Python扩展GDB自动continue致断点仅触发一次的问题求助

我用Python扩展GDB功能时遇到一个异常情况：当断点被频繁触发时，GDB仅在第一次触发时暂停并执行脚本逻辑，后续断点触发时完全不暂停，脚本也不再响应。但如果不在脚本里用gdb.execute('continue', to_string=True)自动恢复执行，而是手动在GDB命令行输入continue，每次断点触发都会正常暂停，脚本也能正常运行。

我的Python脚本代码 #

def CmdCaptureBreaks(event):
    print("capture a event")
    if not isinstance(event, gdb.BreakpointEvent):
        return
    cur = event.breakpoints[0].location
    if cur is None:
        cur = event.breakpoints[0].expr
    o = gdb.execute('bt', to_string=True)
    print(o)
    BreakDispatch(cur)
    #time.sleep(3)
    gdb.execute('continue', to_string=True)

gdb.events.stop.connect(CmdCaptureBreaks)

自动执行continue时的GDB运行日志 #

(gdb) b rte_flow_query
Breakpoint 1 at 0x56190d3ee190
(gdb) c
Continuing.
[Switching to Thread 0x7fd59effd400 (LWP 30636)]

Thread 17 "revalidator12" hit Breakpoint 1, 0x000056190d3ee190 in rte_flow_query ()  ##### first hit
capture a event
#0  0x000056190d3ee190 in rte_flow_query ()
#1  0x000056190d5ea00a in netdev_dpdk_rte_flow_query_count ()
#2  0x000056190d5ea56a in netdev_offload_dpdk_flow_get ()
#3  0x000056190d4b0e03 in dpif_netdev_get_flow_offload_status ()
#4  0x000056190d4b0f44 in get_dpif_flow_status ()
#5  0x000056190d4b1048 in dp_netdev_flow_to_dpif_flow ()
#6  0x000056190d4b2dd0 in dpif_netdev_flow_dump_next ()
#7  0x000056190d4c8301 in dpif_flow_dump_next ()
#8  0x000056190d46fcac in revalidate.isra ()
#9  0x000056190d471d30 in udpif_revalidator ()
#10 0x000056190d566fa0 in ovsthread_wrapper ()
#11 0x00007fd5d0dd9609 in start_thread (arg=<optimized out>) at pthread_create.c:477
#12 0x00007fd5d0acb293 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95



Thread 17 "revalidator12" hit Breakpoint 1, 0x000056190d3ee190 in rte_flow_query () #####no stop
Thread 17 "revalidator12" hit Breakpoint 1, 0x000056190d3ee190 in rte_flow_query () #####no stop
Thread 17 "revalidator12" hit Breakpoint 1, 0x000056190d3ee190 in rte_flow_query () #####no stop
Thread 17 "revalidator12" hit Breakpoint 1, 0x000056190d3ee190 in rte_flow_query ()  #####no stop
Aborted (core dumped)
When execute control+c may cause crash.

手动执行continue时的GDB运行日志 #

(gdb) b rte_flow_query
Breakpoint 1 at 0x5617df277190
(gdb) c
Continuing.
[Switching to Thread 0x7f1627fff400 (LWP 53739)]

Thread 17 "revalidator12" hit Breakpoint 1, 0x00005617df277190 in rte_flow_query ()
capture a event
#0  0x00005617df277190 in rte_flow_query ()
#1  0x00005617df47300a in netdev_dpdk_rte_flow_query_count ()
#2  0x00005617df47356a in netdev_offload_dpdk_flow_get ()
#3  0x00005617df339e03 in dpif_netdev_get_flow_offload_status ()
#4  0x00005617df339f44 in get_dpif_flow_status ()
#5  0x00005617df33a048 in dp_netdev_flow_to_dpif_flow ()
#6  0x00005617df33bdd0 in dpif_netdev_flow_dump_next ()
#7  0x00005617df351301 in dpif_flow_dump_next ()
#8  0x00005617df2f8cac in revalidate.isra ()
#9  0x00005617df2fad30 in udpif_revalidator ()
#10 0x00005617df3effa0 in ovsthread_wrapper ()
#11 0x00007f1669996609 in start_thread (arg=<optimized out>) at pthread_create.c:477
#12 0x00007f1669688293 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95

(gdb) continue   ####### manually execute


Thread 17 "revalidator12" hit Breakpoint 1, 0x00005617df277190 in rte_flow_query ()
capture a event
#0  0x00005617df277190 in rte_flow_query ()
#1  0x00005617df47300a in netdev_dpdk_rte_flow_query_count ()
#2  0x00005617df47356a in netdev_offload_dpdk_flow_get ()
#3  0x00005617df339e03 in dpif_netdev_get_flow_offload_status ()


#4  0x00005617df339f44 in get_dpif_flow_status ()
#5  0x00005617df346770 in dpif_netdev_operate ()
#6  0x00005617df352e80 in dpif_operate ()
#7  0x00005617df2f46a1 in push_dp_ops ()
#8  0x00005617df2f967d in revalidate.isra ()
#9  0x00005617df2fad30 in udpif_revalidator ()
#10 0x00005617df3effa0 in ovsthread_wrapper ()
#11 0x00007f1669996609 in start_thread (arg=<optimized out>) at pthread_create.c:477
#12 0x00007f1669688293 in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:95

环境信息 #

• GDB版本：GNU gdb (Ubuntu 9.2-0ubuntu1~20.04.1) 9.2

恳请各位提供解决建议，提前感谢！

内容的提问来源于stack exchange，提问作者xiaoxi0329