Merge branch 'for-linus' of git://git.open-osd.org/linux-open-osd
Merge commit. Changes below are compared with the first parent: dfa4a423.
Changes
11 files changed · +1583 −374
| File | Status | Add | Del | Content |
|---|---|---|---|---|
| drivers/scsi/osd/Kconfig | modified | +0 | −4 | View file |
| fs/Makefile | modified | +1 | −1 | View file |
| fs/exofs/Kbuild | modified | +2 | −1 | View file |
| fs/exofs/Kconfig | modified | +8 | −1 | View file |
| fs/exofs/exofs.h | modified | +14 | −12 | View file |
| fs/exofs/inode.c | modified | +183 | −50 | View file |
| fs/exofs/ore.c | modified | +473 | −183 | View file |
| fs/exofs/ore_raid.c | added | +660 | −0 | View file |
| fs/exofs/ore_raid.h | added | +79 | −0 | View file |
| fs/exofs/super.c | modified | +86 | −119 | View file |
| include/scsi/osd_ore.h | modified | +77 | −3 | View file |
Only the first 600 diff lines are shown. Download the complete patch.
drivers/scsi/osd/Kconfig
View filediff --git a/drivers/scsi/osd/Kconfig b/drivers/scsi/osd/Kconfig
index 861b5ce..a070351 100644
--- a/drivers/scsi/osd/Kconfig
+++ b/drivers/scsi/osd/Kconfig
@@ -11,10 +11,6 @@
# it under the terms of the GNU General Public version 2 License as
# published by the Free Software Foundation
#
-# FIXME: SCSI_OSD_INITIATOR should select CONFIG (HMAC) SHA1 somehow.
-# How is it done properly?
-#
-
config SCSI_OSD_INITIATOR
tristate "OSD-Initiator library"
depends on SCSI
fs/Makefile
View filediff --git a/fs/Makefile b/fs/Makefile
index afc1096..5c30a13 100644
--- a/fs/Makefile
+++ b/fs/Makefile
@@ -120,6 +120,6 @@ obj-$(CONFIG_DEBUG_FS) += debugfs/
obj-$(CONFIG_OCFS2_FS) += ocfs2/
obj-$(CONFIG_BTRFS_FS) += btrfs/
obj-$(CONFIG_GFS2_FS) += gfs2/
-obj-$(CONFIG_EXOFS_FS) += exofs/
+obj-$(y) += exofs/ # Multiple mods, used by nfs/objlayout
obj-$(CONFIG_CEPH_FS) += ceph/
obj-$(CONFIG_PSTORE) += pstore/
fs/exofs/Kbuild
View filediff --git a/fs/exofs/Kbuild b/fs/exofs/Kbuild
index c5a5855..352ba14 100644
--- a/fs/exofs/Kbuild
+++ b/fs/exofs/Kbuild
@@ -13,7 +13,8 @@
#
# ore module library
-obj-$(CONFIG_ORE) += ore.o
+libore-y := ore.o ore_raid.o
+obj-$(CONFIG_ORE) += libore.o
exofs-y := inode.o file.o symlink.o namei.o dir.o super.o
obj-$(CONFIG_EXOFS_FS) += exofs.o
fs/exofs/Kconfig
View filediff --git a/fs/exofs/Kconfig b/fs/exofs/Kconfig
index 70bae41..fa9a286 100644
--- a/fs/exofs/Kconfig
+++ b/fs/exofs/Kconfig
@@ -1,10 +1,17 @@
+# Note ORE needs to "select ASYNC_XOR". So Not to force multiple selects
+# for every ORE user we do it like this. Any user should add itself here
+# at the "depends on EXOFS_FS || ..." with an ||. The dependencies are
+# selected here, and we default to "ON". So in effect it is like been
+# selected by any of the users.
config ORE
tristate
+ depends on EXOFS_FS
+ select ASYNC_XOR
+ default SCSI_OSD_ULD
config EXOFS_FS
tristate "exofs: OSD based file system support"
depends on SCSI_OSD_ULD
- select ORE
help
EXOFS is a file system that uses an OSD storage device,
as its backing storage.
fs/exofs/exofs.h
View filediff --git a/fs/exofs/exofs.h b/fs/exofs/exofs.h
index f4e442e..51f4b4c 100644
--- a/fs/exofs/exofs.h
+++ b/fs/exofs/exofs.h
@@ -53,6 +53,10 @@
/* u64 has problems with printk this will cast it to unsigned long long */
#define _LLU(x) (unsigned long long)(x)
+struct exofs_dev {
+ struct ore_dev ored;
+ unsigned did;
+};
/*
* our extension to the in-memory superblock
*/
@@ -66,13 +70,9 @@ struct exofs_sb_info {
u32 s_next_generation; /* next gen # to use */
atomic_t s_curr_pending; /* number of pending commands */
- struct pnfs_osd_data_map data_map; /* Default raid to use
- * FIXME: Needed ?
- */
struct ore_layout layout; /* Default files layout */
struct ore_comp one_comp; /* id & cred of partition id=0*/
- struct ore_components comps; /* comps for the partition */
- struct osd_dev *_min_one_dev[1]; /* Place holder for one dev */
+ struct ore_components oc; /* comps for the partition */
};
/*
@@ -86,7 +86,7 @@ struct exofs_i_info {
uint32_t i_dir_start_lookup; /* which page to start lookup */
uint64_t i_commit_size; /* the object's written length */
struct ore_comp one_comp; /* same component for all devices */
- struct ore_components comps; /* inode view of the device table */
+ struct ore_components oc; /* inode view of the device table */
};
static inline osd_id exofs_oi_objno(struct exofs_i_info *oi)
@@ -207,7 +207,7 @@ extern const struct inode_operations exofs_fast_symlink_inode_operations;
* bigger and that the device table repeats twice.
* See: exofs_read_lookup_dev_table()
*/
-static inline void exofs_init_comps(struct ore_components *comps,
+static inline void exofs_init_comps(struct ore_components *oc,
struct ore_comp *one_comp,
struct exofs_sb_info *sbi, osd_id oid)
{
@@ -217,13 +217,15 @@ static inline void exofs_init_comps(struct ore_components *comps,
one_comp->obj.id = oid;
exofs_make_credential(one_comp->cred, &one_comp->obj);
- comps->numdevs = sbi->comps.numdevs;
- comps->single_comp = EC_SINGLE_COMP;
- comps->comps = one_comp;
+ oc->first_dev = 0;
+ oc->numdevs = sbi->layout.group_width * sbi->layout.mirrors_p1 *
+ sbi->layout.group_count;
+ oc->single_comp = EC_SINGLE_COMP;
+ oc->comps = one_comp;
/* Round robin device view of the table */
- first_dev = (dev_mod * sbi->layout.mirrors_p1) % sbi->comps.numdevs;
- comps->ods = sbi->comps.ods + first_dev;
+ first_dev = (dev_mod * sbi->layout.mirrors_p1) % sbi->oc.numdevs;
+ oc->ods = &sbi->oc.ods[first_dev];
}
#endif
fs/exofs/inode.c
View filediff --git a/fs/exofs/inode.c b/fs/exofs/inode.c
index f39a38f..3e5f3a6 100644
--- a/fs/exofs/inode.c
+++ b/fs/exofs/inode.c
@@ -37,11 +37,7 @@
#define EXOFS_DBGMSG2(M...) do {} while (0)
-enum { BIO_MAX_PAGES_KMALLOC =
- (PAGE_SIZE - sizeof(struct bio)) / sizeof(struct bio_vec),
- MAX_PAGES_KMALLOC =
- PAGE_SIZE / sizeof(struct page *),
-};
+enum {MAX_PAGES_KMALLOC = PAGE_SIZE / sizeof(struct page *), };
unsigned exofs_max_io_pages(struct ore_layout *layout,
unsigned expected_pages)
@@ -49,8 +45,7 @@ unsigned exofs_max_io_pages(struct ore_layout *layout,
unsigned pages = min_t(unsigned, expected_pages, MAX_PAGES_KMALLOC);
/* TODO: easily support bio chaining */
- pages = min_t(unsigned, pages,
- layout->group_width * BIO_MAX_PAGES_KMALLOC);
+ pages = min_t(unsigned, pages, layout->max_io_length / PAGE_SIZE);
return pages;
}
@@ -68,6 +63,7 @@ struct page_collect {
bool read_4_write; /* This means two things: that the read is sync
* And the pages should not be unlocked.
*/
+ struct page *that_locked_page;
};
static void _pcol_init(struct page_collect *pcol, unsigned expected_pages,
@@ -86,6 +82,7 @@ static void _pcol_init(struct page_collect *pcol, unsigned expected_pages,
pcol->length = 0;
pcol->pg_first = -1;
pcol->read_4_write = false;
+ pcol->that_locked_page = NULL;
}
static void _pcol_reset(struct page_collect *pcol)
@@ -98,6 +95,7 @@ static void _pcol_reset(struct page_collect *pcol)
pcol->length = 0;
pcol->pg_first = -1;
pcol->ios = NULL;
+ pcol->that_locked_page = NULL;
/* this is probably the end of the loop but in writes
* it might not end here. don't be left with nothing
@@ -149,14 +147,17 @@ static int pcol_add_page(struct page_collect *pcol, struct page *page,
return 0;
}
+enum {PAGE_WAS_NOT_IN_IO = 17};
static int update_read_page(struct page *page, int ret)
{
- if (ret == 0) {
+ switch (ret) {
+ case 0:
/* Everything is OK */
SetPageUptodate(page);
if (PageError(page))
ClearPageError(page);
- } else if (ret == -EFAULT) {
+ break;
+ case -EFAULT:
/* In this case we were trying to read something that wasn't on
* disk yet - return a page full of zeroes. This should be OK,
* because the object should be empty (if there was a write
@@ -167,16 +168,22 @@ static int update_read_page(struct page *page, int ret)
SetPageUptodate(page);
if (PageError(page))
ClearPageError(page);
- ret = 0; /* recovered error */
EXOFS_DBGMSG("recovered read error\n");
- } else /* Error */
+ /* fall through */
+ case PAGE_WAS_NOT_IN_IO:
+ ret = 0; /* recovered error */
+ break;
+ default:
SetPageError(page);
-
+ }
return ret;
}
static void update_write_page(struct page *page, int ret)
{
+ if (unlikely(ret == PAGE_WAS_NOT_IN_IO))
+ return; /* don't pass start don't collect $200 */
+
if (ret) {
mapping_set_error(page->mapping, ret);
SetPageError(page);
@@ -190,15 +197,16 @@ static void update_write_page(struct page *page, int ret)
static int __readpages_done(struct page_collect *pcol)
{
int i;
- u64 resid;
u64 good_bytes;
u64 length = 0;
- int ret = ore_check_io(pcol->ios, &resid);
+ int ret = ore_check_io(pcol->ios, NULL);
- if (likely(!ret))
+ if (likely(!ret)) {
good_bytes = pcol->length;
- else
- good_bytes = pcol->length - resid;
+ ret = PAGE_WAS_NOT_IN_IO;
+ } else {
+ good_bytes = 0;
+ }
EXOFS_DBGMSG2("readpages_done(0x%lx) good_bytes=0x%llx"
" length=0x%lx nr_pages=%u\n",
@@ -259,6 +267,46 @@ static void _unlock_pcol_pages(struct page_collect *pcol, int ret, int rw)
}
}
+static int _maybe_not_all_in_one_io(struct ore_io_state *ios,
+ struct page_collect *pcol_src, struct page_collect *pcol)
+{
+ /* length was wrong or offset was not page aligned */
+ BUG_ON(pcol_src->nr_pages < ios->nr_pages);
+
+ if (pcol_src->nr_pages > ios->nr_pages) {
+ struct page **src_page;
+ unsigned pages_less = pcol_src->nr_pages - ios->nr_pages;
+ unsigned long len_less = pcol_src->length - ios->length;
+ unsigned i;
+ int ret;
+
+ /* This IO was trimmed */
+ pcol_src->nr_pages = ios->nr_pages;
+ pcol_src->length = ios->length;
+
+ /* Left over pages are passed to the next io */
+ pcol->expected_pages += pages_less;
+ pcol->nr_pages = pages_less;
+ pcol->length = len_less;
+ src_page = pcol_src->pages + pcol_src->nr_pages;
+ pcol->pg_first = (*src_page)->index;
+
+ ret = pcol_try_alloc(pcol);
+ if (unlikely(ret))
+ return ret;
+
+ for (i = 0; i < pages_less; ++i)
+ pcol->pages[i] = *src_page++;
+
+ EXOFS_DBGMSG("Length was adjusted nr_pages=0x%x "
+ "pages_less=0x%x expected_pages=0x%x "
+ "next_offset=0x%llx next_len=0x%lx\n",
+ pcol_src->nr_pages, pages_less, pcol->expected_pages,
+ pcol->pg_first * PAGE_SIZE, pcol->length);
+ }
+ return 0;
+}
+
static int read_exec(struct page_collect *pcol)
{
struct exofs_i_info *oi = exofs_i(pcol->inode);
@@ -270,7 +318,7 @@ static int read_exec(struct page_collect *pcol)
return 0;
if (!pcol->ios) {
- int ret = ore_get_rw_state(&pcol->sbi->layout, &oi->comps, true,
+ int ret = ore_get_rw_state(&pcol->sbi->layout, &oi->oc, true,
pcol->pg_first << PAGE_CACHE_SHIFT,
pcol->length, &pcol->ios);
@@ -280,7 +328,6 @@ static int read_exec(struct page_collect *pcol)
ios = pcol->ios;
ios->pages = pcol->pages;
- ios->nr_pages = pcol->nr_pages;
if (pcol->read_4_write) {
ore_read(pcol->ios);
@@ -296,17 +343,23 @@ static int read_exec(struct page_collect *pcol)
*pcol_copy = *pcol;
ios->done = readpages_done;
ios->private = pcol_copy;
+
+ /* pages ownership was passed to pcol_copy */
+ _pcol_reset(pcol);
+
+ ret = _maybe_not_all_in_one_io(ios, pcol_copy, pcol);
+ if (unlikely(ret))
+ goto err;
+
+ EXOFS_DBGMSG2("read_exec(0x%lx) offset=0x%llx length=0x%llx\n",
+ pcol->inode->i_ino, _LLU(ios->offset), _LLU(ios->length));
+
ret = ore_read(ios);
if (unlikely(ret))
goto err;
atomic_inc(&pcol->sbi->s_curr_pending);
- EXOFS_DBGMSG2("read_exec obj=0x%llx start=0x%llx length=0x%lx\n",
- oi->one_comp.obj.id, _LLU(ios->offset), pcol->length);
-
- /* pages ownership was passed to pcol_copy */
- _pcol_reset(pcol);
return 0;
err:
@@ -341,6 +394,8 @@ static int readpage_strip(void *data, struct page *page)
EXOFS_ERR("PageUptodate(0x%lx, 0x%lx)\n", pcol->inode->i_ino,
page->index);
+ pcol->that_locked_page = page;
+
if (page->index < end_index)
len = PAGE_CACHE_SIZE;
else if (page->index == end_index)
@@ -429,6 +484,10 @@ static int exofs_readpages(struct file *file, struct address_space *mapping,
return ret;
}
+ ret = read_exec(&pcol);
+ if (unlikely(ret))
+ return ret;
+
return read_exec(&pcol);
}
@@ -462,17 +521,18 @@ static void writepages_done(struct ore_io_state *ios, void *p)
{
struct page_collect *pcol = p;
int i;
- u64 resid;
u64 good_bytes;
u64 length = 0;
- int ret = ore_check_io(ios, &resid);
+ int ret = ore_check_io(ios, NULL);
atomic_dec(&pcol->sbi->s_curr_pending);
- if (likely(!ret))
+ if (likely(!ret)) {
good_bytes = pcol->length;
- else
- good_bytes = pcol->length - resid;
+ ret = PAGE_WAS_NOT_IN_IO;
+ } else {
+ good_bytes = 0;
+ }
EXOFS_DBGMSG2("writepages_done(0x%lx) good_bytes=0x%llx"
" length=0x%lx nr_pages=%u\n",
@@ -505,6 +565,56 @@ static void writepages_done(struct ore_io_state *ios, void *p)
EXOFS_DBGMSG2("writepages_done END\n");
}
+static struct page *__r4w_get_page(void *priv, u64 offset, bool *uptodate)
+{
+ struct page_collect *pcol = priv;
+ pgoff_t index = offset / PAGE_SIZE;
+
+ if (!pcol->that_locked_page ||
+ (pcol->that_locked_page->index != index)) {
+ struct page *page = find_get_page(pcol->inode->i_mapping, index);
+
+ if (!page) {
+ page = find_or_create_page(pcol->inode->i_mapping,
+ index, GFP_NOFS);
+ if (unlikely(!page)) {
+ EXOFS_DBGMSG("grab_cache_page Failed "
+ "index=0x%llx\n", _LLU(index));
+ return NULL;
+ }
+ unlock_page(page);
+ }
+ if (PageDirty(page) || PageWriteback(page))
+ *uptodate = true;
+ else
+ *uptodate = PageUptodate(page);
+ EXOFS_DBGMSG("index=0x%lx uptodate=%d\n", index, *uptodate);
+ return page;
+ } else {
+ EXOFS_DBGMSG("YES that_locked_page index=0x%lx\n",
+ pcol->that_locked_page->index);
+ *uptodate = true;
+ return pcol->that_locked_page;
+ }
+}
+
+static void __r4w_put_page(void *priv, struct page *page)
+{
+ struct page_collect *pcol = priv;
+
+ if (pcol->that_locked_page != page) {
+ EXOFS_DBGMSG("index=0x%lx\n", page->index);
+ page_cache_release(page);
+ return;
+ }
+ EXOFS_DBGMSG("that_locked_page index=0x%lx\n", page->index);
+}
+
+static const struct _ore_r4w_op _r4w_op = {
+ .get_page = &__r4w_get_page,
+ .put_page = &__r4w_put_page,
+};
+
static int write_exec(struct page_collect *pcol)
{
struct exofs_i_info *oi = exofs_i(pcol->inode);
@@ -516,10 +626,9 @@ static int write_exec(struct page_collect *pcol)
return 0;
BUG_ON(pcol->ios);
- ret = ore_get_rw_state(&pcol->sbi->layout, &oi->comps, false,
+ ret = ore_get_rw_state(&pcol->sbi->layout, &oi->oc, false,
pcol->pg_first << PAGE_CACHE_SHIFT,
pcol->length, &pcol->ios);
-
if (unlikely(ret))
goto err;
@@ -534,10 +643,20 @@ static int write_exec(struct page_collect *pcol)
ios = pcol->ios;
ios->pages = pcol_copy->pages;
- ios->nr_pages = pcol_copy->nr_pages;
ios->done = writepages_done;
+ ios->r4w = &_r4w_op;
ios->private = pcol_copy;
+ /* pages ownership was passed to pcol_copy */
+ _pcol_reset(pcol);
+
+ ret = _maybe_not_all_in_one_io(ios, pcol_copy, pcol);
+ if (unlikely(ret))
+ goto err;
+
+ EXOFS_DBGMSG2("write_exec(0x%lx) offset=0x%llx length=0x%llx\n",
+ pcol->inode->i_ino, _LLU(ios->offset), _LLU(ios->length));
+
ret = ore_write(ios);
if (unlikely(ret)) {
EXOFS_ERR("write_exec: ore_write() Failed\n");
@@ -545,11 +664,6 @@ static int write_exec(struct page_collect *pcol)
}
atomic_inc(&pcol->sbi->s_curr_pending);
- 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);
return 0;
err:
@@ -689,14 +803,33 @@ static int exofs_writepages(struct address_space *mapping,
_pcol_init(&pcol, expected_pages, mapping->host);
ret = write_cache_pages(mapping, wbc, writepage_strip, &pcol);
- if (ret) {
+ if (unlikely(ret)) {
EXOFS_ERR("write_cache_pages => %d\n", ret);
return ret;
}
- return write_exec(&pcol);
+ ret = write_exec(&pcol);
+ if (unlikely(ret))
+ return ret;
+
+ if (wbc->sync_mode == WB_SYNC_ALL) {
+ return write_exec(&pcol); /* pump the last reminder */
+ } else if (pcol.nr_pages) {
+ /* not SYNC let the reminder join the next writeout */
+ unsigned i;
+
+ for (i = 0; i < pcol.nr_pages; i++) {
+ struct page *page = pcol.pages[i];
+
+ end_page_writeback(page);
+ set_page_dirty(page);
+ unlock_page(page);
+ }
+ }
+ return 0;
}
+/*
static int exofs_writepage(struct page *page, struct writeback_control *wbc)
{
struct page_collect pcol;
@@ -712,7 +845,7 @@ static int exofs_writepage(struct page *page, struct writeback_control *wbc)
return write_exec(&pcol);
}
-
+*/
/* i_mutex held using inode->i_size directly */
static void _write_failed(struct inode *inode, loff_t to)
{
@@ -818,7 +951,7 @@ static void exofs_invalidatepage(struct page *page, unsigned long offset)
const struct address_space_operations exofs_aops = {
.readpage = exofs_readpage,
.readpages = exofs_readpages,
- .writepage = exofs_writepage,
+ .writepage = NULL,
.writepages = exofs_writepages,
.write_begin = exofs_write_begin_export,
.write_end = exofs_write_end,
@@ -860,7 +993,7 @@ static int _do_truncate(struct inode *inode, loff_t newsize)
inode->i_mtime = inode->i_ctime = CURRENT_TIME;
- ret = ore_truncate(&sbi->layout, &oi->comps, (u64)newsize);
+ ret = ore_truncate(&sbi->layout, &oi->oc, (u64)newsize);
if (likely(!ret))
truncate_setsize(inode, newsize);
@@ -927,14 +1060,14 @@ static int exofs_get_inode(struct super_block *sb, struct exofs_i_info *oi,
struct exofs_on_disk_inode_layout *layout;
int ret;
- ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
+ ret = ore_get_io_state(&sbi->layout, &oi->oc, &ios);
if (unlikely(ret)) {
EXOFS_ERR("%s: ore_get_io_state failed.\n", __func__);
return ret;
}
- attrs[1].len = exofs_on_disk_inode_layout_size(sbi->comps.numdevs);
- attrs[2].len = exofs_on_disk_inode_layout_size(sbi->comps.numdevs);
+ attrs[1].len = exofs_on_disk_inode_layout_size(sbi->oc.numdevs);
+ attrs[2].len = exofs_on_disk_inode_layout_size(sbi->oc.numdevs);
ios->in_attr = attrs;
ios->in_attr_len = ARRAY_SIZE(attrs);
@@ -1018,7 +1151,7 @@ struct inode *exofs_iget(struct super_block *sb, unsigned long ino)
return inode;
oi = exofs_i(inode);
__oi_init(oi);
- exofs_init_comps(&oi->comps, &oi->one_comp, sb->s_fs_info,
+ exofs_init_comps(&oi->oc, &oi->one_comp, sb->s_fs_info,
exofs_oi_objno(oi));
/* read the inode from the osd */
@@ -1172,13 +1305,13 @@ struct inode *exofs_new_inode(struct inode *dir, int mode)
spin_unlock(&sbi->s_next_gen_lock);
insert_inode_hash(inode);
- exofs_init_comps(&oi->comps, &oi->one_comp, sb->s_fs_info,
+ exofs_init_comps(&oi->oc, &oi->one_comp, sb->s_fs_info,
exofs_oi_objno(oi));
exofs_sbi_write_stats(sbi); /* Make sure new sbi->s_nextid is on disk */
mark_inode_dirty(inode);
- ret = ore_get_io_state(&sbi->layout, &oi->comps, &ios);
+ ret = ore_get_io_state(&sbi->layout, &oi->oc, &ios);
if (unlikely(ret)) {
EXOFS_ERR("exofs_new_inode: ore_get_io_state failed\n");
return ERR_PTR(ret);
@@ -1267,7 +1400,7 @@ static int exofs_update_inode(struct inode *inode, int do_sync)
} else
This diff is incomplete. Download the complete patch.
fs/exofs/ore.c
View fileThis diff is incomplete. Download the complete patch.
fs/exofs/ore_raid.c
View fileThis diff is incomplete. Download the complete patch.
fs/exofs/ore_raid.h
View fileThis diff is incomplete. Download the complete patch.
fs/exofs/super.c
View fileThis diff is incomplete. Download the complete patch.
include/scsi/osd_ore.h
View fileThis diff is incomplete. Download the complete patch.