From: David Gibson <david@gibson.dropbear.id.au>
To: Greg Kurz <groug@kaod.org>
Cc: qemu-ppc@nongnu.org, qemu-devel@nongnu.org
Subject: Re: [PATCH] spapr: Don't request to unplug the same core twice
Date: Mon, 28 Oct 2019 23:20:06 +0100 [thread overview]
Message-ID: <20191028222006.GP3552@umbus.metropole.lan> (raw)
In-Reply-To: <20191024082854.1ea22701@bahia.lan>
[-- Attachment #1: Type: text/plain, Size: 2811 bytes --]
On Thu, Oct 24, 2019 at 08:28:54AM +0200, Greg Kurz wrote:
> On Thu, 24 Oct 2019 09:38:17 +1100
> David Gibson <david@gibson.dropbear.id.au> wrote:
>
> > On Wed, Oct 23, 2019 at 09:17:40PM +0200, Greg Kurz wrote:
> > > We must not call spapr_drc_detach() on a detached DRC otherwise bad things
> > > can happen, ie. QEMU hangs or crashes. This is easily demonstrated with
> > > a CPU hotplug/unplug loop using QMP.
> > >
> > > Signed-off-by: Greg Kurz <groug@kaod.org>
> >
> > Ouch, good catch. Applied.
> >
> > I wonder if we have the same problem with other DRC types.
> >
>
> We don't have it with PHB and PCI types, through the same use of
> spapr_drc_unplug_requested().
>
> LMBs see to avoid it by failing device_del early if an unplug
> request is already in progress:
>
> /*
> * An existing pending dimm state for this DIMM means that there is an
> * unplug operation in progress, waiting for the spapr_lmb_release
> * callback to complete the job (BQL can't cover that far). In this case,
> * bail out to avoid detaching DRCs that were already released.
> */
> if (spapr_pending_dimm_unplugs_find(spapr, dimm)) {
> error_setg(&local_err,
> "Memory unplug already in progress for device %s",
> dev->id);
> goto out;
> }
>
> Not sure why we error out in this case instead of ignoring the unplug
> request.
I suspect no particularly good reason, just history. Everything's a
bit different with LMBs because a single device_del will usually
remove a whole bunch of LMB DRCs. In general the interfacing between
the qemu user side DIMM handling and the PAPR side LMB/DRC handling
is... pretty clunky.
>
> > > ---
> > > hw/ppc/spapr.c | 7 ++++---
> > > 1 file changed, 4 insertions(+), 3 deletions(-)
> > >
> > > diff --git a/hw/ppc/spapr.c b/hw/ppc/spapr.c
> > > index f9410d390a07..94f9d27096af 100644
> > > --- a/hw/ppc/spapr.c
> > > +++ b/hw/ppc/spapr.c
> > > @@ -3741,9 +3741,10 @@ void spapr_core_unplug_request(HotplugHandler *hotplug_dev, DeviceState *dev,
> > > spapr_vcpu_id(spapr, cc->core_id));
> > > g_assert(drc);
> > >
> > > - spapr_drc_detach(drc);
> > > -
> > > - spapr_hotplug_req_remove_by_index(drc);
> > > + if (!spapr_drc_unplug_requested(drc)) {
> > > + spapr_drc_detach(drc);
> > > + spapr_hotplug_req_remove_by_index(drc);
> > > + }
> > > }
> > >
> > > int spapr_core_dt_populate(SpaprDrc *drc, SpaprMachineState *spapr,
> > >
> >
>
--
David Gibson | I'll have my music baroque, and my code
david AT gibson.dropbear.id.au | minimalist, thank you. NOT _the_ _other_
| _way_ _around_!
http://www.ozlabs.org/~dgibson
[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]
prev parent reply other threads:[~2019-10-29 8:01 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-23 19:17 [PATCH] spapr: Don't request to unplug the same core twice Greg Kurz
2019-10-23 22:38 ` David Gibson
2019-10-24 6:28 ` Greg Kurz
2019-10-28 22:20 ` David Gibson [this message]
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=20191028222006.GP3552@umbus.metropole.lan \
--to=david@gibson.dropbear.id.au \
--cc=groug@kaod.org \
--cc=qemu-devel@nongnu.org \
--cc=qemu-ppc@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).