I create 10 fio simultaneously in my server and I found that clat
attribute shows 50% and 95% has a 10 times difference.(each fio write
1GB file with 4KB blocksize with buffered IO)

That result maybe means later IO will cost much time than early IO.(Just guess)

I lookup the generic_perform_write() and the following procedure like below:
write_begin
> prepare_write_begin
> f2fs_wait_on_page_writeback
> zero_segments
write_end
> set_page_dirty

Is there any possible bottleneck in the above function?
Does f2fs_wait_on_page_writeback always wait there?


_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel

Reply via email to