From: David Gibson <david@gibson.dropbear.id.au>
To: Greg Kurz <groug@kaod.org>
Cc: "Aravinda Prasad" <arawinda.p@gmail.com>,
"Alexey Kardashevskiy" <aik@ozlabs.ru>,
"Mahesh Salgaonkar" <mahesh@linux.vnet.ibm.com>,
qemu-devel@nongnu.org, "Nicholas Piggin" <npiggin@gmail.com>,
"Cédric Le Goater" <clg@fr.ibm.com>,
"Ganesh Goudar" <ganeshgr@linux.ibm.com>,
qemu-ppc@nongnu.org
Subject: Re: [PATCH v2 4/4] ppc/spapr: Don't kill the guest if a recovered FWNMI machine check delivery fails
Date: Thu, 26 Mar 2020 11:30:31 +1100 [thread overview]
Message-ID: <20200326003031.GP36889@umbus.fritz.box> (raw)
In-Reply-To: <20200325191332.7da79231@bahia.lan>
[-- Attachment #1: Type: text/plain, Size: 4339 bytes --]
On Wed, Mar 25, 2020 at 07:13:32PM +0100, Greg Kurz wrote:
> On Thu, 26 Mar 2020 00:29:06 +1000
> Nicholas Piggin <npiggin@gmail.com> wrote:
>
> > Try to be tolerant of FWNMI delivery errors if the machine check had been
> > recovered by the host.
> >
> > Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
> > ---
> > hw/ppc/spapr_events.c | 27 ++++++++++++++++++++++-----
> > 1 file changed, 22 insertions(+), 5 deletions(-)
> >
> > diff --git a/hw/ppc/spapr_events.c b/hw/ppc/spapr_events.c
> > index c8964eb25d..b90ecb8afe 100644
> > --- a/hw/ppc/spapr_events.c
> > +++ b/hw/ppc/spapr_events.c
> > @@ -833,13 +833,25 @@ static void spapr_mce_dispatch_elog(PowerPCCPU *cpu, bool recovered)
> > /* get rtas addr from fdt */
> > rtas_addr = spapr_get_rtas_addr();
> > if (!rtas_addr) {
> > - error_report(
> > + if (!recovered) {
> > + error_report(
> > "FWNMI: Unable to deliver machine check to guest: rtas_addr not found.");
> > - qemu_system_guest_panicked(NULL);
> > + qemu_system_guest_panicked(NULL);
> > + } else {
> > + warn_report(
> > +"FWNMI: Unable to deliver machine check to guest: rtas_addr not found. "
> > +"Machine check recovered.");
> > + }
> > g_free(ext_elog);
> > return;
> > }
> >
> > + /*
> > + * Must not set interlock if the MCE does not get delivered to the guest
> > + * in the error case above.
> > + */
>
> It is a bit confusing to read "must not set interlock" and to see the
> interlock being set the line below IM-non-native-speaker-HO... also
> a small clarification of the outcome of taking the interlock without
> delivering the MCE could help people who aren't familiar with FWNMI
> to avoid doing bad things.
That's a good point. That comment made sense to me (and presumably to
Nick) because it directly answers a question I had on an earlier draft
(which moved this without comment). But someone without that context
(say, future me) could well find that confusing.
>
> What about something like the following ?
>
> /*
> * By taking the interlock, we assume that the MCE will be
> * delivered to the guest. CAUTION: don't add anything that
> * could prevent the MCE to be delivered after this line,
> * otherwise the guest won't be able to release the interlock
> * and ultimately hang/crash?
> */
I've substituted your comment inline.
>
> > + spapr->fwnmi_machine_check_interlock = cpu->vcpu_id;
> > +
>
> For improved paranoia, this could even be done just before calling
> ppc_cpu_do_fwnmi_machine_check().
>
> Anyway, the change is good enough so:
>
> Reviewed-by: Greg Kurz <groug@kaod.org>
>
> > stq_be_phys(&address_space_memory, rtas_addr + RTAS_ERROR_LOG_OFFSET,
> > env->gpr[3]);
> > cpu_physical_memory_write(rtas_addr + RTAS_ERROR_LOG_OFFSET +
> > @@ -876,9 +888,15 @@ void spapr_mce_req_event(PowerPCCPU *cpu, bool recovered)
> > * that CPU called "ibm,nmi-interlock")
> > */
> > if (spapr->fwnmi_machine_check_interlock == cpu->vcpu_id) {
> > - error_report(
> > + if (!recovered) {
> > + error_report(
> > "FWNMI: Unable to deliver machine check to guest: nested machine check.");
> > - qemu_system_guest_panicked(NULL);
> > + qemu_system_guest_panicked(NULL);
> > + } else {
> > + warn_report(
> > +"FWNMI: Unable to deliver machine check to guest: nested machine check. "
> > +"Machine check recovered.");
> > + }
> > return;
> > }
> > qemu_cond_wait_iothread(&spapr->fwnmi_machine_check_interlock_cond);
> > @@ -906,7 +924,6 @@ void spapr_mce_req_event(PowerPCCPU *cpu, bool recovered)
> > warn_report("Received a fwnmi while migration was in progress");
> > }
> >
> > - spapr->fwnmi_machine_check_interlock = cpu->vcpu_id;
> > spapr_mce_dispatch_elog(cpu, recovered);
> > }
> >
>
--
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 --]
next prev parent reply other threads:[~2020-03-26 0:43 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-03-25 14:29 [PATCH v2 0/4] FWNMI follow up patches Nicholas Piggin
2020-03-25 14:29 ` [PATCH v2 1/4] ppc/spapr: KVM FWNMI should not be enabled until guest requests it Nicholas Piggin
2020-03-25 17:17 ` Greg Kurz
2020-03-26 0:17 ` David Gibson
2020-03-25 14:29 ` [PATCH v2 2/4] ppc/spapr: Improve FWNMI machine check delivery corner case comments Nicholas Piggin
2020-03-26 0:18 ` David Gibson
2020-03-25 14:29 ` [PATCH v2 3/4] ppc/spapr: Add FWNMI machine check delivery warnings Nicholas Piggin
2020-03-25 17:24 ` Greg Kurz
2020-03-26 0:19 ` David Gibson
2020-03-25 14:29 ` [PATCH v2 4/4] ppc/spapr: Don't kill the guest if a recovered FWNMI machine check delivery fails Nicholas Piggin
2020-03-25 18:13 ` Greg Kurz
2020-03-26 0:30 ` David Gibson [this message]
2020-03-26 0:30 ` David Gibson
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=20200326003031.GP36889@umbus.fritz.box \
--to=david@gibson.dropbear.id.au \
--cc=aik@ozlabs.ru \
--cc=arawinda.p@gmail.com \
--cc=clg@fr.ibm.com \
--cc=ganeshgr@linux.ibm.com \
--cc=groug@kaod.org \
--cc=mahesh@linux.vnet.ibm.com \
--cc=npiggin@gmail.com \
--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).