Hi Daniel,
Let me summarize the takeaway from the bi-weekly MM meeting as I understood
it, that includes Hugh's feedback on per-block tracking vs. mTHP:
Thanks David for the summary. Please, find below some follow up questions.
I want understand if zeropage scanning overhead is preferred over per-block
tracking complexity or if we still need to verify this.
(1) Per-block tracking
Per-block tracking is currently considered unwarranted complexity in
shmem.c. We should try to get it done without that. For any test cases that
fail, we should consider if they are actually valid for shmem.
I agree it was unwarranted complexity but only if this is just to fix lseek() as
we can simply make the test pass by checking if holes are reported as data. That
would be the minimum required for lseek() to be compliant with the syscall.
How can we use per-block tracking for reclaiming memory and what changes would
be needed? Or is per-block really a non-viable option?
The interesting thing is: with mTHP toggles it is opt-in -- like for
PMD-sized THP with shmem_enabled -- and we don't have to be that
concerned about this problem right now.
Clearly, if per-block is viable option, shmem_fault() bug would required to be
fixed first. Any ideas on how to make it reproducible?
The alternatives discussed where sub-page refcounting and zeropage scanning.
Yeah, I don't think sub-page refcounting is a feasible (and certainly
not desired ;) ) option in the folio world.
The first one is not possible (IIUC) because of a refactor years back that
simplified the code and also requires extra complexity. The second option would
require additional overhead as we would involve scanning.
We'll likely need something similar (scanning, tracking?) for anonymous
memory as well. There was a proposal for a THP shrinker some time ago,
that would solve part of the problem.
For example, for shmem you could simply flag folios that failed
splitting during fallocate() as reclaim candidates and try to reclaim
memory later. So you don't have to scan arbitrary folios (which might
also be desired, though).
To optimize FALLOC_FL_PUNCH_HOLE for the cases where splitting+freeing
is not possible at fallcoate() time, detecting zeropages later and
retrying to split+free might be an option, without per-block tracking.
(2) mTHP controls
As a default, we should not be using large folios / mTHP for any shmem, just
like we did with THP via shmem_enabled. This is what this series currently
does, and is aprt of the whole mTHP user-space interface design.
That was clear for me too. But what is the reason we want to boot in 'safe
mode'? What are the implications of not respecing that?
[...]
As I understood from the call, mTHP with sysctl knobs is preferred over
optimistic falloc/write allocation? But is still unclear to me why the former
is preferred.
I think Hugh's point was that this should be an opt-in feature, just
like PMD-sized THP started out, and still is, an opt-in feature.
Problematic interaction with khugepaged (that will be fixed) is one
thing, interaction with memory reclaim (without any kind of memory
reclaim mechanisms in place) might be another one. Controlling and
tuning for specific folio sizes might be another one Hugh raised.
[summarizing what I recall from the discussion, there might be more].
Is large folios a non-viable option?
I think you mean "allocating large folios without user space control".
Because mTHP gives user space full control, to the point where you can
enable all sizes and obtain the same result.
Also, we should properly fallback within the configured sizes, and not jump
"over" configured sizes. Unless there is a good reason.
(3) khugepaged
khugepaged needs to handle larger folios properly as well. Until fixed,
using smaller THP sizes as fallback might prohibit collapsing a PMD-sized
THP later. But really, khugepaged needs to be fixed to handle that.
(4) force/disable
These settings are rather testing artifacts from the old ages. We should not
add them to the per-size toggles. We might "inherit" it from the global one,
though.
"within_size" might have value, and especially for consistency, we should
have them per size.
So, this series only tackles anonymous shmem, which is a good starting
point. Ideally, we'd get support for other shmem (especially during fault
time) soon afterwards, because we won't be adding separate toggles for that
from the interface POV, and having inconsistent behavior between kernel
versions would be a bit unfortunate.
@Baolin, this series likely does not consider (4) yet. And I suggest we have
to take a lot of the "anonymous thp" terminology out of this series,
especially when it comes to documentation.
@Daniel, Pankaj, what are your plans regarding that? It would be great if we
could get an understanding on the next steps on !anon shmem.
I realize I've raised so many questions, but it's essential for us to grasp the
mm concerns and usage scenarios. This understanding will provide clarity on the
direction regarding folios for !anon shmem.
If I understood correctly, Hugh had strong feelings against not
respecting mTHP toggles for shmem. Without per-block tracking, I agree
with that.
--
Cheers,
David / dhildenb
