On Wed, Jan 19, 2011 at 9:57 PM, Paul Pluzhnikov wrote: > Would recording the current jit_event breakpoint in inferior (via > register_inferior_data_with_cleanup ()) be a good solution? That appears to work reasonably well. It fixes both the gdbserver failure in jit.exp, and also as yet undiscovered problem: the old static/global jit_descriptor_addr couldn't possibly work for more than one inferior. Thanks, -- Paul Pluzhnikov 2011-01-20 Paul Pluzhnikov * jit.c (jit_descriptor_addr): Delete. (registering_code): Delete. (clear_int): Delete. (jit_inferior_data): New variable. (struct jit_inferior_data): New type. (get_jit_inferior_data): New function. (jit_inferior_data_cleanup): New function. (jit_read_descriptor): Adjust. (jit_register_code): Adjust. (jit_breakpoint_re_set_internal): New function; move code here ... (jit_inferior_init): ... from here. (jit_breakpoint_re_set): Adjust. (jit_inferior_exit_hook): Adjust. (jit_event_handler): Adjust. (_initialize_jit): Adjust.