On 16.12.2025 01:11, Nicola Vetrini wrote:
> The rule headline states: "A typedef name shall be a unique identifier".
> 
> in x86/mm/shadow/multi.c the typedef names for `guest_va_t` and `guest_pa_t`
> are deliberately reused multiple times due to the file being included with
> different GUEST_PAGING_LEVELS. As a result, a deviation for the rule is
> introduced.
> 
> No functional change.
> 
> Signed-off-by: Nicola Vetrini <[email protected]>

Looks plausible, yet I wonder if we wouldn't better do away with guest_pa_t.
It has one single use (the per-CPU trace_emulate_write_val), and judging
from the involved logic as well as other naming (e.g. the .write_val.l1 field
used in sh_trace_emulate()) it's really guest_l1e_t that's meant there.

Jan

Reply via email to