Re: stat benchmark

From: Zach Brown
Date: Mon Apr 28 2008 - 13:46:55 EST



> I guess it would be difficult to get close to the optimal disk schedule by
> using syslets; if a directory contains 1000 files that would require 1000
> syslets and a good I/O scheduler - that's unlikely to be feasible.

It wouldn't even get to the I/O scheduler. The VFS stat path (see
real_lookup()) is synchronous and serialized. Each stat will hold the
i_mutex of the parent directory while it waits for the file system's
lookup method to populate the inode from disk.

- z
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/