- Timestamp:
-
08/09/23 13:22:18 (9 months ago)
- Author:
-
Utkarsh Verma
- Comment:
-
Add weekly update
Legend:
- Unmodified
- Added
- Removed
- Modified
-
v59
|
v60
|
|
208 | 208 | * Got the mailbox API to work in my bare-metal playground project. |
209 | 209 | * Next week: Port the mailbox API to RTEMS with framebuffer support. |
| 210 | * **August 2**: Coding Period !#10 |
| 211 | * Last week's work: |
| 212 | * Found some edge cases where the mailbox API was breaking. Refactored the code to better support the property channel interface while following the DRY philosophy. |
| 213 | * Figured out that unaligned memory accesses were breaking the API. This is because mailbox messages are unaligned by nature. |
| 214 | * Blocker: The compiler was optimizing two WORD accesses into a single DWORD operation. This was problematic because the messages are not guaranteed to be aligned by nature. Using -malign-strict was the solution. |
| 215 | * Next week: Add framebuffer support. |
| 216 | * **August 9**: Coding Period !#11 |
| 217 | * Last week's work: |
| 218 | * The mailbox API has been tested and works properly for all tested edge cases. However, the RPi is still not sending framebuffer data to the TV. |
| 219 | * The TV shows a black screen indicating that initialization was successful to some extent. |
| 220 | * Started preparing patches to be merged to master. Patches till UART support are done. |
| 221 | * Blocker: Framebuffer support is hard to debug. |
| 222 | * Next week: Pick up I2C and SPI since framebuffer progress is stuck. Prepare a patch for mailbox support. |