I'm wondering if anybody has used the new QEMU plugin feature to write a
plugin which drives the UCSD simpoints engine?
Our approach to this problem was to make the plugin sensitive to the execution
of well-chosen instructions to turn on/turn off the actions of the plugin. The
plugin will still be called (don't unregister it) but should return immediately
if the on/off state is off. This will still slow the emul
also don't see this in (guessing here...)
https://github.com/rth7680/qemu/branches/active branch tgt-arm-sve-3
Is there another repo/branch I could try?
Robert Henry
I'm running on a host ubuntu 21.04 system on a Neoverse ARM64 processor
(Ampere) processor.
I want to run /usr/bin/qemu-system-aarch64 in kvm mode, where the guest is
an ARM64 running ubuntu 20.10. (The guest runs fine in qemu tcg mode.)
I am unable to get the expected invocation to run. The qe
On 3/23/21 4:34 PM, Pranith Kumar wrote:
On Tue, Mar 23, 2021 at 2:04 PM Robert Henry
mailto:rrh.he...@gmail.com>> wrote:
my invocation is:
/usr/bin/qemu-system-aarch64 \
-m size=4096 \
-smp cpus=4 \
-accel kvm \
-cpu host \
-nographic \
-machine virt \
-bios $HOME/QEMU_EFI.fd \
-dr
On 3/23/21 5:14 PM, Pranith Kumar wrote:
On Tue, Mar 23, 2021 at 5:05 PM Robert Henry
mailto:robhe...@microsoft.com>> wrote:
On 3/23/21 4:34 PM, Pranith Kumar wrote:
On Tue, Mar 23, 2021 at 2:04 PM Robert Henry
mailto:rrh.he...@gmail.com>> wrote:
my invocation is:
/usr/bin
On 3/23/21 5:25 PM, Pranith Kumar wrote:
On Tue, Mar 23, 2021 at 5:22 PM Robert Henry
mailto:robhe...@microsoft.com>> wrote:
On ThunderX2, please use version 2. I am not sure if version 3 is supported.
Invoking in thunderx2, machine virt,gic-version=2 gives me the error message:
qemu-
On 3/24/21 4:17 AM, Peter Maydell wrote:
> On Wed, 24 Mar 2021 at 00:05, Robert Henry wrote:
>> Running on ampere/mtjade, this ,gic-version=3 is needed, and makes kvm mode
>> work as expected!
> Usually the best setting is gic-version=host which (like -cpu host) says
> &q
is in user mode, the values for vcpu_index slowly
increase without an apparent upper bound known statically (or when the
plugin is loaded?).
Robert Henry
: Sunday, September 19, 2021 10:54 AM
To: rrh.henry ; qemu-discuss@nongnu.org
Cc: Alex Bennée ; qemu-devel
Subject: [EXTERNAL] Re: Range of vcpu_index to plugin callbacks
(Cc'ing qemu-devel@ mailing list since this is a development question).
On 9/19/21 19:44, Robert Henry wrote:
> What is t
:
https://lists.gnu.org/archive/html/qemu-devel/2021-02/msg04557.html
and a related question I asked regarding the fidelity of memory traces from
ARM64 vector instructions:
https://lists.nongnu.org/archive/html/qemu-discuss/2020-12/msg00013.html
Robert Henry
I'm wondering if there is any prior art in changing the implementation of
the aarch64 store-exclusive instructions so that they fail with some
additional backoff protocol. For example, if the store-exclusive
instruction were to succeed, then it will actually fail with a probability
of P, and thus
Running stock Ubuntu Linux, where host=linux, guest=linux, arch=aarch64, I
want the guest linux kernel to use a fast, high bandwidth, low latency,
insecure and deterministic source for "entropy" as exposed in the guest on
/dev/urandom. Put another way, I do not want the guest to block during
boot
I cannot figure out how to make QEMU TCG emulate aarch64 Graviton G3. I'm
working against master/head of the QEMU source.
I come up with feature names by consulting
https://marcin.juszkiewicz.com.pl/download/tables/arm-socs.html which is
compendium of the /proc/cpuinfo info crowd sourced from
well for your evaluation.
Robert Henry
(robhe...@microsoft.com through 28June2024, but always at
rrh.he...@gmail.com)
as worked "best" for other plugins that are not
currently part of the QEMU source tree?
I will try to get a repo in github.com/microsoft/ but I do not know what is
involved.
Robert Henry
robhe...@microsoft.com
rrh.he...@gmail.com
https://github.com/robertHenry6bev
16 matches
Mail list logo
