From: Andrew Jones <drjones@redhat.com>
To: Laurent Vivier <lvivier@redhat.com>
Cc: Eduardo Habkost <ehabkost@redhat.com>,
Igor Mammedov <imammedo@redhat.com>,
Bharata B Rao <bharata@linux.vnet.ibm.com>,
David Gibson <david@gibson.dropbear.id.au>,
Peter Maydell <peter.maydell@linaro.org>,
Anthony Green <green@moxielogic.com>,
qemu-devel@nongnu.org, Alexander Graf <agraf@suse.de>,
Max Filippov <jcmvbkbc@gmail.com>,
Greg Ungerer <gerg@uclinux.org>,
"Edgar E . Iglesias" <edgar.iglesias@gmail.com>,
Guan Xuetao <gxt@mprc.pku.edu.cn>, Jia Liu <proljc@gmail.com>,
Markus Armbruster <armbru@redhat.com>,
Artyom Tarasenko <atar4qemu@gmail.com>,
Chen Gang <gang.chen.5i5j@gmail.com>, Greg Kurz <groug@kaod.org>,
qemu-arm@nongnu.org, Richard Henderson <rth@twiddle.net>,
Matthew Rosato <mjrosato@linux.vnet.ibm.com>,
Bastian Koppelmann <kbastian@mail.uni-paderborn.de>,
Michael Walle <michael@walle.cc>,
qemu-ppc@nongnu.org, Paolo Bonzini <pbonzini@redhat.com>,
Aurelien Jarno <aurelien@aurel32.net>
Subject: Re: [Qemu-devel] [PATCH v4 2/3] exec: move cpu_exec_init() calls to realize functions
Date: Wed, 19 Oct 2016 10:13:23 +0200 [thread overview]
Message-ID: <20161019081323.b4tle74njsh3locg@hawk.localdomain> (raw)
In-Reply-To: <1476818573-5228-3-git-send-email-lvivier@redhat.com>
On Tue, Oct 18, 2016 at 09:22:52PM +0200, Laurent Vivier wrote:
> Modify all CPUs to call it from XXX_cpu_realizefn() function.
>
> Remove all the cannot_destroy_with_object_finalize_yet as
> unsafe references have been moved to cpu_exec_realizefn().
> (tested with QOM command provided by commit 4c315c27)
>
> for arm:
>
> Setting of cpu->mp_affinity is moved from arm_cpu_initfn()
> to arm_cpu_realizefn() as setting of cpu_index is now done
> in cpu_exec_realizefn(). To avoid to overwrite an user defined
> value, we set it to an invalid value by default, and update
> it in realize function only if the value is still invalid.
>
> Signed-off-by: Laurent Vivier <lvivier@redhat.com>
> Reviewed-by: David Gibson <david@gibson.dropbear.id.au>
> Reviewed-by: Igor Mammedov <imammedo@redhat.com>
> ---
[...]
> diff --git a/target-arm/cpu.c b/target-arm/cpu.c
> index 1b9540e..f0d2074 100644
> --- a/target-arm/cpu.c
> +++ b/target-arm/cpu.c
> @@ -33,6 +33,8 @@
> #include "sysemu/kvm.h"
> #include "kvm_arm.h"
>
> +#define MP_AFFINITY_INVALID (~ARM64_AFFINITY_MASK)
I would have defined this next to ARM64_AFFINITY_MASK in
target-arm/cpu-qom.h
> +
> static void arm_cpu_set_pc(CPUState *cs, vaddr value)
> {
> ARMCPU *cpu = ARM_CPU(cs);
> @@ -441,22 +443,11 @@ static void arm_cpu_initfn(Object *obj)
> CPUState *cs = CPU(obj);
> ARMCPU *cpu = ARM_CPU(obj);
> static bool inited;
> - uint32_t Aff1, Aff0;
>
> cs->env_ptr = &cpu->env;
> - cpu_exec_init(cs, &error_abort);
> cpu->cp_regs = g_hash_table_new_full(g_int_hash, g_int_equal,
> g_free, g_free);
>
> - /* This cpu-id-to-MPIDR affinity is used only for TCG; KVM will override it.
> - * We don't support setting cluster ID ([16..23]) (known as Aff2
> - * in later ARM ARM versions), or any of the higher affinity level fields,
> - * so these bits always RAZ.
> - */
> - Aff1 = cs->cpu_index / ARM_CPUS_PER_CLUSTER;
> - Aff0 = cs->cpu_index % ARM_CPUS_PER_CLUSTER;
> - cpu->mp_affinity = (Aff1 << ARM_AFF1_SHIFT) | Aff0;
> -
> #ifndef CONFIG_USER_ONLY
> /* Our inbound IRQ and FIQ lines */
> if (kvm_enabled()) {
> @@ -576,6 +567,14 @@ static void arm_cpu_realizefn(DeviceState *dev, Error **errp)
> ARMCPU *cpu = ARM_CPU(dev);
> ARMCPUClass *acc = ARM_CPU_GET_CLASS(dev);
> CPUARMState *env = &cpu->env;
> + Error *local_err = NULL;
> + uint32_t Aff1, Aff0;
> +
> + cpu_exec_realizefn(cs, &local_err);
> + if (local_err != NULL) {
> + error_propagate(errp, local_err);
> + return;
> + }
>
> /* Some features automatically imply others: */
> if (arm_feature(env, ARM_FEATURE_V8)) {
> @@ -631,6 +630,17 @@ static void arm_cpu_realizefn(DeviceState *dev, Error **errp)
> set_feature(env, ARM_FEATURE_THUMB_DSP);
> }
>
> + /* This cpu-id-to-MPIDR affinity is used only for TCG; KVM will override it.
> + * We don't support setting cluster ID ([16..23]) (known as Aff2
> + * in later ARM ARM versions), or any of the higher affinity level fields,
> + * so these bits always RAZ.
> + */
> + if (cpu->mp_affinity == MP_AFFINITY_INVALID) {
> + Aff1 = cs->cpu_index / ARM_CPUS_PER_CLUSTER;
> + Aff0 = cs->cpu_index % ARM_CPUS_PER_CLUSTER;
I think ARM_CPUS_PER_CLUSTER should be renamed to
ARM_DEFAULT_CPUS_PER_CLUSTER and either moved from where
it's currently defined (above arm_cpu_initfn) to just above
arm_cpu_realizefn, or to the same place as ARM64_AFFINITY_MASK
and MP_AFFINITY_INVALID.
Aff0 and Aff1 could be declared in this scope, as they're only
used here.
> + cpu->mp_affinity = (Aff1 << ARM_AFF1_SHIFT) | Aff0;
> + }
> +
> if (cpu->reset_hivecs) {
> cpu->reset_sctlr |= (1 << 13);
> }
> @@ -1461,7 +1471,7 @@ static Property arm_cpu_properties[] = {
> DEFINE_PROP_BOOL("start-powered-off", ARMCPU, start_powered_off, false),
> DEFINE_PROP_UINT32("psci-conduit", ARMCPU, psci_conduit, 0),
> DEFINE_PROP_UINT32("midr", ARMCPU, midr, 0),
> - DEFINE_PROP_UINT64("mp-affinity", ARMCPU, mp_affinity, 0),
> + DEFINE_PROP_UINT64("mp-affinity", ARMCPU, mp_affinity, MP_AFFINITY_INVALID),
> DEFINE_PROP_END_OF_LIST()
> };
>
> @@ -1533,17 +1543,6 @@ static void arm_cpu_class_init(ObjectClass *oc, void *data)
> cc->debug_check_watchpoint = arm_debug_check_watchpoint;
>
> cc->disas_set_info = arm_disas_set_info;
> -
> - /*
> - * Reason: arm_cpu_initfn() calls cpu_exec_init(), which saves
> - * the object in cpus -> dangling pointer after final
> - * object_unref().
> - *
> - * Once this is fixed, the devices that create ARM CPUs should be
> - * updated not to set cannot_destroy_with_object_finalize_yet,
> - * unless they still screw up something else.
> - */
> - dc->cannot_destroy_with_object_finalize_yet = true;
> }
>
> static void cpu_register(const ARMCPUInfo *info)
[...]
Otherwise looks good to me.
Thanks,
drew
next prev parent reply other threads:[~2016-10-19 8:13 UTC|newest]
Thread overview: 9+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-10-18 19:22 [Qemu-devel] [PATCH v4 0/3] Split cpu_exec_init() into an init and a realize part Laurent Vivier
2016-10-18 19:22 ` [Qemu-devel] [PATCH v4 1/3] exec: split cpu_exec_init() Laurent Vivier
2016-10-18 19:22 ` [Qemu-devel] [PATCH v4 2/3] exec: move cpu_exec_init() calls to realize functions Laurent Vivier
2016-10-19 8:13 ` Andrew Jones [this message]
2016-10-19 8:30 ` Laurent Vivier
2016-10-19 13:46 ` Andrew Jones
2016-10-19 13:56 ` Laurent Vivier
2016-10-18 19:22 ` [Qemu-devel] [PATCH v4 3/3] exec: call cpu_exec_exit() from a CPU unrealize common function Laurent Vivier
2016-10-18 19:56 ` [Qemu-devel] [PATCH v4 0/3] Split cpu_exec_init() into an init and a realize part Eduardo Habkost
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=20161019081323.b4tle74njsh3locg@hawk.localdomain \
--to=drjones@redhat.com \
--cc=agraf@suse.de \
--cc=armbru@redhat.com \
--cc=atar4qemu@gmail.com \
--cc=aurelien@aurel32.net \
--cc=bharata@linux.vnet.ibm.com \
--cc=david@gibson.dropbear.id.au \
--cc=edgar.iglesias@gmail.com \
--cc=ehabkost@redhat.com \
--cc=gang.chen.5i5j@gmail.com \
--cc=gerg@uclinux.org \
--cc=green@moxielogic.com \
--cc=groug@kaod.org \
--cc=gxt@mprc.pku.edu.cn \
--cc=imammedo@redhat.com \
--cc=jcmvbkbc@gmail.com \
--cc=kbastian@mail.uni-paderborn.de \
--cc=lvivier@redhat.com \
--cc=michael@walle.cc \
--cc=mjrosato@linux.vnet.ibm.com \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=proljc@gmail.com \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=qemu-ppc@nongnu.org \
--cc=rth@twiddle.net \
/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).