Hi all, Linked below is a very preliminary patch to enable SNIA DDF v1 support in dmraid. At least in theory, linear volumes, RAID 0, 1, 4 and 5 support should be working, though at the moment I've only managed to test RAID 0 and 1. I have three Adaptec SAS/SATA RAID controllers that write metadata in DDF format--aic94xx HostRAID, an Adaptec 2410SA SATA, and a ServeRAID 8i card. Note, however, that there are a few areas where these cards seem to go a bit out of spec; I've attempted to capture and make note of these areas in the code. Like last time with HostRAID support, my rudimentary tests involve first booting the Adaptec controller, stamping out partitions, filesystems and various data, and removing the controller. Then, I boot back up and see how far I get with mounting the array, fscking, and comparing the data to the originals. I've done this enough times without data loss to conclude that it's working all right, though I've yet to try out any truly thorny scenarios. As noted by the numerous FIXMEs in the patch, there's still a fair amount of work to do in terms of making the driver more robust. For example, I don't check the table CRCs at this point, stackable RAIDs are not supported, and it's theoretically possible (though the Adaptec controllers don't seem to allow this) to cut up a disk and put various parts in different arrays. I'm not even really sure how to express multiple disk to multiple array relations in dmraid, actually. :) Also, if anyone else has a SW RAID controller that uses DDF, your mileage may vary. Please tread carefully. 96KB patch: http://sweaglesw.net/~djwong/docs/ddf1.patch I also added CVT64 to byteorder.h to facilitate byte swapping of 64-bit values. Questions? Comments? Heinz, if you'd like metadata samples, I'll send you the (somewhat large) tarballs off-list. --Darrick
Attachment:
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Ataraid-list@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/ataraid-list