uhyve: don't panic on initial unmap if page is not mapped #1508
Conversation
| flush.ignore(); | ||
| match page_table.unmap(page) { | ||
| Ok((_frame, flush)) => flush.ignore(), | ||
| Err(UnmapError::PageNotMapped) => {} // If it wasn't mapped, that's not an issue |
There was a problem hiding this comment.
This is similar to what we do in paging::unmap. Would it make sense to log this too?
There was a problem hiding this comment.
I don't think so unless you like ~510 lines of info log output on a normal boot.
There was a problem hiding this comment.
We would not have to use info. But if you don't think a debug or trace is useful, that's fine for me.
There was a problem hiding this comment.
Honestly, I don't think it is worth it.
There was a problem hiding this comment.
Should we add a TODO into the comment or will this be tracked in an issue?
There was a problem hiding this comment.
I think it is sufficient not to panic for a while, and as soon as the Uhyve version that doesn't map the memory anymore (hermit-os/uhyve@700425d btw) is out for a while, we completely remove this part. We can track it in hermit-os/uhyve#426 or create a new issue.
I'm reworking the initial page mapping of uhyve to resolve hermit-os/uhyve#426, but the kernel currently expects the first GiB to be mapped and panics otherwise.
In a transition phase, it makes sense to at least not panic if the page is not mapped.