From: "Andreas Färber" <afaerber@suse.de>
To: Jens Freimann <jfrei@linux.vnet.ibm.com>
Cc: Heinz Graalfs <graalfs@linux.vnet.ibm.com>,
Alexander Graf <agraf@suse.de>,
qemu-devel <qemu-devel@nongnu.org>,
Christian Borntraeger <borntraeger@de.ibm.com>,
Cornelia Huck <cornelia.huck@de.ibm.com>,
Einar Lueck <elelueck@linux.vnet.ibm.com>
Subject: Re: [Qemu-devel] [PATCH 2/3] s390: Add CPU reset handler
Date: Sun, 16 Dec 2012 16:30:21 +0100 [thread overview]
Message-ID: <50CDE90D.9000603@suse.de> (raw)
In-Reply-To: <1355503606-54131-3-git-send-email-jfrei@linux.vnet.ibm.com>
Am 14.12.2012 17:46, schrieb Jens Freimann:
> Add a CPU reset handler to have all CPUs in a PoP compliant
> state.
>
> Signed-off-by: Jens Freimann <jfrei@linux.vnet.ibm.com>
The logic looks okay now. Some comments inline.
> ---
> v2 -> v3:
> * explain in comment which code sets cpu 0 to running during IPL
>
> v1 -> v2:
> * move setting of control registers and psa to s390_cpu_reset
> and call it from the new s390_machine_cpu_reset_cb()
> This makes it more similar to how it is done on x86
> * in s390_cpu_reset() set env->halted state of cpu after
> the memset. This is needed to keep our s390_cpu_running
> counter in sync when s390_cpu_reset is called via the
> qemu_devices_reset path
> * set env->halted state in s390_cpu_initfn to 1 to avoid
> decrementing the cpu counter during first reset
> ---
> target-s390x/cpu.c | 32 ++++++++++++++++++++++++++++++--
> target-s390x/kvm.c | 9 ++++++++-
> 2 files changed, 38 insertions(+), 3 deletions(-)
>
> diff --git a/target-s390x/cpu.c b/target-s390x/cpu.c
> index 619b202..75d4036 100644
> --- a/target-s390x/cpu.c
> +++ b/target-s390x/cpu.c
> @@ -4,6 +4,7 @@
> * Copyright (c) 2009 Ulrich Hecht
> * Copyright (c) 2011 Alexander Graf
> * Copyright (c) 2012 SUSE LINUX Products GmbH
> + * Copyright (c) 2012 IBM Corp.
> *
> * This library is free software; you can redistribute it and/or
> * modify it under the terms of the GNU Lesser General Public
> @@ -18,9 +19,13 @@
> * You should have received a copy of the GNU Lesser General Public
> * License along with this library; if not, see
> * <http://www.gnu.org/licenses/lgpl-2.1.html>
> + * Contributions after 2012-12-11 are licensed under the terms of the
> + * GNU GPL, version 2 or (at your option) any later version.
> + *
> */
>
> #include "cpu.h"
> +#include "hw/hw.h"
> #include "qemu-common.h"
> #include "qemu-timer.h"
>
> @@ -37,12 +42,30 @@ static void s390_cpu_reset(CPUState *s)
> log_cpu_state(env, 0);
> }
>
> - scc->parent_reset(s);
> + s390_del_running_cpu(env);
>
> + scc->parent_reset(s);
If this gets respun, a white line separating the parent reset from the
local reset would be nice. :)
> memset(env, 0, offsetof(CPUS390XState, breakpoints));
> +
> + /* architectured initial values for CR 0 and 14 */
> + env->cregs[0] = 0xE0UL;
> + env->cregs[14] = 0xC2000000UL;
> + /* set to z/Architecture mode */
> + env->psw.mask = 0x0000000180000000ULL;
> + env->psa = 0;
> + /* set halted to 1 to make sure we can add the cpu in
> + * s390_ipl_cpu code, where env->halted is set back to 0
> + * after incrementing the cpu counter */
> + env->halted = 1;
> /* FIXME: reset vector? */
Do the above added cregs/psw/psa reset values resolve this FIXME? Or
does that refer to something different?
> tlb_flush(env, 1);
> - s390_add_running_cpu(env);
> +}
> +
> +static void s390_cpu_machine_reset_cb(void *opaque)
> +{
> + S390CPU *cpu = opaque;
> +
> + cpu_reset(CPU(cpu));
> }
>
> static void s390_cpu_initfn(Object *obj)
> @@ -66,7 +89,12 @@ static void s390_cpu_initfn(Object *obj)
> env->cpu_num = cpu_num++;
> env->ext_index = -1;
>
> + /* set env->halted state to 1 to avoid decrementing the running
> + * cpu counter in s390_cpu_reset to a negative number at
> + * initial ipl */
> + env->halted = 1;
> cpu_reset(CPU(cpu));
> + qemu_register_reset(s390_cpu_machine_reset_cb, cpu);
Since we register the reset handler in instance_init, we should
unregister it in a instance_finalize callback (uninitfn?). Since we do
not hot-unplug s390 CPUs yet to my knowledge, that could be done in a
follow-up.
(For x86 it it registered in the provisional realize function and not
unregistered lacking a matching unrealization mechanism today; elsewhere
reset registration is done in the machine.)
> }
>
> static void s390_cpu_class_init(ObjectClass *oc, void *data)
> diff --git a/target-s390x/kvm.c b/target-s390x/kvm.c
> index 94de764..fda9f1f 100644
> --- a/target-s390x/kvm.c
> +++ b/target-s390x/kvm.c
> @@ -85,7 +85,14 @@ int kvm_arch_init_vcpu(CPUS390XState *env)
>
> void kvm_arch_reset_vcpu(CPUS390XState *env)
> {
Note to Alex: In my upcoming KVM CPUState series, this argument type
changes to CPUState ...
> - /* FIXME: add code to reset vcpu. */
> + /* The initial reset call is needed here to reset in-kernel
> + * vcpu data that we can't access directly from QEMU
> + * (i.e. with older kernels which don't support sync_regs/ONE_REG).
> + * Before this ioctl cpu_synchronize_state() is called in common kvm
> + * code (kvm-all) */
> + if (kvm_vcpu_ioctl(env, KVM_S390_INITIAL_RESET, NULL)) {
... but so does the argument for kvm_vcpu_ioctl(), so merging becomes a
trivial env -> cpu change.
> + perror("Can't reset vcpu\n");
> + }
> }
>
> int kvm_arch_put_registers(CPUS390XState *env, int level)
Andreas
--
SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
GF: Jeff Hawn, Jennifer Guild, Felix Imendörffer; HRB 16746 AG Nürnberg
next prev parent reply other threads:[~2012-12-16 15:30 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-12-14 16:46 [Qemu-devel] [PATCH 0/3] s390: ipl device, cpu reset handler and cpu model support Jens Freimann
2012-12-14 16:46 ` [Qemu-devel] [PATCH 1/3] s390: Move IPL code into a separate device Jens Freimann
2012-12-16 16:26 ` Andreas Färber
2012-12-16 21:15 ` Christian Borntraeger
2012-12-14 16:46 ` [Qemu-devel] [PATCH 2/3] s390: Add CPU reset handler Jens Freimann
2012-12-16 15:30 ` Andreas Färber [this message]
2012-12-17 8:49 ` Jens Freimann
2012-12-17 14:49 ` Alexander Graf
2012-12-17 15:41 ` Jens Freimann
2012-12-17 17:21 ` Andreas Färber
2012-12-17 17:27 ` Alexander Graf
2012-12-14 16:46 ` [Qemu-devel] [PATCH 3/3] S390: Enable -cpu help and QMP query-cpu-definitions Jens Freimann
2012-12-16 15:42 ` Andreas Färber
2012-12-17 14:47 ` Alexander Graf
2012-12-17 17:32 ` Andreas Färber
-- strict thread matches above, loose matches on Subject: below --
2012-12-18 17:50 [Qemu-devel] [PATCH 0/3] s390: ipl device, cpu reset handler and cpu model support Jens Freimann
2012-12-18 17:50 ` [Qemu-devel] [PATCH 2/3] s390: Add CPU reset handler Jens Freimann
2013-01-03 12:50 ` Alexander Graf
2013-01-04 13:55 ` Jens Freimann
2013-01-03 12:55 ` Alexander Graf
2013-01-04 14:09 ` Jens Freimann
2013-01-04 14:11 ` Alexander Graf
2012-12-12 13:08 [Qemu-devel] [PATCH 0/3] s390: ipl device, cpu reset handler and cpu model support Jens Freimann
2012-12-12 13:08 ` [Qemu-devel] [PATCH 2/3] s390: Add CPU reset handler Jens Freimann
2012-12-12 13:38 ` Alexander Graf
2012-12-12 15:04 ` Jens Freimann
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=50CDE90D.9000603@suse.de \
--to=afaerber@suse.de \
--cc=agraf@suse.de \
--cc=borntraeger@de.ibm.com \
--cc=cornelia.huck@de.ibm.com \
--cc=elelueck@linux.vnet.ibm.com \
--cc=graalfs@linux.vnet.ibm.com \
--cc=jfrei@linux.vnet.ibm.com \
--cc=qemu-devel@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).