From: George Dunlap <George.Dunlap@eu.citrix.com>
To: Juergen Gross <jgross@suse.com>
Cc: Ian Jackson <ian.jackson@eu.citrix.com>,
Stefano Stabellini <stefano.stabellini@eu.citrix.com>,
Wei Liu <wei.liu2@citrix.com>, Chunyan Liu <cyliu@suse.com>,
"xen-devel@lists.xen.org" <xen-devel@lists.xen.org>
Subject: Re: [PATCH 1/3] libxl: make libxl__need_xenpv_qemu() operate on domain config
Date: Thu, 17 Mar 2016 16:53:36 +0000 [thread overview]
Message-ID: <CAFLBxZbua-BUcqjbhPcCjQRMqoAaJUGwM-9XOOZmAzV7h-sUxA@mail.gmail.com> (raw)
In-Reply-To: <1457622051-30189-2-git-send-email-jgross@suse.com>
On Thu, Mar 10, 2016 at 3:00 PM, Juergen Gross <jgross@suse.com> wrote:
> libxl__need_xenpv_qemu() is called with configuration data for console,
> vfbs, disks and channels today in order to evaluate the need for
> starting a device model for a pv domain.
>
> The console data is local to the caller and setup in a way to never
> require a device model. All other data is taken from the domain config
> structure.
>
> In order to support other device backends via qemu change the interface
> of libxl__need_xenpv_qemu() to take the domain config structure as
> input instead of the single device arrays.
>
> Signed-off-by: Juergen Gross <jgross@suse.com>
> ---
> tools/libxl/libxl_create.c | 9 +------
> tools/libxl/libxl_dm.c | 59 ++++++++++++--------------------------------
> tools/libxl/libxl_internal.h | 5 +---
> 3 files changed, 18 insertions(+), 55 deletions(-)
>
> diff --git a/tools/libxl/libxl_create.c b/tools/libxl/libxl_create.c
> index 61b5c01..0e2b0a0 100644
> --- a/tools/libxl/libxl_create.c
> +++ b/tools/libxl/libxl_create.c
> @@ -1304,7 +1304,6 @@ static void domcreate_launch_dm(libxl__egc *egc, libxl__multidev *multidev,
> }
> case LIBXL_DOMAIN_TYPE_PV:
> {
> - int need_qemu = 0;
> libxl__device_console console;
> libxl__device device;
>
> @@ -1314,17 +1313,11 @@ static void domcreate_launch_dm(libxl__egc *egc, libxl__multidev *multidev,
> }
>
> init_console_info(gc, &console, 0);
> -
> - need_qemu = libxl__need_xenpv_qemu(gc, 1, &console,
> - d_config->num_vfbs, d_config->vfbs,
> - d_config->num_disks, &d_config->disks[0],
> - d_config->num_channels, &d_config->channels[0]);
> -
> console.backend_domid = state->console_domid;
> libxl__device_console_add(gc, domid, &console, state, &device);
> libxl__device_console_dispose(&console);
>
> - if (need_qemu) {
> + if (libxl__need_xenpv_qemu(gc, d_config)) {
> dcs->dmss.dm.guest_domid = domid;
> libxl__spawn_local_dm(egc, &dcs->dmss.dm);
> return;
> diff --git a/tools/libxl/libxl_dm.c b/tools/libxl/libxl_dm.c
> index 4aca38e..4b840f4 100644
> --- a/tools/libxl/libxl_dm.c
> +++ b/tools/libxl/libxl_dm.c
> @@ -2107,61 +2107,34 @@ int libxl__destroy_device_model(libxl__gc *gc, uint32_t domid)
> GCSPRINTF("/local/domain/%d/image/device-model-pid", domid));
> }
>
> -int libxl__need_xenpv_qemu(libxl__gc *gc,
> - int nr_consoles, libxl__device_console *consoles,
> - int nr_vfbs, libxl_device_vfb *vfbs,
> - int nr_disks, libxl_device_disk *disks,
> - int nr_channels, libxl_device_channel *channels)
> +int libxl__need_xenpv_qemu(libxl__gc *gc, libxl_domain_config *d_config)
> {
> - int i, ret = 0;
> + int i, ret;
> uint32_t domid;
>
> - /*
> - * qemu is required in order to support 2 or more consoles. So switch all
> - * backends to qemu if this is the case
> - */
> - if (nr_consoles > 1) {
> - for (i = 0; i < nr_consoles; i++)
> - consoles[i].consback = LIBXL__CONSOLE_BACKEND_IOEMU;
> + ret = libxl__get_domid(gc, &domid);
> + if (ret) goto out;
Doesn't this mean that if libxl__get_domid() fails,
libxl__need_xenpv_qemu() will effectively return "true"?
I realize this is a bug in the existing code but there's no need to be
bug-for-bug compatible here. :-)
Maybe add an 'rc' variable, and make this:
rc = libxl__get_domid();
if(rc) goto out;
Other than that, looks good to me.
-George
_______________________________________________
Xen-devel mailing list
Xen-devel@lists.xen.org
http://lists.xen.org/xen-devel
next prev parent reply other threads:[~2016-03-17 16:53 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-03-10 15:00 [PATCH 0/3] libxl: add support for qemu base pvusb backend Juergen Gross
2016-03-10 15:00 ` [PATCH 1/3] libxl: make libxl__need_xenpv_qemu() operate on domain config Juergen Gross
2016-03-17 16:53 ` George Dunlap [this message]
2016-03-18 8:23 ` Juergen Gross
2016-03-10 15:00 ` [PATCH 2/3] libxl: add domain config parameter to force start of qemu Juergen Gross
2016-03-17 16:06 ` George Dunlap
2016-03-18 8:11 ` Juergen Gross
2016-03-21 14:28 ` George Dunlap
2016-03-21 14:44 ` Juergen Gross
2016-03-10 15:00 ` [PATCH 3/3] libxl: add new pvusb backend "qusb" provided by qemu Juergen Gross
2016-03-17 16:39 ` George Dunlap
2016-03-18 8:22 ` Juergen Gross
2016-03-17 16:55 ` George Dunlap
2016-03-18 8:24 ` Juergen Gross
2016-03-24 20:07 ` Wei Liu
2016-03-25 6:09 ` Juergen Gross
2016-03-29 9:45 ` George Dunlap
2016-03-29 10:35 ` Juergen Gross
2016-03-30 4:43 ` Juergen Gross
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=CAFLBxZbua-BUcqjbhPcCjQRMqoAaJUGwM-9XOOZmAzV7h-sUxA@mail.gmail.com \
--to=george.dunlap@eu.citrix.com \
--cc=cyliu@suse.com \
--cc=ian.jackson@eu.citrix.com \
--cc=jgross@suse.com \
--cc=stefano.stabellini@eu.citrix.com \
--cc=wei.liu2@citrix.com \
--cc=xen-devel@lists.xen.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).