTransClose(ListenTransConns[i]);
+ for (i = 0; i < ListenTransCount; i++, p++) {
+ if (*p) {
+ _XSERVTransClose(ListenTransConns[i]);
+ *p = NULL;
+ }
+ }
}
static void
Kevin - can you file a bug/send a mail to the Xorg developers for their input on this? Silence on this since September - is this still reproducible? (In reply to comment #3) > Silence on this since September - is this still reproducible? Yes. It still exists in latest build (git rev: master/8657c2fd57600e0b0eaf2a14692de7453c78e1bb) This issue still exists in 1.8 M2 RC2 build git rev: master/22c38e2d8a058619e3d05534b5e830b750aa4241 Thanks to AlanC of xorg fame, there's a patch on the list for this. |