2006-10-04 13:16:22 +04:00
/**
* eCryptfs : Linux filesystem encryption layer
* This is where eCryptfs coordinates the symmetric encryption and
* decryption of the file data as it passes between the lower
* encrypted file and the upper decrypted file .
*
* Copyright ( C ) 1997 - 2003 Erez Zadok
* Copyright ( C ) 2001 - 2003 Stony Brook University
2007-02-12 11:53:46 +03:00
* Copyright ( C ) 2004 - 2007 International Business Machines Corp .
2006-10-04 13:16:22 +04:00
* Author ( s ) : Michael A . Halcrow < mahalcro @ us . ibm . com >
*
* This program is free software ; you can redistribute it and / or
* modify it under the terms of the GNU General Public License as
* published by the Free Software Foundation ; either version 2 of the
* License , or ( at your option ) any later version .
*
* This program is distributed in the hope that it will be useful , but
* WITHOUT ANY WARRANTY ; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE . See the GNU
* General Public License for more details .
*
* You should have received a copy of the GNU General Public License
* along with this program ; if not , write to the Free Software
* Foundation , Inc . , 59 Temple Place - Suite 330 , Boston , MA
* 02111 - 1307 , USA .
*/
# include <linux/pagemap.h>
# include <linux/writeback.h>
# include <linux/page-flags.h>
# include <linux/mount.h>
# include <linux/file.h>
# include <linux/crypto.h>
# include <linux/scatterlist.h>
include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h
percpu.h is included by sched.h and module.h and thus ends up being
included when building most .c files. percpu.h includes slab.h which
in turn includes gfp.h making everything defined by the two files
universally available and complicating inclusion dependencies.
percpu.h -> slab.h dependency is about to be removed. Prepare for
this change by updating users of gfp and slab facilities include those
headers directly instead of assuming availability. As this conversion
needs to touch large number of source files, the following script is
used as the basis of conversion.
http://userweb.kernel.org/~tj/misc/slabh-sweep.py
The script does the followings.
* Scan files for gfp and slab usages and update includes such that
only the necessary includes are there. ie. if only gfp is used,
gfp.h, if slab is used, slab.h.
* When the script inserts a new include, it looks at the include
blocks and try to put the new include such that its order conforms
to its surrounding. It's put in the include block which contains
core kernel includes, in the same order that the rest are ordered -
alphabetical, Christmas tree, rev-Xmas-tree or at the end if there
doesn't seem to be any matching order.
* If the script can't find a place to put a new include (mostly
because the file doesn't have fitting include block), it prints out
an error message indicating which .h file needs to be added to the
file.
The conversion was done in the following steps.
1. The initial automatic conversion of all .c files updated slightly
over 4000 files, deleting around 700 includes and adding ~480 gfp.h
and ~3000 slab.h inclusions. The script emitted errors for ~400
files.
2. Each error was manually checked. Some didn't need the inclusion,
some needed manual addition while adding it to implementation .h or
embedding .c file was more appropriate for others. This step added
inclusions to around 150 files.
3. The script was run again and the output was compared to the edits
from #2 to make sure no file was left behind.
4. Several build tests were done and a couple of problems were fixed.
e.g. lib/decompress_*.c used malloc/free() wrappers around slab
APIs requiring slab.h to be added manually.
5. The script was run on all .h files but without automatically
editing them as sprinkling gfp.h and slab.h inclusions around .h
files could easily lead to inclusion dependency hell. Most gfp.h
inclusion directives were ignored as stuff from gfp.h was usually
wildly available and often used in preprocessor macros. Each
slab.h inclusion directive was examined and added manually as
necessary.
6. percpu.h was updated not to include slab.h.
7. Build test were done on the following configurations and failures
were fixed. CONFIG_GCOV_KERNEL was turned off for all tests (as my
distributed build env didn't work with gcov compiles) and a few
more options had to be turned off depending on archs to make things
build (like ipr on powerpc/64 which failed due to missing writeq).
* x86 and x86_64 UP and SMP allmodconfig and a custom test config.
* powerpc and powerpc64 SMP allmodconfig
* sparc and sparc64 SMP allmodconfig
* ia64 SMP allmodconfig
* s390 SMP allmodconfig
* alpha SMP allmodconfig
* um on x86_64 SMP allmodconfig
8. percpu.h modifications were reverted so that it could be applied as
a separate patch and serve as bisection point.
Given the fact that I had only a couple of failures from tests on step
6, I'm fairly confident about the coverage of this conversion patch.
If there is a breakage, it's likely to be something in one of the arch
headers which should be easily discoverable easily on most builds of
the specific arch.
Signed-off-by: Tejun Heo <tj@kernel.org>
Guess-its-ok-by: Christoph Lameter <cl@linux-foundation.org>
Cc: Ingo Molnar <mingo@redhat.com>
Cc: Lee Schermerhorn <Lee.Schermerhorn@hp.com>
2010-03-24 11:04:11 +03:00
# include <linux/slab.h>
2008-07-24 08:30:07 +04:00
# include <asm/unaligned.h>
2006-10-04 13:16:22 +04:00
# include "ecryptfs_kernel.h"
/**
2007-10-16 12:28:14 +04:00
* ecryptfs_get_locked_page
2006-10-04 13:16:22 +04:00
*
* Get one page from cache or lower f / s , return error otherwise .
*
2007-10-16 12:28:14 +04:00
* Returns locked and up - to - date page ( if ok ) , with increased
2006-10-04 13:16:22 +04:00
* refcnt .
*/
2010-05-21 19:02:14 +04:00
struct page * ecryptfs_get_locked_page ( struct inode * inode , loff_t index )
2006-10-04 13:16:22 +04:00
{
2010-05-21 19:02:14 +04:00
struct page * page = read_mapping_page ( inode - > i_mapping , index , NULL ) ;
2007-10-16 12:28:14 +04:00
if ( ! IS_ERR ( page ) )
lock_page ( page ) ;
return page ;
2006-10-04 13:16:22 +04:00
}
/**
* ecryptfs_writepage
* @ page : Page that is locked before this call is made
*
* Returns zero on success ; non - zero otherwise
2012-01-25 11:40:31 +04:00
*
* This is where we encrypt the data and pass the encrypted data to
* the lower filesystem . In OpenPGP - compatible mode , we operate on
* entire underlying packets .
2006-10-04 13:16:22 +04:00
*/
static int ecryptfs_writepage ( struct page * page , struct writeback_control * wbc )
{
int rc ;
2007-10-16 12:28:08 +04:00
rc = ecryptfs_encrypt_page ( page ) ;
2006-10-04 13:16:22 +04:00
if ( rc ) {
ecryptfs_printk ( KERN_WARNING , " Error encrypting "
2010-11-11 02:46:16 +03:00
" page (upper index [0x%.16lx]) \n " , page - > index ) ;
2006-10-04 13:16:22 +04:00
ClearPageUptodate ( page ) ;
goto out ;
}
SetPageUptodate ( page ) ;
out :
2011-03-09 03:26:03 +03:00
unlock_page ( page ) ;
2006-10-04 13:16:22 +04:00
return rc ;
}
2010-02-11 09:02:32 +03:00
static void strip_xattr_flag ( char * page_virt ,
struct ecryptfs_crypt_stat * crypt_stat )
{
if ( crypt_stat - > flags & ECRYPTFS_METADATA_IN_XATTR ) {
size_t written ;
crypt_stat - > flags & = ~ ECRYPTFS_METADATA_IN_XATTR ;
ecryptfs_write_crypt_stat_flags ( page_virt , crypt_stat ,
& written ) ;
crypt_stat - > flags | = ECRYPTFS_METADATA_IN_XATTR ;
}
}
2007-02-12 11:53:47 +03:00
/**
* Header Extent :
* Octets 0 - 7 : Unencrypted file size ( big - endian )
* Octets 8 - 15 : eCryptfs special marker
* Octets 16 - 19 : Flags
* Octet 16 : File format version number ( between 0 and 255 )
* Octets 17 - 18 : Reserved
* Octet 19 : Bit 1 ( lsb ) : Reserved
* Bit 2 : Encrypted ?
* Bits 3 - 8 : Reserved
* Octets 20 - 23 : Header extent size ( big - endian )
* Octets 24 - 25 : Number of header extents at front of file
* ( big - endian )
* Octet 26 : Begin RFC 2440 authentication token packet set
*/
2006-10-04 13:16:22 +04:00
2007-10-16 12:28:11 +04:00
/**
* ecryptfs_copy_up_encrypted_with_header
* @ page : Sort of a ` ` virtual ' ' representation of the encrypted lower
* file . The actual lower file does not have the metadata in
* the header . This is locked .
* @ crypt_stat : The eCryptfs inode ' s cryptographic context
*
* The ` ` view ' ' is the version of the file that userspace winds up
* seeing , with the header information inserted .
*/
static int
ecryptfs_copy_up_encrypted_with_header ( struct page * page ,
struct ecryptfs_crypt_stat * crypt_stat )
{
loff_t extent_num_in_page = 0 ;
loff_t num_extents_per_page = ( PAGE_CACHE_SIZE
/ crypt_stat - > extent_size ) ;
int rc = 0 ;
while ( extent_num_in_page < num_extents_per_page ) {
2007-10-16 12:28:12 +04:00
loff_t view_extent_num = ( ( ( ( loff_t ) page - > index )
* num_extents_per_page )
2007-10-16 12:28:11 +04:00
+ extent_num_in_page ) ;
2008-02-06 12:38:32 +03:00
size_t num_header_extents_at_front =
2010-02-11 14:09:14 +03:00
( crypt_stat - > metadata_size / crypt_stat - > extent_size ) ;
2007-10-16 12:28:11 +04:00
2008-02-06 12:38:32 +03:00
if ( view_extent_num < num_header_extents_at_front ) {
2007-10-16 12:28:11 +04:00
/* This is a header extent */
char * page_virt ;
2012-02-10 09:39:50 +04:00
page_virt = kmap_atomic ( page ) ;
2007-10-16 12:28:11 +04:00
memset ( page_virt , 0 , PAGE_CACHE_SIZE ) ;
/* TODO: Support more than one header extent */
if ( view_extent_num = = 0 ) {
2010-02-11 16:10:38 +03:00
size_t written ;
2007-10-16 12:28:11 +04:00
rc = ecryptfs_read_xattr_region (
page_virt , page - > mapping - > host ) ;
2010-02-11 09:02:32 +03:00
strip_xattr_flag ( page_virt + 16 , crypt_stat ) ;
2010-02-11 16:10:38 +03:00
ecryptfs_write_header_metadata ( page_virt + 20 ,
crypt_stat ,
& written ) ;
2007-10-16 12:28:11 +04:00
}
2012-02-10 09:39:50 +04:00
kunmap_atomic ( page_virt ) ;
2007-10-16 12:28:11 +04:00
flush_dcache_page ( page ) ;
if ( rc ) {
printk ( KERN_ERR " %s: Error reading xattr "
2008-04-29 11:59:48 +04:00
" region; rc = [%d] \n " , __func__ , rc ) ;
2007-10-16 12:28:11 +04:00
goto out ;
}
} else {
/* This is an encrypted data extent */
loff_t lower_offset =
2008-02-06 12:38:32 +03:00
( ( view_extent_num * crypt_stat - > extent_size )
2010-02-11 14:09:14 +03:00
- crypt_stat - > metadata_size ) ;
2007-10-16 12:28:11 +04:00
rc = ecryptfs_read_lower_page_segment (
page , ( lower_offset > > PAGE_CACHE_SHIFT ) ,
( lower_offset & ~ PAGE_CACHE_MASK ) ,
crypt_stat - > extent_size , page - > mapping - > host ) ;
if ( rc ) {
printk ( KERN_ERR " %s: Error attempting to read "
" extent at offset [%lld] in the lower "
2008-04-29 11:59:48 +04:00
" file; rc = [%d] \n " , __func__ ,
2007-10-16 12:28:11 +04:00
lower_offset , rc ) ;
goto out ;
}
}
extent_num_in_page + + ;
}
out :
return rc ;
}
2006-10-04 13:16:22 +04:00
/**
* ecryptfs_readpage
2007-10-16 12:28:11 +04:00
* @ file : An eCryptfs file
* @ page : Page from eCryptfs inode mapping into which to stick the read data
2006-10-04 13:16:22 +04:00
*
* Read in a page , decrypting if necessary .
*
* Returns zero on success ; non - zero on error .
*/
static int ecryptfs_readpage ( struct file * file , struct page * page )
{
2007-10-16 12:28:11 +04:00
struct ecryptfs_crypt_stat * crypt_stat =
2010-05-21 18:56:12 +04:00
& ecryptfs_inode_to_private ( page - > mapping - > host ) - > crypt_stat ;
2006-10-04 13:16:22 +04:00
int rc = 0 ;
2011-02-23 09:54:20 +03:00
if ( ! crypt_stat | | ! ( crypt_stat - > flags & ECRYPTFS_ENCRYPTED ) ) {
2007-10-16 12:28:11 +04:00
rc = ecryptfs_read_lower_page_segment ( page , page - > index , 0 ,
PAGE_CACHE_SIZE ,
page - > mapping - > host ) ;
2007-02-12 11:53:47 +03:00
} else if ( crypt_stat - > flags & ECRYPTFS_VIEW_AS_ENCRYPTED ) {
if ( crypt_stat - > flags & ECRYPTFS_METADATA_IN_XATTR ) {
2007-10-16 12:28:11 +04:00
rc = ecryptfs_copy_up_encrypted_with_header ( page ,
crypt_stat ) ;
if ( rc ) {
printk ( KERN_ERR " %s: Error attempting to copy "
" the encrypted content from the lower "
" file whilst inserting the metadata "
" from the xattr into the header; rc = "
2008-04-29 11:59:48 +04:00
" [%d] \n " , __func__ , rc ) ;
2007-10-16 12:28:11 +04:00
goto out ;
2007-02-12 11:53:47 +03:00
}
2007-10-16 12:28:11 +04:00
2007-02-12 11:53:47 +03:00
} else {
2007-10-16 12:28:11 +04:00
rc = ecryptfs_read_lower_page_segment (
page , page - > index , 0 , PAGE_CACHE_SIZE ,
page - > mapping - > host ) ;
2007-02-12 11:53:47 +03:00
if ( rc ) {
printk ( KERN_ERR " Error reading page; rc = "
" [%d] \n " , rc ) ;
goto out ;
}
}
2006-10-04 13:16:22 +04:00
} else {
2007-10-16 12:28:08 +04:00
rc = ecryptfs_decrypt_page ( page ) ;
2006-10-04 13:16:22 +04:00
if ( rc ) {
ecryptfs_printk ( KERN_ERR , " Error decrypting page; "
" rc = [%d] \n " , rc ) ;
goto out ;
}
}
out :
2007-10-16 12:28:14 +04:00
if ( rc )
ClearPageUptodate ( page ) ;
else
SetPageUptodate ( page ) ;
2010-11-11 02:46:16 +03:00
ecryptfs_printk ( KERN_DEBUG , " Unlocking page with index = [0x%.16lx] \n " ,
2006-10-04 13:16:22 +04:00
page - > index ) ;
unlock_page ( page ) ;
return rc ;
}
2007-02-12 11:53:46 +03:00
/**
* Called with lower inode mutex held .
*/
2006-10-04 13:16:22 +04:00
static int fill_zeros_to_end_of_page ( struct page * page , unsigned int to )
{
struct inode * inode = page - > mapping - > host ;
int end_byte_in_page ;
2007-02-12 11:53:48 +03:00
if ( ( i_size_read ( inode ) / PAGE_CACHE_SIZE ) ! = page - > index )
goto out ;
end_byte_in_page = i_size_read ( inode ) % PAGE_CACHE_SIZE ;
if ( to > end_byte_in_page )
end_byte_in_page = to ;
2008-02-05 09:28:29 +03:00
zero_user_segment ( page , end_byte_in_page , PAGE_CACHE_SIZE ) ;
2006-10-04 13:16:22 +04:00
out :
2007-02-12 11:53:48 +03:00
return 0 ;
2006-10-04 13:16:22 +04:00
}
2008-03-05 01:29:24 +03:00
/**
2008-10-16 09:02:50 +04:00
* ecryptfs_write_begin
2008-03-05 01:29:24 +03:00
* @ file : The eCryptfs file
2008-10-16 09:02:50 +04:00
* @ mapping : The eCryptfs object
* @ pos : The file offset at which to start writing
* @ len : Length of the write
* @ flags : Various flags
* @ pagep : Pointer to return the page
* @ fsdata : Pointer to return fs data ( unused )
2008-03-05 01:29:24 +03:00
*
* This function must zero any hole we create
*
* Returns zero on success ; non - zero otherwise
*/
2008-10-16 09:02:50 +04:00
static int ecryptfs_write_begin ( struct file * file ,
struct address_space * mapping ,
loff_t pos , unsigned len , unsigned flags ,
struct page * * pagep , void * * fsdata )
2006-10-04 13:16:22 +04:00
{
2008-10-16 09:02:50 +04:00
pgoff_t index = pos > > PAGE_CACHE_SHIFT ;
struct page * page ;
2007-12-18 03:20:10 +03:00
loff_t prev_page_end_size ;
2008-03-05 01:29:24 +03:00
int rc = 0 ;
2006-10-04 13:16:22 +04:00
fs: symlink write_begin allocation context fix
With the write_begin/write_end aops, page_symlink was broken because it
could no longer pass a GFP_NOFS type mask into the point where the
allocations happened. They are done in write_begin, which would always
assume that the filesystem can be entered from reclaim. This bug could
cause filesystem deadlocks.
The funny thing with having a gfp_t mask there is that it doesn't really
allow the caller to arbitrarily tinker with the context in which it can be
called. It couldn't ever be GFP_ATOMIC, for example, because it needs to
take the page lock. The only thing any callers care about is __GFP_FS
anyway, so turn that into a single flag.
Add a new flag for write_begin, AOP_FLAG_NOFS. Filesystems can now act on
this flag in their write_begin function. Change __grab_cache_page to
accept a nofs argument as well, to honour that flag (while we're there,
change the name to grab_cache_page_write_begin which is more instructive
and does away with random leading underscores).
This is really a more flexible way to go in the end anyway -- if a
filesystem happens to want any extra allocations aside from the pagecache
ones in ints write_begin function, it may now use GFP_KERNEL (rather than
GFP_NOFS) for common case allocations (eg. ocfs2_alloc_write_ctxt, for a
random example).
[kosaki.motohiro@jp.fujitsu.com: fix ubifs]
[kosaki.motohiro@jp.fujitsu.com: fix fuse]
Signed-off-by: Nick Piggin <npiggin@suse.de>
Reviewed-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: <stable@kernel.org> [2.6.28.x]
Signed-off-by: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
[ Cleaned up the calling convention: just pass in the AOP flags
untouched to the grab_cache_page_write_begin() function. That
just simplifies everybody, and may even allow future expansion of the
logic. - Linus ]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-04 23:00:53 +03:00
page = grab_cache_page_write_begin ( mapping , index , flags ) ;
2008-10-16 09:02:50 +04:00
if ( ! page )
return - ENOMEM ;
* pagep = page ;
2010-11-15 21:43:22 +03:00
prev_page_end_size = ( ( loff_t ) index < < PAGE_CACHE_SHIFT ) ;
2007-10-16 12:28:14 +04:00
if ( ! PageUptodate ( page ) ) {
2008-03-05 01:29:24 +03:00
struct ecryptfs_crypt_stat * crypt_stat =
2010-05-21 18:56:12 +04:00
& ecryptfs_inode_to_private ( mapping - > host ) - > crypt_stat ;
2008-03-05 01:29:24 +03:00
2011-02-23 09:54:20 +03:00
if ( ! ( crypt_stat - > flags & ECRYPTFS_ENCRYPTED ) ) {
2008-03-05 01:29:24 +03:00
rc = ecryptfs_read_lower_page_segment (
2008-10-16 09:02:50 +04:00
page , index , 0 , PAGE_CACHE_SIZE , mapping - > host ) ;
2008-03-05 01:29:24 +03:00
if ( rc ) {
printk ( KERN_ERR " %s: Error attemping to read "
" lower page segment; rc = [%d] \n " ,
2008-04-29 11:59:48 +04:00
__func__ , rc ) ;
2008-03-05 01:29:24 +03:00
ClearPageUptodate ( page ) ;
goto out ;
} else
SetPageUptodate ( page ) ;
} else if ( crypt_stat - > flags & ECRYPTFS_VIEW_AS_ENCRYPTED ) {
if ( crypt_stat - > flags & ECRYPTFS_METADATA_IN_XATTR ) {
rc = ecryptfs_copy_up_encrypted_with_header (
page , crypt_stat ) ;
if ( rc ) {
printk ( KERN_ERR " %s: Error attempting "
" to copy the encrypted content "
" from the lower file whilst "
" inserting the metadata from "
" the xattr into the header; rc "
2008-04-29 11:59:48 +04:00
" = [%d] \n " , __func__ , rc ) ;
2008-03-05 01:29:24 +03:00
ClearPageUptodate ( page ) ;
goto out ;
}
SetPageUptodate ( page ) ;
} else {
rc = ecryptfs_read_lower_page_segment (
2008-10-16 09:02:50 +04:00
page , index , 0 , PAGE_CACHE_SIZE ,
mapping - > host ) ;
2008-03-05 01:29:24 +03:00
if ( rc ) {
printk ( KERN_ERR " %s: Error reading "
" page; rc = [%d] \n " ,
2008-04-29 11:59:48 +04:00
__func__ , rc ) ;
2008-03-05 01:29:24 +03:00
ClearPageUptodate ( page ) ;
goto out ;
}
SetPageUptodate ( page ) ;
}
} else {
2010-11-15 21:43:22 +03:00
if ( prev_page_end_size
> = i_size_read ( page - > mapping - > host ) ) {
zero_user ( page , 0 , PAGE_CACHE_SIZE ) ;
eCryptfs: Avoid unnecessary disk read and data decryption during writing
ecryptfs_write_begin grabs a page from page cache for writing.
If the page contains invalid data, or data older than the
counterpart on the disk, eCryptfs will read out the
corresponing data from the disk into the page, decrypt them,
then perform writing. However, for this page, if the length
of the data to be written into is equal to page size,
that means the whole page of data will be overwritten,
in which case, it does not matter whatever the data were before,
it is beneficial to perform writing directly rather than bothering
to read and decrypt first.
With this optimization, according to our test on a machine with
Intel Core 2 Duo processor, iozone 'write' operation on an existing
file with write size being multiple of page size will enjoy a steady
3x speedup.
Signed-off-by: Li Wang <wangli@kylinos.com.cn>
Signed-off-by: Yunchuan Wen <wenyunchuan@kylinos.com.cn>
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
2012-10-30 15:52:40 +04:00
SetPageUptodate ( page ) ;
} else if ( len < PAGE_CACHE_SIZE ) {
2010-11-15 21:43:22 +03:00
rc = ecryptfs_decrypt_page ( page ) ;
if ( rc ) {
printk ( KERN_ERR " %s: Error decrypting "
" page at index [%ld]; "
" rc = [%d] \n " ,
__func__ , page - > index , rc ) ;
ClearPageUptodate ( page ) ;
goto out ;
}
eCryptfs: Avoid unnecessary disk read and data decryption during writing
ecryptfs_write_begin grabs a page from page cache for writing.
If the page contains invalid data, or data older than the
counterpart on the disk, eCryptfs will read out the
corresponing data from the disk into the page, decrypt them,
then perform writing. However, for this page, if the length
of the data to be written into is equal to page size,
that means the whole page of data will be overwritten,
in which case, it does not matter whatever the data were before,
it is beneficial to perform writing directly rather than bothering
to read and decrypt first.
With this optimization, according to our test on a machine with
Intel Core 2 Duo processor, iozone 'write' operation on an existing
file with write size being multiple of page size will enjoy a steady
3x speedup.
Signed-off-by: Li Wang <wangli@kylinos.com.cn>
Signed-off-by: Yunchuan Wen <wenyunchuan@kylinos.com.cn>
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
2012-10-30 15:52:40 +04:00
SetPageUptodate ( page ) ;
2008-03-05 01:29:24 +03:00
}
}
2007-10-16 12:28:14 +04:00
}
2008-03-05 01:29:24 +03:00
/* If creating a page or more of holes, zero them out via truncate.
* Note , this will increase i_size . */
2008-10-16 09:02:50 +04:00
if ( index ! = 0 ) {
2007-12-18 03:20:10 +03:00
if ( prev_page_end_size > i_size_read ( page - > mapping - > host ) ) {
2007-06-28 01:09:44 +04:00
rc = ecryptfs_truncate ( file - > f_path . dentry ,
2007-12-18 03:20:10 +03:00
prev_page_end_size ) ;
2007-06-28 01:09:44 +04:00
if ( rc ) {
2008-03-05 01:29:24 +03:00
printk ( KERN_ERR " %s: Error on attempt to "
2007-06-28 01:09:44 +04:00
" truncate to (higher) offset [%lld]; "
2008-04-29 11:59:48 +04:00
" rc = [%d] \n " , __func__ ,
2008-03-05 01:29:24 +03:00
prev_page_end_size , rc ) ;
2007-06-28 01:09:44 +04:00
goto out ;
}
2007-05-24 00:58:15 +04:00
}
2007-12-18 03:20:10 +03:00
}
2008-03-05 01:29:24 +03:00
/* Writing to a new page, and creating a small hole from start
* of page ? Zero it out . */
2008-10-16 09:02:50 +04:00
if ( ( i_size_read ( mapping - > host ) = = prev_page_end_size )
& & ( pos ! = 0 ) )
2008-02-05 09:28:29 +03:00
zero_user ( page , 0 , PAGE_CACHE_SIZE ) ;
2006-10-04 13:16:22 +04:00
out :
2011-03-09 20:49:13 +03:00
if ( unlikely ( rc ) ) {
unlock_page ( page ) ;
page_cache_release ( page ) ;
* pagep = NULL ;
}
2006-10-04 13:16:22 +04:00
return rc ;
}
/**
* ecryptfs_write_inode_size_to_header
*
* Writes the lower file size to the first 8 bytes of the header .
*
* Returns zero on success ; non - zero on error .
*/
2007-10-16 12:28:08 +04:00
static int ecryptfs_write_inode_size_to_header ( struct inode * ecryptfs_inode )
2006-10-04 13:16:22 +04:00
{
2007-10-16 12:28:08 +04:00
char * file_size_virt ;
int rc ;
2006-10-04 13:16:22 +04:00
2007-10-16 12:28:08 +04:00
file_size_virt = kmalloc ( sizeof ( u64 ) , GFP_KERNEL ) ;
if ( ! file_size_virt ) {
rc = - ENOMEM ;
2007-03-01 07:12:16 +03:00
goto out ;
}
2008-07-24 08:30:07 +04:00
put_unaligned_be64 ( i_size_read ( ecryptfs_inode ) , file_size_virt ) ;
2007-10-16 12:28:08 +04:00
rc = ecryptfs_write_lower ( ecryptfs_inode , file_size_virt , 0 ,
sizeof ( u64 ) ) ;
kfree ( file_size_virt ) ;
2009-09-17 04:04:20 +04:00
if ( rc < 0 )
2007-10-16 12:28:08 +04:00
printk ( KERN_ERR " %s: Error writing file size to header; "
2008-04-29 11:59:48 +04:00
" rc = [%d] \n " , __func__ , rc ) ;
2009-09-17 04:04:20 +04:00
else
rc = 0 ;
2006-10-04 13:16:22 +04:00
out :
return rc ;
}
2007-10-16 12:28:08 +04:00
struct kmem_cache * ecryptfs_xattr_cache ;
static int ecryptfs_write_inode_size_to_xattr ( struct inode * ecryptfs_inode )
2007-02-12 11:53:46 +03:00
{
ssize_t size ;
void * xattr_virt ;
2007-10-16 12:28:08 +04:00
struct dentry * lower_dentry =
ecryptfs_inode_to_private ( ecryptfs_inode ) - > lower_file - > f_dentry ;
struct inode * lower_inode = lower_dentry - > d_inode ;
2007-02-12 11:53:46 +03:00
int rc ;
2007-10-16 12:28:08 +04:00
if ( ! lower_inode - > i_op - > getxattr | | ! lower_inode - > i_op - > setxattr ) {
printk ( KERN_WARNING
" No support for setting xattr in lower filesystem \n " ) ;
rc = - ENOSYS ;
goto out ;
}
2007-02-12 11:53:46 +03:00
xattr_virt = kmem_cache_alloc ( ecryptfs_xattr_cache , GFP_KERNEL ) ;
if ( ! xattr_virt ) {
printk ( KERN_ERR " Out of memory whilst attempting to write "
" inode size to xattr \n " ) ;
rc = - ENOMEM ;
goto out ;
}
2007-10-16 12:28:08 +04:00
mutex_lock ( & lower_inode - > i_mutex ) ;
size = lower_inode - > i_op - > getxattr ( lower_dentry , ECRYPTFS_XATTR_NAME ,
xattr_virt , PAGE_CACHE_SIZE ) ;
2007-02-12 11:53:46 +03:00
if ( size < 0 )
size = 8 ;
2008-07-24 08:30:07 +04:00
put_unaligned_be64 ( i_size_read ( ecryptfs_inode ) , xattr_virt ) ;
2007-10-16 12:28:08 +04:00
rc = lower_inode - > i_op - > setxattr ( lower_dentry , ECRYPTFS_XATTR_NAME ,
xattr_virt , size , 0 ) ;
mutex_unlock ( & lower_inode - > i_mutex ) ;
2007-02-12 11:53:46 +03:00
if ( rc )
printk ( KERN_ERR " Error whilst attempting to write inode size "
" to lower file xattr; rc = [%d] \n " , rc ) ;
kmem_cache_free ( ecryptfs_xattr_cache , xattr_virt ) ;
out :
return rc ;
}
2007-10-16 12:28:08 +04:00
int ecryptfs_write_inode_size_to_metadata ( struct inode * ecryptfs_inode )
2007-02-12 11:53:46 +03:00
{
struct ecryptfs_crypt_stat * crypt_stat ;
2007-10-16 12:28:08 +04:00
crypt_stat = & ecryptfs_inode_to_private ( ecryptfs_inode ) - > crypt_stat ;
2009-04-14 00:29:27 +04:00
BUG_ON ( ! ( crypt_stat - > flags & ECRYPTFS_ENCRYPTED ) ) ;
2007-02-12 11:53:46 +03:00
if ( crypt_stat - > flags & ECRYPTFS_METADATA_IN_XATTR )
2007-10-16 12:28:08 +04:00
return ecryptfs_write_inode_size_to_xattr ( ecryptfs_inode ) ;
2007-02-12 11:53:46 +03:00
else
2007-10-16 12:28:08 +04:00
return ecryptfs_write_inode_size_to_header ( ecryptfs_inode ) ;
2007-02-12 11:53:46 +03:00
}
2006-10-04 13:16:22 +04:00
/**
2008-10-16 09:02:50 +04:00
* ecryptfs_write_end
2006-10-04 13:16:22 +04:00
* @ file : The eCryptfs file object
2008-10-16 09:02:50 +04:00
* @ mapping : The eCryptfs object
* @ pos : The file position
* @ len : The length of the data ( unused )
* @ copied : The amount of data copied
2006-10-04 13:16:22 +04:00
* @ page : The eCryptfs page
2008-10-16 09:02:50 +04:00
* @ fsdata : The fsdata ( unused )
2006-10-04 13:16:22 +04:00
*/
2008-10-16 09:02:50 +04:00
static int ecryptfs_write_end ( struct file * file ,
struct address_space * mapping ,
loff_t pos , unsigned len , unsigned copied ,
struct page * page , void * fsdata )
2006-10-04 13:16:22 +04:00
{
2008-10-16 09:02:50 +04:00
pgoff_t index = pos > > PAGE_CACHE_SHIFT ;
unsigned from = pos & ( PAGE_CACHE_SIZE - 1 ) ;
unsigned to = from + copied ;
struct inode * ecryptfs_inode = mapping - > host ;
2007-10-16 12:28:11 +04:00
struct ecryptfs_crypt_stat * crypt_stat =
2010-05-21 18:56:12 +04:00
& ecryptfs_inode_to_private ( ecryptfs_inode ) - > crypt_stat ;
2006-10-04 13:16:22 +04:00
int rc ;
ecryptfs_printk ( KERN_DEBUG , " Calling fill_zeros_to_end_of_page "
2010-11-11 02:46:16 +03:00
" (page w/ index = [0x%.16lx], to = [%d]) \n " , index , to ) ;
2009-04-14 00:29:27 +04:00
if ( ! ( crypt_stat - > flags & ECRYPTFS_ENCRYPTED ) ) {
rc = ecryptfs_write_lower_page_segment ( ecryptfs_inode , page , 0 ,
to ) ;
if ( ! rc ) {
rc = copied ;
fsstack_copy_inode_size ( ecryptfs_inode ,
ecryptfs_inode_to_lower ( ecryptfs_inode ) ) ;
}
goto out ;
}
eCryptfs: Avoid unnecessary disk read and data decryption during writing
ecryptfs_write_begin grabs a page from page cache for writing.
If the page contains invalid data, or data older than the
counterpart on the disk, eCryptfs will read out the
corresponing data from the disk into the page, decrypt them,
then perform writing. However, for this page, if the length
of the data to be written into is equal to page size,
that means the whole page of data will be overwritten,
in which case, it does not matter whatever the data were before,
it is beneficial to perform writing directly rather than bothering
to read and decrypt first.
With this optimization, according to our test on a machine with
Intel Core 2 Duo processor, iozone 'write' operation on an existing
file with write size being multiple of page size will enjoy a steady
3x speedup.
Signed-off-by: Li Wang <wangli@kylinos.com.cn>
Signed-off-by: Yunchuan Wen <wenyunchuan@kylinos.com.cn>
Signed-off-by: Tyler Hicks <tyhicks@canonical.com>
2012-10-30 15:52:40 +04:00
if ( ! PageUptodate ( page ) ) {
if ( copied < PAGE_CACHE_SIZE ) {
rc = 0 ;
goto out ;
}
SetPageUptodate ( page ) ;
}
2007-10-16 12:28:11 +04:00
/* Fills in zeros if 'to' goes beyond inode size */
2006-10-04 13:16:22 +04:00
rc = fill_zeros_to_end_of_page ( page , to ) ;
if ( rc ) {
ecryptfs_printk ( KERN_WARNING , " Error attempting to fill "
2010-11-11 02:46:16 +03:00
" zeros in page with index = [0x%.16lx] \n " , index ) ;
2006-10-04 13:16:22 +04:00
goto out ;
}
2012-07-04 03:50:57 +04:00
rc = ecryptfs_encrypt_page ( page ) ;
if ( rc ) {
ecryptfs_printk ( KERN_WARNING , " Error encrypting page (upper "
" index [0x%.16lx]) \n " , index ) ;
goto out ;
}
2008-10-16 09:02:50 +04:00
if ( pos + copied > i_size_read ( ecryptfs_inode ) ) {
i_size_write ( ecryptfs_inode , pos + copied ) ;
2006-10-04 13:16:22 +04:00
ecryptfs_printk ( KERN_DEBUG , " Expanded file size to "
2010-11-11 02:46:16 +03:00
" [0x%.16llx] \n " ,
( unsigned long long ) i_size_read ( ecryptfs_inode ) ) ;
2006-10-04 13:16:22 +04:00
}
2012-07-04 03:50:57 +04:00
rc = ecryptfs_write_inode_size_to_metadata ( ecryptfs_inode ) ;
if ( rc )
printk ( KERN_ERR " Error writing inode size to metadata; "
" rc = [%d] \n " , rc ) ;
else
rc = copied ;
2006-10-04 13:16:22 +04:00
out :
2012-07-04 03:50:57 +04:00
unlock_page ( page ) ;
2008-10-16 09:02:50 +04:00
page_cache_release ( page ) ;
2006-10-04 13:16:22 +04:00
return rc ;
}
static sector_t ecryptfs_bmap ( struct address_space * mapping , sector_t block )
{
int rc = 0 ;
struct inode * inode ;
struct inode * lower_inode ;
inode = ( struct inode * ) mapping - > host ;
lower_inode = ecryptfs_inode_to_lower ( inode ) ;
if ( lower_inode - > i_mapping - > a_ops - > bmap )
rc = lower_inode - > i_mapping - > a_ops - > bmap ( lower_inode - > i_mapping ,
block ) ;
return rc ;
}
2009-09-22 04:01:10 +04:00
const struct address_space_operations ecryptfs_aops = {
2006-10-04 13:16:22 +04:00
. writepage = ecryptfs_writepage ,
. readpage = ecryptfs_readpage ,
2008-10-16 09:02:50 +04:00
. write_begin = ecryptfs_write_begin ,
. write_end = ecryptfs_write_end ,
2006-10-04 13:16:22 +04:00
. bmap = ecryptfs_bmap ,
} ;