* [PATCH 0/5] pending items for 2.6.37
@ 2010-12-20 19:15 Sage Weil
2010-12-20 19:15 ` [PATCH 1/5] ceph: fix msgr_init error path Sage Weil
0 siblings, 1 reply; 6+ messages in thread
From: Sage Weil @ 2010-12-20 19:15 UTC (permalink / raw)
To: ceph-devel; +Cc: Sage Weil
This is the current set of patches I have queued for the next -rc. The
direct-io page dirty ones could really use some review. Also the buffer
alignment.
Thanks!
sage
---
Henry C Chang (3):
ceph: fix direct-io on non-page-aligned buffers
ceph: mark user pages dirty on direct-io reads
ceph: handle partial result from get_user_pages
Sage Weil (2):
ceph: fix msgr_init error path
ceph: fix null pointer dereference in ceph_init_dentry for nfs
reexport
fs/ceph/dir.c | 3 ++-
fs/ceph/file.c | 39 +++++++++++++++++++++++----------------
include/linux/ceph/libceph.h | 6 ++++--
net/ceph/messenger.c | 8 +++-----
net/ceph/pagevec.c | 15 +++++++++------
5 files changed, 41 insertions(+), 30 deletions(-)
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH 1/5] ceph: fix msgr_init error path
2010-12-20 19:15 [PATCH 0/5] pending items for 2.6.37 Sage Weil
@ 2010-12-20 19:15 ` Sage Weil
2010-12-20 19:15 ` [PATCH 2/5] ceph: fix direct-io on non-page-aligned buffers Sage Weil
0 siblings, 1 reply; 6+ messages in thread
From: Sage Weil @ 2010-12-20 19:15 UTC (permalink / raw)
To: ceph-devel; +Cc: Sage Weil
create_workqueue() returns NULL on failure.
Signed-off-by: Sage Weil <sage@newdream.net>
---
net/ceph/messenger.c | 8 +++-----
1 files changed, 3 insertions(+), 5 deletions(-)
diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index 1c7a2ec..b6ff4a1 100644
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -97,11 +97,9 @@ struct workqueue_struct *ceph_msgr_wq;
int ceph_msgr_init(void)
{
ceph_msgr_wq = create_workqueue("ceph-msgr");
- if (IS_ERR(ceph_msgr_wq)) {
- int ret = PTR_ERR(ceph_msgr_wq);
- pr_err("msgr_init failed to create workqueue: %d\n", ret);
- ceph_msgr_wq = NULL;
- return ret;
+ if (!ceph_msgr_wq) {
+ pr_err("msgr_init failed to create workqueue\n");
+ return -ENOMEM;
}
return 0;
}
--
1.6.6.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 2/5] ceph: fix direct-io on non-page-aligned buffers
2010-12-20 19:15 ` [PATCH 1/5] ceph: fix msgr_init error path Sage Weil
@ 2010-12-20 19:15 ` Sage Weil
2010-12-20 19:15 ` [PATCH 3/5] ceph: fix null pointer dereference in ceph_init_dentry for nfs reexport Sage Weil
0 siblings, 1 reply; 6+ messages in thread
From: Sage Weil @ 2010-12-20 19:15 UTC (permalink / raw)
To: ceph-devel; +Cc: Henry C Chang, Sage Weil
From: Henry C Chang <henry_c_chang@tcloudcomputing.com>
The user buffer may be 512-byte aligned, not page-aligned. We were
assuming the buffer was page-aligned and only accounting for
non-page-aligned io offsets.
Signed-off-by: Henry C Chang <henry_c_chang@tcloudcomputing.com>
Signed-off-by: Sage Weil <sage@newdream.net>
---
fs/ceph/file.c | 31 +++++++++++++++++++------------
1 files changed, 19 insertions(+), 12 deletions(-)
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index 8d79b89..e860d8f 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -282,7 +282,8 @@ int ceph_release(struct inode *inode, struct file *file)
static int striped_read(struct inode *inode,
u64 off, u64 len,
struct page **pages, int num_pages,
- int *checkeof, bool align_to_pages)
+ int *checkeof, bool align_to_pages,
+ unsigned long buf_align)
{
struct ceph_fs_client *fsc = ceph_inode_to_client(inode);
struct ceph_inode_info *ci = ceph_inode(inode);
@@ -307,7 +308,7 @@ static int striped_read(struct inode *inode,
more:
if (align_to_pages)
- page_align = (pos - io_align) & ~PAGE_MASK;
+ page_align = (pos - io_align + buf_align) & ~PAGE_MASK;
else
page_align = pos & ~PAGE_MASK;
this_len = left;
@@ -376,16 +377,18 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data,
struct inode *inode = file->f_dentry->d_inode;
struct page **pages;
u64 off = *poff;
- int num_pages = calc_pages_for(off, len);
- int ret;
+ int num_pages, ret;
dout("sync_read on file %p %llu~%u %s\n", file, off, len,
(file->f_flags & O_DIRECT) ? "O_DIRECT" : "");
- if (file->f_flags & O_DIRECT)
+ if (file->f_flags & O_DIRECT) {
+ num_pages = calc_pages_for((unsigned long)data, len);
pages = ceph_get_direct_page_vector(data, num_pages);
- else
+ } else {
+ num_pages = calc_pages_for(off, len);
pages = ceph_alloc_page_vector(num_pages, GFP_NOFS);
+ }
if (IS_ERR(pages))
return PTR_ERR(pages);
@@ -400,7 +403,8 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data,
goto done;
ret = striped_read(inode, off, len, pages, num_pages, checkeof,
- file->f_flags & O_DIRECT);
+ file->f_flags & O_DIRECT,
+ (unsigned long)data & ~PAGE_MASK);
if (ret >= 0 && (file->f_flags & O_DIRECT) == 0)
ret = ceph_copy_page_vector_to_user(pages, data, off, ret);
@@ -456,6 +460,7 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data,
int do_sync = 0;
int check_caps = 0;
int page_align, io_align;
+ unsigned long buf_align;
int ret;
struct timespec mtime = CURRENT_TIME;
@@ -471,6 +476,7 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data,
pos = *offset;
io_align = pos & ~PAGE_MASK;
+ buf_align = (unsigned long)data & ~PAGE_MASK;
ret = filemap_write_and_wait_range(inode->i_mapping, pos, pos + left);
if (ret < 0)
@@ -496,12 +502,15 @@ static ssize_t ceph_sync_write(struct file *file, const char __user *data,
*/
more:
len = left;
- if (file->f_flags & O_DIRECT)
+ if (file->f_flags & O_DIRECT) {
/* write from beginning of first page, regardless of
io alignment */
- page_align = (pos - io_align) & ~PAGE_MASK;
- else
+ page_align = (pos - io_align + buf_align) & ~PAGE_MASK;
+ num_pages = calc_pages_for((unsigned long)data, len);
+ } else {
page_align = pos & ~PAGE_MASK;
+ num_pages = calc_pages_for(pos, len);
+ }
req = ceph_osdc_new_request(&fsc->client->osdc, &ci->i_layout,
ceph_vino(inode), pos, &len,
CEPH_OSD_OP_WRITE, flags,
@@ -512,8 +521,6 @@ more:
if (!req)
return -ENOMEM;
- num_pages = calc_pages_for(pos, len);
-
if (file->f_flags & O_DIRECT) {
pages = ceph_get_direct_page_vector(data, num_pages);
if (IS_ERR(pages)) {
--
1.6.6.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 3/5] ceph: fix null pointer dereference in ceph_init_dentry for nfs reexport
2010-12-20 19:15 ` [PATCH 2/5] ceph: fix direct-io on non-page-aligned buffers Sage Weil
@ 2010-12-20 19:15 ` Sage Weil
2010-12-20 19:15 ` [PATCH 4/5] ceph: mark user pages dirty on direct-io reads Sage Weil
0 siblings, 1 reply; 6+ messages in thread
From: Sage Weil @ 2010-12-20 19:15 UTC (permalink / raw)
To: ceph-devel; +Cc: Sage Weil
The fh_to_dentry etc. methods use ceph_init_dentry(), which assumes that
d_parent is defined. It isn't for those callers, so check!
Signed-off-by: Sage Weil <sage@newdream.net>
---
fs/ceph/dir.c | 3 ++-
1 files changed, 2 insertions(+), 1 deletions(-)
diff --git a/fs/ceph/dir.c b/fs/ceph/dir.c
index 158c700..d902948 100644
--- a/fs/ceph/dir.c
+++ b/fs/ceph/dir.c
@@ -40,7 +40,8 @@ int ceph_init_dentry(struct dentry *dentry)
if (dentry->d_fsdata)
return 0;
- if (ceph_snap(dentry->d_parent->d_inode) == CEPH_NOSNAP)
+ if (dentry->d_parent == NULL || /* nfs fh_to_dentry */
+ ceph_snap(dentry->d_parent->d_inode) == CEPH_NOSNAP)
dentry->d_op = &ceph_dentry_ops;
else if (ceph_snap(dentry->d_parent->d_inode) == CEPH_SNAPDIR)
dentry->d_op = &ceph_snapdir_dentry_ops;
--
1.6.6.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 4/5] ceph: mark user pages dirty on direct-io reads
2010-12-20 19:15 ` [PATCH 3/5] ceph: fix null pointer dereference in ceph_init_dentry for nfs reexport Sage Weil
@ 2010-12-20 19:15 ` Sage Weil
2010-12-20 19:15 ` [PATCH 5/5] ceph: handle partial result from get_user_pages Sage Weil
0 siblings, 1 reply; 6+ messages in thread
From: Sage Weil @ 2010-12-20 19:15 UTC (permalink / raw)
To: ceph-devel; +Cc: Henry C Chang, Sage Weil
From: Henry C Chang <henry_c_chang@tcloudcomputing.com>
For read operation, we have to set the argument _write_ of get_user_pages
to 1 since we will write data to pages. Also, we need to SetPageDirty before
releasing these pages.
Signed-off-by: Henry C Chang <henry_c_chang@tcloudcomputing.com>
Signed-off-by: Sage Weil <sage@newdream.net>
---
fs/ceph/file.c | 8 ++++----
include/linux/ceph/libceph.h | 6 ++++--
net/ceph/pagevec.c | 11 +++++++----
3 files changed, 15 insertions(+), 10 deletions(-)
diff --git a/fs/ceph/file.c b/fs/ceph/file.c
index e860d8f..7d0e4a8 100644
--- a/fs/ceph/file.c
+++ b/fs/ceph/file.c
@@ -384,7 +384,7 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data,
if (file->f_flags & O_DIRECT) {
num_pages = calc_pages_for((unsigned long)data, len);
- pages = ceph_get_direct_page_vector(data, num_pages);
+ pages = ceph_get_direct_page_vector(data, num_pages, true);
} else {
num_pages = calc_pages_for(off, len);
pages = ceph_alloc_page_vector(num_pages, GFP_NOFS);
@@ -413,7 +413,7 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data,
done:
if (file->f_flags & O_DIRECT)
- ceph_put_page_vector(pages, num_pages);
+ ceph_put_page_vector(pages, num_pages, true);
else
ceph_release_page_vector(pages, num_pages);
dout("sync_read result %d\n", ret);
@@ -522,7 +522,7 @@ more:
return -ENOMEM;
if (file->f_flags & O_DIRECT) {
- pages = ceph_get_direct_page_vector(data, num_pages);
+ pages = ceph_get_direct_page_vector(data, num_pages, false);
if (IS_ERR(pages)) {
ret = PTR_ERR(pages);
goto out;
@@ -572,7 +572,7 @@ more:
}
if (file->f_flags & O_DIRECT)
- ceph_put_page_vector(pages, num_pages);
+ ceph_put_page_vector(pages, num_pages, false);
else if (file->f_flags & O_SYNC)
ceph_release_page_vector(pages, num_pages);
diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h
index 9e76d35..72c72bf 100644
--- a/include/linux/ceph/libceph.h
+++ b/include/linux/ceph/libceph.h
@@ -227,8 +227,10 @@ extern int ceph_open_session(struct ceph_client *client);
extern void ceph_release_page_vector(struct page **pages, int num_pages);
extern struct page **ceph_get_direct_page_vector(const char __user *data,
- int num_pages);
-extern void ceph_put_page_vector(struct page **pages, int num_pages);
+ int num_pages,
+ bool write_page);
+extern void ceph_put_page_vector(struct page **pages, int num_pages,
+ bool dirty);
extern void ceph_release_page_vector(struct page **pages, int num_pages);
extern struct page **ceph_alloc_page_vector(int num_pages, gfp_t flags);
extern int ceph_copy_user_to_page_vector(struct page **pages,
diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c
index ac34fee..01947a5 100644
--- a/net/ceph/pagevec.c
+++ b/net/ceph/pagevec.c
@@ -13,7 +13,7 @@
* build a vector of user pages
*/
struct page **ceph_get_direct_page_vector(const char __user *data,
- int num_pages)
+ int num_pages, bool write_page)
{
struct page **pages;
int rc;
@@ -24,7 +24,7 @@ struct page **ceph_get_direct_page_vector(const char __user *data,
down_read(¤t->mm->mmap_sem);
rc = get_user_pages(current, current->mm, (unsigned long)data,
- num_pages, 0, 0, pages, NULL);
+ num_pages, write_page, 0, pages, NULL);
up_read(¤t->mm->mmap_sem);
if (rc < 0)
goto fail;
@@ -36,12 +36,15 @@ fail:
}
EXPORT_SYMBOL(ceph_get_direct_page_vector);
-void ceph_put_page_vector(struct page **pages, int num_pages)
+void ceph_put_page_vector(struct page **pages, int num_pages, bool dirty)
{
int i;
- for (i = 0; i < num_pages; i++)
+ for (i = 0; i < num_pages; i++) {
+ if (dirty)
+ set_page_dirty_lock(pages[i]);
put_page(pages[i]);
+ }
kfree(pages);
}
EXPORT_SYMBOL(ceph_put_page_vector);
--
1.6.6.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
* [PATCH 5/5] ceph: handle partial result from get_user_pages
2010-12-20 19:15 ` [PATCH 4/5] ceph: mark user pages dirty on direct-io reads Sage Weil
@ 2010-12-20 19:15 ` Sage Weil
0 siblings, 0 replies; 6+ messages in thread
From: Sage Weil @ 2010-12-20 19:15 UTC (permalink / raw)
To: ceph-devel; +Cc: Henry C Chang, Sage Weil
From: Henry C Chang <henry_c_chang@tcloudcomputing.com>
The get_user_pages() helper can return fewer than the requested pages.
Error out in that case, and clean up the partial result.
Signed-off-by: Henry C Chang <henry_c_chang@tcloudcomputing.com>
Signed-off-by: Sage Weil <sage@newdream.net>
---
net/ceph/pagevec.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c
index 01947a5..1a040e6 100644
--- a/net/ceph/pagevec.c
+++ b/net/ceph/pagevec.c
@@ -26,12 +26,12 @@ struct page **ceph_get_direct_page_vector(const char __user *data,
rc = get_user_pages(current, current->mm, (unsigned long)data,
num_pages, write_page, 0, pages, NULL);
up_read(¤t->mm->mmap_sem);
- if (rc < 0)
+ if (rc < num_pages)
goto fail;
return pages;
fail:
- kfree(pages);
+ ceph_put_page_vector(pages, rc > 0 ? rc : 0, false);
return ERR_PTR(rc);
}
EXPORT_SYMBOL(ceph_get_direct_page_vector);
--
1.6.6.1
^ permalink raw reply related [flat|nested] 6+ messages in thread
end of thread, other threads:[~2010-12-20 19:15 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-12-20 19:15 [PATCH 0/5] pending items for 2.6.37 Sage Weil
2010-12-20 19:15 ` [PATCH 1/5] ceph: fix msgr_init error path Sage Weil
2010-12-20 19:15 ` [PATCH 2/5] ceph: fix direct-io on non-page-aligned buffers Sage Weil
2010-12-20 19:15 ` [PATCH 3/5] ceph: fix null pointer dereference in ceph_init_dentry for nfs reexport Sage Weil
2010-12-20 19:15 ` [PATCH 4/5] ceph: mark user pages dirty on direct-io reads Sage Weil
2010-12-20 19:15 ` [PATCH 5/5] ceph: handle partial result from get_user_pages Sage Weil
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.