Commit 22191ac353 ("drivers/crypto/fsl: assign job-rings to
non-TrustZone") breaks HABv4 encrypted boot support in the
following i.MX devices:
- i.MX6UL
- i.MX7S
- i.MX7D
- i.MX7ULP
For preparing a HABv4 encrypted boot image it's necessary to
encapsulate the generated DEK in a blob. In devices listed
above the blob generation function takes into consideration
the Job Ring TrustZone ownership configuration (JROWN_NS)
and can be only decapsulated by the same configuration.
The ROM code expects DEK blobs encapsulated by the Secure World
environments which commonly have JROWN_NS = 0.
As U-Boot is running in Secure World we must have JROWN_NS = 0
so the blobs generated by dek_blob tool can be decapsulated
by the ROM code.
Job-rings assignment is now handled in OP-TEE OS, this commit can
be safely reverted.
https://github.com/OP-TEE/optee_os/pull/2986
This reverts commit 22191ac353.
Signed-off-by: Breno Lima <breno.lima@nxp.com>
Reviewed-by: Fabio Estevam <festevam@gmail.com>
Acked-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
Extend the instantiate_rng() function and the corresponding CAAM job
descriptor to instantiate all RNG state handles. This moves the RNG
instantiation code in line with the CAAM kernel driver.
Previously, only the first state handle was instantiated. The second
one was instantiated by the CAAM kernel driver. This works if the
kernel runs in secure mode, but fails in non-secure mode since the
kernel driver uses DEC0 directly instead of over the job ring
interface. Instantiating all RNG state handles in u-boot removes the
need for using DEC0 in the kernel driver, making it possible to use
the CAAM in non-secure mode.
Signed-off-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de>
Tested-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Reviewed-by: York Sun <york.sun@nxp.com>
After enabling TrustZone various parts of the CAAM silicon become
inaccessible to non TrustZone contexts. The job-ring registers are designed
to allow non TrustZone contexts like Linux to still submit jobs to CAAM
even after TrustZone has been enabled.
The default job-ring permissions after the BootROM look like this for
job-ring zero.
ms=0x00008001 ls=0x00008001
The MS field is JRaMIDR_MS (job ring MID most significant).
Referring to "Security Reference Manual for i.MX 7Dual and 7Solo
Applications Processors, Rev. 0, 03/2017" section 8.10.4 we see that
JROWN_NS controls whether or not a job-ring is accessible from non
TrustZone.
Bit 15 (TrustZone) is the logical inverse of bit 3 hence the above value of
0x8001 shows that JROWN_NS=0 and TrustZone=1.
Clearly then as soon as TrustZone becomes active the job-ring registers are
no longer accessible from Linux, which is not what we want.
This patch explicitly sets all job-ring registers to JROWN_NS=1 (non
TrustZone) by default and to the Non-Secure MID 001. Both settings are
required to successfully assign a job-ring to non-secure mode. If a piece
of TrustZone firmware requires ownership of job-ring registers it can unset
the JROWN_NS bit itself.
This patch in conjunction with a modification of the Linux kernel to skip
HWRNG initialisation makes CAAM usable to Linux with TrustZone enabled.
Signed-off-by: Bryan O'Donoghue <bryan.odonoghue@linaro.org>
Cc: Fabio Estevam <fabio.estevam@nxp.com>
Cc: Peng Fan <peng.fan@nxp.com>
Cc: Alex Porosanu <alexandru.porosanu@nxp.com>
Cc: Ruchika Gupta <ruchika.gupta@nxp.com>
Cc: Aneesh Bansal <aneesh.bansal@nxp.com>
Link: https://github.com/OP-TEE/optee_os/issues/1408
Link: https://tinyurl.com/yam5gv9a
Tested-by: Lukas Auer <lukas.auer@aisec.fraunhofer.de>
For SoCs that contain multiple SEC engines, each of them needs
to be initialized (by means of initializing among others the
random number generator).
Signed-off-by: Alex Porosanu <alexandru.porosanu@nxp.com>
Reviewed-by: York Sun <york.sun@nxp.com>
This commit solves CAAM coherency issue on ls2080. When caches are
enabled and CAAM's DMA's AXI transcations are not made cacheable,
Core reads/writes data from/to caches and CAAM does from main memory.
This forces data flushes to synchronize various data structures. But
even if any data in proximity of these structures is read by core,
these structures again are fetched in caches.
To avoid this problem, either all the data that CAAM accesses can be
made cache line aligned or CAAM transcations can be made cacheable.
So, this commit makes CAAM transcations as write back with write and
read allocate.
Signed-off-by: Saksham Jain <saksham.jain@nxp.com>
Reviewed-by: York Sun <york.sun@nxp.com>
The output ring needs to be invalidated before enqueuing the job to SEC.
While allocation of space to output ring, it should be taken care that the
size is cacheline size aligned inorder to prevent invalidating valid data.
The patch also correct the method of aligning end of structs while flushing caches
Since start = align(start_of_struct), it is incorrect to assign
end = align(start + struct_size). It should instead be,
end = align(start_of_struct + struct_size).
Signed-off-by: Saksham Jain <saksham@nxp.com>
Signed-off-by: Ruchika Gupta <ruchika.gupta@nxp.com>
Reviewed-by: York Sun <york.sun@nxp.com>
The SEC driver code has been cleaned up to work for 64 bit
physical addresses and systems where endianess of SEC block
is different from the Core.
Changes:
1. Descriptor created on Core is modified as per SEC block
endianness before the job is submitted.
2. The read/write of physical addresses to Job Rings will
be depend on endianness of SEC block as 32 bit low and
high part of the 64 bit address will vary.
3. The 32 bit low and high part of the 64 bit address in
descriptor will vary depending on endianness of SEC.
Signed-off-by: Aneesh Bansal <aneesh.bansal@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
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.
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
Fixes: b9eebfade9 ("fsl_sec: Add hardware accelerated SHA256 and SHA1")
Signed-off-by: Horia Geantă <horia.geanta@freescale.com>
Reviewed-by: Aneesh Bansal <aneesh.bansal@freescale.com>
Reviewed-by: Mingkai Hu <Mingkai.Hu@freescale.com>
Acked-by: Ruchika Gupta<ruchika.gupta@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
SHA-256 and SHA-1 accelerated using SEC hardware in Freescale SoC's
The driver for SEC (CAAM) IP is based on linux drivers/crypto/caam.
The platforms needto add the MACRO CONFIG_FSL_CAAM inorder to
enable initialization of this hardware IP.
Signed-off-by: Ruchika Gupta <ruchika.gupta@freescale.com>
Reviewed-by: York Sun <yorksun@freescale.com>
