lpc_handler: add note to copyFrom stub
Add a note explaining the copyFrom() design. A different implementation
was writing to the staged file directly from a memory-mapped region.
This was done by accessing devmem directly after everything was
configured. Using devmem will no longer be available, therefore the
different systems will need drivers that allow accessing that data.
Change-Id: I7fae4c23d88c5fe2075efafc365a1fcc8d7bac74
Signed-off-by: Patrick Venture <venture@google.com>
diff --git a/lpc_handler.cpp b/lpc_handler.cpp
index 0aebb3f..71cf24d 100644
--- a/lpc_handler.cpp
+++ b/lpc_handler.cpp
@@ -23,7 +23,14 @@
std::vector<std::uint8_t> LpcDataHandler::copyFrom(std::uint32_t length)
{
- /* TODO: implement this */
+ /* TODO: implement this -- in an earlier and different version of this that
+ * didn't use BLOBs, the region was memory-mapped and the writes to the data
+ * were just done directly from the memory-mapped region instead of a
+ * copyFrom() first call. The idea with this change is that we may not be
+ * able to get a memory-mapped handle from the driver from which to
+ * automatically read data, but rather must perform some ioctl or other
+ * access to get the data from the driver.
+ */
return {};
}