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

[Daniel P. Berrange]

On Fri, Oct 07, 2016 at 03:46:27PM +0200, Lluís Vilanova wrote:
> 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)

No, this isn't what we want todo. The real problem is that we should
have been registering the trace events in the linux-user/bsd-user
programs. I sent a fix for that.

Regards,
Daniel
-- 
|: http://berrange.com      -o-    http://www.flickr.com/photos/dberrange/ :|
|: http://libvirt.org              -o-             http://virt-manager.org :|
|: http://entangle-photo.org       -o-    http://search.cpan.org/~danberr/ :|