From mboxrd@z Thu Jan  1 00:00:00 1970
From: Andrew Cooper <andrew.cooper3@citrix.com>
Subject: Re: Xen-unstable: Regression (on AMD) host boot stuck
 on (XEN) testing the IO APIC.......................
Date: Tue, 26 Aug 2014 10:32:10 +0100
Message-ID: <53FC541A.9030505@citrix.com>
References: <367097038.20140825191202@eikelenboom.it>
	<53FB783D.5040302@citrix.com>
	<1283801136.20140825202522@eikelenboom.it>
	<53FC54A6020000780002D776@mail.emea.novell.com>
	<53FC572D020000780002D781@mail.emea.novell.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Return-path: <xen-devel-bounces@lists.xen.org>
Received: from mail6.bemta3.messagelabs.com ([195.245.230.39])
	by lists.xen.org with esmtp (Exim 4.72)
	(envelope-from <Andrew.Cooper3@citrix.com>) id 1XMD6t-0004rK-BC
	for xen-devel@lists.xenproject.org; Tue, 26 Aug 2014 09:32:15 +0000
In-Reply-To: <53FC572D020000780002D781@mail.emea.novell.com>
List-Unsubscribe: <http://lists.xen.org/cgi-bin/mailman/options/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=unsubscribe>
List-Post: <mailto:xen-devel@lists.xen.org>
List-Help: <mailto:xen-devel-request@lists.xen.org?subject=help>
List-Subscribe: <http://lists.xen.org/cgi-bin/mailman/listinfo/xen-devel>,
	<mailto:xen-devel-request@lists.xen.org?subject=subscribe>
Sender: xen-devel-bounces@lists.xen.org
Errors-To: xen-devel-bounces@lists.xen.org
To: Jan Beulich <JBeulich@suse.com>, Sander Eikelenboom <linux@eikelenboom.it>
Cc: xen-devel <xen-devel@lists.xenproject.org>
List-Id: xen-devel@lists.xenproject.org

On 26/08/14 08:45, Jan Beulich wrote:
>>>> On 26.08.14 at 09:34, <JBeulich@suse.com> wrote:
>> Considering how early this is, the only theory I have is that
>> rcu_pending() may end up constantly returning true, thus always
>> causing RCU_SOFTIRQ to get raised, not allowing the loop in
>> __do_softirq() to ever exit. If that's the case, we seem to have
>> two options:
>> - pass into __print_IO_APIC() whether it's being called from a key
>>   handler
>> - only call process_pending_softirqs() when system_state >=
>>   SYS_STATE_smp_boot
>>
>> Personally I'd prefer the former option.
> So here's a patch implementing that (the issue clearly isn't limited
> to AMD).
>
> Jan
>
> x86/IO-APIC: don't process softirqs during early boot
>
> Commit e13b320399 ("x86/irq: process softirqs in irq keyhandlers")
> made this unconditional, but the boot time use of __print_IO_APIC()
> (when "apic_verbosity=debug" was given) can't tolerate that.
>
> Reported-by: Sander Eikelenboom <linux@eikelenboom.it>
> Signed-off-by: Jan Beulich <jbeulich@suse.com>

Reviewed-by: Andrew Cooper <andrew.cooper3@citrix.com>

>
> --- a/xen/arch/x86/io_apic.c
> +++ b/xen/arch/x86/io_apic.c
> @@ -1092,7 +1092,7 @@ static inline void UNEXPECTED_IO_APIC(vo
>  {
>  }
>  
> -static void /*__init*/ __print_IO_APIC(void)
> +static void /*__init*/ __print_IO_APIC(bool_t boot)
>  {
>      int apic, i;
>      union IO_APIC_reg_00 reg_00;
> @@ -1113,7 +1113,8 @@ static void /*__init*/ __print_IO_APIC(v
>      printk(KERN_INFO "testing the IO APIC.......................\n");
>  
>      for (apic = 0; apic < nr_ioapics; apic++) {
> -        process_pending_softirqs();
> +        if ( !boot )
> +            process_pending_softirqs();
>  
>          if (!nr_ioapic_entries[apic])
>              continue;
> @@ -1219,7 +1220,7 @@ static void /*__init*/ __print_IO_APIC(v
>      for (i = 0; i < nr_irqs_gsi; i++) {
>          struct irq_pin_list *entry = irq_2_pin + i;
>  
> -        if ( !(i & 0x1f) )
> +        if ( !boot && !(i & 0x1f) )
>              process_pending_softirqs();
>  
>          if (entry->pin < 0)
> @@ -1242,12 +1243,12 @@ static void /*__init*/ __print_IO_APIC(v
>  static void __init print_IO_APIC(void)
>  {
>      if (apic_verbosity != APIC_QUIET)
> -        __print_IO_APIC();
> +        __print_IO_APIC(1);
>  }
>  
>  static void _print_IO_APIC_keyhandler(unsigned char key)
>  {
> -    __print_IO_APIC();
> +    __print_IO_APIC(0);
>  }
>  static struct keyhandler print_IO_APIC_keyhandler = {
>      .diagnostic = 1,
>
>
>