fscrypt currently only supports AES encryption. However, many low-end
mobile devices have older CPUs that don't have AES instructions, e.g.
the ARMv8 Cryptography Extensions. Currently, user data on such devices
is not encrypted at rest because AES is too slow, even when the NEON
bit-sliced implementation of AES is used. Unfortunately, it is
infeasible to encrypt these devices at all when AES is the only option.
Therefore, this patch updates fscrypt to support the Speck block cipher,
which was recently added to the crypto API. The C implementation of
Speck is not especially fast, but Speck can be implemented very
efficiently with general-purpose vector instructions, e.g. ARM NEON.
For example, on an ARMv7 processor, we measured the NEON-accelerated
Speck128/256-XTS at 69 MB/s for both encryption and decryption, while
AES-256-XTS with the NEON bit-sliced implementation was only 22 MB/s
encryption and 19 MB/s decryption.
There are multiple variants of Speck. This patch only adds support for
Speck128/256, which is the variant with a 128-bit block size and 256-bit
key size -- the same as AES-256. This is believed to be the most secure
variant of Speck, and it's only about 6% slower than Speck128/128.
Speck64/128 would be at least 20% faster because it has 20% rounds, and
it can be even faster on CPUs that can't efficiently do the 64-bit
operations needed for Speck128. However, Speck64's 64-bit block size is
not preferred security-wise. ARM NEON also supports the needed 64-bit
operations even on 32-bit CPUs, resulting in Speck128 being fast enough
for our targeted use cases so far.
The chosen modes of operation are XTS for contents and CTS-CBC for
filenames. These are the same modes of operation that fscrypt defaults
to for AES. Note that as with the other fscrypt modes, Speck will not
be used unless userspace chooses to use it. Nor are any of the existing
modes (which are all AES-based) being removed, of course.
We intentionally don't make CONFIG_FS_ENCRYPTION select
CONFIG_CRYPTO_SPECK, so people will have to enable Speck support
themselves if they need it. This is because we shouldn't bloat the
FS_ENCRYPTION dependencies with every new cipher, especially ones that
aren't recommended for most users. Moreover, CRYPTO_SPECK is just the
generic implementation, which won't be fast enough for many users; in
practice, they'll need to enable CRYPTO_SPECK_NEON to get acceptable
performance.
More details about our choice of Speck can be found in our patches that
added Speck to the crypto API, and the follow-on discussion threads.
We're planning a publication that explains the choice in more detail.
But briefly, we can't use ChaCha20 as we previously proposed, since it
would be insecure to use a stream cipher in this context, with potential
IV reuse during writes on f2fs and/or on wear-leveling flash storage.
We also evaluated many other lightweight and/or ARX-based block ciphers
such as Chaskey-LTS, RC5, LEA, CHAM, Threefish, RC6, NOEKEON, SPARX, and
XTEA. However, all had disadvantages vs. Speck, such as insufficient
performance with NEON, much less published cryptanalysis, or an
insufficient security level. Various design choices in Speck make it
perform better with NEON than competing ciphers while still having a
security margin similar to AES, and in the case of Speck128 also the
same available security levels. Unfortunately, Speck does have some
political baggage attached -- it's an NSA designed cipher, and was
rejected from an ISO standard (though for context, as far as I know none
of the above-mentioned alternatives are ISO standards either).
Nevertheless, we believe it is a good solution to the problem from a
technical perspective.
Certain algorithms constructed from ChaCha or the ChaCha permutation,
such as MEM (Masked Even-Mansour) or HPolyC, may also meet our
performance requirements. However, these are new constructions that
need more time to receive the cryptographic review and acceptance needed
to be confident in their security. HPolyC hasn't been published yet,
and we are concerned that MEM makes stronger assumptions about the
underlying permutation than the ChaCha stream cipher does. In contrast,
the XTS mode of operation is relatively well accepted, and Speck has
over 70 cryptanalysis papers. Of course, these ChaCha-based algorithms
can still be added later if they become ready.
The best known attack on Speck128/256 is a differential cryptanalysis
attack on 25 of 34 rounds with 2^253 time complexity and 2^125 chosen
plaintexts, i.e. only marginally faster than brute force. There is no
known attack on the full 34 rounds.
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
(cherry-picked from commit 12d28f7955https://git.kernel.org/pub/scm/linux/kernel/git/tytso/fscrypt.git master)
(dropped Documentation/filesystems/fscrypt.rst change)
(fixed merge conflict in fs/crypto/keyinfo.c)
(also ported change to fs/ext4/, which isn't using fs/crypto/ in this
kernel version)
Change-Id: I62c632044dfd06a2c5b74c2fb058f9c3b8af0add
Signed-off-by: Eric Biggers <ebiggers@google.com>
When uid_base_stuff has no entries, proc_uid_base_readdir tries to
compute an address before the start of the array. Revise this check to
use uid_base_stuff + nents instead, which makes the code valid
regardless of array size.
Bug: 80158484
Test: No more compiler warning with CONFIG_CPU_FREQ_TIMES=n
Change-Id: I6e55b27c3ba8210cee194f6d27bbd62c0b263796
Signed-off-by: Connor O'Brien <connoro@google.com>
The vdso{32,64}.so can fail to build when CC=clang when clang tries to
find a suitable GCC toolchain to link these libraries with.
/usr/bin/ld: arch/x86/entry/vdso/vclock_gettime.o: access beyond end of merged section (782)
This happens because the host environment leaked into the CROSS_COMPILE
environment due to the way clang searches for suitable GCC toolchains.
Most of the time this goes unnoticed because the host linker is new
enough to work anyway, but on this particular machine it was not.
Extract the needed --target and --gcc-toolchain flags added in the top
level Makefile from KBUILD_CFLAGS.
Bug: 63889157
Change-Id: If7d4097d1d2eaf95f18d0295483bde8792a06844
Signed-off-by: Alistair Strachan <astrachan@google.com>
The last upgrade introduced a new build failure, because it had a bug
which caused it to emit PLT relocations, certain types of which cannot
be handled by the reloc tool in the kernel.
See https://bugs.llvm.org/show_bug.cgi?id=36674 for more details.
Bug: 63889157
Change-Id: I813febdbacb0579abcb12dc7f2164cce1e2f5a26
Signed-off-by: Alistair Strachan <astrachan@google.com>
Use the same clang version as hikey-linaro.
Bug: 63889157
Change-Id: I6932d6149642d429086207e63aa8a8d5c2afd6f7
Signed-off-by: Alistair Strachan <astrachan@google.com>
Reconcile with changes made to the kernel manifest. Clang must come from
master because it was not usable for kernel builds in older branches of
the Android platform.
Bug: 63889157
Change-Id: Id0a080fc2f1cba495f37f26afa48e43e736b756a
Signed-off-by: Alistair Strachan <astrachan@google.com>
We've seen in-field reports showing _lots_ (18 in one case, 41 in
another) of tasks all sitting there blocked on:
mutex_lock+0x4c/0x68
dm_bufio_shrink_count+0x38/0x78
shrink_slab.part.54.constprop.65+0x100/0x464
shrink_zone+0xa8/0x198
In the two cases analyzed, we see one task that looks like this:
Workqueue: kverityd verity_prefetch_io
__switch_to+0x9c/0xa8
__schedule+0x440/0x6d8
schedule+0x94/0xb4
schedule_timeout+0x204/0x27c
schedule_timeout_uninterruptible+0x44/0x50
wait_iff_congested+0x9c/0x1f0
shrink_inactive_list+0x3a0/0x4cc
shrink_lruvec+0x418/0x5cc
shrink_zone+0x88/0x198
try_to_free_pages+0x51c/0x588
__alloc_pages_nodemask+0x648/0xa88
__get_free_pages+0x34/0x7c
alloc_buffer+0xa4/0x144
__bufio_new+0x84/0x278
dm_bufio_prefetch+0x9c/0x154
verity_prefetch_io+0xe8/0x10c
process_one_work+0x240/0x424
worker_thread+0x2fc/0x424
kthread+0x10c/0x114
...and that looks to be the one holding the mutex.
The problem has been reproduced on fairly easily:
0. Be running Chrome OS w/ verity enabled on the root filesystem
1. Pick test patch: http://crosreview.com/412360
2. Install launchBalloons.sh and balloon.arm from
http://crbug.com/468342
...that's just a memory stress test app.
3. On a 4GB rk3399 machine, run
nice ./launchBalloons.sh 4 900 100000
...that tries to eat 4 * 900 MB of memory and keep accessing.
4. Login to the Chrome web browser and restore many tabs
With that, I've seen printouts like:
DOUG: long bufio 90758 ms
...and stack trace always show's we're in dm_bufio_prefetch().
The problem is that we try to allocate memory with GFP_NOIO while
we're holding the dm_bufio lock. Instead we should be using
GFP_NOWAIT. Using GFP_NOIO can cause us to sleep while holding the
lock and that causes the above problems.
The current behavior explained by David Rientjes:
It will still try reclaim initially because __GFP_WAIT (or
__GFP_KSWAPD_RECLAIM) is set by GFP_NOIO. This is the cause of
contention on dm_bufio_lock() that the thread holds. You want to
pass GFP_NOWAIT instead of GFP_NOIO to alloc_buffer() when holding a
mutex that can be contended by a concurrent slab shrinker (if
count_objects didn't use a trylock, this pattern would trivially
deadlock).
This change significantly increases responsiveness of the system while
in this state. It makes a real difference because it unblocks kswapd.
In the bug report analyzed, kswapd was hung:
kswapd0 D ffffffc000204fd8 0 72 2 0x00000000
Call trace:
[<ffffffc000204fd8>] __switch_to+0x9c/0xa8
[<ffffffc00090b794>] __schedule+0x440/0x6d8
[<ffffffc00090bac0>] schedule+0x94/0xb4
[<ffffffc00090be44>] schedule_preempt_disabled+0x28/0x44
[<ffffffc00090d900>] __mutex_lock_slowpath+0x120/0x1ac
[<ffffffc00090d9d8>] mutex_lock+0x4c/0x68
[<ffffffc000708e7c>] dm_bufio_shrink_count+0x38/0x78
[<ffffffc00030b268>] shrink_slab.part.54.constprop.65+0x100/0x464
[<ffffffc00030dbd8>] shrink_zone+0xa8/0x198
[<ffffffc00030e578>] balance_pgdat+0x328/0x508
[<ffffffc00030eb7c>] kswapd+0x424/0x51c
[<ffffffc00023f06c>] kthread+0x10c/0x114
[<ffffffc000203dd0>] ret_from_fork+0x10/0x40
By unblocking kswapd memory pressure should be reduced.
Change-Id: I10da1bcb02160d75320c16259a54b5de4aafede1
Suggested-by: David Rientjes <rientjes@google.com>
Reviewed-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
(cherry picked from commit 9ea61cac0b)
Signed-off-by: Minchan Kim <minchan@google.com>
After d_revalidate returns 0, the vfs will call
d_invalidate, which will call d_drop itself, along
with other cleanup.
Bug: 78262592
Change-Id: Idbb30e008c05d62edf2217679cb6a5517d8d1a2c
Signed-off-by: Daniel Rosenberg <drosen@google.com>
goldfish_dma_ioctl_getoff and goldfish_dma_ioctl_create_region
are not used outside of goldfish_pipe_v2.c - mark them
static.
Bug: 72717639
Bug: 66884503
Signed-off-by: Roman Kiryanov <rkir@google.com>
Change-Id: Id4fe524cc4db9d36b0c794ad9f34356f396153b4
This change improves the pipe performance by removing unnesessary
memory copying.
Bug: 72717639
Bug: 66884503
Change-Id: I0d279f682039e411faf4212713d82ec355c3e9ee
Signed-off-by: Roman Kiryanov <rkir@google.com>
Signed-off-by: Lingfeng Yang <lfy@google.com>
This patch clears PageError in some pages tagged by read path, but when we
write the pages with valid contents, writepage should clear the bit likewise
ext4.
Change-Id: I7d599ea65d0e30e5faa1cbfb3e1309f62511202b
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
(cherry picked from commit a44b418c31)
pageout() in MM traslates EAGAIN, so calls handle_write_error()
-> mapping_set_error() -> set_bit(AS_EIO, ...).
file_write_and_wait_range() will see EIO error, which is critical
to return value of fsync() followed by atomic_write failure to user.
Change-Id: I6124ed299baf5e5b8993bb6b54f07bb8771eda8c
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
(cherry picked from commit 520a948618)
cfg80211_chandef_create() expects an 'enum nl80211_channel_type' as
channel type however in ieee80211_sta_join_ibss()
NL80211_CHAN_WIDTH_20_NOHT is passed in two occasions, which is of
the enum type 'nl80211_chan_width'. Change the value to NL80211_CHAN_NO_HT
(20 MHz, non-HT channel) of the channel type enum.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
(cherry picked from commit a4ac6f2e53)
Bug: 78886293
Change-Id: Iab3fc750c2cd8b32cf501b9739445a7b5c83e8cd
Signed-off-by: Alistair Strachan <astrachan@google.com>
When clang detects a non-boolean constant in a logical operation it
generates a 'constant-logical-operand' warning. In
ieee80211_try_rate_control_ops_get() the result of strlen(<const str>)
is used in a logical operation, clang resolves the expression to an
(integer) constant at compile time when clang's builtin strlen function
is used.
Change the condition to check for strlen() > 0 to make the constant
operand boolean and thus avoid the warning.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
(cherry picked from commit 93f56de259)
Bug: 78886293
Change-Id: Ia819eb188699c1d81047c0dfa143da52c6cb490c
Signed-off-by: Alistair Strachan <astrachan@google.com>
rate_flg is of type 'enum nl80211_attrs', however it is assigned with
'enum nl80211_rate_info' values. Change the type of rate_flg accordingly.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Johannes Berg <johannes.berg@intel.com>
(cherry picked from commit bbf67e450a)
Bug: 78886293
Change-Id: I5423e226dc65375245f8ca6261c025d967552f17
Signed-off-by: Alistair Strachan <astrachan@google.com>
sysfs_get_dirent is usually invoked with a string literal, which
have the type char[]. While the toplevel Makefile
disables -Wpointer-sign, other Makefiles like
arch/x86/boot/compressed/Makefile
redefine KBUILD_CFLAGS. Fixes the warning:
In file included from arch/x86/boot/compressed/kaslr.c:17:
In file included from ./include/linux/module.h:17:
In file included from ./include/linux/kobject.h:21:
./include/linux/sysfs.h:517:37: warning: passing 'const unsigned char *'
to parameter of
type 'const char *' converts between pointers to integer types
with different sign
[-Wpointer-sign]
return kernfs_find_and_get(parent, name);
^~~~
./include/linux/kernfs.h:462:57: note: passing argument to parameter
'name' here
kernfs_find_and_get(struct kernfs_node *kn, const char *name)
^
Signed-off-by: Nick Desaulniers <nick.desaulniers@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 89cf2a20c3)
Bug: 78886293
Change-Id: Ic03f7b132fbf67b3543462448aa3f3633eae49a6
Signed-off-by: Alistair Strachan <astrachan@google.com>
This fixes the following clang warning:
kernel/trace/trace.c:3231:12: warning: address of array 'iter->started'
will always evaluate to 'true' [-Wpointer-bool-conversion]
if (iter->started)
Link: http://lkml.kernel.org/r/20170421234110.117075-1-mka@chromium.org
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
(cherry picked from commit 4dbbe2d8e9)
Bug: 78886293
Change-Id: Ib17a68ce55ca80b04bdea2d232f8ca9f88b1b8a3
Signed-off-by: Alistair Strachan <astrachan@google.com>
Apparently netpoll_setup() assumes that netpoll.dev_name is a pointer
when checking if the device name is set:
if (np->dev_name) {
...
However the field is a character array, therefore the condition always
yields true. Check instead whether the first byte of the array has a
non-zero value.
Signed-off-by: Matthias Kaehlcke <mka@chromium.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit 0c3a8f8b8f)
Bug: 78886293
Change-Id: I1a6eec091c4bab5769a3519196f529030a71b6dd
Signed-off-by: Alistair Strachan <astrachan@google.com>
Fix "warning: cast to pointer from integer of different size" when
printing the region shm physical address. Use the %pa conversion
specifier and pass the resource by reference.
Signed-off-by: Alistair Strachan <astrachan@google.com>
[sent upstream via staging https://patchwork.kernel.org/patch/10376967/]
Bug: 702147120
Change-Id: Ibb0c32b461821235155c464bfd5c5963db512734
Map the region shm as write-combining instead of uncachable.
Signed-off-by: Alistair Strachan <astrachan@google.com>
[sent upstream via staging https://patchwork.kernel.org/patch/10376965/]
Bug: 702147120
Change-Id: If587b4f26abaa3580b889b04dbb655f8d5b4a8ee
We have a function around a pair of writel to
simplify 64-bit support.
Bug: 72717639
bug: 66884503
Change-Id: Id23d23f6998e7be348646bb4ea9884f165c1149b
Signed-off-by: Roman Kiryanov <rkir@google.com>
The dev_ functions are the same as the corresponding
pr_ functions, but also print identifying information
about the struct device.
Bug: 72717639
bug: 66884503
Change-Id: I0b1147cf607eb5b0b07dd9753dcf2b60e9700afe
Signed-off-by: Roman Kiryanov <rkir@google.com>
Change MODULE_LICENSE() string to properly match
the license for the driver (goldfish_pipe.h).
Bug: 72717639
Bug: 66884503
Change-Id: I4d1ef778276b26b49c83c967c3c9314b2c0c5da7
Signed-off-by: Roman Kiryanov <rkir@google.com>
Return an error instead of crashing in signalled_pipes_add_locked.
Bug: 72717639
Bug: 66884503
Change-Id: I811ad1932f1600f8bbe4598cdaf206bd96ea921a
Signed-off-by: Roman Kiryanov <rkir@google.com>
Casting twice is not required.
Bug: 72717639
Bug: 66884503
Change-Id: I3420388683a9746f2d2110af51d9d25c12c7eea6
Signed-off-by: Roman Kiryanov <rkir@google.com>
Replace the 'goldfish' prefix with 'goldfish_pipe' to
say they are pipe functions.
Bug: 72717639
Bug: 66884503
Change-Id: I5a5cf7ee38cf2ae193877b1ffac19eadb15a374a
Signed-off-by: Roman Kiryanov <rkir@google.com>
We don't need an array of 1 for pipe_dev and
use better names to distinguish between
goldfish_pipe_dev and miscdevice.
Bug: 72717639
Bug: 66884503
Change-Id: Iab040c158745f034ca8e9569fd49c84933b1c4ba
Signed-off-by: Roman Kiryanov <rkir@google.com>
To separate variable declarations from executable code
and to improve readability.
Bug: 72717639
Bug: 66884503
Change-Id: I46fb70b13b8e3d061dfc2288f5720a379f1f39a9
Signed-off-by: Roman Kiryanov <rkir@google.com>
We get a build error in the irqsoff tracer in some configurations:
kernel/trace/trace_irqsoff.c: In function 'trace_preempt_on':
kernel/trace/trace_irqsoff.c:855:2: error: implicit declaration of function 'trace_preempt_enable_rcuidle'; did you mean 'trace_irq_enable_rcuidle'? [-Werror=implicit-function-declaration]
trace_preempt_enable_rcuidle(a0, a1);
The problem is that trace_preempt_enable_rcuidle() has different
definition based on multiple Kconfig symbols, but not all combinations
have a valid definition.
This changes the conditions so that we always get exactly one
definition of each of the four tracing macros. I have not tried
to verify that these definitions are sensible, but now we
can build all randconfig combinations again.
Link: http://lkml.kernel.org/r/20171019083230.2450779-1-arnd@arndb.de
Change-Id: I28715af208379e993df85c2fb35549290f4fbd6e
Fixes: d59158162e ("tracing: Add support for preempt and irq enable/disable events")
Acked-by: Joel Fernandes <joelaf@google.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Steven Rostedt (VMware) <rostedt@goodmis.org>
Some debugging messages did not use %pK, but since those messages are
not very useful and have been removed upstream, just remove them
instead.
Bug: 77937819
Change-Id: Ie45897fe2d6ec3f842a02883e8ec929ed2e76933
Signed-off-by: Alistair Strachan <astrachan@google.com>
d_make_root will call iput on failure, so we
shouldn't try to do that ourselves.
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug: 77923821
Change-Id: I1abb4afb0f894ab917b7c6be8c833676f436beb7
When an sdcardfs dentry is destroyed, it may not yet
have its fsdata initialized. It must be checked before
we try to access the paths in its private data.
Additionally, when cleaning up the superblock after
a failure, we don't have our sb private data, so
check for that case.
Bug: 77923821
Change-Id: I89caf6e121ed86480b42024664453fe0031bbcf3
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Cherry-pick from origin/upstream-f2fs-stable-linux-4.4.y:
commit f819874f58 ("f2fs: check cap_resource only for data blocks")
This patch changes the rule to check cap_resource for data blocks, not inode
or node blocks in order to avoid selinux denial.
Change-Id: I875d7ccf7cce7b833a1c11cb0eef0b504b823c4a
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Cherry-pick from origin/upstream-f2fs-stable-linux-4.4.y:
commit 3e7a141175 ("Revert "f2fs: introduce f2fs_set_page_dirty_nobuffer"")
This patch reverts copied f2fs_set_page_dirty_nobuffer to use generic function
for stability.
This reverts commit fe76b796fc.
Change-Id: I3d4728d894d1af41a2f1e30ebc375907abd5ffc8
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Cherry-pick from origin/upstream-f2fs-stable-linux-4.4.y:
commit 070da80085 ("f2fs: clear PageError on writepage")
This patch clears PageError in some pages tagged by read path, but when we
write the pages with valid contents, writepage should clear the bit likewise
ext4.
Change-Id: I7272074f2bb9c81fc43e37074b44e9d761756263
Reviewed-by: Chao Yu <yuchao0@huawei.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
hrtimer_init_on_stack() needs a matching call to
destroy_hrtimer_on_stack(), so both need to be exported.
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
(cherry picked from commit c08376ac97)
[astrachan: Fixes i386-allmodconfig build failure in vsoc.ko noticed
by 01.org kbuild-all project building kernel/msm]
Bug: 78036828
Change-Id: If4d5c466255019322ea21ef38ee5b1b382cce969
Signed-off-by: Alistair Strachan <astrachan@google.com>
This allows platforms that are CPU/memory contrained to verify data
blocks only the first time they are read from the data device, rather
than every time. As such, it provides a reduced level of security
because only offline tampering of the data device's content will be
detected, not online tampering.
Hash blocks are still verified each time they are read from the hash
device, since verification of hash blocks is less performance critical
than data blocks, and a hash block will not be verified any more after
all the data blocks it covers have been verified anyway.
This option introduces a bitset that is used to check if a block has
been validated before or not. A block can be validated more than once
as there is no thread protection for the bitset.
These changes were developed and tested on entry-level Android Go
devices.
(change to use vzalloc/vfree)
(change per_bio_data_size variable)
Bug: 72664474
Change-Id: I2db2db26c7897c78407789de2cc4cff3587c84bb
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
(cherry picked from commit 843f38d382)
Signed-off-by: Patrik Torstensson <totte@google.com>
Cherry-pick from origin/upstream-f2fs-stable-linux-4.4.y:
dafecc032e ("f2fs: call unlock_new_inode() before d_instantiate()")
xfstest generic/429 sometimes hangs on f2fs, caused by a thread being
unable to take a directory's i_rwsem for write in vfs_rmdir(). In the
test, one thread repeatedly creates and removes a directory, and other
threads repeatedly look up a file in the directory. The bug is that
f2fs_mkdir() calls d_instantiate() before unlock_new_inode(), resulting
in the directory inode being exposed to lookups before it has been fully
initialized. And with CONFIG_DEBUG_LOCK_ALLOC, unlock_new_inode()
reinitializes ->i_rwsem, corrupting its state when it is already held.
Fix it by calling unlock_new_inode() before d_instantiate(). This
matches what other filesystems do.
Fixes: 57397d86c6 ("f2fs: add inode operations for special inodes")
Change-Id: I1c50a45449b89ce590ae2caccb7d499f4b717665
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
Cherry-pick from origin/upstream-f2fs-stable-linux-4.4.y:
c18b4f60c8 ("f2fs: refactor read path to allow multiple postprocessing steps")
Currently f2fs's ->readpage() and ->readpages() assume that either the
data undergoes no postprocessing, or decryption only. But with
fs-verity, there will be an additional authenticity verification step,
and it may be needed either by itself, or combined with decryption.
To support this, store a 'struct bio_post_read_ctx' in ->bi_private
which contains a work struct, a bitmask of postprocessing steps that are
enabled, and an indicator of the current step. The bio completion
routine, if there was no I/O error, enqueues the first postprocessing
step. When that completes, it continues to the next step. Pages that
fail any postprocessing step have PageError set. Once all steps have
completed, pages without PageError set are set Uptodate, and all pages
are unlocked.
Also replace f2fs_encrypted_file() with a new function
f2fs_post_read_required() in places like direct I/O and garbage
collection that really should be testing whether the file needs special
I/O processing, not whether it is encrypted specifically.
This may also be useful for other future f2fs features such as
compression.
Change-Id: I742be348b9dfc2113200bcc5366a84e978371a54
Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org>
