Re: removing support for kernel stack swapping

On Mon, Jun 03, 2024 at 10:41:22PM +0100, Jessica Clarke wrote:
> On 3 Jun 2024, at 22:39, Konstantin Belousov <[email protected]> wrote:
> > 
> > On Mon, Jun 03, 2024 at 10:15:15PM +0100, Jessica Clarke wrote:
> >> On 3 Jun 2024, at 22:11, Konstantin Belousov <[email protected]> wrote:
> >>> 
> >>> On Sun, Jun 02, 2024 at 07:57:04PM -0400, Mark Johnston wrote:
> >>>> FreeBSD will, when free pages are scarce, try to swap out the kernel
> >>>> stacks (typically 16KB per thread) of sleeping user threads.  I'm told
> >>>> that this mechanism was first implemented in BSD for the VAX port and
> >>>> that stabilizing it was quite an endeavour.
> >>>> 
> >>>> This feature has wide-ranging implications for code in the kernel.  For
> >>>> instance, if a thread allocates a structure on its stack, links it into
> >>>> some data structure visible to other threads, and goes to sleep, it must
> >>>> use PHOLD to ensure that the stack doesn't get swapped out while
> >>>> sleeping.  A missing PHOLD can thus result in a kernel panic, but this
> >>>> kind of mistake is very easy to make and hard to catch without thorough
> >>>> stress testing.  The kernel stack allocator also requires a fair bit of
> >>>> code to implement this feature, and we've had multiple bugs in that
> >>>> area, especially in relation to NUMA support.  Moreover, this feature
> >>>> will leave threads swapped out after the system has recovered, resulting
> >>>> in high scheduling latency once they're ready to run again.
> >>>> 
> >>>> In a very stressed system, it's possible that we can free up something
> >>>> like 1MB of RAM using this mechanism.  I argue that this mechanism is
> >>>> not worth it on modern systems: it isn't going to make the difference
> >>>> between a graceful recovery from memory pressure and a catatonic state
> >>>> which forces a reboot.  The complexity and resulting bugs it induces is
> >>>> not worth it.
> >>> On amd64, 1MB of physical memory for stacks is consumed by 64k threads,
> >> 
> >> To avoid any confusion, you mean 64 kthreads here, right? At least that
> >> makes sense for the story and the maths.
> > I mean 65535 threads (each of which must have kernel stack).
> 
> At 16 KiB each that would be 1 GiB total, not 1 MiB?
Indeed my math was incorrect, 1MB is consumed by 64 threads.
64k threads eat 1GB for stacks today.