Re: [Qemu-devel] [PULL 00/20] Tracing patches

[Lluís Vilanova]

Peter Maydell writes:

> On 7 October 2016 at 11:09, Stefan Hajnoczi <address@hidden> wrote:
>> The following changes since commit e902754e3d0890945ddcc1b33748ed73762ddb8d:
>> 
>> Merge remote-tracking branch 'remotes/dgibson/tags/ppc-for-2.8-20161006' 
>> into staging (2016-10-06 13:34:00 +0100)
>> 
>> are available in the git repository at:
>> 
>> git://github.com/stefanha/qemu.git tags/tracing-pull-request
>> 
>> for you to fetch changes up to bcdc1363dcb9adb5f7ea016a4d0b41f0357c8182:
>> 
>> trace: Add missing execution mode of guest events (2016-10-07 09:17:49 +0100)
>> 
>> ----------------------------------------------------------------
>> 
>> ----------------------------------------------------------------

> This makes all the linux-user binaries dump core on startup:

> $ gdb --args 
> /home/petmay01/linaro/qemu-for-merges/build/all-linux-static/x86_64-linux-user/qemu-x86_64
> /bin/ls
> (gdb) r
> Starting program:
> /home/petmay01/linaro/qemu-for-merges/build/all-linux-static/x86_64-linux-user/qemu-x86_64
> /bin/ls
> [Thread debugging using libthread_db enabled]
> Using host libthread_db library "/lib/x86_64-linux-gnu/libthread_db.so.1".
> [New Thread 0x7ffff7ffa700 (LWP 14965)]

> Thread 1 "qemu-x86_64" received signal SIGABRT, Aborted.
> 0x000000006019fbe8 in raise ()
> (gdb) bt
> #0  0x000000006019fbe8 in raise ()
> #1  0x00000000601a030a in abort ()
> #2  0x00000000600bdd02 in bitmap_new (nbits=0)
>     at /home/petmay01/linaro/qemu-for-merges/include/qemu/bitmap.h:99
> #3  0x00000000600be99f in cpu_common_initfn (obj=0x625f6340)
>     at /home/petmay01/linaro/qemu-for-merges/qom/cpu.c:360
> #4  0x00000000600bfe97 in object_init_with_type (obj=0x625f6340, 
> ti=0x625eb9a0)
>     at /home/petmay01/linaro/qemu-for-merges/qom/object.c:339
> #5  0x00000000600bfe79 in object_init_with_type (obj=0x625f6340, 
> ti=0x625e2d80)
>     at /home/petmay01/linaro/qemu-for-merges/qom/object.c:335
> #6  0x00000000600bfe79 in object_init_with_type (obj=0x625f6340, 
> ti=0x625e78d0)
>     at /home/petmay01/linaro/qemu-for-merges/qom/object.c:335
> #7  0x00000000600c00aa in object_initialize_with_type
> (data=0x625f6340, size=38416, type=0x625e78d0)
>     at /home/petmay01/linaro/qemu-for-merges/qom/object.c:370
> #8  0x00000000600c0508 in object_new_with_type (type=0x625e78d0)
>     at /home/petmay01/linaro/qemu-for-merges/qom/object.c:478
> #9  0x00000000600c0542 in object_new (typename=0x625e2fe0 "qemu64-x86_64-cpu")
>     at /home/petmay01/linaro/qemu-for-merges/qom/object.c:488
> #10 0x00000000600be03f in cpu_generic_init (typename=0x6025ad54
> "x86_64-cpu", cpu_model=0x60247ca5 "qemu64")
>     at /home/petmay01/linaro/qemu-for-merges/qom/cpu.c:76
> #11 0x0000000060089438 in cpu_x86_init (cpu_model=0x60247ca5 "qemu64")
>     at /home/petmay01/linaro/qemu-for-merges/target-i386/cpu.c:2266
> #12 0x0000000060031031 in main (argc=2, argv=0x7fffffffe4b8,
> envp=0x7fffffffe4d0)
>     at /home/petmay01/linaro/qemu-for-merges/linux-user/main.c:4253

Aha! I didn't take that abort into consideration during the review. The fix
should be on patch 13 of the series (trace: dynamically allocate trace_dstate in
CPUState):

diff --git a/qom/cpu.c b/qom/cpu.c
index 484c493..40f2eb1 100644
--- a/qom/cpu.c
+++ b/qom/cpu.c
@@ -356,12 +356,15 @@ static void cpu_common_initfn(Object *obj)
     qemu_mutex_init(&cpu->work_mutex);
     QTAILQ_INIT(&cpu->breakpoints);
     QTAILQ_INIT(&cpu->watchpoints);
-    bitmap_zero(cpu->trace_dstate, TRACE_VCPU_EVENT_COUNT);
+
+    cpu->trace_dstate = bitmap_new(min(trace_get_vcpu_event_count(), 1);
 }
 
 static void cpu_common_finalize(Object *obj)
 {
-    cpu_exec_exit(CPU(obj));
+    CPUState *cpu = CPU(obj);
+    cpu_exec_exit(cpu);
+    g_free(cpu->trace_dstate);
 }
 
 static int64_t cpu_common_get_arch_id(CPUState *cpu)

Beware that min() needs to be defined somewhere on that file first (should it be
available on some generic header?).

Cheers,
  Lluis