The locks are taken in this order during driver registration
(uld_mutex), at: cxgb4_register_uld.part.14+0x49/0xd60 [cxgb4]
(rtnl_mutex), at: rtnetlink_rcv_msg+0x2db/0x400
(uld_mutex), at: cxgb_up+0x3a/0x7b0 [cxgb4]
(rtnl_mutex), at: chcr_add_xfrmops+0x83/0xa0 [chcr](stucked here)
To avoid this now the netdev features are updated after the
cxgb4_register_uld function is completed.
Fixes: 6dad4e8ab3 ("chcr: Add support for Inline IPSec").
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This issue occurs only when multiadapters are present. Hang
happens because assign_chcr_device returns u_ctx pointer of
adapter which is not yet initialized as for this adapter cxgb_up
is not been called yet.
The last_dev pointer is used to determine u_ctx pointer and it
is initialized two times in chcr_uld_add in chcr_dev_add respectively.
The fix here is don't initialize the last_dev pointer during
chcr_uld_add. Only assign to value to it when the adapter's
initialization is completed i.e in chcr_dev_add.
Fixes: fef4912b66 ("crypto: chelsio - Handle PCI shutdown event").
Signed-off-by: Ayush Sawal <ayush.sawal@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Rework CAAM RNG implementation as follows:
- Make use of the fact that HWRNG supports partial reads and will
handle such cases gracefully by removing recursion in caam_read()
- Convert blocking caam_read() codepath to do a single blocking job
read directly into requested buffer, bypassing any intermediary
buffers
- Convert async caam_read() codepath into a simple single
reader/single writer FIFO use-case, thus simplifying concurrency
handling and delegating buffer read/write position management to KFIFO
subsystem.
- Leverage the same low level RNG data extraction code for both async
and blocking caam_read() scenarios, get rid of the shared job
descriptor and make non-shared one as a simple as possible (just
HEADER + ALGORITHM OPERATION + FIFO STORE)
- Split private context from DMA related memory, so that the former
could be allocated without GFP_DMA.
NOTE: On its face value this commit decreased throughput numbers
reported by
dd if=/dev/hwrng of=/dev/null bs=1 count=100K [iflag=nonblock]
by about 15%, however commits that enable prediction resistance and
limit JR total size impact the performance so much and move the
bottleneck such as to make this regression irrelevant.
NOTE: On the bright side, this commit reduces RNG in kernel DMA buffer
memory usage from 2 x RN_BUF_SIZE (~256K) to 32K.
Signed-off-by: Andrey Smirnov <andrew.smirnov@gmail.com>
Reviewed-by: Horia Geantă <horia.geanta@nxp.com>
Cc: Chris Healy <cphealy@gmail.com>
Cc: Lucas Stach <l.stach@pengutronix.de>
Cc: Horia Geantă <horia.geanta@nxp.com>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Iuliana Prodan <iuliana.prodan@nxp.com>
Cc: linux-crypto@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: linux-imx@nxp.com
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Kernel TLS forms TLS header in kernel during encryption and removes
while decryption before giving packet back to user application. The
similar logic is introduced in chtls code as well.
v1->v2:
- tls_proccess_cmsg() uses tls_handle_open_record() which is not required
in TOE-TLS. Don't mix TOE with other TLS types.
Signed-off-by: Vinay Kumar Yadav <vinay.yadav@chelsio.com>
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Acked-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
drivers/crypto/chelsio/chcr_algo.c: In function 'chcr_device_init':
drivers/crypto/chelsio/chcr_algo.c:1440:18: warning:
variable 'adap' set but not used [-Wunused-but-set-variable]
commit 567be3a5d2 ("crypto: chelsio - Use multiple txq/rxq per tfm
to process the requests") involved this unused variable.
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Since snprintf() returns the would-be-output size instead of the
actual output size, the succeeding calls may go beyond the given
buffer limit. Fix it by replacing with scnprintf().
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The wake token cannot be sent without ignoring the nack for the
device address
Signed-off-by: Jianhui Zhao <zhaojh329@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Combine found device and created qp into one operation instead of found
device and create qp both are independent operations. when execute
multiple tasks, the different threads may find same device at the same
time, but the number of queues is insufficient on the device. causing
one of threads fail to create a qp. Now fix this, First find device then
create qp, if result failure. the current thread will find next device.
Signed-off-by: Kai Ye <yekai13@huawei.com>
Signed-off-by: Shukun Tan <tanshukun1@huawei.com>
Reviewed-by: Zhou Wang <wangzhou1@hisilicon.com>
Reviewed-by: Zaibo Xu <xuzaibo@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Encapsulate hisi_qm_alloc_qps_node() to new interface to replace
find_zip_device(), which will fix the bug of creating QP failure
especially in multi-thread scenario.
Signed-off-by: Shukun Tan <tanshukun1@huawei.com>
Reviewed-by: Zhou Wang <wangzhou1@hisilicon.com>
Reviewed-by: Zaibo Xu <xuzaibo@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
drivers/crypto/chelsio/chcr_ktls.c: In function chcr_ktls_cpl_set_tcb_rpl:
drivers/crypto/chelsio/chcr_ktls.c:662:11: warning:
variable status set but not used [-Wunused-but-set-variable]
commit 8a30923e15 ("cxgb4/chcr: Save tx keys and handle HW response")
involved this unused variable, remove it.
Reported-by: Hulk Robot <hulkci@huawei.com>
Signed-off-by: YueHaibing <yuehaibing@huawei.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Our handling of ciphers with IV trailing the AAD was correct
but overly complicated. Refactor to simplify and possibly
save one DMA burst.
This has the added bonus of behaving the same as the generic
rfc4543 implementation for none compliants inputs where the
IV in the iv field was not the same as the IV in the AAD.
There should be no change in behaviour with correct inputs.
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Use crypto_ipsec_check_assoclen() instead of home grown functions.
Clean up some unneeded code as a result. Delete stale comments
while we're at it.
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Move testing of condition to after the point we decide if
we need it or not.
Signed-off-by: Gilad Ben-Yossef <gilad@benyossef.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Instead of using CAP_SYS_ADMIN which is restricted to the root user,
check the file mode for write permissions before executing commands that
can affect the platform. This allows for more fine-grained access
control to the SEV ioctl interface. This would allow a SEV-only user
or group the ability to administer the platform without requiring them
to be root or granting them overly powerful permissions.
For example:
chown root:root /dev/sev
chmod 600 /dev/sev
setfacl -m g:sev:r /dev/sev
setfacl -m g:sev-admin:rw /dev/sev
In this instance, members of the "sev-admin" group have the ability to
perform all ioctl calls (including the ones that modify platform state).
Members of the "sev" group only have access to the ioctls that do not
modify the platform state.
This also makes opening "/dev/sev" more consistent with how file
descriptors are usually handled. By only checking for CAP_SYS_ADMIN,
the file descriptor could be opened read-only but could still execute
ioctls that modify the platform state. This patch enforces that the file
descriptor is opened with write privileges if it is going to be used to
modify the platform state.
This flexibility is completely opt-in, and if it is not desirable by
the administrator then they do not need to give anyone else access to
/dev/sev.
Signed-off-by: Connor Kuehl <ckuehl@redhat.com>
Reviewed-by: Brijesh Singh <brijesh.singh@amd.com>
Reviewed-by: Bandan Das <bsd@redhat.com>
Acked-by: David Rientjes <rientjes@google.com>
Acked-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
In the scenario of SMMU translation, the SEC performance of short messages
(<512Bytes) cannot meet our expectations. To avoid this, we reserve the
plat buffer (PBUF) memory for small packets when creating TFM.
Signed-off-by: Longfang Liu <liulongfang@huawei.com>
Signed-off-by: Zaibo Xu <xuzaibo@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
In order to improve performance of small packets (<512Bytes)
in SMMU translation scenario, we need to identify the type of IOMMU
in the SEC probe to process small packets by a different method.
Signed-off-by: Longfang Liu <liulongfang@huawei.com>
Signed-off-by: Zaibo Xu <xuzaibo@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Allocate one workqueue for each QM instead of one for all QMs,
we found the throughput of SEC engine can be increased to
the hardware limit throughput during testing sec2 performance.
so we added this scheme.
Signed-off-by: Ye Kai <yekai13@huawei.com>
Signed-off-by: Longfang Liu <liulongfang@huawei.com>
Signed-off-by: Zaibo Xu <xuzaibo@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Since SEC need not so many workqueues as our test, we just use
one workqueue created by the device driver of QM if necessary,
which will also reduce CPU waste without any throughput decreasing.
Signed-off-by: Shukun Tan <tanshukun1@huawei.com>
Signed-off-by: Zaibo Xu <xuzaibo@huawei.com>
Reviewed-by: Jonathan Cameron <Jonathan.Cameron@huawei.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
The CONFIG_CRYPTO_DEV_QCE_SOFT_THRESHOLD symbol was renamed during
development, but the stringify reference in the parameter description
sneaked by unnoticed.
Signed-off-by: Eneas U de Queiroz <cotequeiroz@gmail.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Introduce clear_psp_master_device() to ensure that sp_dev_master gets
properly cleared on the release of a psp device.
Fixes: 2a6170dfe7 ("crypto: ccp: Add Platform Security Processor (PSP) device support")
Signed-off-by: John Allen <john.allen@amd.com>
Acked-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Explicitly free and clear misc_dev in sev_exit(). Since devm_kzalloc()
associates misc_dev with the first device that gets probed, change from
devm_kzalloc() to kzalloc() and explicitly free memory in sev_exit() as
the first device probed is not guaranteed to be the last device released.
To ensure that the variable gets properly set to NULL, remove the local
definition of misc_dev.
Fixes: 200664d523 ("crypto: ccp: Add Secure Encrypted Virtualization (SEV) command support")
Signed-off-by: John Allen <john.allen@amd.com>
Acked-by: Tom Lendacky <thomas.lendacky@amd.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Adding ipv6 support and ktls related statistics.
v1->v2:
- added blank lines at 2 places.
v3->v4:
- Replaced atomic_t with atomic64_t
- added few necessary stat counters.
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch contains handling of first part or middle part of the record.
When we get a middle record, we will fetch few already sent bytes to
make packet start 16 byte aligned.
And if the packet has only the header part, we don't need to send it for
packet encryption, send that packet as a plaintext.
v1->v2:
- un-necessary updating left variable.
v3->v4:
- replaced kfree_skb with dev_kfree_skb_any.
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
TCP segment can chop a record in any order. Record can either be
complete or it can be partial (first part which contains header,
middle part which doesn't have header or TAG, and the end part
which contains TAG. This patch handles partial end part of a tx
record. In case of partial end part's, driver will send complete
record to HW, so that HW will calculate GHASH (TAG) of complete
packet.
Also added support to handle multiple records in a segment.
v1->v2:
- miner change in calling chcr_write_cpl_set_tcb_ulp.
- no need of checking return value of chcr_ktls_write_tcp_options.
v3->v4:
- replaced kfree_skb with dev_kfree_skb_any.
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Added tx handling in this patch. This includes handling of segments
contain single complete record.
v1->v2:
- chcr_write_cpl_set_tcb_ulp is added in this patch.
v3->v4:
- mss calculation logic.
- replaced kfree_skb with dev_kfree_skb_any.
- corrected error message reported by kbuild test robot <lkp@intel.com>
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
As part of this patch generated and saved crypto keys, handled HW
response of act_open_req and set_tcb_req. Defined connection state
update.
v1->v2:
- optimized tcb update using control queue.
- state machine handling when earlier states received.
v2->v3:
- Added one empty line after function declaration.
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
A new macro is defined to enable ktls tx offload support on Chelsio
T6 adapter. And if this macro is enabled, cxgb4 will send mailbox to
enable or disable ktls settings on HW.
In chcr, enabled tx offload flag in netdev and registered tls_dev_add
and tls_dev_del.
v1->v2:
- mark tcb state to close in tls_dev_del.
- u_ctx is now picked from adapter structure.
- clear atid in case of failure.
- corrected ULP_CRYPTO_KTLS_INLINE value.
v2->v3:
- add empty line after variable declaration.
- local variable declaration in reverse christmas tree ordering.
Signed-off-by: Rohit Maheshwari <rohitm@chelsio.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since in the software implementation of XTS-AES there is
no notion of sector every input length is processed the same way.
CAAM implementation has the notion of sector which causes different
results between the software implementation and the one in CAAM
for input lengths bigger than 512 bytes.
Increase sector size to maximum value on 16 bits.
Fixes: c6415a6016 ("crypto: caam - add support for acipher xts(aes)")
Cc: <stable@vger.kernel.org> # v4.12+
Signed-off-by: Andrei Botila <andrei.botila@nxp.com>
Reviewed-by: Horia Geantă <horia.geanta@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
HW generates a Data Size error for chacha20 requests that are not
a multiple of 64B, since algorithm state (AS) does not have
the FINAL bit set.
Since updating req->iv (for chaining) is not required,
modify skcipher descriptors to set the FINAL bit for chacha20.
[Note that for skcipher decryption we know that ctx1_iv_off is 0,
which allows for an optimization by not checking algorithm type,
since append_dec_op1() sets FINAL bit for all algorithms except AES.]
Also drop the descriptor operations that save the IV.
However, in order to keep code logic simple, things like
S/G tables generation etc. are not touched.
Cc: <stable@vger.kernel.org> # v5.3+
Fixes: 334d37c9e2 ("crypto: caam - update IV using HW support")
Signed-off-by: Horia Geantă <horia.geanta@nxp.com>
Tested-by: Valentin Ciocoi Radulescu <valentin.ciocoi@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
