[Date Prev][Date Next] [Thread Prev][Thread Next]
Re: [libvirt] [PATCH] nwfilter: intantiate filters on loading driver
- From: John Ferlan <jferlan redhat com>
- To: Nikolay Shirokovskiy <nshirokovskiy virtuozzo com>, "libvir-list redhat com" <libvir-list redhat com>
- Subject: Re: [libvirt] [PATCH] nwfilter: intantiate filters on loading driver
- Date: Wed, 31 Oct 2018 20:48:42 -0400
On 10/31/18 10:41 AM, Nikolay Shirokovskiy wrote:
> On 31.10.2018 16:14, John Ferlan wrote:
>> On 10/30/18 3:24 AM, Nikolay Shirokovskiy wrote:
>>> On 29.10.2018 22:37, John Ferlan wrote:
>>>> On 10/15/18 4:26 AM, Nikolay Shirokovskiy wrote:
>>>>> Before using filters binding filters instantiation was done by hypervisors
>>>>> drivers initialization code (qemu was the only such hypervisor). Now qemu
>>>>> reconnection done supposes it should be done by nwfilter driver probably.
>>>>> Let's add this missing step.
>>>>> Signed-off-by: Nikolay Shirokovskiy <nshirokovskiy virtuozzo com>
>>>>> src/nwfilter/nwfilter_driver.c | 3 +++
>>>>> 1 file changed, 3 insertions(+)
>>>> If there's research you've done where the instantiation was done before
>>>> introduction of the nwfilter bindings that would be really helpful...
>>>> I found that virNWFilterBuildAll was introduced in commit 3df907bfff.
>>>> There were 2 callers for it:
>>>> 1. virNWFilterTriggerRebuildImpl
>>>> 2. nwfilterStateReload
>>>> The former called as part of the virNWFilterConfLayerInit callback
>>>> during nwfilterStateInitialize (about 50 lines earlier).
>> First off let me say you certainly find a lot of interesting bugs/issues
>> that are complex and that's good. Often times I wonder how you trip
>> across things or how to quantify what you've found. Some are simpler
>> than others. This one on the surface would seem to be simple, but I keep
>> wondering is there a downside.
> The issue is related to my recent posts:
>  [RFC] Faster libvirtd restart with nwfilter rules
> which continues here:
> In short if there is lots of VMs with filters then libvirtd restart takes a lot of time
> during which libvirtd is unresponsive. By the way the issue is found in libvirt
> versions 3.9.0 and ealier (don't know of newer versions, Virtuozzo is based on 3.9.0 now,
> just like Centos7 :) )
So many different issues - trying to focus on just the one for this
> And attempts to fix it:
>  [PATCH RFC 0/4] nwfilter: don't reinstantiate filters if they are not changed
>  [PATCH v2 0/2] nwfilter: don't reinstantiate rules if there is no need to
> And the reason I started v2 is Laine mentioned that it is important to
> reinstantiate rules on restart (v1 do not care if somebody messed tables).
I've seen the patches and even read some briefly, but still need to take
this particular patch as separate from those.
> And I discovered that upstream version stops reinstantiating rules at all
> after introducing filters bindings. And this functionality is old:
So the key is "when" did that happen? That is can you pinpoint or
bisect a time in history where the filters were instantiated? And by
instantiated what exactly (call) are you referencing?
> commit cf6f8b9a9720fe5323a84e690de9fbf8ba41f6ac
> Author: Stefan Berger <stefanb us ibm com>
> Date: Mon Aug 16 12:59:54 2010 -0400
> nwfilter: extend nwfilter reload support
Yes, nwfilter is old, brittle, and in need of attention. Not sure anyone
really wants to take on the task though! I just realized I never got the
common object code implemented there. Mostly because of lock issues.
Suffice to say there's more "interesting" changes in the nwfilter space
being discussed internally.
>> The nwfilter processing is kindly said rather strange, complex, and to a
>> degree fragile. Thus when patches come along they are met with greater
>> scrutiny. From just reading the commit message here I don't get a sense
>> for the problem and why the solution fixes it. So I'm left to try and
>> research myself and ask a lot of questions.
>> BTW, some of the detail provided in this response is really useful as
>> either part of a cover or after the --- in the single patch. I would
>> think you'd "know" when you've done lots of research into a problem that
>> providing reviews more than a mere hint would be useful! For nwfilter
>> bindings, it's hard to imagine this is one of those - it just happened
>> type events. There seems to be a very specific sequence that's missing
>> from the commit description.
>>> Right but virNWFilterTriggerRebuildImpl is not actually called, it
>>> is set as rebuild callback. This rebuild callback can be called in
>>> virNWFilterObjTestUnassignDef and virNWFilterObjListAssignDef.
>> Ah yes, the virNWFilterConfLayerInit sets up the context to call the
>> virNWFilterTriggerRebuildImpl for virNWFilterObjTestUnassignDef and
>> virNWFilterObjListAssignDef and no I see it doesn't directly call the
>> Still, I don't see where the processing of a rebuild callback is
>> different than prior to commit 3df907bfff - at least with respect to the
>> two places which would call it.
> Right processing did not changed, I just wanted to show that
> virNWFilterBuildAll is not called now and before on nwfilter init. By the
> way 3df907bfff introduced virNWFilterBuildAll but not its functionality.
So prior to that commit made during v4.5.0 development, were the filters
instantiated? If so, then what changed caused them to no longer be?
>>> The former is not called during nwfilter driver init. The latter
>>> is called as part of virNWFilterObjListLoadAllConfigs but rebuild
>>> callback is never called on this path because the list is empty
>> Which list? nwfilters? nwfilter-bindings?
> nwfilters. And again by the way this is a bit vague statement. The
> matter is not list is not empty but we won't find same name in it
> when virNWFilterObjListAssignDef is called during driver init
> (which is said in below paragraph part in other words too)
If I restart libvirtd I see the same nwfilter list being generated. If I
have a domain running I see the filter binding for that domain. What's
When a domain isn't started, it finds the nwfilter and "loads" it into
nwfilter binding. That rule is applied somewhere, right? That rule
remains in effect for the domain regardless of how many libvirtd
restarts there are doesn't it? It seems as if you're saying we need to
reapply the filter on libvirtd restart.
>>> (callback is called only on updates when filter with same name
>>> is already present in the list). So virNWFilterBuildAll is
>>> not called on nwfilter driver init.
>> And similar logic was run was before commit 3df907bfff? This direct
>> correlation is the part I'm missing. What follows is my understand of
>> the before and after - some of the before is a bit of hand waving.
>> Perhaps Daniel can chime in and "fix up" inaccuracies.
>> Prior to commit 3df907bfff, the virNWFilterDomainFWUpdateCB was only run
>> when the domain was active. IOW: The domain would need to preload the
>> bindings in "hidden" locations such that the various vnetX (or whatever
>> target dev for the <interface>) devices would load the whichever
>> <filterref> was assocated. When libvirtd was restarted the processing
>> was similar and the "magic" of reinstantiation was provided through
>> virNWFilterRegisterCallbackDriver in (for example) qemuStateInitialize.
>> That I believe is the part you describe in your commit as "Before using
>> filters binding filters instantiation was done by hypervisors drivers
>> initialization code (qemu was the only such hypervisor)." - although I
>> see LXC and UML drivers changed as well, so I could be wrong with my
>> assumption of what you meant.
> Not quite correct. virNWFilterRegisterCallbackDriver is used when
> filter definition is updated and we need to reinstantiate filter for
> every active domain that uses it. But libvirtd restart is different,
> in this case reinstantiation occured only in qemu driver in
> reconnection process, the driver of reinstantiation was hypervisor.
>> After that commit, rather than requiring qemuStateInitialize to register
>> a callback driver which somehow magically loaded the filters for running
>> guests, the nwfilter-bindings for running guests are loaded via (for
>> example) /var/run/libvirt/vnet0.xml file processing during
>> virNWFilterBindingObjListLoadAllConfigs (change in the previous commit
>> 57f5621f46). So rather than the rebuild processing magically occurring
>> in the background by some hypervisor driver performing the rebuild
>> callback processing. Since virRegisterStateDriver (and friends) for
>> nwfilter are run before qemu, IIUC that means the filter bindings would
>> be loaded already. It's all a complicated dance.
> Yes loaded, but not reinstatiated.
What call is being done in the normal path that isn't being done in this
libvirtd restart path? Whatever the STEP_APPLY_CURRENT does? Or the
virNWFilterInstantiateFilter from nwfilterBindingCreateXML? If the
latter, then rather than using BuildAll, why not InstantiateFilter from
I keep coming back to once things are instantiated from the initial
libvirtd start why is it that the libvirtd restart processing needs to
do make the same call?
>> So in this after model for running guests it seems to me that the exact
>> same processing occurs. Now if someone during libvirtd's stop period
>> does something "outside the scope" of libvirt to change things that
>> libvirt wouldn't otherwise be notified about, then all bets are off.
>> Similar for other pieces of the code such as CPU's, Memory, Storage,
>> etc.; however, for those there is a query at reinitialization time that
>> can help reconcile differences due to perhaps missed events from qemu
>> because libvirtd wasn't processing. Not sure there's something similar
>> to query for nwfilter and bindings, but I assume there wouldn't have
>> been any before either.
>>>> So how does calling this now w/ @false help things during the state
>>>> initialize processing?
>>> Before filters bindings nwfilter driver only loads filters on it's
>>> init function. Then qemu driver for example on reconnection called:
>>> and filter rules gets [re]instantiated.
>> From commit f14c37ce4c2
>>> Now virNWFilterInstantiateFilter returns without actual instantiating
>>> because virNWFilterBindingLookupByPortDev finds binding which is loaded
>>> on nwfilter driver initialization>
>>> The consequences is that if somebody cleans rules between libvirtd stop
>>> and start then rules won't get instantiated again.
>> and this is the "key point" you are trying to reconcile, true?
>>> The fix is to [re]instantiate bindings in nwfilter driver init right
>>> after binding and filters are loaded. With @false argument virNWFilterBuildAll
>>> call virNWFilterInstantiateFilter for each binding - just what we need
>>> to. @true is used by virNWFilterObjListAssignDef/virNWFilterObjTestUnassignDef
>>> to use @newDef of object filter during instantiation etc.
>> Can you provided a concrete example showing your steps to help clear up
>> things for me? Using just one filter is fine. What does the guest look
>> like before whatever it is you do is done? What steps do you take? Then
>> what does it look like afterwards? What would you expect? If you did the
>> same/similar steps prior to the referenced commits what was the result?
> If you mean a more concrete usecase when such a clearing of firewall tables
> occurs when libvirtd is stopped - I don't have one. I just restore old
> functionality and also motivated by Laine comment to my first patch series.
> Or may be I don't understand you.
Well having a use case would certainly help. It's unclear what the
libvirtd restart processing isn't doing now that it may have been doing
before nwfilter bindings were added. You keep saying instantiate and I
keep asking what was done before that isn't being done now. Then you're
throwing in a few other patches and I'm lost.
> Thanx for a review.
>> As an "aside", it's been noted somewhere admins should not be messing
>> with nwfilter bindings. If someone "cleans rules" during a period when
>> libvirtd is stopped, then what is "expected" to happen afterwards.
>>>>> diff --git a/src/nwfilter/nwfilter_driver.c b/src/nwfilter/nwfilter_driver.c
>>>>> index 1ee5162..1ab906f 100644
>>>>> --- a/src/nwfilter/nwfilter_driver.c
>>>>> +++ b/src/nwfilter/nwfilter_driver.c
>>>>> @@ -264,6 +264,9 @@ nwfilterStateInitialize(bool privileged,
>>>>> if (virNWFilterBindingObjListLoadAllConfigs(driver->bindings, driver->bindingDir) < 0)
>>>>> goto error;
>>>>> + if (virNWFilterBuildAll(driver, false) < 0)
>>>>> + goto error;
>>>>> return 0;
[Date Prev][Date Next] [Thread Prev][Thread Next]