Re: [PATCH 4/5] x86/mm: Decouple kernel TLB flushes from flush_tlb_info

[Chuyi Zhou]

On 2026-09-23 6:50 p.m., Sebastian Andrzej Siewior wrote:
> On 2026-09-23 18:43:55 [+0800], Chuyi Zhou wrote:
>> The ceiling check does not cover all possible inputs, though. For
>> example, with start = ULONG_MAX - PAGE_SIZE + 1 and end = TLB_FLUSH_ALL,
> 
> Well, due to the number of pages when it is considered to do a full
> flush. But "ULONG_MAX - PAGE_SIZE + 1"  isn't a valid start input, is
> it?

Correct. On x86-64 that address falls in the unused hole at the top
of the address space, so it does not represent the start of an actual
kernel mapping.

My example only illustrated a difference for arbitrary input values;
it did not demonstrate a valid use case that needs the explicit check.
I'll drop it.

> 
> Sebastian