On Thu, Nov 08, 2012 at 06:21:23PM -0500, Theodore Ts'o wrote: > On Fri, Oct 26, 2012 at 09:23:39PM +0800, Zheng Liu wrote: > > + * 3. performance analysis > > + * -- overhead > > + * 1. Apart from operations on a delayed extent tree, we need to > > + * down_write(inode->i_data_sem) in delayed write path to maintain delayed > > + * extent tree, this can have impact on parallel read-write and write-write > > Hi Zheng, > > I can fix this up, before I finalize your commit, but I just want to > check. I believe this comment is out of date --- we are now using a > r/w spinlock, i_es_lock, yes? Since we never hold the spinlock for > very long, I would be surprised if this is going to be a scalability > bottleneck (too bad Eric doesn't have access to the big SMP machine > that he used to use to help us do our scalability testing, so we could > check to be sure). Hi Ted, Oops, it is my fault. Indeed it needs to be replaced with i_es_lock. I can do some tests in a server which has 16 cores, but I am afraid that it is not so big as you thought. I am willing to run Eric's tests to ensure that there is no any scalability problem. Regards, Zheng -- To unsubscribe from this list: send the line "unsubscribe linux-ext4" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
