jackcartersmith/blackmagic
1
0
Fork 0
Code Issues Pull Requests Releases Activity

Modified SSP driver to wait for data to be sent before returning. This is critical when controlling device CS# pins via GPIO. Long-term, it might be better to have a different API that permits this level of control.

Browse Source
This commit is contained in:
Jared Boone 2012-06-15 17:20:50 -07:00
parent e28d07b81e
commit 536c922577
1 changed files with 26 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

26
lib/lpc43xx/ssp.c
View File

@ -113,6 +113,21 @@ u16 ssp_read(ssp_num_t ssp_num)
return SSP_DR(ssp_port); return SSP_DR(ssp_port);
} }
void ssp_wait_until_not_busy(ssp_num_t ssp_num)
{
u32 ssp_port;
if(ssp_num == SSP0_NUM)
{
ssp_port = SSP0;
}else
{
ssp_port = SSP1;
}
while( (SSP_SR(ssp_port) & SSP_SR_BSY) );
}
/* This Function Wait Data TX Ready, and Write Data to SSP */ /* This Function Wait Data TX Ready, and Write Data to SSP */
void ssp_write(ssp_num_t ssp_num, u16 data) void ssp_write(ssp_num_t ssp_num, u16 data)
{ {
@ -130,5 +145,16 @@ void ssp_write(ssp_num_t ssp_num, u16 data)
while( (SSP_SR(ssp_port) & SSP_SR_TNF) == 0); while( (SSP_SR(ssp_port) & SSP_SR_TNF) == 0);
SSP_DR(ssp_port) = data; SSP_DR(ssp_port) = data;
/* Wait for not busy, since we're controlling CS# of
* devices manually and need to wait for the data to
* be sent. It may also be important to wait here
* in case we're configuring devices via SPI and also
* with GPIO control -- we need to know when SPI
* commands are effective before altering a device's
* state with GPIO. I'm thinking the MAX2837, for
* example...
*/
ssp_wait_until_not_busy(ssp_num);
} }