From: Sergio Lopez <slp@redhat.com>
To: "Michael S. Tsirkin" <mst@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>,
Eduardo Habkost <ehabkost@redhat.com>,
Maran Wilson <maran.wilson@oracle.com>,
Stefan Hajnoczi <stefanha@gmail.com>,
QEMU Developers <qemu-devel@nongnu.org>,
Gerd Hoffmann <kraxel@redhat.com>,
Paolo Bonzini <pbonzini@redhat.com>,
Stefano Garzarella <sgarzare@redhat.com>,
Richard Henderson <rth@twiddle.net>
Subject: Re: [Qemu-devel] [PATCH v3 0/4] Introduce the microvm machine type
Date: Thu, 25 Jul 2019 17:49:32 +0200 [thread overview]
Message-ID: <87k1c6m7w3.fsf@redhat.com> (raw)
In-Reply-To: <20190725110114-mutt-send-email-mst@kernel.org>
[-- Attachment #1: Type: text/plain, Size: 5230 bytes --]
Michael S. Tsirkin <mst@redhat.com> writes:
> On Thu, Jul 25, 2019 at 10:58:22AM -0400, Michael S. Tsirkin wrote:
>> On Thu, Jul 25, 2019 at 04:42:42PM +0200, Sergio Lopez wrote:
>> >
>> > Paolo Bonzini <pbonzini@redhat.com> writes:
>> >
>> > > On 25/07/19 15:26, Stefan Hajnoczi wrote:
>> > >> The microvm design has a premise and it can be answered definitively
>> > >> through performance analysis.
>> > >>
>> > >> If I had to explain to someone why PCI or ACPI significantly slows
>> > >> things down, I couldn't honestly do so. I say significantly because
>> > >> PCI init definitely requires more vmexits but can it be a small
>> > >> number? For ACPI I have no idea why it would consume significant
>> > >> amounts of time.
>> > >
>> > > My guess is that it's just a lot of code that has to run. :(
>> >
>> > I think I haven't shared any numbers about ACPI.
>> >
>> > I don't have details about where exactly the time is spent, but
>> > compiling a guest kernel without ACPI decreases the average boot time in
>> > ~12ms, and the kernel's unstripped ELF binary size goes down in a
>> > whooping ~300KiB.
>>
>> At least the binary size is hardly surprising.
>>
>> I'm guessing you built in lots of drivers.
>>
>> It would be educational to try to enable ACPI core but disable all
>> optional features.
I just tried disabling everything that menuconfig allowed me to. Saves
~27KiB and doesn't improve boot time.
> Trying with ACPI_REDUCED_HARDWARE_ONLY would also be educational.
I also tried enabling this one in my original config. It saves ~11.5KiB,
and has on impact on boot time either.
>>
>> > On the other hand, removing ACPI from QEMU decreases its initialization
>> > time in ~5ms, and the binary size is ~183KiB smaller.
>>
>> Yes - ACPI generation uses a ton of allocations and data copies.
>>
>> Need to play with pre-allocation strategies. Maybe something
>> as simple as:
>>
>> diff --git a/hw/i386/acpi-build.c b/hw/i386/acpi-build.c
>> index f3fdfefcd5..24becc069e 100644
>> --- a/hw/i386/acpi-build.c
>> +++ b/hw/i386/acpi-build.c
>> @@ -2629,8 +2629,10 @@ void acpi_build(AcpiBuildTables *tables, MachineState *machine)
>> acpi_get_pci_holes(&pci_hole, &pci_hole64);
>> acpi_get_slic_oem(&slic_oem);
>>
>> +#define DEFAULT_ARRAY_SIZE 16
>> table_offsets = g_array_new(false, true /* clear */,
>> - sizeof(uint32_t));
>> + sizeof(uint32_t),
>> + DEFAULT_ARRAY_SIZE);
>> ACPI_BUILD_DPRINTF("init ACPI tables\n");
>>
>> bios_linker_loader_alloc(tables->linker,
>>
>> will already help a bit.
>>
>> >
>> > IMHO, those are pretty relevant savings on both fronts.
>> >
>> > >> Until we have this knowledge, the premise of microvm is unproven and
>> > >> merging it would be premature because maybe we can get into the same
>> > >> ballpark by optimizing existing code.
>> > >>
>> > >> I'm sorry for being a pain. I actually think the analysis will
>> > >> support microvm, but it still needs to be done in order to justify it.
>> > >
>> > > No, you're not a pain, you're explaining your reasoning and that helps.
>> > >
>> > > To me *maintainability is the biggest consideration* when introducing a
>> > > new feature. "We can do just as well with q35" is a good reason to
>> > > deprecate and delete microvm, but not a good reason to reject it now as
>> > > long as microvm is good enough in terms of maintainability. Keeping it
>> > > out of tree only makes it harder to do this kind of experiment. virtio
>> > > 1 seems to be the biggest remaining blocker and I think it'd be a good
>> > > thing to have even for the ARM virt machine type.
>> > >
>> > > FWIW the "PCI tax" seems to be ~10 ms in QEMU, ~10 ms in the firmware(*)
>> > > and ~25 ms in the kernel. I must say that's pretty good, but it's still
>> > > 30% of the whole boot time and reducing it is the hardest part. If
>> > > having microvm in tree can help reducing it, good. Yes, it will get
>> > > users, but most likely they will have to support pc or q35 as a fallback
>> > > so we could still delete microvm at any time with the due deprecation
>> > > period if it turns out to be a failed experiment.
>> > >
>> > > Whether to use qboot or SeaBIOS for microvm is another story, but it's
>> > > an implementation detail as long as the ROM size doesn't change and/or
>> > > we don't do versioned machine types. So we can switch from one to the
>> > > other at any time; we can also include qboot directly in QEMU's tree,
>> > > without going through a submodule, which also reduces the infrastructure
>> > > needed (mirrors, etc.) and makes it easier to delete it.
>> > >
>> > > Paolo
>> > >
>> > > (*) I measured 15ms in SeaBIOS and 5ms in qboot from the first to the
>> > > last write to 0xcf8. I suspect part of qboot's 10ms boot time actually
>> > > end up measured as PCI in SeaBIOS, due to different init order, so the
>> > > real firmware cost of PAM and PCI initialization should be 5ms for qboot
>> > > and 10ms for SeaBIOS.
>> >
>>
>>
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 832 bytes --]
next prev parent reply other threads:[~2019-07-25 15:49 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-02 12:11 [Qemu-devel] [PATCH v3 0/4] Introduce the microvm machine type Sergio Lopez
2019-07-02 12:11 ` [Qemu-devel] [PATCH v3 1/4] hw/virtio: Factorize virtio-mmio headers Sergio Lopez
2019-07-25 9:46 ` Liam Merwick
2019-07-25 9:58 ` Michael S. Tsirkin
2019-07-25 10:03 ` Peter Maydell
2019-07-25 10:36 ` Paolo Bonzini
2019-07-02 12:11 ` [Qemu-devel] [PATCH v3 2/4] hw/i386: Add an Intel MPTable generator Sergio Lopez
2019-07-02 12:11 ` [Qemu-devel] [PATCH v3 3/4] hw/i386: Factorize PVH related functions Sergio Lopez
2019-07-23 8:39 ` Liam Merwick
2019-07-02 12:11 ` [Qemu-devel] [PATCH v3 4/4] hw/i386: Introduce the microvm machine type Sergio Lopez
2019-07-02 13:58 ` Gerd Hoffmann
2019-07-25 10:47 ` Paolo Bonzini
2019-07-02 15:01 ` [Qemu-devel] [PATCH v3 0/4] " no-reply
2019-07-02 15:23 ` Peter Maydell
2019-07-02 17:34 ` Sergio Lopez
2019-07-02 18:04 ` Peter Maydell
2019-07-02 22:04 ` Sergio Lopez
2019-07-25 9:59 ` Michael S. Tsirkin
2019-07-25 10:05 ` Peter Maydell
2019-07-25 10:10 ` Michael S. Tsirkin
2019-07-25 14:52 ` Sergio Lopez
2019-07-25 10:42 ` Sergio Lopez
2019-07-25 11:23 ` Paolo Bonzini
2019-07-25 12:01 ` Stefan Hajnoczi
2019-07-25 12:10 ` Michael S. Tsirkin
2019-07-25 13:26 ` Stefan Hajnoczi
2019-07-25 13:43 ` Paolo Bonzini
2019-07-25 13:54 ` Michael S. Tsirkin
2019-07-25 14:13 ` Paolo Bonzini
2019-07-25 14:42 ` Michael S. Tsirkin
2019-07-25 14:04 ` Peter Maydell
2019-07-25 14:26 ` Paolo Bonzini
2019-07-25 14:35 ` Michael S. Tsirkin
2019-07-25 14:42 ` Sergio Lopez
2019-07-25 14:58 ` Michael S. Tsirkin
2019-07-25 15:01 ` Michael S. Tsirkin
2019-07-25 15:39 ` Paolo Bonzini
2019-07-25 17:38 ` Michael S. Tsirkin
2019-07-26 12:46 ` Igor Mammedov
2019-07-25 15:49 ` Sergio Lopez [this message]
2019-07-25 13:48 ` Michael S. Tsirkin
2019-07-02 15:30 ` no-reply
2019-07-03 9:58 ` Stefan Hajnoczi
2019-07-18 15:21 ` Sergio Lopez
2019-07-19 10:29 ` Stefan Hajnoczi
2019-07-19 13:48 ` Sergio Lopez
2019-07-19 15:09 ` Stefan Hajnoczi
2019-07-19 15:42 ` Montes, Julio
2019-07-23 8:43 ` Sergio Lopez
2019-07-23 9:47 ` Stefan Hajnoczi
2019-07-23 10:01 ` Paolo Bonzini
2019-07-24 11:14 ` Paolo Bonzini
2019-07-25 9:35 ` Sergio Lopez
2019-07-25 10:03 ` Michael S. Tsirkin
2019-07-25 10:55 ` Paolo Bonzini
2019-07-25 14:46 ` Michael S. Tsirkin
2019-07-25 15:35 ` Paolo Bonzini
2019-07-25 17:33 ` Michael S. Tsirkin
2019-07-25 20:30 ` Michael S. Tsirkin
2019-07-26 7:57 ` Paolo Bonzini
2019-07-26 11:10 ` Michael S. Tsirkin
2019-07-23 11:30 ` Stefano Garzarella
2019-07-24 15:23 ` Stefano Garzarella
2019-08-29 9:02 ` Jing Liu
2019-08-29 15:46 ` Sergio Lopez
2019-08-30 4:53 ` Jing Liu
2019-08-30 14:27 ` Sergio Lopez
2019-09-02 5:43 ` Jing Liu
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=87k1c6m7w3.fsf@redhat.com \
--to=slp@redhat.com \
--cc=ehabkost@redhat.com \
--cc=kraxel@redhat.com \
--cc=maran.wilson@oracle.com \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=rth@twiddle.net \
--cc=sgarzare@redhat.com \
--cc=stefanha@gmail.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).