| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Message-ID: <20251107224956.477056-1-gourry@gourry.net>
Date: Fri, 7 Nov 2025 17:49:45 -0500
From: Gregory Price <gourry@...rry.net>
To: linux-mm@...ck.org
Cc: linux-cxl@...r.kernel.org,
linux-kernel@...r.kernel.org,
nvdimm@...ts.linux.dev,
linux-fsdevel@...r.kernel.org,
cgroups@...r.kernel.org,
dave@...olabs.net,
jonathan.cameron@...wei.com,
dave.jiang@...el.com,
alison.schofield@...el.com,
vishal.l.verma@...el.com,
ira.weiny@...el.com,
dan.j.williams@...el.com,
longman@...hat.com,
akpm@...ux-foundation.org,
david@...hat.com,
lorenzo.stoakes@...cle.com,
Liam.Howlett@...cle.com,
vbabka@...e.cz,
rppt@...nel.org,
surenb@...gle.com,
mhocko@...e.com,
osalvador@...e.de,
ziy@...dia.com,
matthew.brost@...el.com,
joshua.hahnjy@...il.com,
rakie.kim@...com,
byungchul@...com,
gourry@...rry.net,
ying.huang@...ux.alibaba.com,
apopple@...dia.com,
mingo@...hat.com,
peterz@...radead.org,
juri.lelli@...hat.com,
vincent.guittot@...aro.org,
dietmar.eggemann@....com,
rostedt@...dmis.org,
bsegall@...gle.com,
mgorman@...e.de,
vschneid@...hat.com,
tj@...nel.org,
hannes@...xchg.org,
mkoutny@...e.com,
kees@...nel.org,
muchun.song@...ux.dev,
roman.gushchin@...ux.dev,
shakeel.butt@...ux.dev,
rientjes@...gle.com,
jackmanb@...gle.com,
cl@...two.org,
harry.yoo@...cle.com,
axelrasmussen@...gle.com,
yuanchu@...gle.com,
weixugc@...gle.com,
zhengqi.arch@...edance.com,
yosry.ahmed@...ux.dev,
nphamcs@...il.com,
chengming.zhou@...ux.dev,
fabio.m.de.francesco@...ux.intel.com,
rrichter@....com,
ming.li@...omail.com,
usamaarif642@...il.com,
brauner@...nel.org,
oleg@...hat.com,
namcao@...utronix.de,
escape@...ux.alibaba.com,
dongjoo.seo1@...sung.com
Subject: [RFC LPC2026 PATCH 0/9] Protected Memory NUMA Nodes
Author Note
-----------
This is a code RFC for discussion related to
"Mempolicy is dead, long live memory policy!"
https://lpc.events/event/19/contributions/2143/
Given the subtlety of some of these changes, and the upcoming holidays
I wanted to publish this well ahead of time for discussion. This is
the baseline patch set which predicates a new kind of mempolicy based
on NUMA node memory features - which can be defined by the components
adding memory to such NUMA nodes.
Included is an example of a Compressed Memory Node, and how compressed
RAM could be managed by zswap. Compressed memory is its own rabbit
hole - I recommend not getting hung up on the example.
The core discussion should be around whether such a "Protected Node"
based system is reasonable - and whether there are sufficient potential
users to warrant support.
Also please do not get hung up on naming. "Protected" just means
"Not-System-RAM". If you see "Default" just assume "Systam RAM".
base-commit: 1c353dc8d962de652bc7ad2ba2e63f553331391c
-----------
With this set, we aim to enable allocation of "special purpose memory"
with the page allocator (mm/page_alloc.c) without exposing the same
memory as "Typical System RAM". Unless a non-userland component
explicitly asks for the node, and does so with a GFP_PROTECTED flag,
memory on that node cannot be "accidentally" used as normal ram.
We present an example of using this mechanism within ZSWAP, as-if
a "compressed memory node" was present. How to describe the features
of memory present on nodes is left up to comment here and at LPC '26.
Important Note: Since userspace interfaces are restricted by the
default_node mask (sysram), nothing in userspace can explicitly
request memory from protected nodes. Instead, the intent is to
create new components which understand different node features,
which abstracts the hardware complexity away from userland.
The ZSWAP example demonstrates this with `mt_compressed_nodemask`
which is simply a hack to simply demonstrate the idea.
There are 4 major changes in this set:
1) Introducing default_sysram_nodes in mm/memory-tiers.c which denotes
the set of default nodes which are eligible for use as normal sysram
Some existing users noew pass default_sysram_nodes into the page
allocator instead of NULL, but passing a NULL pointer in will simply
have it replaced by default_sysram_nodes anyway.
default_sysram_nodes is always guaranteed to contain the N_MEMORY
nodes that were present at boot time, and so it can never be empty.
2) The addition of `cpuset.mems.default` which restricts cgroups to
using `default_sysram_nodes` by default, while allowing non-sysram
nodes into mems_effective (mems_allowed).
This is done to allow separate control over sysram and protected node
sets by cgroups while maintaining the hierarchical rules.
current cpuset configuration
cpuset.mems_allowed
|.mems_effective < (mems_allowed ∩ parent.mems_effective)
|->tasks.mems_allowed < cpuset.mems_effective
new cpuset configuration
cpuset.mems_allowed
|.mems_effective < (mems_allowed ∩ parent.mems_effective)
|.mems_default < (mems_effective ∩ default_sys_nodemask)
|->task.mems_default < cpuset.mems_default - (note renamed)
3) Addition of MHP_PROTECTED_MEMORY flag to denote to memory-hotplug
that the memory capacity being added should mark the node as a
protected memory node. A node is either SysRAM or Protected, and
cannot contain both (adding protected to an existing SysRAM node
will result in EINVAL).
DAX and CXL are made aware of the bit and have `protected_memory`
bits added to their relevant subsystems.
4) Adding GFP_PROTECTED - which allows page_alloc.c to request memory
from the provided node or nodemask. It changes the behavior of
the cpuset mems_allowed check.
Probably there needs to be some additional work done here to
restrict non-cgroup kernels.
Gregory Price (9):
gfp: Add GFP_PROTECTED for protected-node allocations
memory-tiers: create default_sysram_nodes
mm: default slub, oom_kill, compaction, and page_alloc to sysram
mm,cpusets: rename task->mems_allowed to task->mems_default
cpuset: introduce cpuset.mems.default
mm/memory_hotplug: add MHP_PROTECTED_MEMORY flag
drivers/dax: add protected memory bit to dev_dax
drivers/cxl: add protected_memory bit to cxl region
[HACK] mm/zswap: compressed ram integration example
drivers/cxl/core/region.c | 30 ++++++
drivers/cxl/cxl.h | 2 +
drivers/dax/bus.c | 39 ++++++++
drivers/dax/bus.h | 1 +
drivers/dax/cxl.c | 1 +
drivers/dax/dax-private.h | 1 +
drivers/dax/kmem.c | 2 +
fs/proc/array.c | 2 +-
include/linux/cpuset.h | 52 +++++------
include/linux/gfp_types.h | 3 +
include/linux/memory-tiers.h | 4 +
include/linux/memory_hotplug.h | 10 ++
include/linux/mempolicy.h | 2 +-
include/linux/sched.h | 6 +-
init/init_task.c | 2 +-
kernel/cgroup/cpuset-internal.h | 8 ++
kernel/cgroup/cpuset-v1.c | 7 ++
kernel/cgroup/cpuset.c | 157 +++++++++++++++++++++-----------
kernel/fork.c | 2 +-
kernel/sched/fair.c | 4 +-
mm/hugetlb.c | 8 +-
mm/memcontrol.c | 2 +-
mm/memory-tiers.c | 25 ++++-
mm/memory_hotplug.c | 25 +++++
mm/mempolicy.c | 34 +++----
mm/migrate.c | 4 +-
mm/oom_kill.c | 11 ++-
mm/page_alloc.c | 28 +++---
mm/show_mem.c | 2 +-
mm/slub.c | 4 +-
mm/vmscan.c | 2 +-
mm/zswap.c | 65 ++++++++++++-
32 files changed, 411 insertions(+), 134 deletions(-)
--
2.51.1
Powered by blists - more mailing lists