Re: [PATCH v2 03/16] domain: add rendernode attribute on <accel>

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



On Fri, Aug 23, 2019 at 01:04:53PM -0400, Cole Robinson wrote:
On 8/23/19 12:21 PM, Cole Robinson wrote:
From: Marc-André Lureau <marcandre.lureau@xxxxxxxxxx>

vhost-user-gpu helper may accept --render-node option to specify on
which GPU should the renderning be done.


What does it do if the user doesn't pass one? Pick one itself, or just
not use one somehow?

If it picks one, then we may need to treat this like we treat other
rendernode instances and autoallocate one if the user doesn't specify,
otherwise we won't be able to add the path to the cgroup for example, or
selinux label it if necessary. I'm not sure if that actually applies in
this case, but it's worth considering.


Even if we'll chose one automatically, we should at least output it in
live XML.

(by comparison <graphics> rendernode is the target/display rendering)

Signed-off-by: Marc-André Lureau <marcandre.lureau@xxxxxxxxxx>
Signed-off-by: Cole Robinson <crobinso@xxxxxxxxxx>

Also I see now that I accidentally signed off all these patches, that
was not intentional. Please strip those from v3


Per https://libvirt.org/governance.html#codeofconduct we need your
intentional agreement with the Developer Certificate of Origin:
https://developercertificate.org/ to be able to include your changes.

I'd be okay from stripping it from the patches you did not actually
change, but it should stay on those where you actually changed
something.

Also, we do not have that process documented, but the kernel practice
(and what I've been doing) seems to be that the maintainer merging the
patch also adds a sign-off to provide that chain of certificates.

Thirdly, I see some patches committed to libvirt by Marc-André but
no mention in AUTHORS.in history - maybe it's time to get his commit
access restored and documented?

---
 docs/formatdomain.html.in     |  5 +++++
 docs/schemas/domaincommon.rng |  5 +++++
 src/conf/domain_conf.c        | 18 +++++++++++++++++-
 src/conf/domain_conf.h        |  1 +
 4 files changed, 28 insertions(+), 1 deletion(-)


@@ -15474,6 +15481,11 @@ virDomainVideoDefParseXML(virDomainXMLOptionPtr xmlopt,
             goto error;
         }
     }
+    if (!def->vhostuser && def->accel && def->accel->rendernode) {
+        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
+                       _("unsupported rendernode accel attribute without 'vhost-user'"));
+        goto error;
+    }


This function doesn't represent best practices, but this style of check
should be moved to virDomainVideoDefValidate IMO


Yes.

Jano

Thanks,
Cole

--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list

Attachment: signature.asc
Description: PGP signature

--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list

[Index of Archives]     [Virt Tools]     [Libvirt Users]     [Lib OS Info]     [Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Big List of Linux Books]     [Yosemite News]     [KDE Users]     [Fedora Tools]

  Powered by Linux