Hello, I've pushed out my current batch of fixes for the compression code. It introduces async compression for writes and does a better job of filling in pages for compressed extents while reading it. The latest git trees should fix the bad page mapping errors and improve performance even for uncompressed writes. Checksumming is now done via an ordered work queue, which makes sure IO is sent down to the device in the same order the writes were started. On my 4 drive array, it pushes performance from 280MB/s 340MB/s for sequential writes. The same ordered work queues are used to keep ordering sane for the async compression code. -chris -- To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
