2019-12-12 17:13:12

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

This series is based on latest linux-next. The patches are located at:
https://github.com/davidhildenbrand/linux.git virtio-mem-rfc-v4

The basic idea of virtio-mem is to provide a flexible,
cross-architecture memory hot(un)plug solution that avoids many limitations
imposed by existing technologies, architectures, and interfaces. More
details can be found below and in linked material.

This RFC is limited to x86-64, however, should theoretically work on any
architecture that supports virtio and implements memory hot(un)plug under
Linux - like s390x, powerpc64 and arm64. On x86-64, it is currently
possible to add/remove memory to the system in >= 4MB granularity.
Memory hotplug works very reliably. For memory unplug, there are no
guarantees how much memory can actually get unplugged, it depends on the
setup (especially: fragmentation of (unmovable) memory). I have plans to
improve that in the future.

--------------------------------------------------------------------------
1. virtio-mem
--------------------------------------------------------------------------

The basic idea behind virtio-mem was presented at KVM Forum 2018. The
slides can be found at [1]. The previous RFC can be found at [2]. The
first RFC can be found at [3]. However, the concept evolved over time. The
KVM Forum slides roughly match the current design.

Patch #2 ("virtio-mem: Paravirtualized memory hotplug") contains quite some
information, especially in "include/uapi/linux/virtio_mem.h":

Each virtio-mem device manages a dedicated region in physical address
space. Each device can belong to a single NUMA node, multiple devices
for a single NUMA node are possible. A virtio-mem device is like a
"resizable DIMM" consisting of small memory blocks that can be plugged
or unplugged. The device driver is responsible for (un)plugging memory
blocks on demand.

Virtio-mem devices can only operate on their assigned memory region in
order to (un)plug memory. A device cannot (un)plug memory belonging to
other devices.

The "region_size" corresponds to the maximum amount of memory that can
be provided by a device. The "size" corresponds to the amount of memory
that is currently plugged. "requested_size" corresponds to a request
from the device to the device driver to (un)plug blocks. The
device driver should try to (un)plug blocks in order to reach the
"requested_size". It is impossible to plug more memory than requested.

The "usable_region_size" represents the memory region that can actually
be used to (un)plug memory. It is always at least as big as the
"requested_size" and will grow dynamically. It will only shrink when
explicitly triggered (VIRTIO_MEM_REQ_UNPLUG).

Memory in the usable region can usually be read, however, there are no
guarantees. It can happen that the device cannot process a request,
because it is busy. The device driver has to retry later.

Usually, during system resets all memory will get unplugged, so the
device driver can start with a clean state. However, in specific
scenarios (if the device is busy) it can happen that the device still
has memory plugged. The device driver can request to unplug all memory
(VIRTIO_MEM_REQ_UNPLUG) - which might take a while to succeed if the
device is busy.

--------------------------------------------------------------------------
2. Linux Implementation
--------------------------------------------------------------------------

This RFC reuses quite some existing MM infrastructure, however, has to
expose some additional functionality.

Memory blocks (e.g., 128MB) are added/removed on demand. Within these
memory blocks, subblocks (e.g., 4MB) are plugged/unplugged. The sizes
depend on the target architecture, MAX_ORDER + pageblock_order, and
the block size of a virtio-mem device.

add_memory()/try_remove_memory() is used to add/remove memory blocks.
virtio-mem will not online memory blocks itself. This has to be done by
user space, or configured into the kernel
(CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE). virtio-mem will only unplug memory
that was online to the ZONE_NORMAL. Memory is suggested to be onlined to
the ZONE_NORMAL for now.

The memory hotplug notifier is used to properly synchronize against
onlining/offlining of memory blocks and to track the states of memory
blocks (including the zone memory blocks are onlined to).

The set_online_page() callback is used to keep unplugged subblocks
of a memory block fake-offline when onlining the memory block.
generic_online_page() is used to fake-online plugged subblocks. This
handling is similar to the Hyper-V balloon driver.

PG_offline is used to mark unplugged subblocks as offline, so e.g.,
dumping tools (makedumpfile) will skip these pages. This is similar to
other balloon drivers like virtio-balloon and Hyper-V.

Memory offlining code is extended to allow drivers to drop their reference
to PG_offline pages when MEM_GOING_OFFLINE, so these pages can be skipped
when offlining memory blocks. This allows to offline memory blocks that
have partially unplugged (allocated e.g., via alloc_contig_range())
subblocks - or are completely unplugged.

alloc_contig_range()/free_contig_range() [now exposed] is used to
unplug/plug subblocks of memory blocks the are already exposed to Linux.

offline_and_remove_memory() [new] is used to offline a fully unplugged
memory block and remove it from Linux.


A lot of additional information can be found in the separate patches and
as comments in the code itself.

--------------------------------------------------------------------------
3. Changes RFC v2 -> v3
--------------------------------------------------------------------------

A lot of things changed, especially also on the QEMU + virtio side. The
biggest changes on the Linux driver side are:
- Onlining/offlining of subblocks is now emulated on top of memory blocks.
set_online_page()+alloc_contig_range()+free_contig_range() is now used
for that. Core MM does not have to be modified and will continue to
online/offline full memory blocks.
- Onlining/offlining of memory blocks is no longer performed by virtio-mem.
- Pg_offline is upstream and can be used. It is also used to allow
offlining of partially unplugged memory blocks.
- Memory block states + subblocks are now tracked more space-efficient.
- Proper kexec(), kdump(), driver unload, driver reload, ZONE_MOVABLE, ...
handling.

--------------------------------------------------------------------------
4. Changes RFC v3 -> RFC v4
--------------------------------------------------------------------------

Only minor things changed, especially nothing on the QEMU + virtio side.
Interresting changes on the Linux driver side are:
- "mm: Allow to offline unmovable PageOffline() pages via
MEM_GOING_OFFLINE"
-- Rework to Michals suggestion (allow to isolate all PageOffline() pages
by skipping all PageOffline() pages in has_unmovable_pages(). Fail
offlining later if the pages cannot be offlined/migrated).
- "virtio-mem: Allow to offline partially unplugged memory blocks"
-- Adapt to Michals suggestion on core-mm part.
- "virtio-mem: Better retry handling"
-- Optimize retry intervals
- "virtio-mem: Drop slab objects when unplug continues to fail"
-- Call drop_slab()/drop_slab_node() when unplug keeps failing for a longer
time.
- Multiple cleanups and fixes.

--------------------------------------------------------------------------
5. Future work
--------------------------------------------------------------------------

The separate patches contain a lot of future work items. One of the next
steps is to make memory unplug more likely to succeed - currently, there
are no guarantees on how much memory can get unplugged again. I have
various ideas on how to limit fragmentation of all memory blocks that
virtio-mem added.

Memory hotplug:
- Reduce the amount of memory resources if that turnes out to be an
issue. Or try to speed up relevant code paths to deal with many
resources.
- Allocate the vmemmap from the added memory. Makes hotplug more likely
to succeed, the vmemmap is stored on the same NUMA node and that
unmovable memory will later not hinder unplug.

Memory hotunplug:
- Performance improvements:
-- Sense (lockless) if it make sense to try alloc_contig_range() at all
before directly trying to isolate and taking locks.
-- Try to unplug bigger chunks if possible first.
-- Identify free areas first, that don't have to be evacuated.
- Make unplug more likely to succeed:
-- There are various idea to limit fragmentation on memory block
granularity. (e.g., ZONE_PREFER_MOVABLE and smart balancing)
-- Allocate memmap from added memory. This way, less unmovable data can
end up on the memory blocks.
- OOM handling, e.g., via an OOM handler.
- Defragmentation
-- Will require a new virtio-mem CMD to exchange plugged<->unplugged blocks

--------------------------------------------------------------------------
6. Example Usage
--------------------------------------------------------------------------

A very basic QEMU prototype (kept updated) is available at:
https://github.com/davidhildenbrand/qemu.git virtio-mem

It lacks various features, however, works to test the guest driver side:
- No support for resizable memory regions / memory backends yet
- No protection of unplugged memory (esp., userfaultfd-wp) yet
- No dump/migration/XXX optimizations to skip unplugged memory (and avoid
touching it)

Start QEMU with two virtio-mem devices (one per NUMA node):
$ qemu-system-x86_64 -m 4G,maxmem=20G \
-smp sockets=2,cores=2 \
-numa node,nodeid=0,cpus=0-1 -numa node,nodeid=1,cpus=2-3 \
[...]
-object memory-backend-ram,id=mem0,size=8G \
-device virtio-mem-pci,id=vm0,memdev=mem0,node=0,requested-size=128M \
-object memory-backend-ram,id=mem1,size=8G \
-device virtio-mem-pci,id=vm1,memdev=mem1,node=1,requested-size=80M

Query the configuration:
QEMU 4.1.95 monitor - type 'help' for more information
(qemu) info memory-devices
Memory device [virtio-mem]: "vm0"
memaddr: 0x140000000
node: 0
requested-size: 134217728
size: 134217728
max-size: 8589934592
block-size: 2097152
memdev: /objects/mem0
Memory device [virtio-mem]: "vm1"
memaddr: 0x340000000
node: 1
requested-size: 83886080
size: 83886080
max-size: 8589934592
block-size: 2097152
memdev: /objects/mem1

Add some memory to node 1:
QEMU 4.1.95 monitor - type 'help' for more information
(qemu) qom-set vm1 requested-size 1G

Remove some memory from node 0:
QEMU 4.1.95 monitor - type 'help' for more information
(qemu) qom-set vm0 requested-size 64M

Query the configuration again:
(qemu) info memory-devices
Memory device [virtio-mem]: "vm0"
memaddr: 0x140000000
node: 0
requested-size: 67108864
size: 67108864
max-size: 8589934592
block-size: 2097152
memdev: /objects/mem0
Memory device [virtio-mem]: "vm1"
memaddr: 0x340000000
node: 1
requested-size: 1073741824
size: 1073741824
max-size: 8589934592
block-size: 2097152
memdev: /objects/mem1

--------------------------------------------------------------------------
7. Q/A
--------------------------------------------------------------------------

Q: Why add/remove parts ("subblocks") of memory blocks/sections?
A: Flexibility (section size depends on the architecture) - e.g., some
architectures have a section size of 2GB. Also, the memory block size
is variable (e.g., on x86-64). I want to avoid any such restrictions.
Some use cases want to add/remove memory in smaller granularities to a
VM (e.g., the Hyper-V balloon also implements this) - especially smaller
VMs like used for kata containers. Also, on memory unplug, it is more
reliable to free-up and unplug multiple small chunks instead
of one big chunk. E.g., if one page of a DIMM is either unmovable or
pinned, the DIMM can't get unplugged. This approach is basically a
compromise between DIMM-based memory hot(un)plug and balloon
inflation/deflation, which works mostly on page granularity.

Q: Why care about memory blocks?
A: They are the way to tell user space about new memory. This way,
memory can get onlined/offlined by user space. Also, e.g., kdump
relies on udev events to reload kexec when memory blocks are
onlined/offlined. Memory blocks are the "real" memory hot(un)plug
granularity. Everything that's smaller has to be emulated "on top".

Q: Won't memory unplug of subblocks fragment memory?
A: Yes and no. Unplugging e.g., >=4MB subblocks on x86-64 will not really
fragment memory like unplugging random pages like a balloon driver does.
Buddy merging will not be limited. However, any allocation that requires
bigger consecutive memory chunks (e.g., gigantic pages) might observe
the fragmentation. Possible solutions: Allocate gigantic huge pages
before unplugging memory, don't unplug memory, combine virtio-mem with
DIMM based memory or bigger initial memory. Remember, a virtio-mem
device will only unplug on the memory range it manages, not on other
DIMMs. Unplug of single memory blocks will result in similar
fragmentation in respect to gigantic huge pages. I ahve plans for a
virtio-mem defragmentation feature in the future.

Q: How reliable is memory unplug?
A: There are no guarantees on how much memory can get unplugged
again. However, it is more likely to find 4MB chunks to unplug than
e.g., 128MB chunks. If memory is terribly fragmented, there is nothing
we can do - for now. I consider memory hotplug the first primary use
of virtio-mem. Memory unplug might usually work, but we want to improve
the performance and the amount of memory we can actually unplug later.

Q: Why not unplug from the ZONE_MOVABLE?
A: Unplugged memory chunks are unmovable. Unmovable data must not end up
on the ZONE_MOVABLE - similar to gigantic pages - they will never be
allocated from ZONE_MOVABLE. virtio-mem added memory can be onlined
to the ZONE_MOVABLE, but subblocks will not get unplugged from it.

Q: How big should the initial (!virtio-mem) memory of a VM be?
A: virtio-mem memory will not go to the DMA zones. So to avoid running out
of DMA memory, I suggest something like 2-3GB on x86-64. But many
VMs can most probably deal with less DMA memory - depends on the use
case.

[1] https://events.linuxfoundation.org/wp-content/uploads/2017/12/virtio-mem-Paravirtualized-Memory-David-Hildenbrand-Red-Hat-1.pdf
[2] https://lkml.kernel.org/r/[email protected]
[3] https://lkml.kernel.org/r/[email protected]

Cc: Sebastien Boeuf <[email protected]>
Cc: Samuel Ortiz <[email protected]>
Cc: Robert Bradford <[email protected]>
Cc: Luiz Capitulino <[email protected]>

David Hildenbrand (13):
ACPI: NUMA: export pxm_to_node
virtio-mem: Paravirtualized memory hotplug
virtio-mem: Paravirtualized memory hotunplug part 1
mm: Export alloc_contig_range() / free_contig_range()
virtio-mem: Paravirtualized memory hotunplug part 2
mm: Allow to offline unmovable PageOffline() pages via
MEM_GOING_OFFLINE
virtio-mem: Allow to offline partially unplugged memory blocks
mm/memory_hotplug: Introduce offline_and_remove_memory()
virtio-mem: Offline and remove completely unplugged memory blocks
virtio-mem: Better retry handling
mm/vmscan: Move count_vm_event(DROP_SLAB) into drop_slab()
mm/vmscan: Export drop_slab() and drop_slab_node()
virtio-mem: Drop slab objects when unplug continues to fail

drivers/acpi/numa/srat.c | 1 +
drivers/virtio/Kconfig | 18 +
drivers/virtio/Makefile | 1 +
drivers/virtio/virtio_mem.c | 1939 +++++++++++++++++++++++++++++++
fs/drop_caches.c | 4 +-
include/linux/memory_hotplug.h | 1 +
include/linux/mm.h | 4 +-
include/linux/page-flags.h | 10 +
include/uapi/linux/virtio_ids.h | 1 +
include/uapi/linux/virtio_mem.h | 204 ++++
mm/memory_hotplug.c | 76 +-
mm/page_alloc.c | 26 +
mm/page_isolation.c | 9 +
mm/vmscan.c | 3 +
14 files changed, 2282 insertions(+), 15 deletions(-)
create mode 100644 drivers/virtio/virtio_mem.c
create mode 100644 include/uapi/linux/virtio_mem.h

--
2.23.0


2019-12-12 17:13:41

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 03/13] virtio-mem: Paravirtualized memory hotunplug part 1

Unplugging subblocks of memory blocks that are offline is easy. All we
have to do is watch out for concurrent onlining acticity.

Cc: "Michael S. Tsirkin" <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Dave Young <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Stefan Hajnoczi <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
drivers/virtio/virtio_mem.c | 109 +++++++++++++++++++++++++++++++++++-
1 file changed, 107 insertions(+), 2 deletions(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 8f4ceeab3d7c..f1af05def5df 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -119,7 +119,7 @@ struct virtio_mem {
*
* When this lock is held the pointers can't change, ONLINE and
* OFFLINE blocks can't change the state and no subblocks will get
- * plugged.
+ * plugged/unplugged.
*/
struct mutex hotplug_mutex;
bool hotplug_active;
@@ -321,6 +321,19 @@ static bool virtio_mem_mb_test_sb_plugged(struct virtio_mem *vm,
bit + count;
}

+/*
+ * Test if all selected subblocks are unplugged.
+ */
+static bool virtio_mem_mb_test_sb_unplugged(struct virtio_mem *vm,
+ unsigned long mb_id, int sb_id,
+ int count)
+{
+ const int bit = (mb_id - vm->first_mb_id) * vm->nb_sb_per_mb + sb_id;
+
+ /* TODO: Helper similar to bitmap_set() */
+ return find_next_bit(vm->sb_bitmap, bit + count, bit) >= bit + count;
+}
+
/*
* Find the first plugged subblock. Returns vm->nb_sb_per_mb in case there is
* none.
@@ -511,6 +524,9 @@ static void virtio_mem_notify_offline(struct virtio_mem *vm,
BUG();
break;
}
+
+ /* trigger the workqueue, maybe we can now unplug memory. */
+ virtio_mem_retry(vm);
}

static void virtio_mem_notify_online(struct virtio_mem *vm, unsigned long mb_id,
@@ -1123,6 +1139,93 @@ static int virtio_mem_plug_request(struct virtio_mem *vm, uint64_t diff)
return rc;
}

+/*
+ * Unplug the desired number of plugged subblocks of an offline memory block.
+ * Will fail if any subblock cannot get unplugged (instead of skipping it).
+ *
+ * Will modify the state of the memory block. Might temporarily drop the
+ * hotplug_mutex.
+ *
+ * Note: Can fail after some subblocks were successfully unplugged.
+ */
+static int virtio_mem_mb_unplug_any_sb_offline(struct virtio_mem *vm,
+ unsigned long mb_id,
+ uint64_t *nb_sb)
+{
+ int rc;
+
+ rc = virtio_mem_mb_unplug_any_sb(vm, mb_id, nb_sb);
+
+ /* some subblocks might have been unplugged even on failure */
+ if (!virtio_mem_mb_test_sb_plugged(vm, mb_id, 0, vm->nb_sb_per_mb))
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL);
+ if (rc)
+ return rc;
+
+ if (virtio_mem_mb_test_sb_unplugged(vm, mb_id, 0, vm->nb_sb_per_mb)) {
+ /*
+ * Remove the block from Linux - this should never fail.
+ * Hinder the block from getting onlined by marking it
+ * unplugged. Temporarily drop the mutex, so
+ * any pending GOING_ONLINE requests can be serviced/rejected.
+ */
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_UNUSED);
+
+ mutex_unlock(&vm->hotplug_mutex);
+ rc = virtio_mem_mb_remove(vm, mb_id);
+ BUG_ON(rc);
+ mutex_lock(&vm->hotplug_mutex);
+ }
+ return 0;
+}
+
+/*
+ * Try to unplug the requested amount of memory.
+ */
+static int virtio_mem_unplug_request(struct virtio_mem *vm, uint64_t diff)
+{
+ uint64_t nb_sb = diff / vm->subblock_size;
+ unsigned long mb_id;
+ int rc;
+
+ if (!nb_sb)
+ return 0;
+
+ /*
+ * We'll drop the mutex a couple of times when it is safe to do so.
+ * This might result in some blocks switching the state (online/offline)
+ * and we could miss them in this run - we will retry again later.
+ */
+ mutex_lock(&vm->hotplug_mutex);
+
+ /* Try to unplug subblocks of partially plugged offline blocks. */
+ virtio_mem_for_each_mb_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL) {
+ rc = virtio_mem_mb_unplug_any_sb_offline(vm, mb_id,
+ &nb_sb);
+ if (rc || !nb_sb)
+ goto out_unlock;
+ cond_resched();
+ }
+
+ /* Try to unplug subblocks of plugged offline blocks. */
+ virtio_mem_for_each_mb_state(vm, mb_id, VIRTIO_MEM_MB_STATE_OFFLINE) {
+ rc = virtio_mem_mb_unplug_any_sb_offline(vm, mb_id,
+ &nb_sb);
+ if (rc || !nb_sb)
+ goto out_unlock;
+ cond_resched();
+ }
+
+ mutex_unlock(&vm->hotplug_mutex);
+ return 0;
+out_unlock:
+ mutex_unlock(&vm->hotplug_mutex);
+ return rc;
+}
+
/*
* Try to unplug all blocks that couldn't be unplugged before, for example,
* because the hypervisor was busy.
@@ -1203,8 +1306,10 @@ static void virtio_mem_run_wq(struct work_struct *work)
if (vm->requested_size > vm->plugged_size) {
diff = vm->requested_size - vm->plugged_size;
rc = virtio_mem_plug_request(vm, diff);
+ } else {
+ diff = vm->plugged_size - vm->requested_size;
+ rc = virtio_mem_unplug_request(vm, diff);
}
- /* TODO: try to unplug memory */
}

switch (rc) {
--
2.23.0

2019-12-12 17:13:43

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 02/13] virtio-mem: Paravirtualized memory hotplug

Each virtio-mem device owns exactly one memory region. It is responsible
for adding/removing memory from that memory region on request.

When the device driver starts up, the requested amount of memory is
queried and then plugged to Linux. On request, further memory can be
plugged or unplugged. This patch only implements the plugging part.

On x86-64, memory can currently be plugged in 4MB ("subblock") granularity.
When required, a new memory block will be added (e.g., usually 128MB on
x86-64) in order to plug more subblocks. Only x86-64 was tested for now.

The online_page callback is used to keep unplugged subblocks offline
when onlining memory - similar to the Hyper-V balloon driver. Unplugged
pages are marked PG_offline, to tell dump tools (e.g., makedumpfile) to
skip them.

User space is usually responsible for onlining the added memory. The
memory hotplug notifier is used to synchronize virtio-mem activity
against memory onlining/offlining.

Each virtio-mem device can belong to a NUMA node, which allows us to
easily add/remove small chunks of memory to/from a specific NUMA node by
using multiple virtio-mem devices. Something that works even when the
guest has no idea about the NUMA topology.

One way to view virtio-mem is as a "resizable DIMM" or a DIMM with many
"sub-DIMMS".

This patch directly introduces the basic infrastructure to implement memory
unplug. Especially the memory block states and subblock bitmaps will be
heavily used there.

Notes:
- Memory blocks that are partally unplugged must not be onlined to the
MOVABLE_ZONE. They contain unmovable parts and might e.g., result
in warnings when trying to offline them.
- In the kdump kernel, we don't want to load the driver, to not mess
with memory to be dumped.
- In case memory is to be onlined by user space, we limit the amount of
offline memory blocks, to not run out of memory.
- Suspend/Hibernate is not supported due to the way virtio-mem devices
behave. Limited support might be possible in the future.
- Reloading the device driver is not supported.
- When unloading the driver, we have to remove partially plugged offline
memory blocks. Otherwise they could get fully onlined later on, when
we no longer have control over via the online_page callback.
- As the hypervisor might suddenly be busy (during reboot, in-between
requests, when adding of memory fails), we have to take care of some
corner cases - especially virtio_mem_unplug_pending_mb() and
virtio_mem_send_unplug_all_request(). The hypervisor could for example
be busy if it is currently migrating the guest.

Future work:
- Reduce the amount of memory resources if that turnes out to be an
issue. Or try to speed up relevant code paths to deal with many
resources.
- Allocate the vmemmap from the added memory. Makes hotplug more likely
to succeed, the vmemmap is stored on the same NUMA node and that
unmovable memory will later not hinder unplug.

Cc: "Michael S. Tsirkin" <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Dave Young <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Stefan Hajnoczi <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: "Rafael J. Wysocki" <[email protected]>
Cc: Len Brown <[email protected]>
Cc: [email protected]
Signed-off-by: David Hildenbrand <[email protected]>
---
drivers/virtio/Kconfig | 17 +
drivers/virtio/Makefile | 1 +
drivers/virtio/virtio_mem.c | 1569 +++++++++++++++++++++++++++++++
include/uapi/linux/virtio_ids.h | 1 +
include/uapi/linux/virtio_mem.h | 204 ++++
5 files changed, 1792 insertions(+)
create mode 100644 drivers/virtio/virtio_mem.c
create mode 100644 include/uapi/linux/virtio_mem.h

diff --git a/drivers/virtio/Kconfig b/drivers/virtio/Kconfig
index 078615cf2afc..294720d53057 100644
--- a/drivers/virtio/Kconfig
+++ b/drivers/virtio/Kconfig
@@ -64,6 +64,23 @@ config VIRTIO_BALLOON

If unsure, say M.

+config VIRTIO_MEM
+ tristate "Virtio mem driver"
+ default m
+ depends on X86_64
+ depends on VIRTIO
+ depends on MEMORY_HOTPLUG_SPARSE
+ depends on MEMORY_HOTREMOVE
+ help
+ This driver provides access to virtio-mem paravirtualized memory
+ devices, allowing to hotplug and hotunplug memory.
+
+ This driver is was only tested under x86-64, but should
+ theoretically work on all architectures that support memory
+ hotplug and hotremove.
+
+ If unsure, say M.
+
config VIRTIO_INPUT
tristate "Virtio input driver"
depends on VIRTIO
diff --git a/drivers/virtio/Makefile b/drivers/virtio/Makefile
index 3a2b5c5dcf46..906d5a00ac85 100644
--- a/drivers/virtio/Makefile
+++ b/drivers/virtio/Makefile
@@ -6,3 +6,4 @@ virtio_pci-y := virtio_pci_modern.o virtio_pci_common.o
virtio_pci-$(CONFIG_VIRTIO_PCI_LEGACY) += virtio_pci_legacy.o
obj-$(CONFIG_VIRTIO_BALLOON) += virtio_balloon.o
obj-$(CONFIG_VIRTIO_INPUT) += virtio_input.o
+obj-$(CONFIG_VIRTIO_MEM) += virtio_mem.o
diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
new file mode 100644
index 000000000000..8f4ceeab3d7c
--- /dev/null
+++ b/drivers/virtio/virtio_mem.c
@@ -0,0 +1,1569 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Virtio-mem device driver.
+ *
+ * Copyright Red Hat, Inc. 2019
+ *
+ * Author(s): David Hildenbrand <[email protected]>
+ */
+
+#include <linux/virtio.h>
+#include <linux/virtio_mem.h>
+#include <linux/workqueue.h>
+#include <linux/module.h>
+#include <linux/mm.h>
+#include <linux/memory_hotplug.h>
+#include <linux/memory.h>
+#include <linux/hrtimer.h>
+#include <linux/crash_dump.h>
+#include <linux/mutex.h>
+#include <linux/bitmap.h>
+
+#include <acpi/acpi_numa.h>
+
+enum virtio_mem_mb_state {
+ /* Unplugged, not added to Linux. Can be reused later. */
+ VIRTIO_MEM_MB_STATE_UNUSED = 0,
+ /* (Partially) plugged, not added to Linux. Error on add_memory(). */
+ VIRTIO_MEM_MB_STATE_PLUGGED,
+ /* Fully plugged, fully added to Linux, offline. */
+ VIRTIO_MEM_MB_STATE_OFFLINE,
+ /* Paritally plugged, fully added to Linux, offline. */
+ VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL,
+ /* Fully plugged, fully added to Linux, online (!ZONE_MOVABLE). */
+ VIRTIO_MEM_MB_STATE_ONLINE,
+ /* Partially plugged, fully added to Linux, online (!ZONE_MOVABLE). */
+ VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL,
+ /*
+ * Fully plugged, fully added to Linux, online (ZONE_MOVABLE).
+ * We are not allowed to allocate (unplug) parts of this block that
+ * are not movable (similar to gigantic pages). We will never allow
+ * to online OFFLINE_PARTIAL to ZONE_MOVABLE (as they would contain
+ * unmovable parts).
+ */
+ VIRTIO_MEM_MB_STATE_ONLINE_MOVABLE,
+ VIRTIO_MEM_MB_STATE_COUNT
+};
+
+struct virtio_mem {
+ struct virtio_device *vdev;
+
+ /* We might first have to unplug all memory when starting up. */
+ bool unplug_all_required;
+
+ /* Workqueue that processes the plug/unplug requests. */
+ struct work_struct wq;
+ atomic_t config_changed;
+
+ /* Virtqueue for guest->host requests. */
+ struct virtqueue *vq;
+
+ /* Wait for a host response to a guest request. */
+ wait_queue_head_t host_resp;
+
+ /* Space for one guest request and the host response. */
+ struct virtio_mem_req req;
+ struct virtio_mem_resp resp;
+
+ /* The current size of the device. */
+ uint64_t plugged_size;
+ /* The requested size of the device. */
+ uint64_t requested_size;
+
+ /* The device block size (for communicating with the device). */
+ uint32_t device_block_size;
+ /* The translated node id. NUMA_NO_NODE in case not specified. */
+ int nid;
+ /* Physical start address of the memory region. */
+ uint64_t addr;
+
+ /* The subblock size. */
+ uint32_t subblock_size;
+ /* The number of subblocks per memory block. */
+ uint32_t nb_sb_per_mb;
+
+ /* Id of the first memory block of this device. */
+ unsigned long first_mb_id;
+ /* Id of the last memory block of this device. */
+ unsigned long last_mb_id;
+ /* Id of the last usable memory block of this device. */
+ unsigned long last_usable_mb_id;
+ /* Id of the next memory bock to prepare when needed. */
+ unsigned long next_mb_id;
+
+ /* Summary of all memory block states. */
+ unsigned long nb_mb_state[VIRTIO_MEM_MB_STATE_COUNT];
+#define VIRTIO_MEM_NB_OFFLINE_THRESHOLD 10
+
+ /*
+ * One byte state per memory block.
+ *
+ * Allocated via vmalloc(). When preparing new blocks, resized
+ * (alloc+copy+free) when needed (crossing pages with the next mb).
+ * (when crossing pages).
+ *
+ * With 128MB memory blocks, we have states for 512GB of memory in one
+ * page.
+ */
+ uint8_t *mb_state;
+
+ /*
+ * $nb_sb_per_mb bit per memory block. Handled similar to mb_state.
+ *
+ * With 4MB subblocks, we manage 128GB of memory in one page.
+ */
+ unsigned long *sb_bitmap;
+
+ /*
+ * Mutex that protects the nb_mb_state, mb_state, and sb_bitmap.
+ *
+ * When this lock is held the pointers can't change, ONLINE and
+ * OFFLINE blocks can't change the state and no subblocks will get
+ * plugged.
+ */
+ struct mutex hotplug_mutex;
+ bool hotplug_active;
+
+ /* An error ocurred we cannot handle - stop processing requests. */
+ bool broken;
+
+ /* The driver is being removed. */
+ spinlock_t removal_lock;
+ bool removing;
+
+ /* Timer for retrying to plug/unplug memory. */
+ struct hrtimer retry_timer;
+#define VIRTIO_MEM_RETRY_TIMER_MS 30000
+
+ /* Memory notifier (online/offline events). */
+ struct notifier_block memory_notifier;
+
+ /* Next device in the list of virtio-mem devices. */
+ struct list_head next;
+};
+
+/*
+ * We have to share a single online_page callback among all virtio-mem
+ * devices. We use RCU to iterate the list in the callback.
+ */
+static DEFINE_MUTEX(virtio_mem_mutex);
+static LIST_HEAD(virtio_mem_devices);
+
+static void virtio_mem_online_page_cb(struct page *page, unsigned int order);
+
+/*
+ * Register a virtio-mem device so it will be considered for the online_page
+ * callback.
+ */
+static int register_virtio_mem_device(struct virtio_mem *vm)
+{
+ int rc = 0;
+
+ /* First device registers the callback. */
+ mutex_lock(&virtio_mem_mutex);
+ if (list_empty(&virtio_mem_devices))
+ rc = set_online_page_callback(&virtio_mem_online_page_cb);
+ if (!rc)
+ list_add_rcu(&vm->next, &virtio_mem_devices);
+ mutex_unlock(&virtio_mem_mutex);
+
+ return rc;
+}
+
+/*
+ * Unregister a virtio-mem device so it will no longer be considered for the
+ * online_page callback.
+ */
+static void unregister_virtio_mem_device(struct virtio_mem *vm)
+{
+ /* Last device unregisters the callback. */
+ mutex_lock(&virtio_mem_mutex);
+ list_del_rcu(&vm->next);
+ if (list_empty(&virtio_mem_devices))
+ restore_online_page_callback(&virtio_mem_online_page_cb);
+ mutex_unlock(&virtio_mem_mutex);
+
+ synchronize_rcu();
+}
+
+/*
+ * Calculate the memory block id of a given address.
+ */
+static unsigned long virtio_mem_phys_to_mb_id(unsigned long addr)
+{
+ return addr / memory_block_size_bytes();
+}
+
+/*
+ * Calculate the physical start address of a given memory block id.
+ */
+static unsigned long virtio_mem_mb_id_to_phys(unsigned long mb_id)
+{
+ return mb_id * memory_block_size_bytes();
+}
+
+/*
+ * Calculate the subblock id of a given address.
+ */
+static unsigned long virtio_mem_phys_to_sb_id(struct virtio_mem *vm,
+ unsigned long addr)
+{
+ const unsigned long mb_id = virtio_mem_phys_to_mb_id(addr);
+ const unsigned long mb_addr = virtio_mem_mb_id_to_phys(mb_id);
+
+ return (addr - mb_addr) / vm->subblock_size;
+}
+
+/*
+ * Set the state of a memory block, taking care of the state counter.
+ */
+static void virtio_mem_mb_set_state(struct virtio_mem *vm, unsigned long mb_id,
+ enum virtio_mem_mb_state state)
+{
+ const unsigned long idx = mb_id - vm->first_mb_id;
+ enum virtio_mem_mb_state old_state;
+
+ old_state = vm->mb_state[idx];
+ vm->mb_state[idx] = state;
+
+ BUG_ON(vm->nb_mb_state[old_state] == 0);
+ vm->nb_mb_state[old_state]--;
+ vm->nb_mb_state[state]++;
+}
+
+/*
+ * Get the state of a memory block.
+ */
+static enum virtio_mem_mb_state virtio_mem_mb_get_state(struct virtio_mem *vm,
+ unsigned long mb_id)
+{
+ const unsigned long idx = mb_id - vm->first_mb_id;
+
+ return vm->mb_state[idx];
+}
+
+/*
+ * Prepare the state array for the next memory block.
+ */
+static int virtio_mem_mb_state_prepare_next_mb(struct virtio_mem *vm)
+{
+ unsigned long old_bytes = vm->next_mb_id - vm->first_mb_id + 1;
+ unsigned long new_bytes = vm->next_mb_id - vm->first_mb_id + 2;
+ int old_pages = PFN_UP(old_bytes);
+ int new_pages = PFN_UP(new_bytes);
+ uint8_t *new_mb_state;
+
+ if (vm->mb_state && old_pages == new_pages)
+ return 0;
+
+ new_mb_state = vzalloc(new_pages * PAGE_SIZE);
+ if (!new_mb_state)
+ return -ENOMEM;
+
+ mutex_lock(&vm->hotplug_mutex);
+ if (vm->mb_state)
+ memcpy(new_mb_state, vm->mb_state, old_pages * PAGE_SIZE);
+ vfree(vm->mb_state);
+ vm->mb_state = new_mb_state;
+ mutex_unlock(&vm->hotplug_mutex);
+
+ return 0;
+}
+
+#define virtio_mem_for_each_mb_state(_vm, _mb_id, _state) \
+ for (_mb_id = _vm->first_mb_id; \
+ _mb_id < _vm->next_mb_id && _vm->nb_mb_state[_state]; \
+ _mb_id++) \
+ if (virtio_mem_mb_get_state(_vm, _mb_id) == _state)
+
+/*
+ * Mark all selected subblocks plugged.
+ *
+ * Will not modify the state of the memory block.
+ */
+static void virtio_mem_mb_set_sb_plugged(struct virtio_mem *vm,
+ unsigned long mb_id, int sb_id,
+ int count)
+{
+ const int bit = (mb_id - vm->first_mb_id) * vm->nb_sb_per_mb + sb_id;
+
+ __bitmap_set(vm->sb_bitmap, bit, count);
+}
+
+/*
+ * Mark all selected subblocks unplugged.
+ *
+ * Will not modify the state of the memory block.
+ */
+static void virtio_mem_mb_set_sb_unplugged(struct virtio_mem *vm,
+ unsigned long mb_id, int sb_id,
+ int count)
+{
+ const int bit = (mb_id - vm->first_mb_id) * vm->nb_sb_per_mb + sb_id;
+
+ __bitmap_clear(vm->sb_bitmap, bit, count);
+}
+
+/*
+ * Test if all selected subblocks are plugged.
+ */
+static bool virtio_mem_mb_test_sb_plugged(struct virtio_mem *vm,
+ unsigned long mb_id, int sb_id,
+ int count)
+{
+ const int bit = (mb_id - vm->first_mb_id) * vm->nb_sb_per_mb + sb_id;
+
+ if (count == 1)
+ return test_bit(bit, vm->sb_bitmap);
+
+ /* TODO: Helper similar to bitmap_set() */
+ return find_next_zero_bit(vm->sb_bitmap, bit + count, bit) >=
+ bit + count;
+}
+
+/*
+ * Find the first plugged subblock. Returns vm->nb_sb_per_mb in case there is
+ * none.
+ */
+static int virtio_mem_mb_first_plugged_sb(struct virtio_mem *vm,
+ unsigned long mb_id)
+{
+ const int bit = (mb_id - vm->first_mb_id) * vm->nb_sb_per_mb;
+
+ return find_next_bit(vm->sb_bitmap, bit + vm->nb_sb_per_mb, bit) - bit;
+}
+
+/*
+ * Find the first unplugged subblock. Returns vm->nb_sb_per_mb in case there is
+ * none.
+ */
+static int virtio_mem_mb_first_unplugged_sb(struct virtio_mem *vm,
+ unsigned long mb_id)
+{
+ const int bit = (mb_id - vm->first_mb_id) * vm->nb_sb_per_mb;
+
+ return find_next_zero_bit(vm->sb_bitmap, bit + vm->nb_sb_per_mb, bit) -
+ bit;
+}
+
+/*
+ * Prepare the subblock bitmap for the next memory block.
+ */
+static int virtio_mem_sb_bitmap_prepare_next_mb(struct virtio_mem *vm)
+{
+ const unsigned long old_nb_mb = vm->next_mb_id - vm->first_mb_id;
+ const unsigned long old_nb_bits = old_nb_mb * vm->nb_sb_per_mb;
+ const unsigned long new_nb_bits = (old_nb_mb + 1) * vm->nb_sb_per_mb;
+ int old_pages = PFN_UP(BITS_TO_LONGS(old_nb_bits) * sizeof(long));
+ int new_pages = PFN_UP(BITS_TO_LONGS(new_nb_bits) * sizeof(long));
+ unsigned long *new_sb_bitmap, *old_sb_bitmap;
+
+ if (vm->sb_bitmap && old_pages == new_pages)
+ return 0;
+
+ new_sb_bitmap = vzalloc(new_pages * PAGE_SIZE);
+ if (!new_sb_bitmap)
+ return -ENOMEM;
+
+ mutex_lock(&vm->hotplug_mutex);
+ if (new_sb_bitmap)
+ memcpy(new_sb_bitmap, vm->sb_bitmap, old_pages * PAGE_SIZE);
+
+ old_sb_bitmap = vm->sb_bitmap;
+ vm->sb_bitmap = new_sb_bitmap;
+ mutex_unlock(&vm->hotplug_mutex);
+
+ vfree(old_sb_bitmap);
+ return 0;
+}
+
+/*
+ * Try to add a memory block to Linux. This will usually only fail
+ * if out of memory.
+ *
+ * Must not be called with the vm->hotplug_mutex held (possible deadlock with
+ * onlining code).
+ *
+ * Will not modify the state of the memory block.
+ */
+static int virtio_mem_mb_add(struct virtio_mem *vm, unsigned long mb_id)
+{
+ const uint64_t addr = virtio_mem_mb_id_to_phys(mb_id);
+ int nid = vm->nid;
+
+ if (nid == NUMA_NO_NODE)
+ nid = memory_add_physaddr_to_nid(addr);
+
+ dev_dbg(&vm->vdev->dev, "adding memory block: %lu\n", mb_id);
+ return add_memory(nid, addr, memory_block_size_bytes());
+}
+
+/*
+ * Try to remove a memory block from Linux. Will only fail if the memory block
+ * is not offline.
+ *
+ * Must not be called with the vm->hotplug_mutex held (possible deadlock with
+ * onlining code).
+ *
+ * Will not modify the state of the memory block.
+ */
+static int virtio_mem_mb_remove(struct virtio_mem *vm, unsigned long mb_id)
+{
+ const uint64_t addr = virtio_mem_mb_id_to_phys(mb_id);
+ int nid = vm->nid;
+
+ if (nid == NUMA_NO_NODE)
+ nid = memory_add_physaddr_to_nid(addr);
+
+ dev_dbg(&vm->vdev->dev, "removing memory block: %lu\n", mb_id);
+ return remove_memory(nid, addr, memory_block_size_bytes());
+}
+
+/*
+ * Trigger the workqueue so the device can perform its magic.
+ */
+static void virtio_mem_retry(struct virtio_mem *vm)
+{
+ unsigned long flags;
+
+ spin_lock_irqsave(&vm->removal_lock, flags);
+ if (!vm->removing)
+ queue_work(system_freezable_wq, &vm->wq);
+ spin_unlock_irqrestore(&vm->removal_lock, flags);
+}
+
+static int virtio_mem_translate_node_id(struct virtio_mem *vm, uint16_t node_id)
+{
+ int node = NUMA_NO_NODE;
+
+ if (IS_ENABLED(CONFIG_NUMA) && IS_ENABLED(CONFIG_ACPI_NUMA) &&
+ virtio_has_feature(vm->vdev, VIRTIO_MEM_F_ACPI_PXM))
+ node = pxm_to_node(node_id);
+ return node;
+}
+
+/*
+ * Test if a virtio-mem device overlaps with the given range. Can be called
+ * from (notifier) callbacks lockless.
+ */
+static bool virtio_mem_overlaps_range(struct virtio_mem *vm,
+ unsigned long start, unsigned long size)
+{
+ unsigned long dev_start = virtio_mem_mb_id_to_phys(vm->first_mb_id);
+ unsigned long dev_end = virtio_mem_mb_id_to_phys(vm->last_mb_id) +
+ memory_block_size_bytes();
+ unsigned long end = start + size;
+
+ return (start >= dev_start && start < dev_end) ||
+ (end > dev_start && end <= dev_end) ||
+ (start < dev_start && end > dev_end);
+}
+
+/*
+ * Test if a virtio-mem device owns a memory block. Can be called from
+ * (notifier) callbacks lockless.
+ */
+static bool virtio_mem_owned_mb(struct virtio_mem *vm, unsigned long mb_id)
+{
+ return mb_id >= vm->first_mb_id && mb_id <= vm->last_mb_id;
+}
+
+static int virtio_mem_notify_going_online(struct virtio_mem *vm,
+ unsigned long mb_id,
+ enum zone_type zone)
+{
+ switch (virtio_mem_mb_get_state(vm, mb_id)) {
+ case VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL:
+ /*
+ * We won't allow to online a partially plugged memory block
+ * to the MOVABLE zone - it would contain unmovable parts.
+ */
+ if (zone == ZONE_MOVABLE) {
+ dev_warn_ratelimited(&vm->vdev->dev,
+ "memory block has holes, MOVABLE not supported\n");
+ return NOTIFY_BAD;
+ }
+ return NOTIFY_OK;
+ case VIRTIO_MEM_MB_STATE_OFFLINE:
+ return NOTIFY_OK;
+ default:
+ break;
+ }
+ dev_warn_ratelimited(&vm->vdev->dev,
+ "memory block onlining denied\n");
+ return NOTIFY_BAD;
+}
+
+static void virtio_mem_notify_offline(struct virtio_mem *vm,
+ unsigned long mb_id)
+{
+ switch (virtio_mem_mb_get_state(vm, mb_id)) {
+ case VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL:
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL);
+ break;
+ case VIRTIO_MEM_MB_STATE_ONLINE:
+ case VIRTIO_MEM_MB_STATE_ONLINE_MOVABLE:
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE);
+ break;
+ default:
+ BUG();
+ break;
+ }
+}
+
+static void virtio_mem_notify_online(struct virtio_mem *vm, unsigned long mb_id,
+ enum zone_type zone)
+{
+ unsigned long nb_offline;
+
+ switch (virtio_mem_mb_get_state(vm, mb_id)) {
+ case VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL:
+ BUG_ON(zone == ZONE_MOVABLE);
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL);
+ break;
+ case VIRTIO_MEM_MB_STATE_OFFLINE:
+ if (zone == ZONE_MOVABLE)
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_ONLINE_MOVABLE);
+ else
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_ONLINE);
+ break;
+ default:
+ BUG();
+ break;
+ }
+ nb_offline = vm->nb_mb_state[VIRTIO_MEM_MB_STATE_OFFLINE] +
+ vm->nb_mb_state[VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL];
+
+ /* see if we can add new blocks now that we onlined one block */
+ if (nb_offline == VIRTIO_MEM_NB_OFFLINE_THRESHOLD - 1)
+ virtio_mem_retry(vm);
+}
+
+/*
+ * This callback will either be called synchonously from add_memory() or
+ * asynchronously (e.g., triggered via user space). We have to be careful
+ * with locking when calling add_memory().
+ */
+static int virtio_mem_memory_notifier_cb(struct notifier_block *nb,
+ unsigned long action, void *arg)
+{
+ struct virtio_mem *vm = container_of(nb, struct virtio_mem,
+ memory_notifier);
+ struct memory_notify *mhp = arg;
+ const unsigned long start = PFN_PHYS(mhp->start_pfn);
+ const unsigned long size = PFN_PHYS(mhp->nr_pages);
+ const unsigned long mb_id = virtio_mem_phys_to_mb_id(start);
+ enum zone_type zone;
+ int rc = NOTIFY_OK;
+
+ /*
+ * Memory is usually onlined/offlined in memory block granularity
+ * (via device_online()/device_offline()). This means, we
+ * cannot cross virtio-mem device boundaries and memory block
+ * boundaries.
+ *
+ * Only obscure users (powernv/memtrace.c) play dirty games
+ * manually with offline_pages() with multiple memory blocks -
+ * let's NACK these requests in case virtio-mem is involved.
+ */
+ if (size != memory_block_size_bytes() ||
+ !IS_ALIGNED(start, memory_block_size_bytes())) {
+ if (virtio_mem_overlaps_range(vm, start, size))
+ return NOTIFY_BAD;
+ return NOTIFY_DONE;
+ }
+
+ if (!virtio_mem_owned_mb(vm, mb_id))
+ return NOTIFY_DONE;
+
+ switch (action) {
+ case MEM_GOING_OFFLINE:
+ spin_lock_irq(&vm->removal_lock);
+ if (vm->removing)
+ rc = notifier_to_errno(-EBUSY);
+ spin_unlock_irq(&vm->removal_lock);
+ if (rc != NOTIFY_OK)
+ break;
+ mutex_lock(&vm->hotplug_mutex);
+ vm->hotplug_active = true;
+ break;
+ case MEM_GOING_ONLINE:
+ spin_lock_irq(&vm->removal_lock);
+ if (vm->removing)
+ rc = notifier_to_errno(-EBUSY);
+ spin_unlock_irq(&vm->removal_lock);
+ if (rc != NOTIFY_OK)
+ break;
+ mutex_lock(&vm->hotplug_mutex);
+ vm->hotplug_active = true;
+ zone = page_zonenum(pfn_to_page(mhp->start_pfn));
+ rc = virtio_mem_notify_going_online(vm, mb_id, zone);
+ break;
+ case MEM_OFFLINE:
+ virtio_mem_notify_offline(vm, mb_id);
+ vm->hotplug_active = false;
+ mutex_unlock(&vm->hotplug_mutex);
+ break;
+ case MEM_ONLINE:
+ zone = page_zonenum(pfn_to_page(mhp->start_pfn));
+ virtio_mem_notify_online(vm, mb_id, zone);
+ vm->hotplug_active = false;
+ mutex_unlock(&vm->hotplug_mutex);
+ break;
+ case MEM_CANCEL_OFFLINE:
+ case MEM_CANCEL_ONLINE:
+ if (!vm->hotplug_active)
+ break;
+ vm->hotplug_active = false;
+ mutex_unlock(&vm->hotplug_mutex);
+ break;
+ default:
+ break;
+ }
+
+ return rc;
+}
+
+/*
+ * Set a range of pages PG_offline.
+ */
+static void virtio_mem_set_fake_offline(unsigned long pfn,
+ unsigned int nr_pages)
+{
+ for (; nr_pages--; pfn++)
+ __SetPageOffline(pfn_to_page(pfn));
+}
+
+/*
+ * Clear PG_offline from a range of pages.
+ */
+static void virtio_mem_clear_fake_offline(unsigned long pfn,
+ unsigned int nr_pages)
+{
+ for (; nr_pages--; pfn++)
+ __ClearPageOffline(pfn_to_page(pfn));
+}
+
+/*
+ * Release a range of fake-offline pages to the buddy, effectively
+ * fake-onlining them.
+ */
+static void virtio_mem_fake_online(unsigned long pfn, unsigned int nr_pages)
+{
+ const int order = MAX_ORDER - 1;
+ int i;
+
+ /*
+ * We are always called with subblock granularity, which is at least
+ * aligned to MAX_ORDER - 1.
+ */
+ BUG_ON(!IS_ALIGNED(pfn, 1 << order));
+ BUG_ON(!IS_ALIGNED(nr_pages, 1 << order));
+
+ virtio_mem_clear_fake_offline(pfn, nr_pages);
+
+ for (i = 0; i < nr_pages; i += 1 << order)
+ generic_online_page(pfn_to_page(pfn + i), order);
+}
+
+static void virtio_mem_online_page_cb(struct page *page, unsigned int order)
+{
+ const unsigned long addr = page_to_phys(page);
+ const unsigned long mb_id = virtio_mem_phys_to_mb_id(addr);
+ struct virtio_mem *vm;
+ int sb_id;
+
+ /*
+ * We exploit here that subblocks have at least MAX_ORDER - 1
+ * size/alignment and that this callback is is called with such a
+ * size/alignment. So we cannot cross subblocks and therefore
+ * also not memory blocks.
+ */
+ rcu_read_lock();
+ list_for_each_entry_rcu(vm, &virtio_mem_devices, next) {
+ if (!virtio_mem_owned_mb(vm, mb_id))
+ continue;
+
+ sb_id = virtio_mem_phys_to_sb_id(vm, addr);
+ /*
+ * If plugged, online the pages, otherwise, set them fake
+ * offline (PageOffline).
+ */
+ if (virtio_mem_mb_test_sb_plugged(vm, mb_id, sb_id, 1))
+ generic_online_page(page, order);
+ else
+ virtio_mem_set_fake_offline(PFN_DOWN(addr), 1 << order);
+ rcu_read_unlock();
+ return;
+ }
+ rcu_read_unlock();
+
+ /* not virtio-mem memory, but e.g., a DIMM. online it */
+ generic_online_page(page, order);
+}
+
+static uint64_t virtio_mem_send_request(struct virtio_mem *vm,
+ const struct virtio_mem_req *req)
+{
+ struct scatterlist *sgs[2], sg_req, sg_resp;
+ unsigned int len;
+ int rc;
+
+ /* don't use the request residing on the stack (vaddr) */
+ vm->req = *req;
+
+ /* out: buffer for request */
+ sg_init_one(&sg_req, &vm->req, sizeof(vm->req));
+ sgs[0] = &sg_req;
+
+ /* in: buffer for response */
+ sg_init_one(&sg_resp, &vm->resp, sizeof(vm->resp));
+ sgs[1] = &sg_resp;
+
+ rc = virtqueue_add_sgs(vm->vq, sgs, 1, 1, vm, GFP_KERNEL);
+ if (rc < 0)
+ return rc;
+
+ virtqueue_kick(vm->vq);
+
+ /* wait for a response */
+ wait_event(vm->host_resp, virtqueue_get_buf(vm->vq, &len));
+
+ return virtio16_to_cpu(vm->vdev, vm->resp.type);
+}
+
+static int virtio_mem_send_plug_request(struct virtio_mem *vm, uint64_t addr,
+ uint64_t size)
+{
+ const uint64_t nb_vm_blocks = size / vm->device_block_size;
+ const struct virtio_mem_req req = {
+ .type = cpu_to_virtio16(vm->vdev, VIRTIO_MEM_REQ_PLUG),
+ .u.plug.addr = cpu_to_virtio64(vm->vdev, addr),
+ .u.plug.nb_blocks = cpu_to_virtio16(vm->vdev, nb_vm_blocks),
+ };
+
+ if (atomic_read(&vm->config_changed))
+ return -EAGAIN;
+
+ switch (virtio_mem_send_request(vm, &req)) {
+ case VIRTIO_MEM_RESP_ACK:
+ vm->plugged_size += size;
+ return 0;
+ case VIRTIO_MEM_RESP_NACK:
+ return -EAGAIN;
+ case VIRTIO_MEM_RESP_BUSY:
+ return -EBUSY;
+ case VIRTIO_MEM_RESP_ERROR:
+ return -EINVAL;
+ default:
+ return -ENOMEM;
+ }
+}
+
+static int virtio_mem_send_unplug_request(struct virtio_mem *vm, uint64_t addr,
+ uint64_t size)
+{
+ const uint64_t nb_vm_blocks = size / vm->device_block_size;
+ const struct virtio_mem_req req = {
+ .type = cpu_to_virtio16(vm->vdev, VIRTIO_MEM_REQ_UNPLUG),
+ .u.unplug.addr = cpu_to_virtio64(vm->vdev, addr),
+ .u.unplug.nb_blocks = cpu_to_virtio16(vm->vdev, nb_vm_blocks),
+ };
+
+ if (atomic_read(&vm->config_changed))
+ return -EAGAIN;
+
+ switch (virtio_mem_send_request(vm, &req)) {
+ case VIRTIO_MEM_RESP_ACK:
+ vm->plugged_size -= size;
+ return 0;
+ case VIRTIO_MEM_RESP_BUSY:
+ return -EBUSY;
+ case VIRTIO_MEM_RESP_ERROR:
+ return -EINVAL;
+ default:
+ return -ENOMEM;
+ }
+}
+
+static int virtio_mem_send_unplug_all_request(struct virtio_mem *vm)
+{
+ const struct virtio_mem_req req = {
+ .type = cpu_to_virtio16(vm->vdev, VIRTIO_MEM_REQ_UNPLUG_ALL),
+ };
+
+ switch (virtio_mem_send_request(vm, &req)) {
+ case VIRTIO_MEM_RESP_ACK:
+ vm->unplug_all_required = false;
+ vm->plugged_size = 0;
+ /* usable region might have shrunk */
+ atomic_set(&vm->config_changed, 1);
+ return 0;
+ case VIRTIO_MEM_RESP_BUSY:
+ return -EBUSY;
+ default:
+ return -ENOMEM;
+ }
+}
+
+/*
+ * Plug selected subblocks. Updates the plugged state, but not the state
+ * of the memory block.
+ */
+static int virtio_mem_mb_plug_sb(struct virtio_mem *vm, unsigned long mb_id,
+ int sb_id, int count)
+{
+ const uint64_t addr = virtio_mem_mb_id_to_phys(mb_id) +
+ sb_id * vm->subblock_size;
+ const uint64_t size = count * vm->subblock_size;
+ int rc;
+
+ dev_dbg(&vm->vdev->dev, "plugging memory block: %lu : %i - %i\n", mb_id,
+ sb_id, sb_id + count - 1);
+
+ rc = virtio_mem_send_plug_request(vm, addr, size);
+ if (!rc)
+ virtio_mem_mb_set_sb_plugged(vm, mb_id, sb_id, count);
+ return rc;
+}
+
+/*
+ * Unplug selected subblocks. Updates the plugged state, but not the state
+ * of the memory block.
+ */
+static int virtio_mem_mb_unplug_sb(struct virtio_mem *vm, unsigned long mb_id,
+ int sb_id, int count)
+{
+ const uint64_t addr = virtio_mem_mb_id_to_phys(mb_id) +
+ sb_id * vm->subblock_size;
+ const uint64_t size = count * vm->subblock_size;
+ int rc;
+
+ dev_dbg(&vm->vdev->dev, "unplugging memory block: %lu : %i - %i\n",
+ mb_id, sb_id, sb_id + count - 1);
+
+ rc = virtio_mem_send_unplug_request(vm, addr, size);
+ if (!rc)
+ virtio_mem_mb_set_sb_unplugged(vm, mb_id, sb_id, count);
+ return rc;
+}
+
+/*
+ * Unplug the desired number of plugged subblocks of a offline or unadded
+ * memory block. Will fail if any subblock cannot get unplugged (instead of
+ * skipping it).
+ *
+ * Will not modify the state of the memory block.
+ *
+ * Note: can fail after some subblocks were unplugged.
+ */
+static int virtio_mem_mb_unplug_any_sb(struct virtio_mem *vm,
+ unsigned long mb_id, uint64_t *nb_sb)
+{
+ int sb_id, count;
+ int rc;
+
+ while (*nb_sb) {
+ sb_id = virtio_mem_mb_first_plugged_sb(vm, mb_id);
+ if (sb_id >= vm->nb_sb_per_mb)
+ break;
+ count = 1;
+ while (count < *nb_sb &&
+ sb_id + count < vm->nb_sb_per_mb &&
+ virtio_mem_mb_test_sb_plugged(vm, mb_id, sb_id + count,
+ 1))
+ count++;
+
+ rc = virtio_mem_mb_unplug_sb(vm, mb_id, sb_id, count);
+ if (rc)
+ return rc;
+ *nb_sb -= count;
+ }
+
+ return 0;
+}
+
+/*
+ * Unplug all plugged subblocks of an offline or unadded memory block.
+ *
+ * Will not modify the state of the memory block.
+ *
+ * Note: can fail after some subblocks were unplugged.
+ */
+static int virtio_mem_mb_unplug(struct virtio_mem *vm, unsigned long mb_id)
+{
+ uint64_t nb_sb = vm->nb_sb_per_mb;
+
+ return virtio_mem_mb_unplug_any_sb(vm, mb_id, &nb_sb);
+}
+
+/*
+ * Prepare tracking data for the next memory block.
+ */
+static int virtio_mem_prepare_next_mb(struct virtio_mem *vm,
+ unsigned long *mb_id)
+{
+ int rc;
+
+ if (vm->next_mb_id > vm->last_usable_mb_id)
+ return -ENOSPC;
+
+ /* Resize the state array if required. */
+ rc = virtio_mem_mb_state_prepare_next_mb(vm);
+ if (rc)
+ return rc;
+
+ /* Resize the subblock bitmap if required. */
+ rc = virtio_mem_sb_bitmap_prepare_next_mb(vm);
+ if (rc)
+ return rc;
+
+ vm->nb_mb_state[VIRTIO_MEM_MB_STATE_UNUSED]++;
+ *mb_id = vm->next_mb_id++;
+ return 0;
+}
+
+/*
+ * Don't add too many blocks that are not onlined yet to avoid running OOM.
+ */
+static bool virtio_mem_too_many_mb_offline(struct virtio_mem *vm)
+{
+ unsigned long nb_offline;
+
+ nb_offline = vm->nb_mb_state[VIRTIO_MEM_MB_STATE_OFFLINE] +
+ vm->nb_mb_state[VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL];
+ return nb_offline >= VIRTIO_MEM_NB_OFFLINE_THRESHOLD;
+}
+
+/*
+ * Try to plug the desired number of subblocks and add the memory block
+ * to Linux.
+ *
+ * Will modify the state of the memory block.
+ */
+static int virtio_mem_mb_plug_and_add(struct virtio_mem *vm,
+ unsigned long mb_id,
+ uint64_t *nb_sb)
+{
+ const int count = min_t(int, *nb_sb, vm->nb_sb_per_mb);
+ int rc, rc2;
+
+ if (WARN_ON_ONCE(!count))
+ return -EINVAL;
+
+ /*
+ * Plug the requested number of subblocks before adding it to linux,
+ * so that onlining will directly online all plugged subblocks.
+ */
+ rc = virtio_mem_mb_plug_sb(vm, mb_id, 0, count);
+ if (rc)
+ return rc;
+
+ /*
+ * Mark the block properly offline before adding it to Linux,
+ * so the memory notifiers will find the block in the right state.
+ */
+ if (count == vm->nb_sb_per_mb)
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE);
+ else
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL);
+
+ /* Add the memory block to linux - if that fails, try to unplug. */
+ rc = virtio_mem_mb_add(vm, mb_id);
+ if (rc) {
+ enum virtio_mem_mb_state new_state = VIRTIO_MEM_MB_STATE_UNUSED;
+
+ dev_err(&vm->vdev->dev,
+ "adding memory block %lu failed with %d\n", mb_id, rc);
+ rc2 = virtio_mem_mb_unplug_sb(vm, mb_id, 0, count);
+
+ /*
+ * TODO: Linux MM does not properly clean up yet in all cases
+ * where adding of memory failed - especially on -ENOMEM.
+ */
+ if (rc2)
+ new_state = VIRTIO_MEM_MB_STATE_PLUGGED;
+ virtio_mem_mb_set_state(vm, mb_id, new_state);
+ return rc;
+ }
+
+ *nb_sb -= count;
+ return 0;
+}
+
+/*
+ * Try to plug the desired number of subblocks of a memory block that
+ * is already added to Linux.
+ *
+ * Will modify the state of the memory block.
+ *
+ * Note: Can fail after some subblocks were successfully plugged.
+ */
+static int virtio_mem_mb_plug_any_sb(struct virtio_mem *vm, unsigned long mb_id,
+ uint64_t *nb_sb, bool online)
+{
+ unsigned long pfn, nr_pages;
+ int sb_id, count;
+ int rc;
+
+ if (WARN_ON_ONCE(!*nb_sb))
+ return -EINVAL;
+
+ while (*nb_sb) {
+ sb_id = virtio_mem_mb_first_unplugged_sb(vm, mb_id);
+ if (sb_id >= vm->nb_sb_per_mb)
+ break;
+ count = 1;
+ while (count < *nb_sb &&
+ sb_id + count < vm->nb_sb_per_mb &&
+ !virtio_mem_mb_test_sb_plugged(vm, mb_id, sb_id + count,
+ 1))
+ count++;
+
+ rc = virtio_mem_mb_plug_sb(vm, mb_id, sb_id, count);
+ if (rc)
+ return rc;
+ *nb_sb -= count;
+ if (!online)
+ continue;
+
+ /* fake-online the pages if the memory block is online */
+ pfn = PFN_DOWN(virtio_mem_mb_id_to_phys(mb_id) +
+ sb_id * vm->subblock_size);
+ nr_pages = PFN_DOWN(count * vm->subblock_size);
+ virtio_mem_fake_online(pfn, nr_pages);
+ }
+
+ if (virtio_mem_mb_test_sb_plugged(vm, mb_id, 0, vm->nb_sb_per_mb)) {
+ if (online)
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_ONLINE);
+ else
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE);
+ }
+
+ return rc;
+}
+
+/*
+ * Try to plug the requested amount of memory.
+ */
+static int virtio_mem_plug_request(struct virtio_mem *vm, uint64_t diff)
+{
+ uint64_t nb_sb = diff / vm->subblock_size;
+ unsigned long mb_id;
+ int rc;
+
+ if (!nb_sb)
+ return 0;
+
+ /* Don't race with onlining/offlining */
+ mutex_lock(&vm->hotplug_mutex);
+
+ /* Try to plug subblocks of partially plugged online blocks. */
+ virtio_mem_for_each_mb_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL) {
+ rc = virtio_mem_mb_plug_any_sb(vm, mb_id, &nb_sb, true);
+ if (rc || !nb_sb)
+ goto out_unlock;
+ cond_resched();
+ }
+
+ /* Try to plug subblocks of partially plugged offline blocks. */
+ virtio_mem_for_each_mb_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL) {
+ rc = virtio_mem_mb_plug_any_sb(vm, mb_id, &nb_sb, false);
+ if (rc || !nb_sb)
+ goto out_unlock;
+ cond_resched();
+ }
+
+ /*
+ * We won't be working on online/offline memory blocks from this point,
+ * so we can't race with memory onlining/offlining. Drop the mutex.
+ */
+ mutex_unlock(&vm->hotplug_mutex);
+
+ /* Try to plug and add unused blocks */
+ virtio_mem_for_each_mb_state(vm, mb_id, VIRTIO_MEM_MB_STATE_UNUSED) {
+ if (virtio_mem_too_many_mb_offline(vm))
+ return -ENOSPC;
+
+ rc = virtio_mem_mb_plug_and_add(vm, mb_id, &nb_sb);
+ if (rc || !nb_sb)
+ return rc;
+ cond_resched();
+ }
+
+ /* Try to prepare, plug and add new blocks */
+ while (nb_sb) {
+ if (virtio_mem_too_many_mb_offline(vm))
+ return -ENOSPC;
+
+ rc = virtio_mem_prepare_next_mb(vm, &mb_id);
+ if (rc)
+ return rc;
+ rc = virtio_mem_mb_plug_and_add(vm, mb_id, &nb_sb);
+ if (rc)
+ return rc;
+ cond_resched();
+ }
+
+ return 0;
+out_unlock:
+ mutex_unlock(&vm->hotplug_mutex);
+ return rc;
+}
+
+/*
+ * Try to unplug all blocks that couldn't be unplugged before, for example,
+ * because the hypervisor was busy.
+ */
+static int virtio_mem_unplug_pending_mb(struct virtio_mem *vm)
+{
+ unsigned long mb_id;
+ int rc;
+
+ virtio_mem_for_each_mb_state(vm, mb_id, VIRTIO_MEM_MB_STATE_PLUGGED) {
+ rc = virtio_mem_mb_unplug(vm, mb_id);
+ if (rc)
+ return rc;
+ virtio_mem_mb_set_state(vm, mb_id, VIRTIO_MEM_MB_STATE_UNUSED);
+ }
+
+ return 0;
+}
+
+/*
+ * Update all parts of the config that could have changed.
+ */
+static void virtio_mem_refresh_config(struct virtio_mem *vm)
+{
+ const uint64_t phys_limit = 1UL << MAX_PHYSMEM_BITS;
+ uint64_t new_plugged_size, usable_region_size, end_addr;
+
+ /* the plugged_size is just a reflection of what _we_ did previously */
+ virtio_cread(vm->vdev, struct virtio_mem_config, plugged_size,
+ &new_plugged_size);
+ if (WARN_ON_ONCE(new_plugged_size != vm->plugged_size))
+ vm->plugged_size = new_plugged_size;
+
+ /* calculate the last usable memory block id */
+ virtio_cread(vm->vdev, struct virtio_mem_config,
+ usable_region_size, &usable_region_size);
+ end_addr = vm->addr + usable_region_size;
+ end_addr = min(end_addr, phys_limit);
+ vm->last_usable_mb_id = virtio_mem_phys_to_mb_id(end_addr) - 1;
+
+ /* see if there is a request to change the size */
+ virtio_cread(vm->vdev, struct virtio_mem_config, requested_size,
+ &vm->requested_size);
+
+ dev_info(&vm->vdev->dev, "plugged size: 0x%llx", vm->plugged_size);
+ dev_info(&vm->vdev->dev, "requested size: 0x%llx", vm->requested_size);
+}
+
+/*
+ * Workqueue function for handling plug/unplug requests and config updates.
+ */
+static void virtio_mem_run_wq(struct work_struct *work)
+{
+ struct virtio_mem *vm = container_of(work, struct virtio_mem, wq);
+ uint64_t diff;
+ int rc = 0;
+
+ hrtimer_cancel(&vm->retry_timer);
+
+ if (vm->broken)
+ return;
+
+ /* Make sure we start with a clean state if there are leftovers. */
+ if (unlikely(vm->unplug_all_required))
+ rc = virtio_mem_send_unplug_all_request(vm);
+
+config_update:
+ if (atomic_read(&vm->config_changed)) {
+ atomic_set(&vm->config_changed, 0);
+ virtio_mem_refresh_config(vm);
+ }
+
+ /* Unplug any leftovers from previous runs */
+ if (!rc)
+ rc = virtio_mem_unplug_pending_mb(vm);
+
+ if (!rc && vm->requested_size != vm->plugged_size) {
+ if (vm->requested_size > vm->plugged_size) {
+ diff = vm->requested_size - vm->plugged_size;
+ rc = virtio_mem_plug_request(vm, diff);
+ }
+ /* TODO: try to unplug memory */
+ }
+
+ switch (rc) {
+ case 0:
+ break;
+ case -ENOSPC:
+ /*
+ * We cannot add any more memory (alignment, physical limit)
+ * or we have too many offline memory blocks.
+ */
+ break;
+ case -EBUSY:
+ /*
+ * The hypervisor cannot process our request right now
+ * (e.g., out of memory, migrating).
+ */
+ case -ENOMEM:
+ /* Out of memory, try again later. */
+ hrtimer_start(&vm->retry_timer,
+ ms_to_ktime(VIRTIO_MEM_RETRY_TIMER_MS),
+ HRTIMER_MODE_REL);
+ break;
+ case -EAGAIN:
+ /* Retry immediately (e.g., the config changed). */
+ goto config_update;
+ default:
+ /* Unknown error, mark as broken */
+ dev_err(&vm->vdev->dev,
+ "unknown error, marking device broken: %d\n", rc);
+ vm->broken = true;
+ }
+}
+
+static enum hrtimer_restart virtio_mem_timer_expired(struct hrtimer *timer)
+{
+ struct virtio_mem *vm = container_of(timer, struct virtio_mem,
+ retry_timer);
+
+ virtio_mem_retry(vm);
+ return HRTIMER_NORESTART;
+}
+
+static void virtio_mem_handle_response(struct virtqueue *vq)
+{
+ struct virtio_mem *vm = vq->vdev->priv;
+
+ wake_up(&vm->host_resp);
+}
+
+static int virtio_mem_init_vq(struct virtio_mem *vm)
+{
+ struct virtqueue *vq;
+
+ vq = virtio_find_single_vq(vm->vdev, virtio_mem_handle_response,
+ "guest-request");
+ if (IS_ERR(vq))
+ return PTR_ERR(vq);
+ vm->vq = vq;
+
+ return 0;
+}
+
+/*
+ * Test if any memory in the range is present in Linux.
+ */
+static bool virtio_mem_any_memory_present(unsigned long start,
+ unsigned long size)
+{
+ const unsigned long start_pfn = PFN_DOWN(start);
+ const unsigned long end_pfn = PFN_UP(start + size);
+ unsigned long pfn;
+
+ for (pfn = start_pfn; pfn != end_pfn; pfn++)
+ if (present_section_nr(pfn_to_section_nr(pfn)))
+ return true;
+
+ return false;
+}
+
+static int virtio_mem_init(struct virtio_mem *vm)
+{
+ const uint64_t phys_limit = 1UL << MAX_PHYSMEM_BITS;
+ uint64_t region_size;
+ uint16_t node_id;
+
+ if (!vm->vdev->config->get) {
+ dev_err(&vm->vdev->dev, "config access disabled\n");
+ return -EINVAL;
+ }
+
+ /*
+ * We don't want to (un)plug or reuse any memory when in kdump. The
+ * memory is still accessible (but not mapped).
+ */
+ if (is_kdump_kernel()) {
+ dev_warn(&vm->vdev->dev, "disabled in kdump kernel\n");
+ return -EBUSY;
+ }
+
+ /* Fetch all properties that can't change. */
+ virtio_cread(vm->vdev, struct virtio_mem_config, plugged_size,
+ &vm->plugged_size);
+ virtio_cread(vm->vdev, struct virtio_mem_config, block_size,
+ &vm->device_block_size);
+ virtio_cread(vm->vdev, struct virtio_mem_config, node_id,
+ &node_id);
+ vm->nid = virtio_mem_translate_node_id(vm, node_id);
+ virtio_cread(vm->vdev, struct virtio_mem_config, addr, &vm->addr);
+ virtio_cread(vm->vdev, struct virtio_mem_config, region_size,
+ &region_size);
+
+ /*
+ * If we still have memory plugged, we might have to unplug all
+ * memory first. However, if somebody simply unloaded the driver
+ * we would have to reinitialize the old state - something we don't
+ * support yet. Detect if we have any memory in the area present.
+ */
+ if (vm->plugged_size) {
+ uint64_t usable_region_size;
+
+ virtio_cread(vm->vdev, struct virtio_mem_config,
+ usable_region_size, &usable_region_size);
+
+ if (virtio_mem_any_memory_present(vm->addr,
+ usable_region_size)) {
+ dev_err(&vm->vdev->dev,
+ "reloading the driver is not supported\n");
+ return -EINVAL;
+ }
+ /*
+ * Note: it might happen that the device is busy and
+ * unplugging all memory might take some time.
+ */
+ dev_info(&vm->vdev->dev, "unplugging all memory required\n");
+ vm->unplug_all_required = 1;
+ }
+
+ /*
+ * We always hotplug memory in memory block granularity. This way,
+ * we have to wait for exactly one memory block to online.
+ */
+ if (vm->device_block_size > memory_block_size_bytes()) {
+ dev_err(&vm->vdev->dev,
+ "The block size is not supported (too big).\n");
+ return -EINVAL;
+ }
+
+ /* bad device setup - warn only */
+ if (!IS_ALIGNED(vm->addr, memory_block_size_bytes()))
+ dev_warn(&vm->vdev->dev,
+ "The alignment of the physical start address can make some memory unusable.\n");
+ if (!IS_ALIGNED(vm->addr + region_size, memory_block_size_bytes()))
+ dev_warn(&vm->vdev->dev,
+ "The alignment of the physical end address can make some memory unusable.\n");
+ if (vm->addr + region_size > phys_limit)
+ dev_warn(&vm->vdev->dev,
+ "Some memory is not addressable. This can make some memory unusable.\n");
+
+ /*
+ * Calculate the subblock size:
+ * - At least MAX_ORDER - 1 / pageblock_order.
+ * - At least the device block size.
+ * In the worst case, a single subblock per memory block.
+ */
+ vm->subblock_size = PAGE_SIZE * 1u << max_t(uint32_t, MAX_ORDER - 1,
+ pageblock_order);
+ vm->subblock_size = max_t(uint32_t, vm->device_block_size,
+ vm->subblock_size);
+ vm->nb_sb_per_mb = memory_block_size_bytes() / vm->subblock_size;
+
+ /* Round up to the next full memory block */
+ vm->first_mb_id = virtio_mem_phys_to_mb_id(vm->addr - 1 +
+ memory_block_size_bytes());
+ vm->next_mb_id = vm->first_mb_id;
+ vm->last_mb_id = virtio_mem_phys_to_mb_id(vm->addr + region_size) - 1;
+
+ dev_info(&vm->vdev->dev, "start address: 0x%llx", vm->addr);
+ dev_info(&vm->vdev->dev, "region size: 0x%llx", region_size);
+ dev_info(&vm->vdev->dev, "device block size: 0x%x",
+ vm->device_block_size);
+ dev_info(&vm->vdev->dev, "memory block size: 0x%lx",
+ memory_block_size_bytes());
+ dev_info(&vm->vdev->dev, "subblock size: 0x%x",
+ vm->subblock_size);
+
+ return 0;
+}
+
+static int virtio_mem_probe(struct virtio_device *vdev)
+{
+ struct virtio_mem *vm;
+ int rc = -EINVAL;
+
+ vdev->priv = vm = kzalloc(sizeof(*vm), GFP_KERNEL);
+ if (!vm)
+ return -ENOMEM;
+
+ init_waitqueue_head(&vm->host_resp);
+ vm->vdev = vdev;
+ INIT_WORK(&vm->wq, virtio_mem_run_wq);
+ mutex_init(&vm->hotplug_mutex);
+ /*
+ * Avoid circular locking lockdep warnings. We lock the mutex
+ * e.g., in MEM_GOING_ONLINE and unlock it in MEM_ONLINE. The
+ * blocking_notifier_call_chain() has it's own lock, which gets unlocked
+ * between both notifier calls.
+ */
+ lockdep_set_novalidate_class(&vm->hotplug_mutex);
+ INIT_LIST_HEAD(&vm->next);
+ spin_lock_init(&vm->removal_lock);
+ hrtimer_init(&vm->retry_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+ vm->retry_timer.function = virtio_mem_timer_expired;
+
+ /* register the virtqueue */
+ rc = virtio_mem_init_vq(vm);
+ if (rc)
+ goto out_free_vm;
+
+ /* initialize the device by querying the config */
+ rc = virtio_mem_init(vm);
+ if (rc)
+ goto out_del_vq;
+
+ /* register callbacks */
+ vm->memory_notifier.notifier_call = virtio_mem_memory_notifier_cb;
+ rc = register_memory_notifier(&vm->memory_notifier);
+ if (rc)
+ goto out_del_vq;
+ rc = register_virtio_mem_device(vm);
+ if (rc)
+ goto out_unreg_mem;
+
+ virtio_device_ready(vdev);
+
+ /* trigger a config update to start processing the requested_size */
+ atomic_set(&vm->config_changed, 1);
+ queue_work(system_freezable_wq, &vm->wq);
+
+ return 0;
+out_unreg_mem:
+ unregister_memory_notifier(&vm->memory_notifier);
+out_del_vq:
+ vdev->config->del_vqs(vdev);
+out_free_vm:
+ kfree(vm);
+ vdev->priv = NULL;
+
+ return rc;
+}
+
+static void virtio_mem_remove(struct virtio_device *vdev)
+{
+ struct virtio_mem *vm = vdev->priv;
+ unsigned long mb_id;
+ int rc;
+
+ /*
+ * Make sure the workqueue won't be triggered anymore and no memory
+ * blocks can be onlined/offlined until we're finished here.
+ */
+ spin_lock_irq(&vm->removal_lock);
+ vm->removing = true;
+ spin_unlock_irq(&vm->removal_lock);
+
+ /* wait until the workqueue stopped */
+ cancel_work_sync(&vm->wq);
+ hrtimer_cancel(&vm->retry_timer);
+
+ /* make sure all callbacks saw vm->removing */
+ mutex_lock(&vm->hotplug_mutex);
+ mutex_unlock(&vm->hotplug_mutex);
+
+ /*
+ * After we unregistered our callbacks, user space can online partially
+ * plugged offline blocks. Make sure to remove them.
+ */
+ virtio_mem_for_each_mb_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_OFFLINE_PARTIAL) {
+ rc = virtio_mem_mb_remove(vm, mb_id);
+ BUG_ON(rc);
+ }
+
+ /* unregister callbacks */
+ unregister_virtio_mem_device(vm);
+ unregister_memory_notifier(&vm->memory_notifier);
+
+ /*
+ * There is no way we could deterministically remove all memory we have
+ * added to the system. And there is no way to stop the driver/device
+ * from going away.
+ */
+ if (vm->plugged_size)
+ dev_warn(&vdev->dev, "device still has memory plugged\n");
+
+ /* remove all tracking data - no locking needed */
+ vfree(vm->mb_state);
+ vfree(vm->sb_bitmap);
+
+ /* reset the device and cleanup the queues */
+ vdev->config->reset(vdev);
+ vdev->config->del_vqs(vdev);
+
+ kfree(vm);
+ vdev->priv = NULL;
+}
+
+static void virtio_mem_config_changed(struct virtio_device *vdev)
+{
+ struct virtio_mem *vm = vdev->priv;
+
+ atomic_set(&vm->config_changed, 1);
+ virtio_mem_retry(vm);
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int virtio_mem_freeze(struct virtio_device *vdev)
+{
+ /*
+ * When restarting the VM, all memory is usually unplugged. Don't
+ * allow to suspend/hibernate.
+ */
+ dev_err(&vdev->dev, "save/restore not supported.\n");
+ return -EPERM;
+}
+
+static int virtio_mem_restore(struct virtio_device *vdev)
+{
+ return -EPERM;
+}
+#endif
+
+static unsigned int virtio_mem_features[] = {
+#if defined(CONFIG_NUMA) && defined(CONFIG_ACPI_NUMA)
+ VIRTIO_MEM_F_ACPI_PXM,
+#endif
+};
+
+static struct virtio_device_id virtio_mem_id_table[] = {
+ { VIRTIO_ID_MEM, VIRTIO_DEV_ANY_ID },
+ { 0 },
+};
+
+static struct virtio_driver virtio_mem_driver = {
+ .feature_table = virtio_mem_features,
+ .feature_table_size = ARRAY_SIZE(virtio_mem_features),
+ .driver.name = KBUILD_MODNAME,
+ .driver.owner = THIS_MODULE,
+ .id_table = virtio_mem_id_table,
+ .probe = virtio_mem_probe,
+ .remove = virtio_mem_remove,
+ .config_changed = virtio_mem_config_changed,
+#ifdef CONFIG_PM_SLEEP
+ .freeze = virtio_mem_freeze,
+ .restore = virtio_mem_restore,
+#endif
+};
+
+module_virtio_driver(virtio_mem_driver);
+MODULE_DEVICE_TABLE(virtio, virtio_mem_id_table);
+MODULE_AUTHOR("David Hildenbrand <[email protected]>");
+MODULE_DESCRIPTION("Virtio-mem driver");
+MODULE_LICENSE("GPL");
diff --git a/include/uapi/linux/virtio_ids.h b/include/uapi/linux/virtio_ids.h
index 585e07b27333..934e0b444454 100644
--- a/include/uapi/linux/virtio_ids.h
+++ b/include/uapi/linux/virtio_ids.h
@@ -44,6 +44,7 @@
#define VIRTIO_ID_VSOCK 19 /* virtio vsock transport */
#define VIRTIO_ID_CRYPTO 20 /* virtio crypto */
#define VIRTIO_ID_IOMMU 23 /* virtio IOMMU */
+#define VIRTIO_ID_MEM 24 /* virtio mem */
#define VIRTIO_ID_FS 26 /* virtio filesystem */
#define VIRTIO_ID_PMEM 27 /* virtio pmem */

diff --git a/include/uapi/linux/virtio_mem.h b/include/uapi/linux/virtio_mem.h
new file mode 100644
index 000000000000..bac73fca6c0e
--- /dev/null
+++ b/include/uapi/linux/virtio_mem.h
@@ -0,0 +1,204 @@
+/* SPDX-License-Identifier: BSD-3-Clause */
+/*
+ * Virtio Mem Device
+ *
+ * Copyright Red Hat, Inc. 2019
+ *
+ * Authors:
+ * David Hildenbrand <[email protected]>
+ *
+ * This header is BSD licensed so anyone can use the definitions
+ * to implement compatible drivers/servers:
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ * 3. Neither the name of IBM nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+ * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+ * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS
+ * FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL IBM OR
+ * CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
+ * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
+ * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF
+ * USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY,
+ * OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT
+ * OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#ifndef _LINUX_VIRTIO_MEM_H
+#define _LINUX_VIRTIO_MEM_H
+
+#include <linux/types.h>
+#include <linux/virtio_types.h>
+#include <linux/virtio_ids.h>
+#include <linux/virtio_config.h>
+
+/*
+ * Each virtio-mem device manages a dedicated region in physical address
+ * space. Each device can belong to a single NUMA node, multiple devices
+ * for a single NUMA node are possible. A virtio-mem device is like a
+ * "resizable DIMM" consisting of small memory blocks that can be plugged
+ * or unplugged. The device driver is responsible for (un)plugging memory
+ * blocks on demand.
+ *
+ * Virtio-mem devices can only operate on their assigned memory region in
+ * order to (un)plug memory. A device cannot (un)plug memory belonging to
+ * other devices.
+ *
+ * The "region_size" corresponds to the maximum amount of memory that can
+ * be provided by a device. The "size" corresponds to the amount of memory
+ * that is currently plugged. "requested_size" corresponds to a request
+ * from the device to the device driver to (un)plug blocks. The
+ * device driver should try to (un)plug blocks in order to reach the
+ * "requested_size". It is impossible to plug more memory than requested.
+ *
+ * The "usable_region_size" represents the memory region that can actually
+ * be used to (un)plug memory. It is always at least as big as the
+ * "requested_size" and will grow dynamically. It will only shrink when
+ * explicitly triggered (VIRTIO_MEM_REQ_UNPLUG).
+ *
+ * Memory in the usable region can usually be read, however, there are no
+ * guarantees. It can happen that the device cannot process a request,
+ * because it is busy. The device driver has to retry later.
+ *
+ * Usually, during system resets all memory will get unplugged, so the
+ * device driver can start with a clean state. However, in specific
+ * scenarios (if the device is busy) it can happen that the device still
+ * has memory plugged. The device driver can request to unplug all memory
+ * (VIRTIO_MEM_REQ_UNPLUG) - which might take a while to succeed if the
+ * device is busy.
+ */
+
+/* --- virtio-mem: feature bits --- */
+
+/* node_id is an ACPI PXM and is valid */
+#define VIRTIO_MEM_F_ACPI_PXM 0
+
+
+/* --- virtio-mem: guest -> host requests --- */
+
+/* request to plug memory blocks */
+#define VIRTIO_MEM_REQ_PLUG 0
+/* request to unplug memory blocks */
+#define VIRTIO_MEM_REQ_UNPLUG 1
+/* request to unplug all blocks and shrink the usable size */
+#define VIRTIO_MEM_REQ_UNPLUG_ALL 2
+/* request information about the plugged state of memory blocks */
+#define VIRTIO_MEM_REQ_STATE 3
+
+struct virtio_mem_req_plug {
+ __virtio64 addr;
+ __virtio16 nb_blocks;
+};
+
+struct virtio_mem_req_unplug {
+ __virtio64 addr;
+ __virtio16 nb_blocks;
+};
+
+struct virtio_mem_req_state {
+ __virtio64 addr;
+ __virtio16 nb_blocks;
+};
+
+struct virtio_mem_req {
+ __virtio16 type;
+ __virtio16 padding[3];
+
+ union {
+ struct virtio_mem_req_plug plug;
+ struct virtio_mem_req_unplug unplug;
+ struct virtio_mem_req_state state;
+ } u;
+};
+
+
+/* --- virtio-mem: host -> guest response --- */
+
+/*
+ * Request processed successfully, applicable for
+ * - VIRTIO_MEM_REQ_PLUG
+ * - VIRTIO_MEM_REQ_UNPLUG
+ * - VIRTIO_MEM_REQ_UNPLUG_ALL
+ * - VIRTIO_MEM_REQ_STATE
+ */
+#define VIRTIO_MEM_RESP_ACK 0
+/*
+ * Request denied - e.g. trying to plug more than requested, applicable for
+ * - VIRTIO_MEM_REQ_PLUG
+ */
+#define VIRTIO_MEM_RESP_NACK 1
+/*
+ * Request cannot be processed right now, try again later, applicable for
+ * - VIRTIO_MEM_REQ_PLUG
+ * - VIRTIO_MEM_REQ_UNPLUG
+ * - VIRTIO_MEM_REQ_UNPLUG_ALL
+ */
+#define VIRTIO_MEM_RESP_BUSY 2
+/*
+ * Error in request (e.g. addresses/alignemnt), applicable for
+ * - VIRTIO_MEM_REQ_PLUG
+ * - VIRTIO_MEM_REQ_UNPLUG
+ * - VIRTIO_MEM_REQ_STATE
+ */
+#define VIRTIO_MEM_RESP_ERROR 3
+
+
+/* State of memory blocks is "plugged" */
+#define VIRTIO_MEM_STATE_PLUGGED 0
+/* State of memory blocks is "unplugged" */
+#define VIRTIO_MEM_STATE_UNPLUGGED 1
+/* State of memory blocks is "mixed" */
+#define VIRTIO_MEM_STATE_MIXED 2
+
+struct virtio_mem_resp_state {
+ __virtio16 state;
+};
+
+struct virtio_mem_resp {
+ __virtio16 type;
+ __virtio16 padding[3];
+
+ union {
+ struct virtio_mem_resp_state state;
+ } u;
+};
+
+/* --- virtio-mem: configuration --- */
+
+struct virtio_mem_config {
+ /* Block size and alignment. Cannot change. */
+ uint32_t block_size;
+ /* Valid with VIRTIO_MEM_F_ACPI_PXM. Cannot change. */
+ uint16_t node_id;
+ uint16_t padding;
+ /* Start address of the memory region. Cannot change. */
+ uint64_t addr;
+ /* Region size (maximum). Cannot change. */
+ uint64_t region_size;
+ /*
+ * Currently usable region size. Can grow up to region_size. Can
+ * shrink due to VIRTIO_MEM_REQ_UNPLUG_ALL (in which case no config
+ * update will be sent).
+ */
+ uint64_t usable_region_size;
+ /*
+ * Currently used size. Changes due to plug/unplug requests, but no
+ * config updates will be sent.
+ */
+ uint64_t plugged_size;
+ /* Requested size. New plug requests cannot exceed it. Can change. */
+ uint64_t requested_size;
+};
+
+#endif /* _LINUX_VIRTIO_MEM_H */
--
2.23.0

2019-12-12 17:13:48

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 04/13] mm: Export alloc_contig_range() / free_contig_range()

A virtio-mem device wants to allocate memory from the memory region it
manages in order to unplug it in the hypervisor - similar to
a balloon driver. Also, it might want to plug previously unplugged
(allocated) memory and give it back to Linux. alloc_contig_range() /
free_contig_range() seem to be the perfect interface for this task.

In contrast to existing balloon devices, a virtio-mem device operates
on bigger chunks (e.g., 4MB) and only on physical memory it manages. It
tracks which chunks (subblocks) are still plugged, so it can go ahead
and try to alloc_contig_range()+unplug them on unplug request, or
plug+free_contig_range() unplugged chunks on plug requests.

A virtio-mem device will use alloc_contig_range() / free_contig_range()
only on ranges that belong to the same node/zone in at least
MAX(MAX_ORDER - 1, pageblock_order) order granularity - e.g., 4MB on
x86-64. The virtio-mem device added that memory, so the memory
exists and does not contain any holes. virtio-mem will only try to allocate
on ZONE_NORMAL, never on ZONE_MOVABLE, just like when allocating
gigantic pages (we don't put unmovable data into the movable zone).

Cc: Andrew Morton <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Mike Rapoport <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Alexander Duyck <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Alexander Potapenko <[email protected]>
Acked-by: Michal Hocko <[email protected]> # to export contig range allocator API
Signed-off-by: David Hildenbrand <[email protected]>
---
mm/page_alloc.c | 2 ++
1 file changed, 2 insertions(+)

diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 62dcd6b76c80..5334decc9e06 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -8494,6 +8494,7 @@ int alloc_contig_range(unsigned long start, unsigned long end,
pfn_max_align_up(end), migratetype);
return ret;
}
+EXPORT_SYMBOL(alloc_contig_range);

static int __alloc_contig_pages(unsigned long start_pfn,
unsigned long nr_pages, gfp_t gfp_mask)
@@ -8609,6 +8610,7 @@ void free_contig_range(unsigned long pfn, unsigned int nr_pages)
}
WARN(count != 0, "%d pages are still in use!\n", count);
}
+EXPORT_SYMBOL(free_contig_range);

/*
* The zone indicated has a new number of managed_pages; batch sizes and percpu
--
2.23.0

2019-12-12 17:13:50

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 05/13] virtio-mem: Paravirtualized memory hotunplug part 2

We can use alloc_contig_range() to try to unplug subblocks. Unplugged
blocks will be marked PG_offline, however, don't have the PG_reserved
flag set. This way, we can differentiate these allocated subblocks from
subblocks that were never onlined and handle them properly in
virtio_mem_fake_online(). free_contig_range() is used to hand back
subblocks to Linux.

It is worth noting that there are no guarantess on how much memory can
actually get unplugged again. All device memory might completely be
fragmented with unmovable data, such that no subblock can get unplugged.
We might want to improve the unplugging capability in the future.

We are not touching the ZONE_MOVABLE. If memory is onlined to the
ZONE_MOVABLE, it can only get unplugged after that memory was offlined
manually by user space. In normal operation, virtio-mem memory is
suggested to be onlined to ZONE_NORMAL. In the future, we will try to
make unplug more likely to succeed.

Add a module parameter to control if online memory shall be touched.

Future work:
- Performance improvements:
-- Sense (lockless) if it make sense to try alloc_contig_range() at all
before directly trying to isolate and taking locks.
-- Try to unplug bigger chunks if possible first.
-- Identify free areas first, that don't have to be evacuated.
- Make unplug more likely to succeed:
-- There are various idea to limit fragmentation on memory block
granularity. (e.g., ZONE_PREFER_MOVABLE and smart balancing)
-- Allocate memmap from added memory. This way, less unmovable data can
end up on the memory blocks.
- OOM handling, e.g., via an OOM handler.
- Defragmentation
-- Will require a new virtio-mem CMD to exchange plugged<->unplugged blocks

Cc: "Michael S. Tsirkin" <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Dave Young <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Stefan Hajnoczi <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
drivers/virtio/Kconfig | 1 +
drivers/virtio/virtio_mem.c | 156 ++++++++++++++++++++++++++++++++----
2 files changed, 143 insertions(+), 14 deletions(-)

diff --git a/drivers/virtio/Kconfig b/drivers/virtio/Kconfig
index 294720d53057..75a760f32ec7 100644
--- a/drivers/virtio/Kconfig
+++ b/drivers/virtio/Kconfig
@@ -71,6 +71,7 @@ config VIRTIO_MEM
depends on VIRTIO
depends on MEMORY_HOTPLUG_SPARSE
depends on MEMORY_HOTREMOVE
+ select CONTIG_ALLOC
help
This driver provides access to virtio-mem paravirtualized memory
devices, allowing to hotplug and hotunplug memory.
diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index f1af05def5df..5a142a371222 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -21,6 +21,10 @@

#include <acpi/acpi_numa.h>

+static bool unplug_online = true;
+module_param(unplug_online, bool, 0644);
+MODULE_PARM_DESC(unplug_online, "Try to unplug online memory");
+
enum virtio_mem_mb_state {
/* Unplugged, not added to Linux. Can be reused later. */
VIRTIO_MEM_MB_STATE_UNUSED = 0,
@@ -646,23 +650,35 @@ static int virtio_mem_memory_notifier_cb(struct notifier_block *nb,
}

/*
- * Set a range of pages PG_offline.
+ * Set a range of pages PG_offline. Remember pages that were never onlined
+ * (via generic_online_page()) using PageDirty().
*/
static void virtio_mem_set_fake_offline(unsigned long pfn,
- unsigned int nr_pages)
+ unsigned int nr_pages, bool onlined)
{
- for (; nr_pages--; pfn++)
- __SetPageOffline(pfn_to_page(pfn));
+ for (; nr_pages--; pfn++) {
+ struct page *page = pfn_to_page(pfn);
+
+ __SetPageOffline(page);
+ if (!onlined)
+ SetPageDirty(page);
+ }
}

/*
- * Clear PG_offline from a range of pages.
+ * Clear PG_offline from a range of pages. If the pages were never onlined,
+ * (via generic_online_page()), clear PageDirty().
*/
static void virtio_mem_clear_fake_offline(unsigned long pfn,
- unsigned int nr_pages)
+ unsigned int nr_pages, bool onlined)
{
- for (; nr_pages--; pfn++)
- __ClearPageOffline(pfn_to_page(pfn));
+ for (; nr_pages--; pfn++) {
+ struct page *page = pfn_to_page(pfn);
+
+ __ClearPageOffline(page);
+ if (!onlined)
+ ClearPageDirty(page);
+ }
}

/*
@@ -681,10 +697,26 @@ static void virtio_mem_fake_online(unsigned long pfn, unsigned int nr_pages)
BUG_ON(!IS_ALIGNED(pfn, 1 << order));
BUG_ON(!IS_ALIGNED(nr_pages, 1 << order));

- virtio_mem_clear_fake_offline(pfn, nr_pages);
+ for (i = 0; i < nr_pages; i += 1 << order) {
+ struct page *page = pfn_to_page(pfn + i);

- for (i = 0; i < nr_pages; i += 1 << order)
- generic_online_page(pfn_to_page(pfn + i), order);
+ /*
+ * If the page is PageDirty(), it was kept fake-offline when
+ * onlining the memory block. Otherwise, it was allocated
+ * using alloc_contig_range(). All pages in a subblock are
+ * alike.
+ */
+ if (PageDirty(page)) {
+ virtio_mem_clear_fake_offline(pfn + i, 1 << order,
+ false);
+ generic_online_page(page, order);
+ } else {
+ virtio_mem_clear_fake_offline(pfn + i, 1 << order,
+ true);
+ free_contig_range(pfn + i, 1 << order);
+ adjust_managed_page_count(page, 1 << order);
+ }
+ }
}

static void virtio_mem_online_page_cb(struct page *page, unsigned int order)
@@ -713,7 +745,8 @@ static void virtio_mem_online_page_cb(struct page *page, unsigned int order)
if (virtio_mem_mb_test_sb_plugged(vm, mb_id, sb_id, 1))
generic_online_page(page, order);
else
- virtio_mem_set_fake_offline(PFN_DOWN(addr), 1 << order);
+ virtio_mem_set_fake_offline(PFN_DOWN(addr), 1 << order,
+ false);
rcu_read_unlock();
return;
}
@@ -1181,6 +1214,72 @@ static int virtio_mem_mb_unplug_any_sb_offline(struct virtio_mem *vm,
return 0;
}

+/*
+ * Unplug the desired number of plugged subblocks of an online memory block.
+ * Will skip subblock that are busy.
+ *
+ * Will modify the state of the memory block.
+ *
+ * Note: Can fail after some subblocks were successfully unplugged. Can
+ * return 0 even if subblocks were busy and could not get unplugged.
+ */
+static int virtio_mem_mb_unplug_any_sb_online(struct virtio_mem *vm,
+ unsigned long mb_id,
+ uint64_t *nb_sb)
+{
+ const unsigned long nr_pages = PFN_DOWN(vm->subblock_size);
+ unsigned long start_pfn;
+ int rc, sb_id;
+
+ /*
+ * TODO: To increase the performance we want to try bigger, consecutive
+ * subblocks first before falling back to single subblocks. Also,
+ * we should sense via something like is_mem_section_removable()
+ * first if it makes sense to go ahead any try to allocate.
+ */
+ for (sb_id = 0; sb_id < vm->nb_sb_per_mb && *nb_sb; sb_id++) {
+ /* Find the next candidate subblock */
+ while (sb_id < vm->nb_sb_per_mb &&
+ !virtio_mem_mb_test_sb_plugged(vm, mb_id, sb_id, 1))
+ sb_id++;
+ if (sb_id >= vm->nb_sb_per_mb)
+ break;
+
+ start_pfn = PFN_DOWN(virtio_mem_mb_id_to_phys(mb_id) +
+ sb_id * vm->subblock_size);
+ rc = alloc_contig_range(start_pfn, start_pfn + nr_pages,
+ MIGRATE_MOVABLE, GFP_KERNEL);
+ if (rc == -ENOMEM)
+ /* whoops, out of memory */
+ return rc;
+ if (rc)
+ /* memory busy, we can't unplug this chunk */
+ continue;
+
+ /* Mark it as fake-offline before unplugging it */
+ virtio_mem_set_fake_offline(start_pfn, nr_pages, true);
+ adjust_managed_page_count(pfn_to_page(start_pfn), -nr_pages);
+
+ /* Try to unplug the allocated memory */
+ rc = virtio_mem_mb_unplug_sb(vm, mb_id, sb_id, 1);
+ if (rc) {
+ /* Return the memory to the buddy. */
+ virtio_mem_fake_online(start_pfn, nr_pages);
+ return rc;
+ }
+
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL);
+ *nb_sb -= 1;
+ }
+
+ /*
+ * TODO: Once all subblocks of a memory block were unplugged, we want
+ * to offline the memory block and remove it.
+ */
+ return 0;
+}
+
/*
* Try to unplug the requested amount of memory.
*/
@@ -1219,8 +1318,36 @@ static int virtio_mem_unplug_request(struct virtio_mem *vm, uint64_t diff)
cond_resched();
}

+ if (!unplug_online) {
+ mutex_unlock(&vm->hotplug_mutex);
+ return 0;
+ }
+
+ /* Try to unplug subblocks of partially plugged online blocks. */
+ virtio_mem_for_each_mb_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL) {
+ rc = virtio_mem_mb_unplug_any_sb_online(vm, mb_id,
+ &nb_sb);
+ if (rc || !nb_sb)
+ goto out_unlock;
+ mutex_unlock(&vm->hotplug_mutex);
+ cond_resched();
+ mutex_lock(&vm->hotplug_mutex);
+ }
+
+ /* Try to unplug subblocks of plugged online blocks. */
+ virtio_mem_for_each_mb_state(vm, mb_id, VIRTIO_MEM_MB_STATE_ONLINE) {
+ rc = virtio_mem_mb_unplug_any_sb_online(vm, mb_id,
+ &nb_sb);
+ if (rc || !nb_sb)
+ goto out_unlock;
+ mutex_unlock(&vm->hotplug_mutex);
+ cond_resched();
+ mutex_lock(&vm->hotplug_mutex);
+ }
+
mutex_unlock(&vm->hotplug_mutex);
- return 0;
+ return nb_sb ? -EBUSY : 0;
out_unlock:
mutex_unlock(&vm->hotplug_mutex);
return rc;
@@ -1324,7 +1451,8 @@ static void virtio_mem_run_wq(struct work_struct *work)
case -EBUSY:
/*
* The hypervisor cannot process our request right now
- * (e.g., out of memory, migrating).
+ * (e.g., out of memory, migrating) or we cannot free up
+ * any memory to unplug it (all plugged memory is busy).
*/
case -ENOMEM:
/* Out of memory, try again later. */
--
2.23.0

2019-12-12 17:14:06

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages via MEM_GOING_OFFLINE

virtio-mem wants to allow to offline memory blocks of which some parts
were unplugged (allocated via alloc_contig_range()), especially, to later
offline and remove completely unplugged memory blocks. The important part
is that PageOffline() has to remain set until the section is offline, so
these pages will never get accessed (e.g., when dumping). The pages should
not be handed back to the buddy (which would require clearing PageOffline()
and result in issues if offlining fails and the pages are suddenly in the
buddy).

Let's allow to do that by allowing to isolate any PageOffline() page
when offlining. This way, we can reach the memory hotplug notifier
MEM_GOING_OFFLINE, where the driver can signal that he is fine with
offlining this page by dropping its reference count. PageOffline() pages
with a reference count of 0 can then be skipped when offlining the
pages (like if they were free, however they are not in the buddy).

Anybody who uses PageOffline() pages and does not agree to offline them
(e.g., Hyper-V balloon, XEN balloon, VMWare balloon for 2MB pages) will not
decrement the reference count and make offlining fail when trying to
migrate such an unmovable page. So there should be no observerable change.
Same applies to balloon compaction users (movable PageOffline() pages), the
pages will simply be migrated.

Note 1: If offlining fails, a driver has to increment the reference
count again in MEM_CANCEL_OFFLINE.

Note 2: A driver that makes use of this has to be aware that re-onlining
the memory block has to be handled by hooking into onlining code
(online_page_callback_t), resetting the page PageOffline() and
not giving them to the buddy.

Cc: Andrew Morton <[email protected]>
Cc: Juergen Gross <[email protected]>
Cc: Konrad Rzeszutek Wilk <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Alexander Duyck <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: Johannes Weiner <[email protected]>
Cc: Anthony Yznaga <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Mel Gorman <[email protected]>
Cc: Mike Rapoport <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Anshuman Khandual <[email protected]>
Cc: Qian Cai <[email protected]>
Cc: Pingfan Liu <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
include/linux/page-flags.h | 10 ++++++++++
mm/memory_hotplug.c | 41 ++++++++++++++++++++++++++++----------
mm/page_alloc.c | 24 ++++++++++++++++++++++
mm/page_isolation.c | 9 +++++++++
4 files changed, 74 insertions(+), 10 deletions(-)

diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
index 1bf83c8fcaa7..ac1775082343 100644
--- a/include/linux/page-flags.h
+++ b/include/linux/page-flags.h
@@ -761,6 +761,16 @@ PAGE_TYPE_OPS(Buddy, buddy)
* not onlined when onlining the section).
* The content of these pages is effectively stale. Such pages should not
* be touched (read/write/dump/save) except by their owner.
+ *
+ * If a driver wants to allow to offline unmovable PageOffline() pages without
+ * putting them back to the buddy, it can do so via the memory notifier by
+ * decrementing the reference count in MEM_GOING_OFFLINE and incrementing the
+ * reference count in MEM_CANCEL_OFFLINE. When offlining, the PageOffline()
+ * pages (now with a reference count of zero) are treated like free pages,
+ * allowing the containing memory block to get offlined. A driver that
+ * relies on this feature is aware that re-onlining the memory block will
+ * require to re-set the pages PageOffline() and not giving them to the
+ * buddy via online_page_callback_t.
*/
PAGE_TYPE_OPS(Offline, offline)

diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index fc617ad6f035..da01453a04e6 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1231,11 +1231,15 @@ int test_pages_in_a_zone(unsigned long start_pfn, unsigned long end_pfn,

/*
* Scan pfn range [start,end) to find movable/migratable pages (LRU pages,
- * non-lru movable pages and hugepages). We scan pfn because it's much
- * easier than scanning over linked list. This function returns the pfn
- * of the first found movable page if it's found, otherwise 0.
+ * non-lru movable pages and hugepages).
+ *
+ * Returns:
+ * 0 in case a movable page is found and movable_pfn was updated.
+ * -ENOENT in case no movable page was found.
+ * -EBUSY in case a definetly unmovable page was found.
*/
-static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
+static int scan_movable_pages(unsigned long start, unsigned long end,
+ unsigned long *movable_pfn)
{
unsigned long pfn;

@@ -1247,18 +1251,29 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
continue;
page = pfn_to_page(pfn);
if (PageLRU(page))
- return pfn;
+ goto found;
if (__PageMovable(page))
- return pfn;
+ goto found;
+
+ /*
+ * Unmovable PageOffline() pages where somebody still holds
+ * a reference count (after MEM_GOING_OFFLINE) can definetly
+ * not be offlined.
+ */
+ if (PageOffline(page) && page_count(page))
+ return -EBUSY;

if (!PageHuge(page))
continue;
head = compound_head(page);
if (page_huge_active(head))
- return pfn;
+ goto found;
skip = compound_nr(head) - (page - head);
pfn += skip - 1;
}
+ return -ENOENT;
+found:
+ *movable_pfn = pfn;
return 0;
}

@@ -1528,7 +1543,8 @@ static int __ref __offline_pages(unsigned long start_pfn,
}

do {
- for (pfn = start_pfn; pfn;) {
+ pfn = start_pfn;
+ do {
if (signal_pending(current)) {
ret = -EINTR;
reason = "signal backoff";
@@ -1538,14 +1554,19 @@ static int __ref __offline_pages(unsigned long start_pfn,
cond_resched();
lru_add_drain_all();

- pfn = scan_movable_pages(pfn, end_pfn);
- if (pfn) {
+ ret = scan_movable_pages(pfn, end_pfn, &pfn);
+ if (!ret) {
/*
* TODO: fatal migration failures should bail
* out
*/
do_migrate_range(pfn, end_pfn);
}
+ } while (!ret);
+
+ if (ret != -ENOENT) {
+ reason = "unmovable page";
+ goto failed_removal_isolated;
}

/*
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 5334decc9e06..840c0bbe2d9f 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -8256,6 +8256,19 @@ bool has_unmovable_pages(struct zone *zone, struct page *page, int count,
if ((flags & MEMORY_OFFLINE) && PageHWPoison(page))
continue;

+ /*
+ * We treat all PageOffline() pages as movable when offlining
+ * to give drivers a chance to decrement their reference count
+ * in MEM_GOING_OFFLINE in order to signalize that these pages
+ * can be offlined as there are no direct references anymore.
+ * For actually unmovable PageOffline() where the driver does
+ * not support this, we will fail later when trying to actually
+ * move these pages that still have a reference count > 0.
+ * (false negatives in this function only)
+ */
+ if ((flags & MEMORY_OFFLINE) && PageOffline(page))
+ continue;
+
if (__PageMovable(page))
continue;

@@ -8683,6 +8696,17 @@ __offline_isolated_pages(unsigned long start_pfn, unsigned long end_pfn)
offlined_pages++;
continue;
}
+ /*
+ * At this point all remaining PageOffline() pages have a
+ * reference count of 0 and can simply be skipped.
+ */
+ if (PageOffline(page)) {
+ BUG_ON(page_count(page));
+ BUG_ON(PageBuddy(page));
+ pfn++;
+ offlined_pages++;
+ continue;
+ }

BUG_ON(page_count(page));
BUG_ON(!PageBuddy(page));
diff --git a/mm/page_isolation.c b/mm/page_isolation.c
index 04ee1663cdbe..43b4dabfedc8 100644
--- a/mm/page_isolation.c
+++ b/mm/page_isolation.c
@@ -170,6 +170,7 @@ __first_valid_page(unsigned long pfn, unsigned long nr_pages)
* a bit mask)
* MEMORY_OFFLINE - isolate to offline (!allocate) memory
* e.g., skip over PageHWPoison() pages
+ * and PageOffline() pages.
* REPORT_FAILURE - report details about the failure to
* isolate the range
*
@@ -278,6 +279,14 @@ __test_page_isolated_in_pageblock(unsigned long pfn, unsigned long end_pfn,
else if ((flags & MEMORY_OFFLINE) && PageHWPoison(page))
/* A HWPoisoned page cannot be also PageBuddy */
pfn++;
+ else if ((flags & MEMORY_OFFLINE) && PageOffline(page) &&
+ !page_count(page))
+ /*
+ * The responsible driver agreed to offline
+ * PageOffline() pages by dropping its reference in
+ * MEM_GOING_OFFLINE.
+ */
+ pfn++;
else
break;
}
--
2.23.0

2019-12-12 17:14:08

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 07/13] virtio-mem: Allow to offline partially unplugged memory blocks

Dropping the reference count of PageOffline() pages allows offlining
code to skip them. However, we also have to convert PG_reserved to
another flag - let's use PG_dirty - so has_unmovable_pages() will
properly handle them. PG_reserved pages get detected as unmovable right
away.

We need the flag to see if we are onlining pages the first time, or if
we allocated them via alloc_contig_range().

Properly take care of offlining code also modifying the stats and
special handling in case the driver gets unloaded.

Cc: "Michael S. Tsirkin" <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Dave Young <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Stefan Hajnoczi <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
drivers/virtio/virtio_mem.c | 64 ++++++++++++++++++++++++++++++++++++-
1 file changed, 63 insertions(+), 1 deletion(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 5a142a371222..a12a0f9c076b 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -564,6 +564,53 @@ static void virtio_mem_notify_online(struct virtio_mem *vm, unsigned long mb_id,
virtio_mem_retry(vm);
}

+static void virtio_mem_notify_going_offline(struct virtio_mem *vm,
+ unsigned long mb_id)
+{
+ const unsigned long nr_pages = PFN_DOWN(vm->subblock_size);
+ unsigned long pfn;
+ int sb_id, i;
+
+ for (sb_id = 0; sb_id < vm->nb_sb_per_mb; sb_id++) {
+ if (virtio_mem_mb_test_sb_plugged(vm, mb_id, sb_id, 1))
+ continue;
+ /*
+ * Drop our reference to the pages so the memory can get
+ * offlined and add the unplugged pages to the managed
+ * page counters (so offlining code can correctly subtract
+ * them again).
+ */
+ pfn = PFN_DOWN(virtio_mem_mb_id_to_phys(mb_id) +
+ sb_id * vm->subblock_size);
+ adjust_managed_page_count(pfn_to_page(pfn), nr_pages);
+ for (i = 0; i < nr_pages; i++)
+ page_ref_dec(pfn_to_page(pfn + i));
+ }
+}
+
+static void virtio_mem_notify_cancel_offline(struct virtio_mem *vm,
+ unsigned long mb_id)
+{
+ const unsigned long nr_pages = PFN_DOWN(vm->subblock_size);
+ unsigned long pfn;
+ int sb_id, i;
+
+ for (sb_id = 0; sb_id < vm->nb_sb_per_mb; sb_id++) {
+ if (virtio_mem_mb_test_sb_plugged(vm, mb_id, sb_id, 1))
+ continue;
+ /*
+ * Get the reference we dropped when going offline and
+ * subtract the unplugged pages from the managed page
+ * counters.
+ */
+ pfn = PFN_DOWN(virtio_mem_mb_id_to_phys(mb_id) +
+ sb_id * vm->subblock_size);
+ adjust_managed_page_count(pfn_to_page(pfn), -nr_pages);
+ for (i = 0; i < nr_pages; i++)
+ page_ref_inc(pfn_to_page(pfn + i));
+ }
+}
+
/*
* This callback will either be called synchonously from add_memory() or
* asynchronously (e.g., triggered via user space). We have to be careful
@@ -611,6 +658,7 @@ static int virtio_mem_memory_notifier_cb(struct notifier_block *nb,
break;
mutex_lock(&vm->hotplug_mutex);
vm->hotplug_active = true;
+ virtio_mem_notify_going_offline(vm, mb_id);
break;
case MEM_GOING_ONLINE:
spin_lock_irq(&vm->removal_lock);
@@ -636,6 +684,12 @@ static int virtio_mem_memory_notifier_cb(struct notifier_block *nb,
mutex_unlock(&vm->hotplug_mutex);
break;
case MEM_CANCEL_OFFLINE:
+ if (!vm->hotplug_active)
+ break;
+ virtio_mem_notify_cancel_offline(vm, mb_id);
+ vm->hotplug_active = false;
+ mutex_unlock(&vm->hotplug_mutex);
+ break;
case MEM_CANCEL_ONLINE:
if (!vm->hotplug_active)
break;
@@ -660,8 +714,11 @@ static void virtio_mem_set_fake_offline(unsigned long pfn,
struct page *page = pfn_to_page(pfn);

__SetPageOffline(page);
- if (!onlined)
+ if (!onlined) {
SetPageDirty(page);
+ /* FIXME: remove after cleanups */
+ ClearPageReserved(page);
+ }
}
}

@@ -1719,6 +1776,11 @@ static void virtio_mem_remove(struct virtio_device *vdev)
rc = virtio_mem_mb_remove(vm, mb_id);
BUG_ON(rc);
}
+ /*
+ * After we unregistered our callbacks, user space can no longer
+ * offline partially plugged online memory blocks. No need to worry
+ * about them.
+ */

/* unregister callbacks */
unregister_virtio_mem_device(vm);
--
2.23.0

2019-12-12 17:14:13

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 08/13] mm/memory_hotplug: Introduce offline_and_remove_memory()

virtio-mem wants to offline and remove a memory block once it unplugged
all subblocks (e.g., using alloc_contig_range()). Let's provide
an interface to do that from a driver. virtio-mem already supports to
offline partially unplugged memory blocks. Offlining a fully unplugged
memory block will not require to migrate any pages. All unplugged
subblocks are PageOffline() and have a reference count of 0 - so
offlining code will simply skip them.

All we need an interface to trigger the "offlining" and the removing in a
single operation - to make sure the memory block cannot get onlined by
user space again before it gets removed.

To keep things simple, allow to only work on a single memory block.

Cc: Andrew Morton <[email protected]>
Cc: David Hildenbrand <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Wei Yang <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Qian Cai <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
include/linux/memory_hotplug.h | 1 +
mm/memory_hotplug.c | 35 ++++++++++++++++++++++++++++++++++
2 files changed, 36 insertions(+)

diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h
index ba0dca6aac6e..586f5c59c291 100644
--- a/include/linux/memory_hotplug.h
+++ b/include/linux/memory_hotplug.h
@@ -310,6 +310,7 @@ extern void try_offline_node(int nid);
extern int offline_pages(unsigned long start_pfn, unsigned long nr_pages);
extern int remove_memory(int nid, u64 start, u64 size);
extern void __remove_memory(int nid, u64 start, u64 size);
+extern int offline_and_remove_memory(int nid, u64 start, u64 size);

#else
static inline bool is_mem_section_removable(unsigned long pfn,
diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
index da01453a04e6..d04369e6d3cc 100644
--- a/mm/memory_hotplug.c
+++ b/mm/memory_hotplug.c
@@ -1825,4 +1825,39 @@ int remove_memory(int nid, u64 start, u64 size)
return rc;
}
EXPORT_SYMBOL_GPL(remove_memory);
+
+/*
+ * Try to offline and remove a memory block. Might take a long time to
+ * finish in case memory is still in use. Primarily useful for memory devices
+ * that logically unplugged all memory (so it's no longer in use) and want to
+ * offline + remove the memory block.
+ */
+int offline_and_remove_memory(int nid, u64 start, u64 size)
+{
+ struct memory_block *mem;
+ int rc = -EINVAL;
+
+ if (!IS_ALIGNED(start, memory_block_size_bytes()) ||
+ size != memory_block_size_bytes())
+ return rc;
+
+ lock_device_hotplug();
+ mem = find_memory_block(__pfn_to_section(PFN_DOWN(start)));
+ if (mem)
+ rc = device_offline(&mem->dev);
+ /* Ignore if the device is already offline. */
+ if (rc > 0)
+ rc = 0;
+
+ /*
+ * In case we succeeded to offline the memory block, remove it.
+ * This cannot fail as it cannot get onlined in the meantime.
+ */
+ if (!rc && try_remove_memory(nid, start, size))
+ BUG();
+ unlock_device_hotplug();
+
+ return rc;
+}
+EXPORT_SYMBOL_GPL(offline_and_remove_memory);
#endif /* CONFIG_MEMORY_HOTREMOVE */
--
2.23.0

2019-12-12 17:14:44

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 09/13] virtio-mem: Offline and remove completely unplugged memory blocks

Let's offline+remove memory blocks once all subblocks are unplugged. We
can use the new Linux MM interface for that. As no memory is in use
anymore, this shouldn't take a long time and shouldn't fail. There might
be corner cases where the offlining could still fail (especially, if
another notifier NACKs the offlining request).

Cc: "Michael S. Tsirkin" <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Dave Young <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Stefan Hajnoczi <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
drivers/virtio/virtio_mem.c | 47 +++++++++++++++++++++++++++++++++----
1 file changed, 43 insertions(+), 4 deletions(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index a12a0f9c076b..807d4e393427 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -436,6 +436,28 @@ static int virtio_mem_mb_remove(struct virtio_mem *vm, unsigned long mb_id)
return remove_memory(nid, addr, memory_block_size_bytes());
}

+/*
+ * Try to offline and remove a memory block from Linux.
+ *
+ * Must not be called with the vm->hotplug_mutex held (possible deadlock with
+ * onlining code).
+ *
+ * Will not modify the state of the memory block.
+ */
+static int virtio_mem_mb_offline_and_remove(struct virtio_mem *vm,
+ unsigned long mb_id)
+{
+ const uint64_t addr = virtio_mem_mb_id_to_phys(mb_id);
+ int nid = vm->nid;
+
+ if (nid == NUMA_NO_NODE)
+ nid = memory_add_physaddr_to_nid(addr);
+
+ dev_dbg(&vm->vdev->dev, "offlining and removing memory block: %lu\n",
+ mb_id);
+ return offline_and_remove_memory(nid, addr, memory_block_size_bytes());
+}
+
/*
* Trigger the workqueue so the device can perform its magic.
*/
@@ -529,7 +551,13 @@ static void virtio_mem_notify_offline(struct virtio_mem *vm,
break;
}

- /* trigger the workqueue, maybe we can now unplug memory. */
+ /*
+ * Trigger the workqueue, maybe we can now unplug memory. Also,
+ * when we offline and remove a memory block, this will re-trigger
+ * us immediately - which is often nice because the removal of
+ * the memory block (e.g., memmap) might have freed up memory
+ * on other memory blocks we manage.
+ */
virtio_mem_retry(vm);
}

@@ -1275,7 +1303,8 @@ static int virtio_mem_mb_unplug_any_sb_offline(struct virtio_mem *vm,
* Unplug the desired number of plugged subblocks of an online memory block.
* Will skip subblock that are busy.
*
- * Will modify the state of the memory block.
+ * Will modify the state of the memory block. Might temporarily drop the
+ * hotplug_mutex.
*
* Note: Can fail after some subblocks were successfully unplugged. Can
* return 0 even if subblocks were busy and could not get unplugged.
@@ -1331,9 +1360,19 @@ static int virtio_mem_mb_unplug_any_sb_online(struct virtio_mem *vm,
}

/*
- * TODO: Once all subblocks of a memory block were unplugged, we want
- * to offline the memory block and remove it.
+ * Once all subblocks of a memory block were unplugged, offline and
+ * remove it. This will usually not fail, as no memory is in use
+ * anymore - however some other notifiers might NACK the request.
*/
+ if (virtio_mem_mb_test_sb_unplugged(vm, mb_id, 0, vm->nb_sb_per_mb)) {
+ mutex_unlock(&vm->hotplug_mutex);
+ rc = virtio_mem_mb_offline_and_remove(vm, mb_id);
+ mutex_lock(&vm->hotplug_mutex);
+ if (!rc)
+ virtio_mem_mb_set_state(vm, mb_id,
+ VIRTIO_MEM_MB_STATE_UNUSED);
+ }
+
return 0;
}

--
2.23.0

2019-12-12 17:14:57

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 10/13] virtio-mem: Better retry handling

Let's start with a retry interval of 30 seconds and double the time until
we reach 30 minutes, in case we keep getting errors. Reset the retry
interval in case we succeeded.

Cc: "Michael S. Tsirkin" <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Dave Young <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Stefan Hajnoczi <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
drivers/virtio/virtio_mem.c | 12 +++++++++---
1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 807d4e393427..3a57434f92ed 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -137,7 +137,9 @@ struct virtio_mem {

/* Timer for retrying to plug/unplug memory. */
struct hrtimer retry_timer;
-#define VIRTIO_MEM_RETRY_TIMER_MS 30000
+ unsigned int retry_timer_ms;
+#define VIRTIO_MEM_RETRY_TIMER_MIN_MS 30000
+#define VIRTIO_MEM_RETRY_TIMER_MAX_MS 1800000

/* Memory notifier (online/offline events). */
struct notifier_block memory_notifier;
@@ -1537,6 +1539,7 @@ static void virtio_mem_run_wq(struct work_struct *work)

switch (rc) {
case 0:
+ vm->retry_timer_ms = VIRTIO_MEM_RETRY_TIMER_MIN_MS;
break;
case -ENOSPC:
/*
@@ -1552,8 +1555,7 @@ static void virtio_mem_run_wq(struct work_struct *work)
*/
case -ENOMEM:
/* Out of memory, try again later. */
- hrtimer_start(&vm->retry_timer,
- ms_to_ktime(VIRTIO_MEM_RETRY_TIMER_MS),
+ hrtimer_start(&vm->retry_timer, ms_to_ktime(vm->retry_timer_ms),
HRTIMER_MODE_REL);
break;
case -EAGAIN:
@@ -1573,6 +1575,9 @@ static enum hrtimer_restart virtio_mem_timer_expired(struct hrtimer *timer)
retry_timer);

virtio_mem_retry(vm);
+ /* Racy (with reset in virtio_mem_run_wq), we ignore that for now. */
+ vm->retry_timer_ms = min_t(unsigned int, vm->retry_timer_ms * 2,
+ VIRTIO_MEM_RETRY_TIMER_MAX_MS);
return HRTIMER_NORESTART;
}

@@ -1746,6 +1751,7 @@ static int virtio_mem_probe(struct virtio_device *vdev)
spin_lock_init(&vm->removal_lock);
hrtimer_init(&vm->retry_timer, CLOCK_MONOTONIC, HRTIMER_MODE_REL);
vm->retry_timer.function = virtio_mem_timer_expired;
+ vm->retry_timer_ms = VIRTIO_MEM_RETRY_TIMER_MIN_MS;

/* register the virtqueue */
rc = virtio_mem_init_vq(vm);
--
2.23.0

2019-12-12 17:15:35

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 11/13] mm/vmscan: Move count_vm_event(DROP_SLAB) into drop_slab()

Let's count within the function itself, so every invocation (of future
users) will be counted.

Cc: Alexander Viro <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: [email protected]
Signed-off-by: David Hildenbrand <[email protected]>
---
fs/drop_caches.c | 4 +---
mm/vmscan.c | 1 +
2 files changed, 2 insertions(+), 3 deletions(-)

diff --git a/fs/drop_caches.c b/fs/drop_caches.c
index d31b6c72b476..a042da782fcd 100644
--- a/fs/drop_caches.c
+++ b/fs/drop_caches.c
@@ -61,10 +61,8 @@ int drop_caches_sysctl_handler(struct ctl_table *table, int write,
iterate_supers(drop_pagecache_sb, NULL);
count_vm_event(DROP_PAGECACHE);
}
- if (sysctl_drop_caches & 2) {
+ if (sysctl_drop_caches & 2)
drop_slab();
- count_vm_event(DROP_SLAB);
- }
if (!stfu) {
pr_info("%s (%d): drop_caches: %d\n",
current->comm, task_pid_nr(current),
diff --git a/mm/vmscan.c b/mm/vmscan.c
index 5a6445e86328..c3e53502a84a 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -726,6 +726,7 @@ void drop_slab(void)

for_each_online_node(nid)
drop_slab_node(nid);
+ count_vm_event(DROP_SLAB);
}

static inline int is_page_cache_freeable(struct page *page)
--
2.23.0

2019-12-12 17:15:48

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 13/13] virtio-mem: Drop slab objects when unplug continues to fail

Start dropping slab objects after 30 minutes and repeat every 30 minutes
in case we can't unplug more memory using alloc_contig_range().
Log messages and make it configurable. Enable dropping slab objects as
default (especially, reclaimable slab objects that are not movable).

In the future, we might want to implement+use drop_slab_range(), which
will also come in handy for other users (e.g., offlining, gigantic huge
pages).

Cc: Alexander Viro <[email protected]>
Cc: Andrew Morton <[email protected]>
Cc: Michal Hocko <[email protected]>
Cc: "Michael S. Tsirkin" <[email protected]>
Cc: Jason Wang <[email protected]>
Cc: Oscar Salvador <[email protected]>
Cc: Igor Mammedov <[email protected]>
Cc: Dave Young <[email protected]>
Cc: Dan Williams <[email protected]>
Cc: Pavel Tatashin <[email protected]>
Cc: Stefan Hajnoczi <[email protected]>
Cc: Vlastimil Babka <[email protected]>
Cc: [email protected]
Signed-off-by: David Hildenbrand <[email protected]>
---
drivers/virtio/virtio_mem.c | 30 ++++++++++++++++++++++++++++++
1 file changed, 30 insertions(+)

diff --git a/drivers/virtio/virtio_mem.c b/drivers/virtio/virtio_mem.c
index 3a57434f92ed..8f25f7453a08 100644
--- a/drivers/virtio/virtio_mem.c
+++ b/drivers/virtio/virtio_mem.c
@@ -25,6 +25,11 @@ static bool unplug_online = true;
module_param(unplug_online, bool, 0644);
MODULE_PARM_DESC(unplug_online, "Try to unplug online memory");

+static bool drop_slab_objects = true;
+module_param(drop_slab_objects, bool, 0644);
+MODULE_PARM_DESC(drop_slab_objects,
+ "Drop slab objects when unplug continues to fail");
+
enum virtio_mem_mb_state {
/* Unplugged, not added to Linux. Can be reused later. */
VIRTIO_MEM_MB_STATE_UNUSED = 0,
@@ -1384,6 +1389,7 @@ static int virtio_mem_mb_unplug_any_sb_online(struct virtio_mem *vm,
static int virtio_mem_unplug_request(struct virtio_mem *vm, uint64_t diff)
{
uint64_t nb_sb = diff / vm->subblock_size;
+ bool retried = false;
unsigned long mb_id;
int rc;

@@ -1421,6 +1427,7 @@ static int virtio_mem_unplug_request(struct virtio_mem *vm, uint64_t diff)
return 0;
}

+retry_locked:
/* Try to unplug subblocks of partially plugged online blocks. */
virtio_mem_for_each_mb_state(vm, mb_id,
VIRTIO_MEM_MB_STATE_ONLINE_PARTIAL) {
@@ -1445,6 +1452,29 @@ static int virtio_mem_unplug_request(struct virtio_mem *vm, uint64_t diff)
}

mutex_unlock(&vm->hotplug_mutex);
+
+ /*
+ * If we can't unplug the requested amount of memory for a long time,
+ * start freeing up memory in caches. This might harm performance,
+ * is configurable, and we log a message. Retry imemdiately a second
+ * time - then wait another VIRTIO_MEM_RETRY_TIMER_MAX_MS.
+ */
+ if (nb_sb && !retried && drop_slab_objects &&
+ vm->retry_timer_ms == VIRTIO_MEM_RETRY_TIMER_MAX_MS) {
+ if (vm->nid == NUMA_NO_NODE) {
+ dev_info(&vm->vdev->dev, "dropping all slab objects\n");
+ drop_slab();
+ } else {
+ dev_info(&vm->vdev->dev,
+ "dropping all slab objects on node=%d\n",
+ vm->nid);
+ drop_slab_node(vm->nid);
+ }
+ retried = true;
+ mutex_lock(&vm->hotplug_mutex);
+ goto retry_locked;
+ }
+
return nb_sb ? -EBUSY : 0;
out_unlock:
mutex_unlock(&vm->hotplug_mutex);
--
2.23.0

2019-12-12 17:15:49

by David Hildenbrand

[permalink] [raw]
Subject: [PATCH RFC v4 12/13] mm/vmscan: Export drop_slab() and drop_slab_node()

We already have a way to trigger reclaiming of all reclaimable slab objects
from user space (echo 2 > /proc/sys/vm/drop_caches). Let's allow drivers
to also trigger this when they really want to make progress and know what
they are doing.

virtio-mem wants to use these functions when it failed to unplug memory
for quite some time (e.g., after 30 minutes). It will then try to
free up reclaimable objects by dropping the slab caches every now and
then (e.g., every 30 minutes) as long as necessary. There will be a way to
disable this feature and info messages will be logged.

In the future, we want to have a drop_slab_range() functionality
instead. Memory offlining code has similar demands and also other
alloc_contig_range() users (e.g., gigantic pages) could make good use of
this feature. Adding it, however, requires more work/thought.

Cc: Andrew Morton <[email protected]>
Cc: Michal Hocko <[email protected]>
Signed-off-by: David Hildenbrand <[email protected]>
---
include/linux/mm.h | 4 ++--
mm/vmscan.c | 2 ++
2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/include/linux/mm.h b/include/linux/mm.h
index 64799c5cb39f..483300f58be8 100644
--- a/include/linux/mm.h
+++ b/include/linux/mm.h
@@ -2706,8 +2706,8 @@ int drop_caches_sysctl_handler(struct ctl_table *, int,
void __user *, size_t *, loff_t *);
#endif

-void drop_slab(void);
-void drop_slab_node(int nid);
+extern void drop_slab(void);
+extern void drop_slab_node(int nid);

#ifndef CONFIG_MMU
#define randomize_va_space 0
diff --git a/mm/vmscan.c b/mm/vmscan.c
index c3e53502a84a..4e1cdaaec5e6 100644
--- a/mm/vmscan.c
+++ b/mm/vmscan.c
@@ -719,6 +719,7 @@ void drop_slab_node(int nid)
} while ((memcg = mem_cgroup_iter(NULL, memcg, NULL)) != NULL);
} while (freed > 10);
}
+EXPORT_SYMBOL(drop_slab_node);

void drop_slab(void)
{
@@ -728,6 +729,7 @@ void drop_slab(void)
drop_slab_node(nid);
count_vm_event(DROP_SLAB);
}
+EXPORT_SYMBOL(drop_slab);

static inline int is_page_cache_freeable(struct page *page)
{
--
2.23.0

2019-12-13 20:16:32

by Konrad Rzeszutek Wilk

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On Thu, Dec 12, 2019 at 06:11:24PM +0100, David Hildenbrand wrote:
> This series is based on latest linux-next. The patches are located at:
> https://github.com/davidhildenbrand/linux.git virtio-mem-rfc-v4
Heya!

Would there be by any chance a virtio-spec git tree somewhere?

..snip..
> --------------------------------------------------------------------------
> 5. Future work
> --------------------------------------------------------------------------
>
> The separate patches contain a lot of future work items. One of the next
> steps is to make memory unplug more likely to succeed - currently, there
> are no guarantees on how much memory can get unplugged again. I have


Or perhaps tell the caller why we can't and let them sort it out?
For example: "Application XYZ is mlocked. Can't offload'.


2019-12-24 07:00:10

by Hui Zhu

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

Hi David,

Thanks for your work.

I Got following build fail if X86_64_ACPI_NUMA is n with rfc3 and rfc4:
make -j8 bzImage
GEN Makefile
DESCEND objtool
CALL /home/teawater/kernel/linux-upstream3/scripts/atomic/check-atomics.sh
CALL /home/teawater/kernel/linux-upstream3/scripts/checksyscalls.sh
CHK include/generated/compile.h
CC drivers/virtio/virtio_mem.o
/home/teawater/kernel/linux-upstream3/drivers/virtio/virtio_mem.c: In function ??virtio_mem_translate_node_id??:
/home/teawater/kernel/linux-upstream3/drivers/virtio/virtio_mem.c:478:10: error: implicit declaration of function ??pxm_to_node?? [-Werror=implicit-function-declaration]
node = pxm_to_node(node_id);
^~~~~~~~~~~
cc1: some warnings being treated as errors
/home/teawater/kernel/linux-upstream3/scripts/Makefile.build:265: recipe for target 'drivers/virtio/virtio_mem.o' failed
make[3]: *** [drivers/virtio/virtio_mem.o] Error 1
/home/teawater/kernel/linux-upstream3/scripts/Makefile.build:503: recipe for target 'drivers/virtio' failed
make[2]: *** [drivers/virtio] Error 2
/home/teawater/kernel/linux-upstream3/Makefile:1649: recipe for target 'drivers' failed
make[1]: *** [drivers] Error 2
/home/teawater/kernel/linux-upstream3/Makefile:179: recipe for target 'sub-make' failed
make: *** [sub-make] Error 2

Best,
Hui

> ?? 2019??12??13?գ?01:11??David Hildenbrand <[email protected]> д????
>
> This series is based on latest linux-next. The patches are located at:
> https://github.com/davidhildenbrand/linux.git virtio-mem-rfc-v4
>
> The basic idea of virtio-mem is to provide a flexible,
> cross-architecture memory hot(un)plug solution that avoids many limitations
> imposed by existing technologies, architectures, and interfaces. More
> details can be found below and in linked material.
>
> This RFC is limited to x86-64, however, should theoretically work on any
> architecture that supports virtio and implements memory hot(un)plug under
> Linux - like s390x, powerpc64 and arm64. On x86-64, it is currently
> possible to add/remove memory to the system in >= 4MB granularity.
> Memory hotplug works very reliably. For memory unplug, there are no
> guarantees how much memory can actually get unplugged, it depends on the
> setup (especially: fragmentation of (unmovable) memory). I have plans to
> improve that in the future.
>
> --------------------------------------------------------------------------
> 1. virtio-mem
> --------------------------------------------------------------------------
>
> The basic idea behind virtio-mem was presented at KVM Forum 2018. The
> slides can be found at [1]. The previous RFC can be found at [2]. The
> first RFC can be found at [3]. However, the concept evolved over time. The
> KVM Forum slides roughly match the current design.
>
> Patch #2 ("virtio-mem: Paravirtualized memory hotplug") contains quite some
> information, especially in "include/uapi/linux/virtio_mem.h":
>
> Each virtio-mem device manages a dedicated region in physical address
> space. Each device can belong to a single NUMA node, multiple devices
> for a single NUMA node are possible. A virtio-mem device is like a
> "resizable DIMM" consisting of small memory blocks that can be plugged
> or unplugged. The device driver is responsible for (un)plugging memory
> blocks on demand.
>
> Virtio-mem devices can only operate on their assigned memory region in
> order to (un)plug memory. A device cannot (un)plug memory belonging to
> other devices.
>
> The "region_size" corresponds to the maximum amount of memory that can
> be provided by a device. The "size" corresponds to the amount of memory
> that is currently plugged. "requested_size" corresponds to a request
> from the device to the device driver to (un)plug blocks. The
> device driver should try to (un)plug blocks in order to reach the
> "requested_size". It is impossible to plug more memory than requested.
>
> The "usable_region_size" represents the memory region that can actually
> be used to (un)plug memory. It is always at least as big as the
> "requested_size" and will grow dynamically. It will only shrink when
> explicitly triggered (VIRTIO_MEM_REQ_UNPLUG).
>
> Memory in the usable region can usually be read, however, there are no
> guarantees. It can happen that the device cannot process a request,
> because it is busy. The device driver has to retry later.
>
> Usually, during system resets all memory will get unplugged, so the
> device driver can start with a clean state. However, in specific
> scenarios (if the device is busy) it can happen that the device still
> has memory plugged. The device driver can request to unplug all memory
> (VIRTIO_MEM_REQ_UNPLUG) - which might take a while to succeed if the
> device is busy.
>
> --------------------------------------------------------------------------
> 2. Linux Implementation
> --------------------------------------------------------------------------
>
> This RFC reuses quite some existing MM infrastructure, however, has to
> expose some additional functionality.
>
> Memory blocks (e.g., 128MB) are added/removed on demand. Within these
> memory blocks, subblocks (e.g., 4MB) are plugged/unplugged. The sizes
> depend on the target architecture, MAX_ORDER + pageblock_order, and
> the block size of a virtio-mem device.
>
> add_memory()/try_remove_memory() is used to add/remove memory blocks.
> virtio-mem will not online memory blocks itself. This has to be done by
> user space, or configured into the kernel
> (CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE). virtio-mem will only unplug memory
> that was online to the ZONE_NORMAL. Memory is suggested to be onlined to
> the ZONE_NORMAL for now.
>
> The memory hotplug notifier is used to properly synchronize against
> onlining/offlining of memory blocks and to track the states of memory
> blocks (including the zone memory blocks are onlined to).
>
> The set_online_page() callback is used to keep unplugged subblocks
> of a memory block fake-offline when onlining the memory block.
> generic_online_page() is used to fake-online plugged subblocks. This
> handling is similar to the Hyper-V balloon driver.
>
> PG_offline is used to mark unplugged subblocks as offline, so e.g.,
> dumping tools (makedumpfile) will skip these pages. This is similar to
> other balloon drivers like virtio-balloon and Hyper-V.
>
> Memory offlining code is extended to allow drivers to drop their reference
> to PG_offline pages when MEM_GOING_OFFLINE, so these pages can be skipped
> when offlining memory blocks. This allows to offline memory blocks that
> have partially unplugged (allocated e.g., via alloc_contig_range())
> subblocks - or are completely unplugged.
>
> alloc_contig_range()/free_contig_range() [now exposed] is used to
> unplug/plug subblocks of memory blocks the are already exposed to Linux.
>
> offline_and_remove_memory() [new] is used to offline a fully unplugged
> memory block and remove it from Linux.
>
>
> A lot of additional information can be found in the separate patches and
> as comments in the code itself.
>
> --------------------------------------------------------------------------
> 3. Changes RFC v2 -> v3
> --------------------------------------------------------------------------
>
> A lot of things changed, especially also on the QEMU + virtio side. The
> biggest changes on the Linux driver side are:
> - Onlining/offlining of subblocks is now emulated on top of memory blocks.
> set_online_page()+alloc_contig_range()+free_contig_range() is now used
> for that. Core MM does not have to be modified and will continue to
> online/offline full memory blocks.
> - Onlining/offlining of memory blocks is no longer performed by virtio-mem.
> - Pg_offline is upstream and can be used. It is also used to allow
> offlining of partially unplugged memory blocks.
> - Memory block states + subblocks are now tracked more space-efficient.
> - Proper kexec(), kdump(), driver unload, driver reload, ZONE_MOVABLE, ...
> handling.
>
> --------------------------------------------------------------------------
> 4. Changes RFC v3 -> RFC v4
> --------------------------------------------------------------------------
>
> Only minor things changed, especially nothing on the QEMU + virtio side.
> Interresting changes on the Linux driver side are:
> - "mm: Allow to offline unmovable PageOffline() pages via
> MEM_GOING_OFFLINE"
> -- Rework to Michals suggestion (allow to isolate all PageOffline() pages
> by skipping all PageOffline() pages in has_unmovable_pages(). Fail
> offlining later if the pages cannot be offlined/migrated).
> - "virtio-mem: Allow to offline partially unplugged memory blocks"
> -- Adapt to Michals suggestion on core-mm part.
> - "virtio-mem: Better retry handling"
> -- Optimize retry intervals
> - "virtio-mem: Drop slab objects when unplug continues to fail"
> -- Call drop_slab()/drop_slab_node() when unplug keeps failing for a longer
> time.
> - Multiple cleanups and fixes.
>
> --------------------------------------------------------------------------
> 5. Future work
> --------------------------------------------------------------------------
>
> The separate patches contain a lot of future work items. One of the next
> steps is to make memory unplug more likely to succeed - currently, there
> are no guarantees on how much memory can get unplugged again. I have
> various ideas on how to limit fragmentation of all memory blocks that
> virtio-mem added.
>
> Memory hotplug:
> - Reduce the amount of memory resources if that turnes out to be an
> issue. Or try to speed up relevant code paths to deal with many
> resources.
> - Allocate the vmemmap from the added memory. Makes hotplug more likely
> to succeed, the vmemmap is stored on the same NUMA node and that
> unmovable memory will later not hinder unplug.
>
> Memory hotunplug:
> - Performance improvements:
> -- Sense (lockless) if it make sense to try alloc_contig_range() at all
> before directly trying to isolate and taking locks.
> -- Try to unplug bigger chunks if possible first.
> -- Identify free areas first, that don't have to be evacuated.
> - Make unplug more likely to succeed:
> -- There are various idea to limit fragmentation on memory block
> granularity. (e.g., ZONE_PREFER_MOVABLE and smart balancing)
> -- Allocate memmap from added memory. This way, less unmovable data can
> end up on the memory blocks.
> - OOM handling, e.g., via an OOM handler.
> - Defragmentation
> -- Will require a new virtio-mem CMD to exchange plugged<->unplugged blocks
>
> --------------------------------------------------------------------------
> 6. Example Usage
> --------------------------------------------------------------------------
>
> A very basic QEMU prototype (kept updated) is available at:
> https://github.com/davidhildenbrand/qemu.git virtio-mem
>
> It lacks various features, however, works to test the guest driver side:
> - No support for resizable memory regions / memory backends yet
> - No protection of unplugged memory (esp., userfaultfd-wp) yet
> - No dump/migration/XXX optimizations to skip unplugged memory (and avoid
> touching it)
>
> Start QEMU with two virtio-mem devices (one per NUMA node):
> $ qemu-system-x86_64 -m 4G,maxmem=20G \
> -smp sockets=2,cores=2 \
> -numa node,nodeid=0,cpus=0-1 -numa node,nodeid=1,cpus=2-3 \
> [...]
> -object memory-backend-ram,id=mem0,size=8G \
> -device virtio-mem-pci,id=vm0,memdev=mem0,node=0,requested-size=128M \
> -object memory-backend-ram,id=mem1,size=8G \
> -device virtio-mem-pci,id=vm1,memdev=mem1,node=1,requested-size=80M
>
> Query the configuration:
> QEMU 4.1.95 monitor - type 'help' for more information
> (qemu) info memory-devices
> Memory device [virtio-mem]: "vm0"
> memaddr: 0x140000000
> node: 0
> requested-size: 134217728
> size: 134217728
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem0
> Memory device [virtio-mem]: "vm1"
> memaddr: 0x340000000
> node: 1
> requested-size: 83886080
> size: 83886080
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem1
>
> Add some memory to node 1:
> QEMU 4.1.95 monitor - type 'help' for more information
> (qemu) qom-set vm1 requested-size 1G
>
> Remove some memory from node 0:
> QEMU 4.1.95 monitor - type 'help' for more information
> (qemu) qom-set vm0 requested-size 64M
>
> Query the configuration again:
> (qemu) info memory-devices
> Memory device [virtio-mem]: "vm0"
> memaddr: 0x140000000
> node: 0
> requested-size: 67108864
> size: 67108864
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem0
> Memory device [virtio-mem]: "vm1"
> memaddr: 0x340000000
> node: 1
> requested-size: 1073741824
> size: 1073741824
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem1
>
> --------------------------------------------------------------------------
> 7. Q/A
> --------------------------------------------------------------------------
>
> Q: Why add/remove parts ("subblocks") of memory blocks/sections?
> A: Flexibility (section size depends on the architecture) - e.g., some
> architectures have a section size of 2GB. Also, the memory block size
> is variable (e.g., on x86-64). I want to avoid any such restrictions.
> Some use cases want to add/remove memory in smaller granularities to a
> VM (e.g., the Hyper-V balloon also implements this) - especially smaller
> VMs like used for kata containers. Also, on memory unplug, it is more
> reliable to free-up and unplug multiple small chunks instead
> of one big chunk. E.g., if one page of a DIMM is either unmovable or
> pinned, the DIMM can't get unplugged. This approach is basically a
> compromise between DIMM-based memory hot(un)plug and balloon
> inflation/deflation, which works mostly on page granularity.
>
> Q: Why care about memory blocks?
> A: They are the way to tell user space about new memory. This way,
> memory can get onlined/offlined by user space. Also, e.g., kdump
> relies on udev events to reload kexec when memory blocks are
> onlined/offlined. Memory blocks are the "real" memory hot(un)plug
> granularity. Everything that's smaller has to be emulated "on top".
>
> Q: Won't memory unplug of subblocks fragment memory?
> A: Yes and no. Unplugging e.g., >=4MB subblocks on x86-64 will not really
> fragment memory like unplugging random pages like a balloon driver does.
> Buddy merging will not be limited. However, any allocation that requires
> bigger consecutive memory chunks (e.g., gigantic pages) might observe
> the fragmentation. Possible solutions: Allocate gigantic huge pages
> before unplugging memory, don't unplug memory, combine virtio-mem with
> DIMM based memory or bigger initial memory. Remember, a virtio-mem
> device will only unplug on the memory range it manages, not on other
> DIMMs. Unplug of single memory blocks will result in similar
> fragmentation in respect to gigantic huge pages. I ahve plans for a
> virtio-mem defragmentation feature in the future.
>
> Q: How reliable is memory unplug?
> A: There are no guarantees on how much memory can get unplugged
> again. However, it is more likely to find 4MB chunks to unplug than
> e.g., 128MB chunks. If memory is terribly fragmented, there is nothing
> we can do - for now. I consider memory hotplug the first primary use
> of virtio-mem. Memory unplug might usually work, but we want to improve
> the performance and the amount of memory we can actually unplug later.
>
> Q: Why not unplug from the ZONE_MOVABLE?
> A: Unplugged memory chunks are unmovable. Unmovable data must not end up
> on the ZONE_MOVABLE - similar to gigantic pages - they will never be
> allocated from ZONE_MOVABLE. virtio-mem added memory can be onlined
> to the ZONE_MOVABLE, but subblocks will not get unplugged from it.
>
> Q: How big should the initial (!virtio-mem) memory of a VM be?
> A: virtio-mem memory will not go to the DMA zones. So to avoid running out
> of DMA memory, I suggest something like 2-3GB on x86-64. But many
> VMs can most probably deal with less DMA memory - depends on the use
> case.
>
> [1] https://events.linuxfoundation.org/wp-content/uploads/2017/12/virtio-mem-Paravirtualized-Memory-David-Hildenbrand-Red-Hat-1.pdf
> [2] https://lkml.kernel.org/r/[email protected]
> [3] https://lkml.kernel.org/r/[email protected]
>
> Cc: Sebastien Boeuf <[email protected]>
> Cc: Samuel Ortiz <[email protected]>
> Cc: Robert Bradford <[email protected]>
> Cc: Luiz Capitulino <[email protected]>
>
> David Hildenbrand (13):
> ACPI: NUMA: export pxm_to_node
> virtio-mem: Paravirtualized memory hotplug
> virtio-mem: Paravirtualized memory hotunplug part 1
> mm: Export alloc_contig_range() / free_contig_range()
> virtio-mem: Paravirtualized memory hotunplug part 2
> mm: Allow to offline unmovable PageOffline() pages via
> MEM_GOING_OFFLINE
> virtio-mem: Allow to offline partially unplugged memory blocks
> mm/memory_hotplug: Introduce offline_and_remove_memory()
> virtio-mem: Offline and remove completely unplugged memory blocks
> virtio-mem: Better retry handling
> mm/vmscan: Move count_vm_event(DROP_SLAB) into drop_slab()
> mm/vmscan: Export drop_slab() and drop_slab_node()
> virtio-mem: Drop slab objects when unplug continues to fail
>
> drivers/acpi/numa/srat.c | 1 +
> drivers/virtio/Kconfig | 18 +
> drivers/virtio/Makefile | 1 +
> drivers/virtio/virtio_mem.c | 1939 +++++++++++++++++++++++++++++++
> fs/drop_caches.c | 4 +-
> include/linux/memory_hotplug.h | 1 +
> include/linux/mm.h | 4 +-
> include/linux/page-flags.h | 10 +
> include/uapi/linux/virtio_ids.h | 1 +
> include/uapi/linux/virtio_mem.h | 204 ++++
> mm/memory_hotplug.c | 76 +-
> mm/page_alloc.c | 26 +
> mm/page_isolation.c | 9 +
> mm/vmscan.c | 3 +
> 14 files changed, 2282 insertions(+), 15 deletions(-)
> create mode 100644 drivers/virtio/virtio_mem.c
> create mode 100644 include/uapi/linux/virtio_mem.h
>
> --
> 2.23.0

2019-12-24 09:29:29

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory



> Am 24.12.2019 um 08:04 schrieb teawater <[email protected]>:
>
> Hi David,
>
> Thanks for your work.
>
> I Got following build fail if X86_64_ACPI_NUMA is n with rfc3 and rfc4:
> make -j8 bzImage
> GEN Makefile
> DESCEND objtool
> CALL /home/teawater/kernel/linux-upstream3/scripts/atomic/check-atomics.sh
> CALL /home/teawater/kernel/linux-upstream3/scripts/checksyscalls.sh
> CHK include/generated/compile.h
> CC drivers/virtio/virtio_mem.o
> /home/teawater/kernel/linux-upstream3/drivers/virtio/virtio_mem.c: In function ‘virtio_mem_translate_node_id’:
> /home/teawater/kernel/linux-upstream3/drivers/virtio/virtio_mem.c:478:10: error: implicit declaration of function ‘pxm_to_node’ [-Werror=implicit-function-declaration]
> node = pxm_to_node(node_id);
> ^~~~~~~~~~~
> cc1: some warnings being treated as errors
> /home/teawater/kernel/linux-upstream3/scripts/Makefile.build:265: recipe for target 'drivers/virtio/virtio_mem.o' failed
> make[3]: *** [drivers/virtio/virtio_mem.o] Error 1
> /home/teawater/kernel/linux-upstream3/scripts/Makefile.build:503: recipe for target 'drivers/virtio' failed
> make[2]: *** [drivers/virtio] Error 2
> /home/teawater/kernel/linux-upstream3/Makefile:1649: recipe for target 'drivers' failed
> make[1]: *** [drivers] Error 2
> /home/teawater/kernel/linux-upstream3/Makefile:179: recipe for target 'sub-make' failed
> make: *** [sub-make] Error 2
>

Thanks Hui,

So it has to be wrapped in an ifdef, thanks!

Cheers!

2020-01-09 16:01:49

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On 12.12.19 18:11, David Hildenbrand wrote:
> This series is based on latest linux-next. The patches are located at:
> https://github.com/davidhildenbrand/linux.git virtio-mem-rfc-v4
>
> The basic idea of virtio-mem is to provide a flexible,
> cross-architecture memory hot(un)plug solution that avoids many limitations
> imposed by existing technologies, architectures, and interfaces. More
> details can be found below and in linked material.
>
> This RFC is limited to x86-64, however, should theoretically work on any
> architecture that supports virtio and implements memory hot(un)plug under
> Linux - like s390x, powerpc64 and arm64. On x86-64, it is currently
> possible to add/remove memory to the system in >= 4MB granularity.
> Memory hotplug works very reliably. For memory unplug, there are no
> guarantees how much memory can actually get unplugged, it depends on the
> setup (especially: fragmentation of (unmovable) memory). I have plans to
> improve that in the future.
>
> --------------------------------------------------------------------------
> 1. virtio-mem
> --------------------------------------------------------------------------
>
> The basic idea behind virtio-mem was presented at KVM Forum 2018. The
> slides can be found at [1]. The previous RFC can be found at [2]. The
> first RFC can be found at [3]. However, the concept evolved over time. The
> KVM Forum slides roughly match the current design.
>
> Patch #2 ("virtio-mem: Paravirtualized memory hotplug") contains quite some
> information, especially in "include/uapi/linux/virtio_mem.h":
>
> Each virtio-mem device manages a dedicated region in physical address
> space. Each device can belong to a single NUMA node, multiple devices
> for a single NUMA node are possible. A virtio-mem device is like a
> "resizable DIMM" consisting of small memory blocks that can be plugged
> or unplugged. The device driver is responsible for (un)plugging memory
> blocks on demand.
>
> Virtio-mem devices can only operate on their assigned memory region in
> order to (un)plug memory. A device cannot (un)plug memory belonging to
> other devices.
>
> The "region_size" corresponds to the maximum amount of memory that can
> be provided by a device. The "size" corresponds to the amount of memory
> that is currently plugged. "requested_size" corresponds to a request
> from the device to the device driver to (un)plug blocks. The
> device driver should try to (un)plug blocks in order to reach the
> "requested_size". It is impossible to plug more memory than requested.
>
> The "usable_region_size" represents the memory region that can actually
> be used to (un)plug memory. It is always at least as big as the
> "requested_size" and will grow dynamically. It will only shrink when
> explicitly triggered (VIRTIO_MEM_REQ_UNPLUG).
>
> Memory in the usable region can usually be read, however, there are no
> guarantees. It can happen that the device cannot process a request,
> because it is busy. The device driver has to retry later.
>
> Usually, during system resets all memory will get unplugged, so the
> device driver can start with a clean state. However, in specific
> scenarios (if the device is busy) it can happen that the device still
> has memory plugged. The device driver can request to unplug all memory
> (VIRTIO_MEM_REQ_UNPLUG) - which might take a while to succeed if the
> device is busy.
>
> --------------------------------------------------------------------------
> 2. Linux Implementation
> --------------------------------------------------------------------------
>
> This RFC reuses quite some existing MM infrastructure, however, has to
> expose some additional functionality.
>
> Memory blocks (e.g., 128MB) are added/removed on demand. Within these
> memory blocks, subblocks (e.g., 4MB) are plugged/unplugged. The sizes
> depend on the target architecture, MAX_ORDER + pageblock_order, and
> the block size of a virtio-mem device.
>
> add_memory()/try_remove_memory() is used to add/remove memory blocks.
> virtio-mem will not online memory blocks itself. This has to be done by
> user space, or configured into the kernel
> (CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE). virtio-mem will only unplug memory
> that was online to the ZONE_NORMAL. Memory is suggested to be onlined to
> the ZONE_NORMAL for now.
>
> The memory hotplug notifier is used to properly synchronize against
> onlining/offlining of memory blocks and to track the states of memory
> blocks (including the zone memory blocks are onlined to).
>
> The set_online_page() callback is used to keep unplugged subblocks
> of a memory block fake-offline when onlining the memory block.
> generic_online_page() is used to fake-online plugged subblocks. This
> handling is similar to the Hyper-V balloon driver.
>
> PG_offline is used to mark unplugged subblocks as offline, so e.g.,
> dumping tools (makedumpfile) will skip these pages. This is similar to
> other balloon drivers like virtio-balloon and Hyper-V.
>
> Memory offlining code is extended to allow drivers to drop their reference
> to PG_offline pages when MEM_GOING_OFFLINE, so these pages can be skipped
> when offlining memory blocks. This allows to offline memory blocks that
> have partially unplugged (allocated e.g., via alloc_contig_range())
> subblocks - or are completely unplugged.
>
> alloc_contig_range()/free_contig_range() [now exposed] is used to
> unplug/plug subblocks of memory blocks the are already exposed to Linux.
>
> offline_and_remove_memory() [new] is used to offline a fully unplugged
> memory block and remove it from Linux.
>
>
> A lot of additional information can be found in the separate patches and
> as comments in the code itself.
>
> --------------------------------------------------------------------------
> 3. Changes RFC v2 -> v3
> --------------------------------------------------------------------------
>
> A lot of things changed, especially also on the QEMU + virtio side. The
> biggest changes on the Linux driver side are:
> - Onlining/offlining of subblocks is now emulated on top of memory blocks.
> set_online_page()+alloc_contig_range()+free_contig_range() is now used
> for that. Core MM does not have to be modified and will continue to
> online/offline full memory blocks.
> - Onlining/offlining of memory blocks is no longer performed by virtio-mem.
> - Pg_offline is upstream and can be used. It is also used to allow
> offlining of partially unplugged memory blocks.
> - Memory block states + subblocks are now tracked more space-efficient.
> - Proper kexec(), kdump(), driver unload, driver reload, ZONE_MOVABLE, ...
> handling.
>
> --------------------------------------------------------------------------
> 4. Changes RFC v3 -> RFC v4
> --------------------------------------------------------------------------
>
> Only minor things changed, especially nothing on the QEMU + virtio side.
> Interresting changes on the Linux driver side are:
> - "mm: Allow to offline unmovable PageOffline() pages via
> MEM_GOING_OFFLINE"
> -- Rework to Michals suggestion (allow to isolate all PageOffline() pages
> by skipping all PageOffline() pages in has_unmovable_pages(). Fail
> offlining later if the pages cannot be offlined/migrated).
> - "virtio-mem: Allow to offline partially unplugged memory blocks"
> -- Adapt to Michals suggestion on core-mm part.
> - "virtio-mem: Better retry handling"
> -- Optimize retry intervals
> - "virtio-mem: Drop slab objects when unplug continues to fail"
> -- Call drop_slab()/drop_slab_node() when unplug keeps failing for a longer
> time.
> - Multiple cleanups and fixes.
>
> --------------------------------------------------------------------------
> 5. Future work
> --------------------------------------------------------------------------
>
> The separate patches contain a lot of future work items. One of the next
> steps is to make memory unplug more likely to succeed - currently, there
> are no guarantees on how much memory can get unplugged again. I have
> various ideas on how to limit fragmentation of all memory blocks that
> virtio-mem added.
>
> Memory hotplug:
> - Reduce the amount of memory resources if that turnes out to be an
> issue. Or try to speed up relevant code paths to deal with many
> resources.
> - Allocate the vmemmap from the added memory. Makes hotplug more likely
> to succeed, the vmemmap is stored on the same NUMA node and that
> unmovable memory will later not hinder unplug.
>
> Memory hotunplug:
> - Performance improvements:
> -- Sense (lockless) if it make sense to try alloc_contig_range() at all
> before directly trying to isolate and taking locks.
> -- Try to unplug bigger chunks if possible first.
> -- Identify free areas first, that don't have to be evacuated.
> - Make unplug more likely to succeed:
> -- There are various idea to limit fragmentation on memory block
> granularity. (e.g., ZONE_PREFER_MOVABLE and smart balancing)
> -- Allocate memmap from added memory. This way, less unmovable data can
> end up on the memory blocks.
> - OOM handling, e.g., via an OOM handler.
> - Defragmentation
> -- Will require a new virtio-mem CMD to exchange plugged<->unplugged blocks
>
> --------------------------------------------------------------------------
> 6. Example Usage
> --------------------------------------------------------------------------
>
> A very basic QEMU prototype (kept updated) is available at:
> https://github.com/davidhildenbrand/qemu.git virtio-mem
>
> It lacks various features, however, works to test the guest driver side:
> - No support for resizable memory regions / memory backends yet
> - No protection of unplugged memory (esp., userfaultfd-wp) yet
> - No dump/migration/XXX optimizations to skip unplugged memory (and avoid
> touching it)
>
> Start QEMU with two virtio-mem devices (one per NUMA node):
> $ qemu-system-x86_64 -m 4G,maxmem=20G \
> -smp sockets=2,cores=2 \
> -numa node,nodeid=0,cpus=0-1 -numa node,nodeid=1,cpus=2-3 \
> [...]
> -object memory-backend-ram,id=mem0,size=8G \
> -device virtio-mem-pci,id=vm0,memdev=mem0,node=0,requested-size=128M \
> -object memory-backend-ram,id=mem1,size=8G \
> -device virtio-mem-pci,id=vm1,memdev=mem1,node=1,requested-size=80M
>
> Query the configuration:
> QEMU 4.1.95 monitor - type 'help' for more information
> (qemu) info memory-devices
> Memory device [virtio-mem]: "vm0"
> memaddr: 0x140000000
> node: 0
> requested-size: 134217728
> size: 134217728
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem0
> Memory device [virtio-mem]: "vm1"
> memaddr: 0x340000000
> node: 1
> requested-size: 83886080
> size: 83886080
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem1
>
> Add some memory to node 1:
> QEMU 4.1.95 monitor - type 'help' for more information
> (qemu) qom-set vm1 requested-size 1G
>
> Remove some memory from node 0:
> QEMU 4.1.95 monitor - type 'help' for more information
> (qemu) qom-set vm0 requested-size 64M
>
> Query the configuration again:
> (qemu) info memory-devices
> Memory device [virtio-mem]: "vm0"
> memaddr: 0x140000000
> node: 0
> requested-size: 67108864
> size: 67108864
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem0
> Memory device [virtio-mem]: "vm1"
> memaddr: 0x340000000
> node: 1
> requested-size: 1073741824
> size: 1073741824
> max-size: 8589934592
> block-size: 2097152
> memdev: /objects/mem1
>
> --------------------------------------------------------------------------
> 7. Q/A
> --------------------------------------------------------------------------
>
> Q: Why add/remove parts ("subblocks") of memory blocks/sections?
> A: Flexibility (section size depends on the architecture) - e.g., some
> architectures have a section size of 2GB. Also, the memory block size
> is variable (e.g., on x86-64). I want to avoid any such restrictions.
> Some use cases want to add/remove memory in smaller granularities to a
> VM (e.g., the Hyper-V balloon also implements this) - especially smaller
> VMs like used for kata containers. Also, on memory unplug, it is more
> reliable to free-up and unplug multiple small chunks instead
> of one big chunk. E.g., if one page of a DIMM is either unmovable or
> pinned, the DIMM can't get unplugged. This approach is basically a
> compromise between DIMM-based memory hot(un)plug and balloon
> inflation/deflation, which works mostly on page granularity.
>
> Q: Why care about memory blocks?
> A: They are the way to tell user space about new memory. This way,
> memory can get onlined/offlined by user space. Also, e.g., kdump
> relies on udev events to reload kexec when memory blocks are
> onlined/offlined. Memory blocks are the "real" memory hot(un)plug
> granularity. Everything that's smaller has to be emulated "on top".
>
> Q: Won't memory unplug of subblocks fragment memory?
> A: Yes and no. Unplugging e.g., >=4MB subblocks on x86-64 will not really
> fragment memory like unplugging random pages like a balloon driver does.
> Buddy merging will not be limited. However, any allocation that requires
> bigger consecutive memory chunks (e.g., gigantic pages) might observe
> the fragmentation. Possible solutions: Allocate gigantic huge pages
> before unplugging memory, don't unplug memory, combine virtio-mem with
> DIMM based memory or bigger initial memory. Remember, a virtio-mem
> device will only unplug on the memory range it manages, not on other
> DIMMs. Unplug of single memory blocks will result in similar
> fragmentation in respect to gigantic huge pages. I ahve plans for a
> virtio-mem defragmentation feature in the future.
>
> Q: How reliable is memory unplug?
> A: There are no guarantees on how much memory can get unplugged
> again. However, it is more likely to find 4MB chunks to unplug than
> e.g., 128MB chunks. If memory is terribly fragmented, there is nothing
> we can do - for now. I consider memory hotplug the first primary use
> of virtio-mem. Memory unplug might usually work, but we want to improve
> the performance and the amount of memory we can actually unplug later.
>
> Q: Why not unplug from the ZONE_MOVABLE?
> A: Unplugged memory chunks are unmovable. Unmovable data must not end up
> on the ZONE_MOVABLE - similar to gigantic pages - they will never be
> allocated from ZONE_MOVABLE. virtio-mem added memory can be onlined
> to the ZONE_MOVABLE, but subblocks will not get unplugged from it.
>
> Q: How big should the initial (!virtio-mem) memory of a VM be?
> A: virtio-mem memory will not go to the DMA zones. So to avoid running out
> of DMA memory, I suggest something like 2-3GB on x86-64. But many
> VMs can most probably deal with less DMA memory - depends on the use
> case.
>
> [1] https://events.linuxfoundation.org/wp-content/uploads/2017/12/virtio-mem-Paravirtualized-Memory-David-Hildenbrand-Red-Hat-1.pdf
> [2] https://lkml.kernel.org/r/[email protected]
> [3] https://lkml.kernel.org/r/[email protected]
>
> Cc: Sebastien Boeuf <[email protected]>
> Cc: Samuel Ortiz <[email protected]>
> Cc: Robert Bradford <[email protected]>
> Cc: Luiz Capitulino <[email protected]>
>
> David Hildenbrand (13):
> ACPI: NUMA: export pxm_to_node
> virtio-mem: Paravirtualized memory hotplug
> virtio-mem: Paravirtualized memory hotunplug part 1
> mm: Export alloc_contig_range() / free_contig_range()
> virtio-mem: Paravirtualized memory hotunplug part 2
> mm: Allow to offline unmovable PageOffline() pages via
> MEM_GOING_OFFLINE
> virtio-mem: Allow to offline partially unplugged memory blocks
> mm/memory_hotplug: Introduce offline_and_remove_memory()
> virtio-mem: Offline and remove completely unplugged memory blocks
> virtio-mem: Better retry handling
> mm/vmscan: Move count_vm_event(DROP_SLAB) into drop_slab()
> mm/vmscan: Export drop_slab() and drop_slab_node()
> virtio-mem: Drop slab objects when unplug continues to fail

Ping,

I'd love to get some feedback on

a) The remaining MM bits from MM folks (especially, patch #6 and #8).
b) The general virtio infrastructure (esp. uapi in patch #2) from virtio
folks.

I'm planning to send a proper v1 (!RFC) once I have all necessary MM
acks. In the meanwhile, I will do more testing and minor reworks (e.g.,
fix !CONFIG_NUMA compilation).

--
Thanks,

David / dhildenb

2020-01-29 09:43:13

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On 09.01.20 14:48, David Hildenbrand wrote:
> On 12.12.19 18:11, David Hildenbrand wrote:
>> This series is based on latest linux-next. The patches are located at:
>> https://github.com/davidhildenbrand/linux.git virtio-mem-rfc-v4
>>
>> The basic idea of virtio-mem is to provide a flexible,
>> cross-architecture memory hot(un)plug solution that avoids many limitations
>> imposed by existing technologies, architectures, and interfaces. More
>> details can be found below and in linked material.
>>
>> This RFC is limited to x86-64, however, should theoretically work on any
>> architecture that supports virtio and implements memory hot(un)plug under
>> Linux - like s390x, powerpc64 and arm64. On x86-64, it is currently
>> possible to add/remove memory to the system in >= 4MB granularity.
>> Memory hotplug works very reliably. For memory unplug, there are no
>> guarantees how much memory can actually get unplugged, it depends on the
>> setup (especially: fragmentation of (unmovable) memory). I have plans to
>> improve that in the future.
>>
>> --------------------------------------------------------------------------
>> 1. virtio-mem
>> --------------------------------------------------------------------------
>>
>> The basic idea behind virtio-mem was presented at KVM Forum 2018. The
>> slides can be found at [1]. The previous RFC can be found at [2]. The
>> first RFC can be found at [3]. However, the concept evolved over time. The
>> KVM Forum slides roughly match the current design.
>>
>> Patch #2 ("virtio-mem: Paravirtualized memory hotplug") contains quite some
>> information, especially in "include/uapi/linux/virtio_mem.h":
>>
>> Each virtio-mem device manages a dedicated region in physical address
>> space. Each device can belong to a single NUMA node, multiple devices
>> for a single NUMA node are possible. A virtio-mem device is like a
>> "resizable DIMM" consisting of small memory blocks that can be plugged
>> or unplugged. The device driver is responsible for (un)plugging memory
>> blocks on demand.
>>
>> Virtio-mem devices can only operate on their assigned memory region in
>> order to (un)plug memory. A device cannot (un)plug memory belonging to
>> other devices.
>>
>> The "region_size" corresponds to the maximum amount of memory that can
>> be provided by a device. The "size" corresponds to the amount of memory
>> that is currently plugged. "requested_size" corresponds to a request
>> from the device to the device driver to (un)plug blocks. The
>> device driver should try to (un)plug blocks in order to reach the
>> "requested_size". It is impossible to plug more memory than requested.
>>
>> The "usable_region_size" represents the memory region that can actually
>> be used to (un)plug memory. It is always at least as big as the
>> "requested_size" and will grow dynamically. It will only shrink when
>> explicitly triggered (VIRTIO_MEM_REQ_UNPLUG).
>>
>> Memory in the usable region can usually be read, however, there are no
>> guarantees. It can happen that the device cannot process a request,
>> because it is busy. The device driver has to retry later.
>>
>> Usually, during system resets all memory will get unplugged, so the
>> device driver can start with a clean state. However, in specific
>> scenarios (if the device is busy) it can happen that the device still
>> has memory plugged. The device driver can request to unplug all memory
>> (VIRTIO_MEM_REQ_UNPLUG) - which might take a while to succeed if the
>> device is busy.
>>
>> --------------------------------------------------------------------------
>> 2. Linux Implementation
>> --------------------------------------------------------------------------
>>
>> This RFC reuses quite some existing MM infrastructure, however, has to
>> expose some additional functionality.
>>
>> Memory blocks (e.g., 128MB) are added/removed on demand. Within these
>> memory blocks, subblocks (e.g., 4MB) are plugged/unplugged. The sizes
>> depend on the target architecture, MAX_ORDER + pageblock_order, and
>> the block size of a virtio-mem device.
>>
>> add_memory()/try_remove_memory() is used to add/remove memory blocks.
>> virtio-mem will not online memory blocks itself. This has to be done by
>> user space, or configured into the kernel
>> (CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE). virtio-mem will only unplug memory
>> that was online to the ZONE_NORMAL. Memory is suggested to be onlined to
>> the ZONE_NORMAL for now.
>>
>> The memory hotplug notifier is used to properly synchronize against
>> onlining/offlining of memory blocks and to track the states of memory
>> blocks (including the zone memory blocks are onlined to).
>>
>> The set_online_page() callback is used to keep unplugged subblocks
>> of a memory block fake-offline when onlining the memory block.
>> generic_online_page() is used to fake-online plugged subblocks. This
>> handling is similar to the Hyper-V balloon driver.
>>
>> PG_offline is used to mark unplugged subblocks as offline, so e.g.,
>> dumping tools (makedumpfile) will skip these pages. This is similar to
>> other balloon drivers like virtio-balloon and Hyper-V.
>>
>> Memory offlining code is extended to allow drivers to drop their reference
>> to PG_offline pages when MEM_GOING_OFFLINE, so these pages can be skipped
>> when offlining memory blocks. This allows to offline memory blocks that
>> have partially unplugged (allocated e.g., via alloc_contig_range())
>> subblocks - or are completely unplugged.
>>
>> alloc_contig_range()/free_contig_range() [now exposed] is used to
>> unplug/plug subblocks of memory blocks the are already exposed to Linux.
>>
>> offline_and_remove_memory() [new] is used to offline a fully unplugged
>> memory block and remove it from Linux.
>>
>>
>> A lot of additional information can be found in the separate patches and
>> as comments in the code itself.
>>
>> --------------------------------------------------------------------------
>> 3. Changes RFC v2 -> v3
>> --------------------------------------------------------------------------
>>
>> A lot of things changed, especially also on the QEMU + virtio side. The
>> biggest changes on the Linux driver side are:
>> - Onlining/offlining of subblocks is now emulated on top of memory blocks.
>> set_online_page()+alloc_contig_range()+free_contig_range() is now used
>> for that. Core MM does not have to be modified and will continue to
>> online/offline full memory blocks.
>> - Onlining/offlining of memory blocks is no longer performed by virtio-mem.
>> - Pg_offline is upstream and can be used. It is also used to allow
>> offlining of partially unplugged memory blocks.
>> - Memory block states + subblocks are now tracked more space-efficient.
>> - Proper kexec(), kdump(), driver unload, driver reload, ZONE_MOVABLE, ...
>> handling.
>>
>> --------------------------------------------------------------------------
>> 4. Changes RFC v3 -> RFC v4
>> --------------------------------------------------------------------------
>>
>> Only minor things changed, especially nothing on the QEMU + virtio side.
>> Interresting changes on the Linux driver side are:
>> - "mm: Allow to offline unmovable PageOffline() pages via
>> MEM_GOING_OFFLINE"
>> -- Rework to Michals suggestion (allow to isolate all PageOffline() pages
>> by skipping all PageOffline() pages in has_unmovable_pages(). Fail
>> offlining later if the pages cannot be offlined/migrated).
>> - "virtio-mem: Allow to offline partially unplugged memory blocks"
>> -- Adapt to Michals suggestion on core-mm part.
>> - "virtio-mem: Better retry handling"
>> -- Optimize retry intervals
>> - "virtio-mem: Drop slab objects when unplug continues to fail"
>> -- Call drop_slab()/drop_slab_node() when unplug keeps failing for a longer
>> time.
>> - Multiple cleanups and fixes.
>>
>> --------------------------------------------------------------------------
>> 5. Future work
>> --------------------------------------------------------------------------
>>
>> The separate patches contain a lot of future work items. One of the next
>> steps is to make memory unplug more likely to succeed - currently, there
>> are no guarantees on how much memory can get unplugged again. I have
>> various ideas on how to limit fragmentation of all memory blocks that
>> virtio-mem added.
>>
>> Memory hotplug:
>> - Reduce the amount of memory resources if that turnes out to be an
>> issue. Or try to speed up relevant code paths to deal with many
>> resources.
>> - Allocate the vmemmap from the added memory. Makes hotplug more likely
>> to succeed, the vmemmap is stored on the same NUMA node and that
>> unmovable memory will later not hinder unplug.
>>
>> Memory hotunplug:
>> - Performance improvements:
>> -- Sense (lockless) if it make sense to try alloc_contig_range() at all
>> before directly trying to isolate and taking locks.
>> -- Try to unplug bigger chunks if possible first.
>> -- Identify free areas first, that don't have to be evacuated.
>> - Make unplug more likely to succeed:
>> -- There are various idea to limit fragmentation on memory block
>> granularity. (e.g., ZONE_PREFER_MOVABLE and smart balancing)
>> -- Allocate memmap from added memory. This way, less unmovable data can
>> end up on the memory blocks.
>> - OOM handling, e.g., via an OOM handler.
>> - Defragmentation
>> -- Will require a new virtio-mem CMD to exchange plugged<->unplugged blocks
>>
>> --------------------------------------------------------------------------
>> 6. Example Usage
>> --------------------------------------------------------------------------
>>
>> A very basic QEMU prototype (kept updated) is available at:
>> https://github.com/davidhildenbrand/qemu.git virtio-mem
>>
>> It lacks various features, however, works to test the guest driver side:
>> - No support for resizable memory regions / memory backends yet
>> - No protection of unplugged memory (esp., userfaultfd-wp) yet
>> - No dump/migration/XXX optimizations to skip unplugged memory (and avoid
>> touching it)
>>
>> Start QEMU with two virtio-mem devices (one per NUMA node):
>> $ qemu-system-x86_64 -m 4G,maxmem=20G \
>> -smp sockets=2,cores=2 \
>> -numa node,nodeid=0,cpus=0-1 -numa node,nodeid=1,cpus=2-3 \
>> [...]
>> -object memory-backend-ram,id=mem0,size=8G \
>> -device virtio-mem-pci,id=vm0,memdev=mem0,node=0,requested-size=128M \
>> -object memory-backend-ram,id=mem1,size=8G \
>> -device virtio-mem-pci,id=vm1,memdev=mem1,node=1,requested-size=80M
>>
>> Query the configuration:
>> QEMU 4.1.95 monitor - type 'help' for more information
>> (qemu) info memory-devices
>> Memory device [virtio-mem]: "vm0"
>> memaddr: 0x140000000
>> node: 0
>> requested-size: 134217728
>> size: 134217728
>> max-size: 8589934592
>> block-size: 2097152
>> memdev: /objects/mem0
>> Memory device [virtio-mem]: "vm1"
>> memaddr: 0x340000000
>> node: 1
>> requested-size: 83886080
>> size: 83886080
>> max-size: 8589934592
>> block-size: 2097152
>> memdev: /objects/mem1
>>
>> Add some memory to node 1:
>> QEMU 4.1.95 monitor - type 'help' for more information
>> (qemu) qom-set vm1 requested-size 1G
>>
>> Remove some memory from node 0:
>> QEMU 4.1.95 monitor - type 'help' for more information
>> (qemu) qom-set vm0 requested-size 64M
>>
>> Query the configuration again:
>> (qemu) info memory-devices
>> Memory device [virtio-mem]: "vm0"
>> memaddr: 0x140000000
>> node: 0
>> requested-size: 67108864
>> size: 67108864
>> max-size: 8589934592
>> block-size: 2097152
>> memdev: /objects/mem0
>> Memory device [virtio-mem]: "vm1"
>> memaddr: 0x340000000
>> node: 1
>> requested-size: 1073741824
>> size: 1073741824
>> max-size: 8589934592
>> block-size: 2097152
>> memdev: /objects/mem1
>>
>> --------------------------------------------------------------------------
>> 7. Q/A
>> --------------------------------------------------------------------------
>>
>> Q: Why add/remove parts ("subblocks") of memory blocks/sections?
>> A: Flexibility (section size depends on the architecture) - e.g., some
>> architectures have a section size of 2GB. Also, the memory block size
>> is variable (e.g., on x86-64). I want to avoid any such restrictions.
>> Some use cases want to add/remove memory in smaller granularities to a
>> VM (e.g., the Hyper-V balloon also implements this) - especially smaller
>> VMs like used for kata containers. Also, on memory unplug, it is more
>> reliable to free-up and unplug multiple small chunks instead
>> of one big chunk. E.g., if one page of a DIMM is either unmovable or
>> pinned, the DIMM can't get unplugged. This approach is basically a
>> compromise between DIMM-based memory hot(un)plug and balloon
>> inflation/deflation, which works mostly on page granularity.
>>
>> Q: Why care about memory blocks?
>> A: They are the way to tell user space about new memory. This way,
>> memory can get onlined/offlined by user space. Also, e.g., kdump
>> relies on udev events to reload kexec when memory blocks are
>> onlined/offlined. Memory blocks are the "real" memory hot(un)plug
>> granularity. Everything that's smaller has to be emulated "on top".
>>
>> Q: Won't memory unplug of subblocks fragment memory?
>> A: Yes and no. Unplugging e.g., >=4MB subblocks on x86-64 will not really
>> fragment memory like unplugging random pages like a balloon driver does.
>> Buddy merging will not be limited. However, any allocation that requires
>> bigger consecutive memory chunks (e.g., gigantic pages) might observe
>> the fragmentation. Possible solutions: Allocate gigantic huge pages
>> before unplugging memory, don't unplug memory, combine virtio-mem with
>> DIMM based memory or bigger initial memory. Remember, a virtio-mem
>> device will only unplug on the memory range it manages, not on other
>> DIMMs. Unplug of single memory blocks will result in similar
>> fragmentation in respect to gigantic huge pages. I ahve plans for a
>> virtio-mem defragmentation feature in the future.
>>
>> Q: How reliable is memory unplug?
>> A: There are no guarantees on how much memory can get unplugged
>> again. However, it is more likely to find 4MB chunks to unplug than
>> e.g., 128MB chunks. If memory is terribly fragmented, there is nothing
>> we can do - for now. I consider memory hotplug the first primary use
>> of virtio-mem. Memory unplug might usually work, but we want to improve
>> the performance and the amount of memory we can actually unplug later.
>>
>> Q: Why not unplug from the ZONE_MOVABLE?
>> A: Unplugged memory chunks are unmovable. Unmovable data must not end up
>> on the ZONE_MOVABLE - similar to gigantic pages - they will never be
>> allocated from ZONE_MOVABLE. virtio-mem added memory can be onlined
>> to the ZONE_MOVABLE, but subblocks will not get unplugged from it.
>>
>> Q: How big should the initial (!virtio-mem) memory of a VM be?
>> A: virtio-mem memory will not go to the DMA zones. So to avoid running out
>> of DMA memory, I suggest something like 2-3GB on x86-64. But many
>> VMs can most probably deal with less DMA memory - depends on the use
>> case.
>>
>> [1] https://events.linuxfoundation.org/wp-content/uploads/2017/12/virtio-mem-Paravirtualized-Memory-David-Hildenbrand-Red-Hat-1.pdf
>> [2] https://lkml.kernel.org/r/[email protected]
>> [3] https://lkml.kernel.org/r/[email protected]
>>
>> Cc: Sebastien Boeuf <[email protected]>
>> Cc: Samuel Ortiz <[email protected]>
>> Cc: Robert Bradford <[email protected]>
>> Cc: Luiz Capitulino <[email protected]>
>>
>> David Hildenbrand (13):
>> ACPI: NUMA: export pxm_to_node
>> virtio-mem: Paravirtualized memory hotplug
>> virtio-mem: Paravirtualized memory hotunplug part 1
>> mm: Export alloc_contig_range() / free_contig_range()
>> virtio-mem: Paravirtualized memory hotunplug part 2
>> mm: Allow to offline unmovable PageOffline() pages via
>> MEM_GOING_OFFLINE
>> virtio-mem: Allow to offline partially unplugged memory blocks
>> mm/memory_hotplug: Introduce offline_and_remove_memory()
>> virtio-mem: Offline and remove completely unplugged memory blocks
>> virtio-mem: Better retry handling
>> mm/vmscan: Move count_vm_event(DROP_SLAB) into drop_slab()
>> mm/vmscan: Export drop_slab() and drop_slab_node()
>> virtio-mem: Drop slab objects when unplug continues to fail
>
> Ping,
>
> I'd love to get some feedback on
>
> a) The remaining MM bits from MM folks (especially, patch #6 and #8).

Friendly ping again:

Can I get some feedback on the two important MM changes in this series

"[PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages
via MEM_GOING_OFFLINE"

and

"[PATCH RFC v4 08/13] mm/memory_hotplug: Introduce
offline_and_remove_memory()"

--
Thanks,

David / dhildenb

2020-02-25 09:59:43

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On 29.01.20 10:41, David Hildenbrand wrote:
> On 09.01.20 14:48, David Hildenbrand wrote:
>> On 12.12.19 18:11, David Hildenbrand wrote:
>>> This series is based on latest linux-next. The patches are located at:
>>> https://github.com/davidhildenbrand/linux.git virtio-mem-rfc-v4
>>>
>>> The basic idea of virtio-mem is to provide a flexible,
>>> cross-architecture memory hot(un)plug solution that avoids many limitations
>>> imposed by existing technologies, architectures, and interfaces. More
>>> details can be found below and in linked material.
>>>
>>> This RFC is limited to x86-64, however, should theoretically work on any
>>> architecture that supports virtio and implements memory hot(un)plug under
>>> Linux - like s390x, powerpc64 and arm64. On x86-64, it is currently
>>> possible to add/remove memory to the system in >= 4MB granularity.
>>> Memory hotplug works very reliably. For memory unplug, there are no
>>> guarantees how much memory can actually get unplugged, it depends on the
>>> setup (especially: fragmentation of (unmovable) memory). I have plans to
>>> improve that in the future.
>>>
>>> --------------------------------------------------------------------------
>>> 1. virtio-mem
>>> --------------------------------------------------------------------------
>>>
>>> The basic idea behind virtio-mem was presented at KVM Forum 2018. The
>>> slides can be found at [1]. The previous RFC can be found at [2]. The
>>> first RFC can be found at [3]. However, the concept evolved over time. The
>>> KVM Forum slides roughly match the current design.
>>>
>>> Patch #2 ("virtio-mem: Paravirtualized memory hotplug") contains quite some
>>> information, especially in "include/uapi/linux/virtio_mem.h":
>>>
>>> Each virtio-mem device manages a dedicated region in physical address
>>> space. Each device can belong to a single NUMA node, multiple devices
>>> for a single NUMA node are possible. A virtio-mem device is like a
>>> "resizable DIMM" consisting of small memory blocks that can be plugged
>>> or unplugged. The device driver is responsible for (un)plugging memory
>>> blocks on demand.
>>>
>>> Virtio-mem devices can only operate on their assigned memory region in
>>> order to (un)plug memory. A device cannot (un)plug memory belonging to
>>> other devices.
>>>
>>> The "region_size" corresponds to the maximum amount of memory that can
>>> be provided by a device. The "size" corresponds to the amount of memory
>>> that is currently plugged. "requested_size" corresponds to a request
>>> from the device to the device driver to (un)plug blocks. The
>>> device driver should try to (un)plug blocks in order to reach the
>>> "requested_size". It is impossible to plug more memory than requested.
>>>
>>> The "usable_region_size" represents the memory region that can actually
>>> be used to (un)plug memory. It is always at least as big as the
>>> "requested_size" and will grow dynamically. It will only shrink when
>>> explicitly triggered (VIRTIO_MEM_REQ_UNPLUG).
>>>
>>> Memory in the usable region can usually be read, however, there are no
>>> guarantees. It can happen that the device cannot process a request,
>>> because it is busy. The device driver has to retry later.
>>>
>>> Usually, during system resets all memory will get unplugged, so the
>>> device driver can start with a clean state. However, in specific
>>> scenarios (if the device is busy) it can happen that the device still
>>> has memory plugged. The device driver can request to unplug all memory
>>> (VIRTIO_MEM_REQ_UNPLUG) - which might take a while to succeed if the
>>> device is busy.
>>>
>>> --------------------------------------------------------------------------
>>> 2. Linux Implementation
>>> --------------------------------------------------------------------------
>>>
>>> This RFC reuses quite some existing MM infrastructure, however, has to
>>> expose some additional functionality.
>>>
>>> Memory blocks (e.g., 128MB) are added/removed on demand. Within these
>>> memory blocks, subblocks (e.g., 4MB) are plugged/unplugged. The sizes
>>> depend on the target architecture, MAX_ORDER + pageblock_order, and
>>> the block size of a virtio-mem device.
>>>
>>> add_memory()/try_remove_memory() is used to add/remove memory blocks.
>>> virtio-mem will not online memory blocks itself. This has to be done by
>>> user space, or configured into the kernel
>>> (CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE). virtio-mem will only unplug memory
>>> that was online to the ZONE_NORMAL. Memory is suggested to be onlined to
>>> the ZONE_NORMAL for now.
>>>
>>> The memory hotplug notifier is used to properly synchronize against
>>> onlining/offlining of memory blocks and to track the states of memory
>>> blocks (including the zone memory blocks are onlined to).
>>>
>>> The set_online_page() callback is used to keep unplugged subblocks
>>> of a memory block fake-offline when onlining the memory block.
>>> generic_online_page() is used to fake-online plugged subblocks. This
>>> handling is similar to the Hyper-V balloon driver.
>>>
>>> PG_offline is used to mark unplugged subblocks as offline, so e.g.,
>>> dumping tools (makedumpfile) will skip these pages. This is similar to
>>> other balloon drivers like virtio-balloon and Hyper-V.
>>>
>>> Memory offlining code is extended to allow drivers to drop their reference
>>> to PG_offline pages when MEM_GOING_OFFLINE, so these pages can be skipped
>>> when offlining memory blocks. This allows to offline memory blocks that
>>> have partially unplugged (allocated e.g., via alloc_contig_range())
>>> subblocks - or are completely unplugged.
>>>
>>> alloc_contig_range()/free_contig_range() [now exposed] is used to
>>> unplug/plug subblocks of memory blocks the are already exposed to Linux.
>>>
>>> offline_and_remove_memory() [new] is used to offline a fully unplugged
>>> memory block and remove it from Linux.
>>>
>>>
>>> A lot of additional information can be found in the separate patches and
>>> as comments in the code itself.
>>>
>>> --------------------------------------------------------------------------
>>> 3. Changes RFC v2 -> v3
>>> --------------------------------------------------------------------------
>>>
>>> A lot of things changed, especially also on the QEMU + virtio side. The
>>> biggest changes on the Linux driver side are:
>>> - Onlining/offlining of subblocks is now emulated on top of memory blocks.
>>> set_online_page()+alloc_contig_range()+free_contig_range() is now used
>>> for that. Core MM does not have to be modified and will continue to
>>> online/offline full memory blocks.
>>> - Onlining/offlining of memory blocks is no longer performed by virtio-mem.
>>> - Pg_offline is upstream and can be used. It is also used to allow
>>> offlining of partially unplugged memory blocks.
>>> - Memory block states + subblocks are now tracked more space-efficient.
>>> - Proper kexec(), kdump(), driver unload, driver reload, ZONE_MOVABLE, ...
>>> handling.
>>>
>>> --------------------------------------------------------------------------
>>> 4. Changes RFC v3 -> RFC v4
>>> --------------------------------------------------------------------------
>>>
>>> Only minor things changed, especially nothing on the QEMU + virtio side.
>>> Interresting changes on the Linux driver side are:
>>> - "mm: Allow to offline unmovable PageOffline() pages via
>>> MEM_GOING_OFFLINE"
>>> -- Rework to Michals suggestion (allow to isolate all PageOffline() pages
>>> by skipping all PageOffline() pages in has_unmovable_pages(). Fail
>>> offlining later if the pages cannot be offlined/migrated).
>>> - "virtio-mem: Allow to offline partially unplugged memory blocks"
>>> -- Adapt to Michals suggestion on core-mm part.
>>> - "virtio-mem: Better retry handling"
>>> -- Optimize retry intervals
>>> - "virtio-mem: Drop slab objects when unplug continues to fail"
>>> -- Call drop_slab()/drop_slab_node() when unplug keeps failing for a longer
>>> time.
>>> - Multiple cleanups and fixes.
>>>
>>> --------------------------------------------------------------------------
>>> 5. Future work
>>> --------------------------------------------------------------------------
>>>
>>> The separate patches contain a lot of future work items. One of the next
>>> steps is to make memory unplug more likely to succeed - currently, there
>>> are no guarantees on how much memory can get unplugged again. I have
>>> various ideas on how to limit fragmentation of all memory blocks that
>>> virtio-mem added.
>>>
>>> Memory hotplug:
>>> - Reduce the amount of memory resources if that turnes out to be an
>>> issue. Or try to speed up relevant code paths to deal with many
>>> resources.
>>> - Allocate the vmemmap from the added memory. Makes hotplug more likely
>>> to succeed, the vmemmap is stored on the same NUMA node and that
>>> unmovable memory will later not hinder unplug.
>>>
>>> Memory hotunplug:
>>> - Performance improvements:
>>> -- Sense (lockless) if it make sense to try alloc_contig_range() at all
>>> before directly trying to isolate and taking locks.
>>> -- Try to unplug bigger chunks if possible first.
>>> -- Identify free areas first, that don't have to be evacuated.
>>> - Make unplug more likely to succeed:
>>> -- There are various idea to limit fragmentation on memory block
>>> granularity. (e.g., ZONE_PREFER_MOVABLE and smart balancing)
>>> -- Allocate memmap from added memory. This way, less unmovable data can
>>> end up on the memory blocks.
>>> - OOM handling, e.g., via an OOM handler.
>>> - Defragmentation
>>> -- Will require a new virtio-mem CMD to exchange plugged<->unplugged blocks
>>>
>>> --------------------------------------------------------------------------
>>> 6. Example Usage
>>> --------------------------------------------------------------------------
>>>
>>> A very basic QEMU prototype (kept updated) is available at:
>>> https://github.com/davidhildenbrand/qemu.git virtio-mem
>>>
>>> It lacks various features, however, works to test the guest driver side:
>>> - No support for resizable memory regions / memory backends yet
>>> - No protection of unplugged memory (esp., userfaultfd-wp) yet
>>> - No dump/migration/XXX optimizations to skip unplugged memory (and avoid
>>> touching it)
>>>
>>> Start QEMU with two virtio-mem devices (one per NUMA node):
>>> $ qemu-system-x86_64 -m 4G,maxmem=20G \
>>> -smp sockets=2,cores=2 \
>>> -numa node,nodeid=0,cpus=0-1 -numa node,nodeid=1,cpus=2-3 \
>>> [...]
>>> -object memory-backend-ram,id=mem0,size=8G \
>>> -device virtio-mem-pci,id=vm0,memdev=mem0,node=0,requested-size=128M \
>>> -object memory-backend-ram,id=mem1,size=8G \
>>> -device virtio-mem-pci,id=vm1,memdev=mem1,node=1,requested-size=80M
>>>
>>> Query the configuration:
>>> QEMU 4.1.95 monitor - type 'help' for more information
>>> (qemu) info memory-devices
>>> Memory device [virtio-mem]: "vm0"
>>> memaddr: 0x140000000
>>> node: 0
>>> requested-size: 134217728
>>> size: 134217728
>>> max-size: 8589934592
>>> block-size: 2097152
>>> memdev: /objects/mem0
>>> Memory device [virtio-mem]: "vm1"
>>> memaddr: 0x340000000
>>> node: 1
>>> requested-size: 83886080
>>> size: 83886080
>>> max-size: 8589934592
>>> block-size: 2097152
>>> memdev: /objects/mem1
>>>
>>> Add some memory to node 1:
>>> QEMU 4.1.95 monitor - type 'help' for more information
>>> (qemu) qom-set vm1 requested-size 1G
>>>
>>> Remove some memory from node 0:
>>> QEMU 4.1.95 monitor - type 'help' for more information
>>> (qemu) qom-set vm0 requested-size 64M
>>>
>>> Query the configuration again:
>>> (qemu) info memory-devices
>>> Memory device [virtio-mem]: "vm0"
>>> memaddr: 0x140000000
>>> node: 0
>>> requested-size: 67108864
>>> size: 67108864
>>> max-size: 8589934592
>>> block-size: 2097152
>>> memdev: /objects/mem0
>>> Memory device [virtio-mem]: "vm1"
>>> memaddr: 0x340000000
>>> node: 1
>>> requested-size: 1073741824
>>> size: 1073741824
>>> max-size: 8589934592
>>> block-size: 2097152
>>> memdev: /objects/mem1
>>>
>>> --------------------------------------------------------------------------
>>> 7. Q/A
>>> --------------------------------------------------------------------------
>>>
>>> Q: Why add/remove parts ("subblocks") of memory blocks/sections?
>>> A: Flexibility (section size depends on the architecture) - e.g., some
>>> architectures have a section size of 2GB. Also, the memory block size
>>> is variable (e.g., on x86-64). I want to avoid any such restrictions.
>>> Some use cases want to add/remove memory in smaller granularities to a
>>> VM (e.g., the Hyper-V balloon also implements this) - especially smaller
>>> VMs like used for kata containers. Also, on memory unplug, it is more
>>> reliable to free-up and unplug multiple small chunks instead
>>> of one big chunk. E.g., if one page of a DIMM is either unmovable or
>>> pinned, the DIMM can't get unplugged. This approach is basically a
>>> compromise between DIMM-based memory hot(un)plug and balloon
>>> inflation/deflation, which works mostly on page granularity.
>>>
>>> Q: Why care about memory blocks?
>>> A: They are the way to tell user space about new memory. This way,
>>> memory can get onlined/offlined by user space. Also, e.g., kdump
>>> relies on udev events to reload kexec when memory blocks are
>>> onlined/offlined. Memory blocks are the "real" memory hot(un)plug
>>> granularity. Everything that's smaller has to be emulated "on top".
>>>
>>> Q: Won't memory unplug of subblocks fragment memory?
>>> A: Yes and no. Unplugging e.g., >=4MB subblocks on x86-64 will not really
>>> fragment memory like unplugging random pages like a balloon driver does.
>>> Buddy merging will not be limited. However, any allocation that requires
>>> bigger consecutive memory chunks (e.g., gigantic pages) might observe
>>> the fragmentation. Possible solutions: Allocate gigantic huge pages
>>> before unplugging memory, don't unplug memory, combine virtio-mem with
>>> DIMM based memory or bigger initial memory. Remember, a virtio-mem
>>> device will only unplug on the memory range it manages, not on other
>>> DIMMs. Unplug of single memory blocks will result in similar
>>> fragmentation in respect to gigantic huge pages. I ahve plans for a
>>> virtio-mem defragmentation feature in the future.
>>>
>>> Q: How reliable is memory unplug?
>>> A: There are no guarantees on how much memory can get unplugged
>>> again. However, it is more likely to find 4MB chunks to unplug than
>>> e.g., 128MB chunks. If memory is terribly fragmented, there is nothing
>>> we can do - for now. I consider memory hotplug the first primary use
>>> of virtio-mem. Memory unplug might usually work, but we want to improve
>>> the performance and the amount of memory we can actually unplug later.
>>>
>>> Q: Why not unplug from the ZONE_MOVABLE?
>>> A: Unplugged memory chunks are unmovable. Unmovable data must not end up
>>> on the ZONE_MOVABLE - similar to gigantic pages - they will never be
>>> allocated from ZONE_MOVABLE. virtio-mem added memory can be onlined
>>> to the ZONE_MOVABLE, but subblocks will not get unplugged from it.
>>>
>>> Q: How big should the initial (!virtio-mem) memory of a VM be?
>>> A: virtio-mem memory will not go to the DMA zones. So to avoid running out
>>> of DMA memory, I suggest something like 2-3GB on x86-64. But many
>>> VMs can most probably deal with less DMA memory - depends on the use
>>> case.
>>>
>>> [1] https://events.linuxfoundation.org/wp-content/uploads/2017/12/virtio-mem-Paravirtualized-Memory-David-Hildenbrand-Red-Hat-1.pdf
>>> [2] https://lkml.kernel.org/r/[email protected]
>>> [3] https://lkml.kernel.org/r/[email protected]
>>>
>>> Cc: Sebastien Boeuf <[email protected]>
>>> Cc: Samuel Ortiz <[email protected]>
>>> Cc: Robert Bradford <[email protected]>
>>> Cc: Luiz Capitulino <[email protected]>
>>>
>>> David Hildenbrand (13):
>>> ACPI: NUMA: export pxm_to_node
>>> virtio-mem: Paravirtualized memory hotplug
>>> virtio-mem: Paravirtualized memory hotunplug part 1
>>> mm: Export alloc_contig_range() / free_contig_range()
>>> virtio-mem: Paravirtualized memory hotunplug part 2
>>> mm: Allow to offline unmovable PageOffline() pages via
>>> MEM_GOING_OFFLINE
>>> virtio-mem: Allow to offline partially unplugged memory blocks
>>> mm/memory_hotplug: Introduce offline_and_remove_memory()
>>> virtio-mem: Offline and remove completely unplugged memory blocks
>>> virtio-mem: Better retry handling
>>> mm/vmscan: Move count_vm_event(DROP_SLAB) into drop_slab()
>>> mm/vmscan: Export drop_slab() and drop_slab_node()
>>> virtio-mem: Drop slab objects when unplug continues to fail
>>
>> Ping,
>>
>> I'd love to get some feedback on
>>
>> a) The remaining MM bits from MM folks (especially, patch #6 and #8).
>
> Friendly ping again:
>
> Can I get some feedback on the two important MM changes in this series
>
> "[PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages
> via MEM_GOING_OFFLINE"
>
> and
>
> "[PATCH RFC v4 08/13] mm/memory_hotplug: Introduce
> offline_and_remove_memory()"
>

Yet another ping.

--
Thanks,

David / dhildenb

2020-02-25 14:13:20

by Michal Hocko

[permalink] [raw]
Subject: Re: [PATCH RFC v4 08/13] mm/memory_hotplug: Introduce offline_and_remove_memory()

On Thu 12-12-19 18:11:32, David Hildenbrand wrote:
> virtio-mem wants to offline and remove a memory block once it unplugged
> all subblocks (e.g., using alloc_contig_range()). Let's provide
> an interface to do that from a driver. virtio-mem already supports to
> offline partially unplugged memory blocks. Offlining a fully unplugged
> memory block will not require to migrate any pages. All unplugged
> subblocks are PageOffline() and have a reference count of 0 - so
> offlining code will simply skip them.
>
> All we need an interface to trigger the "offlining" and the removing in a
> single operation - to make sure the memory block cannot get onlined by
> user space again before it gets removed.

Why does that matter? Is it really likely that the userspace would
interfere? What would be the scenario?

Or is still mostly about not requiring callers to open code this general
patter?

> To keep things simple, allow to only work on a single memory block.
>
> Cc: Andrew Morton <[email protected]>
> Cc: David Hildenbrand <[email protected]>
> Cc: Oscar Salvador <[email protected]>
> Cc: Michal Hocko <[email protected]>
> Cc: Pavel Tatashin <[email protected]>
> Cc: Wei Yang <[email protected]>
> Cc: Dan Williams <[email protected]>
> Cc: Qian Cai <[email protected]>
> Signed-off-by: David Hildenbrand <[email protected]>
> ---
> include/linux/memory_hotplug.h | 1 +
> mm/memory_hotplug.c | 35 ++++++++++++++++++++++++++++++++++
> 2 files changed, 36 insertions(+)
>
> diff --git a/include/linux/memory_hotplug.h b/include/linux/memory_hotplug.h
> index ba0dca6aac6e..586f5c59c291 100644
> --- a/include/linux/memory_hotplug.h
> +++ b/include/linux/memory_hotplug.h
> @@ -310,6 +310,7 @@ extern void try_offline_node(int nid);
> extern int offline_pages(unsigned long start_pfn, unsigned long nr_pages);
> extern int remove_memory(int nid, u64 start, u64 size);
> extern void __remove_memory(int nid, u64 start, u64 size);
> +extern int offline_and_remove_memory(int nid, u64 start, u64 size);
>
> #else
> static inline bool is_mem_section_removable(unsigned long pfn,
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index da01453a04e6..d04369e6d3cc 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -1825,4 +1825,39 @@ int remove_memory(int nid, u64 start, u64 size)
> return rc;
> }
> EXPORT_SYMBOL_GPL(remove_memory);
> +
> +/*
> + * Try to offline and remove a memory block. Might take a long time to
> + * finish in case memory is still in use. Primarily useful for memory devices
> + * that logically unplugged all memory (so it's no longer in use) and want to
> + * offline + remove the memory block.
> + */
> +int offline_and_remove_memory(int nid, u64 start, u64 size)
> +{
> + struct memory_block *mem;
> + int rc = -EINVAL;
> +
> + if (!IS_ALIGNED(start, memory_block_size_bytes()) ||
> + size != memory_block_size_bytes())
> + return rc;
> +
> + lock_device_hotplug();
> + mem = find_memory_block(__pfn_to_section(PFN_DOWN(start)));
> + if (mem)
> + rc = device_offline(&mem->dev);
> + /* Ignore if the device is already offline. */
> + if (rc > 0)
> + rc = 0;
> +
> + /*
> + * In case we succeeded to offline the memory block, remove it.
> + * This cannot fail as it cannot get onlined in the meantime.
> + */
> + if (!rc && try_remove_memory(nid, start, size))
> + BUG();
> + unlock_device_hotplug();
> +
> + return rc;
> +}
> +EXPORT_SYMBOL_GPL(offline_and_remove_memory);
> #endif /* CONFIG_MEMORY_HOTREMOVE */
> --
> 2.23.0

--
Michal Hocko
SUSE Labs

2020-02-25 14:14:19

by Michal Hocko

[permalink] [raw]
Subject: Re: [PATCH RFC v4 11/13] mm/vmscan: Move count_vm_event(DROP_SLAB) into drop_slab()

On Thu 12-12-19 18:11:35, David Hildenbrand wrote:
> Let's count within the function itself, so every invocation (of future
> users) will be counted.
>
> Cc: Alexander Viro <[email protected]>
> Cc: Andrew Morton <[email protected]>
> Cc: [email protected]
> Signed-off-by: David Hildenbrand <[email protected]>

Slight inconsistency with the page cache droppint but nothing earth
shattering.

Acked-by: Michal Hocko <[email protected]>

> ---
> fs/drop_caches.c | 4 +---
> mm/vmscan.c | 1 +
> 2 files changed, 2 insertions(+), 3 deletions(-)
>
> diff --git a/fs/drop_caches.c b/fs/drop_caches.c
> index d31b6c72b476..a042da782fcd 100644
> --- a/fs/drop_caches.c
> +++ b/fs/drop_caches.c
> @@ -61,10 +61,8 @@ int drop_caches_sysctl_handler(struct ctl_table *table, int write,
> iterate_supers(drop_pagecache_sb, NULL);
> count_vm_event(DROP_PAGECACHE);
> }
> - if (sysctl_drop_caches & 2) {
> + if (sysctl_drop_caches & 2)
> drop_slab();
> - count_vm_event(DROP_SLAB);
> - }
> if (!stfu) {
> pr_info("%s (%d): drop_caches: %d\n",
> current->comm, task_pid_nr(current),
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index 5a6445e86328..c3e53502a84a 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -726,6 +726,7 @@ void drop_slab(void)
>
> for_each_online_node(nid)
> drop_slab_node(nid);
> + count_vm_event(DROP_SLAB);
> }
>
> static inline int is_page_cache_freeable(struct page *page)
> --
> 2.23.0

--
Michal Hocko
SUSE Labs

2020-02-25 14:28:19

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 08/13] mm/memory_hotplug: Introduce offline_and_remove_memory()

On 25.02.20 15:11, Michal Hocko wrote:
> On Thu 12-12-19 18:11:32, David Hildenbrand wrote:
>> virtio-mem wants to offline and remove a memory block once it unplugged
>> all subblocks (e.g., using alloc_contig_range()). Let's provide
>> an interface to do that from a driver. virtio-mem already supports to
>> offline partially unplugged memory blocks. Offlining a fully unplugged
>> memory block will not require to migrate any pages. All unplugged
>> subblocks are PageOffline() and have a reference count of 0 - so
>> offlining code will simply skip them.
>>
>> All we need an interface to trigger the "offlining" and the removing in a
>> single operation - to make sure the memory block cannot get onlined by
>> user space again before it gets removed.
>
> Why does that matter? Is it really likely that the userspace would
> interfere? What would be the scenario?

I guess it's not that relevant after all (I think this comment dates
back to the times where we didn't have try_remove_memory() and could
actually BUG_ON() in remove_memory() if there would have been a race).
Can drop that part.

>
> Or is still mostly about not requiring callers to open code this general
> patter?

From kernel module context, I cannot get access to the actual memory
block device (find_memory_block()) and call the device_unregister().

Especially, also the device hotplug lock is not exported. So this is a
clean helper function to be used from kernel module context. (e.g., also
hyper-v showed interest for using that)

--
Thanks,

David / dhildenb

2020-02-25 15:10:01

by Michal Hocko

[permalink] [raw]
Subject: Re: [PATCH RFC v4 12/13] mm/vmscan: Export drop_slab() and drop_slab_node()

On Thu 12-12-19 18:11:36, David Hildenbrand wrote:
> We already have a way to trigger reclaiming of all reclaimable slab objects
> from user space (echo 2 > /proc/sys/vm/drop_caches). Let's allow drivers
> to also trigger this when they really want to make progress and know what
> they are doing.

I cannot say I would be fan of this. This is a global action with user
visible performance impact. I am worried that we will find out that all
sorts of drivers have a very good idea that dropping slab caches is
going to help their problem whatever it is. We have seen the same patter
in the userspace already and that is the reason we are logging the usage
to the log and count invocations in the counter.

> virtio-mem wants to use these functions when it failed to unplug memory
> for quite some time (e.g., after 30 minutes). It will then try to
> free up reclaimable objects by dropping the slab caches every now and
> then (e.g., every 30 minutes) as long as necessary. There will be a way to
> disable this feature and info messages will be logged.
>
> In the future, we want to have a drop_slab_range() functionality
> instead. Memory offlining code has similar demands and also other
> alloc_contig_range() users (e.g., gigantic pages) could make good use of
> this feature. Adding it, however, requires more work/thought.

We already do have a memory_notify(MEM_GOING_OFFLINE) for that purpose
and slab allocator implements a callback (slab_mem_going_offline_callback).
The callback is quite dumb and it doesn't really try to free objects
from the given memory range or even try to drop active objects which
might turn out to be hard but this sounds like a more robust way to
achieve what you want.

> Cc: Andrew Morton <[email protected]>
> Cc: Michal Hocko <[email protected]>
> Signed-off-by: David Hildenbrand <[email protected]>
> ---
> include/linux/mm.h | 4 ++--
> mm/vmscan.c | 2 ++
> 2 files changed, 4 insertions(+), 2 deletions(-)
>
> diff --git a/include/linux/mm.h b/include/linux/mm.h
> index 64799c5cb39f..483300f58be8 100644
> --- a/include/linux/mm.h
> +++ b/include/linux/mm.h
> @@ -2706,8 +2706,8 @@ int drop_caches_sysctl_handler(struct ctl_table *, int,
> void __user *, size_t *, loff_t *);
> #endif
>
> -void drop_slab(void);
> -void drop_slab_node(int nid);
> +extern void drop_slab(void);
> +extern void drop_slab_node(int nid);
>
> #ifndef CONFIG_MMU
> #define randomize_va_space 0
> diff --git a/mm/vmscan.c b/mm/vmscan.c
> index c3e53502a84a..4e1cdaaec5e6 100644
> --- a/mm/vmscan.c
> +++ b/mm/vmscan.c
> @@ -719,6 +719,7 @@ void drop_slab_node(int nid)
> } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL)) != NULL);
> } while (freed > 10);
> }
> +EXPORT_SYMBOL(drop_slab_node);
>
> void drop_slab(void)
> {
> @@ -728,6 +729,7 @@ void drop_slab(void)
> drop_slab_node(nid);
> count_vm_event(DROP_SLAB);
> }
> +EXPORT_SYMBOL(drop_slab);
>
> static inline int is_page_cache_freeable(struct page *page)
> {
> --
> 2.23.0

--
Michal Hocko
SUSE Labs

2020-02-25 15:12:56

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 12/13] mm/vmscan: Export drop_slab() and drop_slab_node()

On 25.02.20 15:58, Michal Hocko wrote:
> On Thu 12-12-19 18:11:36, David Hildenbrand wrote:
>> We already have a way to trigger reclaiming of all reclaimable slab objects
>> from user space (echo 2 > /proc/sys/vm/drop_caches). Let's allow drivers
>> to also trigger this when they really want to make progress and know what
>> they are doing.
>
> I cannot say I would be fan of this. This is a global action with user
> visible performance impact. I am worried that we will find out that all
> sorts of drivers have a very good idea that dropping slab caches is
> going to help their problem whatever it is. We have seen the same patter
> in the userspace already and that is the reason we are logging the usage
> to the log and count invocations in the counter.

Yeah, I decided to hold back patch 11-13 for the v1 (which I am planning
to post in March after more testing). What we really want is to make
memory offlining an alloc_contig_range() work better with reclaimable
objects.

>
>> virtio-mem wants to use these functions when it failed to unplug memory
>> for quite some time (e.g., after 30 minutes). It will then try to
>> free up reclaimable objects by dropping the slab caches every now and
>> then (e.g., every 30 minutes) as long as necessary. There will be a way to
>> disable this feature and info messages will be logged.
>>
>> In the future, we want to have a drop_slab_range() functionality
>> instead. Memory offlining code has similar demands and also other
>> alloc_contig_range() users (e.g., gigantic pages) could make good use of
>> this feature. Adding it, however, requires more work/thought.
>
> We already do have a memory_notify(MEM_GOING_OFFLINE) for that purpose
> and slab allocator implements a callback (slab_mem_going_offline_callback).
> The callback is quite dumb and it doesn't really try to free objects
> from the given memory range or even try to drop active objects which
> might turn out to be hard but this sounds like a more robust way to
> achieve what you want.

Two things:

1. memory_notify(MEM_GOING_OFFLINE) is called after trying to isolate
the page range and checking if we only have movable pages. Won't help
much I guess.

2. alloc_contig_range() won't benefit from that.

Something like drop_slab_range() would be better, and calling it from
the right spots in the core (e.g., set_migratetype_isolate() see below).

Especially, have a look at mm/page_isolation.c:set_migratetype_isolate()

"FIXME: Now, memory hotplug doesn't call shrink_slab() by itself. We
just check MOVABLE pages."

--
Thanks,

David / dhildenb

2020-02-25 17:26:54

by Michal Hocko

[permalink] [raw]
Subject: Re: [PATCH RFC v4 12/13] mm/vmscan: Export drop_slab() and drop_slab_node()

On Tue 25-02-20 16:09:29, David Hildenbrand wrote:
> On 25.02.20 15:58, Michal Hocko wrote:
> > On Thu 12-12-19 18:11:36, David Hildenbrand wrote:
> >> We already have a way to trigger reclaiming of all reclaimable slab objects
> >> from user space (echo 2 > /proc/sys/vm/drop_caches). Let's allow drivers
> >> to also trigger this when they really want to make progress and know what
> >> they are doing.
> >
> > I cannot say I would be fan of this. This is a global action with user
> > visible performance impact. I am worried that we will find out that all
> > sorts of drivers have a very good idea that dropping slab caches is
> > going to help their problem whatever it is. We have seen the same patter
> > in the userspace already and that is the reason we are logging the usage
> > to the log and count invocations in the counter.
>
> Yeah, I decided to hold back patch 11-13 for the v1 (which I am planning
> to post in March after more testing). What we really want is to make
> memory offlining an alloc_contig_range() work better with reclaimable
> objects.
>
> >
> >> virtio-mem wants to use these functions when it failed to unplug memory
> >> for quite some time (e.g., after 30 minutes). It will then try to
> >> free up reclaimable objects by dropping the slab caches every now and
> >> then (e.g., every 30 minutes) as long as necessary. There will be a way to
> >> disable this feature and info messages will be logged.
> >>
> >> In the future, we want to have a drop_slab_range() functionality
> >> instead. Memory offlining code has similar demands and also other
> >> alloc_contig_range() users (e.g., gigantic pages) could make good use of
> >> this feature. Adding it, however, requires more work/thought.
> >
> > We already do have a memory_notify(MEM_GOING_OFFLINE) for that purpose
> > and slab allocator implements a callback (slab_mem_going_offline_callback).
> > The callback is quite dumb and it doesn't really try to free objects
> > from the given memory range or even try to drop active objects which
> > might turn out to be hard but this sounds like a more robust way to
> > achieve what you want.
>
> Two things:
>
> 1. memory_notify(MEM_GOING_OFFLINE) is called after trying to isolate
> the page range and checking if we only have movable pages. Won't help
> much I guess.

You are right, I have missed that. Can we reorder those two calls?

> 2. alloc_contig_range() won't benefit from that.

True.

> Something like drop_slab_range() would be better, and calling it from
> the right spots in the core (e.g., set_migratetype_isolate() see below).
>
> Especially, have a look at mm/page_isolation.c:set_migratetype_isolate()
>
> "FIXME: Now, memory hotplug doesn't call shrink_slab() by itself. We
> just check MOVABLE pages."

shrink_slab is really a large hammer for this purpose. The notifier
mechanism sounds more appropriate to me. If that means to move it
outside of its current position then let's try to experiment with that.
But there is a long route to have per pfn range reclaim.
--
Michal Hocko
SUSE Labs

2020-02-25 17:30:06

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 12/13] mm/vmscan: Export drop_slab() and drop_slab_node()

On 25.02.20 18:06, Michal Hocko wrote:
> On Tue 25-02-20 16:09:29, David Hildenbrand wrote:
>> On 25.02.20 15:58, Michal Hocko wrote:
>>> On Thu 12-12-19 18:11:36, David Hildenbrand wrote:
>>>> We already have a way to trigger reclaiming of all reclaimable slab objects
>>>> from user space (echo 2 > /proc/sys/vm/drop_caches). Let's allow drivers
>>>> to also trigger this when they really want to make progress and know what
>>>> they are doing.
>>>
>>> I cannot say I would be fan of this. This is a global action with user
>>> visible performance impact. I am worried that we will find out that all
>>> sorts of drivers have a very good idea that dropping slab caches is
>>> going to help their problem whatever it is. We have seen the same patter
>>> in the userspace already and that is the reason we are logging the usage
>>> to the log and count invocations in the counter.
>>
>> Yeah, I decided to hold back patch 11-13 for the v1 (which I am planning
>> to post in March after more testing). What we really want is to make
>> memory offlining an alloc_contig_range() work better with reclaimable
>> objects.
>>
>>>
>>>> virtio-mem wants to use these functions when it failed to unplug memory
>>>> for quite some time (e.g., after 30 minutes). It will then try to
>>>> free up reclaimable objects by dropping the slab caches every now and
>>>> then (e.g., every 30 minutes) as long as necessary. There will be a way to
>>>> disable this feature and info messages will be logged.
>>>>
>>>> In the future, we want to have a drop_slab_range() functionality
>>>> instead. Memory offlining code has similar demands and also other
>>>> alloc_contig_range() users (e.g., gigantic pages) could make good use of
>>>> this feature. Adding it, however, requires more work/thought.
>>>
>>> We already do have a memory_notify(MEM_GOING_OFFLINE) for that purpose
>>> and slab allocator implements a callback (slab_mem_going_offline_callback).
>>> The callback is quite dumb and it doesn't really try to free objects
>>> from the given memory range or even try to drop active objects which
>>> might turn out to be hard but this sounds like a more robust way to
>>> achieve what you want.
>>
>> Two things:
>>
>> 1. memory_notify(MEM_GOING_OFFLINE) is called after trying to isolate
>> the page range and checking if we only have movable pages. Won't help
>> much I guess.
>
> You are right, I have missed that. Can we reorder those two calls?

AFAIK no (would have to look up the details, but there was a good reason
for the order, e.g., avoid races with other users of page isolation like
alloc_contig_range()).

Especially, "[PATCH RFC v4 06/13] mm: Allow to offline unmovable
PageOffline() pages via MEM_GOING_OFFLINE" (which is still impatiently
waiting for an ACK ;) ) also works around that ordering issue in a way
we discussed back then.

--
Thanks,

David / dhildenb

2020-02-25 18:27:53

by Alexander Duyck

[permalink] [raw]
Subject: Re: [PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages via MEM_GOING_OFFLINE

On Thu, 2019-12-12 at 18:11 +0100, David Hildenbrand wrote:
> virtio-mem wants to allow to offline memory blocks of which some parts
> were unplugged (allocated via alloc_contig_range()), especially, to later
> offline and remove completely unplugged memory blocks. The important part
> is that PageOffline() has to remain set until the section is offline, so
> these pages will never get accessed (e.g., when dumping). The pages should
> not be handed back to the buddy (which would require clearing PageOffline()
> and result in issues if offlining fails and the pages are suddenly in the
> buddy).
>
> Let's allow to do that by allowing to isolate any PageOffline() page
> when offlining. This way, we can reach the memory hotplug notifier
> MEM_GOING_OFFLINE, where the driver can signal that he is fine with
> offlining this page by dropping its reference count. PageOffline() pages
> with a reference count of 0 can then be skipped when offlining the
> pages (like if they were free, however they are not in the buddy).
>
> Anybody who uses PageOffline() pages and does not agree to offline them
> (e.g., Hyper-V balloon, XEN balloon, VMWare balloon for 2MB pages) will not
> decrement the reference count and make offlining fail when trying to
> migrate such an unmovable page. So there should be no observerable change.
> Same applies to balloon compaction users (movable PageOffline() pages), the
> pages will simply be migrated.
>
> Note 1: If offlining fails, a driver has to increment the reference
> count again in MEM_CANCEL_OFFLINE.
>
> Note 2: A driver that makes use of this has to be aware that re-onlining
> the memory block has to be handled by hooking into onlining code
> (online_page_callback_t), resetting the page PageOffline() and
> not giving them to the buddy.
>
> Cc: Andrew Morton <[email protected]>
> Cc: Juergen Gross <[email protected]>
> Cc: Konrad Rzeszutek Wilk <[email protected]>
> Cc: Pavel Tatashin <[email protected]>
> Cc: Alexander Duyck <[email protected]>
> Cc: Vlastimil Babka <[email protected]>
> Cc: Johannes Weiner <[email protected]>
> Cc: Anthony Yznaga <[email protected]>
> Cc: Michal Hocko <[email protected]>
> Cc: Oscar Salvador <[email protected]>
> Cc: Mel Gorman <[email protected]>
> Cc: Mike Rapoport <[email protected]>
> Cc: Dan Williams <[email protected]>
> Cc: Anshuman Khandual <[email protected]>
> Cc: Qian Cai <[email protected]>
> Cc: Pingfan Liu <[email protected]>
> Signed-off-by: David Hildenbrand <[email protected]>
> ---
> include/linux/page-flags.h | 10 ++++++++++
> mm/memory_hotplug.c | 41 ++++++++++++++++++++++++++++----------
> mm/page_alloc.c | 24 ++++++++++++++++++++++
> mm/page_isolation.c | 9 +++++++++
> 4 files changed, 74 insertions(+), 10 deletions(-)
>
> diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h
> index 1bf83c8fcaa7..ac1775082343 100644
> --- a/include/linux/page-flags.h
> +++ b/include/linux/page-flags.h
> @@ -761,6 +761,16 @@ PAGE_TYPE_OPS(Buddy, buddy)
> * not onlined when onlining the section).
> * The content of these pages is effectively stale. Such pages should not
> * be touched (read/write/dump/save) except by their owner.
> + *
> + * If a driver wants to allow to offline unmovable PageOffline() pages without
> + * putting them back to the buddy, it can do so via the memory notifier by
> + * decrementing the reference count in MEM_GOING_OFFLINE and incrementing the
> + * reference count in MEM_CANCEL_OFFLINE. When offlining, the PageOffline()
> + * pages (now with a reference count of zero) are treated like free pages,
> + * allowing the containing memory block to get offlined. A driver that
> + * relies on this feature is aware that re-onlining the memory block will
> + * require to re-set the pages PageOffline() and not giving them to the
> + * buddy via online_page_callback_t.
> */
> PAGE_TYPE_OPS(Offline, offline)
>
> diff --git a/mm/memory_hotplug.c b/mm/memory_hotplug.c
> index fc617ad6f035..da01453a04e6 100644
> --- a/mm/memory_hotplug.c
> +++ b/mm/memory_hotplug.c
> @@ -1231,11 +1231,15 @@ int test_pages_in_a_zone(unsigned long start_pfn, unsigned long end_pfn,
>
> /*
> * Scan pfn range [start,end) to find movable/migratable pages (LRU pages,
> - * non-lru movable pages and hugepages). We scan pfn because it's much
> - * easier than scanning over linked list. This function returns the pfn
> - * of the first found movable page if it's found, otherwise 0.
> + * non-lru movable pages and hugepages).
> + *
> + * Returns:
> + * 0 in case a movable page is found and movable_pfn was updated.
> + * -ENOENT in case no movable page was found.
> + * -EBUSY in case a definetly unmovable page was found.
> */
> -static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
> +static int scan_movable_pages(unsigned long start, unsigned long end,
> + unsigned long *movable_pfn)
> {
> unsigned long pfn;
>
> @@ -1247,18 +1251,29 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
> continue;
> page = pfn_to_page(pfn);
> if (PageLRU(page))
> - return pfn;
> + goto found;
> if (__PageMovable(page))
> - return pfn;
> + goto found;
> +
> + /*
> + * Unmovable PageOffline() pages where somebody still holds
> + * a reference count (after MEM_GOING_OFFLINE) can definetly
> + * not be offlined.
> + */
> + if (PageOffline(page) && page_count(page))
> + return -EBUSY;

So the comment confused me a bit because technically this function isn't
about offlining memory, it is about finding movable pages. I had to do a
bit of digging to find the only consumer is __offline_pages, but if we are
going to talk about "offlining" instead of "moving" in this function it
might make sense to rename it.

>
> if (!PageHuge(page))
> continue;
> head = compound_head(page);
> if (page_huge_active(head))
> - return pfn;
> + goto found;
> skip = compound_nr(head) - (page - head);
> pfn += skip - 1;
> }
> + return -ENOENT;
> +found:
> + *movable_pfn = pfn;
> return 0;
> }

So I am looking at this function and it seems like your change completely
changes the behavior. The code before would walk the entire range and if
at least 1 page was available to move you would return the PFN of that
page. Now what seems to happen is that you will return -EBUSY as soon as
you encounter an offline page with a page count. I would think that would
slow down the offlining process since you have made the Unmovable
PageOffline() page a head of line blocker that you have to wait to get
around.

Would it perhaps make more sense to add a return value initialized to
ENOENT, and if you encounter one of these offline pages you change the
return value to EBUSY, and then if you walk through the entire list
without finding a movable page you just return the value?

Otherwise you might want to add a comment explaining why the function
should stall instead of skipping over the unmovable section that will
hopefully become movable later.

> @@ -1528,7 +1543,8 @@ static int __ref __offline_pages(unsigned long start_pfn,
> }
>
> do {
> - for (pfn = start_pfn; pfn;) {
> + pfn = start_pfn;
> + do {
> if (signal_pending(current)) {
> ret = -EINTR;
> reason = "signal backoff";
> @@ -1538,14 +1554,19 @@ static int __ref __offline_pages(unsigned long start_pfn,
> cond_resched();
> lru_add_drain_all();
>
> - pfn = scan_movable_pages(pfn, end_pfn);
> - if (pfn) {
> + ret = scan_movable_pages(pfn, end_pfn, &pfn);
> + if (!ret) {
> /*
> * TODO: fatal migration failures should bail
> * out
> */
> do_migrate_range(pfn, end_pfn);
> }
> + } while (!ret);
> +
> + if (ret != -ENOENT) {
> + reason = "unmovable page";
> + goto failed_removal_isolated;
> }
>
> /*
> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> index 5334decc9e06..840c0bbe2d9f 100644
> --- a/mm/page_alloc.c
> +++ b/mm/page_alloc.c
> @@ -8256,6 +8256,19 @@ bool has_unmovable_pages(struct zone *zone, struct page *page, int count,
> if ((flags & MEMORY_OFFLINE) && PageHWPoison(page))
> continue;
>
> + /*
> + * We treat all PageOffline() pages as movable when offlining
> + * to give drivers a chance to decrement their reference count
> + * in MEM_GOING_OFFLINE in order to signalize that these pages

You can probably just use "signal" or "indicate" instead of "signalize".

> + * can be offlined as there are no direct references anymore.
> + * For actually unmovable PageOffline() where the driver does
> + * not support this, we will fail later when trying to actually
> + * move these pages that still have a reference count > 0.
> + * (false negatives in this function only)
> + */
> + if ((flags & MEMORY_OFFLINE) && PageOffline(page))
> + continue;
> +
> if (__PageMovable(page))
> continue;
>
> @@ -8683,6 +8696,17 @@ __offline_isolated_pages(unsigned long start_pfn, unsigned long end_pfn)
> offlined_pages++;
> continue;
> }
> + /*
> + * At this point all remaining PageOffline() pages have a
> + * reference count of 0 and can simply be skipped.
> + */
> + if (PageOffline(page)) {
> + BUG_ON(page_count(page));
> + BUG_ON(PageBuddy(page));
> + pfn++;
> + offlined_pages++;
> + continue;
> + }
>
> BUG_ON(page_count(page));
> BUG_ON(!PageBuddy(page));
> diff --git a/mm/page_isolation.c b/mm/page_isolation.c
> index 04ee1663cdbe..43b4dabfedc8 100644
> --- a/mm/page_isolation.c
> +++ b/mm/page_isolation.c
> @@ -170,6 +170,7 @@ __first_valid_page(unsigned long pfn, unsigned long nr_pages)
> * a bit mask)
> * MEMORY_OFFLINE - isolate to offline (!allocate) memory
> * e.g., skip over PageHWPoison() pages
> + * and PageOffline() pages.
> * REPORT_FAILURE - report details about the failure to
> * isolate the range
> *
> @@ -278,6 +279,14 @@ __test_page_isolated_in_pageblock(unsigned long pfn, unsigned long end_pfn,
> else if ((flags & MEMORY_OFFLINE) && PageHWPoison(page))
> /* A HWPoisoned page cannot be also PageBuddy */
> pfn++;
> + else if ((flags & MEMORY_OFFLINE) && PageOffline(page) &&
> + !page_count(page))
> + /*
> + * The responsible driver agreed to offline
> + * PageOffline() pages by dropping its reference in
> + * MEM_GOING_OFFLINE.
> + */
> + pfn++;
> else
> break;
> }


2020-02-25 19:31:33

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages via MEM_GOING_OFFLINE

>> /*
>> * Scan pfn range [start,end) to find movable/migratable pages (LRU pages,
>> - * non-lru movable pages and hugepages). We scan pfn because it's much
>> - * easier than scanning over linked list. This function returns the pfn
>> - * of the first found movable page if it's found, otherwise 0.
>> + * non-lru movable pages and hugepages).
>> + *
>> + * Returns:
>> + * 0 in case a movable page is found and movable_pfn was updated.
>> + * -ENOENT in case no movable page was found.
>> + * -EBUSY in case a definetly unmovable page was found.
>> */
>> -static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
>> +static int scan_movable_pages(unsigned long start, unsigned long end,
>> + unsigned long *movable_pfn)
>> {
>> unsigned long pfn;
>>
>> @@ -1247,18 +1251,29 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
>> continue;
>> page = pfn_to_page(pfn);
>> if (PageLRU(page))
>> - return pfn;
>> + goto found;
>> if (__PageMovable(page))
>> - return pfn;
>> + goto found;
>> +
>> + /*
>> + * Unmovable PageOffline() pages where somebody still holds
>> + * a reference count (after MEM_GOING_OFFLINE) can definetly
>> + * not be offlined.
>> + */
>> + if (PageOffline(page) && page_count(page))
>> + return -EBUSY;
>
> So the comment confused me a bit because technically this function isn't
> about offlining memory, it is about finding movable pages. I had to do a
> bit of digging to find the only consumer is __offline_pages, but if we are
> going to talk about "offlining" instead of "moving" in this function it
> might make sense to rename it.

Well, it's contained in memory_hotplug.c, and the only user of moving
pages around in there is offlining code :) And it's job is to locate
movable pages, skip over some (temporary? unmovable ones) and (now)
indicate definitely unmovable ones.

Any idea for a better name?
scan_movable_pages_and_stop_on_definitely_unmovable() is not so nice :)


>
>>
>> if (!PageHuge(page))
>> continue;
>> head = compound_head(page);
>> if (page_huge_active(head))
>> - return pfn;
>> + goto found;
>> skip = compound_nr(head) - (page - head);
>> pfn += skip - 1;
>> }
>> + return -ENOENT;
>> +found:
>> + *movable_pfn = pfn;
>> return 0;
>> }
>
> So I am looking at this function and it seems like your change completely
> changes the behavior. The code before would walk the entire range and if
> at least 1 page was available to move you would return the PFN of that
> page. Now what seems to happen is that you will return -EBUSY as soon as
> you encounter an offline page with a page count. I would think that would
> slow down the offlining process since you have made the Unmovable
> PageOffline() page a head of line blocker that you have to wait to get
> around.

So, the comment says "Unmovable PageOffline() pages where somebody still
holds a reference count (after MEM_GOING_OFFLINE) can definitely not be
offlined". And the doc "-EBUSY in case a definitely unmovable page was
found."

So why would this make offlining slow? Offlining will be aborted,
because offlining is not possible.

Please note that this is the exact old behavior, where isolating the
page range would have failed directly and offlining would have been
aborted early. The old offlining failure in the case in the offlining
path would have been "failure to isolate range".

Also, note that the users of PageOffline() with unmovable pages are very
rare (only balloon drivers for now).

>
> Would it perhaps make more sense to add a return value initialized to
> ENOENT, and if you encounter one of these offline pages you change the
> return value to EBUSY, and then if you walk through the entire list
> without finding a movable page you just return the value?

Did you have a look in which context this function is used, especially
[1] and [2]?

>
> Otherwise you might want to add a comment explaining why the function
> should stall instead of skipping over the unmovable section that will
> hopefully become movable later.

So we have "-EBUSY in case a definitely unmovable page was found.". Do
you have a better suggestion?

>
>> @@ -1528,7 +1543,8 @@ static int __ref __offline_pages(unsigned long start_pfn,
>> }
>>
>> do {
>> - for (pfn = start_pfn; pfn;) {
>> + pfn = start_pfn;
>> + do {
>> if (signal_pending(current)) {
>> ret = -EINTR;
>> reason = "signal backoff";
>> @@ -1538,14 +1554,19 @@ static int __ref __offline_pages(unsigned long start_pfn,
>> cond_resched();
>> lru_add_drain_all();
>>
>> - pfn = scan_movable_pages(pfn, end_pfn);
>> - if (pfn) {
>> + ret = scan_movable_pages(pfn, end_pfn, &pfn);
>> + if (!ret) {
>> /*
>> * TODO: fatal migration failures should bail
>> * out
>> */
>> do_migrate_range(pfn, end_pfn);
>> }

[1] we detect a definite offlining blocker and

>> + } while (!ret);
>> +
>> + if (ret != -ENOENT) {
>> + reason = "unmovable page";

[2] we abort offlining

>> + goto failed_removal_isolated;
>> }
>>
>> /*
>> diff --git a/mm/page_alloc.c b/mm/page_alloc.c
>> index 5334decc9e06..840c0bbe2d9f 100644
>> --- a/mm/page_alloc.c
>> +++ b/mm/page_alloc.c
>> @@ -8256,6 +8256,19 @@ bool has_unmovable_pages(struct zone *zone, struct page *page, int count,
>> if ((flags & MEMORY_OFFLINE) && PageHWPoison(page))
>> continue;
>>
>> + /*
>> + * We treat all PageOffline() pages as movable when offlining
>> + * to give drivers a chance to decrement their reference count
>> + * in MEM_GOING_OFFLINE in order to signalize that these pages
>
> You can probably just use "signal" or "indicate" instead of "signalize".

Makes sense, "indicate" it is!

Thanks!

--
Thanks,

David / dhildenb

2020-02-25 21:46:39

by Alexander Duyck

[permalink] [raw]
Subject: Re: [PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages via MEM_GOING_OFFLINE

On Tue, 2020-02-25 at 19:49 +0100, David Hildenbrand wrote:
> > > /*
> > > * Scan pfn range [start,end) to find movable/migratable pages (LRU pages,
> > > - * non-lru movable pages and hugepages). We scan pfn because it's much
> > > - * easier than scanning over linked list. This function returns the pfn
> > > - * of the first found movable page if it's found, otherwise 0.
> > > + * non-lru movable pages and hugepages).
> > > + *
> > > + * Returns:
> > > + * 0 in case a movable page is found and movable_pfn was updated.
> > > + * -ENOENT in case no movable page was found.
> > > + * -EBUSY in case a definetly unmovable page was found.
> > > */
> > > -static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
> > > +static int scan_movable_pages(unsigned long start, unsigned long end,
> > > + unsigned long *movable_pfn)
> > > {
> > > unsigned long pfn;
> > >
> > > @@ -1247,18 +1251,29 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
> > > continue;
> > > page = pfn_to_page(pfn);
> > > if (PageLRU(page))
> > > - return pfn;
> > > + goto found;
> > > if (__PageMovable(page))
> > > - return pfn;
> > > + goto found;
> > > +
> > > + /*
> > > + * Unmovable PageOffline() pages where somebody still holds
> > > + * a reference count (after MEM_GOING_OFFLINE) can definetly
> > > + * not be offlined.
> > > + */
> > > + if (PageOffline(page) && page_count(page))
> > > + return -EBUSY;
> >
> > So the comment confused me a bit because technically this function isn't
> > about offlining memory, it is about finding movable pages. I had to do a
> > bit of digging to find the only consumer is __offline_pages, but if we are
> > going to talk about "offlining" instead of "moving" in this function it
> > might make sense to rename it.
>
> Well, it's contained in memory_hotplug.c, and the only user of moving
> pages around in there is offlining code :) And it's job is to locate
> movable pages, skip over some (temporary? unmovable ones) and (now)
> indicate definitely unmovable ones.
>
> Any idea for a better name?
> scan_movable_pages_and_stop_on_definitely_unmovable() is not so nice :)

I dunno. What I was getting at is that the wording here would make it
clearer if you simply stated that these pages "can definately not be
moved". Saying you cannot offline a page that is PageOffline seems kind of
redundant, then again calling it an Unmovable and then saying it cannot be
moves is also redundant I suppose. In the end you don't move them, but
they can be switched to offline if the page count hits 0. When that
happens you simply end up skipping over them in the code for
__test_page_isolated_in_pageblock and __offline_isolated_pages.

> > >
> > > if (!PageHuge(page))
> > > continue;
> > > head = compound_head(page);
> > > if (page_huge_active(head))
> > > - return pfn;
> > > + goto found;
> > > skip = compound_nr(head) - (page - head);
> > > pfn += skip - 1;
> > > }
> > > + return -ENOENT;
> > > +found:
> > > + *movable_pfn = pfn;
> > > return 0;
> > > }
> >
> > So I am looking at this function and it seems like your change completely
> > changes the behavior. The code before would walk the entire range and if
> > at least 1 page was available to move you would return the PFN of that
> > page. Now what seems to happen is that you will return -EBUSY as soon as
> > you encounter an offline page with a page count. I would think that would
> > slow down the offlining process since you have made the Unmovable
> > PageOffline() page a head of line blocker that you have to wait to get
> > around.
>
> So, the comment says "Unmovable PageOffline() pages where somebody still
> holds a reference count (after MEM_GOING_OFFLINE) can definitely not be
> offlined". And the doc "-EBUSY in case a definitely unmovable page was
> found."
>
> So why would this make offlining slow? Offlining will be aborted,
> because offlining is not possible.

Okay, my reading of the code was a bit off. In my mind I was thinking you
were effectively treating it almost like an EAGAIN and continuing the
loop. I didn't catch the part where you had rewritten the for loop as a
do-while in __offline_pages.

> Please note that this is the exact old behavior, where isolating the
> page range would have failed directly and offlining would have been
> aborted early. The old offlining failure in the case in the offlining
> path would have been "failure to isolate range".
>
> Also, note that the users of PageOffline() with unmovable pages are very
> rare (only balloon drivers for now).
>
> > Would it perhaps make more sense to add a return value initialized to
> > ENOENT, and if you encounter one of these offline pages you change the
> > return value to EBUSY, and then if you walk through the entire list
> > without finding a movable page you just return the value?
>
> Did you have a look in which context this function is used, especially
> [1] and [2]?
>
> > Otherwise you might want to add a comment explaining why the function
> > should stall instead of skipping over the unmovable section that will
> > hopefully become movable later.
>
> So we have "-EBUSY in case a definitely unmovable page was found.". Do
> you have a better suggestion?
>
> > > @@ -1528,7 +1543,8 @@ static int __ref __offline_pages(unsigned long start_pfn,
> > > }
> > >
> > > do {
> > > - for (pfn = start_pfn; pfn;) {
> > > + pfn = start_pfn;
> > > + do {
> > > if (signal_pending(current)) {
> > > ret = -EINTR;
> > > reason = "signal backoff";
> > > @@ -1538,14 +1554,19 @@ static int __ref __offline_pages(unsigned long start_pfn,
> > > cond_resched();
> > > lru_add_drain_all();
> > >
> > > - pfn = scan_movable_pages(pfn, end_pfn);
> > > - if (pfn) {
> > > + ret = scan_movable_pages(pfn, end_pfn, &pfn);
> > > + if (!ret) {
> > > /*
> > > * TODO: fatal migration failures should bail
> > > * out
> > > */
> > > do_migrate_range(pfn, end_pfn);
> > > }
>
> [1] we detect a definite offlining blocker and
>
> > > + } while (!ret);
> > > +
> > > + if (ret != -ENOENT) {
> > > + reason = "unmovable page";
>
> [2] we abort offlining
>
> > > + goto failed_removal_isolated;
> > > }
> > >
> > > /*

Yeah, this is the piece I misread. I knew the loop this was in previously
was looping when returning -ENOENT so for some reason I had it in my head
that you were still looping on -EBUSY.

So the one question I would have is if at this point are we guaranteed
that the balloon drivers have already taken care of the page count for all
the pages they set to PageOffline? Based on the patch description I was
thinking that this was going to be looping for a while waiting for the
driver to clear the pages and then walking through them at the end of the
loop via check_pages_isolated_cb.

> > > diff --git a/mm/page_alloc.c b/mm/page_alloc.c
> > > index 5334decc9e06..840c0bbe2d9f 100644
> > > --- a/mm/page_alloc.c
> > > +++ b/mm/page_alloc.c
> > > @@ -8256,6 +8256,19 @@ bool has_unmovable_pages(struct zone *zone, struct page *page, int count,
> > > if ((flags & MEMORY_OFFLINE) && PageHWPoison(page))
> > > continue;
> > >
> > > + /*
> > > + * We treat all PageOffline() pages as movable when offlining
> > > + * to give drivers a chance to decrement their reference count
> > > + * in MEM_GOING_OFFLINE in order to signalize that these pages
> >
> > You can probably just use "signal" or "indicate" instead of "signalize".
>
> Makes sense, "indicate" it is!
>
> Thanks!
>

No problem.

2020-02-25 22:20:08

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages via MEM_GOING_OFFLINE

On 25.02.20 22:46, Alexander Duyck wrote:
> On Tue, 2020-02-25 at 19:49 +0100, David Hildenbrand wrote:
>>>> /*
>>>> * Scan pfn range [start,end) to find movable/migratable pages (LRU pages,
>>>> - * non-lru movable pages and hugepages). We scan pfn because it's much
>>>> - * easier than scanning over linked list. This function returns the pfn
>>>> - * of the first found movable page if it's found, otherwise 0.
>>>> + * non-lru movable pages and hugepages).
>>>> + *
>>>> + * Returns:
>>>> + * 0 in case a movable page is found and movable_pfn was updated.
>>>> + * -ENOENT in case no movable page was found.
>>>> + * -EBUSY in case a definetly unmovable page was found.
>>>> */
>>>> -static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
>>>> +static int scan_movable_pages(unsigned long start, unsigned long end,
>>>> + unsigned long *movable_pfn)
>>>> {
>>>> unsigned long pfn;
>>>>
>>>> @@ -1247,18 +1251,29 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
>>>> continue;
>>>> page = pfn_to_page(pfn);
>>>> if (PageLRU(page))
>>>> - return pfn;
>>>> + goto found;
>>>> if (__PageMovable(page))
>>>> - return pfn;
>>>> + goto found;
>>>> +
>>>> + /*
>>>> + * Unmovable PageOffline() pages where somebody still holds
>>>> + * a reference count (after MEM_GOING_OFFLINE) can definetly
>>>> + * not be offlined.
>>>> + */
>>>> + if (PageOffline(page) && page_count(page))
>>>> + return -EBUSY;
>>>
>>> So the comment confused me a bit because technically this function isn't
>>> about offlining memory, it is about finding movable pages. I had to do a
>>> bit of digging to find the only consumer is __offline_pages, but if we are
>>> going to talk about "offlining" instead of "moving" in this function it
>>> might make sense to rename it.
>>
>> Well, it's contained in memory_hotplug.c, and the only user of moving
>> pages around in there is offlining code :) And it's job is to locate
>> movable pages, skip over some (temporary? unmovable ones) and (now)
>> indicate definitely unmovable ones.
>>
>> Any idea for a better name?
>> scan_movable_pages_and_stop_on_definitely_unmovable() is not so nice :)
>
> I dunno. What I was getting at is that the wording here would make it
> clearer if you simply stated that these pages "can definately not be
> moved". Saying you cannot offline a page that is PageOffline seems kind of
> redundant, then again calling it an Unmovable and then saying it cannot be
> moves is also redundant I suppose. In the end you don't move them, but

So, in summary, there are
- PageOffline() pages that are movable (balloon compaction).
- PageOffline() pages that cannot be moved and cannot be offlined (e.g.,
no balloon compaction enabled, XEN, HyperV, ...) . page_count(page) >=
0
- PageOffline() pages that cannot be moved, but can be offlined.
page_count(page) == 0.


> they can be switched to offline if the page count hits 0. When that
> happens you simply end up skipping over them in the code for
> __test_page_isolated_in_pageblock and __offline_isolated_pages.

Yes. The thing with the wording is that pages with (PageOffline(page) &&
!page_count(page)) can also not really be moved, but they can be skipped
when offlining. If we call that "moving them to /dev/null", then yes,
they can be moved to some degree :)

I can certainly do here e.g.,

/*
* PageOffline() pages that are not marked __PageMovable() and have a
* reference count > 0 (after MEM_GOING_OFFLINE) are definitely
* unmovable. If their reference count would be 0, they could be skipped
* when offlining memory sections.
*/

And maybe I'll add to the function doc, that unmovable pages that are
skipped in this function can include pages that can be skipped when
offlining (moving them to nirvana).

Other suggestions?

[...]

>>
>> [1] we detect a definite offlining blocker and
>>
>>>> + } while (!ret);
>>>> +
>>>> + if (ret != -ENOENT) {
>>>> + reason = "unmovable page";
>>
>> [2] we abort offlining
>>
>>>> + goto failed_removal_isolated;
>>>> }
>>>>
>>>> /*
>
> Yeah, this is the piece I misread. I knew the loop this was in previously
> was looping when returning -ENOENT so for some reason I had it in my head
> that you were still looping on -EBUSY.

Ah okay, I see. Yeah, that wouldn't make sense for the use case I have :)

>
> So the one question I would have is if at this point are we guaranteed
> that the balloon drivers have already taken care of the page count for all
> the pages they set to PageOffline? Based on the patch description I was
> thinking that this was going to be looping for a while waiting for the
> driver to clear the pages and then walking through them at the end of the
> loop via check_pages_isolated_cb.

So, e.g., the patch description states

"Let's allow to do that by allowing to isolate any PageOffline() page
when offlining. This way, we can reach the memory hotplug notifier
MEM_GOING_OFFLINE, where the driver can signal that he is fine with
offlining this page by dropping its reference count."

Any balloon driver that does not allow offlining (e.g., XEN, HyperV,
virtio-balloon), will always have a refcount of (at least) 1. Drivers
that want to make use of that (esp. virtio-mem, but eventually also
HyperV), will drop their refcount via the MEM_GOING_OFFLINE call.

So yes, at this point, all applicable users were notified via
MEM_GOING_OFFLINE and had their chance to decrement the refcount. If
they didn't, offlining will be aborted.

Thanks again!

--
Thanks,

David / dhildenb

2020-02-26 16:27:46

by Alexander Duyck

[permalink] [raw]
Subject: Re: [PATCH RFC v4 06/13] mm: Allow to offline unmovable PageOffline() pages via MEM_GOING_OFFLINE

On Tue, 2020-02-25 at 23:19 +0100, David Hildenbrand wrote:
> On 25.02.20 22:46, Alexander Duyck wrote:
> > On Tue, 2020-02-25 at 19:49 +0100, David Hildenbrand wrote:
> > > > > /*
> > > > > * Scan pfn range [start,end) to find movable/migratable pages (LRU pages,
> > > > > - * non-lru movable pages and hugepages). We scan pfn because it's much
> > > > > - * easier than scanning over linked list. This function returns the pfn
> > > > > - * of the first found movable page if it's found, otherwise 0.
> > > > > + * non-lru movable pages and hugepages).
> > > > > + *
> > > > > + * Returns:
> > > > > + * 0 in case a movable page is found and movable_pfn was updated.
> > > > > + * -ENOENT in case no movable page was found.
> > > > > + * -EBUSY in case a definetly unmovable page was found.
> > > > > */
> > > > > -static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
> > > > > +static int scan_movable_pages(unsigned long start, unsigned long end,
> > > > > + unsigned long *movable_pfn)
> > > > > {
> > > > > unsigned long pfn;
> > > > >
> > > > > @@ -1247,18 +1251,29 @@ static unsigned long scan_movable_pages(unsigned long start, unsigned long end)
> > > > > continue;
> > > > > page = pfn_to_page(pfn);
> > > > > if (PageLRU(page))
> > > > > - return pfn;
> > > > > + goto found;
> > > > > if (__PageMovable(page))
> > > > > - return pfn;
> > > > > + goto found;
> > > > > +
> > > > > + /*
> > > > > + * Unmovable PageOffline() pages where somebody still holds
> > > > > + * a reference count (after MEM_GOING_OFFLINE) can definetly
> > > > > + * not be offlined.
> > > > > + */
> > > > > + if (PageOffline(page) && page_count(page))
> > > > > + return -EBUSY;
> > > >
> > > > So the comment confused me a bit because technically this function isn't
> > > > about offlining memory, it is about finding movable pages. I had to do a
> > > > bit of digging to find the only consumer is __offline_pages, but if we are
> > > > going to talk about "offlining" instead of "moving" in this function it
> > > > might make sense to rename it.
> > >
> > > Well, it's contained in memory_hotplug.c, and the only user of moving
> > > pages around in there is offlining code :) And it's job is to locate
> > > movable pages, skip over some (temporary? unmovable ones) and (now)
> > > indicate definitely unmovable ones.
> > >
> > > Any idea for a better name?
> > > scan_movable_pages_and_stop_on_definitely_unmovable() is not so nice :)
> >
> > I dunno. What I was getting at is that the wording here would make it
> > clearer if you simply stated that these pages "can definately not be
> > moved". Saying you cannot offline a page that is PageOffline seems kind of
> > redundant, then again calling it an Unmovable and then saying it cannot be
> > moves is also redundant I suppose. In the end you don't move them, but
>
> So, in summary, there are
> - PageOffline() pages that are movable (balloon compaction).
> - PageOffline() pages that cannot be moved and cannot be offlined (e.g.,
> no balloon compaction enabled, XEN, HyperV, ...) . page_count(page) >=
> 0
> - PageOffline() pages that cannot be moved, but can be offlined.
> page_count(page) == 0.
>
>
> > they can be switched to offline if the page count hits 0. When that
> > happens you simply end up skipping over them in the code for
> > __test_page_isolated_in_pageblock and __offline_isolated_pages.
>
> Yes. The thing with the wording is that pages with (PageOffline(page) &&
> !page_count(page)) can also not really be moved, but they can be skipped
> when offlining. If we call that "moving them to /dev/null", then yes,
> they can be moved to some degree :)
>
> I can certainly do here e.g.,
>
> /*
> * PageOffline() pages that are not marked __PageMovable() and have a
> * reference count > 0 (after MEM_GOING_OFFLINE) are definitely
> * unmovable. If their reference count would be 0, they could be skipped
> * when offlining memory sections.
> */
>
> And maybe I'll add to the function doc, that unmovable pages that are
> skipped in this function can include pages that can be skipped when
> offlining (moving them to nirvana).
>
> Other suggestions?

No, this sounds good and makes it much clearer.

> [...]
>
> > > [1] we detect a definite offlining blocker and
> > >
> > > > > + } while (!ret);
> > > > > +
> > > > > + if (ret != -ENOENT) {
> > > > > + reason = "unmovable page";
> > >
> > > [2] we abort offlining
> > >
> > > > > + goto failed_removal_isolated;
> > > > > }
> > > > >
> > > > > /*
> >
> > Yeah, this is the piece I misread. I knew the loop this was in previously
> > was looping when returning -ENOENT so for some reason I had it in my head
> > that you were still looping on -EBUSY.
>
> Ah okay, I see. Yeah, that wouldn't make sense for the use case I have :)
>
> > So the one question I would have is if at this point are we guaranteed
> > that the balloon drivers have already taken care of the page count for all
> > the pages they set to PageOffline? Based on the patch description I was
> > thinking that this was going to be looping for a while waiting for the
> > driver to clear the pages and then walking through them at the end of the
> > loop via check_pages_isolated_cb.
>
> So, e.g., the patch description states
>
> "Let's allow to do that by allowing to isolate any PageOffline() page
> when offlining. This way, we can reach the memory hotplug notifier
> MEM_GOING_OFFLINE, where the driver can signal that he is fine with
> offlining this page by dropping its reference count."
>
> Any balloon driver that does not allow offlining (e.g., XEN, HyperV,
> virtio-balloon), will always have a refcount of (at least) 1. Drivers
> that want to make use of that (esp. virtio-mem, but eventually also
> HyperV), will drop their refcount via the MEM_GOING_OFFLINE call.
>
> So yes, at this point, all applicable users were notified via
> MEM_GOING_OFFLINE and had their chance to decrement the refcount. If
> they didn't, offlining will be aborted.
>
> Thanks again!

Thank you as well. I'm still getting up to speed on the inner workings of
much of this and so discussions such as this usually prove to be quite
beneficial for me.

2020-03-02 12:49:14

by Michal Hocko

[permalink] [raw]
Subject: Re: [PATCH RFC v4 08/13] mm/memory_hotplug: Introduce offline_and_remove_memory()

On Tue 25-02-20 15:27:28, David Hildenbrand wrote:
> On 25.02.20 15:11, Michal Hocko wrote:
> > On Thu 12-12-19 18:11:32, David Hildenbrand wrote:
> >> virtio-mem wants to offline and remove a memory block once it unplugged
> >> all subblocks (e.g., using alloc_contig_range()). Let's provide
> >> an interface to do that from a driver. virtio-mem already supports to
> >> offline partially unplugged memory blocks. Offlining a fully unplugged
> >> memory block will not require to migrate any pages. All unplugged
> >> subblocks are PageOffline() and have a reference count of 0 - so
> >> offlining code will simply skip them.
> >>
> >> All we need an interface to trigger the "offlining" and the removing in a
> >> single operation - to make sure the memory block cannot get onlined by
> >> user space again before it gets removed.
> >
> > Why does that matter? Is it really likely that the userspace would
> > interfere? What would be the scenario?
>
> I guess it's not that relevant after all (I think this comment dates
> back to the times where we didn't have try_remove_memory() and could
> actually BUG_ON() in remove_memory() if there would have been a race).
> Can drop that part.
>
> >
> > Or is still mostly about not requiring callers to open code this general
> > patter?
>
> From kernel module context, I cannot get access to the actual memory
> block device (find_memory_block()) and call the device_unregister().
>
> Especially, also the device hotplug lock is not exported. So this is a
> clean helper function to be used from kernel module context. (e.g., also
> hyper-v showed interest for using that)

Fair enough.
--
Michal Hocko
SUSE Labs

2020-03-02 12:53:46

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 08/13] mm/memory_hotplug: Introduce offline_and_remove_memory()

On 02.03.20 13:48, Michal Hocko wrote:
> On Tue 25-02-20 15:27:28, David Hildenbrand wrote:
>> On 25.02.20 15:11, Michal Hocko wrote:
>>> On Thu 12-12-19 18:11:32, David Hildenbrand wrote:
>>>> virtio-mem wants to offline and remove a memory block once it unplugged
>>>> all subblocks (e.g., using alloc_contig_range()). Let's provide
>>>> an interface to do that from a driver. virtio-mem already supports to
>>>> offline partially unplugged memory blocks. Offlining a fully unplugged
>>>> memory block will not require to migrate any pages. All unplugged
>>>> subblocks are PageOffline() and have a reference count of 0 - so
>>>> offlining code will simply skip them.
>>>>
>>>> All we need an interface to trigger the "offlining" and the removing in a
>>>> single operation - to make sure the memory block cannot get onlined by
>>>> user space again before it gets removed.
>>>
>>> Why does that matter? Is it really likely that the userspace would
>>> interfere? What would be the scenario?
>>
>> I guess it's not that relevant after all (I think this comment dates
>> back to the times where we didn't have try_remove_memory() and could
>> actually BUG_ON() in remove_memory() if there would have been a race).
>> Can drop that part.
>>
>>>
>>> Or is still mostly about not requiring callers to open code this general
>>> patter?
>>
>> From kernel module context, I cannot get access to the actual memory
>> block device (find_memory_block()) and call the device_unregister().
>>
>> Especially, also the device hotplug lock is not exported. So this is a
>> clean helper function to be used from kernel module context. (e.g., also
>> hyper-v showed interest for using that)
>
> Fair enough.
>

I'll send a v1 shortly, I rephrased the description to make this clear.
Thanks!

--
Thanks,

David / dhildenb

2020-06-05 09:00:19

by Alex Shi

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory



在 2020/1/9 下午9:48, David Hildenbrand 写道:
> Ping,
>
> I'd love to get some feedback on
>
> a) The remaining MM bits from MM folks (especially, patch #6 and #8).
> b) The general virtio infrastructure (esp. uapi in patch #2) from virtio
> folks.
>
> I'm planning to send a proper v1 (!RFC) once I have all necessary MM
> acks. In the meanwhile, I will do more testing and minor reworks (e.g.,
> fix !CONFIG_NUMA compilation).


Hi David,

Thanks for your work!

I am trying your https://github.com/davidhildenbrand/linux.git virtio-mem-v5
which works fine for me, but just a 'DMA error' happens when a vm start with
less than 2GB memory, Do I missed sth?

Thanks
Alex


(qemu) qom-set vm0 requested-size 1g
(qemu) [ 26.560026] virtio_mem virtio0: plugged size: 0x0
[ 26.560648] virtio_mem virtio0: requested size: 0x40000000
[ 26.561730] systemd-journald[167]: no db file to read /run/udev/data/+virtio:virtio0: No such file or directory
[ 26.563138] systemd-journald[167]: no db file to read /run/udev/data/+virtio:virtio0: No such file or directory
[ 26.569122] Built 1 zonelists, mobility grouping on. Total pages: 513141
[ 26.570039] Policy zone: Normal

(qemu) [ 32.175838] e1000 0000:00:03.0: swiotlb buffer is full (sz: 81 bytes), total 0 (slots), used 0 (slots)
[ 32.176922] e1000 0000:00:03.0: TX DMA map failed
[ 32.177488] e1000 0000:00:03.0: swiotlb buffer is full (sz: 81 bytes), total 0 (slots), used 0 (slots)
[ 32.178535] e1000 0000:00:03.0: TX DMA map failed

my qemu command is like this:
qemu-system-x86_64 --enable-kvm \
-m 2G,maxmem=16G -kernel /root/linux-next/$1/arch/x86/boot/bzImage \
-smp 4 \
-append "earlyprintk=ttyS0 root=/dev/sda1 console=ttyS0 debug psi=1 nokaslr ignore_loglevel" \
-hda /root/CentOS-7-x86_64-Azure-1703.qcow2 \
-net user,hostfwd=tcp::2222-:22 -net nic -s \
-object memory-backend-ram,id=mem0,size=3G \
-device virtio-mem-pci,id=vm0,memdev=mem0,node=0,requested-size=0M \
--nographic

2020-06-05 09:12:05

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On 05.06.20 10:55, Alex Shi wrote:
>
>
> 在 2020/1/9 下午9:48, David Hildenbrand 写道:
>> Ping,
>>
>> I'd love to get some feedback on
>>
>> a) The remaining MM bits from MM folks (especially, patch #6 and #8).
>> b) The general virtio infrastructure (esp. uapi in patch #2) from virtio
>> folks.
>>
>> I'm planning to send a proper v1 (!RFC) once I have all necessary MM
>> acks. In the meanwhile, I will do more testing and minor reworks (e.g.,
>> fix !CONFIG_NUMA compilation).
>
>
> Hi David,
>
> Thanks for your work!
>
> I am trying your https://github.com/davidhildenbrand/linux.git virtio-mem-v5
> which works fine for me, but just a 'DMA error' happens when a vm start with
> less than 2GB memory, Do I missed sth?

Please use the virtio-mem-v4 branch for now, v5 is still under
construction (and might be scrapped completely if v4 goes upstream as is).

Looks like a DMA issue. Your're hotplugging 1GB, which should not really
eat too much memory. There was a similar issue reported by Hui in [1],
which boiled down to wrong usage of the swiotlb parameter.

In such cases you should always try to reproduce with hotplug of a
sam-sized DIMM. E.g., hotplugging a 1GB DIMM should result in the same
issue.

What does your .config specify for CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE?

I'll try to reproduce with v4 briefly.

[1]
https://lkml.kernel.org/r/[email protected]

>
> Thanks
> Alex
>
>
> (qemu) qom-set vm0 requested-size 1g
> (qemu) [ 26.560026] virtio_mem virtio0: plugged size: 0x0
> [ 26.560648] virtio_mem virtio0: requested size: 0x40000000
> [ 26.561730] systemd-journald[167]: no db file to read /run/udev/data/+virtio:virtio0: No such file or directory
> [ 26.563138] systemd-journald[167]: no db file to read /run/udev/data/+virtio:virtio0: No such file or directory
> [ 26.569122] Built 1 zonelists, mobility grouping on. Total pages: 513141
> [ 26.570039] Policy zone: Normal
>
> (qemu) [ 32.175838] e1000 0000:00:03.0: swiotlb buffer is full (sz: 81 bytes), total 0 (slots), used 0 (slots)
> [ 32.176922] e1000 0000:00:03.0: TX DMA map failed
> [ 32.177488] e1000 0000:00:03.0: swiotlb buffer is full (sz: 81 bytes), total 0 (slots), used 0 (slots)
> [ 32.178535] e1000 0000:00:03.0: TX DMA map failed
>
> my qemu command is like this:
> qemu-system-x86_64 --enable-kvm \
> -m 2G,maxmem=16G -kernel /root/linux-next/$1/arch/x86/boot/bzImage \
> -smp 4 \
> -append "earlyprintk=ttyS0 root=/dev/sda1 console=ttyS0 debug psi=1 nokaslr ignore_loglevel" \
> -hda /root/CentOS-7-x86_64-Azure-1703.qcow2 \
> -net user,hostfwd=tcp::2222-:22 -net nic -s \
> -object memory-backend-ram,id=mem0,size=3G \
> -device virtio-mem-pci,id=vm0,memdev=mem0,node=0,requested-size=0M \
> --nographic
>
>


--
Thanks,

David / dhildenb

2020-06-05 09:38:58

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On 05.06.20 11:08, David Hildenbrand wrote:
> On 05.06.20 10:55, Alex Shi wrote:
>>
>>
>> 在 2020/1/9 下午9:48, David Hildenbrand 写道:
>>> Ping,
>>>
>>> I'd love to get some feedback on
>>>
>>> a) The remaining MM bits from MM folks (especially, patch #6 and #8).
>>> b) The general virtio infrastructure (esp. uapi in patch #2) from virtio
>>> folks.
>>>
>>> I'm planning to send a proper v1 (!RFC) once I have all necessary MM
>>> acks. In the meanwhile, I will do more testing and minor reworks (e.g.,
>>> fix !CONFIG_NUMA compilation).
>>
>>
>> Hi David,
>>
>> Thanks for your work!
>>
>> I am trying your https://github.com/davidhildenbrand/linux.git virtio-mem-v5
>> which works fine for me, but just a 'DMA error' happens when a vm start with
>> less than 2GB memory, Do I missed sth?
>
> Please use the virtio-mem-v4 branch for now, v5 is still under
> construction (and might be scrapped completely if v4 goes upstream as is).
>
> Looks like a DMA issue. Your're hotplugging 1GB, which should not really
> eat too much memory. There was a similar issue reported by Hui in [1],
> which boiled down to wrong usage of the swiotlb parameter.
>
> In such cases you should always try to reproduce with hotplug of a
> sam-sized DIMM. E.g., hotplugging a 1GB DIMM should result in the same
> issue.
>
> What does your .config specify for CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE?
>
> I'll try to reproduce with v4 briefly.

I guess I know what's happening here. In case we only have DMA memory
when booting, we don't reserve swiotlb buffers. Once we hotplug memory
and online ZONE_NORMAL, we don't have any swiotlb DMA bounce buffers to
map such PFNs (total 0 (slots), used 0 (slots)).

Can you try with "swiotlb=force" on the kernel cmdline?

--
Thanks,

David / dhildenb

2020-06-05 10:08:26

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On 05.06.20 11:36, David Hildenbrand wrote:
> On 05.06.20 11:08, David Hildenbrand wrote:
>> On 05.06.20 10:55, Alex Shi wrote:
>>>
>>>
>>> 在 2020/1/9 下午9:48, David Hildenbrand 写道:
>>>> Ping,
>>>>
>>>> I'd love to get some feedback on
>>>>
>>>> a) The remaining MM bits from MM folks (especially, patch #6 and #8).
>>>> b) The general virtio infrastructure (esp. uapi in patch #2) from virtio
>>>> folks.
>>>>
>>>> I'm planning to send a proper v1 (!RFC) once I have all necessary MM
>>>> acks. In the meanwhile, I will do more testing and minor reworks (e.g.,
>>>> fix !CONFIG_NUMA compilation).
>>>
>>>
>>> Hi David,
>>>
>>> Thanks for your work!
>>>
>>> I am trying your https://github.com/davidhildenbrand/linux.git virtio-mem-v5
>>> which works fine for me, but just a 'DMA error' happens when a vm start with
>>> less than 2GB memory, Do I missed sth?
>>
>> Please use the virtio-mem-v4 branch for now, v5 is still under
>> construction (and might be scrapped completely if v4 goes upstream as is).
>>
>> Looks like a DMA issue. Your're hotplugging 1GB, which should not really
>> eat too much memory. There was a similar issue reported by Hui in [1],
>> which boiled down to wrong usage of the swiotlb parameter.
>>
>> In such cases you should always try to reproduce with hotplug of a
>> sam-sized DIMM. E.g., hotplugging a 1GB DIMM should result in the same
>> issue.
>>
>> What does your .config specify for CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE?
>>
>> I'll try to reproduce with v4 briefly.
>
> I guess I know what's happening here. In case we only have DMA memory
> when booting, we don't reserve swiotlb buffers. Once we hotplug memory
> and online ZONE_NORMAL, we don't have any swiotlb DMA bounce buffers to
> map such PFNs (total 0 (slots), used 0 (slots)).
>
> Can you try with "swiotlb=force" on the kernel cmdline?

Alternative, looks like you can specify "-m 2G,maxmem=16G,slots=1", to
create proper ACPI tables that indicate hotpluggable memory. (I'll have
to look into QEMU to figure out to always indicate hotpluggable memory
that way).


--
Thanks,

David / dhildenb

2020-06-05 10:08:33

by Alex Shi

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory



在 2020/6/5 下午5:08, David Hildenbrand 写道:
> Please use the virtio-mem-v4 branch for now, v5 is still under
> construction (and might be scrapped completely if v4 goes upstream as is).
>
> Looks like a DMA issue. Your're hotplugging 1GB, which should not really
> eat too much memory. There was a similar issue reported by Hui in [1],
> which boiled down to wrong usage of the swiotlb parameter.

I have no swiotbl=noforce set, and sometime no swiotlb error reported, like
(qemu) [ 41.591308] e1000 0000:00:03.0: dma_direct_map_page: overflow 0x000000011fd470da+54 of device mask ffffffff
[ 41.592431] e1000 0000:00:03.0: TX DMA map failed
[ 41.593031] e1000 0000:00:03.0: dma_direct_map_page: overflow 0x000000011fd474da+54 of device mask ffffff
...
[ 63.049464] ata_piix 0000:00:01.1: dma_direct_map_sg: overflow 0x0000000107db2000+4096 of device mask ffffffff
[ 63.068297] ata1.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0
[ 63.069057] ata1.00: failed command: READ DMA
[ 63.069580] ata1.00: cmd c8/00:20:40:bd:d2/00:00:00:00:00/e0 tag 0 dma 16384 in
[ 63.069580] res 50/00:00:3f:30:80/00:00:00:00:00/a0 Emask 0x40 (internal error)
>
> In such cases you should always try to reproduce with hotplug of a
> sam-sized DIMM. E.g., hotplugging a 1GB DIMM should result in the same
> issue.
>
> What does your .config specify for CONFIG_MEMORY_HOTPLUG_DEFAULT_ONLINE?

Yes, it's set.

I had tried the v2/v4 version, which has the same issue.
Is this related with virtio-mem start address too low?

Thanks a lot!
>
> I'll try to reproduce with v4 briefly.
>
> [1]
> https://lkml.kernel.org/r/[email protected]

2020-06-05 10:11:13

by Alex Shi

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory



在 2020/6/5 下午5:36, David Hildenbrand 写道:
> I guess I know what's happening here. In case we only have DMA memory
> when booting, we don't reserve swiotlb buffers. Once we hotplug memory
> and online ZONE_NORMAL, we don't have any swiotlb DMA bounce buffers to
> map such PFNs (total 0 (slots), used 0 (slots)).
>
> Can you try with "swiotlb=force" on the kernel cmdline?

Yes, it works fine with this cmdline. problems gone,

2020-06-05 10:49:17

by Alex Shi

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory



在 2020/6/5 下午6:05, David Hildenbrand 写道:
>> I guess I know what's happening here. In case we only have DMA memory
>> when booting, we don't reserve swiotlb buffers. Once we hotplug memory
>> and online ZONE_NORMAL, we don't have any swiotlb DMA bounce buffers to
>> map such PFNs (total 0 (slots), used 0 (slots)).
>>
>> Can you try with "swiotlb=force" on the kernel cmdline?
> Alternative, looks like you can specify "-m 2G,maxmem=16G,slots=1", to
> create proper ACPI tables that indicate hotpluggable memory. (I'll have
> to look into QEMU to figure out to always indicate hotpluggable memory
> that way).
>


That works too. Yes, better resolved in qemu, maybe. :)

Thanks!

2020-06-05 12:21:07

by David Hildenbrand

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory

On 05.06.20 12:46, Alex Shi wrote:
>
>
> 在 2020/6/5 下午6:05, David Hildenbrand 写道:
>>> I guess I know what's happening here. In case we only have DMA memory
>>> when booting, we don't reserve swiotlb buffers. Once we hotplug memory
>>> and online ZONE_NORMAL, we don't have any swiotlb DMA bounce buffers to
>>> map such PFNs (total 0 (slots), used 0 (slots)).
>>>
>>> Can you try with "swiotlb=force" on the kernel cmdline?
>> Alternative, looks like you can specify "-m 2G,maxmem=16G,slots=1", to
>> create proper ACPI tables that indicate hotpluggable memory. (I'll have
>> to look into QEMU to figure out to always indicate hotpluggable memory
>> that way).
>>
>
>
> That works too. Yes, better resolved in qemu, maybe. :)
>

You can checkout

[email protected]:davidhildenbrand/qemu.git virtio-mem-v4

(prone to change before officially sent), which will create srat tables
also if no "slots" parameter was defined (and no -numa config was
specified).

Your original example should work with that.

--
Thanks,

David / dhildenb

2020-06-09 03:08:33

by Alex Shi

[permalink] [raw]
Subject: Re: [PATCH RFC v4 00/13] virtio-mem: paravirtualized memory



在 2020/6/5 下午8:18, David Hildenbrand 写道:
> On 05.06.20 12:46, Alex Shi wrote:
>>
>>
>> 在 2020/6/5 下午6:05, David Hildenbrand 写道:
>>>> I guess I know what's happening here. In case we only have DMA memory
>>>> when booting, we don't reserve swiotlb buffers. Once we hotplug memory
>>>> and online ZONE_NORMAL, we don't have any swiotlb DMA bounce buffers to
>>>> map such PFNs (total 0 (slots), used 0 (slots)).
>>>>
>>>> Can you try with "swiotlb=force" on the kernel cmdline?
>>> Alternative, looks like you can specify "-m 2G,maxmem=16G,slots=1", to
>>> create proper ACPI tables that indicate hotpluggable memory. (I'll have
>>> to look into QEMU to figure out to always indicate hotpluggable memory
>>> that way).
>>>
>>
>>
>> That works too. Yes, better resolved in qemu, maybe. :)
>>
>
> You can checkout
>
> [email protected]:davidhildenbrand/qemu.git virtio-mem-v4

yes, it works for me. Thanks!

>
> (prone to change before officially sent), which will create srat tables
> also if no "slots" parameter was defined (and no -numa config was
> specified).
>
> Your original example should work with that.
>