Simulator dv-sockser.o patch testing

Simulator dv-sockser.o patch testing

[Joel Sherrill]

`Hi

I tested three configurations:

+ gdb 7.5.91 as baseline
+ gdb head with my last acinclude.m4 patch
+ gdb head with Mike's acinclude.m4 patch

I tested these 10 targets [1]:

bfin-elf cris-elf frv-elf iq2000-elf lm32-elf m32r-elf
m68hc11-elf mips-elf mn10300-elf sh64-elf

with: (1) no simulator option, (2) --disable-sim-hardware
and (3) --enable-sim-hardware

This is a total of 30 builds.

gdb 7.5.91 and my patch built 27 of the 30 and failed to
bfin-elf, m68hc11-elf, and mn10300-elf with --disable-sim-hardware.
Mike's patch let them all build.

The only discrepancy in the make check summary was that
WITH Mike's patch cris-elf with --enable-sim-hardware has this:

=== sim Summary ===
# of expected passes 304
# of untested testcases 136

In the 7.5.91 log, the 136 was 170.

I know absolutely nothing about cris and only ran this because
the configure file changed after running autoconf. Nothing specific
to the cris was changed. This could be something else going on.

As far as I can tell, Mike's patch is an improvement over mine
and what was in 7.5.91. If he regenerates the configure scripts
that need it, I would be perfectly happy with his patch being
applied.

This is the set of files which were touched by his patch or
regenerated in my tree:

#    modified:   sim/bfin/configure
#    modified:   sim/common/acinclude.m4
#    modified:   sim/cris/configure
#    modified:   sim/frv/configure
#    modified:   sim/iq2000/configure
#    modified:   sim/lm32/configure
#    modified:   sim/m32r/configure
#    modified:   sim/m68hc11/configure
#    modified:   sim/mips/configure
#    modified:   sim/mn10300/configure
#    modified:   sim/sh64/configure

Mike, can you commit yours? Please. :)

[1] I forgot to include mipstx39-elf in the targets built against
all three versions. I did run it against Mike's by hand and it is OK.

-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill@OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985

Re: Simulator dv-sockser.o patch testing

[Hans-Peter Nilsson]

On Sun, 24 Mar 2013, Joel Sherrill wrote:
> The only discrepancy in the make check summary was that
> WITH Mike's patch cris-elf with --enable-sim-hardware has this:
>
> === sim Summary ===
> # of expected passes 304
> # of untested testcases 136
>
> In the 7.5.91 log, the 136 was 170.


The last time HEAD built, without --enable-sim-hardware there
was:

# of expected passes            304
# of untested testcases         170

(no C compiler installed, hence the untested cases.)

> I know absolutely nothing about cris and only ran this because
> the configure file changed after running autoconf. Nothing specific
> to the cris was changed. This could be something else going on.

I guess so.  Nothing special about CRIS.  Grep is your friend:

src/sim $ grep SIM_AC_OPTION_HARDWARE */configure.ac

bfin/configure.ac:SIM_AC_OPTION_HARDWARE(yes,,\
cris/configure.ac:SIM_AC_OPTION_HARDWARE(no,,rv cris)
frv/configure.ac:SIM_AC_OPTION_HARDWARE(always,"","")
iq2000/configure.ac:SIM_AC_OPTION_HARDWARE(always,"","")
lm32/configure.ac:SIM_AC_OPTION_HARDWARE(yes,,lm32cpu lm32timer lm32uart)
m32r/configure.ac:SIM_AC_OPTION_HARDWARE(always,"","")
m68hc11/configure.ac:SIM_AC_OPTION_HARDWARE($hw_enabled,$hw_devices,$hw_extra_devices)
mips/configure.ac:SIM_AC_OPTION_HARDWARE($hw_enabled,$hw_devices,$hw_extra_devices)
mn10300/configure.ac:SIM_AC_OPTION_HARDWARE(yes,,mn103cpu
mn103int mn103tim mn103ser mn103iop)
mn10300/configure.ac:SIM_AC_OPTION_HARDWARE(always,"","")
sh64/configure.ac:SIM_AC_OPTION_HARDWARE(always,"","")

That is, it's the only sim where the default config is "no": "no
hardware enabled by default".  If that number *decreases* due to
a patch supposed to *enable* things, the patch seems wrong.

> As far as I can tell, Mike's patch is an improvement over mine
> and what was in 7.5.91. If he regenerates the configure scripts
> that need it, I would be perfectly happy with his patch being
> applied.
>
> This is the set of files which were touched by his patch or
> regenerated in my tree:
>
> #    modified:   sim/bfin/configure
> #    modified:   sim/common/acinclude.m4
> #    modified:   sim/cris/configure
> #    modified:   sim/frv/configure
> #    modified:   sim/iq2000/configure
> #    modified:   sim/lm32/configure
> #    modified:   sim/m32r/configure
> #    modified:   sim/m68hc11/configure
> #    modified:   sim/mips/configure
> #    modified:   sim/mn10300/configure
> #    modified:   sim/sh64/configure
>
> Mike, can you commit yours? Please. :)

First, please look into why the above number changes.

brgds, H-P

Re: Simulator dv-sockser.o patch testing

[Hans-Peter Nilsson]

On Mon, 25 Mar 2013, Hans-Peter Nilsson wrote:
> That is, it's the only sim where the default config is "no": "no
> hardware enabled by default".  If that number *decreases* due to
> a patch supposed to *enable* things, the patch seems wrong.

Oops: it's the number of *untested* cases that decrease when
enabling default-disabled hardware (yes, I do test and account
for that in the test-suite)... *blush*

So ignore my previous post while I remove my foot from my mouth.

brgds, H-P

Re: Simulator dv-sockser.o patch testing

[Joel Sherrill]

On 3/25/2013 10:22 PM, Hans-Peter Nilsson wrote:
> On Mon, 25 Mar 2013, Hans-Peter Nilsson wrote:
>> That is, it's the only sim where the default config is "no": "no
>> hardware enabled by default".  If that number *decreases* due to
>> a patch supposed to *enable* things, the patch seems wrong.
> Oops: it's the number of *untested* cases that decrease when
> enabling default-disabled hardware (yes, I do test and account
> for that in the test-suite)... *blush*
>
> So ignore my previous post while I remove my foot from my mouth.
It is really complicated and a lot of details. No apology needed. :)

Does Mike's patch look ok based upon my testing?

If so, he has some committing to do. :)
> brgds, H-P


-- 
Joel Sherrill, Ph.D.             Director of Research & Development
joel.sherrill@OARcorp.com        On-Line Applications Research
Ask me about RTEMS: a free RTOS  Huntsville AL 35805
Support Available                (256) 722-9985

Re: Simulator dv-sockser.o patch testing

[Hans-Peter Nilsson]

On Tue, 26 Mar 2013, Joel Sherrill wrote:
> On 3/25/2013 10:22 PM, Hans-Peter Nilsson wrote:
> > On Mon, 25 Mar 2013, Hans-Peter Nilsson wrote:
> > > That is, it's the only sim where the default config is "no": "no
> > > hardware enabled by default".  If that number *decreases* due to
> > > a patch supposed to *enable* things, the patch seems wrong.
> > Oops: it's the number of *untested* cases that decrease when
> > enabling default-disabled hardware (yes, I do test and account
> > for that in the test-suite)... *blush*
> >
> > So ignore my previous post while I remove my foot from my mouth.
> It is really complicated and a lot of details. No apology needed. :)
>
> Does Mike's patch look ok based upon my testing?

Dunno, there was a bit of churn there and it wasn't clear what
the intent was, but as Mike has the required rights no review is
required AFAIK.

brgds, H-P