Linux KVM/arm64 development list
 help / color / mirror / Atom feed
From: Marc Zyngier <maz@kernel.org>
To: Itaru Kitayama <itaru.kitayama@linux.dev>
Cc: kvmarm@lists.linux.dev
Subject: Re: KVM selftest for L2 guest execution fails
Date: Wed, 30 Jul 2025 13:24:24 +0100	[thread overview]
Message-ID: <861ppx9893.wl-maz@kernel.org> (raw)
In-Reply-To: <C37BCE4D-0958-4C74-BEFC-F3A7A689A848@linux.dev>

On Wed, 30 Jul 2025 07:44:48 +0100,
Itaru Kitayama <itaru.kitayama@linux.dev> wrote:
> 
> Hi,
> My small kvm sefltest tested on the kvmarm/next with NV2+VHE on FVP fails:
> 
> # ./arm64/test5
> Random seed: 0x6b8b4567
> ==== Test Assertion Failure ====
>   lib/kvm_util.c:1746: !ret
>   pid=257 tid=257 errno=11 - Resource temporarily unavailable
> sh: line 1: addr2line: command not found
>   KVM_RUN failed, rc: -1 errno: 11 (Resource temporarily unavailable)
> 
> Below is the test code, am I doing still wrong?

Quite a lot is wrong there.

> 
> Thanks,
> Itaru.
> 
>  // SPDX-License-Identifier: GPL-2.0-only
> 
> #include "test_util.h"
> #include "kvm_util.h"
> #include "processor.h"
> #include "ucall.h"
> 
> #define GUEST_SYNC_VAL 0x1234
> #define UCALL_GPA 0x500000
> 
> static void l1_guest_code(void)
> {
>         /*
>         * Set up HCR_EL2 to enable virtualization and nested support:
>         *   - HCR_EL2.VM  = 1 (enable stage-2 translation)
>         *   - HCR_EL2.NV  = 1 (trap to EL2 on guest HVC, SMC, WFI, etc.)
>         *   - HCR_EL2.TGE = 1 (trap general exceptions in EL1 to EL2)

You really want to read the spec again -- this is not what this means.

>         */
>        uint64_t hcr = (1UL << 0)   |  // VM
>                       (1UL << 27)  |  // TGE
>                       (1UL << 31);    // NV

- You enable S2 translation, and yet don't setup anything related to it

- You have TGE set, and yet (try to) return to EL1

- You set NV, and yet don't have a VNCR_EL2 populated

I don't know what you are trying to do here, but this is pretty badly
thought out.

I'd suggest you start by having a simple L2, without making use of NV
at all in the guest. You will still have to have a stage-2 and a
correct HCR_EL2 configuration -- none of which is as trivial as this
test suggests.

Once that is actually working, you can go wild and enable NV. But not
before.

	M.

-- 
Without deviation from the norm, progress is not possible.

  reply	other threads:[~2025-07-30 12:24 UTC|newest]

Thread overview: 3+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-07-30  6:44 KVM selftest for L2 guest execution fails Itaru Kitayama
2025-07-30 12:24 ` Marc Zyngier [this message]
2025-07-30 21:40   ` Itaru Kitayama

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=861ppx9893.wl-maz@kernel.org \
    --to=maz@kernel.org \
    --cc=itaru.kitayama@linux.dev \
    --cc=kvmarm@lists.linux.dev \
    /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