Incorrect behaviour with INT_MAX pmpaddr #739
Labels
bug
Something isn't working
Comments
Closed
Timmmm
added a commit
to Timmmm/sail-riscv
that referenced
this issue
Mar 7, 2025
This fixes a bug if you set `pmpaddr` to `ones()`. For NAPOT it would return `(zeros(), zeros())`. For NA4 it would return `(ones(), zeros())`. In both cases it will match no addresses, whereas it should match all memory, or the top 4 bytes respectively. A test is included for NAPOT that fails before and passes afterwards. Fixes riscv#739
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
If you set
pmpaddrtoones()then it will misbehave. For NAPOT it will return(zeros(), zeros()). For NA4 it will return(ones(), zeros()). In both cases it will match no addresses, whereas it should match all memory, or the top 4 bytes respectively.I have a fix for this, and a test (I know!). The test is written in Zig... I was experimenting with that. I'll demo it in the next meeting I'm at (24th).
The text was updated successfully, but these errors were encountered: