[webkit-changes] [118038] trunk/Source/JavaScriptCore

[commit-queue]

Title: [118038] trunk/Source/_javascript_Core

Revision

118038

Author

commit-qu...@webkit.org

Date

2012-05-22 13:23:44 -0700 (Tue, 22 May 2012)

Log Message

[BlackBerry] getPlatformThreadRegisters() should fetch target thread's registers
https://bugs.webkit.org/show_bug.cgi?id=87148

Patch by Yong Li <y...@rim.com> on 2012-05-22
Reviewed by George Staikos.

Our previous implementation of getPlatformThreadRegisters() read registers in current
thread's context but it is supposed to read the target thread's registers.

* heap/MachineStackMarker.cpp:
(JSC::getPlatformThreadRegisters):

Modified Paths

• trunk/Source/_javascript_Core/ChangeLog
• trunk/Source/_javascript_Core/heap/MachineStackMarker.cpp

Diff

Modified: trunk/Source/_javascript_Core/ChangeLog (118037 => 118038)

--- trunk/Source/_javascript_Core/ChangeLog	2012-05-22 20:21:58 UTC (rev 118037)
+++ trunk/Source/_javascript_Core/ChangeLog	2012-05-22 20:23:44 UTC (rev 118038)
@@ -1,3 +1,16 @@
+2012-05-22  Yong Li  <y...@rim.com>
+
+        [BlackBerry] getPlatformThreadRegisters() should fetch target thread's registers
+        https://bugs.webkit.org/show_bug.cgi?id=87148
+
+        Reviewed by George Staikos.
+
+        Our previous implementation of getPlatformThreadRegisters() read registers in current
+        thread's context but it is supposed to read the target thread's registers.
+
+        * heap/MachineStackMarker.cpp:
+        (JSC::getPlatformThreadRegisters):
+
 2012-05-05  Filip Pizlo  <fpi...@apple.com>
 
         DFG should support reflective arguments access

Modified: trunk/Source/_javascript_Core/heap/MachineStackMarker.cpp (118037 => 118038)

--- trunk/Source/_javascript_Core/heap/MachineStackMarker.cpp	2012-05-22 20:21:58 UTC (rev 118037)
+++ trunk/Source/_javascript_Core/heap/MachineStackMarker.cpp	2012-05-22 20:23:44 UTC (rev 118038)
@@ -356,14 +356,20 @@
     return sizeof(CONTEXT);
 #elif OS(QNX)
     memset(&regs, 0, sizeof(regs));
-    regs.tid = pthread_self();
-    int fd = open("/proc/self", O_RDONLY);
+    regs.tid = platformThread;
+    // FIXME: If we find this hurts performance, we can consider caching the fd and keeping it open.
+    int fd = open("/proc/self/as", O_RDONLY);
     if (fd == -1) {
-        LOG_ERROR("Unable to open /proc/self (errno: %d)", errno);
+        LOG_ERROR("Unable to open /proc/self/as (errno: %d)", errno);
         CRASH();
     }
-    devctl(fd, DCMD_PROC_TIDSTATUS, &regs, sizeof(regs), 0);
+    int rc = devctl(fd, DCMD_PROC_TIDSTATUS, &regs, sizeof(regs), 0);
+    if (rc != EOK) {
+        LOG_ERROR("devctl(DCMD_PROC_TIDSTATUS) failed (error: %d)", rc);
+        CRASH();
+    }
     close(fd);
+    return sizeof(struct _debug_thread_info);
 #elif USE(PTHREADS)
     pthread_attr_init(&regs);
 #if HAVE(PTHREAD_NP_H) || OS(NETBSD)

_______________________________________________
webkit-changes mailing list
webkit-changes@lists.webkit.org
http://lists.webkit.org/mailman/listinfo.cgi/webkit-changes