exofs: Move all operations to an io_engine
In anticipation for multi-device operations, we separate osd operations
into an abstract I/O API. Currently only one device is used but later
when adding more devices, we will drive all devices in parallel according
to a "data_map" that describes how data is arranged on multiple devices.
The file system level operates, like before, as if there is one object
(inode-number) and an i_size. The io engine will split this to the same
object-number but on multiple device.
At first we introduce Mirror (raid 1) layout. But at the final outcome
we intend to fully implement the pNFS-Objects data-map, including
raid 0,4,5,6 over mirrored devices, over multiple device-groups. And
more. See: http://tools.ietf.org/html/draft-ietf-nfsv4-pnfs-obj-12
* Define an io_state based API for accessing osd storage devices
in an abstract way.
Usage:
First a caller allocates an io state with:
exofs_get_io_state(struct exofs_sb_info *sbi,
struct exofs_io_state** ios);
Then calles one of:
exofs_sbi_create(struct exofs_io_state *ios);
exofs_sbi_remove(struct exofs_io_state *ios);
exofs_sbi_write(struct exofs_io_state *ios);
exofs_sbi_read(struct exofs_io_state *ios);
exofs_oi_truncate(struct exofs_i_info *oi, u64 new_len);
And when done
exofs_put_io_state(struct exofs_io_state *ios);
* Convert all source files to use this new API
* Convert from bio_alloc to bio_kmalloc
* In io engine we make use of the now fixed osd_req_decode_sense
There are no functional changes or on disk additions after this patch.
Signed-off-by: Boaz Harrosh <bharrosh@panasas.com>
diff --git a/fs/exofs/super.c b/fs/exofs/super.c
index 28add3e..4cd97f5 100644
--- a/fs/exofs/super.c
+++ b/fs/exofs/super.c
@@ -203,49 +203,40 @@
{
struct exofs_sb_info *sbi;
struct exofs_fscb *fscb;
- struct osd_request *or;
- struct osd_obj_id obj;
+ struct exofs_io_state *ios;
int ret = -ENOMEM;
- fscb = kzalloc(sizeof(struct exofs_fscb), GFP_KERNEL);
- if (!fscb) {
- EXOFS_ERR("exofs_write_super: memory allocation failed.\n");
- return -ENOMEM;
- }
-
lock_super(sb);
sbi = sb->s_fs_info;
+ fscb = &sbi->s_fscb;
+
+ ret = exofs_get_io_state(sbi, &ios);
+ if (ret)
+ goto out;
+
+ ios->length = sizeof(*fscb);
+ memset(fscb, 0, ios->length);
fscb->s_nextid = cpu_to_le64(sbi->s_nextid);
fscb->s_numfiles = cpu_to_le32(sbi->s_numfiles);
fscb->s_magic = cpu_to_le16(sb->s_magic);
fscb->s_newfs = 0;
- or = osd_start_request(sbi->s_dev, GFP_KERNEL);
- if (unlikely(!or)) {
- EXOFS_ERR("exofs_write_super: osd_start_request failed.\n");
- goto out;
- }
+ ios->obj.id = EXOFS_SUPER_ID;
+ ios->offset = 0;
+ ios->kern_buff = fscb;
+ ios->cred = sbi->s_cred;
- obj.partition = sbi->s_pid;
- obj.id = EXOFS_SUPER_ID;
- ret = osd_req_write_kern(or, &obj, 0, fscb, sizeof(*fscb));
+ ret = exofs_sbi_write(ios);
if (unlikely(ret)) {
- EXOFS_ERR("exofs_write_super: osd_req_write_kern failed.\n");
- goto out;
- }
-
- ret = exofs_sync_op(or, sbi->s_timeout, sbi->s_cred);
- if (unlikely(ret)) {
- EXOFS_ERR("exofs_write_super: exofs_sync_op failed.\n");
+ EXOFS_ERR("%s: exofs_sbi_write failed.\n", __func__);
goto out;
}
sb->s_dirt = 0;
out:
- if (or)
- osd_end_request(or);
+ EXOFS_DBGMSG("s_nextid=0x%llx ret=%d\n", _LLU(sbi->s_nextid), ret);
+ exofs_put_io_state(ios);
unlock_super(sb);
- kfree(fscb);
return ret;
}
@@ -302,24 +293,23 @@
struct inode *root;
struct exofs_mountopt *opts = data;
struct exofs_sb_info *sbi; /*extended info */
+ struct osd_dev *od; /* Master device */
struct exofs_fscb fscb; /*on-disk superblock info */
- struct osd_request *or = NULL;
struct osd_obj_id obj;
int ret;
sbi = kzalloc(sizeof(*sbi), GFP_KERNEL);
if (!sbi)
return -ENOMEM;
- sb->s_fs_info = sbi;
/* use mount options to fill superblock */
- sbi->s_dev = osduld_path_lookup(opts->dev_name);
- if (IS_ERR(sbi->s_dev)) {
- ret = PTR_ERR(sbi->s_dev);
- sbi->s_dev = NULL;
+ od = osduld_path_lookup(opts->dev_name);
+ if (IS_ERR(od)) {
+ ret = PTR_ERR(od);
goto free_sbi;
}
+ sbi->s_dev = od;
sbi->s_pid = opts->pid;
sbi->s_timeout = opts->timeout;
@@ -333,35 +323,13 @@
sb->s_bdev = NULL;
sb->s_dev = 0;
- /* read data from on-disk superblock object */
obj.partition = sbi->s_pid;
obj.id = EXOFS_SUPER_ID;
exofs_make_credential(sbi->s_cred, &obj);
- or = osd_start_request(sbi->s_dev, GFP_KERNEL);
- if (unlikely(!or)) {
- if (!silent)
- EXOFS_ERR(
- "exofs_fill_super: osd_start_request failed.\n");
- ret = -ENOMEM;
+ ret = exofs_read_kern(od, sbi->s_cred, &obj, 0, &fscb, sizeof(fscb));
+ if (unlikely(ret))
goto free_sbi;
- }
- ret = osd_req_read_kern(or, &obj, 0, &fscb, sizeof(fscb));
- if (unlikely(ret)) {
- if (!silent)
- EXOFS_ERR(
- "exofs_fill_super: osd_req_read_kern failed.\n");
- ret = -ENOMEM;
- goto free_sbi;
- }
-
- ret = exofs_sync_op(or, sbi->s_timeout, sbi->s_cred);
- if (unlikely(ret)) {
- if (!silent)
- EXOFS_ERR("exofs_fill_super: exofs_sync_op failed.\n");
- ret = -EIO;
- goto free_sbi;
- }
sb->s_magic = le16_to_cpu(fscb.s_magic);
sbi->s_nextid = le64_to_cpu(fscb.s_nextid);
@@ -380,6 +348,7 @@
spin_lock_init(&sbi->s_next_gen_lock);
/* set up operation vectors */
+ sb->s_fs_info = sbi;
sb->s_op = &exofs_sops;
sb->s_export_op = &exofs_export_ops;
root = exofs_iget(sb, EXOFS_ROOT_ID - EXOFS_OBJ_OFF);
@@ -406,16 +375,14 @@
}
_exofs_print_device("Mounting", opts->dev_name, sbi->s_dev, sbi->s_pid);
- ret = 0;
-out:
- if (or)
- osd_end_request(or);
- return ret;
+ return 0;
free_sbi:
+ EXOFS_ERR("Unable to mount exofs on %s pid=0x%llx err=%d\n",
+ opts->dev_name, sbi->s_pid, ret);
osduld_put_device(sbi->s_dev); /* NULL safe */
kfree(sbi);
- goto out;
+ return ret;
}
/*
@@ -444,7 +411,7 @@
{
struct super_block *sb = dentry->d_sb;
struct exofs_sb_info *sbi = sb->s_fs_info;
- struct osd_obj_id obj = {sbi->s_pid, 0};
+ struct exofs_io_state *ios;
struct osd_attr attrs[] = {
ATTR_DEF(OSD_APAGE_PARTITION_QUOTAS,
OSD_ATTR_PQ_CAPACITY_QUOTA, sizeof(__be64)),
@@ -453,26 +420,25 @@
};
uint64_t capacity = ULLONG_MAX;
uint64_t used = ULLONG_MAX;
- struct osd_request *or;
uint8_t cred_a[OSD_CAP_LEN];
int ret;
- /* get used/capacity attributes */
- exofs_make_credential(cred_a, &obj);
-
- or = osd_start_request(sbi->s_dev, GFP_KERNEL);
- if (unlikely(!or)) {
- EXOFS_DBGMSG("exofs_statfs: osd_start_request failed.\n");
- return -ENOMEM;
+ ret = exofs_get_io_state(sbi, &ios);
+ if (ret) {
+ EXOFS_DBGMSG("exofs_get_io_state failed.\n");
+ return ret;
}
- osd_req_get_attributes(or, &obj);
- osd_req_add_get_attr_list(or, attrs, ARRAY_SIZE(attrs));
- ret = exofs_sync_op(or, sbi->s_timeout, cred_a);
+ exofs_make_credential(cred_a, &ios->obj);
+ ios->cred = sbi->s_cred;
+ ios->in_attr = attrs;
+ ios->in_attr_len = ARRAY_SIZE(attrs);
+
+ ret = exofs_sbi_read(ios);
if (unlikely(ret))
goto out;
- ret = extract_attr_from_req(or, &attrs[0]);
+ ret = extract_attr_from_ios(ios, &attrs[0]);
if (likely(!ret)) {
capacity = get_unaligned_be64(attrs[0].val_ptr);
if (unlikely(!capacity))
@@ -480,7 +446,7 @@
} else
EXOFS_DBGMSG("exofs_statfs: get capacity failed.\n");
- ret = extract_attr_from_req(or, &attrs[1]);
+ ret = extract_attr_from_ios(ios, &attrs[1]);
if (likely(!ret))
used = get_unaligned_be64(attrs[1].val_ptr);
else
@@ -497,7 +463,7 @@
buf->f_namelen = EXOFS_NAME_LEN;
out:
- osd_end_request(or);
+ exofs_put_io_state(ios);
return ret;
}