zonefs: Do not ignore explicit_open with active zone limit

A zoned device may have no limit on the number of open zones but may
have a limit on the number of active zones it can support. In such
case, the explicit_open mount option should not be ignored to ensure
that the open() system call activates the zone with an explicit zone
open command, thus guaranteeing that the zone can be written.

Enforce this by ignoring the explicit_open mount option only for
devices that have both the open and active zone limits equal to 0.

Fixes: 87c9ce3ffec9 ("zonefs: Add active seq file accounting")
Signed-off-by: Damien Le Moal <damien.lemoal@opensource.wdc.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>

diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index ecce84909ca17f38414923b4ca7e94b0f6e12b70..123464d2145aa8e25569e24ffe2f22e790f80d0c 100644 (file)
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -1085,7 +1085,8 @@ static int zonefs_seq_file_write_open(struct inode *inode)
 
                if (sbi->s_mount_opts & ZONEFS_MNTOPT_EXPLICIT_OPEN) {
 
-                       if (wro > sbi->s_max_wro_seq_files) {
+                       if (sbi->s_max_wro_seq_files
+                           && wro > sbi->s_max_wro_seq_files) {
                                atomic_dec(&sbi->s_wro_seq_files);
                                ret = -EBUSY;
                                goto unlock;
@@ -1785,8 +1786,10 @@ static int zonefs_fill_super(struct super_block *sb, void *data, int silent)
                    blkdev_nr_zones(sb->s_bdev->bd_disk));
 
        if (!sbi->s_max_wro_seq_files &&
+           !sbi->s_max_active_seq_files &&
            sbi->s_mount_opts & ZONEFS_MNTOPT_EXPLICIT_OPEN) {
-               zonefs_info(sb, "No open zones limit. Ignoring explicit_open mount option\n");
+               zonefs_info(sb,
+                       "No open and active zone limits. Ignoring explicit_open mount option\n");
                sbi->s_mount_opts &= ~ZONEFS_MNTOPT_EXPLICIT_OPEN;
        }