SAMD20: Some progress debugging SPI. Currently hangs

This commit is contained in:
Gregory Nutt 2014-02-20 15:20:42 -06:00
parent d3a6ad82b4
commit 7a83c06d62

View File

@ -618,8 +618,8 @@ Configurations
reconfiguration process.
2. Unless stated otherwise, all configurations generate console
output of on SERCOM4 which is available on EXT1 or EXT3 (see the
section "Serial Consoles" above). The virtual COM port could
output of on SERCOM4 which is available on EXT1, EXT2, or EXT3 (see
the section "Serial Consoles" above). The virtual COM port could
be used, instead, by reconfiguring to use SERCOM3 instead of
SERCOM4:
@ -678,8 +678,8 @@ Configurations
use to set or PATH variable. The path in the that file may not,
however, be correct for your installation.
See also the "NOTE about Windows native toolchains" in the section call
"GNU Toolchain Options" above.
See also the "NOTE about Windows native toolchains" in the section
called "GNU Toolchain Options" above.
Configuration sub-directories
-----------------------------
@ -883,8 +883,11 @@ Configuration sub-directories
be clock related???
- The program seems to be running normally, just producing bad output.
3. The configuration suggests CONFIG_MMCSD_HAVECARDDETECT=y, but as of
3. SPI current hangs so no much progress has been made tested the I/O1
module.
The configuration suggests CONFIG_MMCSD_HAVECARDDETECT=y, but as of
this writing, there is no support for EIC pin interrupts.
4. OLED1 module is untested. These intructions were just ported from
4. OLED1 module is untested. These instructions were just lifted from
the SAM4L Xplained Pro README.txt file.