Remove code handling old ARM aliases from GDB

Remove code handling old ARM aliases from GDB

[Joseph S. Myers]

There used to be various alternative target triplets for some ARM
targets such as strongarm*-*-*, thumb*-*-* and xscale*-*-*.  These are
now handled by config.sub as aliases for arm*-*-* (or in the case of
ep9312*-*-*, rejected outright by config.sub), so no code in configure
scripts or testsuites needs to allow for the old alternative names any
more; this patch removes code handling them from GDB.  I previously
removed such code from binutils in
<http://sourceware.org/ml/binutils/2011-03/msg00576.html>.  OK to
commit?

gdb:
2011-05-04  Joseph Myers  <joseph@codesourcery.com>

	* configure.host (xscale*): Don't handle target.
	* configure.tgt (thumb*-*-* | strongarm*-*-* | xscale-*-*): Don't
	handle targets.

gdb/gdbserver:
2011-05-04  Joseph Myers  <joseph@codesourcery.com>

	* README: Don't mention xscale*-*-linux*.
	* configure.srv (xscale*-*-linux*): Don't handle target.

gdb/testsuite:
2011-05-04  Joseph Myers  <joseph@codesourcery.com>

	* gdb.base/a2-run.exp (strongarm-*-coff): Don't handle target.
	* gdb.base/float.exp (xscale*-*-*, strongarm*-*-*): Don't handle
	targets.
	* gdb.base/long_long.exp (xscale*-*-*, strongarm*-*-*): Don't
	handle targets.

sim:
2011-05-04  Joseph Myers  <joseph@codesourcery.com>

	* configure.tgt (thumb*-*-* | strongarm*-*-* | xscale-*-*): Don't
	handle targets.
	* configure: Regenerate.

sim/testsuite:
2011-05-04  Joseph Myers  <joseph@codesourcery.com>

	* configure: Regenerate.
	* sim/arm/allinsn.exp (xscale*-*-*): Don't handle target.
	* sim/arm/misc.exp (thumb*-*-*, xscale*-*-*): Don't handle
	targets.
	* sim/arm/iwmmxt/iwmmxt.exp: Test for arm*-*-* instead of
	xscale*-*-*.
	* sim/arm/thumb/allthumb.exp (thumb*-*-*): Don't handle target.
	* sim/arm/xscale/xscale.exp: Test for arm*-*-* instead of
	xscale*-*-*.

Index: gdb/configure.host
===================================================================
RCS file: /cvs/src/src/gdb/configure.host,v
retrieving revision 1.105
diff -u -r1.105 configure.host
--- gdb/configure.host	13 Jan 2011 16:24:12 -0000	1.105
+++ gdb/configure.host	4 May 2011 18:52:01 -0000
@@ -51,7 +51,6 @@
 s390*)			gdb_host_cpu=s390 ;;
 sh*)			gdb_host_cpu=sh ;;
 x86_64*)		gdb_host_cpu=i386 ;;
-xscale*)		gdb_host_cpu=arm ;;
 m32r*)			gdb_host_cpu=m32r ;;
 xtensa*)		gdb_host_cpu=xtensa ;;
 *)			gdb_host_cpu=$host_cpu ;;
Index: gdb/configure.tgt
===================================================================
RCS file: /cvs/src/src/gdb/configure.tgt,v
retrieving revision 1.243
diff -u -r1.243 configure.tgt
--- gdb/configure.tgt	6 Mar 2011 00:22:38 -0000	1.243
+++ gdb/configure.tgt	4 May 2011 18:52:01 -0000
@@ -94,7 +94,7 @@
 	# Target: SymbianOS/arm
 	gdb_target_obs="arm-tdep.o solib-target.o arm-symbian-tdep.o"
 	;;
-arm*-*-* | thumb*-*-* | strongarm*-*-* | xscale-*-*)
+arm*-*-*)
 	# Target: ARM embedded system
 	gdb_target_obs="arm-tdep.o"
 	gdb_sim=../sim/arm/libsim.a
Index: gdb/gdbserver/README
===================================================================
RCS file: /cvs/src/src/gdb/gdbserver/README,v
retrieving revision 1.6
diff -u -r1.6 README
--- gdb/gdbserver/README	29 Dec 2010 16:15:42 -0000	1.6
+++ gdb/gdbserver/README	4 May 2011 18:52:01 -0000
@@ -99,7 +99,6 @@
 	sh-*-linux*
 	spu*-*-*
 	x86_64-*-linux*
-	xscale*-*-linux*
 
 Configuring GDBserver you should specify the same machine for host and
 target (which are the machine that GDBserver is going to run on.  This
Index: gdb/gdbserver/configure.srv
===================================================================
RCS file: /cvs/src/src/gdb/gdbserver/configure.srv,v
retrieving revision 1.59
diff -u -r1.59 configure.srv
--- gdb/gdbserver/configure.srv	29 Dec 2010 16:15:42 -0000	1.59
+++ gdb/gdbserver/configure.srv	4 May 2011 18:52:01 -0000
@@ -257,11 +257,6 @@
 			srv_mingw=yes
 			;;
 
-  xscale*-*-linux*)	srv_regobj=reg-arm.o
-			srv_tgtobj="linux-low.o linux-arm-low.o"
-			srv_linux_usrregs=yes
-			srv_linux_thread_db=yes
-			;;
   xtensa*-*-linux*)	srv_regobj=reg-xtensa.o
 			srv_tgtobj="linux-low.o linux-xtensa-low.o"
 			srv_linux_regsets=yes
Index: gdb/testsuite/gdb.base/a2-run.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/a2-run.exp,v
retrieving revision 1.18
diff -u -r1.18 a2-run.exp
--- gdb/testsuite/gdb.base/a2-run.exp	7 Mar 2011 16:03:02 -0000	1.18
+++ gdb/testsuite/gdb.base/a2-run.exp	4 May 2011 18:52:01 -0000
@@ -111,7 +111,7 @@
     verbose "Timeout is now $timeout seconds" 2
     gdb_expect -re "$gdb_prompt $" {}
 } else {
-	setup_xfail "mips-idt-*" "arm-*-coff strongarm-*-coff"
+	setup_xfail "mips-idt-*" "arm-*-coff"
 	gdb_run_cmd 5
 	gdb_expect {
 	    -re ".*120.*$gdb_prompt $"\
@@ -139,7 +139,7 @@
     verbose "Timeout is now $timeout seconds" 2
     gdb_expect -re "$gdb_prompt $" {}
 } else {
-    setup_xfail "arm-*-coff strongarm-*-coff"
+    setup_xfail "arm-*-coff"
     gdb_expect {
 	    -re ".*120.*$gdb_prompt $"\
 				{ pass "run \"$testfile\" again with same args" }
@@ -216,7 +216,7 @@
     verbose "Timeout is now $timeout seconds" 2
     gdb_expect -re "$gdb_prompt $" {}
 } else {
-    setup_xfail "arm-*-coff strongarm-*-coff"
+    setup_xfail "arm-*-coff"
     gdb_expect {
 	    -re ".*720.*$gdb_prompt $" {
 		pass "run \"$testfile\" again after setting args"
Index: gdb/testsuite/gdb.base/float.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/float.exp,v
retrieving revision 1.23
diff -u -r1.23 float.exp
--- gdb/testsuite/gdb.base/float.exp	1 Jan 2011 15:33:41 -0000	1.23
+++ gdb/testsuite/gdb.base/float.exp	4 May 2011 18:52:01 -0000
@@ -42,9 +42,7 @@
 
 if { [istarget "alpha*-*-*"] } then {
     gdb_test "info float" "f0.*" "info float"
-} elseif { [istarget "arm*-*-*"] || \
-	[istarget "xscale*-*-*"] || \
-	[istarget "strongarm*-*-*"] } then {
+} elseif { [istarget "arm*-*-*"] } then {
     gdb_test_multiple "info float" "info float" {
 	-re "Software FPU type.*mask:.*flags:.*$gdb_prompt $" {
 	    pass "info float (FPA)"
Index: gdb/testsuite/gdb.base/long_long.exp
===================================================================
RCS file: /cvs/src/src/gdb/testsuite/gdb.base/long_long.exp,v
retrieving revision 1.34
diff -u -r1.34 long_long.exp
--- gdb/testsuite/gdb.base/long_long.exp	1 Jan 2011 15:33:42 -0000	1.34
+++ gdb/testsuite/gdb.base/long_long.exp	4 May 2011 18:52:01 -0000
@@ -165,9 +165,7 @@
     # words swapped in big endian format.
     # EABI targets default to natural-endian VFP format.
 
-    if { ([istarget "arm*-*-*"] || \
-	  [istarget "xscale*-*-*"] || \
-	  [istarget "strongarm*-*-*"]) \
+    if { ([istarget "arm*-*-*"]) \
 	 && !([istarget "*-*-*eabi"] || \
 	      [istarget "*-*-mingw32ce*"] || \
 	      [istarget "*-*-openbsd*"] || \
Index: sim/configure
===================================================================
RCS file: /cvs/src/src/sim/configure,v
retrieving revision 1.35
diff -u -r1.35 configure
--- sim/configure	6 Mar 2011 00:20:17 -0000	1.35
+++ sim/configure	4 May 2011 18:52:03 -0000
@@ -3623,7 +3623,7 @@
 sim_igen=no
 sim_arch=
 case "${target}" in
-   arm*-*-* | thumb*-*-* | strongarm*-*-* | xscale-*-*)
+   arm*-*-*)
 
   sim_arch=arm
 
Index: sim/configure.tgt
===================================================================
RCS file: /cvs/src/src/sim/configure.tgt,v
retrieving revision 1.2
diff -u -r1.2 configure.tgt
--- sim/configure.tgt	6 Mar 2011 00:20:19 -0000	1.2
+++ sim/configure.tgt	4 May 2011 18:52:03 -0000
@@ -16,7 +16,7 @@
 sim_igen=no
 sim_arch=
 case "${target}" in
-   arm*-*-* | thumb*-*-* | strongarm*-*-* | xscale-*-*)
+   arm*-*-*)
        SIM_ARCH(arm)
        sim_testsuite=yes
        ;;
Index: sim/testsuite/configure
===================================================================
RCS file: /cvs/src/src/sim/testsuite/configure,v
retrieving revision 1.6
diff -u -r1.6 configure
--- sim/testsuite/configure	26 Apr 2010 16:23:23 -0000	1.6
+++ sim/testsuite/configure	4 May 2011 18:52:03 -0000
@@ -1823,13 +1823,16 @@
 sim_igen=no
 sim_arch=
 case "${target}" in
-   arm*-*-* | thumb*-*-* | strongarm*-*-* | xscale-*-*)
+   arm*-*-*)
        sim_arch=arm
        sim_testsuite=yes
        ;;
    avr*-*-*)
        sim_arch=avr
        ;;
+   bfin-*-*)
+       sim_arch=bfin
+       ;;
    cr16*-*-*)
        sim_arch=cr16
        sim_testsuite=yes
Index: sim/testsuite/sim/arm/allinsn.exp
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/allinsn.exp,v
retrieving revision 1.1
diff -u -r1.1 allinsn.exp
--- sim/testsuite/sim/arm/allinsn.exp	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/allinsn.exp	4 May 2011 18:52:03 -0000
@@ -1,6 +1,6 @@
 # ARM simulator testsuite.
 
-if { [istarget arm*-*-*] || [istarget xscale*-*-*] } {
+if { [istarget arm*-*-*] } {
     # load support procs (none yet)
     # load_lib cgen.exp
 
Index: sim/testsuite/sim/arm/misc.exp
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/misc.exp,v
retrieving revision 1.1
diff -u -r1.1 misc.exp
--- sim/testsuite/sim/arm/misc.exp	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/misc.exp	4 May 2011 18:52:03 -0000
@@ -1,6 +1,6 @@
 # Miscellaneous ARM simulator testcases
 
-if { [istarget arm*-*-*] || [istarget thumb*-*-*] || [istarget xscale*-*-*] } {
+if { [istarget arm*-*-*] } {
     # load support procs
     # load_lib cgen.exp
 
Index: sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp,v
retrieving revision 1.1
diff -u -r1.1 iwmmxt.exp
--- sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp	4 May 2011 18:52:03 -0000
@@ -1,6 +1,6 @@
 # Intel(r) Wireless MMX(tm) technology simulator testsuite.
 
-if { [istarget xscale*-*-*] } {
+if { [istarget arm*-*-*] } {
     # load support procs (none yet)
     # load_lib cgen.exp
 
Index: sim/testsuite/sim/arm/thumb/allthumb.exp
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/thumb/allthumb.exp,v
retrieving revision 1.1
diff -u -r1.1 allthumb.exp
--- sim/testsuite/sim/arm/thumb/allthumb.exp	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/thumb/allthumb.exp	4 May 2011 18:52:03 -0000
@@ -1,7 +1,6 @@
 # ARM simulator testsuite.
 
-if { [istarget arm*-*-*]
-     || [istarget thumb*-*-*] } {
+if { [istarget arm*-*-*] } {
     # load support procs (none yet)
     # load_lib cgen.exp
 
Index: sim/testsuite/sim/arm/xscale/xscale.exp
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/xscale/xscale.exp,v
retrieving revision 1.1
diff -u -r1.1 xscale.exp
--- sim/testsuite/sim/arm/xscale/xscale.exp	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/xscale/xscale.exp	4 May 2011 18:52:03 -0000
@@ -1,6 +1,6 @@
 # XSCALE simulator testsuite.
 
-if { [istarget xscale*-*-*] } {
+if { [istarget arm*-*-*] } {
     # load support procs (none yet)
     # load_lib cgen.exp
 

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: Remove code handling old ARM aliases from GDB

[Mark Kettenis]

> Date: Wed, 4 May 2011 19:00:14 +0000 (UTC)
> From: "Joseph S. Myers" <joseph@codesourcery.com>
> 
> There used to be various alternative target triplets for some ARM
> targets such as strongarm*-*-*, thumb*-*-* and xscale*-*-*.  These are
> now handled by config.sub as aliases for arm*-*-* (or in the case of
> ep9312*-*-*, rejected outright by config.sub), so no code in configure
> scripts or testsuites needs to allow for the old alternative names any
> more; this patch removes code handling them from GDB.  I previously
> removed such code from binutils in
> <http://sourceware.org/ml/binutils/2011-03/msg00576.html>.  OK to
> commit?

Makes sense to me.  Go ahead.

> gdb:
> 2011-05-04  Joseph Myers  <joseph@codesourcery.com>
> 
> 	* configure.host (xscale*): Don't handle target.
> 	* configure.tgt (thumb*-*-* | strongarm*-*-* | xscale-*-*): Don't
> 	handle targets.
> 
> gdb/gdbserver:
> 2011-05-04  Joseph Myers  <joseph@codesourcery.com>
> 
> 	* README: Don't mention xscale*-*-linux*.
> 	* configure.srv (xscale*-*-linux*): Don't handle target.
> 
> gdb/testsuite:
> 2011-05-04  Joseph Myers  <joseph@codesourcery.com>
> 
> 	* gdb.base/a2-run.exp (strongarm-*-coff): Don't handle target.
> 	* gdb.base/float.exp (xscale*-*-*, strongarm*-*-*): Don't handle
> 	targets.
> 	* gdb.base/long_long.exp (xscale*-*-*, strongarm*-*-*): Don't
> 	handle targets.
> 
> sim:
> 2011-05-04  Joseph Myers  <joseph@codesourcery.com>
> 
> 	* configure.tgt (thumb*-*-* | strongarm*-*-* | xscale-*-*): Don't
> 	handle targets.
> 	* configure: Regenerate.
> 
> sim/testsuite:
> 2011-05-04  Joseph Myers  <joseph@codesourcery.com>
> 
> 	* configure: Regenerate.
> 	* sim/arm/allinsn.exp (xscale*-*-*): Don't handle target.
> 	* sim/arm/misc.exp (thumb*-*-*, xscale*-*-*): Don't handle
> 	targets.
> 	* sim/arm/iwmmxt/iwmmxt.exp: Test for arm*-*-* instead of
> 	xscale*-*-*.
> 	* sim/arm/thumb/allthumb.exp (thumb*-*-*): Don't handle target.
> 	* sim/arm/xscale/xscale.exp: Test for arm*-*-* instead of
> 	xscale*-*-*.

Re: Remove code handling old ARM aliases from GDB

[Pedro Alves]

On Wednesday 04 May 2011 20:00:14, Joseph S. Myers wrote:
> There used to be various alternative target triplets for some ARM
> targets such as strongarm*-*-*, thumb*-*-* and xscale*-*-*.  These are
> now handled by config.sub as aliases for arm*-*-* (or in the case of
> ep9312*-*-*, rejected outright by config.sub), so no code in configure
> scripts or testsuites needs to allow for the old alternative names any
> more; this patch removes code handling them from GDB.  I previously
> removed such code from binutils in
> <http://sourceware.org/ml/binutils/2011-03/msg00576.html>.  OK to
> commit?

Sure, thanks.

> --- sim/testsuite/configure	26 Apr 2010 16:23:23 -0000	1.6
> +++ sim/testsuite/configure	4 May 2011 18:52:03 -0000
> @@ -1823,13 +1823,16 @@
>  sim_igen=no
>  sim_arch=
>  case "${target}" in
> -   arm*-*-* | thumb*-*-* | strongarm*-*-* | xscale-*-*)
> +   arm*-*-*)
>         sim_arch=arm
>         sim_testsuite=yes
>         ;;
>     avr*-*-*)
>         sim_arch=avr
>         ;;
> +   bfin-*-*)
> +       sim_arch=bfin
> +       ;;

Eh, looks like someone forgot to regenerate this file
before.  I see that missing on the branch as well.
I'll fix it.

>     cr16*-*-*)
>         sim_arch=cr16
>         sim_testsuite=yes


-- 
Pedro Alves

Re: Remove code handling old ARM aliases from GDB

[Pedro Alves]

On Wednesday 04 May 2011 20:23:01, Pedro Alves wrote:
> > +   bfin-*-*)
> > +       sim_arch=bfin
> > +       ;;
> 
> Eh, looks like someone forgot to regenerate this file
> before.  I see that missing on the branch as well.
> I'll fix it.

Done.

-- 
Pedro Alves

Index: ChangeLog
===================================================================
RCS file: /cvs/src/src/sim/testsuite/ChangeLog,v
retrieving revision 1.66
diff -u -p -r1.66 ChangeLog
--- ChangeLog	6 Oct 2010 23:41:20 -0000	1.66
+++ ChangeLog	4 May 2011 19:21:57 -0000
@@ -1,3 +1,7 @@
+2011-05-04  Pedro Alves  <pedro@codesourcery.com>
+
+	* configure: Regenerate.
+
 2010-10-07  Hans-Peter Nilsson  <hp@axis.com>
 
 	* sim/cris/c/seek3.c, sim/cris/c/seek4.c: New tests.
Index: configure
===================================================================
RCS file: /cvs/src/src/sim/testsuite/configure,v
retrieving revision 1.6
diff -u -p -r1.6 configure
--- configure	26 Apr 2010 16:23:23 -0000	1.6
+++ configure	4 May 2011 19:21:58 -0000
@@ -1830,6 +1830,9 @@ case "${target}" in
    avr*-*-*)
        sim_arch=avr
        ;;
+   bfin-*-*)
+       sim_arch=bfin
+       ;;
    cr16*-*-*)
        sim_arch=cr16
        sim_testsuite=yes

Re: Remove code handling old ARM aliases from GDB

[Hans-Peter Nilsson]

On Wed, 4 May 2011, Joseph S. Myers wrote:

> sim/testsuite:
> 2011-05-04  Joseph Myers  <joseph@codesourcery.com>
>
> 	* configure: Regenerate.
> 	* sim/arm/allinsn.exp (xscale*-*-*): Don't handle target.
> 	* sim/arm/misc.exp (thumb*-*-*, xscale*-*-*): Don't handle
> 	targets.
> 	* sim/arm/iwmmxt/iwmmxt.exp: Test for arm*-*-* instead of
> 	xscale*-*-*.
> 	* sim/arm/thumb/allthumb.exp (thumb*-*-*): Don't handle target.
> 	* sim/arm/xscale/xscale.exp: Test for arm*-*-* instead of
> 	xscale*-*-*.

> Index: sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp
> ===================================================================
> RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp,v
> retrieving revision 1.1
> diff -u -r1.1 iwmmxt.exp
> --- sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp	1 Apr 2003 11:07:58 -0000	1.1
> +++ sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp	4 May 2011 18:52:03 -0000
> @@ -1,6 +1,6 @@
>  # Intel(r) Wireless MMX(tm) technology simulator testsuite.
>
> -if { [istarget xscale*-*-*] } {
> +if { [istarget arm*-*-*] } {
>      # load support procs (none yet)
>      # load_lib cgen.exp

> Index: sim/testsuite/sim/arm/xscale/xscale.exp
> ===================================================================
> RCS file: /cvs/src/src/sim/testsuite/sim/arm/xscale/xscale.exp,v
> retrieving revision 1.1
> diff -u -r1.1 xscale.exp
> --- sim/testsuite/sim/arm/xscale/xscale.exp	1 Apr 2003 11:07:58 -0000	1.1
> +++ sim/testsuite/sim/arm/xscale/xscale.exp	4 May 2011 18:52:03 -0000
> @@ -1,6 +1,6 @@
>  # XSCALE simulator testsuite.
>
> -if { [istarget xscale*-*-*] } {
> +if { [istarget arm*-*-*] } {
>      # load support procs (none yet)
>      # load_lib cgen.exp

How did you test these changes?

These changes broke testing for arm-elfo, on ang x86_64 host
using current binutils CVS like so:

Schedule of variations:
    arm-sim

Running target arm-sim
Using /usr/share/dejagnu/baseboards/arm-sim.exp as board description file for target.
Using /usr/share/dejagnu/config/sim.exp as generic interface file for target.
Using /usr/share/dejagnu/baseboards/basic-sim.exp as board description file for target.
Using /tmp/hpautotest-sim/src/sim/testsuite/config/default.exp
as tool-and-target-specific interface file.
Running /tmp/hpautotest-sim/src/sim/testsuite/sim/arm/allinsn.exp ...
Running /tmp/hpautotest-sim/src/sim/testsuite/sim/arm/iwmmxt/iwmmxt.exp ...
FAIL: xscale wcmpgt.cgs (execution)
FAIL: xscale wmac.cgs (execution)
FAIL: xscale wsra.cgs (execution)
Running /tmp/hpautotest-sim/src/sim/testsuite/sim/arm/misc.exp ...
Running /tmp/hpautotest-sim/src/sim/testsuite/sim/arm/thumb/allthumb.exp ...
Running /tmp/hpautotest-sim/src/sim/testsuite/sim/arm/xscale/xscale.exp ...
WARNING: program timed out.
FAIL: xscale blx.cgs (execution)

Please fix, suggesting by adding needed options to the assembler
and sim invocations.  I don't know if src/sim can handle
whatever is tested by iwmmxt.exp and xscale.exp, but if it can't
at all (i.e. these tests are not intended to be run with src/sim
but with real hw or qemu), an alternative would be to
sufficiently gate the tests.  If all else fails, reverting your
change is an option.

brgds, H-P

Re: Remove code handling old ARM aliases from GDB

[Hans-Peter Nilsson]

On Thu, 5 May 2011, Hans-Peter Nilsson wrote:
> These changes broke testing for arm-elfo, on ang x86_64 host

JFTR, that target should be "arm-elf" (a typo type "o")

brgds, H-P

Re: Remove code handling old ARM aliases from GDB

[Joseph S. Myers]

On Thu, 5 May 2011, Hans-Peter Nilsson wrote:

> > -if { [istarget xscale*-*-*] } {
> > +if { [istarget arm*-*-*] } {

> How did you test these changes?

I considered them sufficiently obviously syntactically correct not to need 
testing.

> These changes broke testing for arm-elfo, on ang x86_64 host
> using current binutils CVS like so:

> FAIL: xscale wcmpgt.cgs (execution)
> FAIL: xscale wmac.cgs (execution)
> FAIL: xscale wsra.cgs (execution)

> WARNING: program timed out.
> FAIL: xscale blx.cgs (execution)

I don't call that breaking testing; I call that exposing bugs (whether in 
the simulator or testsuite) that were previously hidden by the use of an 
obsolete target triplet name.  Since the simulator features do not depend 
on the target triplet name, it seems clear that these tests should be run 
for all ARM targets - and so it seems better to expose the problems by 
using a current target name than to make the formally no-op change of 
using if { 0 } to replace an always-false conditional.

If the issue were actually missing assembler or simulator options a lot 
more tests in those directories would fail, so it seems much more likely 
that the relevant simulator functionality has become bitrotten over time.

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: Remove code handling old ARM aliases from GDB

[Hans-Peter Nilsson]

On Thu, 5 May 2011, Joseph S. Myers wrote:
> On Thu, 5 May 2011, Hans-Peter Nilsson wrote:
>
> > > -if { [istarget xscale*-*-*] } {
> > > +if { [istarget arm*-*-*] } {
>
> > How did you test these changes?
>
> I considered them sufficiently obviously syntactically correct not to need
> testing.

It might have seemed like that at the time, but obviously it
wasn't so.

> > These changes broke testing for arm-elfo, on ang x86_64 host
> > using current binutils CVS like so:
>
> > FAIL: xscale wcmpgt.cgs (execution)
> > FAIL: xscale wmac.cgs (execution)
> > FAIL: xscale wsra.cgs (execution)
>
> > WARNING: program timed out.
> > FAIL: xscale blx.cgs (execution)
>
> I don't call that breaking testing; I call that exposing bugs (whether in
> the simulator or testsuite) that were previously hidden by the use of an
> obsolete target triplet name.

arm-elf passed before and does not anymore after those changes.
That's breaking testing.  Any other definition falls prey to
"latent bugs" weaseling.

If you can't invest the time to verify whether there are indeed
options missing, I'd be ok with an "if { 0 }" replacement plus a
big comment.  It'd also be prudent to open a PR (obviously
timing out is a bad choice unless that's what the default ARM
target hw would do).  I unfortunately doubt we have a
sufficiently interested target maintainer for src/sim/arm.

brgds, H-P

Re: Remove code handling old ARM aliases from GDB

[Joseph S. Myers]

On Thu, 5 May 2011, Hans-Peter Nilsson wrote:

> On Thu, 5 May 2011, Joseph S. Myers wrote:
> > On Thu, 5 May 2011, Hans-Peter Nilsson wrote:
> >
> > > > -if { [istarget xscale*-*-*] } {
> > > > +if { [istarget arm*-*-*] } {
> >
> > > How did you test these changes?
> >
> > I considered them sufficiently obviously syntactically correct not to need
> > testing.
> 
> It might have seemed like that at the time, but obviously it
> wasn't so.

Well, it was *syntactically correct*, as shown by the tests running in 
your log rather than hitting a Tcl syntax error.

> > I don't call that breaking testing; I call that exposing bugs (whether in
> > the simulator or testsuite) that were previously hidden by the use of an
> > obsolete target triplet name.
> 
> arm-elf passed before and does not anymore after those changes.

"arm-elf" is not a testcase.

A testsuite regression is when a test assertion fails before and passes 
after a patch.  A test assertion corresponds to particular text that may 
appear after "PASS: " or "FAIL: " in the .sum file (for a properly 
functioning DejaGnu testsuite, a slightly looser definition may be 
appropriate for GDB at present).  A failure of a test that was not run 
before the patch is not a regression.

> That's breaking testing.

No, breaking testing would be introducing a syntax error so that tests 
that previously run stop running or run incorrectly.  Output with FAILs in 
it isn't broken, it simply gives information about what works and what 
doesn't.  If a FAIL replaced a PASS that would be a regression, not 
breaking testing, but that isn't the case here either.

> Any other definition falls prey to
> "latent bugs" weaseling.

The simulator bugs exposed weren't latent since the simulator probably had 
those bugs for years.  So far as there were testsuite bugs (the tests not 
running) those were *fixed* by the patch.

I consider the situation here to be exactly the same as if I'd added 
target-independent tests to the GDB testsuite for a feature that should 
work on all targets but that would require target-specific work for most 
targets to make it work properly.  Such tests would, of course, expose 
bugs in various targets, but it would be up to those interested in 
improving results for each target to fix them there; I don't think there's 
any expectation that such new tests will pass everywhere.  (An example of 
potential new tests that would have such an issue would be tests for how 
GDB handles argument passing and return involving complex numbers; that 
needs ABI support for each target, and very few GDB targets have such ABI 
support that actually corresponds to what GCC does.)  I think enabling 
tests that were wrongly disabled is just like adding new tests - if 
someone fixing them is unclear about the semantics of instructions being 
tested, I'll try to advise on specific questions, just as I would on 
questions about the hypothetical tests for complex number ABIs.

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: Remove code handling old ARM aliases from GDB

[Hans-Peter Nilsson]

On Thu, 5 May 2011, Joseph S. Myers wrote:
> On Thu, 5 May 2011, Hans-Peter Nilsson wrote:
>
> > On Thu, 5 May 2011, Joseph S. Myers wrote:
> > > On Thu, 5 May 2011, Hans-Peter Nilsson wrote:
> > >
> > > > > -if { [istarget xscale*-*-*] } {
> > > > > +if { [istarget arm*-*-*] } {
> > >
> > > > How did you test these changes?
> > >
> > > I considered them sufficiently obviously syntactically correct not to need
> > > testing.
> >
> > It might have seemed like that at the time, but obviously it
> > wasn't so.
>
> Well, it was *syntactically correct*, as shown by the tests running in
> your log rather than hitting a Tcl syntax error.

Please, we're not talking about Tcl syntax.

> > > I don't call that breaking testing; I call that exposing bugs (whether in
> > > the simulator or testsuite) that were previously hidden by the use of an
> > > obsolete target triplet name.
> >
> > arm-elf passed before and does not anymore after those changes.
>
> "arm-elf" is not a testcase.

I'm not sure what you mean, but "arm-elf passed" was a
contraction for "when running the sim tests for the arm-elf
target, they all passed", which should be obvious from the
context.

> A testsuite regression is when a test assertion fails before and passes
> after a patch.  A test assertion corresponds to particular text that may
> appear after "PASS: " or "FAIL: " in the .sum file (for a properly
> functioning DejaGnu testsuite, a slightly looser definition may be
> appropriate for GDB at present).  A failure of a test that was not run
> before the patch is not a regression.

I did not say that a specific test regressed; I didn't even
mention the word regression.

But now that you mention it, the end result of running the
test-suite for the arm-elf is certainly a regression from before
those changes.

> > That's breaking testing.
>
> No, breaking testing would be introducing a syntax error so that tests
> that previously run stop running or run incorrectly.  Output with FAILs in
> it isn't broken, it simply gives information about what works and what
> doesn't.  If a FAIL replaced a PASS that would be a regression, not
> breaking testing, but that isn't the case here either.

Splitting of hairs, choice of words.  I guess we have to
disagree on the words, but I'm sure we can agree that running
the tests does exit with an error after, but didn't before your
changes.

As having introduced this, you're on the hook to investigate and
rectify.  Another option to those I mentioned is to xfail or
kfailing the failing tests.  Please.

brgds, H-P

Re: Remove code handling old ARM aliases from GDB

[Joseph S. Myers]

On Thu, 5 May 2011, Hans-Peter Nilsson wrote:

> Splitting of hairs, choice of words.  I guess we have to
> disagree on the words, but I'm sure we can agree that running
> the tests does exit with an error after, but didn't before your
> changes.

sim tests on their own, yes - I just consider this to be an *improvement* 
(better testsuite coverage, exit status more accurately reflecting the 
existence of defects).  I think of sim as part of GDB, and I've never seen 
the gdb testsuite anywhere near to having completely clean results for any 
target.

> As having introduced this, you're on the hook to investigate and
> rectify.

I'm not expert on GDB policies.  Where is this documented and would this 
apply to the case I described of tests for complex numbers ABI handling, 
which I consider exactly analogous?  I don't believe this would apply 
under GCC practice, as it's not a regression but (effectively) a new test.

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: Remove code handling old ARM aliases from GDB

[Joseph S. Myers]

Please see if this xfailing patch is effective for you.  I still maintain 
that problems exposed by a new test (as opposed to defects in the test 
itself) should be considered completely independent of the test itself and 
something to be fixed or xfailed by those interested in the buggy area of 
the tools.

Index: sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs,v
retrieving revision 1.1
diff -u -r1.1 wcmpgt.cgs
--- sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs	5 May 2011 23:48:33 -0000
@@ -1,6 +1,7 @@
 # Intel(r) Wireless MMX(tm) technology testcase for WCMPGT
 # mach: xscale
 # as: -mcpu=xscale+iwmmxt
+# xfail: *-*-*
 
 	.include "testutils.inc"
 
Index: sim/testsuite/sim/arm/iwmmxt/wmac.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/wmac.cgs,v
retrieving revision 1.1
diff -u -r1.1 wmac.cgs
--- sim/testsuite/sim/arm/iwmmxt/wmac.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/iwmmxt/wmac.cgs	5 May 2011 23:48:33 -0000
@@ -1,6 +1,7 @@
 # Intel(r) Wireless MMX(tm) technology testcase for WMAC
 # mach: xscale
 # as: -mcpu=xscale+iwmmxt
+# xfail: *-*-*
 
 	.include "testutils.inc"
 
Index: sim/testsuite/sim/arm/iwmmxt/wsra.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/wsra.cgs,v
retrieving revision 1.1
diff -u -r1.1 wsra.cgs
--- sim/testsuite/sim/arm/iwmmxt/wsra.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/iwmmxt/wsra.cgs	5 May 2011 23:48:33 -0000
@@ -1,6 +1,7 @@
 # Intel(r) Wireless MMX(tm) technology testcase for WSRA
 # mach: xscale
 # as: -mcpu=xscale+iwmmxt
+# xfail: *-*-*
 
 	.include "testutils.inc"
 
Index: sim/testsuite/sim/arm/xscale/blx.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/xscale/blx.cgs,v
retrieving revision 1.1
diff -u -r1.1 blx.cgs
--- sim/testsuite/sim/arm/xscale/blx.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/xscale/blx.cgs	5 May 2011 23:48:33 -0000
@@ -1,5 +1,6 @@
 # arm testcase for bl$cond $offset24
 # mach: all
+# xfail: *-*-*
 
 	.include "testutils.inc"
 

-- 
Joseph S. Myers
joseph@codesourcery.com

Re: Remove code handling old ARM aliases from GDB

[Joel Brobecker]

> As having introduced this, you're on the hook to investigate and
> rectify.  Another option to those I mentioned is to xfail or
> kfailing the failing tests.  Please.

I've tried to understand the problem, and it seems to me that Joseph's
position is entirely reasonable in this case. He has effectively
introduced new tests that happen to fail in the arm-elf case, is that
correct? If they fail on arm-elf, someone who cares about this platform
should investigate them and determine whether to fix them, or whether
they are a sim/gdb problem (kfail), or a problem from an external
dependency (xfail).

-- 
Joel

Re: Remove code handling old ARM aliases from GDB

[Hans-Peter Nilsson]

On Thu, 5 May 2011, Joel Brobecker wrote:
> > As having introduced this, you're on the hook to investigate and
> > rectify.  Another option to those I mentioned is to xfail or
> > kfailing the failing tests.  Please.
>
> I've tried to understand the problem, and it seems to me that Joseph's
> position is entirely reasonable in this case. He has effectively
> introduced new tests that happen to fail in the arm-elf case, is that
> correct?

A gating test within the arm testsuite was changed from matching
xscale*-*-* to arm*-*-*, so going from never-match to
always-match which at a glance seems correct there.  Except that
some of the uncovered tests fail.

The theory is that way back then, a tuple matching xscale*-*-*
enabled stuff that was disabled on arm*-*-*.  (It seems this has
some bearing to reality, see sim/arm/ChangeLog).  Hopefully the
tests passed at that time, so no, they're not *new* tests.
Since then, at least parts of the xscale stuff have been folded
into arm*-*, but the testsuite hasn't been updated and the tests
and/or the simulator has rotted.

My take is that when you modify stuff and there's related
breakage where there was none apparent before your change, you
should at least put a minimum of effort into checking why, and
even fix it, regardless of whether it was a just a test that
never ran before or whatever.

> If they fail on arm-elf, someone who cares about this platform
> should investigate them and determine whether to fix them, or whether
> they are a sim/gdb problem (kfail), or a problem from an external
> dependency (xfail).

I opened PR 12737.

In my autotester for the src/sim simulators for various targets,
I'm using the thumbs-up from make check from the arm-elf
testsuite as a sign that it's at least no worse off than before
for that target.  Having new, old or re-discovered tests that
fail where none did before, doesn't help.  And no, keeping track
of new vs. old fails is overkill in the src/sim testsuite.

brgds, H-P

Re: Remove code handling old ARM aliases from GDB

[Hans-Peter Nilsson]

(Taking the liberty of CC:ing the list when replying to an
off-list message; I didn't see anything better kept off-list.)

On Thu, 5 May 2011, Joseph S. Myers wrote:

> Please see if this xfailing patch is effective for you.

Not that I have anything you couldn't test, but yes it was,
thanks.  Though, only *some* iwmmxt tests fail and all tests
pass the same GAS option seen below, so this is a sim problem,
thus it should as Joel Brobecker points out, be kfail, which
requires a PR, where the aforementioned PR12737 fits the bill.
I'll commit the following, so we can move on.  Though, I'll give
it another 24h for someone to shout in case I understood wrong.

sim/testsuite:
	PR sim/12737
	* sim/arm/iwmmxt/wcmpgt.cgs, sim/arm/iwmmxt/wmac.cgs,
	sim/arm/iwmmxt/wsra.cgs, sim/arm/xscale/blx.cgs: Kfail.


Index: sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs,v
retrieving revision 1.1
diff -u -r1.1 wcmpgt.cgs
--- sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/iwmmxt/wcmpgt.cgs	5 May 2011 23:48:33 -0000
@@ -1,6 +1,7 @@
 # Intel(r) Wireless MMX(tm) technology testcase for WCMPGT
 # mach: xscale
 # as: -mcpu=xscale+iwmmxt
+# kfail: sim/12737 *-*-*

 	.include "testutils.inc"

Index: sim/testsuite/sim/arm/iwmmxt/wmac.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/wmac.cgs,v
retrieving revision 1.1
diff -u -r1.1 wmac.cgs
--- sim/testsuite/sim/arm/iwmmxt/wmac.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/iwmmxt/wmac.cgs	5 May 2011 23:48:33 -0000
@@ -1,6 +1,7 @@
 # Intel(r) Wireless MMX(tm) technology testcase for WMAC
 # mach: xscale
 # as: -mcpu=xscale+iwmmxt
+# kfail: sim/12737 *-*-*

 	.include "testutils.inc"

Index: sim/testsuite/sim/arm/iwmmxt/wsra.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/iwmmxt/wsra.cgs,v
retrieving revision 1.1
diff -u -r1.1 wsra.cgs
--- sim/testsuite/sim/arm/iwmmxt/wsra.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/iwmmxt/wsra.cgs	5 May 2011 23:48:33 -0000
@@ -1,6 +1,7 @@
 # Intel(r) Wireless MMX(tm) technology testcase for WSRA
 # mach: xscale
 # as: -mcpu=xscale+iwmmxt
+# kfail: sim/12737 *-*-*

 	.include "testutils.inc"

Index: sim/testsuite/sim/arm/xscale/blx.cgs
===================================================================
RCS file: /cvs/src/src/sim/testsuite/sim/arm/xscale/blx.cgs,v
retrieving revision 1.1
diff -u -r1.1 blx.cgs
--- sim/testsuite/sim/arm/xscale/blx.cgs	1 Apr 2003 11:07:58 -0000	1.1
+++ sim/testsuite/sim/arm/xscale/blx.cgs	5 May 2011 23:48:33 -0000
@@ -1,5 +1,6 @@
 # arm testcase for bl$cond $offset24
 # mach: all
+# kfail: sim/12737 *-*-*

 	.include "testutils.inc"


brgds, H-P