qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Paolo Bonzini <pbonzini@redhat.com>
To: Anthony Liguori <anthony@codemonkey.ws>
Cc: qemu-devel@nongnu.org
Subject: [Qemu-devel] Re: [PATCH 11/19] use a bottom half to run timers
Date: Tue, 05 Jan 2010 13:48:10 +0100	[thread overview]
Message-ID: <4B43350A.7000309@redhat.com> (raw)
In-Reply-To: <4B4280FC.4020602@codemonkey.ws>

On 01/05/2010 12:59 AM, Anthony Liguori wrote:
>
> When I think of a main loop, I think of something that provides the
> following three services
>
> 1) fd based events
> 2) time based events
> 3) an idle callback
>
> The problem we have is that bottom halves aren't quite idle callbacks.
> They have some really weird properties about the guarantees of when
> they're executed.  At any rate, you really cannot express a time based
> event as an idle callback because you need to setup the select() timeout
> based on the next deadline and then dispatch timer event based on
> selects return.  idle functions have none of this ability.

I see bottom halves as a fourth service, providing the ability to 
synchronize stuff that needs to execute in a particular thread 
(guaranteeing thread-safety and especially signal-safety).

In some sense, the problem is that bottom halves conflate this service 
_and_ idle callbacks, and that is the weird property.  Idle callbacks 
are used basically only for DMA, and real-world DMA does not have at all 
the semantics that qemu_bh_schedule_idle provides (which in turn is very 
tricky and not exactly what comments in qemu_bh_update_timeout promise).

Compared to qemu_bh_schedule_idle, bottom halves have sane semantics. 
It would be nicer IMO to remove qemu_bh_schedule_idle, have first-class 
idle callbacks, and leave bottom halves as they are now.  I'll put that 
on the todo list. :-)

That said, I'll try to replace this patch with one that doesn't use a 
bottom half.

Paolo

  reply	other threads:[~2010-01-05 19:28 UTC|newest]

Thread overview: 41+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2009-12-21  8:09 [Qemu-devel] [PATCH 00/19][RFC] Cleanups + split timer handling out of vl.o Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 01/19] centralize handling of -icount Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 02/19] add qemu_icount_round Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 03/19] avoid dubiously clever code in win32_start_timer Paolo Bonzini
2010-01-04 19:34   ` Anthony Liguori
2010-01-04 18:39     ` Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 04/19] fix error in win32_rearm_timer Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 05/19] only one flag is needed for alarm_timer Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 06/19] more alarm timer cleanup Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 07/19] add qemu_get_clock_ns Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 08/19] move kbd/mouse events to event.c Paolo Bonzini
2010-01-04 20:19   ` Anthony Liguori
2009-12-21  8:09 ` [Qemu-devel] [PATCH 09/19] remove qemu_rearm_alarm_timer from main loop Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 10/19] add qemu_bh_scheduled Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 11/19] use a bottom half to run timers Paolo Bonzini
2010-01-04 20:24   ` Anthony Liguori
2010-01-04 19:38     ` Jamie Lokier
2010-01-05  8:38       ` Paolo Bonzini
2010-01-04 20:01     ` [Qemu-devel] " Michael S. Tsirkin
2010-01-04 23:54       ` Anthony Liguori
2010-01-05 12:07         ` Michael S. Tsirkin
2010-01-05 15:23           ` malc
2010-01-05 15:23             ` Michael S. Tsirkin
2010-01-05 15:32               ` malc
2010-01-05 15:33                 ` Michael S. Tsirkin
2010-01-05 15:39                   ` malc
2010-01-04 20:01     ` Paolo Bonzini
2010-01-04 23:59       ` Anthony Liguori
2010-01-05 12:48         ` Paolo Bonzini [this message]
2010-01-05 13:06           ` Anthony Liguori
2010-01-06  1:20             ` Jamie Lokier
2009-12-21  8:09 ` [Qemu-devel] [PATCH 12/19] new function qemu_icount_delta Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 13/19] move tcg_has_work to cpu-exec.c and rename it Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 14/19] disentangle tcg and deadline calculation Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 15/19] do not provide qemu_event_increment if iothread not used Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 16/19] tweak qemu_notify_event Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 17/19] move vmstate registration of vmstate_timers earlier Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 18/19] introduce qemu_clock_enable Paolo Bonzini
2009-12-21  8:09 ` [Qemu-devel] [PATCH 19/19] split out qemu-timer.c Paolo Bonzini
2010-01-04 20:26   ` Anthony Liguori
2010-01-04 19:26 ` [Qemu-devel] [PATCH 00/19][RFC] Cleanups + split timer handling out of vl.o Anthony Liguori

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=4B43350A.7000309@redhat.com \
    --to=pbonzini@redhat.com \
    --cc=anthony@codemonkey.ws \
    --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).