Re: [RFA/hppa] fix for skipping prologues

[Andrew Cagney <cagney at gnu dot org>]

Randolph Chung wrote:
> If gcc schedules non-prologue insns into the prologue, the hppa frame
> unwinder gets confused. Here's a fix. It's also interesting to note that
> the problems are pretty much described by Mark's email to the gdb list
> recently: http://sources.redhat.com/ml/gdb/2004-11/msg00055.html
>
> In this case, we need both types of prologue skipping, one for placing
> breakpoints on functions and one for actually analyzing the prologue for
> unwinding.
>
> With Mark's fix, after_prologue() in hppa-tdep.c can go away. but that's
> for another patch....

Ok, also ok for the 6.3 branch (and if you can identify other HP/UX 
specific patches that should be back ported - I think there was one 
involving configure?) feel free to backport them.

Andrew

> 2004-11-09  Randolph Chung  <tausq@debian.org>
>
> 	* hppa-tdep.c (skip_prologue_hard_way): Make static, add
> 	stop_after_branch parameter and use it to determine if we should look
> 	at the delay slot of the first branch insn to find prologue insns.
> 	(hppa_skip_prologue): Update comment and call to skip_prologue_hard_way.
> 	(hppa_frame_cache): Likewise.
>
> Index: hppa-tdep.c
> ===================================================================
> RCS file: /cvs/src/src/gdb/hppa-tdep.c,v
> retrieving revision 1.176
> diff -u -p -r1.176 hppa-tdep.c
> --- hppa-tdep.c	3 Nov 2004 23:20:43 -0000	1.176
> +++ hppa-tdep.c	9 Nov 2004 16:42:53 -0000
> @@ -1247,14 +1247,15 @@ inst_saves_fr (unsigned long inst)
>     be in the prologue.  */
>  
>  
> -CORE_ADDR
> -skip_prologue_hard_way (CORE_ADDR pc)
> +static CORE_ADDR
> +skip_prologue_hard_way (CORE_ADDR pc, int stop_before_branch)
>  {
>    char buf[4];
>    CORE_ADDR orig_pc = pc;
>    unsigned long inst, stack_remaining, save_gr, save_fr, save_rp, save_sp;
>    unsigned long args_stored, status, i, restart_gr, restart_fr;
>    struct unwind_table_entry *u;
> +  int final_iteration;
>  
>    restart_gr = 0;
>    restart_fr = 0;
> @@ -1298,6 +1299,8 @@ restart:
>      save_fr |= (1 << i);
>    save_fr &= ~restart_fr;
>  
> +  final_iteration = 0;
> +
>    /* Loop until we find everything of interest or hit a branch.
>  
>       For unoptimized GCC code and for any HP CC code this will never ever
> @@ -1434,7 +1437,7 @@ restart:
>  
>        /* Quit if we hit any kind of branch.  This can happen if a prologue
>           instruction is in the delay slot of the first call/branch.  */
> -      if (is_branch (inst))
> +      if (is_branch (inst) && stop_before_branch)
>  	break;
>  
>        /* What a crock.  The HP compilers set args_stored even if no
> @@ -1455,6 +1458,13 @@ restart:
>  
>        /* Bump the PC.  */
>        pc += 4;
> +
> +      /* !stop_before_branch, so also look at the insn in the delay slot 
> +         of the branch.  */
> +      if (final_iteration)
> +	break;
> +      if (is_branch (inst))
> +	final_iteration = 1;
>      }
>  
>    /* We've got a tenative location for the end of the prologue.  However
> @@ -1513,12 +1523,13 @@ after_prologue (CORE_ADDR pc)
>  
>  /* To skip prologues, I use this predicate.  Returns either PC itself
>     if the code at PC does not look like a function prologue; otherwise
> -   returns an address that (if we're lucky) follows the prologue.  If
> -   LENIENT, then we must skip everything which is involved in setting
> -   up the frame (it's OK to skip more, just so long as we don't skip
> -   anything which might clobber the registers which are being saved.
> -   Currently we must not skip more on the alpha, but we might the lenient
> -   stuff some day.  */
> +   returns an address that (if we're lucky) follows the prologue.  
> +   
> +   hppa_skip_prologue is called by gdb to place a breakpoint in a function.
> +   It doesn't necessarily skips all the insns in the prologue. In fact
> +   we might not want to skip all the insns because a prologue insn may
> +   appear in the delay slot of the first branch, and we don't want to
> +   skip over the branch in that case.  */
>  
>  static CORE_ADDR
>  hppa_skip_prologue (CORE_ADDR pc)
> @@ -1543,7 +1554,7 @@ hppa_skip_prologue (CORE_ADDR pc)
>    if (post_prologue_pc != 0)
>      return max (pc, post_prologue_pc);
>    else
> -    return (skip_prologue_hard_way (pc));
> +    return (skip_prologue_hard_way (pc, 1));
>  }
>  
>  struct hppa_frame_cache
> @@ -1625,18 +1636,24 @@ hppa_frame_cache (struct frame_info *nex
>      int looking_for_rp = u->Save_RP;
>      int fp_loc = -1;
>  
> -    /* We have to use hppa_skip_prologue instead of just 
> +    /* We have to use skip_prologue_hard_way instead of just 
>         skip_prologue_using_sal, in case we stepped into a function without
>         symbol information.  hppa_skip_prologue also bounds the returned
>         pc by the passed in pc, so it will not return a pc in the next
> -       function.  */
> +       function.  
> +       
> +       We used to call hppa_skip_prologue to find the end of the prologue,
> +       but if some non-prologue instructions get scheduled into the prologue,
> +       and the program is compiled with debug information, the "easy" way
> +       in hppa_skip_prologue will return a prologue end that is too early
> +       for us to notice any potential frame adjustments.  */
>  
>      /* We used to use frame_func_unwind () to locate the beginning of the
>         function to pass to skip_prologue ().  However, when objects are 
>         compiled without debug symbols, frame_func_unwind can return the wrong 
>         function (or 0).  We can do better than that by using unwind records.  */
>  
> -    prologue_end = hppa_skip_prologue (u->region_start);
> +    prologue_end = skip_prologue_hard_way (u->region_start, 0);
>      end_pc = frame_pc_unwind (next_frame);
>  
>      if (prologue_end != 0 && end_pc > prologue_end)