Flavio Castelli:
> Unfortunately this does not solve the issue.

Does it mean you got those messages again?
"aufs_migratepage:666:mergesolv[17456]: This operation is not supported...."


> Moreover the debugging messages you asked me to add are not providing
> any useful message. As I told you in my previous mail I had to load the
> aufs module with debug=1 in order to see more messages inside of the
> kernel logs. That however slowed down the whole system. It would be
> definitely better to leave debug=0 when loading the kernel module.

I am afraid you are misunderstanding the debug message I want. What I
want is something like

aufs aufs_migratepage:XXX:xxx[XXX]: This operation is not supported. Please 
report this application to aufs-users ML.
Call stack:
        ...
        ...
f-1: ...
f0: ...
f1: ...
d-1: ...
d0: ...
d1: ...
i-1: ...
i0: ...
i1: ...

We may not get all these lines. It is OK. The point is to see how
"file", "dentry" and "inode" are when aufs_migratepage() is called.
The messages in aufs_readpage() were helpful a little, butit may not
help anymore.

The lots of messages you saw (when the module parameter debug is 1) like
h_d_revalidate:...
dump_opts:...
are not important.


> According to the output of numastats my machine has a single memory node:

Hmm...
On a single memory node system, I wonder why aufs_migrate() is called...


J. R. Okajima

------------------------------------------------------------------------------
October Webinars: Code for Performance
Free Intel webinars can help you accelerate application performance.
Explore tips for MPI, OpenMP, advanced profiling, and more. Get the most from 
the latest Intel processors and coprocessors. See abstracts and register >
http://pubads.g.doubleclick.net/gampad/clk?id=60133471&iu=/4140/ostg.clktrk

Reply via email to