qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Richard Henderson <richard.henderson@linaro.org>
To: Cornelia Huck <cohuck@redhat.com>,
	Peter Maydell <peter.maydell@linaro.org>,
	Thomas Huth <thuth@redhat.com>,
	Laurent Vivier <lvivier@redhat.com>
Cc: qemu-arm@nongnu.org, qemu-devel@nongnu.org, kvm@vger.kernel.org,
	Eric Auger <eauger@redhat.com>,
	"Dr. David Alan Gilbert" <dgilbert@redhat.com>,
	Juan Quintela <quintela@redhat.com>,
	Gavin Shan <gshan@redhat.com>
Subject: Re: [PATCH v4 1/2] arm/kvm: add support for MTE
Date: Tue, 17 Jan 2023 09:37:42 -1000	[thread overview]
Message-ID: <b92e6786-acc6-50ef-8804-e4e3ef4eb2d6@linaro.org> (raw)
In-Reply-To: <20230111161317.52250-2-cohuck@redhat.com>

On 1/11/23 06:13, Cornelia Huck wrote:
> @@ -2136,7 +2136,7 @@ static void machvirt_init(MachineState *machine)
>   
>       if (vms->mte && (kvm_enabled() || hvf_enabled())) {
>           error_report("mach-virt: %s does not support providing "
> -                     "MTE to the guest CPU",
> +                     "emulated MTE to the guest CPU",
>                        kvm_enabled() ? "KVM" : "HVF");

Not your bug, but noticing this should use current_accel_name().

> +static inline bool arm_machine_has_tag_memory(void)
> +{
> +#ifndef CONFIG_USER_ONLY
> +    Object *obj = object_dynamic_cast(qdev_get_machine(), TYPE_VIRT_MACHINE);
> +
> +    /* so far, only the virt machine has support for tag memory */
> +    if (obj) {
> +        VirtMachineState *vms = VIRT_MACHINE(obj);
> +
> +        return vms->mte;
> +    }
> +#endif
> +    return false;
> +}

True for CONFIG_USER_ONLY, via page_get_target_data().
You should have seen check-tcg test failures...

> +void arm_cpu_mte_finalize(ARMCPU *cpu, Error **errp)
> +{
> +    bool enable_mte;
> +
> +    switch (cpu->prop_mte) {
> +    case ON_OFF_AUTO_OFF:
> +        enable_mte = false;
> +        break;
> +    case ON_OFF_AUTO_ON:
> +        if (!kvm_enabled()) {

tcg_enabled(), here and everywhere else you test for !kvm.

> +#ifdef CONFIG_KVM
> +        if (kvm_enabled() && !kvm_arm_mte_supported()) {

kvm_arm.h should get a stub inline returning false, so that the ifdef is removed.
See e.g. kvm_arm_sve_supported().

> +    default: /* AUTO */
> +        if (!kvm_enabled()) {

tcg_enabled.

> +    /* accelerator-specific enablement */
> +    if (kvm_enabled()) {
> +#ifdef CONFIG_KVM
> +        if (kvm_vm_enable_cap(kvm_state, KVM_CAP_ARM_MTE, 0)) {
> +            error_setg(errp, "Failed to enable KVM_CAP_ARM_MTE");

Ideally this ifdef could go away as well.

> +        } else {
> +            /* TODO: add proper migration support with MTE enabled */
> +            if (!mte_migration_blocker) {

Move the global variable here, as a local static?

I guess this check is to avoid adding one blocker per cpu?
I would guess the cap doesn't need enabling more than once either?


> +                error_setg(&mte_migration_blocker,
> +                           "Live migration disabled due to MTE enabled");
> +                if (migrate_add_blocker(mte_migration_blocker, NULL)) {

You pass NULL to the migrate_add_blocker errp argument...

> +                    error_setg(errp, "Failed to add MTE migration blocker");

... then make up your own generic reason for why it failed.
In this case it seems only related to another command-line option: --only-migratable.


Anyway, I wonder about hiding all of this in target/arm/kvm.c:

bool kvm_arm_enable_mte(Error *errp)
{
     static bool once = false;
     Error *blocker;

     if (once) {
         return;
     }
     once = true;

     if (kvm_vm_enable_cap(kvm_state, KVM_CAP_ARM_MTE, 0)) {
         error_setg_errno(errp, "Failed to enable KVM_CAP_ARM_MTE");
         return false;
     }

     blocker = g_new0(Error);
     error_setg(blocker, "Live migration disabled....");
     return !migrate_add_blocker(blocker, errp);
}

with

static inline bool kvm_arm_enable_mte(Error *errp)
{
     g_assert_not_reached();
}

in the !CONFIG_KVM block in kvm_arm.h.




  parent reply	other threads:[~2023-01-17 19:38 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-01-11 16:13 [PATCH v4 0/2] arm: enable MTE for QEMU + kvm Cornelia Huck
2023-01-11 16:13 ` [PATCH v4 1/2] arm/kvm: add support for MTE Cornelia Huck
2023-01-17 16:16   ` Peter Maydell
2023-01-17 16:50     ` Dr. David Alan Gilbert
2023-01-17 16:59       ` Cornelia Huck
2023-01-17 17:11         ` Dr. David Alan Gilbert
2023-01-17 17:01       ` Peter Maydell
2023-01-17 17:53         ` Dr. David Alan Gilbert
2023-01-17 16:52     ` Cornelia Huck
2023-01-17 19:37   ` Richard Henderson [this message]
2023-01-18 17:37     ` Cornelia Huck
2023-01-23 13:50   ` Eric Auger
2023-01-26 11:47     ` Cornelia Huck
2023-01-26 12:15       ` Cornelia Huck
2023-01-11 16:13 ` [PATCH v4 2/2] qtests/arm: add some mte tests Cornelia Huck
2023-01-17  7:21   ` Philippe Mathieu-Daudé
2023-01-23 14:29   ` Eric Auger
2023-01-26 10:57     ` Cornelia Huck

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=b92e6786-acc6-50ef-8804-e4e3ef4eb2d6@linaro.org \
    --to=richard.henderson@linaro.org \
    --cc=cohuck@redhat.com \
    --cc=dgilbert@redhat.com \
    --cc=eauger@redhat.com \
    --cc=gshan@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=lvivier@redhat.com \
    --cc=peter.maydell@linaro.org \
    --cc=qemu-arm@nongnu.org \
    --cc=qemu-devel@nongnu.org \
    --cc=quintela@redhat.com \
    --cc=thuth@redhat.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).