| Message ID | 20210728165632.11813-1-dsterba@suse.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | btrfs: sysfs: advertise zoned support among features | expand |
On 29/07/2021 00:56, David Sterba wrote: > We've hidden the zoned support in sysfs under debug config for the first > releases but now the stability is reasonable, though not all features > have been implemented. > > As this depends on a config option, the per-filesystem feature won't > exist as such filesystem can't be mounted. The static feature will print > 1 when the support is built-in, 0 otherwise. > > Signed-off-by: David Sterba <dsterba@suse.com> > --- > > The merge target is not set, depends if everybody thinks it's the time > even though there are still known bugs. We're also waiting for > util-linux support (blkid, wipefs), so that needs to be synced too. > > fs/btrfs/sysfs.c | 12 +++++++++--- > 1 file changed, 9 insertions(+), 3 deletions(-) > > diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c > index bfe5e27617b0..7ad8f802ab88 100644 > --- a/fs/btrfs/sysfs.c > +++ b/fs/btrfs/sysfs.c > @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); > BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); > BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); > BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); > -/* Remove once support for zoned allocation is feature complete */ > -#ifdef CONFIG_BTRFS_DEBUG > +#ifdef CONFIG_BLK_DEV_ZONED > BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); > #endif > #ifdef CONFIG_FS_VERITY > @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { > BTRFS_FEAT_ATTR_PTR(metadata_uuid), > BTRFS_FEAT_ATTR_PTR(free_space_tree), > BTRFS_FEAT_ATTR_PTR(raid1c34), > -#ifdef CONFIG_BTRFS_DEBUG > +#ifdef CONFIG_BLK_DEV_ZONED > BTRFS_FEAT_ATTR_PTR(zoned), > #endif > #ifdef CONFIG_FS_VERITY Looks good until here. > @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, > BTRFS_ATTR(static_feature, supported_sectorsizes, > supported_sectorsizes_show); > > +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) > +{ > + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); > +} > +BTRFS_ATTR(static_feature, zoned, zoned_show); > + > static struct attribute *btrfs_supported_static_feature_attrs[] = { > BTRFS_ATTR_PTR(static_feature, rmdir_subvol), > BTRFS_ATTR_PTR(static_feature, supported_checksums), > BTRFS_ATTR_PTR(static_feature, send_stream_version), > BTRFS_ATTR_PTR(static_feature, supported_rescue_options), > BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), > + BTRFS_ATTR_PTR(static_feature, zoned), > NULL > }; We don't need this part. btrfs_supported_feature_attrs will take care of showing zoned if when enabled on the mounted btrfs. As of now, this patch fails with [ 44.464597] sysfs: cannot create duplicate filename '/fs/btrfs/features/zoned' Thanks, Anand
On Thu, Aug 05, 2021 at 08:13:03AM +0800, Anand Jain wrote: > On 29/07/2021 00:56, David Sterba wrote: > > We've hidden the zoned support in sysfs under debug config for the first > > releases but now the stability is reasonable, though not all features > > have been implemented. > > > > As this depends on a config option, the per-filesystem feature won't > > exist as such filesystem can't be mounted. The static feature will print > > 1 when the support is built-in, 0 otherwise. > > > > Signed-off-by: David Sterba <dsterba@suse.com> > > --- > > > > The merge target is not set, depends if everybody thinks it's the time > > even though there are still known bugs. We're also waiting for > > util-linux support (blkid, wipefs), so that needs to be synced too. > > > > fs/btrfs/sysfs.c | 12 +++++++++--- > > 1 file changed, 9 insertions(+), 3 deletions(-) > > > > diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c > > index bfe5e27617b0..7ad8f802ab88 100644 > > --- a/fs/btrfs/sysfs.c > > +++ b/fs/btrfs/sysfs.c > > @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); > > BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); > > BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); > > BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); > > -/* Remove once support for zoned allocation is feature complete */ > > -#ifdef CONFIG_BTRFS_DEBUG > > +#ifdef CONFIG_BLK_DEV_ZONED > > BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); > > #endif > > #ifdef CONFIG_FS_VERITY > > @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { > > BTRFS_FEAT_ATTR_PTR(metadata_uuid), > > BTRFS_FEAT_ATTR_PTR(free_space_tree), > > BTRFS_FEAT_ATTR_PTR(raid1c34), > > -#ifdef CONFIG_BTRFS_DEBUG > > +#ifdef CONFIG_BLK_DEV_ZONED > > BTRFS_FEAT_ATTR_PTR(zoned), > > #endif > > #ifdef CONFIG_FS_VERITY > > > Looks good until here. > > > > @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, > > BTRFS_ATTR(static_feature, supported_sectorsizes, > > supported_sectorsizes_show); > > > > +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) > > +{ > > + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); > > +} > > +BTRFS_ATTR(static_feature, zoned, zoned_show); > > + > > static struct attribute *btrfs_supported_static_feature_attrs[] = { > > BTRFS_ATTR_PTR(static_feature, rmdir_subvol), > > BTRFS_ATTR_PTR(static_feature, supported_checksums), > > BTRFS_ATTR_PTR(static_feature, send_stream_version), > > BTRFS_ATTR_PTR(static_feature, supported_rescue_options), > > BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), > > + BTRFS_ATTR_PTR(static_feature, zoned), > > NULL > > }; > > We don't need this part. btrfs_supported_feature_attrs will > take care of showing zoned if when enabled on the mounted btrfs. The idea is to put zoned also to the static features so the zoned support can be determined before mount. > > As of now, this patch fails with > [ 44.464597] sysfs: cannot create duplicate filename > '/fs/btrfs/features/zoned' I'll have a look.
On 09/08/2021 17:19, David Sterba wrote: > On Thu, Aug 05, 2021 at 08:13:03AM +0800, Anand Jain wrote: >> On 29/07/2021 00:56, David Sterba wrote: >>> We've hidden the zoned support in sysfs under debug config for the first >>> releases but now the stability is reasonable, though not all features >>> have been implemented. >>> >>> As this depends on a config option, the per-filesystem feature won't >>> exist as such filesystem can't be mounted. The static feature will print >>> 1 when the support is built-in, 0 otherwise. >>> >>> Signed-off-by: David Sterba <dsterba@suse.com> >>> --- >>> >>> The merge target is not set, depends if everybody thinks it's the time >>> even though there are still known bugs. We're also waiting for >>> util-linux support (blkid, wipefs), so that needs to be synced too. >>> >>> fs/btrfs/sysfs.c | 12 +++++++++--- >>> 1 file changed, 9 insertions(+), 3 deletions(-) >>> >>> diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c >>> index bfe5e27617b0..7ad8f802ab88 100644 >>> --- a/fs/btrfs/sysfs.c >>> +++ b/fs/btrfs/sysfs.c >>> @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); >>> BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); >>> BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); >>> BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); >>> -/* Remove once support for zoned allocation is feature complete */ >>> -#ifdef CONFIG_BTRFS_DEBUG >>> +#ifdef CONFIG_BLK_DEV_ZONED >>> BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); >>> #endif >>> #ifdef CONFIG_FS_VERITY >>> @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { >>> BTRFS_FEAT_ATTR_PTR(metadata_uuid), >>> BTRFS_FEAT_ATTR_PTR(free_space_tree), >>> BTRFS_FEAT_ATTR_PTR(raid1c34), >>> -#ifdef CONFIG_BTRFS_DEBUG >>> +#ifdef CONFIG_BLK_DEV_ZONED >>> BTRFS_FEAT_ATTR_PTR(zoned), >>> #endif >>> #ifdef CONFIG_FS_VERITY >> >> >> Looks good until here. >> >> >>> @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, >>> BTRFS_ATTR(static_feature, supported_sectorsizes, >>> supported_sectorsizes_show); >>> >>> +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) >>> +{ >>> + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); >>> +} >>> +BTRFS_ATTR(static_feature, zoned, zoned_show); >>> + >>> static struct attribute *btrfs_supported_static_feature_attrs[] = { >>> BTRFS_ATTR_PTR(static_feature, rmdir_subvol), >>> BTRFS_ATTR_PTR(static_feature, supported_checksums), >>> BTRFS_ATTR_PTR(static_feature, send_stream_version), >>> BTRFS_ATTR_PTR(static_feature, supported_rescue_options), >>> BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), >>> + BTRFS_ATTR_PTR(static_feature, zoned), >>> NULL >>> }; >> >> We don't need this part. btrfs_supported_feature_attrs will >> take care of showing zoned if when enabled on the mounted btrfs. > > The idea is to put zoned also to the static features so the zoned > support can be determined before mount. As in the proposed patch which adds a table to figure out the correct table to add the attribute, adding to the btrfs_supported_static_feature_attrs attribute list will add only to /sys/fs/btrfs/features, however adding the attribute to btrfs_supported_feature_attrs adds to both /sys/fs/btrfs/features and /sys/fs/btrfs/uuid/features. ------------- btrfs_supported_static_feature_attrs /sys/fs/btrfs/features btrfs_supported_feature_attrs /sys/fs/btrfs/features and /sys/fs/btrfs/uuid/features (if visible) ------------- > >> >> As of now, this patch fails with >> [ 44.464597] sysfs: cannot create duplicate filename >> '/fs/btrfs/features/zoned' > > I'll have a look. >
On Tue, Aug 10, 2021 at 08:30:51AM +0800, Anand Jain wrote: > As in the proposed patch which adds a table to figure out the correct > table to add the attribute, adding to the > btrfs_supported_static_feature_attrs attribute list will add only to > /sys/fs/btrfs/features, however adding the attribute to > btrfs_supported_feature_attrs adds to both /sys/fs/btrfs/features and > /sys/fs/btrfs/uuid/features. I can't see it in the code right now, but that would mean that eg. zoned would show up in static features once such filesystem is mounted. And that does not happen or I'm missing something.
On 19.08.21 г. 2:48, David Sterba wrote: > On Tue, Aug 10, 2021 at 08:30:51AM +0800, Anand Jain wrote: >> As in the proposed patch which adds a table to figure out the correct >> table to add the attribute, adding to the >> btrfs_supported_static_feature_attrs attribute list will add only to >> /sys/fs/btrfs/features, however adding the attribute to >> btrfs_supported_feature_attrs adds to both /sys/fs/btrfs/features and >> /sys/fs/btrfs/uuid/features. > > I can't see it in the code right now, but that would mean that eg. zoned > would show up in static features once such filesystem is mounted. And > that does not happen or I'm missing something. > This happens because when initialising sysfs we create the btrfs_feature_attr_group which contains btrfs_supported_feature_attrs attributes, which have #ifdef CONFIG_BLK_DEV_ZONED BTRFS_FEAT_ATTR_PTR(zoned), #endif Subsequently you define the static feature via BTRFS_ATTR(static_feature, zoned, zoned_show); and finally we call: ret = sysfs_merge_group(&btrfs_kset->kobj, &btrfs_static_feature_attr_group); Which merges the static feature to the feature group. That's why the message about duplication. So one of the 2 definitions needs to go.
On 19.08.21 г. 12:08, Nikolay Borisov wrote: > > > On 19.08.21 г. 2:48, David Sterba wrote: >> On Tue, Aug 10, 2021 at 08:30:51AM +0800, Anand Jain wrote: >>> As in the proposed patch which adds a table to figure out the correct >>> table to add the attribute, adding to the >>> btrfs_supported_static_feature_attrs attribute list will add only to >>> /sys/fs/btrfs/features, however adding the attribute to >>> btrfs_supported_feature_attrs adds to both /sys/fs/btrfs/features and >>> /sys/fs/btrfs/uuid/features. >> >> I can't see it in the code right now, but that would mean that eg. zoned >> would show up in static features once such filesystem is mounted. And >> that does not happen or I'm missing something. >> > > This happens because when initialising sysfs we create the > btrfs_feature_attr_group which contains btrfs_supported_feature_attrs > attributes, which have > #ifdef CONFIG_BLK_DEV_ZONED > > BTRFS_FEAT_ATTR_PTR(zoned), > > #endif > > Subsequently you define the static feature via > BTRFS_ATTR(static_feature, zoned, zoned_show); > > and finally we call: > > ret = sysfs_merge_group(&btrfs_kset->kobj, > &btrfs_static_feature_attr_group); > > > Which merges the static feature to the feature group. That's why the > message about duplication. So one of the 2 definitions needs to go. > Looking at the maze that btrfs' sysfs code is it seems the decision which of the two sets to use when defining a feature is whether it can be enabled or not at runtime. I think for zoned block device we can't really disable the support at runtime? If so then it needs to only be defined as a static feature? Just because the kernel supports it doesn't necessarily mean it's going to be used if the underlying device is not zoned, right, it should depend on the runtime characteristics of the underlying device?
On 28.07.21 г. 19:56, David Sterba wrote: > We've hidden the zoned support in sysfs under debug config for the first > releases but now the stability is reasonable, though not all features > have been implemented. > > As this depends on a config option, the per-filesystem feature won't > exist as such filesystem can't be mounted. The static feature will print > 1 when the support is built-in, 0 otherwise. > > Signed-off-by: David Sterba <dsterba@suse.com> > --- > > The merge target is not set, depends if everybody thinks it's the time > even though there are still known bugs. We're also waiting for > util-linux support (blkid, wipefs), so that needs to be synced too. > > fs/btrfs/sysfs.c | 12 +++++++++--- > 1 file changed, 9 insertions(+), 3 deletions(-) > > diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c > index bfe5e27617b0..7ad8f802ab88 100644 > --- a/fs/btrfs/sysfs.c > +++ b/fs/btrfs/sysfs.c > @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); > BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); > BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); > BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); > -/* Remove once support for zoned allocation is feature complete */ > -#ifdef CONFIG_BTRFS_DEBUG > +#ifdef CONFIG_BLK_DEV_ZONED > BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); > #endif > #ifdef CONFIG_FS_VERITY > @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { > BTRFS_FEAT_ATTR_PTR(metadata_uuid), > BTRFS_FEAT_ATTR_PTR(free_space_tree), > BTRFS_FEAT_ATTR_PTR(raid1c34), > -#ifdef CONFIG_BTRFS_DEBUG > +#ifdef CONFIG_BLK_DEV_ZONED > BTRFS_FEAT_ATTR_PTR(zoned), > #endif > #ifdef CONFIG_FS_VERITY > @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, > BTRFS_ATTR(static_feature, supported_sectorsizes, > supported_sectorsizes_show); > > +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) > +{ > + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); > +} > +BTRFS_ATTR(static_feature, zoned, zoned_show); > + > static struct attribute *btrfs_supported_static_feature_attrs[] = { > BTRFS_ATTR_PTR(static_feature, rmdir_subvol), > BTRFS_ATTR_PTR(static_feature, supported_checksums), > BTRFS_ATTR_PTR(static_feature, send_stream_version), > BTRFS_ATTR_PTR(static_feature, supported_rescue_options), > BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), > + BTRFS_ATTR_PTR(static_feature, zoned), > NULL > }; Why isn't the above hunk predicated on CONFIG_BLK_DEV_ZONED the same as the ATTR_INCOMPAT zoneed bit, but as explained in my earlier email one of these should go and whichever remains must be predicated on CONFIG_BLK_DEV_ZONED. > >
On 2021/08/19 18:21, Nikolay Borisov wrote: > > > On 28.07.21 г. 19:56, David Sterba wrote: >> We've hidden the zoned support in sysfs under debug config for the first >> releases but now the stability is reasonable, though not all features >> have been implemented. >> >> As this depends on a config option, the per-filesystem feature won't >> exist as such filesystem can't be mounted. The static feature will print >> 1 when the support is built-in, 0 otherwise. >> >> Signed-off-by: David Sterba <dsterba@suse.com> >> --- >> >> The merge target is not set, depends if everybody thinks it's the time >> even though there are still known bugs. We're also waiting for >> util-linux support (blkid, wipefs), so that needs to be synced too. >> >> fs/btrfs/sysfs.c | 12 +++++++++--- >> 1 file changed, 9 insertions(+), 3 deletions(-) >> >> diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c >> index bfe5e27617b0..7ad8f802ab88 100644 >> --- a/fs/btrfs/sysfs.c >> +++ b/fs/btrfs/sysfs.c >> @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); >> BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); >> BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); >> BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); >> -/* Remove once support for zoned allocation is feature complete */ >> -#ifdef CONFIG_BTRFS_DEBUG >> +#ifdef CONFIG_BLK_DEV_ZONED >> BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); >> #endif >> #ifdef CONFIG_FS_VERITY >> @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { >> BTRFS_FEAT_ATTR_PTR(metadata_uuid), >> BTRFS_FEAT_ATTR_PTR(free_space_tree), >> BTRFS_FEAT_ATTR_PTR(raid1c34), >> -#ifdef CONFIG_BTRFS_DEBUG >> +#ifdef CONFIG_BLK_DEV_ZONED >> BTRFS_FEAT_ATTR_PTR(zoned), >> #endif >> #ifdef CONFIG_FS_VERITY >> @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, >> BTRFS_ATTR(static_feature, supported_sectorsizes, >> supported_sectorsizes_show); >> >> +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) >> +{ >> + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); >> +} >> +BTRFS_ATTR(static_feature, zoned, zoned_show); >> + >> static struct attribute *btrfs_supported_static_feature_attrs[] = { >> BTRFS_ATTR_PTR(static_feature, rmdir_subvol), >> BTRFS_ATTR_PTR(static_feature, supported_checksums), >> BTRFS_ATTR_PTR(static_feature, send_stream_version), >> BTRFS_ATTR_PTR(static_feature, supported_rescue_options), >> BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), >> + BTRFS_ATTR_PTR(static_feature, zoned), >> NULL >> }; > > Why isn't the above hunk predicated on CONFIG_BLK_DEV_ZONED the same as > the ATTR_INCOMPAT zoneed bit, but as explained in my earlier email one > of these should go and whichever remains must be predicated on > CONFIG_BLK_DEV_ZONED. zoned-btrfs can be used with regular devices too. In that case, zones are emulated, all of them being conventional. So btrfs zoned feature should definitely not be dependent on CONFIG_BLK_DEV_ZONED. If CONFIG_BLK_DEV_ZONED is not defined, then zoned btrfs will be usable only on regular devices. But since in that case zoned devices will not show up, it is all consistent.
On 19.08.21 г. 12:31, Damien Le Moal wrote: > On 2021/08/19 18:21, Nikolay Borisov wrote: >> >> >> On 28.07.21 г. 19:56, David Sterba wrote: >>> We've hidden the zoned support in sysfs under debug config for the first >>> releases but now the stability is reasonable, though not all features >>> have been implemented. >>> >>> As this depends on a config option, the per-filesystem feature won't >>> exist as such filesystem can't be mounted. The static feature will print >>> 1 when the support is built-in, 0 otherwise. >>> >>> Signed-off-by: David Sterba <dsterba@suse.com> >>> --- >>> >>> The merge target is not set, depends if everybody thinks it's the time >>> even though there are still known bugs. We're also waiting for >>> util-linux support (blkid, wipefs), so that needs to be synced too. >>> >>> fs/btrfs/sysfs.c | 12 +++++++++--- >>> 1 file changed, 9 insertions(+), 3 deletions(-) >>> >>> diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c >>> index bfe5e27617b0..7ad8f802ab88 100644 >>> --- a/fs/btrfs/sysfs.c >>> +++ b/fs/btrfs/sysfs.c >>> @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); >>> BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); >>> BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); >>> BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); >>> -/* Remove once support for zoned allocation is feature complete */ >>> -#ifdef CONFIG_BTRFS_DEBUG >>> +#ifdef CONFIG_BLK_DEV_ZONED >>> BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); >>> #endif >>> #ifdef CONFIG_FS_VERITY >>> @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { >>> BTRFS_FEAT_ATTR_PTR(metadata_uuid), >>> BTRFS_FEAT_ATTR_PTR(free_space_tree), >>> BTRFS_FEAT_ATTR_PTR(raid1c34), >>> -#ifdef CONFIG_BTRFS_DEBUG >>> +#ifdef CONFIG_BLK_DEV_ZONED >>> BTRFS_FEAT_ATTR_PTR(zoned), >>> #endif >>> #ifdef CONFIG_FS_VERITY >>> @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, >>> BTRFS_ATTR(static_feature, supported_sectorsizes, >>> supported_sectorsizes_show); >>> >>> +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) >>> +{ >>> + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); >>> +} >>> +BTRFS_ATTR(static_feature, zoned, zoned_show); >>> + >>> static struct attribute *btrfs_supported_static_feature_attrs[] = { >>> BTRFS_ATTR_PTR(static_feature, rmdir_subvol), >>> BTRFS_ATTR_PTR(static_feature, supported_checksums), >>> BTRFS_ATTR_PTR(static_feature, send_stream_version), >>> BTRFS_ATTR_PTR(static_feature, supported_rescue_options), >>> BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), >>> + BTRFS_ATTR_PTR(static_feature, zoned), >>> NULL >>> }; >> >> Why isn't the above hunk predicated on CONFIG_BLK_DEV_ZONED the same as >> the ATTR_INCOMPAT zoneed bit, but as explained in my earlier email one >> of these should go and whichever remains must be predicated on >> CONFIG_BLK_DEV_ZONED. > > zoned-btrfs can be used with regular devices too. In that case, zones are > emulated, all of them being conventional. So btrfs zoned feature should > definitely not be dependent on CONFIG_BLK_DEV_ZONED. > > If CONFIG_BLK_DEV_ZONED is not defined, then zoned btrfs will be usable only on > regular devices. But since in that case zoned devices will not show up, it is > all consistent. Then we should discuss what the semantics of the ZONED flag under features should be? I.e do we need to explicitly distinguish between "btrfs supports zoned AND the kernel is compiled with BLK_DEV_ZONED so we say we support it" and "btrfs has support for zoned devices but you have to figure on your own if BLK_DEV_ZONED is enabled" ? I.e by having sys/fs/btrfs/zoned set to 1 what information do we want to convey to the user?
On 2021/08/19 18:37, Nikolay Borisov wrote: > > > On 19.08.21 г. 12:31, Damien Le Moal wrote: >> On 2021/08/19 18:21, Nikolay Borisov wrote: >>> >>> >>> On 28.07.21 г. 19:56, David Sterba wrote: >>>> We've hidden the zoned support in sysfs under debug config for the first >>>> releases but now the stability is reasonable, though not all features >>>> have been implemented. >>>> >>>> As this depends on a config option, the per-filesystem feature won't >>>> exist as such filesystem can't be mounted. The static feature will print >>>> 1 when the support is built-in, 0 otherwise. >>>> >>>> Signed-off-by: David Sterba <dsterba@suse.com> >>>> --- >>>> >>>> The merge target is not set, depends if everybody thinks it's the time >>>> even though there are still known bugs. We're also waiting for >>>> util-linux support (blkid, wipefs), so that needs to be synced too. >>>> >>>> fs/btrfs/sysfs.c | 12 +++++++++--- >>>> 1 file changed, 9 insertions(+), 3 deletions(-) >>>> >>>> diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c >>>> index bfe5e27617b0..7ad8f802ab88 100644 >>>> --- a/fs/btrfs/sysfs.c >>>> +++ b/fs/btrfs/sysfs.c >>>> @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); >>>> BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); >>>> BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); >>>> BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); >>>> -/* Remove once support for zoned allocation is feature complete */ >>>> -#ifdef CONFIG_BTRFS_DEBUG >>>> +#ifdef CONFIG_BLK_DEV_ZONED >>>> BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); >>>> #endif >>>> #ifdef CONFIG_FS_VERITY >>>> @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { >>>> BTRFS_FEAT_ATTR_PTR(metadata_uuid), >>>> BTRFS_FEAT_ATTR_PTR(free_space_tree), >>>> BTRFS_FEAT_ATTR_PTR(raid1c34), >>>> -#ifdef CONFIG_BTRFS_DEBUG >>>> +#ifdef CONFIG_BLK_DEV_ZONED >>>> BTRFS_FEAT_ATTR_PTR(zoned), >>>> #endif >>>> #ifdef CONFIG_FS_VERITY >>>> @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, >>>> BTRFS_ATTR(static_feature, supported_sectorsizes, >>>> supported_sectorsizes_show); >>>> >>>> +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) >>>> +{ >>>> + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); >>>> +} >>>> +BTRFS_ATTR(static_feature, zoned, zoned_show); >>>> + >>>> static struct attribute *btrfs_supported_static_feature_attrs[] = { >>>> BTRFS_ATTR_PTR(static_feature, rmdir_subvol), >>>> BTRFS_ATTR_PTR(static_feature, supported_checksums), >>>> BTRFS_ATTR_PTR(static_feature, send_stream_version), >>>> BTRFS_ATTR_PTR(static_feature, supported_rescue_options), >>>> BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), >>>> + BTRFS_ATTR_PTR(static_feature, zoned), >>>> NULL >>>> }; >>> >>> Why isn't the above hunk predicated on CONFIG_BLK_DEV_ZONED the same as >>> the ATTR_INCOMPAT zoneed bit, but as explained in my earlier email one >>> of these should go and whichever remains must be predicated on >>> CONFIG_BLK_DEV_ZONED. >> >> zoned-btrfs can be used with regular devices too. In that case, zones are >> emulated, all of them being conventional. So btrfs zoned feature should >> definitely not be dependent on CONFIG_BLK_DEV_ZONED. >> >> If CONFIG_BLK_DEV_ZONED is not defined, then zoned btrfs will be usable only on >> regular devices. But since in that case zoned devices will not show up, it is >> all consistent. > > Then we should discuss what the semantics of the ZONED flag under > features should be? I.e do we need to explicitly distinguish between > "btrfs supports zoned AND the kernel is compiled with BLK_DEV_ZONED so > we say we support it" and "btrfs has support for zoned devices but you > have to figure on your own if BLK_DEV_ZONED is enabled" ? I.e by having > sys/fs/btrfs/zoned set to 1 what information do we want to convey to the > user? That this version of btrfs supports the zoned mode. What device you can actually apply it to then depends on CONFIG_BLK_DEV_ZONED. If that one is set, the zoned btrfs feature applies to all devices, regular and zoned. If it is not set, then the feature applies to regular devices only, since zoned devices will NOT be present. The user will not see them :) Rather all simple I think. But I am heavily biased with this since I have been doing zoned stuff for years. So someone totally new to zoned storage needs to chime in !
On 19/08/2021 17:14, Nikolay Borisov wrote: > > > On 19.08.21 г. 12:08, Nikolay Borisov wrote: >> >> >> On 19.08.21 г. 2:48, David Sterba wrote: >>> On Tue, Aug 10, 2021 at 08:30:51AM +0800, Anand Jain wrote: >>>> As in the proposed patch which adds a table to figure out the correct >>>> table to add the attribute, adding to the >>>> btrfs_supported_static_feature_attrs attribute list will add only to >>>> /sys/fs/btrfs/features, however adding the attribute to >>>> btrfs_supported_feature_attrs adds to both /sys/fs/btrfs/features and >>>> /sys/fs/btrfs/uuid/features. >>> >>> I can't see it in the code right now, but that would mean that eg. zoned >>> would show up in static features once such filesystem is mounted. And >>> that does not happen or I'm missing something. >>> >> >> This happens because when initialising sysfs we create the >> btrfs_feature_attr_group which contains btrfs_supported_feature_attrs >> attributes, which have >> #ifdef CONFIG_BLK_DEV_ZONED >> >> BTRFS_FEAT_ATTR_PTR(zoned), >> >> #endif >> >> Subsequently you define the static feature via >> BTRFS_ATTR(static_feature, zoned, zoned_show); >> >> and finally we call: >> >> ret = sysfs_merge_group(&btrfs_kset->kobj, >> &btrfs_static_feature_attr_group); >> >> >> Which merges the static feature to the feature group. That's why the >> message about duplication. So one of the 2 definitions needs to go. >> > > Looking at the maze that btrfs' sysfs code is it seems the decision > which of the two sets to use when defining a feature is whether it can ...be enabled per fsid. :: > I think for zoned block device we can't > really disable the support at runtime? Hmm. It is not like that. Suppose there are two btrfs filesystems on a system fsid1 and fsid2. fsid1 is zoned. fsid2 is a normal conventional device. Then on fsid1 zoned incompatible feature/flag is enabled and, on fsid2 it is not enabled. So IMO zoned should be added to btrfs_supported_feature_attrs. By doing this, on /sys/fs/btrfs/feature zoned is shown on /sys/fs/btrfs/fsid1/feature zoned is shown on /sys/fs/btrfs/faid2/feature zoned is NOT shown btrfs_feature_visible() and get_features() manages to do that dynamically. > If so then it needs to only be > defined as a static feature? No. pls. > Just because the kernel supports it doesn't > necessarily mean it's going to be used if the underlying device is not > zoned, right, it should depend on the runtime characteristics of the > underlying device? Right. As in the above example. Imagine runtime == mkfs time in this context, so to say.
On 19/08/2021 19:16, Anand Jain wrote: > > > On 19/08/2021 17:14, Nikolay Borisov wrote: >> >> >> On 19.08.21 г. 12:08, Nikolay Borisov wrote: >>> >>> >>> On 19.08.21 г. 2:48, David Sterba wrote: >>>> On Tue, Aug 10, 2021 at 08:30:51AM +0800, Anand Jain wrote: >>>>> As in the proposed patch which adds a table to figure out the >>>>> correct >>>>> table to add the attribute, adding to the >>>>> btrfs_supported_static_feature_attrs attribute list will add only to >>>>> /sys/fs/btrfs/features, however adding the attribute to >>>>> btrfs_supported_feature_attrs adds to both /sys/fs/btrfs/features and >>>>> /sys/fs/btrfs/uuid/features. >>>> >>>> I can't see it in the code right now, but that would mean that eg. >>>> zoned >>>> would show up in static features once such filesystem is mounted. And >>>> that does not happen or I'm missing something. >>>> >>> >>> This happens because when initialising sysfs we create the >>> btrfs_feature_attr_group which contains btrfs_supported_feature_attrs >>> attributes, which have >>> #ifdef CONFIG_BLK_DEV_ZONED >>> >>> BTRFS_FEAT_ATTR_PTR(zoned), >>> >>> #endif >>> >>> Subsequently you define the static feature via >>> BTRFS_ATTR(static_feature, zoned, zoned_show); >>> >>> and finally we call: >>> >>> ret = sysfs_merge_group(&btrfs_kset->kobj, >>> &btrfs_static_feature_attr_group); >>> >>> >>> Which merges the static feature to the feature group. That's why the >>> message about duplication. So one of the 2 definitions needs to go. >>> >> >> Looking at the maze that btrfs' sysfs code is it seems the decision > > >> which of the two sets to use when defining a feature is whether it can > > ...be enabled per fsid. > :: > >> I think for zoned block device we can't >> really disable the support at runtime? > > Hmm. It is not like that. Suppose there are two btrfs filesystems on a > system fsid1 and fsid2. fsid1 is zoned. fsid2 is a normal conventional > device. Then on fsid1 zoned incompatible feature/flag is enabled and, > on fsid2 it is not enabled. > > So IMO zoned should be added to btrfs_supported_feature_attrs. Provided /sys/fs/btrfs/feature/zoned will return nothing in specific values like sectorsize. > > By doing this, > on /sys/fs/btrfs/feature zoned is shown > on /sys/fs/btrfs/fsid1/feature zoned is shown > on /sys/fs/btrfs/faid2/feature zoned is NOT shown > > btrfs_feature_visible() and get_features() manages to do that > dynamically. > > >> If so then it needs to only be >> defined as a static feature? > > No. pls. > >> Just because the kernel supports it doesn't >> necessarily mean it's going to be used if the underlying device is not >> zoned, right, it should depend on the runtime characteristics of the >> underlying device? > > Right. As in the above example. Imagine runtime == mkfs time in this > context, so to say. >
diff --git a/fs/btrfs/sysfs.c b/fs/btrfs/sysfs.c index bfe5e27617b0..7ad8f802ab88 100644 --- a/fs/btrfs/sysfs.c +++ b/fs/btrfs/sysfs.c @@ -263,8 +263,7 @@ BTRFS_FEAT_ATTR_INCOMPAT(no_holes, NO_HOLES); BTRFS_FEAT_ATTR_INCOMPAT(metadata_uuid, METADATA_UUID); BTRFS_FEAT_ATTR_COMPAT_RO(free_space_tree, FREE_SPACE_TREE); BTRFS_FEAT_ATTR_INCOMPAT(raid1c34, RAID1C34); -/* Remove once support for zoned allocation is feature complete */ -#ifdef CONFIG_BTRFS_DEBUG +#ifdef CONFIG_BLK_DEV_ZONED BTRFS_FEAT_ATTR_INCOMPAT(zoned, ZONED); #endif #ifdef CONFIG_FS_VERITY @@ -285,7 +284,7 @@ static struct attribute *btrfs_supported_feature_attrs[] = { BTRFS_FEAT_ATTR_PTR(metadata_uuid), BTRFS_FEAT_ATTR_PTR(free_space_tree), BTRFS_FEAT_ATTR_PTR(raid1c34), -#ifdef CONFIG_BTRFS_DEBUG +#ifdef CONFIG_BLK_DEV_ZONED BTRFS_FEAT_ATTR_PTR(zoned), #endif #ifdef CONFIG_FS_VERITY @@ -384,12 +383,19 @@ static ssize_t supported_sectorsizes_show(struct kobject *kobj, BTRFS_ATTR(static_feature, supported_sectorsizes, supported_sectorsizes_show); +static ssize_t zoned_show(struct kobject *kobj, struct kobj_attribute *a, char *buf) +{ + return scnprintf(buf, PAGE_SIZE, "%d\n", IS_ENABLED(CONFIG_BLK_DEV_ZONED)); +} +BTRFS_ATTR(static_feature, zoned, zoned_show); + static struct attribute *btrfs_supported_static_feature_attrs[] = { BTRFS_ATTR_PTR(static_feature, rmdir_subvol), BTRFS_ATTR_PTR(static_feature, supported_checksums), BTRFS_ATTR_PTR(static_feature, send_stream_version), BTRFS_ATTR_PTR(static_feature, supported_rescue_options), BTRFS_ATTR_PTR(static_feature, supported_sectorsizes), + BTRFS_ATTR_PTR(static_feature, zoned), NULL };
We've hidden the zoned support in sysfs under debug config for the first releases but now the stability is reasonable, though not all features have been implemented. As this depends on a config option, the per-filesystem feature won't exist as such filesystem can't be mounted. The static feature will print 1 when the support is built-in, 0 otherwise. Signed-off-by: David Sterba <dsterba@suse.com> --- The merge target is not set, depends if everybody thinks it's the time even though there are still known bugs. We're also waiting for util-linux support (blkid, wipefs), so that needs to be synced too. fs/btrfs/sysfs.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-)