Fix: PDK-12550: C7x events overlaps between multiple UDMA instances.
Out of num of IR interrupts allocated to core via boardconfig, UDMA driver
will use most of them after reserving some interrupts for other drivers
like IPC. Even after reserving the interrupt to other drivers, if interrupts
available for UDMA driver are more than C7x events available for UDMA driver
then driver tries to program C7x event which is not available, results in failure.
This commit makes sure that the Ir interrupts allocated for core,
after removing Ir interrupts reserved for other drivers
(i.e. rmDefBoardCfgResp[UDMA_RM_RES_ID_IR_INTR].rangeNum - totalResrvIntr),
does not exceed available C7x Event for UDMA driver. And also makes sure
the 2 different UDMA instances will never try to program same event of C7x.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Out of num of IR interrupts allocated to core via boardconfig, UDMA driver
will use most of them after reserving some interrupts for other drivers
like IPC. Even after reserving the interrupt to other drivers, if interrupts
available for UDMA driver are more than C7x events available for UDMA driver
then driver tries to program C7x event which is not available, results in failure.
This commit makes sure that the Ir interrupts allocated for core,
after removing Ir interrupts reserved for other drivers
(i.e. rmDefBoardCfgResp[UDMA_RM_RES_ID_IR_INTR].rangeNum - totalResrvIntr),
does not exceed available C7x Event for UDMA driver. And also makes sure
the 2 different UDMA instances will never try to program same event of C7x.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
QNX: IPC: Examples: set the btcm compiler flag for echo_testb
The ipc_echo_testb needs to be built with BTCM support enabled to
ensure proper boot flow. The compile time switch for enabling this
was missing. Added the same.
Signed-off-by: Subbu Chanderashekarapuram <s-chanderashekarapuram@ti.com>
The ipc_echo_testb needs to be built with BTCM support enabled to
ensure proper boot flow. The compile time switch for enabling this
was missing. Added the same.
Signed-off-by: Subbu Chanderashekarapuram <s-chanderashekarapuram@ti.com>
[PDK-12549]: Include mcu4_0, mcu4_1 while generating lateapp2
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
PDK-12440: Updated pinmux unlock/lock sequence for optimal execution
- MMR unlock and lock added in board pinmux config function is increasing
the boot time by 5msec due to multiple MMR unlock and lock.
Moved the MMR unlock and lock to top level functions to reduce the
number of MMR unlock and lock operations.
- MMR unlock and lock added in board pinmux config function is increasing
the boot time by 5msec due to multiple MMR unlock and lock.
Moved the MMR unlock and lock to top level functions to reduce the
number of MMR unlock and lock operations.
lwip:ut:add missing interrupt macros for ping and pong tasks
add missing interrupt macros for ping and pong tasks for J784s4
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
add missing interrupt macros for ping and pong tasks for J784s4
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
2 days ago[PDK-12327] osal: arch: c66: bypass osalArch_TimestampInit for freertos REL.CORESDK.08.06.00.12 REL.CORESDK.08.06.00.13
[PDK-12327] osal: arch: c66: bypass osalArch_TimestampInit for freertos
- With usage of TSC counter in freertos portable layer for runtime
calculation, bypass osalArch_TimestampInit for freertos
to avoid re-initialization of TSC counter.
- FreeRTOS initialize TSC counter as part of schedular start
- Also remove bypassing CycleprofilerP_init for FreeRTOS, since this is
now internally handled in corresponding osalArch implementations
Signed-off-by: Don Dominic <a0486429@ti.com>
- With usage of TSC counter in freertos portable layer for runtime
calculation, bypass osalArch_TimestampInit for freertos
to avoid re-initialization of TSC counter.
- FreeRTOS initialize TSC counter as part of schedular start
- Also remove bypassing CycleprofilerP_init for FreeRTOS, since this is
now internally handled in corresponding osalArch implementations
Signed-off-by: Don Dominic <a0486429@ti.com>
[PDK-12327] osal: arch: r5: updates in timestamp implementation for freertos
- With usage of PMU counter in freertos portable layer for runtime
calculation, osal timestamp related updates are required
to avoid cases like re-initialization of PMU counter and duplicate
counter overflow handling
- Bypass osalArch_TimestampInit for FreeRTOS, since this is already
initialized by during freertos schedular start
- Reuse freertos portable layer implementation of reading PMU counter
value in osalArch_TimestampGet64 to avoid duplicate overflow handle
- To avoid duplicate overflow handle, bypass
osalArch_TimestampCcntAutoRefresh for FreeRTOS, since
its already handled by freertos portable layer in every OS ticks.
Signed-off-by: Don Dominic <a0486429@ti.com>
- With usage of PMU counter in freertos portable layer for runtime
calculation, osal timestamp related updates are required
to avoid cases like re-initialization of PMU counter and duplicate
counter overflow handling
- Bypass osalArch_TimestampInit for FreeRTOS, since this is already
initialized by during freertos schedular start
- Reuse freertos portable layer implementation of reading PMU counter
value in osalArch_TimestampGet64 to avoid duplicate overflow handle
- To avoid duplicate overflow handle, bypass
osalArch_TimestampCcntAutoRefresh for FreeRTOS, since
its already handled by freertos portable layer in every OS ticks.
Signed-off-by: Don Dominic <a0486429@ti.com>
[PDK-12327]freertos: r5f: port: use pmu counter for runtime calc in microseconds
Background:-
- PMU counter can't be used for runtime calc since PMU counter halts in
wfi(IDLE) mode. Idle task runtime will be incorrect due to this and
CPU Load will always be close to 100%
- Hence OS Tick Timer (DM Timer) was used for runtime calc
- milliseconds was calculated based on no. of OS ticks. OS ticks gets
incremented based on DM Timer Interrupts
- microseconds was calculated based on DM Timer count
Issue/Impact:-
- DM Timer based runtime calculation adds overhead of reading DM Timer
register each time to calculate microseconds
- This could also go wrong (off by ~1 ms) when DM Timer overflow occurred
and the current run time calculation is invoked before servicing the
DM Timer interrupt
- This delta can get accumulated in total runtime counter for a task and
could result in total CPU load measurements to be > 100%
- Fixing DM Timer based Implementation to calculate microseconds would be
hacky and may need additional DM Timer register read which will
further increase the overhead
Fix:-
- De-scope usage of DM Timer count and use PMU Counter to calculate
microseconds elapsed after last OS tick
- Initialize PMU counter as part of scheduler start in
vPortConfigTimerForRunTimeStats
- Save PMU counter TS during each OS tick
- This will also handle 32-bit PMU counter overflow
- Hence remove usage of CycleprofilerP_refreshCounter in every
OS tick interrupt
- For IDLE task, since PMU counter halts in wfi(IDLE) mode, idle task
runtime counter value could be off by 1 OS tick
Signed-off-by: Don Dominic <a0486429@ti.com>
Background:-
- PMU counter can't be used for runtime calc since PMU counter halts in
wfi(IDLE) mode. Idle task runtime will be incorrect due to this and
CPU Load will always be close to 100%
- Hence OS Tick Timer (DM Timer) was used for runtime calc
- milliseconds was calculated based on no. of OS ticks. OS ticks gets
incremented based on DM Timer Interrupts
- microseconds was calculated based on DM Timer count
Issue/Impact:-
- DM Timer based runtime calculation adds overhead of reading DM Timer
register each time to calculate microseconds
- This could also go wrong (off by ~1 ms) when DM Timer overflow occurred
and the current run time calculation is invoked before servicing the
DM Timer interrupt
- This delta can get accumulated in total runtime counter for a task and
could result in total CPU load measurements to be > 100%
- Fixing DM Timer based Implementation to calculate microseconds would be
hacky and may need additional DM Timer register read which will
further increase the overhead
Fix:-
- De-scope usage of DM Timer count and use PMU Counter to calculate
microseconds elapsed after last OS tick
- Initialize PMU counter as part of scheduler start in
vPortConfigTimerForRunTimeStats
- Save PMU counter TS during each OS tick
- This will also handle 32-bit PMU counter overflow
- Hence remove usage of CycleprofilerP_refreshCounter in every
OS tick interrupt
- For IDLE task, since PMU counter halts in wfi(IDLE) mode, idle task
runtime counter value could be off by 1 OS tick
Signed-off-by: Don Dominic <a0486429@ti.com>
Migrating to SYSFW version v08.06.01a
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
[PDK-12541][J721E]: Fix SBL profiling logs
- Since j721e doesn't have mcu timer9, SBL profile info for j721e is one index ahead of other SOC's
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Since j721e doesn't have mcu timer9, SBL profile info for j721e is one index ahead of other SOC's
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
J721S2: J784S4: Add BCDMA CH API Testcases to UDMA UT.
- Add BCDMA CH params and API Tests to check UDMA API able to program BCDMA channels.
- BCDMA doesn't support global flow event, updated the TCs which checks programming
of these events to support only UDMA instances.
- Add core specific checks to check whether BCDMA instance is applicable for specific core.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
- Add BCDMA CH params and API Tests to check UDMA API able to program BCDMA channels.
- BCDMA doesn't support global flow event, updated the TCs which checks programming
of these events to support only UDMA instances.
- Add core specific checks to check whether BCDMA instance is applicable for specific core.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Remove Linux Binaries used to boot Linux from SBL
- Since these binaries already present under
sbl/tools/combined_appimage/bin/<board_name> removing here.
- Remove k3-<board_name>-common-proc-board.dtb since this needs to be
generated from ti-linux-kernel.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Since these binaries already present under
sbl/tools/combined_appimage/bin/<board_name> removing here.
- Remove k3-<board_name>-common-proc-board.dtb since this needs to be
generated from ti-linux-kernel.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
lwip: unit test: added initial version of lwip OS port test app
added test app that validates lwip port api's for both safertos and freertos
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
added test app that validates lwip port api's for both safertos and freertos
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
lwip: safertos: added support of safertos for lwip-port
- added sys_arch.c for safertos, implements native api similar to freertos
- added required lwippools.h and lwipopts.h,
to add required macros and lwip memory pools
- added sys_arch.h, perf.h and cc.h,
to add required structures and function definitions
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
- added sys_arch.c for safertos, implements native api similar to freertos
- added required lwippools.h and lwipopts.h,
to add required macros and lwip memory pools
- added sys_arch.h, perf.h and cc.h,
to add required structures and function definitions
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
lwip: freertos: renamed macros to make it OS agnostic, api sys_mbox_trypost_fromisr updated
- renamed macros LWIP_FREERTOS_* to LWIP_* to make it OS agnostic
- changed the return type of sys_mbox_trypost_fromisr to ERR_OK in case there is a context switch
- called portYIELD_FROM_ISR in sys_mbox_trypost_fromisr
so the application doesn't need to do a context switch
- added FreeRTOS.h and task.h header files to make consistent with other OS
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
- renamed macros LWIP_FREERTOS_* to LWIP_* to make it OS agnostic
- changed the return type of sys_mbox_trypost_fromisr to ERR_OK in case there is a context switch
- called portYIELD_FROM_ISR in sys_mbox_trypost_fromisr
so the application doesn't need to do a context switch
- added FreeRTOS.h and task.h header files to make consistent with other OS
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
diag: enet_icssg: removed unneeded enet_cfgserver from COMP_LIST_COMMON
enet_icssg is a baremetal application and has no dependency on enet_cfgserver
hence removed it from COMP_LIST_COMMON
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
enet_icssg is a baremetal application and has no dependency on enet_cfgserver
hence removed it from COMP_LIST_COMMON
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
PDK-12495: Board: Added support for spare area write in NAND flash library
- App images flashed to OSPI NAND memory are not booting as RBL is
checking spare area and might be stopping the boot process due to
invalid data in spare area.
Updated NAND flash library to include spare area write during data
write operation.
- App images flashed to OSPI NAND memory are not booting as RBL is
checking spare area and might be stopping the boot process due to
invalid data in spare area.
Updated NAND flash library to include spare area write during data
write operation.
Board: Fix for J721S2 & J784S4 board library klocwork error
[PDK-12327] freertos: c66x: port: use TSC counter for runtime calc
- Use TSC counter in C66x CPU for calculating current run time
- This reverts commit c6e34646389c5910fda8e1a2787bcc53a98aff24
"[PDK-10107] FreeRTOS: C66x: Use OS Tick Timer instead of TSC Counter"
which was updated along with R5F so that the portable layer would be
having similar implementation for both ISA.
- Unlike R5F PMU counter halts during wfi(IDLE) mode,
There is no such limitation in C66x to use TSC counter
- Also usage of TSC counter available in the CPU itself will be having
lesser overhead compared to reading DM Timer (OS Tick Timer) registers
each time for run time calculations
- This also fixes PDK-12327 on C66x, FreeRTOS task stats reporting total
CPU load > 100%
Signed-off-by: Don Dominic <a0486429@ti.com>
- Use TSC counter in C66x CPU for calculating current run time
- This reverts commit c6e34646389c5910fda8e1a2787bcc53a98aff24
"[PDK-10107] FreeRTOS: C66x: Use OS Tick Timer instead of TSC Counter"
which was updated along with R5F so that the portable layer would be
having similar implementation for both ISA.
- Unlike R5F PMU counter halts during wfi(IDLE) mode,
There is no such limitation in C66x to use TSC counter
- Also usage of TSC counter available in the CPU itself will be having
lesser overhead compared to reading DM Timer (OS Tick Timer) registers
each time for run time calculations
- This also fixes PDK-12327 on C66x, FreeRTOS task stats reporting total
CPU load > 100%
Signed-off-by: Don Dominic <a0486429@ti.com>
UDMA UT: Fixed inconsistent hang issue on C7x Cores
Global variable gUtils_prfObj is having garbage values
before profiling getting started initializing it with
0 resolved the inconsistent hangs on C7x.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Global variable gUtils_prfObj is having garbage values
before profiling getting started initializing it with
0 resolved the inconsistent hangs on C7x.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
OSPI: Configure Phy operating mode basing on input provided.
Configure Phy to operate on master mode or bypass mode basing on
input provided by user. By default the operating mode of PHY will
be programmed basing on OSPI clk.
User can override phyOpMode parameter in OSPI_Params to choose
the operating mode of PHY.
Fixes: PDK-12406
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Configure Phy to operate on master mode or bypass mode basing on
input provided by user. By default the operating mode of PHY will
be programmed basing on OSPI clk.
User can override phyOpMode parameter in OSPI_Params to choose
the operating mode of PHY.
Fixes: PDK-12406
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[PDK-12531] freertos: c7x: port: account for scheduler start time
Background:-
- TSC Counter is used to implement uiPortGetRunTimeCounterValue
- TSC can be accessed only in secure supervisor mode and TSC counter
gets reset during boot time in secure supervisor mode.
- Scheduler start is done supervisor mode. So we don’t have option to
reset TSC (will cause permission exception).
- But current implementation of uiPortGetRunTimeCounterValue is not
taking into consideration when the scheduler is started
Impact:-
- This will result in the Timer Task Run time to be very high.
This is because time from boot to scheduler start will is also get
accounted in the Timer Task Run time, since this is the first task
that will be executed after scheduler start.
Fix:-
- Account for Scheduler start time in uiPortGetRunTimeCounterValue
calculations.
- Store scheduler start TSC counter timestamp
- Use the same in uiPortGetRunTimeCounterValue implementation to
account for Scheduler start time
Signed-off-by: Don Dominic <a0486429@ti.com>
Background:-
- TSC Counter is used to implement uiPortGetRunTimeCounterValue
- TSC can be accessed only in secure supervisor mode and TSC counter
gets reset during boot time in secure supervisor mode.
- Scheduler start is done supervisor mode. So we don’t have option to
reset TSC (will cause permission exception).
- But current implementation of uiPortGetRunTimeCounterValue is not
taking into consideration when the scheduler is started
Impact:-
- This will result in the Timer Task Run time to be very high.
This is because time from boot to scheduler start will is also get
accounted in the Timer Task Run time, since this is the first task
that will be executed after scheduler start.
Fix:-
- Account for Scheduler start time in uiPortGetRunTimeCounterValue
calculations.
- Store scheduler start TSC counter timestamp
- Use the same in uiPortGetRunTimeCounterValue implementation to
account for Scheduler start time
Signed-off-by: Don Dominic <a0486429@ti.com>
6 days ago[PDK-10891][PDK-10957][UART]: UART_Baremetal_DMA_TestApp FIFO Trigger level fix. REL.CORESDK.08.06.00.09
[PDK-10891][PDK-10957][UART]: UART_Baremetal_DMA_TestApp FIFO Trigger level fix.
- Performed Cache invalidation of the FIFO trigger level data.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- Performed Cache invalidation of the FIFO trigger level data.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
Fix: PDK-12501: UDMA init fails on C7x if MAIN UDMA/BCDMA runs on same core
UDMA init failing during interrupt registration on C7x cores if MAIN UDMA/BCDMA runs on same core.
This is due to usage of same interrupt number for both instances. Making use of different ranges
for different instances solves the problem.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
UDMA init failing during interrupt registration on C7x cores if MAIN UDMA/BCDMA runs on same core.
This is due to usage of same interrupt number for both instances. Making use of different ranges
for different instances solves the problem.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
MCSPI: Add 48 bit address support
Currrently MCSPI dma transfers resticted to 32 bit address space,
since the txBuf/rxBuf is void*, converting it to uint64_t to support
transfers upto 48 bit.
Fixes: PDK-12258
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Currrently MCSPI dma transfers resticted to 32 bit address space,
since the txBuf/rxBuf is void*, converting it to uint64_t to support
transfers upto 48 bit.
Fixes: PDK-12258
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[PDK-12512][OSAL_SAFERTOS]: Enable LPSC_SPARE_0 for main domain cores on J7AHP
- For J7AHP, main-domain timers 8 to 19 are connected to LPSC_PER_SPARE_0.
- MCU4_0 and MCU4_1 use DMTimer 8 and DMTimer 9 as tick timers.
- LPSC_SPARE_0 is not enabled by default.
- Added code to enable LPSC_SPARE_0 in prvSetupTimerPSC() for main domain cores.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- For J7AHP, main-domain timers 8 to 19 are connected to LPSC_PER_SPARE_0.
- MCU4_0 and MCU4_1 use DMTimer 8 and DMTimer 9 as tick timers.
- LPSC_SPARE_0 is not enabled by default.
- Added code to enable LPSC_SPARE_0 in prvSetupTimerPSC() for main domain cores.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
[PDK] Static Analysis fixes for Board, GPIO, LPM, UART, OSAL
Signed-off-by: Jahnavi_Guvvala <j-guvvala@ti.com>
Signed-off-by: Jahnavi_Guvvala <j-guvvala@ti.com>
Bug Fix: PDK-12324
- Enabled UART for all RTOS apps
- This is needed for IPC negative tests
Signed-off-by: Rishabh Garg <rishabh@ti.com>
- Enabled UART for all RTOS apps
- This is needed for IPC negative tests
Signed-off-by: Rishabh Garg <rishabh@ti.com>
[PDK-12503][SBL][XIP]: Add separate target for SBL to boot XIP application
- Use SBL_updateXIPFlag() to update the global variable isXIPdefined
- isXIPdefined is used to check whether SBL is booting xip application
or not
- Add PASS string at the end of XIP application
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Use SBL_updateXIPFlag() to update the global variable isXIPdefined
- isXIPdefined is used to check whether SBL is booting xip application
or not
- Add PASS string at the end of XIP application
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Revert "IPC: stack corruption of taskWaiter used in RPMessage_getRemoteEndPt"
This reverts commit a9c8e4160d737b27be34fe8462fbe20e3cf19e08.
This reverts commit a9c8e4160d737b27be34fe8462fbe20e3cf19e08.
Fix: UDMA: KW critical issues
KW throws UNINIT.STACK.MIGHT since udmapChanStats, bcdmaChanStats, udmapRtEnable,
bcdmaRtEnable, have not been initialised in the same function call.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
KW throws UNINIT.STACK.MIGHT since udmapChanStats, bcdmaChanStats, udmapRtEnable,
bcdmaRtEnable, have not been initialised in the same function call.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[PDK-12508]: Change sbl_<boot_mode>_img_hs path to $(PDK_SBL_COMP_PATH)/board/k3
- Earlier We have same APP names for both GP and HS. Since APP names
were same we change the path to store the generated files.
- Now since we have differnet APP names for GP and HS same path can be
used
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Earlier We have same APP names for both GP and HS. Since APP names
were same we change the path to store the generated files.
- Now since we have differnet APP names for GP and HS same path can be
used
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Fix: UDMA UT: Hanging on C7x cores.
- Issue caused due to optimisation of var `loadStatsTask` by C7x complier
in release mode. Making the variable global fixes the issue.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
- Issue caused due to optimisation of var `loadStatsTask` by C7x complier
in release mode. Making the variable global fixes the issue.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
keywriter: Add variable check in generate_test_binaries.sh
The generate_test_binaries.sh requires several exported variables.
PDK_INSTALL_PATH and SOC must be defined. PDK_KEYWR_COMP_PATH can
be derived.
Fixes: PDK-12459
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
The generate_test_binaries.sh requires several exported variables.
PDK_INSTALL_PATH and SOC must be defined. PDK_KEYWR_COMP_PATH can
be derived.
Fixes: PDK-12459
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
keywriter: Set scripts to use bash
Explicitly declare the script to be a bash script.
Fixes: PDK-12458
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
Explicitly declare the script to be a bash script.
Fixes: PDK-12458
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
keywriter: Add j784s4 SoC files
Add boardcfg snapshot based on latest j784s4 firmware build
Add j784s4 utils to enable VPP efuse via WKUP_GPIO_54
Add empty ti-fs-keywriter.bin. Fill expected size with zeros.
Fixes: PDK-12441
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
Add boardcfg snapshot based on latest j784s4 firmware build
Add j784s4 utils to enable VPP efuse via WKUP_GPIO_54
Add empty ti-fs-keywriter.bin. Fill expected size with zeros.
Fixes: PDK-12441
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
keywriter: Simplify macros in main.h
The macros are the same for J721E, J7200, and J721S2. Combine the
macros into the same code block.
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
The macros are the same for J721E, J7200, and J721S2. Combine the
macros into the same code block.
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
keywriter: aesenc_smek.enc must be generated
Re-enable gen_sym_key_x509_extension to generate aesenc_smek.enc. Keep
the encrypt smek message if smek is given.
Fixes: PDK-12456
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
Re-enable gen_sym_key_x509_extension to generate aesenc_smek.enc. Keep
the encrypt smek message if smek is given.
Fixes: PDK-12456
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
9 days ago[FreeRTOS][SafeRTOS]: Pushing out the C7X support files into csl/arch. REL.CORESDK.08.06.00.06 REL.CORESDK.08.06.00.07
[FreeRTOS][SafeRTOS]: Pushing out the C7X support files into csl/arch.
- SafeRTOS uses the same C7X support files as FreeRTOS.
- To avoid redundant code, we push the common support files to CSL.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- SafeRTOS uses the same C7X support files as FreeRTOS.
- To avoid redundant code, we push the common support files to CSL.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
[SafeRTOS]: Added SafeRTOS support for J784S4 and integrated new WHIS packages for J721E and J721S2.
- Version of WHIS Packages:
<J721E,R5F>: 009-004-199-024-219-001.
<J721E,C7X>: 009-004-230-005-219-001.
<J721E,C66>: 009-002-201-005-219-002.
<J7200,R5F>: 009-002-199-024-243-001.
<J721S2,R5F>: 009-004-199-024-235-002.
<J721S2,C7X>: 009-004-230-005-235-001.
<J784S4,R5F>: 009-004-199-024-251-001.
<J784S4,C7X>: 009-004-230-005-251-001.
- **For the R5F on TDA4V-X packages, a lot of APIs have their names changed.**
- This is not the case for DRA821(J7200), and hence the respective API calls is SOC protected accordingly.
- There is a new SafeRTOS_utils_j7200_r5f.asm file, as the vDataAbort function has changed into vApplicationDataAbortHook for TDA4V-X SOCs, but not for DRA821.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- Version of WHIS Packages:
<J721E,R5F>: 009-004-199-024-219-001.
<J721E,C7X>: 009-004-230-005-219-001.
<J721E,C66>: 009-002-201-005-219-002.
<J7200,R5F>: 009-002-199-024-243-001.
<J721S2,R5F>: 009-004-199-024-235-002.
<J721S2,C7X>: 009-004-230-005-235-001.
<J784S4,R5F>: 009-004-199-024-251-001.
<J784S4,C7X>: 009-004-230-005-251-001.
- **For the R5F on TDA4V-X packages, a lot of APIs have their names changed.**
- This is not the case for DRA821(J7200), and hence the respective API calls is SOC protected accordingly.
- There is a new SafeRTOS_utils_j7200_r5f.asm file, as the vDataAbort function has changed into vApplicationDataAbortHook for TDA4V-X SOCs, but not for DRA821.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
Migrating to TIFS version v08_06_01
Signed-off-by: Don Dominic <a0486429@ti.com>
Signed-off-by: Don Dominic <a0486429@ti.com>
[PDK-10977] : Linker file changes to build sbl_boot_xip_test for j721s2,
j784s4
- Change OCMRAM_SBL_SYSFW memory to 0x41C83000 since 0x41C3E000 is used
by SBL
- Issue mentioned in the JIRA is a validation issue
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
j784s4
- Change OCMRAM_SBL_SYSFW memory to 0x41C83000 since 0x41C3E000 is used
by SBL
- Issue mentioned in the JIRA is a validation issue
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Reverted MCSPI driver changes
[PDK-12106][OSAL_SafeRTOS]: HwiP_disable/restore can only be called from Privileged Mode on R5.
- Hwip_disable/restore should only be called from privileged mode.
- Added an Osal utils API: Osal_isInPrivilegeMode.
- The API returns true if called from privileged mode, and false otherwise.
- Used Osal_isInPrivilegedMode to find the privilege mode of the task calling HwiP_disable/restore.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- Hwip_disable/restore should only be called from privileged mode.
- Added an Osal utils API: Osal_isInPrivilegeMode.
- The API returns true if called from privileged mode, and false otherwise.
- Used Osal_isInPrivilegedMode to find the privilege mode of the task calling HwiP_disable/restore.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
Safertos: UDMA UT: Build fix
Safertos doesnot support LoadP_getCPULoad and LoadP_getTaskLoad API's.
Using utils_prf_baremetal.c to fix build.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Safertos doesnot support LoadP_getCPULoad and LoadP_getTaskLoad API's.
Using utils_prf_baremetal.c to fix build.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Fix: MCSPI Build failure
fix build failure due to modification of structure from
void * to uint64_t.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
fix build failure due to modification of structure from
void * to uint64_t.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Fix: MCSPI build failure.
MCSPI facing build failures after modifying the transaction
structure's txBuf/rxBuf from void * to uint64_t. Fix all the
places where txBuf/rxBuf getting typecasted to void *.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
MCSPI facing build failures after modifying the transaction
structure's txBuf/rxBuf from void * to uint64_t. Fix all the
places where txBuf/rxBuf getting typecasted to void *.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
FVID2 klocwork fixes for the MISRAC checker code MISRA.FUNC.UNUSEDPAR.2012
Signed-off-by: Asha <x1101668@ti.com>
Signed-off-by: Asha <x1101668@ti.com>
Fix: UDMA UT linking failures.
Add udma_test_proxy.c udma_test_ring_monitor.c those got removed
from makefile with commit 956aaad9e19, creating build failures
for UDMA UT.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Add udma_test_proxy.c udma_test_ring_monitor.c those got removed
from makefile with commit 956aaad9e19, creating build failures
for UDMA UT.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[PDK-12487][SBL]: Add separate target for SBL to measure boot performance.
- Add sbl_boot_perf_cust_img to measure boot performance using
sbl_boot_perf_test.
- Remove dead code.
- Remove unused linker files
- cleaned up sbl_boot_perf_test code
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Add sbl_boot_perf_cust_img to measure boot performance using
sbl_boot_perf_test.
- Remove dead code.
- Remove unused linker files
- cleaned up sbl_boot_perf_test code
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Bug Fix: PDK-12324
- Board init was added to all IPC libraries
- Intent was to enable DM traces and hence Board Init should be done
only for mcu1_0
- Made the fix in example
Signed-off-by: Rishabh Garg <rishabh@ti.com>
- Board init was added to all IPC libraries
- Intent was to enable DM traces and hence Board Init should be done
only for mcu1_0
- Made the fix in example
Signed-off-by: Rishabh Garg <rishabh@ti.com>
Fix: PDK-12437: UDMA UT prints CPU load as 100%
- UDMA UT was using baremetal calculation instead of
using freeRTOS API's to profile CPU usage and
always prints 100% cpu usage. This commit uses freeRTOS
API's to profile UDMA UT.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
- UDMA UT was using baremetal calculation instead of
using freeRTOS API's to profile CPU usage and
always prints 100% cpu usage. This commit uses freeRTOS
API's to profile UDMA UT.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[PDK-12392]: Making xUsingFPU as pdTRUE in SafeRTOS_TaskP_create.
- SafeRTOS requires xUsingFPU to be pdTrue for using the FPU.
- SafeRTOS requires xUsingFPU to be pdTrue for using the FPU.
[PDK-12457][Boot Perf]: Add logic to print time taken by specific API's
- Remove SBL_ADD_PROFILE_POINT at multiple places wherever not required
- Calculate RBL execution time by reading mcu timer9
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Remove SBL_ADD_PROFILE_POINT at multiple places wherever not required
- Calculate RBL execution time by reading mcu timer9
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
[PDK-12449] : Fix SBL doesn't print logs with CUST SBL flags.
- Issue was application overwriting memory allocated to TIFS.
- Allocate memory for application after the memory reserved for TIFS.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Issue was application overwriting memory allocated to TIFS.
- Allocate memory for application after the memory reserved for TIFS.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Fix: PDK-12435: UDMA UT does not report correct statistics on C7x cores
- GT traces not working properly on C7x cores.
- Bypass GT traces and calling UdmaUt_printf makes logs print properly.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
- GT traces not working properly on C7x cores.
- Bypass GT traces and calling UdmaUt_printf makes logs print properly.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[PDK-12085]: Add HPD_PULSE event Support
- The sink can generate HPD_PULSE event.
- This event needs to handled by the driver DP handler.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- The sink can generate HPD_PULSE event.
- This event needs to handled by the driver DP handler.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
[BUILD_FIX]: Fix Keywriter build error due to b6374d278e3
- Rename 'ti-fs-kewriter.h' to 'tifs_keywriter.h'.
- This header was auto-generated by the bin2c.out which will include #ifdef <filename>.
- Using '-' will cause violation of -Wextra-tokens and -Wc99-extensions.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
Signed-off-by: Don Dominic <dondmadappallil@ti.com>
- Rename 'ti-fs-kewriter.h' to 'tifs_keywriter.h'.
- This header was auto-generated by the bin2c.out which will include #ifdef <filename>.
- Using '-' will cause violation of -Wextra-tokens and -Wc99-extensions.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
Signed-off-by: Don Dominic <dondmadappallil@ti.com>
osal: TaskP: Add support for storing user-defined data
Add support for applications to pass a user-defined data (also called
TLS) to be associated with the task to be created. This is being added
for FreeRTOS and SafeRTOS OSAL implementations only.
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
Add support for applications to pass a user-defined data (also called
TLS) to be associated with the task to be created. This is being added
for FreeRTOS and SafeRTOS OSAL implementations only.
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
PDK-12485 - osal: safertos: ClockP: Initialize xTaskToNotify to avoid conflict with pxCallbackFunction
SafeRTOS timer can be provided with either a task or callback function,
but not both which is treated as an invalid parameter error.
The OSAL implementation for SafeRTOS uses pxCallbackFunction, hence it
must not use task by explicitly clearing xTaskToNotify timer parameter.
xTaskToNotify was not explicitly set which caused ClockP_create().
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
SafeRTOS timer can be provided with either a task or callback function,
but not both which is treated as an invalid parameter error.
The OSAL implementation for SafeRTOS uses pxCallbackFunction, hence it
must not use task by explicitly clearing xTaskToNotify timer parameter.
xTaskToNotify was not explicitly set which caused ClockP_create().
Signed-off-by: Misael Lopez Cruz <misael.lopez@ti.com>
Fixed build error due to autogen
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
Updated Autogen Script for Boardcfg Hex File
To resolve the MisraC error, update the auto-gen script to
include the file name in the header.
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
To resolve the MisraC error, update the auto-gen script to
include the file name in the header.
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
[BUILD]: Added all safertos examples to the safertos_examples list.
- All examples ending in `_safertos` is added to the safertos_examples list.
- Change in APP_NAME of the following applications:
- safertos_test_task_switch-->task_switch_safertos
- safertos_test_ut-->unit_test_safertos
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- All examples ending in `_safertos` is added to the safertos_examples list.
- Change in APP_NAME of the following applications:
- safertos_test_task_switch-->task_switch_safertos
- safertos_test_ut-->unit_test_safertos
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
IPC: stack corruption of taskWaiter used in RPMessage_getRemoteEndPt
Signed-off-by: Likhitha <x1097556@ti.com>
Signed-off-by: Likhitha <x1097556@ti.com>
MCSPI: Add 48 bit address support
Currrently MCSPI dma transfers resticted to 32 bit address space,
since the txBuf/rxBuf is void*, converting it to uint64_t to support
transfers upto 48 bit.
Fixes: PDK-12258
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Currrently MCSPI dma transfers resticted to 32 bit address space,
since the txBuf/rxBuf is void*, converting it to uint64_t to support
transfers upto 48 bit.
Fixes: PDK-12258
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[SBL] : Add prebuilt binaries to test booting linux/U-Boot from SBL
and booting Linux.QNX from BootApp
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
and booting Linux.QNX from BootApp
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
J721S2: Update Indirect trigger address location.
The Indirect AHB Trigger Address for j721s2 is currently programmed as 0x50000000 + 0x0U which is cachable, so all the memcpy writes may not landing up in actual memory. Location of SRAM will start from this trigger address. Making this address to point to a non-cachable location fixes the problem.
Fixes: PDK-12255
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
The Indirect AHB Trigger Address for j721s2 is currently programmed as 0x50000000 + 0x0U which is cachable, so all the memcpy writes may not landing up in actual memory. Location of SRAM will start from this trigger address. Making this address to point to a non-cachable location fixes the problem.
Fixes: PDK-12255
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
[PDK-12466][SafeRTOS UT]: Refreshing PMU counter in the SafeRTOS tick timer hook.
- PMU counter overflows after 1 minute and the safertos_test_ut fails.
- Refreshing the PMU counter in the SafeRTOS OS tick hook function.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- PMU counter overflows after 1 minute and the safertos_test_ut fails.
- Refreshing the PMU counter in the SafeRTOS OS tick hook function.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
[PDK-12483][SBL] : Change SBL_SCRATCH_MEM_START to 0xC2000000
- Since vision apps were using 0xB8000000, changing the
SBL_SCRATCH_MEM_START to 0xC2000000
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Since vision apps were using 0xB8000000, changing the
SBL_SCRATCH_MEM_START to 0xC2000000
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
MISRAC Fixes For OSAL Module
Osal Module Klocwork Fixes for Below Misrac Checker Codes:
-> misra.asm.encaps,misra.builtin_numeric, misra.stmt.cond.not_boolean.2012
-> misra.logic.sideeff,misra.literal.unsigned.suffix,misra.init.size.implicit.2012
-> misra.init.partial.2012,misra.init.multiple.2012,misra.if.no_else,misra.if.no_compound
-> misra.func.unmatched.params,misra.func.no_params,misra.expr.parens.2012,
-> misra.etype.inappr.operand.binop.2012,misra.etype.category.different.2012
Signed-off-by: Talluri Sankaiah <x1101061@ti.com>
Osal Module Klocwork Fixes for Below Misrac Checker Codes:
-> misra.asm.encaps,misra.builtin_numeric, misra.stmt.cond.not_boolean.2012
-> misra.logic.sideeff,misra.literal.unsigned.suffix,misra.init.size.implicit.2012
-> misra.init.partial.2012,misra.init.multiple.2012,misra.if.no_else,misra.if.no_compound
-> misra.func.unmatched.params,misra.func.no_params,misra.expr.parens.2012,
-> misra.etype.inappr.operand.binop.2012,misra.etype.category.different.2012
Signed-off-by: Talluri Sankaiah <x1101061@ti.com>
I2C MISRAC fixes for VA_UNUSED.GEN and MISRA.FUNC.UNMATCHED.PARAMS
Signed-off-by: Asha <x1101668@ti.com>
Signed-off-by: Asha <x1101668@ti.com>
[PDK-12233]Removing sciserver_hwiData.h header for sciserver.c
In the sciserver_hwiData.h, the structures which hold the
HWI and Task data are defined as static.Removing the header
from sciserver.c to avoid multiple creations of the same data.
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
In the sciserver_hwiData.h, the structures which hold the
HWI and Task data are defined as static.Removing the header
from sciserver.c to avoid multiple creations of the same data.
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
3 weeks ago[PDK-12358]Removing Non-supported Sciclient APIs REL.CORESDK.08.06.00.01 REL.CORESDK.08.06.00.02 REL.CORESDK.08.06.00.03
[PDK-12358]Removing Non-supported Sciclient APIs
Sciclient_pmDevicePowerOff() and Sciclient_pmGetWakeupReason()
APIs are not supported on TDA4V platform,removing from
Scilcient drive.
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
Sciclient_pmDevicePowerOff() and Sciclient_pmGetWakeupReason()
APIs are not supported on TDA4V platform,removing from
Scilcient drive.
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
Migrating to SYSFW version v08.06.00
drv: sciclient: Update Migration script for J7ES2.0
Update the sysfw_migration.sh script to enable J721E_2.0 ( GP & HS)
Upadte the firmwareHeader to inclue J721E 2.0 and HS-FS binaries.
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
Update the sysfw_migration.sh script to enable J721E_2.0 ( GP & HS)
Upadte the firmwareHeader to inclue J721E 2.0 and HS-FS binaries.
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
[PDK-12469][J784S4]sciserver: Enable local DRUs in sciserver
-In J784s4,local DRU's are removed from main_always_on domain.
-Therefore DRU's are enabled as a part of sciserver
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
-In J784s4,local DRU's are removed from main_always_on domain.
-Therefore DRU's are enabled as a part of sciserver
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
[PDK-12470][SAFERTOS][OSAL]: Any api call to QueueP for safertos goes to QueueP_freertos.c solved
Issue:
- QueueP_freertos.c included in safertos lib
Resolution:
- Renaming and moving QueueP_freertos.c to nonos
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
Issue:
- QueueP_freertos.c included in safertos lib
Resolution:
- Renaming and moving QueueP_freertos.c to nonos
Signed-off-by: Vaibhav Jindal <v-jindal@ti.com>
[SBL]: Fix booting mcu1_0 application.
- This error occured after SBL code cleanup.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- This error occured after SBL code cleanup.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
[SBL]: Cleanup the dead code.
- Cleanup the code related to am65xx, am64xx.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Cleanup the code related to am65xx, am64xx.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
8 weeks agosciclient: tools: Add new script to integrate updated BoardCfg RM REL.CORESDK.08.05.00.34 REL.CORESDK.08.05.00.35 REL.CORESDK.08.05.00.36 REL.CORESDK.08.05.00.37
sciclient: tools: Add new script to integrate updated BoardCfg RM
- Utility script to integrate updated BoardCfg RM to PDK.
Usage : boardcfg_update.sh <BOARD> --boardcfg=path/to/boardcfg/rm/c/file
Examples:-
./boardcfg_update.sh j721e_evm --boardcfg=path/to/j721e/sciclient/defaultBoardcfg/rm/c/file
./boardcfg_update.sh j7200_evm --boardcfg=path/to/j7200/sciclient/defaultBoardcfg/rm/c/file
./boardcfg_update.sh j721s2_evm --boardcfg=path/to/j721s2/sciclient/defaultBoardcfg/rm/c/file
./boardcfg_update.sh j784s4_evm --boardcfg=path/to/j784s4/sciclient/defaultBoardcfg/rm/c/file
Signed-off-by: Don Dominic <a0486429@ti.com>
- Utility script to integrate updated BoardCfg RM to PDK.
Usage : boardcfg_update.sh <BOARD> --boardcfg=path/to/boardcfg/rm/c/file
Examples:-
./boardcfg_update.sh j721e_evm --boardcfg=path/to/j721e/sciclient/defaultBoardcfg/rm/c/file
./boardcfg_update.sh j7200_evm --boardcfg=path/to/j7200/sciclient/defaultBoardcfg/rm/c/file
./boardcfg_update.sh j721s2_evm --boardcfg=path/to/j721s2/sciclient/defaultBoardcfg/rm/c/file
./boardcfg_update.sh j784s4_evm --boardcfg=path/to/j784s4/sciclient/defaultBoardcfg/rm/c/file
Signed-off-by: Don Dominic <a0486429@ti.com>
[J7200][LPM]: Remove Race condition in IO retention app
Remove race condition which results in dead-sleep condition
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
Remove race condition which results in dead-sleep condition
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
[SBL][HSM]: Add appimage in sbl/tools to to boot M4_1 core.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Add CPU load profiling support for OSPI Flash FreeRTOS TestApp
- Profile CPU load in main_ospi_flash_test.c if build for FreeRTOS.
- Remove TIRTOS specific profiling from application.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
- Profile CPU load in main_ospi_flash_test.c if build for FreeRTOS.
- Remove TIRTOS specific profiling from application.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
Disabling McASP build for MCU1_0
Signed-off-by: Nagpal <x1080849@ti.com>
Signed-off-by: Nagpal <x1080849@ti.com>
PDK-12411 Board: Updated CSI-Tx to CSI-Rx loopback test
- Enabled the CSI-Rx stream for the capture mode
Signed-off-by: Jayanth BR <x1080850@ti.com>
- Enabled the CSI-Rx stream for the capture mode
Signed-off-by: Jayanth BR <x1080850@ti.com>
PDK-12412 Board : Updated the board config header file with proper LIN
UART instance
Signed-off-by: Jayanth BR <x1080850@ti.com>
UART instance
Signed-off-by: Jayanth BR <x1080850@ti.com>
PDK-12413 Board: Updated board name in the pmic diagnostic test for
J784s4_evm
Signed-off-by: Jayanth BR <x1080850@ti.com>
J784s4_evm
Signed-off-by: Jayanth BR <x1080850@ti.com>
PDK-12409 : Board - Updated EEPROM diagnostic test with the proper Board
ID's mapping
Signed-off-by: Jayanth BR <x1080850@ti.com>
ID's mapping
Signed-off-by: Jayanth BR <x1080850@ti.com>
PDK-12414 Board : Updated diagnostic tests for the failure on MCU core
for J784s4_evm
- Enabled MAIN I2C instance for R5 core in current
Monitor and RTC tests.
- Updated proper UART instance for Boot Switch test
Signed-off-by: Jayanth BR <x1080850@ti.com>
for J784s4_evm
- Enabled MAIN I2C instance for R5 core in current
Monitor and RTC tests.
- Updated proper UART instance for Boot Switch test
Signed-off-by: Jayanth BR <x1080850@ti.com>
[SBL] : Add prebuilt linux images used to generate linux appimages.
- These images can be used to generate combined_opt.appimage,
combined_dev.appimage which are used to boot linux from SBL.
- These images can be used to atf_optee.appimage, tidtb_linux.appimage,
tikernalimage_linux.appimage which are used to boot linux from BootApp.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- These images can be used to generate combined_opt.appimage,
combined_dev.appimage which are used to boot linux from SBL.
- These images can be used to atf_optee.appimage, tidtb_linux.appimage,
tikernalimage_linux.appimage which are used to boot linux from BootApp.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
[PDK-12362]Updating the sciclient design doc
Updating the sciclient design doc to remove internal
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
Updating the sciclient design doc to remove internal
Signed-off-by: Kunal Lahoti <k-lahoti@ti.com>
Fixed MCASP packaging for J721S2
Signed-off-by: Rishabh Garg <rishabh@ti.com>
Signed-off-by: Rishabh Garg <rishabh@ti.com>
2 months agoJ721S2: C7x: Fix: UDMA UT crashes while calling HeapP_Create REL.CORESDK.08.05.00.23 REL.CORESDK.08.05.00.24 REL.CORESDK.08.05.00.25 REL.CORESDK.08.05.00.26 REL.CORESDK.08.05.00.27 REL.CORESDK.08.05.00.28 REL.CORESDK.08.05.00.29 REL.CORESDK.08.05.00.30 REL.CORESDK.08.05.00.31 REL.CORESDK.08.05.00.32
J721S2: C7x: Fix: UDMA UT crashes while calling HeapP_Create
On J721S2 we have only 4MB MSMC ranges 0x70000000 - 0x703FFFFF
but UDMA UT c7x linker MSMC origins at 0x70400000 which is out of
MSMC region making UT crash in MSMC testcases.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
On J721S2 we have only 4MB MSMC ranges 0x70000000 - 0x703FFFFF
but UDMA UT c7x linker MSMC origins at 0x70400000 which is out of
MSMC region making UT crash in MSMC testcases.
Signed-off-by: Lohith Kumar <l-kumar@ti.com>
PDK-11085: Removed SPI instance offsets in board library
- OSPI instance offsets (offset from SPI/MCSPI instances) are defined in board
library. This information is SoC specific and need to be retrieved from
the SPI driver. These offsets are no longer needed with dedicated SPI
driver APIs for each type of the SPI (SPI, McSPI & OSPI).
Removed the offset macros from the board flash library.
- OSPI instance offsets (offset from SPI/MCSPI instances) are defined in board
library. This information is SoC specific and need to be retrieved from
the SPI driver. These offsets are no longer needed with dedicated SPI
driver APIs for each type of the SPI (SPI, McSPI & OSPI).
Removed the offset macros from the board flash library.
[Docs][PDK-12361] Update APIs Not supported on TDA4V Platforms.
Sciclient_pmDevicePowerOff() and Sciclient_pmGetWakeupReason()
APIs are not supported on TDA4V platform.
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
Sciclient_pmDevicePowerOff() and Sciclient_pmGetWakeupReason()
APIs are not supported on TDA4V platform.
Signed-off-by: Chandru Dhavamani <chandru@ti.com>
[Board] Introduced timeout for I2C transaction in Board_getIDInfo_v2()
-In Board_getIDInfo_v2(), added timeout for the I2C transaction, so that
if the slave is inaccessible, the I2C transfer returns failure.
Signed-off-by: Jahnavi_Guvvala <j-guvvala@ti.com>
-In Board_getIDInfo_v2(), added timeout for the I2C transaction, so that
if the slave is inaccessible, the I2C transfer returns failure.
Signed-off-by: Jahnavi_Guvvala <j-guvvala@ti.com>
keywriter: Do not run SMEK encryption if input is omitted
In keywriter generation script, if --smek input is not entered, then
do not run the SMEK encryption.
Also fixed the comment for gen_sym_key_x509_extension.
Fixes: PDK-11240
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
In keywriter generation script, if --smek input is not entered, then
do not run the SMEK encryption.
Also fixed the comment for gen_sym_key_x509_extension.
Fixes: PDK-11240
Signed-off-by: Sheng Zhao <shengzhao@ti.com>
[PDK-12342]: Fix CUST sbl failure on J721E
- Remove NULL check since when tried to copy to 0x0 (ATCM address local to CPU) it is treating as
NULL and doesn't copy to that location.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
- Remove NULL check since when tried to copy to 0x0 (ATCM address local to CPU) it is treating as
NULL and doesn't copy to that location.
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
[PDK-12316][PMU Counter]: Removing DM Timer, and checking the PMU overflow in the Freertos OS tick timer instead.
- DM Timer should not be used to check on PMU overflow status.
- DMT(for PMU counter) may conflict with the OS Tick Timer or even other applications.
- Checking the PMU overflow status in the FreeRTOS tick timer instead.
- For SafeRTOS and Baremetal, application should take this responsibility.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
- DM Timer should not be used to check on PMU overflow status.
- DMT(for PMU counter) may conflict with the OS Tick Timer or even other applications.
- Checking the PMU overflow status in the FreeRTOS tick timer instead.
- For SafeRTOS and Baremetal, application should take this responsibility.
Signed-off-by: Mriganka Chakravarty <m-chakravarty@ti.com>
[SBL][eMMC]: Add separate rules to boot from uda and boot0 partition
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>
Signed-off-by: Sai Ramakurthi <s-ramakurthi@ti.com>