Re: Subject: Regression: CONFIG_ASYNC_KERNEL_PGTABLE_FREE causes memory exhaustion and stalls on busy Cascade Lake server (6.18.7 only)

[Robert Dinse <nanook@eskimo.com>]

[-- Attachment #1: Type: text/plain, Size: 2121 bytes --]


      I've tested 6.18.8 on both machines that were having issues on 6.18.7 and 
they are both running well, so don't know what changed but the issue seems to
be addressed in 6.18.8 so I would like to close this bug report.

-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-_-
  Eskimo North Linux Friendly Internet Access, Shell Accounts, and Hosting.
    Knowledgeable human assistance, not telephone trees or script readers.
  See our web site: http://www.eskimo.com/ (206) 812-0051 or (800) 246-6874.

On Sun, 1 Feb 2026, Christophe Leroy (CS GROUP) wrote:

> Date: Sun, 1 Feb 2026 19:16:43 +0100
> From: "Christophe Leroy (CS GROUP)" <chleroy@kernel.org>
> To: Robert Dinse <nanook@eskimo.com>, akpm@linux-foundation.org
> Cc: linux-mm@kvack.org
> Subject: Re: Subject: Regression: CONFIG_ASYNC_KERNEL_PGTABLE_FREE causes
>     memory exhaustion and stalls on busy Cascade Lake server (6.18.7 only)
> 
>
>
> Le 29/01/2026 à 07:10, Robert Dinse a écrit :
>> 
>> 
>> Reverting to 6.18.6 immediately restores the previous stable behavior.
>> 
>> Attempt to disable CONFIG_ASYNC_KERNEL_PGTABLE_FREE
>> ---------------------------------------------------
>> I attempted to disable the new async kernel page table freeing feature:
>> 
>> - The symbol `CONFIG_ASYNC_KERNEL_PGTABLE_FREE` appears in `.config`
>> - However, it does not appear in xconfig or other configuration frontends
>> - Manually editing `.config` to disable it works only until the next 
>> `make`:
>>    - As soon as I re-run the build, the option is silently re-enabled
>> - I tried to chase the Kconfig dependencies, but the chain was too 
>> convoluted; it appears to be effectively non-user-selectable and forced on 
>> by default for my architecture.
>>
>>  From an operator perspective, this feature as currently implemented is not 
>> workable on a busy machine like this, and the inability to disable it makes 
>> it difficult to bisect or run with a known-good configuration.
>> 
>
> Try reverting commit 9f0a7ab700f8 ("iommu/sva: invalidate stale IOTLB entries 
> for kernel address space") ?
>
> Christophe
>