Altera DisplayPort MegaCore Function - Manual (Page 103)

The following figure shows a more detailed view of these operations. For a sink application, the user
application must initialize the DPCD content and the EDID. Additionally, for both source and sink
applications, an interrupt ISR must be registered.
Figure 8-2: Typical Source and Sink User Application Library Calls
btc_dprx_syslib_add_rx(...);
btc_dprx_syslib_init(...);
btc_dprx_dpcd_gpu_access(...);
btc_dprx_edid_set(...);
<Register the RX ISR>
BTC_DPRX_ENABLE_IRQ(...);
btc_dprx_syslib_monitor(...);
btc_dptx_syslib_init(...);
<Register the TX ISR>
BTC_DPTX_ENABLE_HPD_IRQ(...);
btc_dptx_syslib_monitor(...);
Source Sink
Sink instantiations issue an interrupt to the GPU when an AUX channel Request is received from the
connected source. Source instantiations issue an interrupt to the GPU when a logic state change is
detected on the HPD signal generated by the connected DisplayPort sink.
Because sources always act as AUX channel masters, they can manage AUX communication by initiating
a transaction (by sending a request) and then polling the IP core registers waiting to receive a reply.
Optionally, source instantiations can also issue an interrupt to the GPU when an AUX channel reply is
received from the connected DisplayPort sink, allowing the GPU to execute other tasks while waiting for
AUX channel replies.
Enable or disable source and sink interrupts with the following library macros:
•
BTC_DPTX_ENABLE_HPD_ IRQ()
•
BTC_DPTX_DISABLE_HPD_ IRQ()
•
BTC_DPTX_ENABLE_AUX_ IRQ()
•
BTC_DPTX_DISABLE_AUX_ IRQ()
•
BTC_DPRX_ENABLE_ IRQ()
•
BTC_DPRX_DISABLE_ IRQ()
btc_dprx_syslib
manages one to four sink instances by disabling all GPU interrupts when invoked and
restoring them to their previous state on exiting. Therefore, most of the library public functions
implement critical sections.
The GPU main program should minimize overhead when serving interrupts generated by sink instances
(i.e., interrupts related to a connected source’s AUX channel requests).
Interrupts generated by source instances (i.e., interrupts related to a connected sink’s HPD activity) can be
served with a lower priority. In designs where the same GPU handles both source and sink instances, the
GPU must allow for nested interrupts originated by sinks. That is, a sink must be allowed to interrupt a
source interrupt service routine (but not another sink interrupt service routine).
8-2 Using the Library
UG-01131
2015.05.04
Altera Corporation DisplayPort API Reference
Send Feedback