Re: [PATCH v12 2/7] s390x/cpu topology: reporting the CPU topology to the guest

From: Pierre Morel <pmorel@linux.ibm.com>
To: Janis Schoetterl-Glausch <scgl@linux.ibm.com>, qemu-s390x@nongnu.org
Cc: qemu-devel@nongnu.org, borntraeger@de.ibm.com,
	pasic@linux.ibm.com, richard.henderson@linaro.org,
	david@redhat.com, thuth@redhat.com, cohuck@redhat.com,
	mst@redhat.com, pbonzini@redhat.com, kvm@vger.kernel.org,
	ehabkost@redhat.com, marcel.apfelbaum@gmail.com,
	eblake@redhat.com, armbru@redhat.com, seiden@linux.ibm.com,
	nrb@linux.ibm.com, frankja@linux.ibm.com, berrange@redhat.com,
	clg@kaod.org
Subject: Re: [PATCH v12 2/7] s390x/cpu topology: reporting the CPU topology to the guest
Date: Tue, 6 Dec 2022 11:38:24 +0100	[thread overview]
Message-ID: <d883f3be-0ded-72bb-a08a-29a43ffa3a09@linux.ibm.com> (raw)
In-Reply-To: <be6e4c3a2a3b1b4a944ce0558d3e852f78bd9645.camel@linux.ibm.com>

On 12/6/22 10:48, Janis Schoetterl-Glausch wrote:
> On Tue, 2022-11-29 at 18:42 +0100, Pierre Morel wrote:
>> The guest uses the STSI instruction to get information on the
>> CPU topology.
>>
>> Let us implement the STSI instruction for the basis CPU topology
>> level, level 2.
>>
>> Signed-off-by: Pierre Morel <pmorel@linux.ibm.com>
>> ---
>>   target/s390x/cpu.h          |  77 +++++++++++++++
>>   hw/s390x/s390-virtio-ccw.c  |  12 +--
>>   target/s390x/cpu_topology.c | 186 ++++++++++++++++++++++++++++++++++++
>>   target/s390x/kvm/kvm.c      |   6 +-
>>   target/s390x/meson.build    |   1 +
>>   5 files changed, 274 insertions(+), 8 deletions(-)
>>   create mode 100644 target/s390x/cpu_topology.c
>>
>> diff --git a/target/s390x/cpu.h b/target/s390x/cpu.h
>> index 7d6d01325b..dd878ac916 100644
>> --- a/target/s390x/cpu.h
>> +++ b/target/s390x/cpu.h
>>
> [...]
> 
>> +/* Configuration topology */
>> +typedef struct SysIB_151x {
>> +    uint8_t  reserved0[2];
>> +    uint16_t length;
>> +    uint8_t  mag[S390_TOPOLOGY_MAG];
>> +    uint8_t  reserved1;
>> +    uint8_t  mnest;
>> +    uint32_t reserved2;
>> +    char tle[0];
> 
> AFAIK [] is preferred over [0].

grr, yes, I think I have been already told so :)

> 
>> +} QEMU_PACKED QEMU_ALIGNED(8) SysIB_151x;
>> +QEMU_BUILD_BUG_ON(sizeof(SysIB_151x) != 16);
> 
> [...]
>>
>> +/*
>> + * s390_topology_add_cpu:
>> + * @topo: pointer to the topology
>> + * @cpu : pointer to the new CPU
>> + *
>> + * The topology pointed by S390CPU, gives us the CPU topology
>> + * established by the -smp QEMU aruments.
>> + * The core-id is used to calculate the position of the CPU inside
>> + * the topology:
>> + *  - the socket, container TLE, containing the CPU, we have one socket
>> + *    for every num_cores cores.
>> + *  - the CPU TLE inside the socket, we have potentionly up to 4 CPU TLE
>> + *    in a container TLE with the assumption that all CPU are identical
>> + *    with the same polarity and entitlement because we have maximum 256
>> + *    CPUs and each TLE can hold up to 64 identical CPUs.
>> + *  - the bit in the 64 bit CPU TLE core mask
>> + */
>> +static void s390_topology_add_cpu(S390Topology *topo, S390CPU *cpu)
>> +{
>> +    int core_id = cpu->env.core_id;
>> +    int bit, origin;
>> +    int socket_id;
>> +
>> +    cpu->machine_data = topo;
>> +    socket_id = core_id / topo->num_cores;
>> +    /*
>> +     * At the core level, each CPU is represented by a bit in a 64bit
>> +     * uint64_t which represent the presence of a CPU.
>> +     * The firmware assume that all CPU in a CPU TLE have the same
>> +     * type, polarization and are all dedicated or shared.
>> +     * In that case the origin variable represents the offset of the first
>> +     * CPU in the CPU container.
>> +     * More than 64 CPUs per socket are represented in several CPU containers
>> +     * inside the socket container.
>> +     * The only reason to have several S390TopologyCores inside a socket is
>> +     * to have more than 64 CPUs.
>> +     * In that case the origin variable represents the offset of the first CPU
>> +     * in the CPU container. More than 64 CPUs per socket are represented in
>> +     * several CPU containers inside the socket container.
>> +     */
> 
> This comment still contains redundant sentences.
> Did you have a look at my suggestion in v10 patch 1?

Yes, I had, and sorry, I forgot to report here inside the patch 11.
I will take it, thanks for it.

> 
>> +    bit = core_id;
>> +    origin = bit / 64;
>> +    bit %= 64;
>> +    bit = 63 - bit;
>> +
>> +    topo->socket[socket_id].active_count++;
>> +    set_bit(bit, &topo->socket[socket_id].mask[origin]);
>> +}
>> +
>> +/*
>> + * s390_prepare_topology:
>> + * @s390ms : pointer to the S390CcwMachite State
>> + *
>> + * Calls s390_topology_add_cpu to organize the topology
>> + * inside the topology device before writing the SYSIB.
>> + *
>> + * The topology is currently fixed on boot and do not change
> 
> does not change

yes, thanks

regards,
Pierre

-- 
Pierre Morel
IBM Lab Boeblingen