The poll() changes were not well thought out, and completely
unexplained. They also caused a huge performance regression, because
"->poll()" was no longer a trivial file operation that just called down
to the underlying file operations, but instead did at least two indirect
calls.
Indirect calls are sadly slow now with the Spectre mitigation, but the
performance problem could at least be largely mitigated by changing the
"->get_poll_head()" operation to just have a per-file-descriptor pointer
to the poll head instead. That gets rid of one of the new indirections.
But that doesn't fix the new complexity that is completely unwarranted
for the regular case. The (undocumented) reason for the poll() changes
was some alleged AIO poll race fixing, but we don't make the common case
slower and more complex for some uncommon special case, so this all
really needs way more explanations and most likely a fundamental
redesign.
[ This revert is a revert of about 30 different commits, not reverted
individually because that would just be unnecessarily messy - Linus ]
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: Christoph Hellwig <hch@lst.de>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull poll annotations from Al Viro:
"This introduces a __bitwise type for POLL### bitmap, and propagates
the annotations through the tree. Most of that stuff is as simple as
'make ->poll() instances return __poll_t and do the same to local
variables used to hold the future return value'.
Some of the obvious brainos found in process are fixed (e.g. POLLIN
misspelled as POLL_IN). At that point the amount of sparse warnings is
low and most of them are for genuine bugs - e.g. ->poll() instance
deciding to return -EINVAL instead of a bitmap. I hadn't touched those
in this series - it's large enough as it is.
Another problem it has caught was eventpoll() ABI mess; select.c and
eventpoll.c assumed that corresponding POLL### and EPOLL### were
equal. That's true for some, but not all of them - EPOLL### are
arch-independent, but POLL### are not.
The last commit in this series separates userland POLL### values from
the (now arch-independent) kernel-side ones, converting between them
in the few places where they are copied to/from userland. AFAICS, this
is the least disruptive fix preserving poll(2) ABI and making epoll()
work on all architectures.
As it is, it's simply broken on sparc - try to give it EPOLLWRNORM and
it will trigger only on what would've triggered EPOLLWRBAND on other
architectures. EPOLLWRBAND and EPOLLRDHUP, OTOH, are never triggered
at all on sparc. With this patch they should work consistently on all
architectures"
* 'misc.poll' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (37 commits)
make kernel-side POLL... arch-independent
eventpoll: no need to mask the result of epi_item_poll() again
eventpoll: constify struct epoll_event pointers
debugging printk in sg_poll() uses %x to print POLL... bitmap
annotate poll(2) guts
9p: untangle ->poll() mess
->si_band gets POLL... bitmap stored into a user-visible long field
ring_buffer_poll_wait() return value used as return value of ->poll()
the rest of drivers/*: annotate ->poll() instances
media: annotate ->poll() instances
fs: annotate ->poll() instances
ipc, kernel, mm: annotate ->poll() instances
net: annotate ->poll() instances
apparmor: annotate ->poll() instances
tomoyo: annotate ->poll() instances
sound: annotate ->poll() instances
acpi: annotate ->poll() instances
crypto: annotate ->poll() instances
block: annotate ->poll() instances
x86: annotate ->poll() instances
...
This variable was increased and decreased without any protection.
Result was an occasional misscount and negative wrap around resulting
in false resource allocation failures.
Fixes: 7d2c3f54e6 ("crypto: af_alg - remove locking in async callback")
Signed-off-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Reviewed-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The code paths protected by the socket-lock do not use or modify the
socket in a non-atomic fashion. The actions pertaining the socket do not
even need to be handled as an atomic operation. Thus, the socket-lock
can be safely ignored.
This fixes a bug regarding scheduling in atomic as the callback function
may be invoked in interrupt context.
In addition, the sock_hold is moved before the AIO encrypt/decrypt
operation to ensure that the socket is always present. This avoids a
tiny race window where the socket is unprotected and yet used by the AIO
operation.
Finally, the release of resources for a crypto operation is moved into a
common function of af_alg_free_resources.
Cc: <stable@vger.kernel.org>
Fixes: e870456d8e ("crypto: algif_skcipher - overhaul memory management")
Fixes: d887c52d6a ("crypto: algif_aead - overhaul memory management")
Reported-by: Romain Izard <romain.izard.pro@gmail.com>
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Tested-by: Romain Izard <romain.izard.pro@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
algif starts several async crypto ops and waits for their completion.
Move it over to generic code doing the same.
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Consolidate following data structures:
skcipher_async_req, aead_async_req -> af_alg_async_req
skcipher_rsgl, aead_rsql -> af_alg_rsgl
skcipher_tsgl, aead_tsql -> af_alg_tsgl
skcipher_ctx, aead_ctx -> af_alg_ctx
Consolidate following functions:
skcipher_sndbuf, aead_sndbuf -> af_alg_sndbuf
skcipher_writable, aead_writable -> af_alg_writable
skcipher_rcvbuf, aead_rcvbuf -> af_alg_rcvbuf
skcipher_readable, aead_readable -> af_alg_readable
aead_alloc_tsgl, skcipher_alloc_tsgl -> af_alg_alloc_tsgl
aead_count_tsgl, skcipher_count_tsgl -> af_alg_count_tsgl
aead_pull_tsgl, skcipher_pull_tsgl -> af_alg_pull_tsgl
aead_free_areq_sgls, skcipher_free_areq_sgls -> af_alg_free_areq_sgls
aead_wait_for_wmem, skcipher_wait_for_wmem -> af_alg_wait_for_wmem
aead_wmem_wakeup, skcipher_wmem_wakeup -> af_alg_wmem_wakeup
aead_wait_for_data, skcipher_wait_for_data -> af_alg_wait_for_data
aead_data_wakeup, skcipher_data_wakeup -> af_alg_data_wakeup
aead_sendmsg, skcipher_sendmsg -> af_alg_sendmsg
aead_sendpage, skcipher_sendpage -> af_alg_sendpage
aead_async_cb, skcipher_async_cb -> af_alg_async_cb
aead_poll, skcipher_poll -> af_alg_poll
Split out the following common code from recvmsg:
af_alg_alloc_areq: allocation of the request data structure for the
cipher operation
af_alg_get_rsgl: creation of the RX SGL anchored in the request data
structure
The following changes to the implementation without affecting the
functionality have been applied to synchronize slightly different code
bases in algif_skcipher and algif_aead:
The wakeup in af_alg_wait_for_data is triggered when either more data
is received or the indicator that more data is to be expected is
released. The first is triggered by user space, the second is
triggered by the kernel upon finishing the processing of data
(i.e. the kernel is ready for more).
af_alg_sendmsg uses size_t in min_t calculation for obtaining len.
Return code determination is consistent with algif_skcipher. The
scope of the variable i is reduced to match algif_aead. The type of the
variable i is switched from int to unsigned int to match algif_aead.
af_alg_sendpage does not contain the superfluous err = 0 from
aead_sendpage.
af_alg_async_cb requires to store the number of output bytes in
areq->outlen before the AIO callback is triggered.
The POLLIN / POLLRDNORM is now set when either not more data is given or
the kernel is supplied with data. This is consistent to the wakeup from
sleep when the kernel waits for data.
The request data structure is extended by the field last_rsgl which
points to the last RX SGL list entry. This shall help recvmsg
implementation to chain the RX SGL to other SG(L)s if needed. It is
currently used by algif_aead which chains the tag SGL to the RX SGL
during decryption.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Lockdep issues a circular dependency warning when AFS issues an operation
through AF_RXRPC from a context in which the VFS/VM holds the mmap_sem.
The theory lockdep comes up with is as follows:
(1) If the pagefault handler decides it needs to read pages from AFS, it
calls AFS with mmap_sem held and AFS begins an AF_RXRPC call, but
creating a call requires the socket lock:
mmap_sem must be taken before sk_lock-AF_RXRPC
(2) afs_open_socket() opens an AF_RXRPC socket and binds it. rxrpc_bind()
binds the underlying UDP socket whilst holding its socket lock.
inet_bind() takes its own socket lock:
sk_lock-AF_RXRPC must be taken before sk_lock-AF_INET
(3) Reading from a TCP socket into a userspace buffer might cause a fault
and thus cause the kernel to take the mmap_sem, but the TCP socket is
locked whilst doing this:
sk_lock-AF_INET must be taken before mmap_sem
However, lockdep's theory is wrong in this instance because it deals only
with lock classes and not individual locks. The AF_INET lock in (2) isn't
really equivalent to the AF_INET lock in (3) as the former deals with a
socket entirely internal to the kernel that never sees userspace. This is
a limitation in the design of lockdep.
Fix the general case by:
(1) Double up all the locking keys used in sockets so that one set are
used if the socket is created by userspace and the other set is used
if the socket is created by the kernel.
(2) Store the kern parameter passed to sk_alloc() in a variable in the
sock struct (sk_kern_sock). This informs sock_lock_init(),
sock_init_data() and sk_clone_lock() as to the lock keys to be used.
Note that the child created by sk_clone_lock() inherits the parent's
kern setting.
(3) Add a 'kern' parameter to ->accept() that is analogous to the one
passed in to ->create() that distinguishes whether kernel_accept() or
sys_accept4() was the caller and can be passed to sk_alloc().
Note that a lot of accept functions merely dequeue an already
allocated socket. I haven't touched these as the new socket already
exists before we get the parameter.
Note also that there are a couple of places where I've made the accepted
socket unconditionally kernel-based:
irda_accept()
rds_rcp_accept_one()
tcp_accept_from_sock()
because they follow a sock_create_kern() and accept off of that.
Whilst creating this, I noticed that lustre and ocfs don't create sockets
through sock_create_kern() and thus they aren't marked as for-kernel,
though they appear to be internal. I wonder if these should do that so
that they use the new set of lock keys.
Signed-off-by: David Howells <dhowells@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch allows af_alg_release_parent to be called even for
nokey sockets.
Cc: stable@vger.kernel.org
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
This patch adds a compatibility path to support old applications
that do acept(2) before setkey.
Cc: stable@vger.kernel.org
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Each af_alg parent socket obtained by socket(2) corresponds to a
tfm object once bind(2) has succeeded. An accept(2) call on that
parent socket creates a context which then uses the tfm object.
Therefore as long as any child sockets created by accept(2) exist
the parent socket must not be modified or freed.
This patch guarantees this by using locks and a reference count
on the parent socket. Any attempt to modify the parent socket will
fail with EBUSY.
Cc: stable@vger.kernel.org
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Pull crypto update from Herbert Xu:
"Here is the crypto update for 3.20:
- Added 192/256-bit key support to aesni GCM.
- Added MIPS OCTEON MD5 support.
- Fixed hwrng starvation and race conditions.
- Added note that memzero_explicit is not a subsitute for memset.
- Added user-space interface for crypto_rng.
- Misc fixes"
* git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6: (71 commits)
crypto: tcrypt - do not allocate iv on stack for aead speed tests
crypto: testmgr - limit IV copy length in aead tests
crypto: tcrypt - fix buflen reminder calculation
crypto: testmgr - mark rfc4106(gcm(aes)) as fips_allowed
crypto: caam - fix resource clean-up on error path for caam_jr_init
crypto: caam - pair irq map and dispose in the same function
crypto: ccp - terminate ccp_support array with empty element
crypto: caam - remove unused local variable
crypto: caam - remove dead code
crypto: caam - don't emit ICV check failures to dmesg
hwrng: virtio - drop extra empty line
crypto: replace scatterwalk_sg_next with sg_next
crypto: atmel - Free memory in error path
crypto: doc - remove colons in comments
crypto: seqiv - Ensure that IV size is at least 8 bytes
crypto: cts - Weed out non-CBC algorithms
MAINTAINERS: add linux-crypto to hw random
crypto: cts - Remove bogus use of seqiv
crypto: qat - don't need qat_auth_state struct
crypto: algif_rng - fix sparse non static symbol warning
...
With that, all ->sendmsg() instances are converted to iov_iter primitives
and are agnostic wrt the kind of iov_iter they are working with.
So's the last remaining ->recvmsg() instance that wasn't kind-agnostic yet.
All ->sendmsg() and ->recvmsg() advance ->msg_iter by the amount actually
copied and none of them modifies the underlying iovec, etc.
Cc: linux-crypto@vger.kernel.org
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Use setsockopt on the tfm FD to provide the authentication tag size for
an AEAD cipher. This is achieved by adding a callback function which is
intended to be used by the AEAD AF_ALG implementation.
The optlen argument of the setsockopt specifies the authentication tag
size to be used with the AEAD tfm.
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
AEAD requires the caller to specify the following information separate
from the data stream. This information allows the AEAD interface handler
to identify the AAD, ciphertext/plaintext and the authentication tag:
* Associated authentication data of arbitrary length and
length
* Length of authentication tag for encryption
Signed-off-by: Stephan Mueller <smueller@chronox.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Remove linux/mm.h inclusion from netdevice.h -- it's unused (I've checked manually).
To prevent mm.h inclusion via other channels also extract "enum dma_data_direction"
definition into separate header. This tiny piece is what gluing netdevice.h with mm.h
via "netdevice.h => dmaengine.h => dma-mapping.h => scatterlist.h => mm.h".
Removal of mm.h from scatterlist.h was tried and was found not feasible
on most archs, so the link was cutoff earlier.
Hope people are OK with tiny include file.
Note, that mm_types.h is still dragged in, but it is a separate story.
Signed-off-by: Alexey Dobriyan <adobriyan@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch creates the backbone of the user-space interface for
the Crypto API, through a new socket family AF_ALG.
Each session corresponds to one or more connections obtained from
that socket. The number depends on the number of inputs/outputs
of that particular type of operation. For most types there will
be a s ingle connection/file descriptor that is used for both input
and output. AEAD is one of the few that require two inputs.
Each algorithm type will provide its own implementation that plugs
into af_alg. They're keyed using a string such as "skcipher" or
"hash".
IOW this patch only contains the boring bits that is required
to hold everything together.
Thakns to Miloslav Trmac for reviewing this and contributing
fixes and improvements.
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Acked-by: David S. Miller <davem@davemloft.net>
Tested-by: Martin Willi <martin@strongswan.org>
