From: Eric Auger <eric.auger@redhat.com>
To: Cornelia Huck <cohuck@redhat.com>,
eric.auger.pro@gmail.com, qemu-devel@nongnu.org,
qemu-arm@nongnu.org, kvmarm@lists.linux.dev,
peter.maydell@linaro.org, richard.henderson@linaro.org,
alex.bennee@linaro.org, maz@kernel.org, oliver.upton@linux.dev,
sebott@redhat.com, shameerali.kolothum.thodi@huawei.com,
armbru@redhat.com, berrange@redhat.com, abologna@redhat.com,
jdenemar@redhat.com
Cc: agraf@csgraf.de, shahuang@redhat.com, mark.rutland@arm.com,
philmd@linaro.org, pbonzini@redhat.com
Subject: Re: [PATCH v3 00/10] kvm/arm: Introduce a customizable aarch64 KVM host model
Date: Tue, 13 May 2025 17:29:44 +0200 [thread overview]
Message-ID: <dadc4acd-97c8-4647-9467-89cd1966a6f2@redhat.com> (raw)
In-Reply-To: <20250414163849.321857-1-cohuck@redhat.com>
Hi,
On 4/14/25 6:38 PM, Cornelia Huck wrote:
> [This is on top of v5 of the ID register storage rework:
> https://lore.kernel.org/qemu-devel/20250409144304.912325-1-cohuck@redhat.com/T/#t]
>
> It's been a while, but here's finally a respin of the series to make ID
> registers configurable directly via the command line.
>
> Major changes from v2 include:
> - split out the rework of ID register storage (see above) and rebased on top
> of that
> - hopefully improved the documentation
> - fixed some bugs along the way (including output of the cpu model expansion
> command, and compilation with HVF on)
>
> Decided against zeroing unknown registers; it's mostly a "dammed if you do,
> dammed if you don't" situation as one case or the other will not quite work
> as desired, even disregarding fields like AA64PFR1_EL1.MTE_frac where 0
> might indicate things we do not support. You'll need to be careful when
> doing kernel updates and be explicit with configuring fields.
This is not totally clear to me. Is it possible to come with examples
and also derive some guidelines to avoid pitfalls? I guess this kind of
aknowledgement can be scary.
>
> The current cpu properties stay as they are; we can work on converting them
> to compatibility props once we have support for configuring the ID register
> fields on the command line for the other accelerators (this is still KVM only.)
At least what can we do to make sure they are not inconsistent and work
well together. Looks like a must have before being non RFC.
>
> The FEAT_xxx features only support a subset of what we need to configure in
> real life; for example, different AltraMax machines differ in CTR_EL0, which
> is not covered by any FEAT_. It might make sense to provide them as syntactic
> sugar on top.
I agree that anyway we will need an ID reg field granularity for some
real life cases. On top of that we shall try to build higher level props
matching to features as requested by an overwhelming majority in earlier
comments.
>
> We still have to deal with MIDR/REVIDR/AIDR differences by exploiting
> https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit?id=d300b0168ea8fd5022a1413bd37ab63f4e5a7d4d
> (in a different series.)
Shameer, do you plan to contribute the qemu integration of this feature
you developped on kernel side; or do you allow us to integrate it in
this series?
Thanks
Eric
>
> I have not ignored the issue of named models on Arm, I just wanted to get the
> base infrastructure into place first :)
>
> Real world example (migration Graviton 3 -> 4, thx to Sebastian):
>
> -cpu host,pauth=off,SYSREG_ID_AA64PFR0_EL1_SEL2=0,SYSREG_ID_AA64PFR0_EL1_EL0=1,
> SYSREG_ID_AA64ISAR0_EL1_TLB=0,SYSREG_ID_AA64ISAR0_EL1_TS=0,
> SYSREG_ID_AA64ISAR0_EL1_SM4=0,SYSREG_ID_AA64ISAR0_EL1_SM3=0,
> SYSREG_ID_AA64ISAR1_EL1_SPECRES=0,SYSREG_ID_AA64ISAR1_EL1_SB=0,
> SYSREG_ID_AA64ISAR1_EL1_FRINTTS=0,SYSREG_ID_AA64MMFR0_EL1_TGRAN4_2=1,
> SYSREG_ID_AA64MMFR0_EL1_TGRAN16_2=1,SYSREG_ID_AA64MMFR0_EL1_TGRAN64_2=1
>
> (not including handling MIDR differences, which is out of scope for this series)
>
> Code also available at
> https://gitlab.com/cohuck/qemu/-/tree/arm-cpu-model-rfcv3?ref_type=heads
>
>
> Cornelia Huck (5):
> arm/cpu: Add generated sysreg properties
> kvm: kvm_get_writable_id_regs
> arm/cpu: accessors for writable id registers
> arm-qmp-cmds: introspection for ID register props
> arm/cpu-features: document ID reg properties
>
> Eric Auger (5):
> arm/cpu: Add infra to handle generated ID register definitions
> arm/cpu: Add sysreg properties generation
> arm/kvm: Allow reading all the writable ID registers
> arm/kvm: write back modified ID regs to KVM
> arm/cpu: more customization for the kvm host cpu model
>
> docs/system/arm/cpu-features.rst | 104 +++-
> scripts/gen-cpu-sysreg-properties.awk | 325 ++++++++++++
> scripts/update-aarch64-sysreg-code.sh | 5 +-
> target/arm/arm-qmp-cmds.c | 19 +
> target/arm/cpu-custom.h | 60 +++
> target/arm/cpu-sysreg-properties.c | 713 ++++++++++++++++++++++++++
> target/arm/cpu-sysregs.h | 2 +
> target/arm/cpu.c | 12 +
> target/arm/cpu.h | 47 ++
> target/arm/cpu64.c | 24 +-
> target/arm/kvm.c | 289 ++++++++++-
> target/arm/kvm_arm.h | 26 +-
> target/arm/meson.build | 1 +
> target/arm/trace-events | 6 +
> 14 files changed, 1617 insertions(+), 16 deletions(-)
> create mode 100755 scripts/gen-cpu-sysreg-properties.awk
> create mode 100644 target/arm/cpu-custom.h
> create mode 100644 target/arm/cpu-sysreg-properties.c
>
next prev parent reply other threads:[~2025-05-13 15:29 UTC|newest]
Thread overview: 59+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-04-14 16:38 [PATCH v3 00/10] kvm/arm: Introduce a customizable aarch64 KVM host model Cornelia Huck
2025-04-14 16:38 ` [PATCH v3 01/10] arm/cpu: Add infra to handle generated ID register definitions Cornelia Huck
2025-05-13 13:52 ` Eric Auger
2025-05-13 14:05 ` Cornelia Huck
2025-05-13 15:12 ` Eric Auger
2025-04-14 16:38 ` [PATCH v3 02/10] arm/cpu: Add sysreg properties generation Cornelia Huck
2025-04-15 7:09 ` Philippe Mathieu-Daudé
2025-04-15 7:20 ` Philippe Mathieu-Daudé
2025-05-19 14:49 ` Cornelia Huck
2025-05-13 15:23 ` Daniel P. Berrangé
2025-05-14 15:25 ` Cornelia Huck
2025-05-14 15:29 ` Daniel P. Berrangé
2025-04-14 16:38 ` [PATCH v3 03/10] arm/cpu: Add generated sysreg properties Cornelia Huck
2025-04-14 16:38 ` [PATCH v3 04/10] kvm: kvm_get_writable_id_regs Cornelia Huck
2025-05-13 14:20 ` Eric Auger
2025-05-13 14:42 ` Cornelia Huck
2025-05-13 15:16 ` Eric Auger
2025-04-14 16:38 ` [PATCH v3 05/10] arm/cpu: accessors for writable id registers Cornelia Huck
2025-04-29 16:27 ` Sebastian Ott
2025-04-30 13:48 ` Cornelia Huck
2025-04-14 16:38 ` [PATCH v3 06/10] arm/kvm: Allow reading all the writable ID registers Cornelia Huck
2025-05-13 14:31 ` Eric Auger
2025-05-16 14:17 ` Cornelia Huck
2025-05-20 14:05 ` Cornelia Huck
2025-05-23 8:27 ` Shameerali Kolothum Thodi
2025-05-26 12:37 ` Cornelia Huck
2025-04-14 16:38 ` [PATCH v3 07/10] arm/kvm: write back modified ID regs to KVM Cornelia Huck
2025-04-15 7:03 ` Philippe Mathieu-Daudé
2025-04-15 9:54 ` Cornelia Huck
2025-05-13 14:33 ` Eric Auger
2025-07-02 4:01 ` Jinqian Yang
2025-07-02 8:46 ` Cornelia Huck
2025-04-14 16:38 ` [PATCH v3 08/10] arm/cpu: more customization for the kvm host cpu model Cornelia Huck
2025-05-13 14:47 ` Eric Auger
2025-05-13 15:56 ` Daniel P. Berrangé
2025-05-16 14:42 ` Cornelia Huck
2025-05-13 15:59 ` Daniel P. Berrangé
2025-05-14 15:36 ` Cornelia Huck
2025-05-14 18:22 ` Daniel P. Berrangé
2025-05-16 14:51 ` Cornelia Huck
2025-05-16 14:57 ` Daniel P. Berrangé
2025-05-16 15:13 ` Cornelia Huck
2025-09-18 7:40 ` Jinqian Yang
2025-04-14 16:38 ` [PATCH v3 09/10] arm-qmp-cmds: introspection for ID register props Cornelia Huck
2025-05-13 14:50 ` Eric Auger
2025-04-14 16:38 ` [PATCH v3 10/10] arm/cpu-features: document ID reg properties Cornelia Huck
2025-05-13 15:09 ` Eric Auger
2025-05-13 16:23 ` Daniel P. Berrangé
2025-05-13 15:29 ` Eric Auger [this message]
2025-05-14 13:47 ` [PATCH v3 00/10] kvm/arm: Introduce a customizable aarch64 KVM host model Shameerali Kolothum Thodi
2025-05-14 14:47 ` Eric Auger
2025-05-23 13:23 ` Shameerali Kolothum Thodi
2025-05-26 12:44 ` Cornelia Huck
2025-05-27 10:06 ` Cornelia Huck
2025-06-03 15:14 ` Cornelia Huck
2025-06-04 10:58 ` Shameerali Kolothum Thodi
2025-06-04 12:35 ` Cornelia Huck
2025-06-04 13:45 ` Shameerali Kolothum Thodi
2025-06-05 16:31 ` 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=dadc4acd-97c8-4647-9467-89cd1966a6f2@redhat.com \
--to=eric.auger@redhat.com \
--cc=abologna@redhat.com \
--cc=agraf@csgraf.de \
--cc=alex.bennee@linaro.org \
--cc=armbru@redhat.com \
--cc=berrange@redhat.com \
--cc=cohuck@redhat.com \
--cc=eric.auger.pro@gmail.com \
--cc=jdenemar@redhat.com \
--cc=kvmarm@lists.linux.dev \
--cc=mark.rutland@arm.com \
--cc=maz@kernel.org \
--cc=oliver.upton@linux.dev \
--cc=pbonzini@redhat.com \
--cc=peter.maydell@linaro.org \
--cc=philmd@linaro.org \
--cc=qemu-arm@nongnu.org \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
--cc=sebott@redhat.com \
--cc=shahuang@redhat.com \
--cc=shameerali.kolothum.thodi@huawei.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