Re: [PATCH] seqlock: serialize against writers

[Gregory Haskins <gregory.haskins@gmail.com>]

[-- Attachment #1: Type: text/plain, Size: 1633 bytes --]

Steven Rostedt wrote:
> On Fri, 29 Aug 2008, Gregory Haskins wrote:
>
>   
>> Andi Kleen wrote:
>>     
>>>> Im running it on a x86_64 box as we speak.  How can I tell if there is a
>>>> certain mode that is permitting this?
>>>>     
>>>>         
>>> If the boot up says you're running with PMtimer then it uses the fallback
>>> (usually happens on pre Fam10h AMD boxes). A typical Intel box
>>> would use the faster ring 3 only TSC path and then explode with your
>>> change I bet. 
>>>
>>> Or step with gdb through gettimeofday() and see if it does a syscall.
>>>
>>> -Andi
>>>   
>>>       
>> It seems to be running fine with no indication it has fallen back. 
>> Perhaps I need a certain workload to bring out the issue?
>>     
>
> Perhaps you never hit the slow path in userland. That's the only place it 
> would write. Perhaps add a dummy static variable in the fast path, and 
> write to it. See if that crashes you apps.
>
> -- Steve
>   

Yeah, ideas crossed in the mail ;)

I could just force all of the seqbegins to hit the slowpath by hacking
the code and see what happens (aside from slowing down, of course ;)

Question: Which seqlock_t does userspace use?  I assume it uses
seqlock_t and not raw_seqlock_t.  But the only reason that I ask is that
I converted raw_seqlock_t to use the new style as well to be consistent,
even though it is not strictly necessary for the same reasons.  So if
perchance userspace uses the raw variant, I could solve this issue by
only re-working the seqlock_t variant.  Kind of a long shot, but figured
I would mention it :)

-Greg




[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 257 bytes --]