Hi Ying Han,
On Wed, Apr 08, 2009 at 09:17:26AM +0800, Ying Han wrote:
> On Tue, Apr 7, 2009 at 4:27 PM, Wu Fengguang <fengguang.wu@xxxxxxxxx> wrote:
> > On Wed, Apr 08, 2009 at 04:03:36AM +0800, Ying Han wrote:
> >> On Tue, Apr 7, 2009 at 12:17 AM, Wu Fengguang <fengguang.wu@xxxxxxxxx> wrote:
> >> > Cc: Ying Han <yinghan@xxxxxxxxxx>
> >> > Signed-off-by: Wu Fengguang <fengguang.wu@xxxxxxxxx>
> >> > ---
> >> > mm/memory.c | 4 +---
> >> > 1 file changed, 1 insertion(+), 3 deletions(-)
> >> >
> >> > --- mm.orig/mm/memory.c
> >> > +++ mm/mm/memory.c
> >> > @@ -2766,10 +2766,8 @@ static int do_linear_fault(struct mm_str
> >> > {
> >> > pgoff_t pgoff = (((address & PAGE_MASK)
> >> > - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff;
> >> > - int write = write_access & ~FAULT_FLAG_RETRY;
> >> > - unsigned int flags = (write ? FAULT_FLAG_WRITE : 0);
> >> > + unsigned int flags = (write_access ? FAULT_FLAG_WRITE : 0);
> >> >
> >> > - flags |= (write_access & FAULT_FLAG_RETRY);
> >> > pte_unmap(page_table);
> >> > return __do_fault(mm, vma, address, pmd, pgoff, flags, orig_pte);
> >> > }
> >> So, we got rid of FAULT_FLAG_RETRY flag?
> >
> > Seems yes for the current mm tree, see the following two commits.
> >
> > I did this patch on seeing 761fe7bc8193b7. But a closer look
> > indicates that the following two patches disable the filemap
> > VM_FAULT_RETRY part totally...
> >
> > Anyway, if these two patches are to be reverted somehow(I guess yes),
> > this patch shall be _ignored_.
> >
> > btw, do you have any test case and performance numbers for
> > FAULT_FLAG_RETRY? And possible overheads for (the worst case)
> > sparse random mmap reads on a sparse file? I cannot find any
> > in your changelogs..
>
> here is the benchmark i posted on [V1] but somehow missed in [V2] describtion
>
> Benchmarks:
> case 1. one application has a high count of threads each faulting in
> different pages of a hugefile. Benchmark indicate that this double data
> structure walking in case of major fault results in << 1% performance hit.
>
> case 2. add another thread in the above application which in a tight loop of
> mmap()/munmap(). Here we measure loop count in the new thread while other
> threads doing the same amount of work as case one. we got << 3% performance
> hit on the Complete Time(benchmark value for case one) and 10% performance
> improvement on the mmap()/munmap() counter.
>
> This patch helps a lot in cases we have writer which is waitting behind all
> readers, so it could execute much faster.
>
Just tested the sparse-random-read-on-sparse-file case, and found the
performance impact to be 0.4% (8.706s vs 8.744s). Kind of acceptable.
without FAULT_FLAG_RETRY:
iotrace.rb --load stride-100 --mplay /mnt/btrfs-ram/sparse 3.28s user 5.39s system 99% cpu 8.692 total
iotrace.rb --load stride-100 --mplay /mnt/btrfs-ram/sparse 3.17s user 5.54s system 99% cpu 8.742 total
iotrace.rb --load stride-100 --mplay /mnt/btrfs-ram/sparse 3.18s user 5.48s system 99% cpu 8.684 total
FAULT_FLAG_RETRY:
iotrace.rb --load stride-100 --mplay /mnt/btrfs-ram/sparse 3.18s user 5.63s system 99% cpu 8.825 total
iotrace.rb --load stride-100 --mplay /mnt/btrfs-ram/sparse 3.22s user 5.47s system 99% cpu 8.718 total
iotrace.rb --load stride-100 --mplay /mnt/btrfs-ram/sparse 3.13s user 5.55s system 99% cpu 8.690 total
In the above faked workload, the mmap read page offsets are loaded from
stride-100 and performed on /mnt/btrfs-ram/sparse, which are created by:
seq 0 100 1000000 > stride-100
dd if=/dev/zero of=/mnt/btrfs-ram/sparse bs=1M count=1 seek=1024000
Thanks,
Fengguang
--
To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
