From: Glauber de Oliveira Costa <gcosta@redhat.com>
To: Anthony Liguori <aliguori@us.ibm.com>
Cc: jan kiszka <jan.kiszka@siemens.com>,
Glauber Costa <glommer@redhat.com>,
jes@sgi.com, qemu-devel@nongnu.org, avi@qumranet.com,
dmitry baryshkov <dmitry.baryshkov@siemens.com>
Subject: [Qemu-devel] Re: [PATCH 20/32] accel_trace_io
Date: Thu, 23 Oct 2008 13:26:32 -0400 (EDT) [thread overview]
Message-ID: <391190915.2897471224782791753.JavaMail.root@zmail01.collab.prod.int.phx2.redhat.com> (raw)
In-Reply-To: <4900882A.30909@us.ibm.com>
----- "Anthony Liguori" <aliguori@us.ibm.com> wrote:
> Glauber Costa wrote:
> > From: Glauber Costa <gcosta@redhat.com>
> >
> > kqemu keeps trace of the last io done. Do it through
> > an accel_wrapper.
> >
>
> The hooks here are probably not the best they could be. The purpose
> of
> this code in kqemu, from what I can tell, is to control how long we
> run
> in softmmu before trying to run in the kernel again. I presume this
> is
> mainly for kernel-kqemu.
>
> When kernel-kqemu encounters an instruction it can't handle, it drops
> to
> userspace and runs in softmmu for a while. The question is how long
> should it run in softmmu and this is a heuristic. This is what the
> code
> is doing.
>
> Regards,
My proposal is:
* add a hook for in and another for out operations. It's generic enough to be useful
for other purposes in the future. KVM does not need to do anything in here, and kqemu can
just record the last io time
* add a hook at th end of cpu_exec, that allows the accel to shut off current execution based
on any criteria it wants to. accel_break_loop (the current one) seems a fine name for it.
To be even more explicit, the accel can just return 1 or 0 indicating whether or not the current
execution loop should be stopped, and cpu_loop_exit() is called based on the return value.
What do you think?
next prev parent reply other threads:[~2008-10-23 17:26 UTC|newest]
Thread overview: 80+ messages / expand[flat|nested] mbox.gz Atom feed top
2008-10-23 14:18 [Qemu-devel] [PATCH 0/32] New shot at accelerators Glauber Costa
2008-10-23 13:35 ` [Qemu-devel] " Jan Kiszka
2008-10-23 14:07 ` Glauber Costa
2008-10-23 14:15 ` Avi Kivity
2008-10-23 13:44 ` Anthony Liguori
2008-10-23 14:18 ` [Qemu-devel] [PATCH 01/32] use anonymous memory for kqemu Glauber Costa
2008-10-23 13:35 ` [Qemu-devel] " Jan Kiszka
2008-10-23 13:48 ` Anthony Liguori
2008-10-23 14:17 ` Jan Kiszka
2008-10-23 14:25 ` Anthony Liguori
2008-10-23 15:08 ` Leonardo Reiter
2008-10-23 15:20 ` Leonardo Reiter
2008-10-24 19:30 ` Andreas Färber
2008-10-24 19:59 ` Ben Taylor
2008-10-25 10:17 ` [Qemu-devel] QEMU on Solaris 10 (was: [PATCH 01/32] use anonymous memory for kqemu.) Andreas Färber
2008-10-25 10:27 ` Andreas Färber
2008-10-25 10:45 ` Blue Swirl
2008-10-24 15:37 ` [Qemu-devel] Re: [PATCH 01/32] use anonymous memory for kqemu Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 02/32] protect exec-all.h frm multiple inclusion Glauber Costa
2008-10-23 13:52 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:18 ` [Qemu-devel] [PATCH 03/32] change definition of FILE for linux Glauber Costa
2008-10-23 13:52 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:13 ` Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 04/32] move kqemu_cpu_exec to kqemu.c Glauber Costa
2008-10-23 13:55 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:21 ` Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 05/32] use more meaningful values for kqemu_cpu_exec Glauber Costa
2008-10-23 13:57 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:23 ` Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 06/32] split kqemu_init into two Glauber Costa
2008-10-23 13:58 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:28 ` Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 07/32] introduce QEMUAccel and fill it with interrupt specific driver Glauber Costa
2008-10-23 14:00 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:18 ` [Qemu-devel] [PATCH 08/32] init env made accel driver Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 09/32] wrap cache flushing functions into accel drivers Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 10/32] turn info kqemu into generic info accelerator Glauber Costa
2008-10-23 14:03 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:24 ` Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 11/32] separate accelerator part of info profiler Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 12/32] move kqemu externs to kqemu.h Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 13/32] move disabling code to kqemu.c instead of vl.c Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 14/32] set_notdirty goes through accel wrapper Glauber Costa
2008-10-23 14:18 ` [Qemu-devel] [PATCH 15/32] wrap modify_page through accel calls Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 16/32] remove kqemu reference from hw/pc.c Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 17/32] build list of available accelerators Glauber Costa
2008-10-23 13:45 ` [Qemu-devel] " Avi Kivity
2008-10-23 15:09 ` Glauber Costa
2008-10-23 15:15 ` Avi Kivity
2008-10-23 14:19 ` [Qemu-devel] [PATCH 18/32] provide --accel option Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 19/32] add hook to cpu_register_physical_memory Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 20/32] accel_trace_io Glauber Costa
2008-10-23 14:20 ` [Qemu-devel] " Anthony Liguori
2008-10-23 17:26 ` Glauber de Oliveira Costa [this message]
2008-10-25 11:10 ` [Qemu-devel] " andrzej zaborowski
2008-10-25 11:14 ` Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 21/32] get_env accel wrapper Glauber Costa
2008-10-23 13:36 ` [Qemu-devel] " Avi Kivity
2008-10-23 14:19 ` [Qemu-devel] [PATCH 22/32] add next_cpu_index Glauber Costa
2008-10-23 14:21 ` [Qemu-devel] " Anthony Liguori
2008-10-23 14:37 ` Glauber Costa
2008-10-23 14:40 ` Jan Kiszka
2008-10-23 14:55 ` Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 23/32] move cpu_get_time_fast to kqemu.c Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 24/32] check wether kqemu is enabled in open code Glauber Costa
2008-10-23 13:38 ` [Qemu-devel] " Jan Kiszka
2008-10-23 14:49 ` Glauber Costa
2008-10-23 14:23 ` Anthony Liguori
2008-10-23 14:31 ` Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 25/32] provide accel hook for cpu_exec Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 26/32] provide two accelerators for kqemu Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 27/32] arch-specific hooks for accelerator Glauber Costa
2008-10-23 13:30 ` [Qemu-devel] " Avi Kivity
2008-10-23 13:35 ` Jan Kiszka
2008-10-23 13:47 ` Avi Kivity
2008-10-23 14:19 ` [Qemu-devel] [PATCH 28/32] iret arch specific accelerator Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 29/32] sysret/sysexit " Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 30/32] lcall/lret arch specific accel hooks Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 31/32] remove kqemu_is_ok tests Glauber Costa
2008-10-23 14:19 ` [Qemu-devel] [PATCH 32/32] clean up kqemu code Glauber Costa
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=391190915.2897471224782791753.JavaMail.root@zmail01.collab.prod.int.phx2.redhat.com \
--to=gcosta@redhat.com \
--cc=aliguori@us.ibm.com \
--cc=avi@qumranet.com \
--cc=dmitry.baryshkov@siemens.com \
--cc=glommer@redhat.com \
--cc=jan.kiszka@siemens.com \
--cc=jes@sgi.com \
--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).