| Age | Commit message (Collapse) | Author |
|---|
|
fmc_read_32() and fmc_write_32() get called often enough that it's
worth eliminating unnecessary function call overhead.
|
|
We're no longer using a preemptive multitasker, and we're on the
Alpha, not the old bridge board, so all we really need to do is read
or write one word at a time via a pointer then wait for the GPIO idle
pin to light. Well, other than needing to read twice because of a
known hardware problem with the STM32.
So we dont need to fiddle with IRQ, or use the CMSIS SRAM lock, or....
|
|
Note: This affects libhal/ks_token.c, which uses the keystore driver directly.
|
|
Clean up Makefiles and initialization code.
|
|
|
|
|
|
Integrated into the cli-test program as such:
cryptech> test sdram
Initializing SDRAM
Starting SDRAM test (n = 0)
Run sequential write-then-read test for the first chip
Run random write-then-read test for the first chip
Run sequential write-then-read test for the second chip
Run random write-then-read test for the second chip
Run interleaved write-then-read test for both chips at once
SDRAM test (n = 0) completed
SDRAM test completed successfully
cryptech>
|
|
|
|
|
|
Clean up and simplify(?) Makefile.
Add copyrights as needed.
Add include guard to stm-fmc.h.
Move MX_USART2_UART_Init back to stm-init.c for possible copyright reasons.
Move libc, src, and include files to top level.
|
