LKML Archive on lore.kernel.org
help / color / mirror / Atom feed
From: "Maciej S. Szmigiero" <mail@maciej.szmigiero.name>
To: Paul Menzel <pmenzel@molgen.mpg.de>,
	Sean Christopherson <seanjc@google.com>
Cc: kvm@vger.kernel.org, linux-kernel@vger.kernel.org,
	Paolo Bonzini <pbonzini@redhat.com>,
	Vitaly Kuznetsov <vkuznets@redhat.com>,
	Wanpeng Li <wanpengli@tencent.com>,
	Jim Mattson <jmattson@google.com>, Joerg Roedel <joro@8bytes.org>,
	Ingo Molnar <mingo@redhat.com>,
	Thomas Gleixner <tglx@linutronix.de>,
	Borislav Petkov <bp@alien8.de>,
	x86@kernel.org, "H. Peter Anvin" <hpa@zytor.com>
Subject: Re: [PATCH] x86: kvm: Demote level of already loaded message from error to info
Date: Thu, 19 Aug 2021 16:05:47 +0200	[thread overview]
Message-ID: <36208232-f936-9eed-22cf-88a61b294c7e@maciej.szmigiero.name> (raw)
In-Reply-To: <1c5ac4f8-4c39-a969-9ffa-2f527535a4b1@molgen.mpg.de>

On 19.08.2021 08:39, Paul Menzel wrote:
> Am 19.08.21 um 00:23 schrieb Sean Christopherson:
>> On Wed, Aug 18, 2021, Maciej S. Szmigiero wrote:
>>> On 18.08.2021 13:49, Paul Menzel wrote:
>>>> In scripts, running
>>>>
>>>>       modprobe kvm_amd     2>/dev/null
>>>>       modprobe kvm_intel   2>/dev/null
>>>>
>>>> to ensure the modules are loaded causes Linux to log errors.
>>>>
>>>>       $ dmesg --level=err
>>>>       [    0.641747] [Firmware Bug]: TSC_DEADLINE disabled due to Errata; please update microcode to version: 0x3a (or later)
>>>>       [   40.196868] kvm: already loaded the other module
>>>>       [   40.219857] kvm: already loaded the other module
>>>>       [   55.501362] kvm [1177]: vcpu0, guest rIP: 0xffffffff96e5b644 disabled perfctr wrmsr: 0xc2 data 0xffff
>>>>       [   56.397974] kvm [1418]: vcpu0, guest rIP: 0xffffffff81046158 disabled perfctr wrmsr: 0xc1 data 0xabcd
>>>>       [1007981.827781] kvm: already loaded the other module
>>>>       [1008000.394089] kvm: already loaded the other module
>>>>       [1008030.706999] kvm: already loaded the other module
>>>>       [1020396.054470] kvm: already loaded the other module
>>>>       [1020405.614774] kvm: already loaded the other module
>>>>       [1020410.140069] kvm: already loaded the other module
>>>>       [1020704.049231] kvm: already loaded the other module
>>>>
>>>> As one of the two KVM modules is already loaded, KVM is functioning, and
>>>> their is no error condition. Therefore, demote the log message level to
>>>> informational.
>>
>> Hrm, but there is an error condition.  Userspace explicitly requested something
>> and KVM couldn't satisfy the request.
> 
> Yes, that’s the other perspective. ;-) I’d argue, as the Intel/AMD module can’t work on AMD/Intel, the load failure is expected and error. But as “error condition” is not well defined:
> 
>      $ dmesg -h
>      […]
>      Supported log levels (priorities):
>         emerg - system is unusable
>         alert - action must be taken immediately
>          crit - critical conditions
>           err - error conditions
>          warn - warning conditions
>        notice - normal but significant condition
>          info - informational
>         debug - debug-level messages
> 

Why is that script repeatably trying to load kvm_amd and kvm_intel
modules?
I would assume these would be loaded once at system boot time (either
manually or based on their modalias).
If your script absolutely has to load them manually, it could check first
whether /dev/kvm already exists.

>>> Shouldn't this return ENODEV when loading one of these modules instead
>>> as there is no hardware that supports both VMX and SVM?
>>
>> Probably not, as KVM would effectively be speculating, e.g. someone could load an
>> out-of-tree variant of kvm_{intel,amd}.  Maybe instead of switching to ENODEV,
>> reword the comment, make it ratelimited, and shove it down?  That way the message
>> and -EEXIST fires iff the vendor module actually has some chance of being loaded.
>>
>>  From 3528e66bd5107d5ac4f6a6ae50503cf64446866a Mon Sep 17 00:00:00 2001
>> From: Sean Christopherson <seanjc@google.com>
>> Date: Wed, 18 Aug 2021 15:17:43 -0700
>> Subject: [PATCH] KVM: x86: Tweak handling and message when vendor module is
>>   already loaded
>>
>> Reword KVM's error message if a vendor module is already loaded to state
>> exactly that instead of assuming "the other" module is loaded,
> 
> The rewording is definitely an improvement.
> 
>> ratelimit
>> said message to match the other errors, and move the check down below the
>> basic functionality checks so that attempting to load an unsupported
>> module provides the same result regardless of whether or not a supported
>> vendor module is already loaded.
> 
> Maybe add an example, how it would log the error before, and how it’s done now.
> 
>> Reported-by: Paul Menzel <pmenzel@molgen.mpg.de>
>> Cc: Maciej S. Szmigiero <mail@maciej.szmigiero.name>
>> Signed-off-by: Sean Christopherson <seanjc@google.com>
>> ---
>>   arch/x86/kvm/x86.c | 12 ++++++------
>>   1 file changed, 6 insertions(+), 6 deletions(-)
>>
>> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
>> index fdc0c18339fb..15bd4bd3c81d 100644
>> --- a/arch/x86/kvm/x86.c
>> +++ b/arch/x86/kvm/x86.c
>> @@ -8357,12 +8357,6 @@ int kvm_arch_init(void *opaque)
>>       struct kvm_x86_init_ops *ops = opaque;
>>       int r;
>>
>> -    if (kvm_x86_ops.hardware_enable) {
>> -        printk(KERN_ERR "kvm: already loaded the other module\n");
>> -        r = -EEXIST;
>> -        goto out;
>> -    }
>> -
>>       if (!ops->cpu_has_kvm_support()) {
>>           pr_err_ratelimited("kvm: no hardware support\n");
>>           r = -EOPNOTSUPP;
>> @@ -8374,6 +8368,12 @@ int kvm_arch_init(void *opaque)
>>           goto out;
>>       }
>>
>> +    if (kvm_x86_ops.hardware_enable) {
>> +        pr_err_ratelimited("kvm: already loaded a vendor module\n");
>> +        r = -EEXIST;
>> +        goto out;
>> +    }
>> +
>>       /*
>>        * KVM explicitly assumes that the guest has an FPU and
>>        * FXSAVE/FXRSTOR. For example, the KVM_GET_FPU explicitly casts the
>> -- 
>> 2.33.0.rc2.250.ged5fa647cd-goog
> 
> Sounds also good at first sight. No idea, if monitoring scripts in userspace would get confused now.

This definitely looks more informative than the existing message.

It would be even better if it wasn't "kvm: no hardware support" (as this
message is technically incorrect), but something like
"kvm: no VMX hardware support" or "kvm: no SVM hardware support".

> Kind regards,
> 
> Paul

Thanks,
Maciej

      reply	other threads:[~2021-08-19 14:06 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2021-08-18 11:49 Paul Menzel
2021-08-18 13:36 ` Maciej S. Szmigiero
2021-08-18 22:23   ` Sean Christopherson
2021-08-19  6:39     ` Paul Menzel
2021-08-19 14:05       ` Maciej S. Szmigiero [this message]

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=36208232-f936-9eed-22cf-88a61b294c7e@maciej.szmigiero.name \
    --to=mail@maciej.szmigiero.name \
    --cc=bp@alien8.de \
    --cc=hpa@zytor.com \
    --cc=jmattson@google.com \
    --cc=joro@8bytes.org \
    --cc=kvm@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mingo@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=pmenzel@molgen.mpg.de \
    --cc=seanjc@google.com \
    --cc=tglx@linutronix.de \
    --cc=vkuznets@redhat.com \
    --cc=wanpengli@tencent.com \
    --cc=x86@kernel.org \
    --subject='Re: [PATCH] x86: kvm: Demote level of already loaded message from error to info' \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).