| Message ID | 20220826020400.144377-1-shinichiro.kawasaki@wdc.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | btrfs: zoned: set pseudo max append zone limit in zone emulation mode | expand |
On Fri, Aug 26, 2022 at 11:04:00AM +0900, Shin'ichiro Kawasaki wrote: > The commit 7d7672bc5d10 ("btrfs: convert count_max_extents() to use > fs_info->max_extent_size") introduced a division by > fs_info->max_extent_size. This max_extent_size is initialized with max > zone append limit size of the device btrfs runs on. However, in zone > emulation mode, the device is not zoned then its zone append limit is > zero. This resulted in zero value of fs_info->max_extent_size and caused > zero division error. > > Fix the error by setting non-zero pseudo value to max append zone limit > in zone emulation mode. Set the pseudo value based on max_segments as > suggested in the commit c2ae7b772ef4 ("btrfs: zoned: revive > max_zone_append_bytes"). > > Fixes: 7d7672bc5d10 ("btrfs: convert count_max_extents() to use fs_info->max_extent_size") > Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com> > --- > fs/btrfs/zoned.c | 15 ++++++++++++--- > 1 file changed, 12 insertions(+), 3 deletions(-) > > diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c > index b150b07ba1a7..560dd0a67536 100644 > --- a/fs/btrfs/zoned.c > +++ b/fs/btrfs/zoned.c > @@ -421,10 +421,19 @@ int btrfs_get_dev_zone_info(struct btrfs_device *device, bool populate_cache) > * since btrfs adds the pages one by one to a bio, and btrfs cannot > * increase the metadata reservation even if it increases the number of > * extents, it is safe to stick with the limit. > + * > + * When zoned btrfs is in zone emulation mode, bdev is a non-zoned > + * device and does not have valid max zone append size. In this case, > + * use max_segments * PAGE_SIZE as the pseudo max_zone_append_size. Using 'zoned filesystem' or 'zoned mode' is preferred to 'zoned btrfs'. And, 'zoned btrfs is in zone emulation mode' looks a bit wordy. So, how about this? With the zoned emulation, we can have non-zoned device on the zoned mode. In this case, we don't have a valid max zone append size. So, use ... > */ > - zone_info->max_zone_append_size = > - min_t(u64, (u64)bdev_max_zone_append_sectors(bdev) << SECTOR_SHIFT, > - (u64)bdev_max_segments(bdev) << PAGE_SHIFT); > + if (bdev_is_zoned(bdev)) { > + zone_info->max_zone_append_size = min_t(u64, > + (u64)bdev_max_zone_append_sectors(bdev) << SECTOR_SHIFT, > + (u64)bdev_max_segments(bdev) << PAGE_SHIFT); > + } else { > + zone_info->max_zone_append_size = > + bdev_max_segments(bdev) << PAGE_SHIFT; We need to cast it to "(u64)" to keep enough bit width. > + } > if (!IS_ALIGNED(nr_sectors, zone_sectors)) > zone_info->nr_zones++; > > -- > 2.37.2 >
On Aug 26, 2022 / 05:39, Naohiro Aota wrote: > On Fri, Aug 26, 2022 at 11:04:00AM +0900, Shin'ichiro Kawasaki wrote: > > The commit 7d7672bc5d10 ("btrfs: convert count_max_extents() to use > > fs_info->max_extent_size") introduced a division by > > fs_info->max_extent_size. This max_extent_size is initialized with max > > zone append limit size of the device btrfs runs on. However, in zone > > emulation mode, the device is not zoned then its zone append limit is > > zero. This resulted in zero value of fs_info->max_extent_size and caused > > zero division error. > > > > Fix the error by setting non-zero pseudo value to max append zone limit > > in zone emulation mode. Set the pseudo value based on max_segments as > > suggested in the commit c2ae7b772ef4 ("btrfs: zoned: revive > > max_zone_append_bytes"). > > > > Fixes: 7d7672bc5d10 ("btrfs: convert count_max_extents() to use fs_info->max_extent_size") > > Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com> > > --- > > fs/btrfs/zoned.c | 15 ++++++++++++--- > > 1 file changed, 12 insertions(+), 3 deletions(-) > > > > diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c > > index b150b07ba1a7..560dd0a67536 100644 > > --- a/fs/btrfs/zoned.c > > +++ b/fs/btrfs/zoned.c > > @@ -421,10 +421,19 @@ int btrfs_get_dev_zone_info(struct btrfs_device *device, bool populate_cache) > > * since btrfs adds the pages one by one to a bio, and btrfs cannot > > * increase the metadata reservation even if it increases the number of > > * extents, it is safe to stick with the limit. > > + * > > + * When zoned btrfs is in zone emulation mode, bdev is a non-zoned > > + * device and does not have valid max zone append size. In this case, > > + * use max_segments * PAGE_SIZE as the pseudo max_zone_append_size. > > Using 'zoned filesystem' or 'zoned mode' is preferred to 'zoned > btrfs'. And, 'zoned btrfs is in zone emulation mode' looks a bit wordy. So, > how about this? > > With the zoned emulation, we can have non-zoned device on the zoned > mode. In this case, we don't have a valid max zone append size. So, use > ... Okay, will rephrase as suggested. > > > */ > > - zone_info->max_zone_append_size = > > - min_t(u64, (u64)bdev_max_zone_append_sectors(bdev) << SECTOR_SHIFT, > > - (u64)bdev_max_segments(bdev) << PAGE_SHIFT); > > + if (bdev_is_zoned(bdev)) { > > + zone_info->max_zone_append_size = min_t(u64, > > + (u64)bdev_max_zone_append_sectors(bdev) << SECTOR_SHIFT, > > + (u64)bdev_max_segments(bdev) << PAGE_SHIFT); > > + } else { > > + zone_info->max_zone_append_size = > > + bdev_max_segments(bdev) << PAGE_SHIFT; > > We need to cast it to "(u64)" to keep enough bit width. Oops, thanks for the catch. Will fix it in v2. > > > + } > > if (!IS_ALIGNED(nr_sectors, zone_sectors)) > > zone_info->nr_zones++; > > > > -- > > 2.37.2 > >
diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c index b150b07ba1a7..560dd0a67536 100644 --- a/fs/btrfs/zoned.c +++ b/fs/btrfs/zoned.c @@ -421,10 +421,19 @@ int btrfs_get_dev_zone_info(struct btrfs_device *device, bool populate_cache) * since btrfs adds the pages one by one to a bio, and btrfs cannot * increase the metadata reservation even if it increases the number of * extents, it is safe to stick with the limit. + * + * When zoned btrfs is in zone emulation mode, bdev is a non-zoned + * device and does not have valid max zone append size. In this case, + * use max_segments * PAGE_SIZE as the pseudo max_zone_append_size. */ - zone_info->max_zone_append_size = - min_t(u64, (u64)bdev_max_zone_append_sectors(bdev) << SECTOR_SHIFT, - (u64)bdev_max_segments(bdev) << PAGE_SHIFT); + if (bdev_is_zoned(bdev)) { + zone_info->max_zone_append_size = min_t(u64, + (u64)bdev_max_zone_append_sectors(bdev) << SECTOR_SHIFT, + (u64)bdev_max_segments(bdev) << PAGE_SHIFT); + } else { + zone_info->max_zone_append_size = + bdev_max_segments(bdev) << PAGE_SHIFT; + } if (!IS_ALIGNED(nr_sectors, zone_sectors)) zone_info->nr_zones++;
The commit 7d7672bc5d10 ("btrfs: convert count_max_extents() to use fs_info->max_extent_size") introduced a division by fs_info->max_extent_size. This max_extent_size is initialized with max zone append limit size of the device btrfs runs on. However, in zone emulation mode, the device is not zoned then its zone append limit is zero. This resulted in zero value of fs_info->max_extent_size and caused zero division error. Fix the error by setting non-zero pseudo value to max append zone limit in zone emulation mode. Set the pseudo value based on max_segments as suggested in the commit c2ae7b772ef4 ("btrfs: zoned: revive max_zone_append_bytes"). Fixes: 7d7672bc5d10 ("btrfs: convert count_max_extents() to use fs_info->max_extent_size") Signed-off-by: Shin'ichiro Kawasaki <shinichiro.kawasaki@wdc.com> --- fs/btrfs/zoned.c | 15 ++++++++++++--- 1 file changed, 12 insertions(+), 3 deletions(-)