summaryrefslogtreecommitdiff
path: root/boards/qemu-virt
AgeCommit message (Collapse)Author
2024-09-15Start of serious allocator work.Nathan Ringo
2024-09-14Start of a platform-independent paging interface.Nathan Ringo
2024-09-04Enables paging.Nathan Ringo
2024-09-02Split hart0_boot in half, so stacks can be switched in the middle.Nathan Ringo
2024-09-01Move everything over to the new organization.Nathan Ringo
2024-08-27[boards/qemu-virt] Adds a canary for stack overflow.Nathan Ringo
2024-08-25[boards/qemu-virt] Set up timer interrupts.Nathan Ringo
2024-08-25Sets up PMP to allow access to all of physical memory.Nathan Ringo
2024-08-03uncommitted changes...Nathan Ringo
2024-02-25[boards/qemu-virt] Adds a strict flush routine.Nathan Ringo
This writes to the UART, ignoring any notion of readiness. This will probably need to be reworked once a real UART driver exists that is used by this platform.
2024-02-25[console] Adds a "strict flush" routine.Nathan Ringo
When this routine is set by the platform-specific code, it is run after every log message. This is a permanently temporary measure to make it easier to debug boot problems with the kernel when bringing it up on a new platform.
2024-02-25[console] Adds the start of a console subsystem to handle kernel logs.Nathan Ringo
The idea is that this is an output-only facility that log messages from the log crate are written to. It has a small buffer, and relies on a log destination regularly reading from the buffer. In the next commit, platform-specific code will be able to provide a "strict flush" routine, which runs after every write to the buffer and drains the buffer. This is mainly to help kernel development, where it's not assured that the kernel will even get as far in boot as initializing its allocator. Eventually, there will be a console fd that userspace can request, wait on, and read from.
2024-02-24Initial commitNathan Ringo