qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Akihiko Odaki <akihiko.odaki@gmail.com>
To: Paolo Bonzini <pbonzini@redhat.com>
Cc: Peter Maydell <peter.maydell@linaro.org>,
	qemu Developers <qemu-devel@nongnu.org>,
	Gerd Hoffmann <kraxel@redhat.com>
Subject: Re: [PATCH v2 0/2] Create menus in iothread
Date: Tue, 8 Mar 2022 00:59:01 +0900	[thread overview]
Message-ID: <CAMVc7JUBMhmLVAD8Zac4tJCoCrjHDqoev9Q1-q2es1aVsMCNGg@mail.gmail.com> (raw)
In-Reply-To: <085b9a6d-5deb-9225-f7d1-81627576020f@redhat.com>

On Tue, Mar 8, 2022 at 12:32 AM Paolo Bonzini <pbonzini@redhat.com> wrote:
>
> On 3/7/22 14:49, Akihiko Odaki wrote:
> > ui/cocoa: Create menus in iothread
> >
> > Commit 0439c5a4623d674efa0c72abd62ca6e98bb7cf87 introduced an
> > assertion that blk_all_next is called in the main thread. The function
> > is called in the following chain:
> > - blk_all_next
> > - qmp_query_block
> > - addRemovableDevicesMenuItems
> > - main
> >
> > This change moves the menu creation to the iothread. This also changes
> > the menu creation procedure to construct the entire menu tree before
> > setting to NSApp, which is necessary because a menu set once cannot be
> > modified if NSApp is already running.
>
> I wonder if you actually need the iothread/secondary thread separation
> during initialization.  It's needed to run the "secondary" main loop,
> but until that point nobody should care what thread things run on.
> cocoa_display_init() is close enough to the end of qemu_init() that I
> think you can just do:
>
>    main()
>      qemu_init()
>        /* takes iothread lock */
>        cocoa_display_init()
>          /* just save a few values from opts */
>      ... create menus ...
>      [NSApp run]
>        applicationDidFinishLaunching:
>          /* do what was in cocoa_display_init() */
>          qemu_unlock_mutex_iothread();
>          qemu_thread_create(call_qemu_main_loop)
>                                                     call_qemu_main_loop()
>                                                       qemu_main_loop()
>
> This might even obsolete the allow_events hack, because now the main
> thread has the iothread lock until applicationDidFinishLaunching:.

allow_events was necessary not because of the separation of the
thread, but because cocoa_display_init waits the main thread for
finishing the initialization. It can be simply removed if it doesn't
wait for app_started_sem.

Regards,
Akihiko Odaki

>
> Paolo
>
> > v2: Separate a change moving create_initial_menus (Peter Maydell)
> >
> > Akihiko Odaki (2):
> >    ui/cocoa: Move create_initial_menus
> >    ui/cocoa: Create menus in iothread
> >
> >   ui/cocoa.m | 209 +++++++++++++++++++++++++----------------------------
> >   1 file changed, 98 insertions(+), 111 deletions(-)
> >
>


  reply	other threads:[~2022-03-07 16:01 UTC|newest]

Thread overview: 6+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-03-07 13:49 [PATCH v2 0/2] Create menus in iothread Akihiko Odaki
2022-03-07 13:49 ` [PATCH v2 1/2] ui/cocoa: Move create_initial_menus Akihiko Odaki
2022-03-07 13:49 ` [PATCH v2 2/2] ui/cocoa: Create menus in iothread Akihiko Odaki
2022-03-07 15:32 ` [PATCH v2 0/2] " Paolo Bonzini
2022-03-07 15:59   ` Akihiko Odaki [this message]
2022-03-20 21:46 ` Philippe Mathieu-Daudé

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=CAMVc7JUBMhmLVAD8Zac4tJCoCrjHDqoev9Q1-q2es1aVsMCNGg@mail.gmail.com \
    --to=akihiko.odaki@gmail.com \
    --cc=kraxel@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=peter.maydell@linaro.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).