qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Igor Mammedov <imammedo@redhat.com>
To: Babu Moger <babu.moger@amd.com>
Cc: qemu-devel@nongnu.org, pbonzini@redhat.com, mst@redhat.com,
	ehabkost@redhat.com, rth@twiddle.net
Subject: Re: [PATCH v3 1/3] i386: Simplify CPUID_8000_001E for AMD
Date: Fri, 7 Aug 2020 20:11:00 +0200	[thread overview]
Message-ID: <20200807201100.5c141fe4@redhat.com> (raw)
In-Reply-To: <159681796533.9679.8403552849052625336.stgit@naples-babu.amd.com>

On Fri, 07 Aug 2020 11:32:45 -0500
Babu Moger <babu.moger@amd.com> wrote:

> apic_id contains all the information required to build CPUID_8000_001E.
> Also remove the restriction on number bits on core_id and node_id.
> Remove all the hardcoded values and replace with generalized
> fields.
> 
> Refer the Processor Programming Reference (PPR) documentation
> available from the bugzilla Link below.
> 
> Signed-off-by: Babu Moger <babu.moger@amd.com>

Reviewed-by: Igor Mammedov <imammedo@redhat.com>

> ---
>  target/i386/cpu.c |   77 +++++++++++++++++++++++++----------------------------
>  1 file changed, 36 insertions(+), 41 deletions(-)
> 
> diff --git a/target/i386/cpu.c b/target/i386/cpu.c
> index 588f32e136..c892432cae 100644
> --- a/target/i386/cpu.c
> +++ b/target/i386/cpu.c
> @@ -388,57 +388,52 @@ static void encode_topo_cpuid8000001e(X86CPUTopoInfo *topo_info, X86CPU *cpu,
>  {
>      X86CPUTopoIDs topo_ids = {0};
>      unsigned long nodes = MAX(topo_info->nodes_per_pkg, 1);
> -    int shift;
>  
>      x86_topo_ids_from_apicid_epyc(cpu->apic_id, topo_info, &topo_ids);
>  
>      *eax = cpu->apic_id;
> +
>      /*
> +     * CPUID_Fn8000001E_EBX [Core Identifiers] (CoreId)
> +     * Read-only. Reset: 0000_XXXXh.
> +     * See Core::X86::Cpuid::ExtApicId.
> +     * Core::X86::Cpuid::CoreId_lthree[1:0]_core[3:0]_thread[1:0];
>       * CPUID_Fn8000001E_EBX
> -     * 31:16 Reserved
> -     * 15:8  Threads per core (The number of threads per core is
> -     *       Threads per core + 1)
> -     *  7:0  Core id (see bit decoding below)
> -     *       SMT:
> -     *           4:3 node id
> -     *             2 Core complex id
> -     *           1:0 Core id
> -     *       Non SMT:
> -     *           5:4 node id
> -     *             3 Core complex id
> -     *           1:0 Core id
> +     * Bits Description
> +     * 31:16 Reserved.
> +     * 15:8 ThreadsPerCore: threads per core. Read-only. Reset: XXh.
> +     *      The number of threads per core is ThreadsPerCore+1.
> +     *  7:0 CoreId: core ID. Read-only. Reset: XXh.
> +     *
> +     *  NOTE: CoreId is already part of apic_id. Just use it. We can
> +     *  use all the 8 bits to represent the core_id here.
>       */
> -    *ebx = ((topo_info->threads_per_core - 1) << 8) | (topo_ids.node_id << 3) |
> -            (topo_ids.core_id);
> +    *ebx = ((topo_info->threads_per_core - 1) << 8) | (topo_ids.core_id & 0xFF);
> +
>      /*
> +     * CPUID_Fn8000001E_ECX [Node Identifiers] (NodeId)
> +     * Read-only. Reset: 0000_0XXXh.
> +     * Core::X86::Cpuid::NodeId_lthree[1:0]_core[3:0]_thread[1:0];
>       * CPUID_Fn8000001E_ECX
> -     * 31:11 Reserved
> -     * 10:8  Nodes per processor (Nodes per processor is number of nodes + 1)
> -     *  7:0  Node id (see bit decoding below)
> -     *         2  Socket id
> -     *       1:0  Node id
> +     * Bits Description
> +     * 31:11 Reserved.
> +     * 10:8 NodesPerProcessor: Node per processor. Read-only. Reset: XXXb.
> +     *      ValidValues:
> +     *      Value Description
> +     *      000b  1 node per processor.
> +     *      001b  2 nodes per processor.
> +     *      010b Reserved.
> +     *      011b 4 nodes per processor.
> +     *      111b-100b Reserved.
> +     *  7:0 NodeId: Node ID. Read-only. Reset: XXh.
> +     *
> +     * NOTE: Hardware reserves 3 bits for number of nodes per processor.
> +     * But users can create more nodes than the actual hardware can
> +     * support. To genaralize we can use all the upper 8 bits for nodes.
> +     * NodeId is combination of node and socket_id which is already decoded
> +     * in apic_id. Just use it by shifting.
>       */
> -    if (nodes <= 4) {
> -        *ecx = ((nodes - 1) << 8) | (topo_ids.pkg_id << 2) | topo_ids.node_id;
> -    } else {
> -        /*
> -         * Node id fix up. Actual hardware supports up to 4 nodes. But with
> -         * more than 32 cores, we may end up with more than 4 nodes.
> -         * Node id is a combination of socket id and node id. Only requirement
> -         * here is that this number should be unique accross the system.
> -         * Shift the socket id to accommodate more nodes. We dont expect both
> -         * socket id and node id to be big number at the same time. This is not
> -         * an ideal config but we need to to support it. Max nodes we can have
> -         * is 32 (255/8) with 8 cores per node and 255 max cores. We only need
> -         * 5 bits for nodes. Find the left most set bit to represent the total
> -         * number of nodes. find_last_bit returns last set bit(0 based). Left
> -         * shift(+1) the socket id to represent all the nodes.
> -         */
> -        nodes -= 1;
> -        shift = find_last_bit(&nodes, 8);
> -        *ecx = (nodes << 8) | (topo_ids.pkg_id << (shift + 1)) |
> -               topo_ids.node_id;
> -    }
> +    *ecx = ((nodes - 1) << 8) | cpu->node_id;
>      *edx = 0;
>  }
>  
> 
> 



  reply	other threads:[~2020-08-07 18:12 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-08-07 16:32 [PATCH v3 0/3] Fix couple of issues with AMD topology Babu Moger
2020-08-07 16:32 ` [PATCH v3 1/3] i386: Simplify CPUID_8000_001E for AMD Babu Moger
2020-08-07 18:11   ` Igor Mammedov [this message]
2020-08-07 16:32 ` [PATCH v3 2/3] hw/i386: Add a new check to configure smp dies for EPYC Babu Moger
2020-08-07 16:52   ` Daniel P. Berrangé
2020-08-07 17:23     ` Babu Moger
2020-08-07 19:11     ` Igor Mammedov
2020-08-11 21:03       ` Babu Moger
2020-08-13 13:56         ` Igor Mammedov
2020-08-13 21:10           ` Babu Moger
2020-08-07 19:27   ` Igor Mammedov
2020-08-11 21:04     ` Babu Moger
2020-08-07 16:32 ` [PATCH v3 3/3] hw/i386: Remove node_id, nr_nodes and nodes_per_pkg from topology Babu Moger

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=20200807201100.5c141fe4@redhat.com \
    --to=imammedo@redhat.com \
    --cc=babu.moger@amd.com \
    --cc=ehabkost@redhat.com \
    --cc=mst@redhat.com \
    --cc=pbonzini@redhat.com \
    --cc=qemu-devel@nongnu.org \
    --cc=rth@twiddle.net \
    /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).