F. Wessels wrote: >Hi, > >I've got a amd/ati sb700 sata/ide controller to which are connected an ide >disk to boot snv90 from and several sata disks for a zfs raidz pool. > >If i put the controller in the bios in ide mode I can boot into 64 bit mode >and everything works. But the sata disks run in ide mode and are accessed one >at a time. The total bandwidth doesn't exceed that of a single drive. The >getspeed script shows that no drive is limiting. > >If I but the controller in AHCI mode I can NOT boot in 64 bit mode. This is >described in bug 6665032 , all kinds of timeouts happen. But I can boot in 32 >bit mode and everything is fine, although I need run run ZTest and some other >test. At least all disks are used at the same time. But now I'm stuck in 32 >bit and my other apps don't run in 32 bit. The other workarounds suit me even >less, less than 4gb and no dma. (But I didn't try them) > >I'm aware of the doubtfull AHCI implementation in the sb600 and sb700. But >the sb700 runs fine under linux. It can do 64bit dma! The only issue >remaining, Im aware of, is the port multiplier support. But apart from that >everything working ok, under linux 2.6.25 with sufficient patches. > >Perhaps everything is solved in snv93 which takes care of bug 6665032 , but >that specifically takes care of the sb600 which can't do 64 bit dma. > > Hi Frederik,
Have you seen the packet timeout on your ATI SB700 with the 64-bit kernel mode? The current ahci driver only handled the 64-bit dma issue with ATI SB600. Thanks! >No action required, but I hope that the sb600 and sb700 will make it to the >HCL. > >Frederik. > > >This message posted from opensolaris.org >_______________________________________________ >storage-discuss mailing list >[email protected] >http://mail.opensolaris.org/mailman/listinfo/storage-discuss > _______________________________________________ storage-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/storage-discuss
