eb7ab85fd8
The two interrupt controller registers on 818x got spoonerized, oops. Hilariously it actually worked (mostly) becuase the interrupt would still be delivered, the controller isn't involved in the ACK/clear process on thsi device, the and the set of pending interrupts was masked against enabled ones. Found it by accident when I moved things around and caught a "spurious" interrupt. Also the 8195 DRAM region was all packed in a single block that doesn't honor the two regions defined in the host devicetree. That's benign from Zephyr (the memory in the hole actually is usable) but dangerous for SOF as the second region is used by the host driver to allocate DMA buffers and we can't touch it except as directed. Honor the scheme used by the other devices. Signed-off-by: Andy Ross <andyross@google.com> |
||
|---|---|---|
| .. | ||
| mt8186 | ||
| mt8188 | ||
| mt8195 | ||
| mt8196 | ||
| index.rst | ||
| twister.yaml | ||