qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Eduardo Habkost <ehabkost@redhat.com>
To: "Philippe Mathieu-Daudé" <f4bug@amsat.org>
Cc: Paolo Bonzini <pbonzini@redhat.com>,
	Thomas Huth <thuth@redhat.com>, Peter Xu <peterx@redhat.com>,
	Markus Armbruster <armbru@redhat.com>,
	Stefano Stabellini <sstabellini@kernel.org>,
	Anthony Perard <anthony.perard@citrix.com>,
	qemu-devel@nongnu.org, qemu-trivial@nongnu.org,
	"Emilio G . Cota" <cota@braap.org>,
	Eric Blake <eblake@redhat.com>,
	Richard Henderson <rth@twiddle.net>
Subject: Re: [Qemu-devel] [RFC PATCH] vl: only display available accelerators
Date: Wed, 8 Nov 2017 14:56:19 -0200	[thread overview]
Message-ID: <20171108165619.GM3111@localhost.localdomain> (raw)
In-Reply-To: <9adefce7-1c65-9539-1961-fa981f0ecccc@amsat.org>

On Wed, Nov 08, 2017 at 01:20:49PM -0300, Philippe Mathieu-Daudé wrote:
> On 11/02/2017 09:35 PM, Eduardo Habkost wrote:
> > On Mon, Oct 30, 2017 at 03:14:59PM -0300, Philippe Mathieu-Daudé wrote:
> >> examples configuring with '--enable-kvm --disable-tcg'
> >>
> >> - before
> >>
> >>   $ qemu-system-x86_64 -accel help
> >>   Possible accelerators: kvm, xen, hax, tcg
> >>
> >>   $ qemu-system-x86_64 -accel tcg
> >>   qemu-system-x86_64: -machine accel=tcg: No accelerator found
> >>
> >>   # qemu-system-x86_64 -accel hax
> >>   qemu-system-x86_64: -machine accel=hax: No accelerator found
> >>
> >> - after
> >>
> >>   $ qemu-system-x86_64 -accel help
> >>   Possible accelerators:
> >>     xen
> >>     kvm
> >>
> >> Suggested-by: Eduardo Habkost <ehabkost@redhat.com>
> >> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
> >> ---
> >> since RFC:
> >>   - use much cleaner object_class_get_list(TYPE_ACCEL, false)
> >>
> >>  vl.c | 39 ++++++++++++++++++++++++++++++++++-----
> >>  1 file changed, 34 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/vl.c b/vl.c
> >> index ec299099ff..0f13641715 100644
> >> --- a/vl.c
> >> +++ b/vl.c
> >> @@ -2764,6 +2764,39 @@ static gint machine_class_cmp(gconstpointer a, gconstpointer b)
> >>      exit(!name || !is_help_option(name));
> >>  }
> >>  
> >> +static void accel_list_entry(gpointer data, gpointer user_data)
> >> +{
> >> +    ObjectClass *oc = data;
> >> +    const char *typename = object_class_get_name(oc);
> >> +    int len;
> >> +
> >> +    if (!qtest_driver() && !g_strcmp0(typename, ACCEL_CLASS_NAME("qtest"))) {
> >> +        return; /* used by test cases */
> >> +    }
> > 
> > I would prefer to not hardcode the accel name here.  If you want
> > qtest to have special behavior, you can add a new field to
> > AccelClass to implement that.
> > 
> > However, I don't think I agree we should hide qtest from the
> > user: it is accepted on the command-line, so why should we lie?
> 
> Ok.
> 
> > Why not add a description field and indicate that qtest is useful
> > only for testing?
> 
> This is cleaner indeed.
> 
> >> +
> >> +    len = strlen(typename) - strlen("-" TYPE_ACCEL);
> > 
> > You can use AccelClass::name here.
> 
> Hmm yes but currently this is not what the user should use, since the
> accel_find() matches the ACCEL_CLASS_NAME:
> 
> $ ppc64-softmmu/qemu-system-ppc64 -accel QTest
> qemu-system-ppc64: -machine accel=QTest: No accelerator found

Oops, you're correct.

In this case, I suggest using ACCEL_CLASS_SUFFIX instead of
hardcoding ("-" TYPE_ACCEL).

Maybe a accel_class_get_accel_name() helper similar to
x86_cpu_class_get_model_name() would be even better.  It would
require an extra memory allocation, but the code would be
cleaner.  We might add a QMP command returning the accel list in
the future, so a common helper would be useful.


> 
> I'll respin but the change will be bigger so I is probably too late for
> 2.11.
> 
> > 
> >> +    if (len > 0) {
> >> +        error_printf("  %.*s\n", len, typename);
> >> +    }
> >> +}
> >> +
> >> +static void accel_parse(const char *name, QemuOpts *accel_opts)
> >> +{
> >> +    const char *optarg = qemu_opt_get(accel_opts, "accel");
> >> +    GSList *list;
> >> +
> >> +    if (!is_help_option(optarg)) {
> >> +        return;
> >> +    }
> >> +
> >> +    list = object_class_get_list(TYPE_ACCEL, false);
> >> +    error_printf("Possible accelerators:\n");
> >> +    g_slist_foreach(list, accel_list_entry, NULL);
> >> +    g_slist_free(list);
> >> +
> >> +    exit(0);
> >> +}
> >> +
> >>  void qemu_add_exit_notifier(Notifier *notify)
> >>  {
> >>      notifier_list_add(&exit_notifiers, notify);
> >> @@ -3881,11 +3914,7 @@ int main(int argc, char **argv, char **envp)
> >>              case QEMU_OPTION_accel:
> >>                  accel_opts = qemu_opts_parse_noisily(qemu_find_opts("accel"),
> >>                                                       optarg, true);
> >> -                optarg = qemu_opt_get(accel_opts, "accel");
> >> -                if (!optarg || is_help_option(optarg)) {
> >> -                    error_printf("Possible accelerators: kvm, xen, hax, tcg\n");
> >> -                    exit(0);
> >> -                }
> >> +                accel_parse(optarg, accel_opts);
> >>                  opts = qemu_opts_create(qemu_find_opts("machine"), NULL,
> >>                                          false, &error_abort);
> >>                  qemu_opt_set(opts, "accel", optarg, &error_abort);
> >> -- 
> >> 2.15.0.rc2
> >>
> > 

-- 
Eduardo

  reply	other threads:[~2017-11-08 16:56 UTC|newest]

Thread overview: 19+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2017-10-30 18:14 [Qemu-devel] [RFC PATCH] vl: only display available accelerators Philippe Mathieu-Daudé
2017-10-30 18:18 ` Philippe Mathieu-Daudé
2017-11-03  0:35 ` Eduardo Habkost
2017-11-08 16:20   ` Philippe Mathieu-Daudé
2017-11-08 16:56     ` Eduardo Habkost [this message]
  -- strict thread matches above, loose matches on Subject: below --
2017-10-30  4:00 Philippe Mathieu-Daudé
2017-10-30  4:13 ` no-reply
2017-10-30  8:20 ` Eduardo Habkost
2017-10-30 15:01   ` Philippe Mathieu-Daudé
2017-10-30 15:34     ` Peter Maydell
2017-11-08 13:28   ` Daniel P. Berrange
2017-11-08 16:21     ` Philippe Mathieu-Daudé
2017-11-08 16:59       ` Eduardo Habkost
2017-11-08 17:06         ` Daniel P. Berrange
2017-11-08 17:25           ` Eduardo Habkost
2017-11-08 17:28             ` Daniel P. Berrange
2017-11-08 18:03             ` Philippe Mathieu-Daudé
2017-11-08 18:35               ` Eduardo Habkost
2017-11-08 19:41                 ` 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=20171108165619.GM3111@localhost.localdomain \
    --to=ehabkost@redhat.com \
    --cc=anthony.perard@citrix.com \
    --cc=armbru@redhat.com \
    --cc=cota@braap.org \
    --cc=eblake@redhat.com \
    --cc=f4bug@amsat.org \
    --cc=pbonzini@redhat.com \
    --cc=peterx@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=qemu-trivial@nongnu.org \
    --cc=rth@twiddle.net \
    --cc=sstabellini@kernel.org \
    --cc=thuth@redhat.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).