* [PATCH] mkfs: enable reflink and rmap by default
@ 2019-05-03 0:57 Darrick J. Wong
2019-05-03 15:55 ` Eric Sandeen
0 siblings, 1 reply; 3+ messages in thread
From: Darrick J. Wong @ 2019-05-03 0:57 UTC (permalink / raw)
To: Eric Sandeen, Dave Chinner; +Cc: xfs
From: Darrick J. Wong <darrick.wong@oracle.com>
Enable reflink and reverse mapping by default.
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
---
mkfs/xfs_mkfs.c | 8 ++++----
1 file changed, 4 insertions(+), 4 deletions(-)
diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
index 0862621a..3874f7dd 100644
--- a/mkfs/xfs_mkfs.c
+++ b/mkfs/xfs_mkfs.c
@@ -2021,14 +2021,14 @@ _("sparse inodes not supported without CRC support\n"));
}
cli->sb_feat.spinodes = false;
- if (cli->sb_feat.rmapbt) {
+ if (cli->sb_feat.rmapbt && cli_opt_set(&mopts, M_RMAPBT)) {
fprintf(stderr,
_("rmapbt not supported without CRC support\n"));
usage();
}
cli->sb_feat.rmapbt = false;
- if (cli->sb_feat.reflink) {
+ if (cli->sb_feat.reflink && cli_opt_set(&mopts, M_REFLINK)) {
fprintf(stderr,
_("reflink not supported without CRC support\n"));
usage();
@@ -3934,8 +3934,8 @@ main(
.dirftype = true,
.finobt = true,
.spinodes = true,
- .rmapbt = false,
- .reflink = false,
+ .rmapbt = true,
+ .reflink = true,
.parent_pointers = false,
.nodalign = false,
.nortalign = false,
^ permalink raw reply related [flat|nested] 3+ messages in thread* Re: [PATCH] mkfs: enable reflink and rmap by default
2019-05-03 0:57 [PATCH] mkfs: enable reflink and rmap by default Darrick J. Wong
@ 2019-05-03 15:55 ` Eric Sandeen
2019-05-03 21:49 ` Darrick J. Wong
0 siblings, 1 reply; 3+ messages in thread
From: Eric Sandeen @ 2019-05-03 15:55 UTC (permalink / raw)
To: Darrick J. Wong, Eric Sandeen, Dave Chinner; +Cc: xfs
On 5/2/19 7:57 PM, Darrick J. Wong wrote:
> From: Darrick J. Wong <darrick.wong@oracle.com>
>
> Enable reflink and reverse mapping by default.
Soooo until now, I thought I had understood that rmap was unlikely
to be default ever, due to performance impacts. Is this no longer the case?
Of course I can't find any emails or IRC logs to indicate I didn't
just dream this up...
Also, at this point, remind me again what uses rmap and why it should
be enabled now?
-Eric
> Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> ---
> mkfs/xfs_mkfs.c | 8 ++++----
> 1 file changed, 4 insertions(+), 4 deletions(-)
>
> diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
> index 0862621a..3874f7dd 100644
> --- a/mkfs/xfs_mkfs.c
> +++ b/mkfs/xfs_mkfs.c
> @@ -2021,14 +2021,14 @@ _("sparse inodes not supported without CRC support\n"));
> }
> cli->sb_feat.spinodes = false;
>
> - if (cli->sb_feat.rmapbt) {
> + if (cli->sb_feat.rmapbt && cli_opt_set(&mopts, M_RMAPBT)) {
> fprintf(stderr,
> _("rmapbt not supported without CRC support\n"));
> usage();
> }
> cli->sb_feat.rmapbt = false;
>
> - if (cli->sb_feat.reflink) {
> + if (cli->sb_feat.reflink && cli_opt_set(&mopts, M_REFLINK)) {
> fprintf(stderr,
> _("reflink not supported without CRC support\n"));
> usage();
> @@ -3934,8 +3934,8 @@ main(
> .dirftype = true,
> .finobt = true,
> .spinodes = true,
> - .rmapbt = false,
> - .reflink = false,
> + .rmapbt = true,
> + .reflink = true,
> .parent_pointers = false,
> .nodalign = false,
> .nortalign = false,
>
^ permalink raw reply [flat|nested] 3+ messages in thread* Re: [PATCH] mkfs: enable reflink and rmap by default
2019-05-03 15:55 ` Eric Sandeen
@ 2019-05-03 21:49 ` Darrick J. Wong
0 siblings, 0 replies; 3+ messages in thread
From: Darrick J. Wong @ 2019-05-03 21:49 UTC (permalink / raw)
To: Eric Sandeen; +Cc: Eric Sandeen, Dave Chinner, xfs
On Fri, May 03, 2019 at 10:55:05AM -0500, Eric Sandeen wrote:
> On 5/2/19 7:57 PM, Darrick J. Wong wrote:
> > From: Darrick J. Wong <darrick.wong@oracle.com>
> >
> > Enable reflink and reverse mapping by default.
>
> Soooo until now, I thought I had understood that rmap was unlikely
> to be default ever, due to performance impacts. Is this no longer the case?
No, it's still the case.
> Of course I can't find any emails or IRC logs to indicate I didn't
> just dream this up...
>
> Also, at this point, remind me again what uses rmap and why it should
> be enabled now?
TBH it's most useful for enhancing xfs_scrub's powers, so there's no
reason to enable it by default until at least 5.2 when we finally land
the last of the scrubbers. Probably. One could make a solid argument
that we ought to wait until the scrub/health code has been stable for
~6mo before even thinking about rmap by default.
Anyway I'll break this into separate pieces and resend the reflink part
for 5.1.
--D
>
> -Eric
>
> > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > ---
> > mkfs/xfs_mkfs.c | 8 ++++----
> > 1 file changed, 4 insertions(+), 4 deletions(-)
> >
> > diff --git a/mkfs/xfs_mkfs.c b/mkfs/xfs_mkfs.c
> > index 0862621a..3874f7dd 100644
> > --- a/mkfs/xfs_mkfs.c
> > +++ b/mkfs/xfs_mkfs.c
> > @@ -2021,14 +2021,14 @@ _("sparse inodes not supported without CRC support\n"));
> > }
> > cli->sb_feat.spinodes = false;
> >
> > - if (cli->sb_feat.rmapbt) {
> > + if (cli->sb_feat.rmapbt && cli_opt_set(&mopts, M_RMAPBT)) {
> > fprintf(stderr,
> > _("rmapbt not supported without CRC support\n"));
> > usage();
> > }
> > cli->sb_feat.rmapbt = false;
> >
> > - if (cli->sb_feat.reflink) {
> > + if (cli->sb_feat.reflink && cli_opt_set(&mopts, M_REFLINK)) {
> > fprintf(stderr,
> > _("reflink not supported without CRC support\n"));
> > usage();
> > @@ -3934,8 +3934,8 @@ main(
> > .dirftype = true,
> > .finobt = true,
> > .spinodes = true,
> > - .rmapbt = false,
> > - .reflink = false,
> > + .rmapbt = true,
> > + .reflink = true,
> > .parent_pointers = false,
> > .nodalign = false,
> > .nortalign = false,
> >
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2019-05-03 21:49 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2019-05-03 0:57 [PATCH] mkfs: enable reflink and rmap by default Darrick J. Wong
2019-05-03 15:55 ` Eric Sandeen
2019-05-03 21:49 ` Darrick J. Wong
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox