Age | Commit message (Collapse) | Author |
|
|
|
|
|
Signed-off-by: Olivier Masse <olivier.masse@nxp.com>
|
|
|
|
Signed-off-by: Olivier Masse <olivier.masse@nxp.com>
|
|
Signed-off-by: Olivier Masse <olivier.masse@nxp.com>
|
|
Signed-off-by: Olivier Masse <olivier.masse@nxp.com>
|
|
Signed-off-by: Olivier Masse <olivier.masse@nxp.com>
|
|
When Secure Hantro VPU is enabled, allow decoding into non protected memory
Signed-off-by: Olivier Masse <olivier.masse@nxp.com>
|
|
Signed-off-by: Olivier Masse <olivier.masse@nxp.com>
|
|
Normally, the wdog1 is used by A53 side, and it should
be stopped when A53 domain enter STOP mode. when system
out of PoR, this watchdog is owned by both M7 & A53 side,
then this watchdog can only enter STOP mode only when
both A53 & M7 enter STOP mode. it is not reasonable as
this watchdog is only used by A53 side, so assign wdog1
to domain0(a53 side) only.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Anson Huang <anson.huang@nxp.com>
(cherry picked from commit 25ead41a194b87ff9d30e7c4a5c05875a33c432d)
|
|
|
|
NXP secure boot flow uses ROM & SPL to load BL3x instead of
TF(Trust Frimware's BL1/BL2) therefore boot chain relies on
BL31's initialization procedure enabled by RESET_TO_BL31 flag.
It might seem not aligned with its documentation about reset
framework however it can be undertstood as BL31's init procedure
still need to enabled even if BL31 doesn't act as reset handler.
In above case, since RESET_TO_BL31 doesn't mean BL31 is reset
handler, it should allow pre-bl31 bootloader to pass config
data into itself. Refer to below mainline patch
https://github.com/ARM-software/arm-trusted-firmware/commit/
25844ff728e4a0e5430ba2032457aba7b780a701
|
|
|
|
Trusty spd implements the memory share subset of FF-A, this
commit enables related configs. As the space to contain TF-A
is limited on some platforms, this commit refines the TF-A size
by:
1. Reduce the TRUSTY_SHARED_MEMORY_OBJ_SIZE from 512KB to
12KB, widevine playback stress pass on imx8mp.
2. Move the xlat table to ocram for imx8mp, imx8mn, imx8mm.
3. Disable FF-A for imx8mq as no more ocram can be used.
Change-Id: I86be12baf96dfbe5ae80c09d672941b233423cad
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
On some ocram limited platforms, the xlat table and stack are
moved to separate address to avoid build break. This commit makes
the address configurable (instead of 0x180000) so it can apply
to other platforms.
Change-Id: I8db72c86d7fa8da9345fb0c132277d1ed8ce9e59
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
If M7 audio enabled, system suspend will changed from DSM to "fast
wakeup" suspend mode, at this phase origal ATF design kept DRAM
,PLLs, clocks same as Linux runtime environment, this will cause even
system is under "fast wakeup" suspend, the power is simlar as run
time. This patch will switch DDR freqs from "normal run time" freq to
400MTS, then make DDR into rentention mode, disable all unused PLLs,
clocks, bypass M7 required clocks to 24M in 'fast wakeup' suspend mode.
Change-Id: Ibc21f250d47475bdb42d9018b9be9ca503387c86
Signed-off-by: Zhipeng Wang <zhipeng.wang_1@nxp.com>
|
|
The imx8mq platform only reserves 64KB (0x910000~0x91ffff)
for ATF usage, which is not enough and will cause build
break like below:
build/imx8mq/release/bl31/bl31.elf section `coherent_ram' will not fit in region `RAM'
BL31 image has exceeded its limit.
region `RAM' overflowed by 4096 bytes
Makefile:846: recipe for target 'build/imx8mq/release/bl31/bl31.elf' failed
make: *** [build/imx8mq/release/bl31/bl31.elf] Error 1
This commit will reduce the code size for imx8mq by doing:
1. Remove ddr4 dvfs support, it's not used on imx8mq.
2. Switch back to xlat_tables v1 instead of xlat_tables v2, which will
occupy more memory. Comment the dynamic mmap in trusty for imx8mq
as it can be removed.
3. Set the 'LOG_LEVEL' as '0' by default so some debug string won't
be built into the final image.
Change-Id: I68f2d6629918a5e7dd616822b5d41ef148f6664b
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
Build with trusty will take much more space, which may cause
build break. Reduce the trusty stack size for each cpu to save
some space.
Change-Id: I3bbf21747a4abaef82707a929907575e909ac4d7
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
HAB verify in ATF will touch TCM memory, this commit
adds the memory mapping.
Change-Id: I7511f74420fac769d94db9b1a3cae366bcad6b8e
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
Add security configuration(CSU/RDC/TZASC) for android imx8m
platforms.
Change-Id: I76466500f129542a7405ee167251330ed1d770c4
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
Add individual config for ddr4 dvfs, it's enabled for imx8m
by default.
Change-Id: Ic4d3b41fd46e7d182b44d76cf38a13e3dc7ad1d4
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
Allow non-privileged access to all SNVS registers when no
any TEE is available.
Change-Id: I42906e4ddda0ff484d94323b4ed3c9d28245a5f0
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
Update the csu config for imx8m.
Change-Id: Iec3aba43fc765e457fa4f2c4a54a9c18c2565749
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
Enable SNVS_LPCR_LPTA_EN and SNVS_LPCR_LPWUI_EN during shutdown
proces.
Change-Id: I566ce759ce7d121a476209836429aeee07cfe49d
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
Remove the SDEI suppor temprarily to make sure the
DDR DVFS support ok.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
|
|
imx8mq/mm/mn"
need to get more clear if the mr12/14 update is also apply to 8mq/mm/mn,
so revert this patch and keep this feature for 8MP only for now.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Acked-by: Peng Fan <peng.fan@nxp.com>
|
|
the DDR_IPS size also need to be updated for i.MX8MQ/MM/MN,
so fix it.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
|
|
Backup the mr12/14 value as the actual value used is not the
one we configured in the ddrc config timing.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
|
|
Fix Coverity issue: CID 16300011 caused by unnecessary header file.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
Remove unnecessary header file to fix below Coverity issue CID 16300011
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
Fix Coverity issue:
CID 6474575: Out-of-bounds access (OVERRUN)
CID 11014855: Unused value (UNUSED_VALUE)
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
Mapping the TCM memory in ATF so HAB verify can pass.
Restrict the MCU (Domain 1) write access to RDC and setup
the MCU RDC config when MCU starts/stops.
Change-Id: I1a90598095171abc9f57d305907b25ce66441cf3
Signed-off-by: Ji Luo <ji.luo@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
|
|
Configure the force_incr programmable bit in GPV_5 of PL301_display,
which fixes partial write issue. This workaround was done in MCU FW
before, move it to TF-A now as MCU should not touch secure world.
Change-Id: I2e5bbc764640afeab6ac2f4b202939b59bd3b3f2
Signed-off-by: Ji Luo <ji.luo@nxp.com>
|
|
The ocram_a, ocram & ocram_s csu csl index should be 114,
118, 119, so fix it.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
Check the domain_id to make sure the index passed by
the Rich-OS does not exceed the range of the domain
arrays.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
Lock the ocram secure access for imx8mn/imx8mp.
Signed-off-by: Ji Luo <ji.luo@nxp.com>
Change-Id: Ida3bba565c0f94948b7e04c90075cd3352bbf6e9
|
|
Add back the idx define for csu/rdc.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
In case JR0 is used by the HAB for secure boot, it can be used later
for authenticating kernel or other binaries.
We are checking if the HAB is using the JR by the DID set.
Signed-off-by: Franck LENORMAND <franck.lenormand@nxp.com>
|
|
Per i.MX8MP SRM, current idx is wrong, so update them.
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
Signed-off-by: Peng Fan <peng.fan@nxp.com>
|
|
With rpmsg dts, BT can only work while audio is playing.
The reason is that mcu will enter stop mode if no audio
task running, this will affect some peripherals, like uart1.
Use RDC to set: only domain0(A core) can access uart1.
Signed-off-by: Jindong Yue <jindong.yue@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
|
|
Enable necessary clocks for HDMIMIX before adb port handshake.
Signed-off-by: Jindong Yue <jindong.yue@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
|
|
Correct the csu sa & hpctrl setting.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
|
|
APB Write data corruption following MRCTRL0.mr_wr=1 while
hardware-driven MR access is occurring
When performing a software driven MR access, the following
sequence must be done automatically before performing other
APB register accesses:
1. Set MRCTRL0.mr_wr=1
2. Check for MRSTAT.mr_wr_busy=0. If not, go to step (2)
3. Check for MRSTAT.mr_wr_busy=0 again (for the second time). If not, go to step (2)
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
|
|
Add the PLL frequency config for 4000mts.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Ye Li <ye.li@nxp.com>
|
|
Need to increase ISP NoC priority to 0x7 (same as LCDIF panic priority)
to avoid overflow in DDR4 EVK board
Signed-off-by: Jian Li <jian.li@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
|
|
Enable the OCRAM_S TZ for secure protection by default on i.MX8MN/i.MX8MP.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
iMX8MQ B2 chip uses same OCOTP magic value with B1. So read the ROM
version to distinguish it with B1.
Signed-off-by: Ye Li <ye.li@nxp.com>
Reviewed-by: Jacky Bai <ping.bai@nxp.com>
(cherry picked from commit dcfce7731ed19c6235f9b1ad3a68fbfe33b69f9a)
|
|
This is log is just for debug purpose only. Change the
ddr4 dvfs debug log print level to disable this log print
by default.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Peng Fan <peng.fan@nxp.com>
|
|
Normally, the wdog1 is used by A53 side, and it should
be stopped when A53 domain enter STOP mode. when system
out of PoR, this watchdog is owned by both M7 & A53 side,
then this watchdog can only enter STOP mode only when
both A53 & M7 enter STOP mode. it is not reasonable as
this watchdog is only used by A53 side, so assign wdog1
to domain0(a53 side) only.
Signed-off-by: Jacky Bai <ping.bai@nxp.com>
Reviewed-by: Anson Huang <anson.huang@nxp.com>
|