Re: [PATCH 4/6] Share parts of gdb/gdbthread.h with gdbserver

[Sergio Durigan Junior <sergiodj at redhat dot com>]

Thanks for the review.

Comments below.

On Monday, December 26 2016, Luis Machado wrote:

> On 12/22/2016 09:39 PM, Sergio Durigan Junior wrote:
>> Again, it was necessary to share a few functions declared on
>> gdb/gdbthread.h with gdbserver, because they are needed by
>> fork_inferior.  I decided to implement them on
>> gdb/gdbserver/inferiors.c because that's where the thread functions
>> are also implemented on gdbserver.  As a side note, due to the way
>> gdbserver marks a thread as executing, a new argument needed to be
>> added on set_executing.  This argument is a 'struct target_waitstatus
>> *', and is not necessary on GDB, so it was marked with
>> ATTRIBUTE_UNUSED accordingly.
>>
>> gdb/ChangeLog:
>> 2016-12-22  Sergio Durigan Junior  <sergiodj@redhat.com>
>>
>> 	* Makefile.in (HFILES_NO_SRCDIR): Add "common/common-gdbthread.h".
>> 	* common/common-gdbthread.h: New file, with parts from
>> 	"gdb/gdbthread.h".
>> 	* fork-child.c (fork_inferior): Update call of "set_executing".
>> 	* gdbthread.h: Include "common-gdbthread.h".
>> 	(init_thread_list): Moved to "common/common-gdbthread.h".
>> 	(add_thread_silent): Likewise.
>> 	(switch_to_thread): Likewise.
>> 	(set_executing): Likewise.  Added an extra argument, "struct
>> 	target_waitstatus *".
>> 	* infrun.c (handle_inferior_event_1): Update call of
>> 	"set_executing".
>> 	* linux-nat.c (attach_proc_task_lwp_callback): Likewise.
>> 	(linux_handle_extended_wait): Likewise.
>> 	* record-btrace.c (get_thread_current_frame): Likewise.
>> 	* record-full.c (record_full_wait_1): Likewise.
>> 	* remote.c (remote_add_thread): Likewise.
>> 	(process_initial_stop_replies): Likewise.
>> 	* solib-spu.c (spu_skip_standalone_loader): Likewise.
>> 	* target.c (target_resume): Likewise.
>> 	* thread.c (set_executing): Update function declaration to add
>> 	third argument.  Mark it as unused.
>>
>> gdb/gdbserver/ChangeLog:
>> 2016-12-22  Sergio Durigan Junior  <sergiodj@redhat.com>
>>
>> 	* inferiors.c (init_thread_list): New function.
>> 	(switch_to_thread): Likewise.
>> 	(set_executing): Likewise.
>> 	(add_thread_silent): Likewise.
>> ---
>>  gdb/Makefile.in               |  1 +
>>  gdb/common/common-gdbthread.h | 50 +++++++++++++++++++++++++++++++++++++++++++
>>  gdb/fork-child.c              |  2 +-
>>  gdb/gdbserver/inferiors.c     | 42 ++++++++++++++++++++++++++++++++++++
>>  gdb/gdbthread.h               | 20 +----------------
>>  gdb/infrun.c                  |  2 +-
>>  gdb/linux-nat.c               |  4 ++--
>>  gdb/record-btrace.c           |  6 +++---
>>  gdb/record-full.c             |  4 ++--
>>  gdb/remote.c                  |  4 ++--
>>  gdb/solib-spu.c               |  2 +-
>>  gdb/target.c                  |  2 +-
>>  gdb/thread.c                  |  5 ++++-
>>  13 files changed, 111 insertions(+), 33 deletions(-)
>>  create mode 100644 gdb/common/common-gdbthread.h
>>
>> diff --git a/gdb/Makefile.in b/gdb/Makefile.in
>> index 8372b4a..ca13a80 100644
>> --- a/gdb/Makefile.in
>> +++ b/gdb/Makefile.in
>> @@ -1467,6 +1467,7 @@ HFILES_NO_SRCDIR = \
>>  	common/common-debug.h \
>>  	common/common-defs.h \
>>  	common/common-exceptions.h \
>> +	common/common-gdbthread.h \
>>  	common/common-regcache.h \
>>  	common/common-types.h \
>>  	common/common-utils.h \
>> diff --git a/gdb/common/common-gdbthread.h b/gdb/common/common-gdbthread.h
>> new file mode 100644
>> index 0000000..3560892
>> --- /dev/null
>> +++ b/gdb/common/common-gdbthread.h
>> @@ -0,0 +1,50 @@
>> +/* Multi-process/thread control defs for GDB, the GNU debugger.
>> +   Copyright (C) 1987-2016 Free Software Foundation, Inc.
>> +   Contributed by Lynx Real-Time Systems, Inc.  Los Gatos, CA.
>> +
>> +
>> +   This file is part of GDB.
>> +
>> +   This program is free software; you can redistribute it and/or modify
>> +   it under the terms of the GNU General Public License as published by
>> +   the Free Software Foundation; either version 3 of the License, or
>> +   (at your option) any later version.
>> +
>> +   This program is distributed in the hope that it will be useful,
>> +   but WITHOUT ANY WARRANTY; without even the implied warranty of
>> +   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
>> +   GNU General Public License for more details.
>> +
>> +   You should have received a copy of the GNU General Public License
>> +   along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
>> +
>> +#ifndef COMMON_THREAD_H
>> +#define COMMON_THREAD_H
>> +
>> +struct target_waitstatus;
>> +
>> +/* Create an empty thread list, or empty the existing one.  */
>> +
>
> Spurious newline, multiple cases of this throughout.

Yeah, now you're right:

  https://sourceware.org/gdb/wiki/Internals%20GDB-C-Coding-Standards#Empty_line_between_subprogram_description_and_the_subprogram_implementation

  "Note that this only applies to the case where the comment is placed
  besides the subprogram implementation (typically in a .c file). In the
  case of the documentation being placed next to the subprogram
  declaration, then the comment should be placed immediately before the
  declaration."

I'll fix these.  Thanks.

>> +extern void init_thread_list (void);
>> +
>> +/* Switch from one thread to another.  */
>> +
>> +extern void switch_to_thread (ptid_t ptid);
>> +
>> +/* Marks thread PTID as executing, or not.  If PTID is minus_one_ptid,
>> +   marks all threads.
>> +
>> +   Note that this is different from the running state.  See the
>> +   description of state and executing fields of struct
>> +   thread_info.  */
>> +
>> +extern void set_executing (ptid_t ptid, int executing,
>> +			   struct target_waitstatus *ws);
>> +
>> +/* Add a thread to the thread list and return the pointer to the new
>> +   thread.  Caller may use this pointer to initialize the private
>> +   thread data.  */
>> +
>> +extern struct thread_info *add_thread_silent (ptid_t ptid);
>> +
>> +#endif /* ! COMMON_THREAD_H */
>> diff --git a/gdb/fork-child.c b/gdb/fork-child.c
>> index 15f8249..38fca60 100644
>> --- a/gdb/fork-child.c
>> +++ b/gdb/fork-child.c
>> @@ -545,7 +545,7 @@ startup_inferior (int ntraps)
>>      }
>>
>>    /* Mark all threads non-executing.  */
>> -  set_executing (resume_ptid, 0);
>> +  set_executing (resume_ptid, 0, NULL);
>>  }
>>
>>  /* Implement the "unset exec-wrapper" command.  */
>> diff --git a/gdb/gdbserver/inferiors.c b/gdb/gdbserver/inferiors.c
>> index 574a7ba..6b981d0 100644
>> --- a/gdb/gdbserver/inferiors.c
>> +++ b/gdb/gdbserver/inferiors.c
>> @@ -468,3 +468,45 @@ make_cleanup_restore_current_thread (void)
>>  {
>>    return make_cleanup (do_restore_current_thread_cleanup, current_thread);
>>  }
>> +
>> +/* See common/common-gdbthread.h.  */
>> +
>> +void
>> +init_thread_list (void)
>> +{
>> +  /* To be implemented.  */
>> +}
>> +
>> +/* See common/common-gdbthread.h.  */
>> +
>> +void
>> +switch_to_thread (ptid_t ptid)
>> +{
>> +  if (!ptid_equal (ptid, minus_one_ptid))
>> +    current_thread = find_thread_ptid (ptid);
>> +}
>> +
>> +/* See common/common-gdbthread.h.  */
>> +
>> +void
>> +set_executing (ptid_t ptid ATTRIBUTE_UNUSED, int executing ATTRIBUTE_UNUSED,
>> +	       struct target_waitstatus *ws)
>> +{
>> +  gdb_assert (current_thread != NULL);
>> +  current_thread->last_resume_kind = resume_stop;
>> +  current_thread->last_status = *ws;
>
> Is there a chance ws will be NULL and cause a crash here? Should we assert?

Yeah, I was also thinking about that, but currently there is no chance
of it being NULL.  The only place that calls this function for gdbserver
is fork_inferior, and if the function is called without a current_thread
then something really wrong happened.

>> +}
>> +
>> +/* See common/common-gdbthread.h.  */
>> +
>> +struct thread_info *
>> +add_thread_silent (ptid_t ptid)
>> +{
>> +  pid_t pid = ptid_get_pid (ptid);
>> +
>> +  /* Check if there is a process already.  */
>> +  if (find_process_pid (pid) == NULL)
>> +    add_process (pid, 0);
>> +
>> +  return add_thread (ptid_build (pid, pid, 0), NULL);
>> +}
>> diff --git a/gdb/gdbthread.h b/gdb/gdbthread.h
>> index 8f37fbb..d6fdc42 100644
>> --- a/gdb/gdbthread.h
>> +++ b/gdb/gdbthread.h
>> @@ -31,6 +31,7 @@ struct symtab;
>>  #include "common/vec.h"
>>  #include "target/waitstatus.h"
>>  #include "cli/cli-utils.h"
>> +#include "common-gdbthread.h"
>>
>>  /* Frontend view of the thread state.  Possible extensions: stepping,
>>     finishing, until(ling),...  */
>> @@ -344,19 +345,12 @@ struct thread_info
>>    struct thread_info *step_over_next;
>>  };
>>
>> -/* Create an empty thread list, or empty the existing one.  */
>> -extern void init_thread_list (void);
>> -
>>  /* Add a thread to the thread list, print a message
>>     that a new thread is found, and return the pointer to
>>     the new thread.  Caller my use this pointer to
>>     initialize the private thread data.  */
>>  extern struct thread_info *add_thread (ptid_t ptid);
>>
>> -/* Same as add_thread, but does not print a message
>> -   about new thread.  */
>> -extern struct thread_info *add_thread_silent (ptid_t ptid);
>> -
>>  /* Same as add_thread, and sets the private info.  */
>>  extern struct thread_info *add_thread_with_info (ptid_t ptid,
>>  						 struct private_thread_info *);
>> @@ -469,10 +463,6 @@ extern struct thread_info *iterate_over_threads (thread_callback_func, void *);
>>
>>  extern int thread_count (void);
>>
>> -/* Switch from one thread to another.  Also sets the STOP_PC
>> -   global.  */
>> -extern void switch_to_thread (ptid_t ptid);
>> -
>>  /* Switch from one thread to another.  Does not read registers and
>>     sets STOP_PC to -1.  */
>>  extern void switch_to_thread_no_regs (struct thread_info *thread);
>> @@ -518,14 +508,6 @@ extern int is_exited (ptid_t ptid);
>>  /* In the frontend's perpective, is this thread stopped?  */
>>  extern int is_stopped (ptid_t ptid);
>>
>> -/* Marks thread PTID as executing, or not.  If PTID is minus_one_ptid,
>> -   marks all threads.
>> -
>> -   Note that this is different from the running state.  See the
>> -   description of state and executing fields of struct
>> -   thread_info.  */
>> -extern void set_executing (ptid_t ptid, int executing);
>> -
>>  /* Reports if thread PTID is executing.  */
>>  extern int is_executing (ptid_t ptid);
>>
>> diff --git a/gdb/infrun.c b/gdb/infrun.c
>> index bf0632e..f5cc965 100644
>> --- a/gdb/infrun.c
>> +++ b/gdb/infrun.c
>> @@ -4947,7 +4947,7 @@ handle_inferior_event_1 (struct execution_control_state *ecs)
>>      else
>>        mark_ptid = ecs->ptid;
>>
>> -    set_executing (mark_ptid, 0);
>> +    set_executing (mark_ptid, 0, NULL);
>>
>>      /* Likewise the resumed flag.  */
>>      set_resumed (mark_ptid, 0);
>> diff --git a/gdb/linux-nat.c b/gdb/linux-nat.c
>> index cbf94ed..9db620c 100644
>> --- a/gdb/linux-nat.c
>> +++ b/gdb/linux-nat.c
>> @@ -1190,7 +1190,7 @@ attach_proc_task_lwp_callback (ptid_t ptid)
>>  	     raw clone).  */
>>  	  add_thread (lp->ptid);
>>  	  set_running (lp->ptid, 1);
>> -	  set_executing (lp->ptid, 1);
>> +	  set_executing (lp->ptid, 1, NULL);
>>  	}
>>
>>        return 1;
>> @@ -2083,7 +2083,7 @@ linux_handle_extended_wait (struct lwp_info *lp, int status)
>>  	     and the user/frontend, this new thread is running until
>>  	     it next reports a stop.  */
>>  	  set_running (new_lp->ptid, 1);
>> -	  set_executing (new_lp->ptid, 1);
>> +	  set_executing (new_lp->ptid, 1, NULL);
>>
>>  	  if (WSTOPSIG (status) != SIGSTOP)
>>  	    {
>> diff --git a/gdb/record-btrace.c b/gdb/record-btrace.c
>> index 7c0e39f..80cf274 100644
>> --- a/gdb/record-btrace.c
>> +++ b/gdb/record-btrace.c
>> @@ -1928,7 +1928,7 @@ get_thread_current_frame (struct thread_info *tp)
>>       move the thread.  Since we need to recompute the stack, we temporarily
>>       set EXECUTING to flase.  */
>>    executing = is_executing (inferior_ptid);
>> -  set_executing (inferior_ptid, 0);
>> +  set_executing (inferior_ptid, 0, NULL);
>>
>>    frame = NULL;
>>    TRY
>> @@ -1938,7 +1938,7 @@ get_thread_current_frame (struct thread_info *tp)
>>    CATCH (except, RETURN_MASK_ALL)
>>      {
>>        /* Restore the previous execution state.  */
>> -      set_executing (inferior_ptid, executing);
>> +      set_executing (inferior_ptid, executing, NULL);
>>
>>        /* Restore the previous inferior_ptid.  */
>>        inferior_ptid = old_inferior_ptid;
>> @@ -1948,7 +1948,7 @@ get_thread_current_frame (struct thread_info *tp)
>>    END_CATCH
>>
>>    /* Restore the previous execution state.  */
>> -  set_executing (inferior_ptid, executing);
>> +  set_executing (inferior_ptid, executing, NULL);
>>
>>    /* Restore the previous inferior_ptid.  */
>>    inferior_ptid = old_inferior_ptid;
>> diff --git a/gdb/record-full.c b/gdb/record-full.c
>> index 5608e70..3ad9a08 100644
>> --- a/gdb/record-full.c
>> +++ b/gdb/record-full.c
>> @@ -1156,12 +1156,12 @@ record_full_wait_1 (struct target_ops *ops,
>>  			{
>>  			  /* Try to insert the software single step breakpoint.
>>  			     If insert success, set step to 0.  */
>> -			  set_executing (inferior_ptid, 0);
>> +			  set_executing (inferior_ptid, 0, NULL);
>>  			  reinit_frame_cache ();
>>
>>  			  step = !insert_single_step_breakpoints (gdbarch);
>>
>> -			  set_executing (inferior_ptid, 1);
>> +			  set_executing (inferior_ptid, 1, NULL);
>>  			}
>>
>>  		      if (record_debug)
>> diff --git a/gdb/remote.c b/gdb/remote.c
>> index ef6c54e..0b5f837 100644
>> --- a/gdb/remote.c
>> +++ b/gdb/remote.c
>> @@ -1847,7 +1847,7 @@ remote_add_thread (ptid_t ptid, int running, int executing)
>>      thread = add_thread (ptid);
>>
>>    get_private_info_thread (thread)->vcont_resumed = executing;
>> -  set_executing (ptid, executing);
>> +  set_executing (ptid, executing, NULL);
>>    set_running (ptid, running);
>>  }
>>
>> @@ -3949,7 +3949,7 @@ process_initial_stop_replies (int from_tty)
>>  	  || ws.value.sig != GDB_SIGNAL_0)
>>  	thread->suspend.waitstatus_pending_p = 1;
>>
>> -      set_executing (event_ptid, 0);
>> +      set_executing (event_ptid, 0, NULL);
>>        set_running (event_ptid, 0);
>>        thread->priv->vcont_resumed = 0;
>>      }
>> diff --git a/gdb/solib-spu.c b/gdb/solib-spu.c
>> index fa2977e..f55e0a0 100644
>> --- a/gdb/solib-spu.c
>> +++ b/gdb/solib-spu.c
>> @@ -88,7 +88,7 @@ spu_skip_standalone_loader (void)
>>
>>        target_resume (inferior_ptid, 1, GDB_SIGNAL_0);
>>        target_wait (minus_one_ptid, &ws, 0);
>> -      set_executing (minus_one_ptid, 0);
>> +      set_executing (minus_one_ptid, 0, NULL);
>>
>>        inferior_thread ()->control.in_infcall = 0;
>>      }
>> diff --git a/gdb/target.c b/gdb/target.c
>> index 246d292..e7c68bf 100644
>> --- a/gdb/target.c
>> +++ b/gdb/target.c
>> @@ -2325,7 +2325,7 @@ target_resume (ptid_t ptid, int step, enum gdb_signal signal)
>>    registers_changed_ptid (ptid);
>>    /* We only set the internal executing state here.  The user/frontend
>>       running state is set at a higher level.  */
>> -  set_executing (ptid, 1);
>> +  set_executing (ptid, 1, NULL);
>>    clear_inline_frame_state (ptid);
>>  }
>>
>> diff --git a/gdb/thread.c b/gdb/thread.c
>> index e5d6c5f..1114aac 100644
>> --- a/gdb/thread.c
>> +++ b/gdb/thread.c
>> @@ -1011,8 +1011,11 @@ is_executing (ptid_t ptid)
>>    return tp->executing;
>>  }
>>
>> +/* See common/common-gdbthread.h.  */
>> +
>>  void
>> -set_executing (ptid_t ptid, int executing)
>> +set_executing (ptid_t ptid, int executing,
>> +	       struct target_waitstatus *ws ATTRIBUTE_UNUSED)
>
> I'm thinking about this change. Isn't there a better way to do this
> without adding gdb/gdbserver-specific parameters and then passing
> dummy values for the unused parameters depending on who's calling?
>
> Maybe making gdbserver accept the same parameters as gdb and then do
> some extra processing in gdbserver to get at the waitstatus data if it
> still needs that? Or maybe unifying the function and making gdb and
> gdbserver operate the same way in this regard.

Hm.  Re-reading the code now, I've realized there is a way to make
things work.

startup_inferior already sets the global "last_status" (from
gdbserver/server.c) when it's executing.  Therefore, it is possible to
access this variable (maybe through a getter on server.c), and there is
no need to pass it to set_executing.  Of course, that'd mean relying on
a global value to set current_thread->last_status, but that doesn't
really change what is already being done today.

With this modification, there is no need to extend set_executing and
make it receive a target_waitstatus.  However, for gdbserver, both the
PTID and the EXECUTING arguments will still be useless.

Thanks,

-- 
Sergio
GPG key ID: 237A 54B1 0287 28BF 00EF  31F4 D0EB 7628 65FC 5E36
Please send encrypted e-mail if possible
http://sergiodj.net/