qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Keith Busch <keith.busch@intel.com>
To: Markus Armbruster <armbru@redhat.com>
Cc: Bug 1529449 <1529449@bugs.launchpad.net>,
	qemu-devel@nongnu.org, Tom Yan <tom.ty89@gmail.com>
Subject: Re: [Qemu-devel] [Bug 1529449] [NEW] serial is required for -device nvme
Date: Mon, 11 Jan 2016 19:59:21 +0000	[thread overview]
Message-ID: <20160111195920.GA8743@localhost.localdomain> (raw)
In-Reply-To: <87pox8t6op.fsf@blackfin.pond.sub.org>

On Mon, Jan 11, 2016 at 05:35:50PM +0100, Markus Armbruster wrote:
> Tom Yan <tom.ty89@gmail.com> writes:
> > Public bug reported:
> >
> > I am not exactly sure if this is a bug, but I don't see why the option
> > "serial" should be required for -device nvme like the option "drive".
> > Truth is it seem to accept random string as its value anyway, if that's
> > the case, couldn't qemu just generate one for it when it's not
> > specified?
>
> You should've included a reproducer.  Here are mine:
> 
> 1. Bad error reporting on missing drive:
> 
>      $ upstream-qemu -nodefaults -device nvme
>      upstream-qemu: -device nvme: Device initialization failed
> 
>    Expected: error reported like for other devices, e.g.
> 
>      $ upstream-qemu -nodefaults -device virtio-blk
>      upstream-qemu: -device virtio-blk: drive property not set
> 
> 2. Bad error reporting on empty drive:
> 
>      $ upstream-qemu -nodefaults -drive if=none,id=foo -device nvme,drive=foo
>      upstream-qemu: -device nvme,drive=foo: Device initialization failed
> 
>    Expected: error is reported like for other devices, e.g.
> 
>      $ upstream-qemu -nodefaults -drive if=none,id=foo -device virtio-blk,drive=foo
>      upstream-qemu: -device virtio-blk,drive=foo: Device needs media, but drive is empty
> 
> 3. Bad handling of missing serial:
> 
>       $ upstream-qemu -nodefaults -drive if=none,id=foo,file=tmp.qcow2 -device nvme,drive=foo
>       upstream-qemu: -device nvme,drive=foo: Device initialization failed
> 
>    Expected: either default the serial number, like some other devices
>    do, or a decent error message.
> 
> I recommend to convert the device to realize(), and add the missing
> error_setg().  Keith?

Requiring a serial was a concious choice to push that responsibility
on the user, but I don't see a problem having the code provide default
serial string if the user does not over ride it.

If you've multiple nvme devices in your guest, creating the same serial
could cause problems with multipathing if they're basing end device
uniqueness on the serial (some do). If we have the code provide the
serial, perhaps it would be best to make each unique. That's easy enough
to append an incrementing number to the end of the serial.

  reply	other threads:[~2016-01-11 19:59 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2015-12-27  7:03 [Qemu-devel] [Bug 1529449] [NEW] serial is required for -device nvme Tom Yan
2016-01-11 16:35 ` Markus Armbruster
2016-01-11 19:59   ` Keith Busch [this message]
2016-01-12 10:12     ` Markus Armbruster
2016-04-28 18:07 ` [Qemu-devel] [Bug 1529449] " Tom Yan
2016-04-28 19:06   ` Laszlo Ersek (Red Hat)
2016-05-05  9:44 ` Tom Yan
2016-05-05  9:44 ` Tom Yan
2016-05-05  9:45 ` Tom Yan
2020-08-12 12:43 ` Laszlo Ersek (Red Hat)

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=20160111195920.GA8743@localhost.localdomain \
    --to=keith.busch@intel.com \
    --cc=1529449@bugs.launchpad.net \
    --cc=armbru@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=tom.ty89@gmail.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).