On 06.12, Andrea Arcangeli wrote:
> This version has some experimental change to the blkdev layer (latency
> fixes from Chris and Nick too plus the backout of the rc6 latency change
> to see if we can fix it w/o generating overscheduling, especially
> because it doesn't sound the right fix), so I would recommend some
> beating before doing anything critical with it. I would expect it as
> worse to deadlock with some task in D state. It worked fine for me so
> far but I didn't run big stress yet. In theory it should be better, but
> I just wanted to give a warning until it is better tested ;).
In sg.c, gcc shouts on:
shp->unique_id, shp->host_busy, shp->cmd_per_lun,
shp->host_busy should be accessed through atomic_read(), I think.
-- J.A. Magallon <email@example.com> \ Software is like sex: werewolf.able.es \ It's better when it's free Mandrake Linux release 9.2 (Cooker) for i586 Linux 2.4.21-jam1 (gcc 3.3 (Mandrake Linux 9.2 3.3-2mdk)) - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to firstname.lastname@example.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/