This is the mail archive of the gdb-patches@sourceware.org mailing list for the GDB project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Crash of GDB with gdbserver btrace enabled [Re: [patch v9 00/23] branch tracing support for Atom]


On Thu, 07 Mar 2013 11:31:39 +0100, Metzger, Markus T wrote:
> > OK, I agree target_close seems excessive here, it also does not correspond to
> > the current description of target_close:
> > 	This routine is automatically always called after popping the target
> > 	off the target stack
> > 
> > While it is nice cleanup I find it a separate patch, not required for btrace.
> 
> With this patch, the crash is gone with only minimal changes to btrace.

It is only a coincidence and workaround of it.

The problem here is not the specific crash but the violation of the current
to_close definition of operations, as discussed in the referenced Pedro's
mail.  It may cause different problems later.


> OK to add it as separate patch to the btrace series?

No.  But it would be interesting to post it as unrelated patch, for gdb-7.7+.


Thanks,
Jan


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]