/*
* Copyright (C) 2005, 2006
- * Avishay Traeger (avishay@gmail.com) (avishay@il.ibm.com)
- * Copyright (C) 2005, 2006
- * International Business Machines
+ * Avishay Traeger (avishay@gmail.com)
* Copyright (C) 2008, 2009
* Boaz Harrosh <bharrosh@panasas.com>
*
#include "exofs.h"
-#ifdef CONFIG_EXOFS_DEBUG
-# define EXOFS_DEBUG_OBJ_ISIZE 1
-#endif
+#define EXOFS_DBGMSG2(M...) do {} while (0)
+
+enum { BIO_MAX_PAGES_KMALLOC =
+ (PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec),
+};
struct page_collect {
struct exofs_sb_info *sbi;
struct request_queue *req_q;
struct inode *inode;
unsigned expected_pages;
+ struct exofs_io_state *ios;
struct bio *bio;
unsigned nr_pages;
};
static void _pcol_init(struct page_collect *pcol, unsigned expected_pages,
- struct inode *inode)
+ struct inode *inode)
{
struct exofs_sb_info *sbi = inode->i_sb->s_fs_info;
- struct request_queue *req_q = sbi->s_dev->scsi_device->request_queue;
pcol->sbi = sbi;
- pcol->req_q = req_q;
+ /* Create master bios on first Q, later on cloning, each clone will be
+ * allocated on it's destination Q
+ */
+ pcol->req_q = osd_request_queue(sbi->s_ods[0]);
pcol->inode = inode;
pcol->expected_pages = expected_pages;
+ pcol->ios = NULL;
pcol->bio = NULL;
pcol->nr_pages = 0;
pcol->length = 0;
pcol->pg_first = -1;
-
- EXOFS_DBGMSG("_pcol_init ino=0x%lx expected_pages=%u\n", inode->i_ino,
- expected_pages);
}
static void _pcol_reset(struct page_collect *pcol)
pcol->nr_pages = 0;
pcol->length = 0;
pcol->pg_first = -1;
- EXOFS_DBGMSG("_pcol_reset ino=0x%lx expected_pages=%u\n",
- pcol->inode->i_ino, pcol->expected_pages);
+ pcol->ios = NULL;
/* this is probably the end of the loop but in writes
* it might not end here. don't be left with nothing
*/
if (!pcol->expected_pages)
- pcol->expected_pages = 128;
+ pcol->expected_pages = BIO_MAX_PAGES_KMALLOC;
}
static int pcol_try_alloc(struct page_collect *pcol)
{
- int pages = min_t(unsigned, pcol->expected_pages, BIO_MAX_PAGES);
+ int pages = min_t(unsigned, pcol->expected_pages,
+ BIO_MAX_PAGES_KMALLOC);
+
+ if (!pcol->ios) { /* First time allocate io_state */
+ int ret = exofs_get_io_state(pcol->sbi, &pcol->ios);
+
+ if (ret)
+ return ret;
+ }
for (; pages; pages >>= 1) {
- pcol->bio = bio_alloc(GFP_KERNEL, pages);
+ pcol->bio = bio_kmalloc(GFP_KERNEL, pages);
if (likely(pcol->bio))
return 0;
}
- EXOFS_ERR("Failed to kcalloc expected_pages=%u\n",
+ EXOFS_ERR("Failed to bio_kmalloc expected_pages=%u\n",
pcol->expected_pages);
return -ENOMEM;
}
static void pcol_free(struct page_collect *pcol)
{
- bio_put(pcol->bio);
- pcol->bio = NULL;
+ if (pcol->bio) {
+ bio_put(pcol->bio);
+ pcol->bio = NULL;
+ }
+
+ if (pcol->ios) {
+ exofs_put_io_state(pcol->ios);
+ pcol->ios = NULL;
+ }
}
static int pcol_add_page(struct page_collect *pcol, struct page *page,
/* Called at the end of reads, to optionally unlock pages and update their
* status.
*/
-static int __readpages_done(struct osd_request *or, struct page_collect *pcol,
- bool do_unlock)
+static int __readpages_done(struct page_collect *pcol, bool do_unlock)
{
struct bio_vec *bvec;
int i;
u64 resid;
u64 good_bytes;
u64 length = 0;
- int ret = exofs_check_ok_resid(or, &resid, NULL);
-
- osd_end_request(or);
+ int ret = exofs_check_io(pcol->ios, &resid);
if (likely(!ret))
good_bytes = pcol->length;
- else if (!resid)
- good_bytes = 0;
else
good_bytes = pcol->length - resid;
- EXOFS_DBGMSG("readpages_done(0x%lx) good_bytes=0x%llx"
+ EXOFS_DBGMSG2("readpages_done(0x%lx) good_bytes=0x%llx"
" length=0x%lx nr_pages=%u\n",
pcol->inode->i_ino, _LLU(good_bytes), pcol->length,
pcol->nr_pages);
else
page_stat = ret;
- EXOFS_DBGMSG(" readpages_done(0x%lx, 0x%lx) %s\n",
+ EXOFS_DBGMSG2(" readpages_done(0x%lx, 0x%lx) %s\n",
inode->i_ino, page->index,
page_stat ? "bad_bytes" : "good_bytes");
}
pcol_free(pcol);
- EXOFS_DBGMSG("readpages_done END\n");
+ EXOFS_DBGMSG2("readpages_done END\n");
return ret;
}
/* callback of async reads */
-static void readpages_done(struct osd_request *or, void *p)
+static void readpages_done(struct exofs_io_state *ios, void *p)
{
struct page_collect *pcol = p;
- __readpages_done(or, pcol, true);
+ __readpages_done(pcol, true);
atomic_dec(&pcol->sbi->s_curr_pending);
- kfree(p);
+ kfree(pcol);
}
static void _unlock_pcol_pages(struct page_collect *pcol, int ret, int rw)
unlock_page(page);
}
- pcol_free(pcol);
}
static int read_exec(struct page_collect *pcol, bool is_sync)
{
struct exofs_i_info *oi = exofs_i(pcol->inode);
- struct osd_obj_id obj = {pcol->sbi->s_pid,
- pcol->inode->i_ino + EXOFS_OBJ_OFF};
- struct osd_request *or = NULL;
+ struct exofs_io_state *ios = pcol->ios;
struct page_collect *pcol_copy = NULL;
- loff_t i_start = pcol->pg_first << PAGE_CACHE_SHIFT;
int ret;
if (!pcol->bio)
/* see comment in _readpage() about sync reads */
WARN_ON(is_sync && (pcol->nr_pages != 1));
- or = osd_start_request(pcol->sbi->s_dev, GFP_KERNEL);
- if (unlikely(!or)) {
- ret = -ENOMEM;
- goto err;
- }
-
- osd_req_read(or, &obj, pcol->bio, i_start);
+ ios->bio = pcol->bio;
+ ios->length = pcol->length;
+ ios->offset = pcol->pg_first << PAGE_CACHE_SHIFT;
if (is_sync) {
- exofs_sync_op(or, pcol->sbi->s_timeout, oi->i_cred);
- return __readpages_done(or, pcol, false);
+ exofs_oi_read(oi, pcol->ios);
+ return __readpages_done(pcol, false);
}
pcol_copy = kmalloc(sizeof(*pcol_copy), GFP_KERNEL);
}
*pcol_copy = *pcol;
- ret = exofs_async_op(or, readpages_done, pcol_copy, oi->i_cred);
+ ios->done = readpages_done;
+ ios->private = pcol_copy;
+ ret = exofs_oi_read(oi, ios);
if (unlikely(ret))
goto err;
atomic_inc(&pcol->sbi->s_curr_pending);
- EXOFS_DBGMSG("read_exec obj=0x%llx start=0x%llx length=0x%lx\n",
- obj.id, _LLU(i_start), pcol->length);
+ EXOFS_DBGMSG2("read_exec obj=0x%llx start=0x%llx length=0x%lx\n",
+ ios->obj.id, _LLU(ios->offset), pcol->length);
/* pages ownership was passed to pcol_copy */
_pcol_reset(pcol);
err:
if (!is_sync)
_unlock_pcol_pages(pcol, ret, READ);
+
+ pcol_free(pcol);
+
kfree(pcol_copy);
- if (or)
- osd_end_request(or);
return ret;
}
if (len != PAGE_CACHE_SIZE)
zero_user(page, len, PAGE_CACHE_SIZE - len);
- EXOFS_DBGMSG(" readpage_strip(0x%lx, 0x%lx) len=0x%zx\n",
+ EXOFS_DBGMSG2(" readpage_strip(0x%lx, 0x%lx) len=0x%zx\n",
inode->i_ino, page->index, len);
ret = pcol_add_page(pcol, page, len);
if (ret) {
- EXOFS_DBGMSG("Failed pcol_add_page pages[i]=%p "
+ EXOFS_DBGMSG2("Failed pcol_add_page pages[i]=%p "
"this_len=0x%zx nr_pages=%u length=0x%lx\n",
page, len, pcol->nr_pages, pcol->length);
_pcol_init(&pcol, 1, page->mapping->host);
- /* readpage_strip might call read_exec(,async) inside at several places
- * but this is safe for is_async=0 since read_exec will not do anything
- * when we have a single page.
+ /* readpage_strip might call read_exec(,is_sync==false) at several
+ * places but not if we have a single page.
*/
ret = readpage_strip(&pcol, page);
if (ret) {
return _readpage(page, false);
}
-/* Callback for osd_write. All writes are asynchronouse */
-static void writepages_done(struct osd_request *or, void *p)
+/* Callback for osd_write. All writes are asynchronous */
+static void writepages_done(struct exofs_io_state *ios, void *p)
{
struct page_collect *pcol = p;
struct bio_vec *bvec;
u64 resid;
u64 good_bytes;
u64 length = 0;
+ int ret = exofs_check_io(ios, &resid);
- int ret = exofs_check_ok_resid(or, NULL, &resid);
-
- osd_end_request(or);
atomic_dec(&pcol->sbi->s_curr_pending);
if (likely(!ret))
good_bytes = pcol->length;
- else if (!resid)
- good_bytes = 0;
else
good_bytes = pcol->length - resid;
- EXOFS_DBGMSG("writepages_done(0x%lx) good_bytes=0x%llx"
+ EXOFS_DBGMSG2("writepages_done(0x%lx) good_bytes=0x%llx"
" length=0x%lx nr_pages=%u\n",
pcol->inode->i_ino, _LLU(good_bytes), pcol->length,
pcol->nr_pages);
update_write_page(page, page_stat);
unlock_page(page);
- EXOFS_DBGMSG(" writepages_done(0x%lx, 0x%lx) status=%d\n",
+ EXOFS_DBGMSG2(" writepages_done(0x%lx, 0x%lx) status=%d\n",
inode->i_ino, page->index, page_stat);
length += bvec->bv_len;
pcol_free(pcol);
kfree(pcol);
- EXOFS_DBGMSG("writepages_done END\n");
+ EXOFS_DBGMSG2("writepages_done END\n");
}
static int write_exec(struct page_collect *pcol)
{
struct exofs_i_info *oi = exofs_i(pcol->inode);
- struct osd_obj_id obj = {pcol->sbi->s_pid,
- pcol->inode->i_ino + EXOFS_OBJ_OFF};
- struct osd_request *or = NULL;
+ struct exofs_io_state *ios = pcol->ios;
struct page_collect *pcol_copy = NULL;
- loff_t i_start = pcol->pg_first << PAGE_CACHE_SHIFT;
int ret;
if (!pcol->bio)
return 0;
- or = osd_start_request(pcol->sbi->s_dev, GFP_KERNEL);
- if (unlikely(!or)) {
- EXOFS_ERR("write_exec: Faild to osd_start_request()\n");
- ret = -ENOMEM;
- goto err;
- }
-
pcol_copy = kmalloc(sizeof(*pcol_copy), GFP_KERNEL);
if (!pcol_copy) {
EXOFS_ERR("write_exec: Faild to kmalloc(pcol)\n");
*pcol_copy = *pcol;
- osd_req_write(or, &obj, pcol_copy->bio, i_start);
- ret = exofs_async_op(or, writepages_done, pcol_copy, oi->i_cred);
+ pcol_copy->bio->bi_rw |= (1 << BIO_RW); /* FIXME: bio_set_dir() */
+
+ ios->bio = pcol_copy->bio;
+ ios->offset = pcol_copy->pg_first << PAGE_CACHE_SHIFT;
+ ios->length = pcol_copy->length;
+ ios->done = writepages_done;
+ ios->private = pcol_copy;
+
+ ret = exofs_oi_write(oi, ios);
if (unlikely(ret)) {
- EXOFS_ERR("write_exec: exofs_async_op() Faild\n");
+ EXOFS_ERR("write_exec: exofs_oi_write() Faild\n");
goto err;
}
atomic_inc(&pcol->sbi->s_curr_pending);
- EXOFS_DBGMSG("write_exec(0x%lx, 0x%llx) start=0x%llx length=0x%lx\n",
- pcol->inode->i_ino, pcol->pg_first, _LLU(i_start),
+ EXOFS_DBGMSG2("write_exec(0x%lx, 0x%llx) start=0x%llx length=0x%lx\n",
+ pcol->inode->i_ino, pcol->pg_first, _LLU(ios->offset),
pcol->length);
/* pages ownership was passed to pcol_copy */
_pcol_reset(pcol);
err:
_unlock_pcol_pages(pcol, ret, WRITE);
+ pcol_free(pcol);
kfree(pcol_copy);
- if (or)
- osd_end_request(or);
+
return ret;
}
if (PageError(page))
ClearPageError(page);
unlock_page(page);
+ EXOFS_DBGMSG("writepage_strip(0x%lx, 0x%lx) "
+ "outside the limits\n",
+ inode->i_ino, page->index);
return 0;
}
}
ret = write_exec(pcol);
if (unlikely(ret))
goto fail;
+
+ EXOFS_DBGMSG("writepage_strip(0x%lx, 0x%lx) Discontinuity\n",
+ inode->i_ino, page->index);
goto try_again;
}
goto fail;
}
- EXOFS_DBGMSG(" writepage_strip(0x%lx, 0x%lx) len=0x%zx\n",
+ EXOFS_DBGMSG2(" writepage_strip(0x%lx, 0x%lx) len=0x%zx\n",
inode->i_ino, page->index, len);
ret = pcol_add_page(pcol, page, len);
if (unlikely(ret)) {
- EXOFS_DBGMSG("Failed pcol_add_page "
+ EXOFS_DBGMSG2("Failed pcol_add_page "
"nr_pages=%u total_length=0x%lx\n",
pcol->nr_pages, pcol->length);
return 0;
fail:
+ EXOFS_DBGMSG("Error: writepage_strip(0x%lx, 0x%lx)=>%d\n",
+ inode->i_ino, page->index, ret);
set_bit(AS_EIO, &page->mapping->flags);
unlock_page(page);
return ret;
wbc->range_end >> PAGE_CACHE_SHIFT;
if (start || end)
- expected_pages = min(end - start + 1, 32L);
+ expected_pages = end - start + 1;
else
expected_pages = mapping->nrpages;
- EXOFS_DBGMSG("inode(0x%lx) wbc->start=0x%llx wbc->end=0x%llx"
- " m->nrpages=%lu start=0x%lx end=0x%lx\n",
+ if (expected_pages < 32L)
+ expected_pages = 32L;
+
+ EXOFS_DBGMSG2("inode(0x%lx) wbc->start=0x%llx wbc->end=0x%llx "
+ "nrpages=%lu start=0x%lx end=0x%lx expected_pages=%ld\n",
mapping->host->i_ino, wbc->range_start, wbc->range_end,
- mapping->nrpages, start, end);
+ mapping->nrpages, start, end, expected_pages);
_pcol_init(&pcol, expected_pages, mapping->host);
fsdata);
}
+static int exofs_write_end(struct file *file, struct address_space *mapping,
+ loff_t pos, unsigned len, unsigned copied,
+ struct page *page, void *fsdata)
+{
+ struct inode *inode = mapping->host;
+ /* According to comment in simple_write_end i_mutex is held */
+ loff_t i_size = inode->i_size;
+ int ret;
+
+ ret = simple_write_end(file, mapping,pos, len, copied, page, fsdata);
+ if (i_size != inode->i_size)
+ mark_inode_dirty(inode);
+ return ret;
+}
+
const struct address_space_operations exofs_aops = {
.readpage = exofs_readpage,
.readpages = exofs_readpages,
.writepage = exofs_writepage,
.writepages = exofs_writepages,
.write_begin = exofs_write_begin_export,
- .write_end = simple_write_end,
+ .write_end = exofs_write_end,
};
/******************************************************************************
const struct osd_attr g_attr_logical_length = ATTR_DEF(
OSD_APAGE_OBJECT_INFORMATION, OSD_ATTR_OI_LOGICAL_LENGTH, 8);
+static int _do_truncate(struct inode *inode)
+{
+ struct exofs_i_info *oi = exofs_i(inode);
+ loff_t isize = i_size_read(inode);
+ int ret;
+
+ inode->i_mtime = inode->i_ctime = CURRENT_TIME;
+
+ nobh_truncate_page(inode->i_mapping, isize, exofs_get_block);
+
+ ret = exofs_oi_truncate(oi, (u64)isize);
+ EXOFS_DBGMSG("(0x%lx) size=0x%llx\n", inode->i_ino, isize);
+ return ret;
+}
+
/*
* Truncate a file to the specified size - all we have to do is set the size
* attribute. We make sure the object exists first.
*/
void exofs_truncate(struct inode *inode)
{
- struct exofs_sb_info *sbi = inode->i_sb->s_fs_info;
struct exofs_i_info *oi = exofs_i(inode);
- struct osd_obj_id obj = {sbi->s_pid, inode->i_ino + EXOFS_OBJ_OFF};
- struct osd_request *or;
- struct osd_attr attr;
- loff_t isize = i_size_read(inode);
- __be64 newsize;
int ret;
if (!(S_ISREG(inode->i_mode) || S_ISDIR(inode->i_mode)
return;
if (IS_APPEND(inode) || IS_IMMUTABLE(inode))
return;
- inode->i_mtime = inode->i_ctime = CURRENT_TIME;
-
- nobh_truncate_page(inode->i_mapping, isize, exofs_get_block);
-
- or = osd_start_request(sbi->s_dev, GFP_KERNEL);
- if (unlikely(!or)) {
- EXOFS_ERR("ERROR: exofs_truncate: osd_start_request failed\n");
- goto fail;
- }
-
- osd_req_set_attributes(or, &obj);
-
- newsize = cpu_to_be64((u64)isize);
- attr = g_attr_logical_length;
- attr.val_ptr = &newsize;
- osd_req_add_set_attr_list(or, &attr, 1);
/* if we are about to truncate an object, and it hasn't been
* created yet, wait
if (unlikely(wait_obj_created(oi)))
goto fail;
- ret = exofs_sync_op(or, sbi->s_timeout, oi->i_cred);
- osd_end_request(or);
+ ret = _do_truncate(inode);
if (ret)
goto fail;
error = inode_setattr(inode, iattr);
return error;
}
+
+/*
+ * Read an inode from the OSD, and return it as is. We also return the size
+ * attribute in the 'obj_size' argument.
+ */
+static int exofs_get_inode(struct super_block *sb, struct exofs_i_info *oi,
+ struct exofs_fcb *inode, uint64_t *obj_size)
+{
+ struct exofs_sb_info *sbi = sb->s_fs_info;
+ struct osd_attr attrs[2];
+ struct exofs_io_state *ios;
+ int ret;
+
+ *obj_size = ~0;
+ ret = exofs_get_io_state(sbi, &ios);
+ if (unlikely(ret)) {
+ EXOFS_ERR("%s: exofs_get_io_state failed.\n", __func__);
+ return ret;
+ }
+
+ ios->obj.id = exofs_oi_objno(oi);
+ exofs_make_credential(oi->i_cred, &ios->obj);
+ ios->cred = oi->i_cred;
+
+ attrs[0] = g_attr_inode_data;
+ attrs[1] = g_attr_logical_length;
+ ios->in_attr = attrs;
+ ios->in_attr_len = ARRAY_SIZE(attrs);
+
+ ret = exofs_sbi_read(ios);
+ if (ret)
+ goto out;
+
+ ret = extract_attr_from_ios(ios, &attrs[0]);
+ if (ret) {
+ EXOFS_ERR("%s: extract_attr of inode_data failed\n", __func__);
+ goto out;
+ }
+ WARN_ON(attrs[0].len != EXOFS_INO_ATTR_SIZE);
+ memcpy(inode, attrs[0].val_ptr, EXOFS_INO_ATTR_SIZE);
+
+ ret = extract_attr_from_ios(ios, &attrs[1]);
+ if (ret) {
+ EXOFS_ERR("%s: extract_attr of logical_length failed\n",
+ __func__);
+ goto out;
+ }
+ *obj_size = get_unaligned_be64(attrs[1].val_ptr);
+
+out:
+ exofs_put_io_state(ios);
+ return ret;
+}
+
+static void __oi_init(struct exofs_i_info *oi)
+{
+ init_waitqueue_head(&oi->i_wq);
+ oi->i_flags = 0;
+}
+/*
+ * Fill in an inode read from the OSD and set it up for use
+ */
+struct inode *exofs_iget(struct super_block *sb, unsigned long ino)
+{
+ struct exofs_i_info *oi;
+ struct exofs_fcb fcb;
+ struct inode *inode;
+ uint64_t obj_size;
+ int ret;
+
+ inode = iget_locked(sb, ino);
+ if (!inode)
+ return ERR_PTR(-ENOMEM);
+ if (!(inode->i_state & I_NEW))
+ return inode;
+ oi = exofs_i(inode);
+ __oi_init(oi);
+
+ /* read the inode from the osd */
+ ret = exofs_get_inode(sb, oi, &fcb, &obj_size);
+ if (ret)
+ goto bad_inode;
+
+ set_obj_created(oi);
+
+ /* copy stuff from on-disk struct to in-memory struct */
+ inode->i_mode = le16_to_cpu(fcb.i_mode);
+ inode->i_uid = le32_to_cpu(fcb.i_uid);
+ inode->i_gid = le32_to_cpu(fcb.i_gid);
+ inode->i_nlink = le16_to_cpu(fcb.i_links_count);
+ inode->i_ctime.tv_sec = (signed)le32_to_cpu(fcb.i_ctime);
+ inode->i_atime.tv_sec = (signed)le32_to_cpu(fcb.i_atime);
+ inode->i_mtime.tv_sec = (signed)le32_to_cpu(fcb.i_mtime);
+ inode->i_ctime.tv_nsec =
+ inode->i_atime.tv_nsec = inode->i_mtime.tv_nsec = 0;
+ oi->i_commit_size = le64_to_cpu(fcb.i_size);
+ i_size_write(inode, oi->i_commit_size);
+ inode->i_blkbits = EXOFS_BLKSHIFT;
+ inode->i_generation = le32_to_cpu(fcb.i_generation);
+
+ if ((inode->i_size != obj_size) &&
+ (!exofs_inode_is_fast_symlink(inode))) {
+ EXOFS_ERR("WARNING: Size of inode=%llu != object=%llu\n",
+ inode->i_size, _LLU(obj_size));
+ /* FIXME: call exofs_inode_recovery() */
+ }
+
+ oi->i_dir_start_lookup = 0;
+
+ if ((inode->i_nlink == 0) && (inode->i_mode == 0)) {
+ ret = -ESTALE;
+ goto bad_inode;
+ }
+
+ if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) {
+ if (fcb.i_data[0])
+ inode->i_rdev =
+ old_decode_dev(le32_to_cpu(fcb.i_data[0]));
+ else
+ inode->i_rdev =
+ new_decode_dev(le32_to_cpu(fcb.i_data[1]));
+ } else {
+ memcpy(oi->i_data, fcb.i_data, sizeof(fcb.i_data));
+ }
+
+ if (S_ISREG(inode->i_mode)) {
+ inode->i_op = &exofs_file_inode_operations;
+ inode->i_fop = &exofs_file_operations;
+ inode->i_mapping->a_ops = &exofs_aops;
+ } else if (S_ISDIR(inode->i_mode)) {
+ inode->i_op = &exofs_dir_inode_operations;
+ inode->i_fop = &exofs_dir_operations;
+ inode->i_mapping->a_ops = &exofs_aops;
+ } else if (S_ISLNK(inode->i_mode)) {
+ if (exofs_inode_is_fast_symlink(inode))
+ inode->i_op = &exofs_fast_symlink_inode_operations;
+ else {
+ inode->i_op = &exofs_symlink_inode_operations;
+ inode->i_mapping->a_ops = &exofs_aops;
+ }
+ } else {
+ inode->i_op = &exofs_special_inode_operations;
+ if (fcb.i_data[0])
+ init_special_inode(inode, inode->i_mode,
+ old_decode_dev(le32_to_cpu(fcb.i_data[0])));
+ else
+ init_special_inode(inode, inode->i_mode,
+ new_decode_dev(le32_to_cpu(fcb.i_data[1])));
+ }
+
+ unlock_new_inode(inode);
+ return inode;
+
+bad_inode:
+ iget_failed(inode);
+ return ERR_PTR(ret);
+}
+
+int __exofs_wait_obj_created(struct exofs_i_info *oi)
+{
+ if (!obj_created(oi)) {
+ BUG_ON(!obj_2bcreated(oi));
+ wait_event(oi->i_wq, obj_created(oi));
+ }
+ return unlikely(is_bad_inode(&oi->vfs_inode)) ? -EIO : 0;
+}
+/*
+ * Callback function from exofs_new_inode(). The important thing is that we
+ * set the obj_created flag so that other methods know that the object exists on
+ * the OSD.
+ */
+static void create_done(struct exofs_io_state *ios, void *p)
+{
+ struct inode *inode = p;
+ struct exofs_i_info *oi = exofs_i(inode);
+ struct exofs_sb_info *sbi = inode->i_sb->s_fs_info;
+ int ret;
+
+ ret = exofs_check_io(ios, NULL);
+ exofs_put_io_state(ios);
+
+ atomic_dec(&sbi->s_curr_pending);
+
+ if (unlikely(ret)) {
+ EXOFS_ERR("object=0x%llx creation faild in pid=0x%llx",
+ _LLU(exofs_oi_objno(oi)), _LLU(sbi->s_pid));
+ /*TODO: When FS is corrupted creation can fail, object already
+ * exist. Get rid of this asynchronous creation, if exist
+ * increment the obj counter and try the next object. Until we
+ * succeed. All these dangling objects will be made into lost
+ * files by chkfs.exofs
+ */
+ }
+
+ set_obj_created(oi);
+
+ atomic_dec(&inode->i_count);
+ wake_up(&oi->i_wq);
+}
+
+/*
+ * Set up a new inode and create an object for it on the OSD
+ */
+struct inode *exofs_new_inode(struct inode *dir, int mode)
+{
+ struct super_block *sb;
+ struct inode *inode;
+ struct exofs_i_info *oi;
+ struct exofs_sb_info *sbi;
+ struct exofs_io_state *ios;
+ int ret;
+
+ sb = dir->i_sb;
+ inode = new_inode(sb);
+ if (!inode)
+ return ERR_PTR(-ENOMEM);
+
+ oi = exofs_i(inode);
+ __oi_init(oi);
+
+ set_obj_2bcreated(oi);
+
+ sbi = sb->s_fs_info;
+
+ sb->s_dirt = 1;
+ inode->i_uid = current->cred->fsuid;
+ if (dir->i_mode & S_ISGID) {
+ inode->i_gid = dir->i_gid;
+ if (S_ISDIR(mode))
+ mode |= S_ISGID;
+ } else {
+ inode->i_gid = current->cred->fsgid;
+ }
+ inode->i_mode = mode;
+
+ inode->i_ino = sbi->s_nextid++;
+ inode->i_blkbits = EXOFS_BLKSHIFT;
+ inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME;
+ oi->i_commit_size = inode->i_size = 0;
+ spin_lock(&sbi->s_next_gen_lock);
+ inode->i_generation = sbi->s_next_generation++;
+ spin_unlock(&sbi->s_next_gen_lock);
+ insert_inode_hash(inode);
+
+ mark_inode_dirty(inode);
+
+ ret = exofs_get_io_state(sbi, &ios);
+ if (unlikely(ret)) {
+ EXOFS_ERR("exofs_new_inode: exofs_get_io_state failed\n");
+ return ERR_PTR(ret);
+ }
+
+ ios->obj.id = exofs_oi_objno(oi);
+ exofs_make_credential(oi->i_cred, &ios->obj);
+
+ /* increment the refcount so that the inode will still be around when we
+ * reach the callback
+ */
+ atomic_inc(&inode->i_count);
+
+ ios->done = create_done;
+ ios->private = inode;
+ ios->cred = oi->i_cred;
+ ret = exofs_sbi_create(ios);
+ if (ret) {
+ atomic_dec(&inode->i_count);
+ exofs_put_io_state(ios);
+ return ERR_PTR(ret);
+ }
+ atomic_inc(&sbi->s_curr_pending);
+
+ return inode;
+}
+
+/*
+ * struct to pass two arguments to update_inode's callback
+ */
+struct updatei_args {
+ struct exofs_sb_info *sbi;
+ struct exofs_fcb fcb;
+};
+
+/*
+ * Callback function from exofs_update_inode().
+ */
+static void updatei_done(struct exofs_io_state *ios, void *p)
+{
+ struct updatei_args *args = p;
+
+ exofs_put_io_state(ios);
+
+ atomic_dec(&args->sbi->s_curr_pending);
+
+ kfree(args);
+}
+
+/*
+ * Write the inode to the OSD. Just fill up the struct, and set the attribute
+ * synchronously or asynchronously depending on the do_sync flag.
+ */
+static int exofs_update_inode(struct inode *inode, int do_sync)
+{
+ struct exofs_i_info *oi = exofs_i(inode);
+ struct super_block *sb = inode->i_sb;
+ struct exofs_sb_info *sbi = sb->s_fs_info;
+ struct exofs_io_state *ios;
+ struct osd_attr attr;
+ struct exofs_fcb *fcb;
+ struct updatei_args *args;
+ int ret;
+
+ args = kzalloc(sizeof(*args), GFP_KERNEL);
+ if (!args) {
+ EXOFS_DBGMSG("Faild kzalloc of args\n");
+ return -ENOMEM;
+ }
+
+ fcb = &args->fcb;
+
+ fcb->i_mode = cpu_to_le16(inode->i_mode);
+ fcb->i_uid = cpu_to_le32(inode->i_uid);
+ fcb->i_gid = cpu_to_le32(inode->i_gid);
+ fcb->i_links_count = cpu_to_le16(inode->i_nlink);
+ fcb->i_ctime = cpu_to_le32(inode->i_ctime.tv_sec);
+ fcb->i_atime = cpu_to_le32(inode->i_atime.tv_sec);
+ fcb->i_mtime = cpu_to_le32(inode->i_mtime.tv_sec);
+ oi->i_commit_size = i_size_read(inode);
+ fcb->i_size = cpu_to_le64(oi->i_commit_size);
+ fcb->i_generation = cpu_to_le32(inode->i_generation);
+
+ if (S_ISCHR(inode->i_mode) || S_ISBLK(inode->i_mode)) {
+ if (old_valid_dev(inode->i_rdev)) {
+ fcb->i_data[0] =
+ cpu_to_le32(old_encode_dev(inode->i_rdev));
+ fcb->i_data[1] = 0;
+ } else {
+ fcb->i_data[0] = 0;
+ fcb->i_data[1] =
+ cpu_to_le32(new_encode_dev(inode->i_rdev));
+ fcb->i_data[2] = 0;
+ }
+ } else
+ memcpy(fcb->i_data, oi->i_data, sizeof(fcb->i_data));
+
+ ret = exofs_get_io_state(sbi, &ios);
+ if (unlikely(ret)) {
+ EXOFS_ERR("%s: exofs_get_io_state failed.\n", __func__);
+ goto free_args;
+ }
+
+ attr = g_attr_inode_data;
+ attr.val_ptr = fcb;
+ ios->out_attr_len = 1;
+ ios->out_attr = &attr;
+
+ if (!obj_created(oi)) {
+ EXOFS_DBGMSG("!obj_created\n");
+ BUG_ON(!obj_2bcreated(oi));
+ wait_event(oi->i_wq, obj_created(oi));
+ EXOFS_DBGMSG("wait_event done\n");
+ }
+
+ if (!do_sync) {
+ args->sbi = sbi;
+ ios->done = updatei_done;
+ ios->private = args;
+ }
+
+ ret = exofs_oi_write(oi, ios);
+ if (!do_sync && !ret) {
+ atomic_inc(&sbi->s_curr_pending);
+ goto out; /* deallocation in updatei_done */
+ }
+
+ exofs_put_io_state(ios);
+free_args:
+ kfree(args);
+out:
+ EXOFS_DBGMSG("(0x%lx) do_sync=%d ret=>%d\n",
+ inode->i_ino, do_sync, ret);
+ return ret;
+}
+
+int exofs_write_inode(struct inode *inode, int wait)
+{
+ return exofs_update_inode(inode, wait);
+}
+
+/*
+ * Callback function from exofs_delete_inode() - don't have much cleaning up to
+ * do.
+ */
+static void delete_done(struct exofs_io_state *ios, void *p)
+{
+ struct exofs_sb_info *sbi = p;
+
+ exofs_put_io_state(ios);
+
+ atomic_dec(&sbi->s_curr_pending);
+}
+
+/*
+ * Called when the refcount of an inode reaches zero. We remove the object
+ * from the OSD here. We make sure the object was created before we try and
+ * delete it.
+ */
+void exofs_delete_inode(struct inode *inode)
+{
+ struct exofs_i_info *oi = exofs_i(inode);
+ struct super_block *sb = inode->i_sb;
+ struct exofs_sb_info *sbi = sb->s_fs_info;
+ struct exofs_io_state *ios;
+ int ret;
+
+ truncate_inode_pages(&inode->i_data, 0);
+
+ if (is_bad_inode(inode))
+ goto no_delete;
+
+ mark_inode_dirty(inode);
+ exofs_update_inode(inode, inode_needs_sync(inode));
+
+ inode->i_size = 0;
+ if (inode->i_blocks)
+ exofs_truncate(inode);
+
+ clear_inode(inode);
+
+ ret = exofs_get_io_state(sbi, &ios);
+ if (unlikely(ret)) {
+ EXOFS_ERR("%s: exofs_get_io_state failed\n", __func__);
+ return;
+ }
+
+ /* if we are deleting an obj that hasn't been created yet, wait */
+ if (!obj_created(oi)) {
+ BUG_ON(!obj_2bcreated(oi));
+ wait_event(oi->i_wq, obj_created(oi));
+ }
+
+ ios->obj.id = exofs_oi_objno(oi);
+ ios->done = delete_done;
+ ios->private = sbi;
+ ios->cred = oi->i_cred;
+ ret = exofs_sbi_remove(ios);
+ if (ret) {
+ EXOFS_ERR("%s: exofs_sbi_remove failed\n", __func__);
+ exofs_put_io_state(ios);
+ return;
+ }
+ atomic_inc(&sbi->s_curr_pending);
+
+ return;
+
+no_delete:
+ clear_inode(inode);
+}
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff