On Sat, Sep 25, 2010 at 10:01 AM, Christoph Hellwig <hch@xxxxxxxxxxxxx> wrote: > On Sat, Sep 25, 2010 at 09:54:46AM -0700, Richard Sharpe wrote: >> fileio uses submit_bio to submit the data and it defaults to >> WRITE_THROUGH, NV_CACHE and DIRECT_IO (at least in the trunk, but I >> suspect this has been the case for a long while) however, the person >> making the complaint might have switched off WRITE_THROUGH in the >> pursuit of performance, in which case a crash could corrupt things >> badly but it would depend on whether or not clearing WRITE_THROUGH >> also clears NV_CACHE and what the code assembling the caching mode >> page does (and I have only had a cursory glance at the vdisk code). > > If the target uses submit_bio for logical files inside a filesystems > there are hundreds of ways to get exactly the corruption that Slawomir > sees. How does it obtain the logical to physical mapping? What locking > does it use agains other access to the file? Actually, I was wrong. for fileio it does vfs_writev. I started wondering the same questions and went back through the code and noticed that I had missed a test. We really need to know what parameters the person making the complaint is using, and the version of SCST. In addition, I don't understand when this corruption occurred ... It really might be a case of don't do that. -- Regards, Richard Sharpe _______________________________________________ xfs mailing list xfs@xxxxxxxxxxx http://oss.sgi.com/mailman/listinfo/xfs
