qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Juan Quintela <quintela@redhat.com>
To: kvm-devel <kvm@vger.kernel.org>, qemu-devel@nongnu.org
Subject: Fortnightly KVM call minutes for 2023-01-24
Date: Tue, 24 Jan 2023 15:56:15 +0100	[thread overview]
Message-ID: <87zga8f0c0.fsf@secure.mitica> (raw)





First part stolen from: https://etherpad.opendev.org/p/qemu-emulation-bof%402023-01-24
      thanks Phillippe

Single QEMU-system binary and Dynamic Machine Models

Meeting link: https://meet.jit.si/kvmcallmeeting
Just want to dial in on your phone? Dial-in: +1.512.647.1431 PIN: 1518845548#
Click this link to see the dial in phone numbers for this meeting https://meet.jit.si/static/dialInInfo.html?room=kvmcallmeeting

What needs to be done?

    TCG

    How to use different page sizes

    -> convert to page-vary (already used by ARM/MIPS)

    ref: https://lore.kernel.org/qemu-devel/20221228171617.059750c3@orange/


    HW models / machine


    How to create/realize 2 QOM objects which depend on each other?

    what (properties) need to be wired? IRQ, reset lines, MR?


    Sysbus: Demote it to plain qdev?

    con:

    sysbus helpful to remove qdev boilerplace/verbose code

    sysbus tree does the resets [currently blocking qdev conversion]

    pro:

    sysbus IRQ API

    too abstract, not very helpful, "named gpios" API is clearer

    sysbus MMIO API

    also kinda abstract, MRs indexed. No qdev equivalent

    sysbus IO API:

    not very used. first we need to get rid of ISA bus singleton


    Single 32/64 *targets* binary

    Which 32-bit hosts are still used? OK to deprecate them?

    Some targets need special care i.e. KVM 32-bit ARM vCPU on 64-bit Aarch64 host


Previous notes:
    https://etherpad.opendev.org/p/qemu-emulation-bof%402022-12-13
    https://etherpad.opendev.org/p/qemu-emulation-bof%402022-11-29


Do we care about this?

64 bits guests on 32 bits hosts: OK to deprecate
32 bits hosts: still in (some) use

Creating and realizing two objects than need to be linked together.

We can't do it with realize, perhaps we need an intermediate state to
do the link, and then realize.

Can we get Peter or Markus or Paolo for the next call?
waiting to get some patches into the list for discussing.

Problem for Phillipe is that he has to do changes to the API's and
want to be sure that they are agreed before he changes all
devices/targets, a multi month task.

For removing sysbus, we need to wait for reset rework from Peter?

Expose the memory API to external processes.

Under what circumstances you should be able to create/destroy a memory
region?



Later, Juan.



             reply	other threads:[~2023-01-24 14:56 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-24 14:56 Juan Quintela [this message]
2023-01-26 21:26 ` Fortnightly KVM call minutes for 2023-01-24 Bernhard Beschow

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=87zga8f0c0.fsf@secure.mitica \
    --to=quintela@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=qemu-devel@nongnu.org \
    /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).