Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: Paul Gilliam <pgilliam@us.ibm.com>
To: gdb-patches@sourceware.org
Cc: Kevin Buettner <kevinb@redhat.com>,
	        Mark Kettenis <mark.kettenis@xs4all.nl>
Subject: Re: [PATCH] add 'rs6000_in_function_epilogue_p()' (Revised)
Date: Thu, 12 Jan 2006 23:53:00 -0000	[thread overview]
Message-ID: <200601121623.15363.pgilliam@us.ibm.com> (raw)
In-Reply-To: <200601111511.27699.pgilliam@us.ibm.com>

[-- Attachment #1: Type: text/plain, Size: 985 bytes --]

[I am reposting this because it did not show up in the mail-list archive]

On Wednesday 11 January 2006 10:13, Paul Gilliam wrote:
> Kevin and Mark,
> 
> I kind of lost track of this.  Here is a link to the revised patch.  It should address Mark's concerns,
> but I thought, sense it's been a while, I should run it by you again.
> 
> http://sourceware.org/ml/gdb-patches/2005-12/msg00072.html
> 
> OK to commit?
> 
> -=# Paul #=-
> 
> On Friday 02 December 2005 12:32, Kevin Buettner wrote:
> > On Fri, 2 Dec 2005 11:20:22 -0800
> > Paul Gilliam <pgilliam@us.ibm.com> wrote:
> > 
> > >         * rs6000-tdep.c: Add new subroutine, 'rs6000_in_function_epilogue_p()'
> > >         and put it into the architecture vector.
> > 
> > Your patch is okay to commit after you address Mark's concerns.
> > 
> > Kevin
> > 
> > 
> 
> 
> 

I found I had a more up-to-date version of the patch that should address both Mark's concerns and Kevin's

OK to commit?


[-- Attachment #2: watch-locals.patch --]
[-- Type: text/x-diff, Size: 5107 bytes --]

2006-01-11  Paul Gilliam  <pgilliam@us.ibm.com>

	* ppc-tdep.h: Add a define for the hard limit used when scanning an
	epilogue.
        * rs6000-tdep.c: Add new subroutine, 'rs6000_in_function_epilogue_p()'
        and put it into the architecture vector.

Index: ppc-tdep.h
===================================================================
RCS file: /cvs/src/src/gdb/ppc-tdep.h,v
retrieving revision 1.48
diff -a -u -r1.48 ppc-tdep.h
--- ppc-tdep.h	28 Oct 2005 18:23:32 -0000	1.48
+++ ppc-tdep.h	11 Jan 2006 23:03:43 -0000
@@ -384,4 +384,7 @@
 /* Instruction size.  */
 #define PPC_INSN_SIZE 4
 
+/* Estimate for the maximum number of instrctions in a function epilogue.  */
+#define PPC_MAX_EPILOGUE_INSTRUCTIONS  52
+
 #endif /* ppc-tdep.h */
Index: rs6000-tdep.c
===================================================================
RCS file: /cvs/src/src/gdb/rs6000-tdep.c,v
retrieving revision 1.248
diff -a -u -r1.248 rs6000-tdep.c
--- rs6000-tdep.c	1 Nov 2005 19:32:36 -0000	1.248
+++ rs6000-tdep.c	11 Jan 2006 23:03:43 -0000
@@ -502,6 +502,114 @@
   return pc;
 }
 
+static int
+insn_changes_sp_or_jumps (unsigned long insn)
+{
+  int opcode = (insn >> 26) & 0x03f;
+  int sd = (insn >> 21) & 0x01f;
+  int a = (insn >> 16) & 0x01f;
+  /*  b = (insn >> 11) & 0x01f  */
+  int subcode = (insn >> 1) & 0x3ff;
+  /*  rc = insn & 0x001  */
+
+  /* Changes the stack pointer.  */
+
+  /* NOTE: There are many ways to change the value of a given register.
+           The ways below are those used when the register is R1, the SP,
+           in a funtion's epilogue.  */
+
+  if (opcode == 31 && subcode == 444 && a == 1)
+    return 1;  /* mr R1,Rn */
+  if (opcode == 14 && sd == 1)
+    return 1;  /* addi R1,Rn,simm */
+  if (opcode == 58 && sd == 1)
+    return 1;  /* ld R1,ds(Rn) */
+
+  /* Transfers control.  */
+
+  if (opcode == 18)
+    return 1;  /* b */
+  if (opcode == 16)
+    return 1;  /* bc */
+  if (opcode == 19 && subcode == 16)
+    return 1;  /* bclr */
+  if (opcode == 19 && subcode == 528)
+    return 1;  /* bcctr */
+
+  return 0;
+}
+
+/* Return true if we are in the function's epilogue, i.e. after the
+   instruction that destroyed the function's stack frame.
+
+   1) scan forward from the point of execution:
+       a) If you find an instruction that modifies the stack pointer
+          or transfers control (except a return), execution is not in
+          an epilogue, return.
+       b) Stop scanning if you find a return instruction or reach the
+          end of the function or reach the hard limit for the size of
+          an epilogue.
+   2) scan backward from the point of execution:
+        a) If you find an instruction that modifies the stack pointer,
+            execution *is* in an epilogue, return.
+        b) Stop scanning if you reach an instruction that transfers
+           control or the beginning of the function or reach the hard
+           limit for the size of an epilogue.  */
+
+static int
+rs6000_in_function_epilogue_p (struct gdbarch *gdbarch, CORE_ADDR pc)
+{
+  bfd_byte insn_buf[PPC_INSN_SIZE];
+  CORE_ADDR scan_pc, func_start, func_end, epilogue_start, epilogue_end;
+  unsigned long insn;
+  struct frame_info *fr;
+
+  /* Find the search limits based on function boundaries and hard limit.  */
+
+  if (!find_pc_partial_function (pc, NULL, &func_start, &func_end))
+    return 0;
+
+  epilogue_start = pc - PPC_MAX_EPILOGUE_INSTRUCTIONS * PPC_INSN_SIZE;
+  if (epilogue_start < func_start) epilogue_start = func_start;
+
+  epilogue_end = pc + PPC_MAX_EPILOGUE_INSTRUCTIONS * PPC_INSN_SIZE;
+  if (epilogue_end > func_end) epilogue_end = func_end;
+
+  /* Get the current frame.  This may be cheap, since we might have
+     just called it in watchpoint_check, before calling
+     gdbarch_in_function_epilogue_p.  */
+
+  fr = get_current_frame ();
+
+  /* Scan forward until next 'blr'.  */
+
+  for (scan_pc = pc; scan_pc < epilogue_end; scan_pc += PPC_INSN_SIZE)
+    {
+      if (!safe_frame_unwind_memory (fr, scan_pc, insn_buf, PPC_INSN_SIZE))
+        return 0;
+      insn = extract_signed_integer (insn_buf, PPC_INSN_SIZE);
+      if (insn == 0x4e800020)
+        break;
+      if (insn_changes_sp_or_jumps (insn))
+        return 0;
+    }
+
+  /* Scan backward until adjustment to stack pointer (R1).  */
+
+  for (scan_pc = pc - PPC_INSN_SIZE;
+       scan_pc >= epilogue_start;
+       scan_pc -= PPC_INSN_SIZE)
+    {
+      if (!safe_frame_unwind_memory (fr, scan_pc, insn_buf, PPC_INSN_SIZE))
+        return 0;
+      insn = extract_signed_integer (insn_buf, PPC_INSN_SIZE);
+      if (insn_changes_sp_or_jumps (insn))
+        return 1;
+    }
+
+  return 0;
+}
+
 
 /* Fill in fi->saved_regs */
 
@@ -3342,6 +3450,8 @@
   set_gdbarch_deprecated_extract_struct_value_address (gdbarch, rs6000_extract_struct_value_address);
 
   set_gdbarch_skip_prologue (gdbarch, rs6000_skip_prologue);
+  set_gdbarch_in_function_epilogue_p (gdbarch, rs6000_in_function_epilogue_p);
+
   set_gdbarch_inner_than (gdbarch, core_addr_lessthan);
   set_gdbarch_breakpoint_from_pc (gdbarch, rs6000_breakpoint_from_pc);
 

  reply	other threads:[~2006-01-12 23:53 UTC|newest]

Thread overview: 39+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2005-11-30 23:56 [PATCH] add 'rs6000_in_function_epilogue_p()' Paul Gilliam
2005-12-01  5:21 ` Jim Blandy
2005-12-01 18:27   ` Paul Gilliam
2005-12-01 20:14     ` Paul Gilliam
2005-12-02  1:13 ` Jim Blandy
2005-12-02  1:23   ` Daniel Jacobowitz
2005-12-02 20:12     ` Paul Gilliam
2005-12-02 20:17       ` Paul Gilliam
2005-12-03  3:05       ` Jim Blandy
2005-12-02 23:38         ` Jim Blandy
2005-12-04 20:19         ` Daniel Jacobowitz
2005-12-04 18:59           ` Daniel Jacobowitz
2005-12-04 20:48           ` Jim Blandy
2005-12-04 21:12             ` Jim Blandy
2005-12-04 21:16             ` Daniel Jacobowitz
2005-12-04 21:22               ` Jim Blandy
2005-12-02  4:02   ` Joel Brobecker
2005-12-02 18:44   ` Mark Kettenis
2005-12-02 19:15   ` [PATCH] add 'rs6000_in_function_epilogue_p()' (Revised) Paul Gilliam
2005-12-02 20:28     ` Mark Kettenis
2005-12-02 21:19       ` Daniel Jacobowitz
2005-12-02 21:21         ` Mark Kettenis
2005-12-03  4:53       ` [PATCH] add 'rs6000_in_function_epilogue_p()' (Revised, again) Paul Gilliam
2005-12-03  5:43         ` Paul Gilliam
2005-12-02 21:44     ` [PATCH] add 'rs6000_in_function_epilogue_p()' (Revised) Kevin Buettner
2005-12-06 15:20       ` Paul Gilliam
2005-12-06 15:15         ` Paul Gilliam
2005-12-08  4:42         ` Kevin Buettner
2006-01-11 17:44       ` Paul Gilliam
2006-01-12  0:12         ` Paul Gilliam
2006-01-12 23:53           ` Paul Gilliam [this message]
2006-01-13 21:05             ` Mark Kettenis
2006-01-17  3:46               ` Paul Gilliam
2006-01-17 19:29                 ` Mark Kettenis
2006-02-09 17:46                 ` Kevin Buettner
2005-12-02 22:19     ` Jim Blandy
2005-12-02 22:28       ` Daniel Jacobowitz
2005-12-02 23:20         ` Jim Blandy
2005-12-03 12:48       ` Paul Gilliam

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=200601121623.15363.pgilliam@us.ibm.com \
    --to=pgilliam@us.ibm.com \
    --cc=gdb-patches@sourceware.org \
    --cc=kevinb@redhat.com \
    --cc=mark.kettenis@xs4all.nl \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox