summaryrefslogtreecommitdiff
path: root/drivers
AgeCommit message (Collapse)AuthorFilesLines
2015-08-04Merge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6Herbert Xu3-10/+17
Merge the crypto tree to pull in the qat registration bug fix.
2015-08-04crypto: talitos - Remove zero_entry static initializerAaron Sierra1-6/+1
Compiling the talitos driver with my GCC 4.3.1 e500v2 cross-compiler resulted in a failed build due to the anonymous union/structures introduced in this commit: crypto: talitos - enhanced talitos_desc struct for SEC1 The build error was: drivers/crypto/talitos.h:56: error: unknown field 'len' specified in initializer drivers/crypto/talitos.h:56: warning: missing braces around initializer drivers/crypto/talitos.h:56: warning: (near initialization for 'zero_entry.<anonymous>') drivers/crypto/talitos.h:57: error: unknown field 'j_extent' specified in initializer drivers/crypto/talitos.h:58: error: unknown field 'eptr' specified in initializer drivers/crypto/talitos.h:58: warning: excess elements in struct initializer drivers/crypto/talitos.h:58: warning: (near initialization for 'zero_entry') make[2]: *** [drivers/crypto/talitos.o] Error 1 make[1]: *** [drivers/crypto] Error 2 make: *** [drivers] Error 2 This patch eliminates the errors by relying on the C standard's implicit assignment of zero to static variables. Signed-off-by: Aaron Sierra <asierra@xes-inc.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-08-04crypto: img-hash - fix spelling mistake in dev_err error messageColin Ian King1-1/+1
Trival change, fix spelling mistake 'aquire' -> 'acquire' in dev_err message. Signed-off-by: Colin Ian King <colin.king@canonical.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-31crypto: caam - fix rfc4106 encap shared descriptorHoria Geant?1-9/+15
The encap shared descriptor was changed to use the new IV convention. In the process some commands were shifted, making the output length zero, caam effectively writing garbage in dst. While here, update the decap descriptor to execute the "write" commands before the "read"s (as it previously was). This makes sure the input fifo is drained before becoming full. Fixes: 46218750d523 ("crypto: caam - Use new IV convention") Signed-off-by: Horia Geant? <horia.geanta@freescale.com> Signed-off-by: Tudor Ambarus <tudor.ambarus@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-30crypto: nx - use be32_to_cpu for __be32 field in debug msgDan Streetman1-1/+2
One of the debug messages in the NX 842 PowerNV driver is missing the required be32_to_cpu() wrapper when accessing the __be32 field csb->count. Add the wrapper so the message will show the correct count. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-30crypto: nx - don't err if compressed output > inputDan Streetman1-2/+5
Return success instead of error if compression succeeds but the output is larger than the input. It's unlikely that the caller will use the compressed data since it's larger than the original uncompressed data, but there was no error and returning an error code is incorrect. Further, for testing small input buffers, the output is likely to be larger than the input and success needs to be returned to verify the test. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-28crypto: qat - remove unnecessary list iterationBruce Allan1-6/+1
There's no need to iterate through the list for instances in the accel_table since the number of devices is already known in this file. Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-28crypto: qat - Fix typo othewise->otherwiseAhsan Atta5-11/+11
Signed-off-by: Ahsan Atta <ahsan.atta@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-28crypto: qat - remove unused defineBruce Allan1-1/+0
Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-28crypto: qat - fix bug in ADF_RING_SIZE_BYTES_MIN macroBruce Allan1-2/+3
The subject macro mistakenly compares the passed-in ring size in bytes with ADF_RING_SIZE_4K which is 0x6 (an internal representation of 4KB) rather than comparing with the intended value of 4096. Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-28crypto: qat - remove redundant struct elemBruce Allan3-3/+1
The element pci_dev_id in the struct adf_hw_device_data is redundant since the PCI device id can be retrieved from the struct pci_dev. Signed-off-by: Bruce Allan <bruce.w.allan@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-28hwrng: core - correct error check of kthread_run callMartin Schwidefsky1-1/+1
The kthread_run() function can return two different error values but the hwrng core only checks for -ENOMEM. If the other error value -EINTR is returned it is assigned to hwrng_fill and later used on a kthread_stop() call which naturally crashes. Cc: stable@vger.kernel.org Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - make platform drivers directly register with cryptoDan Streetman6-191/+114
Remove the common 'platform' registration module, and move the crypto compression driver registration into each of the pSeries and PowerNV platform NX 842 drivers. Change the nx-842.c code into simple common functions that each platform driver uses to perform constraints-based buffer changes, i.e. realigning and/or resizing buffers to match the driver's hardware requirements. The common 'platform' module was my mistake to create - since each platform driver will only load/operate when running on its own platform (i.e. a pSeries platform or a PowerNV platform), they can directly register with the crypto subsystem, using the same alg and driver name. This removes unneeded complexity. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - rename nx-842-crypto.c to nx-842.cDan Streetman2-1/+1
The last commit merged nx-842.c's code into nx-842-crypto.c. It did not rename nx-842-crypto.c to nx-842.c, in order to let the patch more clearly show what was merged. This just renames nx-842-crypto.c to nx-842.c, with no changes to its code. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - merge nx-compress and nx-compress-cryptoDan Streetman5-162/+60
Merge the nx-842.c code into nx-842-crypto.c. This allows later patches to remove the 'platform' driver, and instead allow each platform driver to directly register with the crypto compression api. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - use common code for both NX decompress success casesDan Streetman1-4/+2
Replace the duplicated finishing code (set destination buffer length and set return code to 0) in the case of decompressing a buffer with no header with a goto to the success case of decompressing a buffer with a header. This is a trivial change that allows both success cases to use common code, and includes the pr_debug() msg in both cases as well. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - don't register pSeries driver if ENODEVDan Streetman1-4/+1
Don't register the pSeries driver when parsing the device tree returns ENODEV. The nx842_probe() function in the pSeries driver returns error instead of registering as a crypto compression driver, when it receives an error return value from the nx842_OF_upd() function that probes the device tree nodes, except when ENODEV is returned. However ENODEV should not be a special case and the driver should not register when there is no hw device, or the hw device is disabled. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - move kzalloc() out of spinlockDan Streetman1-22/+16
Move the kzalloc() calls in nx842_probe() and nx842_OF_upd() to the top of the functions, before taking the devdata spinlock. Since kzalloc() without GFP_ATOMIC can sleep, it can't be called while holding a spinlock. Move the calls to before taking the lock. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - remove pSeries NX 'status' fieldDan Streetman1-29/+10
Remove the 'status' field from the pSeries NX driver data. The 'status' field isn't used by the driver at all; it simply checks the devicetree status node at initialization, and returns success if 'okay' and failure otherwise. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: nx - remove __init/__exit from VIO functionsDan Streetman1-4/+4
Remove the __init and __exit modifiers from the VIO driver probe and remove functions. The driver functions should not be marked __init/__exit because they can/will be called during runtime, not only at module init and exit. Signed-off-by: Dan Streetman <ddstreet@ieee.org> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: qat - Don't attempt to register algorithm multiple timesTadeusz Struk1-3/+16
When multiple devices are present in the system the driver attempts to register the same algorithm many times. Changes in v2: - use proper synchronization mechanizm between register and unregister Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: ixp4xx - Remove bogus BUG_ON on scattered dst bufferHerbert Xu1-1/+0
This patch removes a bogus BUG_ON in the ablkcipher path that triggers when the destination buffer is different from the source buffer and is scattered. Cc: stable@vger.kernel.org Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-23crypto: qat - Fix invalid synchronization between register/unregister sym algsTadeusz Struk1-8/+16
The synchronization method used atomic was bogus. Use a proper synchronization with mutex. Cc: stable@vger.kernel.org Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-21crypto: qat - fix invalid check for RSA keylen in fips modeTadeusz Struk1-2/+2
The condition checking allowed key length was invalid. Reported-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-20crypto: sunxi-ss - Add Allwinner Security System crypto acceleratorLABBE Corentin7-0/+1656
Add support for the Security System included in Allwinner SoC A20. The Security System is a hardware cryptographic accelerator that support: - MD5 and SHA1 hash algorithms - AES block cipher in CBC/ECB mode with 128/196/256bits keys. - DES and 3DES block cipher in CBC/ECB mode Signed-off-by: LABBE Corentin <clabbe.montjoie@gmail.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-20crypto: caam - fix warning in APPEND_MATH_IMM_u64Tudor Ambarus1-1/+1
An implicit truncation is done when using a variable of 64 bits in MATH command: warning: large integer implicitly truncated to unsigned type [-Woverflow] Silence the compiler by feeding it with an explicit truncated value. Signed-off-by: Tudor Ambarus <tudor.ambarus@freescale.com> Signed-off-by: Horia Geant? <horia.geanta@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-20crypto: caam - fix RNG init descriptor ret. code checkingHoria Geant?1-2/+3
When successful, the descriptor that performs RNG initialization is allowed to return a status code of 7000_0000h, since last command in the descriptor is a JUMP HALT. Signed-off-by: Horia Geant? <horia.geanta@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-20crypto: caam - fix snooping for write transactionsHoria Geant?2-2/+9
HW coherency won't work properly for CAAM write transactions if AWCACHE is left to default (POR) value - 4'b0001. It has to be programmed to 4'b0010, i.e. AXI3 Cacheable bit set. For platforms that have HW coherency support: -PPC-based: the update has no effect; CAAM coherency already works due to the IOMMU (PAMU) driver setting the correct memory coherency attributes -ARM-based: the update fixes cache coherency issues, since IOMMU (SMMU) driver is not programmed to behave similar to PAMU Signed-off-by: Horia Geant? <horia.geanta@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-20crypto: caam - fix ERA property readingAlex Porosanu1-7/+7
In order to ensure that the ERA property is properly read from DT on all platforms, of_property_read* function needs to be used. Signed-off-by: Alex Porosanu <alexandru.porosanu@freescale.com> Signed-off-by: Horia Geant? <horia.geanta@freescale.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-17crypto: qat - Add support for RSA algorithmTadeusz Struk13-39/+779
Add RSA support to QAT driver. Removed unused RNG rings. Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-17crypto: qat - add MMP FW support to accel engineTadeusz Struk4-5/+42
Add code that loads the MMP firmware Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-17crypto: qat - add support for MMP FWPingchao Yang6-26/+24
Load Modular Math Processor(MMP) firmware into QAT devices to support public key algorithm acceleration. Signed-off-by: Pingchao Yang <pingchao.yang@intel.com> Signed-off-by: Tadeusz Struk <tadeusz.struk@intel.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-17crypto: nx - Convert ccm to new AEAD interfaceHerbert Xu3-95/+85
This patch converts the nx ccm and 4309 implementations to the new AEAD interface. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-14crypto: vmx - Adding enable_kernel_vsx() to access VSX instructionsLeonidas Da Silva Barbosa4-0/+13
vmx-crypto driver make use of some VSX instructions which are only available if VSX is enabled. Running in cases where VSX are not enabled vmx-crypto fails in a VSX exception. In order to fix this enable_kernel_vsx() was added to turn on VSX instructions for vmx-crypto. Signed-off-by: Leonidas S. Barbosa <leosilva@linux.vnet.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-14crypto: marvell/cesa - Drop owner assignment from platform_driverKrzysztof Kozlowski1-1/+0
platform_driver does not need to set an owner because platform_driver_register() will set it. Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com> Acked-by: Boris Brezillon <boris.brezillon@free-electrons.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-14crypto: caam - Use new IV conventionHerbert Xu1-26/+49
This patch converts rfc4106 to the new calling convention where the IV is now part of the AD and needs to be skipped. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-14crypto: nx - Use new IV conventionHerbert Xu1-26/+40
This patch converts rfc4106 to the new calling convention where the IV is now part of the AD and needs to be skipped. This patch also makes use of type-safe AEAD functions where possible. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-14crypto: nx/842 - Fix context corruptionHerbert Xu1-13/+35
The transform context is shared memory and must not be written to without locking. This patch adds locking to nx-842 to prevent context corruption. Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-14Merge git://git.kernel.org/pub/scm/linux/kernel/git/herbert/crypto-2.6Herbert Xu8-79/+125
Merge the crypto tree to pull in the nx reentrancy patch.
2015-07-08crypto: omap-aes - Use BIT() macroVutla, Lokesh1-17/+17
Use BIT()/GENMASK() macros for all register definitions instead of hand-writing bit masks. Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-08crypto: omap-aes - Fix configuring of AES modeVutla, Lokesh1-8/+5
AES_CTRL_REG is used to configure AES mode. Before configuring any mode we need to make sure all other modes are reset or else driver will misbehave. So mask all modes before configuring any AES mode. Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-08crypto: omap-aes - Increase priority of hw acceleratorVutla, Lokesh1-3/+3
Increasing the priority of omap-aes hw algos, in order to take precedence over sw algos. Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-08crypto: omap-aes - Fix CTR modeVutla, Lokesh1-17/+16
Algo self tests are failing for CTR mode with omap-aes driver, giving the following error: [ 150.053644] omap_aes_crypt: request size is not exact amount of AES blocks [ 150.061262] alg: skcipher: encryption failed on test 5 for ctr-aes-omap: ret=22 This is because the input length is not aligned with AES_BLOCK_SIZE. Adding support for omap-aes driver for inputs with length not aligned with AES_BLOCK_SIZE. Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-08crypto: nx - Fix reentrancy bugsHerbert Xu8-79/+125
This patch fixes a host of reentrancy bugs in the nx driver. The following algorithms are affected: * CCM * GCM * CTR * XCBC * SHA256 * SHA512 The crypto API allows a single transform to be used by multiple threads simultaneously. For example, IPsec will use a single tfm to process packets for a given SA. As packets may arrive on multiple CPUs that tfm must be reentrant. The nx driver does try to deal with this by using a spin lock. Unfortunately only the basic AES/CBC/ECB algorithms do this in the correct way. The symptom of these bugs may range from the generation of incorrect output to memory corruption. Cc: stable@vger.kernel.org Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-07crypto: nx - reduce chattiness of platform driversNishanth Aravamudan2-12/+1
While we never would successfully load on the wrong machine type, there is extra output by default regardless of machine type. For instance, on a PowerVM LPAR, we see the following: nx_compress_powernv: loading nx_compress_powernv: no coprocessors found even though those coprocessors could never be found. Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> Cc: Dan Streetman <ddstreet@us.ibm.com> Cc: Herbert Xu <herbert@gondor.apana.org.au> Cc: "David S. Miller" <davem@davemloft.net> Cc: linux-crypto@vger.kernel.org Cc: linuxppc-dev@lists.ozlabs.org Acked-by: Michael Ellerman <mpe@ellerman.id.au> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-06crypto: nx - do not emit extra output if status is disabledNishanth Aravamudan1-1/+7
If the device-tree indicates the nx-842 device's status is 'disabled', we emit two messages: nx_compress_pseries ibm,compression-v1: nx842_OF_upd_status: status 'disabled' is not 'okay'. nx_compress_pseries ibm,compression-v1: nx842_OF_upd: device disabled Given that 'disabled' is a valid state, and we are going to emit that the device is disabled, only print out a non-'okay' status if it is not 'disabled'. Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-06crypto: nx - rename nx842_{init, exit} to nx842_pseries_{init, exit}Nishanth Aravamudan1-4/+4
While there is no technical reason that both nx-842.c and nx-842-pseries.c can have the same name for the init/exit functions, it is a bit confusing with initcall_debug. Rename the pseries specific functions appropriately Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-06crypto: nx - nx842_OF_upd_status should return ENODEV if device is not 'okay'Nishanth Aravamudan1-1/+2
The current documention mentions explicitly that EINVAL should be returned if the device is not available, but nx842_OF_upd_status() always returns 0. However, nx842_probe() specifically checks for non-ENODEV returns from nx842_of_upd() (which in turn calls nx842_OF_upd_status()) and emits an extra error in that case. It seems like the proper return code of a disabled device is ENODEV. Signed-off-by: Nishanth Aravamudan <nacc@linux.vnet.ibm.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-06crypto: ccp - Provide support to autoload CCP driverTom Lendacky1-0/+2
Add the necessary module device tables to the platform support to allow for autoloading of the CCP driver. This will allow for the CCP's hwrng support to be available without having to manually load the driver. The module device table entry for the pci support is already present. Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
2015-07-06crypto: omap-des - Fix unmapping of dma channelsVutla, Lokesh1-3/+0
dma_unmap_sg() is being called twice after completing the task. Looks like this is a copy paste error when creating des driver. With this the following warn appears during boot: [ 4.210457] ------------[ cut here ]------------ [ 4.215114] WARNING: CPU: 0 PID: 0 at lib/dma-debug.c:1080 check_unmap+0x710/0x9a0() [ 4.222899] omap-des 480a5000.des: DMA-API: device driver tries to free DMA memory it has not allocated [device address=0x00000000ab2ce000] [size=8 bytes] [ 4.236785] Modules linked in: [ 4.239860] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.14.39-02999-g1bc045a-dirty #182 [ 4.247918] [<c001678c>] (unwind_backtrace) from [<c0012574>] (show_stack+0x10/0x14) [ 4.255710] [<c0012574>] (show_stack) from [<c05a37e8>] (dump_stack+0x84/0xb8) [ 4.262977] [<c05a37e8>] (dump_stack) from [<c0046464>] (warn_slowpath_common+0x68/0x8c) [ 4.271107] [<c0046464>] (warn_slowpath_common) from [<c004651c>] (warn_slowpath_fmt+0x30/0x40) [ 4.279854] [<c004651c>] (warn_slowpath_fmt) from [<c02d50a4>] (check_unmap+0x710/0x9a0) [ 4.287991] [<c02d50a4>] (check_unmap) from [<c02d5478>] (debug_dma_unmap_sg+0x90/0x19c) [ 4.296128] [<c02d5478>] (debug_dma_unmap_sg) from [<c04a77d8>] (omap_des_done_task+0x1cc/0x3e4) [ 4.304963] [<c04a77d8>] (omap_des_done_task) from [<c004a090>] (tasklet_action+0x84/0x124) [ 4.313370] [<c004a090>] (tasklet_action) from [<c004a4ac>] (__do_softirq+0xf0/0x20c) [ 4.321235] [<c004a4ac>] (__do_softirq) from [<c004a840>] (irq_exit+0x98/0xec) [ 4.328500] [<c004a840>] (irq_exit) from [<c000f9ac>] (handle_IRQ+0x50/0xb0) [ 4.335589] [<c000f9ac>] (handle_IRQ) from [<c0008688>] (gic_handle_irq+0x28/0x5c) Removing the duplicate call to dma_unmap_sg(). Cc: stable@vger.kernel.org Reported-by: Tomi Valkeinen <tomi.valkeinen@ti.com> Signed-off-by: Lokesh Vutla <lokeshvutla@ti.com> Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>