Re: [PATCH v2] mm/vmalloc: reject overflowing vm_map_ram page counts

From: Shubhang

Date: Mon Nov 09 2026 - 20:59:04 EST


Hi Andrew,

On Sun, 1 Nov 2026, Andrew Morton wrote:

Is there a case where userspace can actually hit this, or is it a
theoretical case which can't actually be encountered, or something else?


Reaching vm_map_ram() with an overflowing count from userspace is real
(as noted in the udmabuf path earlier), but actually hitting an oops or
corruption remains theoretical because in-tree callers don't access
beyond the truncated size.

The changelog described what happens if a caller does access the full
requested range. I can send a v3 clarifying that the failure mode is theoretical if you'd prefer.

Thanks,
Shubhang Kaushik