From: David Hildenbrand <david@redhat.com>
To: "Michael S. Tsirkin" <mst@redhat.com>, Yangming <yangming73@huawei.com>
Cc: "qemu-devel@nongnu.org" <qemu-devel@nongnu.org>,
"wangzhigang (O)" <wangzhigang17@huawei.com>,
"zhangliang (AG)" <zhangliang5@huawei.com>,
xiqi <xiqi2@huawei.com>
Subject: Re: [PATCH] virtio-balloon: optimize the virtio-balloon on the ARM platform.
Date: Tue, 28 Feb 2023 09:58:52 +0100 [thread overview]
Message-ID: <0bba1105-e576-5037-2158-79c00440606b@redhat.com> (raw)
In-Reply-To: <20230228035341-mutt-send-email-mst@kernel.org>
On 28.02.23 09:56, Michael S. Tsirkin wrote:
> On Tue, Feb 28, 2023 at 03:26:56AM +0000, Yangming wrote:
>>>> Optimize the virtio-balloon feature on the ARM platform by adding a
>>>> variable to keep track of the current hot-plugged pc-dimm size,
>>>> instead of traversing the virtual machine's memory modules to count
>>>> the current RAM size during the balloon inflation or deflation
>>>> process. This variable can be updated only when plugging or unplugging
>>>> the device, which will result in an increase of approximately 60%
>>>> efficiency of balloon process on the ARM platform.
>>>>
>>>> We tested the total amount of time required for the balloon inflation
>>> process on ARM:
>>>> inflate the balloon to 64GB of a 128GB guest under stress.
>>>> Before: 102 seconds
>>>> After: 42 seconds
>>>>
>>>> Signed-off-by: Qi Xi <xiqi2@huawei.com>
>>>> Signed-off-by: Ming Yang yangming73@huawei.com
>>>> ---
>>>> hw/mem/pc-dimm.c | 2 ++
>>>> hw/virtio/virtio-balloon.c | 33 +++++----------------------------
>>>> include/hw/boards.h | 1 +
>>>> 3 files changed, 8 insertions(+), 28 deletions(-)
>>>>
>>>> diff --git a/hw/mem/pc-dimm.c b/hw/mem/pc-dimm.c index
>>>> 50ef83215c..192fc7922c 100644
>>>> --- a/hw/mem/pc-dimm.c
>>>> +++ b/hw/mem/pc-dimm.c
>>>> @@ -81,6 +81,7 @@ void pc_dimm_plug(PCDIMMDevice *dimm,
>>> MachineState
>>>> *machine)
>>>>
>>>> memory_device_plug(MEMORY_DEVICE(dimm), machine);
>>>> vmstate_register_ram(vmstate_mr, DEVICE(dimm));
>>>> + machine->device_memory->dimm_size += vmstate_mr->size;
>>>> }
>>>>
>>>> void pc_dimm_unplug(PCDIMMDevice *dimm, MachineState *machine)
>>> @@
>>>> -90,6 +91,7 @@ void pc_dimm_unplug(PCDIMMDevice *dimm,
>>> MachineState
>>>> *machine)
>>>>
>>>> memory_device_unplug(MEMORY_DEVICE(dimm), machine);
>>>> vmstate_unregister_ram(vmstate_mr, DEVICE(dimm));
>>>> + machine->device_memory->dimm_size -= vmstate_mr->size;
>>>> }
>>>
>>> Ahh, missed that my previous comment was not addressed: we only want to
>>> track "real" DIMMs, not NVDIMMs.
>>>
>>> --
>>> Thanks,
>>>
>>> David / dhildenb
>>
>> Optimize the virtio-balloon feature on the ARM platform by adding
>> a variable to keep track of the current hot-plugged pc-dimm size,
>> instead of traversing the virtual machine's memory modules to count
>> the current RAM size during the balloon inflation or deflation
>> process. This variable can be updated only when plugging or unplugging
>> the device, which will result in an increase of approximately 60%
>> efficiency of balloon process on the ARM platform.
>>
>> We tested the total amount of time required for the balloon inflation process on ARM:
>> inflate the balloon to 64GB of a 128GB guest under stress.
>> Before: 102 seconds
>> After: 42 seconds
>>
>> Signed-off-by: Qi Xi <xiqi2@huawei.com>
>> Signed-off-by: Ming Yang yangming73@huawei.com
>> ---
>> hw/mem/pc-dimm.c | 8 ++++++++
>> hw/virtio/virtio-balloon.c | 33 +++++----------------------------
>> include/hw/boards.h | 1 +
>> 3 files changed, 14 insertions(+), 28 deletions(-)
>>
>> diff --git a/hw/mem/pc-dimm.c b/hw/mem/pc-dimm.c
>> index 50ef83215c..2107615016 100644
>> --- a/hw/mem/pc-dimm.c
>> +++ b/hw/mem/pc-dimm.c
>> @@ -81,6 +81,10 @@ void pc_dimm_plug(PCDIMMDevice *dimm, MachineState *machine)
>>
>> memory_device_plug(MEMORY_DEVICE(dimm), machine);
>> vmstate_register_ram(vmstate_mr, DEVICE(dimm));
>> + bool is_nvdimm = object_dynamic_cast(OBJECT(dimm), TYPE_NVDIMM);
>> + if (!is_nvdimm) {
>> + machine->device_memory->dimm_size += vmstate_mr->size;
>> + }
>> }
>>
>> void pc_dimm_unplug(PCDIMMDevice *dimm, MachineState *machine)
>> @@ -90,6 +94,10 @@ void pc_dimm_unplug(PCDIMMDevice *dimm, MachineState *machine)
>>
>> memory_device_unplug(MEMORY_DEVICE(dimm), machine);
>> vmstate_unregister_ram(vmstate_mr, DEVICE(dimm));
>> + bool is_nvdimm = object_dynamic_cast(OBJECT(dimm), TYPE_NVDIMM);
>> + if (!is_nvdimm) {
>> + machine->device_memory->dimm_size -= vmstate_mr->size;
>> + }
>> }
>>
>
> add comments here explaining why are nvdimms excluded?
>
I really prefer to avoid mixing declaration and initialization where it
an be avoided.
Further, the local variable can be easily avoided completely.
if (!object_dynamic_cast(OBJECT(dimm), TYPE_NVDIMM))
With that and with the other comments addressed
Acked-by: David Hildenbrand <david@redhat.com>
--
Thanks,
David / dhildenb
next prev parent reply other threads:[~2023-02-28 8:59 UTC|newest]
Thread overview: 5+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-02-28 3:26 [PATCH] virtio-balloon: optimize the virtio-balloon on the ARM platform Yangming via
2023-02-28 8:56 ` Michael S. Tsirkin
2023-02-28 8:58 ` David Hildenbrand [this message]
[not found] <20230224074624.1531-1-xiqi2@huawei.com>
2023-02-24 8:23 ` Yangming via
2023-02-24 8:28 ` Michael S. Tsirkin
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=0bba1105-e576-5037-2158-79c00440606b@redhat.com \
--to=david@redhat.com \
--cc=mst@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=wangzhigang17@huawei.com \
--cc=xiqi2@huawei.com \
--cc=yangming73@huawei.com \
--cc=zhangliang5@huawei.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).