From mboxrd@z Thu Jan  1 00:00:00 1970
From: Paolo Bonzini <pbonzini@redhat.com>
Subject: Re: [kvm-unit-tests PATCH 6/6] runtime: make the MAX_SMP loop a
 function
Date: Tue, 26 Jan 2016 13:38:29 +0100
Message-ID: <56A768C5.3070305@redhat.com>
References: <1453474709-10679-1-git-send-email-drjones@redhat.com>
 <1453474709-10679-7-git-send-email-drjones@redhat.com>
 <56A753C6.1020604@redhat.com> <20160126122845.GD10697@hawk.localdomain>
Mime-Version: 1.0
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Cc: kvm@vger.kernel.org, rkrcmar@redhat.com
To: Andrew Jones <drjones@redhat.com>
Return-path: <kvm-owner@vger.kernel.org>
Received: from mx1.redhat.com ([209.132.183.28]:39215 "EHLO mx1.redhat.com"
	rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
	id S965058AbcAZMie (ORCPT <rfc822;kvm@vger.kernel.org>);
	Tue, 26 Jan 2016 07:38:34 -0500
Received: from int-mx09.intmail.prod.int.phx2.redhat.com (int-mx09.intmail.prod.int.phx2.redhat.com [10.5.11.22])
	by mx1.redhat.com (Postfix) with ESMTPS id C47DF96B1
	for <kvm@vger.kernel.org>; Tue, 26 Jan 2016 12:38:34 +0000 (UTC)
In-Reply-To: <20160126122845.GD10697@hawk.localdomain>
Sender: kvm-owner@vger.kernel.org
List-ID: <kvm.vger.kernel.org>

 26/01/2016 13:28, Andrew Jones wrote:
> On Tue, Jan 26, 2016 at 12:08:54PM +0100, Paolo Bonzini wrote:
>>
>>
>> On 22/01/2016 15:58, Andrew Jones wrote:
>>> We can now source scripts/runtime.bash multiple times without
>>> always running the MAX_SMP loop. runtime.bash is now ready to
>>> be sourced by arch run scripts (when not running in standalone
>>> mode). Patches coming soon will add a couple of arch-neutral
>>> functions to the arch run scripts.
>>>
>>> Signed-off-by: Andrew Jones <drjones@redhat.com>
>>> ---
>>>  run_tests.sh            |  2 ++
>>>  scripts/mkstandalone.sh |  1 +
>>>  scripts/runtime.bash    | 12 +++++++++---
>>>  3 files changed, 12 insertions(+), 3 deletions(-)
>>>
>>> diff --git a/run_tests.sh b/run_tests.sh
>>> index 558b8e7431d8e..c0c5a72b1ceee 100755
>>> --- a/run_tests.sh
>>> +++ b/run_tests.sh
>>> @@ -28,6 +28,8 @@ EOF
>>>  RUNTIME_arch_run="./$TEST_DIR/run"
>>>  source scripts/runtime.bash
>>>  
>>> +probe_max_smp
>>> +
>>>  while getopts "g:hv" opt; do
>>>      case $opt in
>>>          g)
>>> diff --git a/scripts/mkstandalone.sh b/scripts/mkstandalone.sh
>>> index b0f1e7c098afb..c31a5f10519a1 100755
>>> --- a/scripts/mkstandalone.sh
>>> +++ b/scripts/mkstandalone.sh
>>> @@ -60,6 +60,7 @@ generate_test ()
>>>  
>>>  	cat scripts/runtime.bash
>>>  
>>> +	echo probe_max_smp
>>>  	echo "run ${args[@]}"
>>>  }
>>>  
>>> diff --git a/scripts/runtime.bash b/scripts/runtime.bash
>>> index 63d1b9653007b..dd700f24cffa3 100644
>>> --- a/scripts/runtime.bash
>>> +++ b/scripts/runtime.bash
>>> @@ -1,5 +1,9 @@
>>>  : "${RUNTIME_arch_run?}"
>>>  
>>> +if [ -z "$MAX_SMP" ]; then
>>> +    MAX_SMP=$(getconf _NPROCESSORS_CONF)
>>> +fi
>>> +
>>>  qemu=${QEMU:-qemu-system-$ARCH}
>>>  
>>>  function run()
>>> @@ -57,7 +61,6 @@ function run()
>>>      return $ret
>>>  }
>>>  
>>> -MAX_SMP=$(getconf _NPROCESSORS_CONF)
>>>  #
>>>  # Probe for MAX_SMP, in case it's less than the number of host cpus.
>>>  #
>>> @@ -67,7 +70,10 @@ MAX_SMP=$(getconf _NPROCESSORS_CONF)
>>>  # "arm/arm64: KVM: Remove 'config KVM_ARM_MAX_VCPUS'". So, at some
>>>  # point when maintaining the while loop gets too tiresome, we can
>>>  # just remove it...
>>> -while $RUNTIME_arch_run _NO_FILE_4Uhere_ -smp $MAX_SMP \
>>> +function probe_max_smp()
>>> +{
>>> +    while $RUNTIME_arch_run _NO_FILE_4Uhere_ -smp $MAX_SMP \
>>>  		|& grep -qi 'exceeds max CPUs'; do
>>>  	((--MAX_SMP))
>>> -done
>>> +    done
>>> +}
>>>
>>
>> This kinda goes against the idea of removing duplication between
>> runtests and mkstandalone...
>>
>> I wonder if you should instead remove the
>>
>>     : "${RUNTIME_arch_run?}"
>>
>> and conditionalize the MAX_SMP test on the existence of
>> RUNTIME_arch_run.  Is it set during execution of the arch run scripts?
> 
> Let's drop this patch. Originally I planned on adding more functions to
> scripts/runtime.bash, in my next series (coming soon), which would require
> the arch-run scripts to source scripts/runtime.bash. Now, I think it's
> cleaner to just create a new file to source, scripts/arch-run.bash.

Good.  Applying 5 then.

Paolo