http://people.sistina.com/~thornber/dm/patches/2.6-unstable/2.6.0/2.6.0-udm6.tar.bz2 I'm about to release a 2.6.2-udm1 patchset, but the patches in that tree have been refactored a lot (24 patches rather than nearly 100). So I thought I ought make a final 2.6.0 release so people can see the final few patches. I've made a lot of changes to the multipath target in the last week, so I wouldn't be surprised if it's taken a backwards step in stability. Changes since 2.6.0-udm5 ------------------------ Revision 79: Get event notification working in multipath. Revision 80: Add in the qlogic drivers for 2.6.0 from http://sourceforge.net/projects/linux-qla2xxx/ Revision 81: I just noticed that bio_clone copies the BIO_SEG_VALID bit from the original bio when it was set. When we modify bi_idx or bi_vcnt afterwards the segment counts are invalid and the bit must be dropped (though it is fairly unlikely that it has already been set). [Christophe Saout] Revision 82: Copy the restrictions from the request queue into a temporary local variable and use combine_restrictions_low() instead of duplicating the code. [Christoph Saout] Revision 83: Move the sync count from the mirror set to the log. Newly synced regions weren't being marked clean. [AJ Lewis] Revision 84: Lift min_not_zero to dm.h Revision 85: Change dm-mpath.c:do_work() return a wakeup hint. Revision 86: path->test_sector is unused. Remove it. [Kevin Corry] Revision 87: Remove unnecessary log messages. [Kevin Corry] Revision 88: dm-ioctl-v4.c: No need to check the parameters for DM_LIST_VERSIONS_CMD Revision 89: drop HZ back to 100 (for qemu) Revision 90: A huge patch, this started out as Kevin Corrys patch to merge the valid/invalid path lists. But I kept seeing things I wanted to change. Use with care. Revision 91: Defer erroring an io until we know that all the paths have been tested *since* the last path failed. Large patch because some functions moved around.