Re: xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU

[Don Slutz <dslutz@verizon.com>]

On 03/19/14 16:20, Konrad Rzeszutek Wilk wrote:
> On Wed, Mar 19, 2014 at 08:07:59PM +0000, Andrew Cooper wrote:
>> On 19/03/14 20:04, Konrad Rzeszutek Wilk wrote:
>>> On Wed, Mar 19, 2014 at 03:44:13PM -0400, Don Slutz wrote:
>>>> I just noticed these messages from a 4.5-unstable build (a03984e + xenctx changes):
>>>>
>>>> (XEN) Xen version 4.5-unstable (don@culpepper.cloudswitch.com) (gcc (GCC) 4.7.2 20120921 (Red Hat 4.7.2-2)) debug=y Tue Mar 18 14:00:04 EDT 2014
>>>> (XEN) Latest ChangeSet: Thu Dec 12 19:17:03 2013 +0000 git:a03984e
>>>> ...
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479270] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU1
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479328] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU2
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479346] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU3
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479366] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU4
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479381] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU5
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479398] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU6
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479411] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU7
>>>> Mar 19 15:31:16 dcs-xen-54 kernel: [   92.479427] xen-acpi-processor: (CX): Hypervisor error (-38) for ACPI CPU8
>>>>
>>>>
>>>> This related to no-cpuidle that I have specified.  Is this a bug or is it expected?
>>> Does this fix it for you (inline and attached)?

Thanks for the quick response.  My 1st attempt to build linux 
3.14.0-rc7+ does not boot under xen:


(XEN) mcheck_poll: Machine check polling timer started.
(XEN) *** LOADING DOMAIN 0 ***
(XEN) elf_parse_binary: phdr: paddr=0x1000000 memsz=0xada000
(XEN) elf_parse_binary: phdr: paddr=0x1c00000 memsz=0xfa0f0
(XEN) elf_parse_binary: phdr: paddr=0x1cfb000 memsz=0x13d00
(XEN) elf_parse_binary: phdr: paddr=0x1d0f000 memsz=0x2ed000
(XEN) elf_parse_binary: memory: 0x1000000 -> 0x1ffc000
(XEN) elf_xen_note_check: ERROR: Will only load images built for the 
generic loader or Linux images (Not '' and '')
(XEN)
(XEN) ****************************************
(XEN) Panic on CPU 0:
(XEN) Could not set up DOM0 guest OS
(XEN) ****************************************
(XEN)
(XEN) Reboot in five seconds...


Clearly what I did to build it was incorrect.  Will poke around and try 
to figure out what I have done wrong.

     -Don Slutz



>>>  From 3b2c8fd5e3a8722983f93e790d609d0d5b38054b Mon Sep 17 00:00:00 2001
>>> From: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
>>> Date: Wed, 19 Mar 2014 16:03:23 -0400
>>> Subject: [PATCH] xen-acpi-processor: Don't display errors when we get -ENOSYS
>>>
>>> which is a perfectly legal error. This can be triggered if the
>>> user has booted Xen with the no-cpuidle parameter.
>>>
>>> Reported-by: Don Slutz <dslutz@verizon.com>
>>> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
>> An ENOSYS should imply that dom0 shouldn't try setting any further CPUs
>> worth of information.
> The return value is still returned it is just not spamming the console.
>> ~Andrew
>>
>>> ---
>>>   drivers/xen/xen-acpi-processor.c | 4 ++--
>>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/drivers/xen/xen-acpi-processor.c b/drivers/xen/xen-acpi-processor.c
>>> index 82358d1..59fc190 100644
>>> --- a/drivers/xen/xen-acpi-processor.c
>>> +++ b/drivers/xen/xen-acpi-processor.c
>>> @@ -127,7 +127,7 @@ static int push_cxx_to_hypervisor(struct acpi_processor *_pr)
>>>   			pr_debug("     C%d: %s %d uS\n",
>>>   				 cx->type, cx->desc, (u32)cx->latency);
>>>   		}
>>> -	} else if (ret != -EINVAL)
>>> +	} else if ((ret != -EINVAL) && (ret != -ENOSYS))
>>>   		/* EINVAL means the ACPI ID is incorrect - meaning the ACPI
>>>   		 * table is referencing a non-existing CPU - which can happen
>>>   		 * with broken ACPI tables. */
>>> @@ -259,7 +259,7 @@ static int push_pxx_to_hypervisor(struct acpi_processor *_pr)
>>>   			(u32) perf->states[i].power,
>>>   			(u32) perf->states[i].transition_latency);
>>>   		}
>>> -	} else if (ret != -EINVAL)
>>> +	} else if ((ret != -EINVAL) && (ret != -ENOSYS))
>>>   		/* EINVAL means the ACPI ID is incorrect - meaning the ACPI
>>>   		 * table is referencing a non-existing CPU - which can happen
>>>   		 * with broken ACPI tables. */
>>>
>>>
>>> _______________________________________________
>>> Xen-devel mailing list
>>> Xen-devel@lists.xen.org
>>> http://lists.xen.org/xen-devel