XFS a pad kernelu

[Jan Derfinak]

On Mon, 15 Aug 2005, Kasparek Tomas wrote:

> Ja skoncil na tomhle:
> 
> http://oss.sgi.com/archives/linux-xfs/2005-08/msg00016.html
> 
> resp. 
> http://oss.sgi.com/archives/linux-xfs/2005-07/msg00041.html
> 
> Konkretne jde o problemy XFS s velikosti zasobniku u hodne zatizenych
> serveru a slozitejsi strukture ovladacu (md,lvm,raid,libata,nfsd ...)

Este konkretnejsie ide o problem vrstvy blokovych zariadeni so zasobnikom.
Tento problem sa moze prejavit pri vsetkych suborovych systemoch (vratne
nfs), pokial je navrstvene dostatocne mnozsto blokovych vrstiev na sebe.
Vzhladom na zlozitost XFS sa u neho tato chyba vyskytne najskor, ale su
reportovane pripady s ext3. Skontrolujte si, ci nepouzivate 4k stack a
pokial ano tak ho zmente na 8k. Co sa tyka generickeho riesenie, tak som
zachytil tuto diskusiu medzi Andi Kleen a Jens Axboe:
--------------------------------------------------------------------------
> A generic way to solve this would be to let the block layer
> who calls into the various stacking layers check how much stack is left
> first  and when it is too low push the work to another thread using
> a workqueue.
> 
> Jens, do you think that would be feasible?

(sorry for the late reply, vacation)

Sounds like a possible solution for the problem. 4kb stack is never
going to be completely enough for some block layer stacking setups.
Would need some careful work, I don't want to see each and every io
pushed to a worker for processing and potentially incurring 2 context
switches per io.
--------------------------------------------------------------------------

Bohuzial neviem, ako daleko je tento navrh od realizacie.

jano

--