Patchwork [v2] Don't throw an error in 'info registers' for unavailable MIPS registers.

login
register
mail settings
Submitter John Baldwin
Date June 16, 2017, 5:09 p.m.
Message ID <20170616170944.99055-1-jhb@FreeBSD.org>
Download mbox | patch
Permalink /patch/21054/
State New
Headers show

Comments

John Baldwin - June 16, 2017, 5:09 p.m.
'info registers' for MIPS throws an error and when it first encounters
an unavailable register.  This does not match other architectures
which annotate unavailable registers and continue to print out the
values of subsequent registers.  Replace the error by displaying an
aligned "<unavailable>".  This string is truncated to "<unavl>" when
displaying a 32-bit register.

gdb/ChangeLog:

	* mips-tdep.c (print_gp_register_row): Don't error for unavailable
	registers.
---
 gdb/ChangeLog   |  5 +++++
 gdb/mips-tdep.c | 10 ++++++++--
 2 files changed, 13 insertions(+), 2 deletions(-)
Maciej W. Rozycki - June 16, 2017, 8:08 p.m.
On Fri, 16 Jun 2017, John Baldwin wrote:

> index 3ff2dab491..adc843a5be 100644
> --- a/gdb/mips-tdep.c
> +++ b/gdb/mips-tdep.c
> @@ -6542,8 +6542,14 @@ print_gp_register_row (struct ui_file *file, struct frame_info *frame,
>        value = get_frame_register_value (frame, regnum);
>        if (value_optimized_out (value)
>  	|| !value_entirely_available (value))
> -	error (_("can't read register %d (%s)"),
> -	       regnum, gdbarch_register_name (gdbarch, regnum));
> +	{
> +	  fprintf_filtered (file, "%*s ",
> +			    (int) mips_abi_regsize (gdbarch) * 2,
> +			    mips_abi_regsize (gdbarch) == 4 ? "<unavl>"
> +			    : "<unavailable>");

 Formatting nit here -- you need to wrap an expression in parentheses when 
splitting between lines and indent accordingly, i.e. either:

+			    (mips_abi_regsize (gdbarch) == 4 ? "<unavl>"
+			     : "<unavailable>"));

or:

+			    (mips_abi_regsize (gdbarch) == 4
+			     ? "<unavl>" : "<unavailable>"));

> +	  col++;
> +	  continue;
> +	}

 Otherwise LGTM, so please consider your patch with such a change applied 
preapproved.

 I think that duplicate `col++' could be avoided, as could be the repeated 
calls to `mips_abi_regsize' and `register_size'.  That can be a separate 
clean-up though.

 Thanks for your work on this problem.

  Maciej

Patch

diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index 9062067bcf..26ea6fb8c7 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,8 @@ 
+2017-06-16  John Baldwin  <jhb@FreeBSD.org>
+
+	* mips-tdep.c (print_gp_register_row): Don't error for unavailable
+	registers.
+
 2017-06-16  Alan Hayward  <alan.hayward@arm.com>
 	    Pedro Alves  <palves@redhat.com>
 	    Yao Qi  <yao.qi@linaro.org>
diff --git a/gdb/mips-tdep.c b/gdb/mips-tdep.c
index 3ff2dab491..adc843a5be 100644
--- a/gdb/mips-tdep.c
+++ b/gdb/mips-tdep.c
@@ -6542,8 +6542,14 @@  print_gp_register_row (struct ui_file *file, struct frame_info *frame,
       value = get_frame_register_value (frame, regnum);
       if (value_optimized_out (value)
 	|| !value_entirely_available (value))
-	error (_("can't read register %d (%s)"),
-	       regnum, gdbarch_register_name (gdbarch, regnum));
+	{
+	  fprintf_filtered (file, "%*s ",
+			    (int) mips_abi_regsize (gdbarch) * 2,
+			    mips_abi_regsize (gdbarch) == 4 ? "<unavl>"
+			    : "<unavailable>");
+	  col++;
+	  continue;
+	}
       raw_buffer = value_contents_all (value);
       /* pad small registers */
       for (byte = 0;