| Message ID | 1483076488-2690-1-git-send-email-andreas.gerstmayr@catalysts.cc (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
> On 30 Dec 2016, at 13:37, Andreas Gerstmayr <andreas.gerstmayr@catalysts.cc> wrote: > > This patch sets the io_pages bdi hint based on the rsize mount option. > Without this patch large buffered reads (request size > max readahead) > are processed sequentially in chunks of the readahead size (i.e. read > requests are sent out up to the readahead size, then the > do_generic_file_read() function waits until the first page is received). > > This patch removes this cap and enables parallel reads up to the > specified maximum read size mount option (rsize). > > Signed-off-by: Andreas Gerstmayr <andreas.gerstmayr@catalysts.cc> > --- > > Feedback is appreciated. Maybe we should apply a sensible default value > for rsize instead of unlimited? > > Please note: This patch depends on commit #9491ae4, which is not yet > merged in the testing branch of the ceph-client repository (this commit > is included in kernel version 4.10-rc1). > > > fs/ceph/super.c | 7 +++++++ > 1 file changed, 7 insertions(+) > > diff --git a/fs/ceph/super.c b/fs/ceph/super.c > index 6bd20d7..3c50477 100644 > --- a/fs/ceph/super.c > +++ b/fs/ceph/super.c > @@ -952,6 +952,13 @@ static int ceph_register_bdi(struct super_block *sb, > fsc->backing_dev_info.ra_pages = > VM_MAX_READAHEAD * 1024 / PAGE_SIZE; > > + if (fsc->mount_options->rsize) > + fsc->backing_dev_info.io_pages = > + (fsc->mount_options->rsize + PAGE_SIZE - 1) > + >> PAGE_SHIFT; > + else > + fsc->backing_dev_info.io_pages = ULONG_MAX; > + unlimited by default does not seem like a good idea. I think we should set CEPH_RSIZE_DEFAULT to reasonable value (such as 64M) Regards Yan, Zheng > err = bdi_register(&fsc->backing_dev_info, NULL, "ceph-%ld", > atomic_long_inc_return(&bdi_seq)); > if (!err) > -- > 1.8.3.1 > -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
diff --git a/fs/ceph/super.c b/fs/ceph/super.c index 6bd20d7..3c50477 100644 --- a/fs/ceph/super.c +++ b/fs/ceph/super.c @@ -952,6 +952,13 @@ static int ceph_register_bdi(struct super_block *sb, fsc->backing_dev_info.ra_pages = VM_MAX_READAHEAD * 1024 / PAGE_SIZE; + if (fsc->mount_options->rsize) + fsc->backing_dev_info.io_pages = + (fsc->mount_options->rsize + PAGE_SIZE - 1) + >> PAGE_SHIFT; + else + fsc->backing_dev_info.io_pages = ULONG_MAX; + err = bdi_register(&fsc->backing_dev_info, NULL, "ceph-%ld", atomic_long_inc_return(&bdi_seq)); if (!err)
This patch sets the io_pages bdi hint based on the rsize mount option. Without this patch large buffered reads (request size > max readahead) are processed sequentially in chunks of the readahead size (i.e. read requests are sent out up to the readahead size, then the do_generic_file_read() function waits until the first page is received). This patch removes this cap and enables parallel reads up to the specified maximum read size mount option (rsize). Signed-off-by: Andreas Gerstmayr <andreas.gerstmayr@catalysts.cc> --- Feedback is appreciated. Maybe we should apply a sensible default value for rsize instead of unlimited? Please note: This patch depends on commit #9491ae4, which is not yet merged in the testing branch of the ceph-client repository (this commit is included in kernel version 4.10-rc1). fs/ceph/super.c | 7 +++++++ 1 file changed, 7 insertions(+)