Spurious testsuite failures on multi-cpu sparc-solaris2.9

Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Kaveh R. Ghazi]

I'm trying to test GCC on a 4 cpu sparc-sun-solaris2.9 box using
expect-5.42.1, tcl8.4.7 and dejagnu-1.4.4.  I'm getting lots of
spurious testsuite failures probably related to the solaris kernel bug
mentioned here: http://gcc.gnu.org/ml/gcc/2004-04/msg01480.html

Does any one know if bug 4927647 is addressed in any patch from Sun
yet?  Is the expect patch mentioned here helpful at all?
http://kegel.com/crosstool/current/patches/expect-5.39/pr12096.patch

Any help would be appreciated, my GCC testsuite results are useless at
the moment.  I'd really like to take full advantage of this box.

		Thanks,
		--Kaveh

PS: I'm not a Sun customer, so I can't pester them to fix anything.  I
just want to know if there's a patch yet.  Maybe I can cajole my
sysadmin into getting it installed.

--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu

Re: Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Dan Kegel]

Kaveh R. Ghazi wrote:
> I'm trying to test GCC on a 4 cpu sparc-sun-solaris2.9 box using
> expect-5.42.1, tcl8.4.7 and dejagnu-1.4.4.  I'm getting lots of
> spurious testsuite failures probably related to the solaris kernel bug
> mentioned here: http://gcc.gnu.org/ml/gcc/2004-04/msg01480.html
> 
> Does any one know if bug 4927647 is addressed in any patch from Sun
> yet?  Is the expect patch mentioned here helpful at all?
> http://kegel.com/crosstool/current/patches/expect-5.39/pr12096.patch

Go ahead and try the expect patch, it's got some chance of fixing
your problem.  If that doesn't do it, I guess it's the Solaris bug...

-- 
My technical stuff: http://kegel.com
My politics: see http://www.misleader.org for examples of why I'm for regime change

Re: Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Christian Ehrhardt]

On Tue, Sep 07, 2004 at 09:28:00PM -0400, Kaveh R. Ghazi wrote:
> I'm trying to test GCC on a 4 cpu sparc-sun-solaris2.9 box using
> expect-5.42.1, tcl8.4.7 and dejagnu-1.4.4.  I'm getting lots of
> spurious testsuite failures probably related to the solaris kernel bug
> mentioned here: http://gcc.gnu.org/ml/gcc/2004-04/msg01480.html
> 
> Does any one know if bug 4927647 is addressed in any patch from Sun
> yet?

According to SUN a fix has been integrated into Sol10 and a patch for
Sol9 will be release RSN. We were provided with a T-patch (raw binaries
not an officially released patch) in response to my bug report.

> Is the expect patch mentioned here helpful at all?
> http://kegel.com/crosstool/current/patches/expect-5.39/pr12096.patch

It didn't fix things for me, but it doesn't hurt either.

    regards   Christian

-- 
THAT'S ALL FOLKS!

Re: Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Kaveh R. Ghazi]

 > On Tue, Sep 07, 2004 at 09:28:00PM -0400, Kaveh R. Ghazi wrote:
 > > I'm trying to test GCC on a 4 cpu sparc-sun-solaris2.9 box using
 > > expect-5.42.1, tcl8.4.7 and dejagnu-1.4.4.  I'm getting lots of
 > > spurious testsuite failures probably related to the solaris
 > > kernel bug mentioned here:
 > > http://gcc.gnu.org/ml/gcc/2004-04/msg01480.html
 > > 
 > > Does any one know if bug 4927647 is addressed in any patch from Sun
 > > yet?
 > 
 > According to SUN a fix has been integrated into Sol10 and a patch for
 > Sol9 will be release RSN. We were provided with a T-patch (raw
 > binaries not an officially released patch) in response to my bug
 > report.

I'm curious about whether this patch will be part of the "recommended"
bundle, and thus available to those without a sunsolve account.  It'll
likely be part of the "recommended" set if the fix is located in files
already in that bundle.  Are you able to figure that out?  (Or simply
list the files and I'll try and determine it.)

At one point I worked for a company who received a "T" patch and I was
able to coax sun into releasing an official patch sooner.  I did this
at the time because I felt that the more people using the code (more
than just me!), the more likely it didn't have some other horrible
problem. :-)

Do you think you could prod sun into releasing this officially more
quickly?  I know that I personally would be able to be more productive
if I could run the testsuite in parallel and post results more
quickly.


 > > Is the expect patch mentioned here helpful at all?
 > > http://kegel.com/crosstool/current/patches/expect-5.39/pr12096.patch
 > 
 > It didn't fix things for me, but it doesn't hurt either.
 >     regards   Christian

That patch didn't fix the problem for me either.

		Thanks,
		--Kaveh
--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu

Re: Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Christian Ehrhardt]

On Sat, Sep 18, 2004 at 07:19:09PM -0400, Kaveh R. Ghazi wrote:
>  > > Does any one know if bug 4927647 is addressed in any patch from Sun
>  > > yet?
> 
> I'm curious about whether this patch will be part of the "recommended"
> bundle, and thus available to those without a sunsolve account.

Patch 117171-11 for Sol9 is publicly availiable. This patch should fix
Sun-Bug 4927647 and the spurious testsuite failures. If others can confirm
this, we should probably recommend the patch for those running the
testsuite.

    regards  Christian

-- 
THAT'S ALL FOLKS!

Re: Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Kaveh R. Ghazi]

 > Patch 117171-11 for Sol9 is publicly availiable. This patch should fix
 > Sun-Bug 4927647 and the spurious testsuite failures. If others can
 > confirm this, we should probably recommend the patch for those running
 > the testsuite.
 >     regards  Christian

Works just great for me.  However, I also have the expect patch
installed.  What I'd like to determine is whether both are needed or
simply just the solaris2.9 patch.  I'll try recompiling a clean expect
and see what happens.

Also, are previous version of solaris2 affected?  I only saw the patch
for 2.9.

		Thanks,
		--Kaveh
--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu

Re: Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Christian Ehrhardt]

On Mon, Oct 04, 2004 at 09:50:47AM -0400, Kaveh R. Ghazi wrote:
> Works just great for me.  However, I also have the expect patch
> installed.  What I'd like to determine is whether both are needed or
> simply just the solaris2.9 patch.  I'll try recompiling a clean expect
> and see what happens.

AFAIR the expect-Patch helps only if the testcase generates large amounts
of output. Most if not all of your Problems will go away with the
Solaris-Patch alone.

> Also, are previous version of solaris2 affected?  I only saw the patch
> for 2.9.

I have no idea about Solaris 7 or earlier, according to my tests, Solaris 8
is affected, however, a Patch is not yet publicly availiable.

    regards  Christian

-- 
THAT'S ALL FOLKS!

Re: Spurious testsuite failures on multi-cpu sparc-solaris2.9

[Kaveh R. Ghazi]

 > > Also, are previous version of solaris2 affected?  I only saw the
 > > patch for 2.9.
 > 
 > I have no idea about Solaris 7 or earlier, according to my tests,
 > Solaris 8 is affected, however, a Patch is not yet publicly
 > availiable.
 >     regards  Christian

I've encountered the problem on solaris2.7:
http://gcc.gnu.org/ml/gcc/2004-03/msg00056.html

I hope Sun backports this kernel fix for earlier OSes.

		--Kaveh
--
Kaveh R. Ghazi			ghazi@caip.rutgers.edu