From: Igor Mammedov <imammedo@redhat.com>
To: Tao Xu <tao3.xu@intel.com>
Cc: eblake@redhat.com, mst@redhat.com, xiaoguangrong.eric@gmail.com,
pbonzini@redhat.com, rth@twiddle.net, ehabkost@redhat.com,
karl.heubaum@oracle.com, jingqi.liu@intel.com,
danmei.wei@intel.com, qemu-devel@nongnu.org
Subject: Re: [Qemu-devel] [PATCH v3 2/8] hmat acpi: Build System Locality Latency and Bandwidth Information Structure(s) in ACPI HMAT
Date: Wed, 6 Feb 2019 10:17:02 +0100 [thread overview]
Message-ID: <20190206101702.3cf834bc@Igors-MacBook-Pro.local> (raw)
In-Reply-To: <20190131071658.29120-3-tao3.xu@intel.com>
On Thu, 31 Jan 2019 15:16:52 +0800
Tao Xu <tao3.xu@intel.com> wrote:
> From: Liu Jingqi <jingqi.liu@intel.com>
>
> This structure describes the memory access latency and bandwidth
> information from various memory access initiator proximity domains.
> The latency and bandwidth numbers represented in this structure
> correspond to rated latency and bandwidth for the platform.
> The software could use this information as hint for optimization.
>
> Signed-off-by: Liu Jingqi <jingqi.liu@intel.com>
> Signed-off-by: Tao Xu <tao3.xu@intel.com>
> ---
> hw/acpi/hmat.c | 98 ++++++++++++++++++++++++++++++++++++++++++++++++++
> hw/acpi/hmat.h | 76 +++++++++++++++++++++++++++++++++++++++
> 2 files changed, 174 insertions(+)
>
> diff --git a/hw/acpi/hmat.c b/hw/acpi/hmat.c
> index 7e0fc0a9ae..e3deeaa36b 100644
> --- a/hw/acpi/hmat.c
> +++ b/hw/acpi/hmat.c
> @@ -29,6 +29,11 @@
> #include "hw/acpi/hmat.h"
> #include "hw/nvram/fw_cfg.h"
>
> +struct numa_hmat_lb_info *hmat_lb_info[HMAT_LB_LEVELS][HMAT_LB_TYPES] = {0};
Another global,
it's not acceptable to add new globals unless you have a very good reason to do so.
So try to get by without using them.
> +
> +static uint32_t initiator_pxm[MAX_NODES], target_pxm[MAX_NODES];
> +static uint32_t num_initiator, num_target;
> +
> /* Build Memory Subsystem Address Range Structure */
> static void build_hmat_spa(GArray *table_data,
> uint64_t base, uint64_t length, int node)
> @@ -110,10 +115,103 @@ static void hmat_build_spa(GArray *table_data, PCMachineState *pcms)
> }
> }
>
> +static void classify_proximity_domains(void)
> +{
> + int node;
> +
> + for (node = 0; node < nb_numa_nodes; node++) {
> + if (numa_info[node].is_initiator) {
> + initiator_pxm[num_initiator++] = node;
> + }
> + if (numa_info[node].is_target) {
> + target_pxm[num_target++] = node;
> + }
> + }
> +}
> +
> +static void hmat_build_lb(GArray *table_data)
> +{
> + AcpiHmatLBInfo *hmat_lb;
> + struct numa_hmat_lb_info *numa_hmat_lb;
> + int i, j, hrchy, type;
> +
> + if (!num_initiator && !num_target) {
> + classify_proximity_domains();
> + }
> +
> + for (hrchy = HMAT_LB_MEM_MEMORY;
> + hrchy <= HMAT_LB_MEM_CACHE_3RD_LEVEL; hrchy++) {
> + for (type = HMAT_LB_DATA_ACCESS_LATENCY;
> + type <= HMAT_LB_DATA_WRITE_BANDWIDTH; type++) {
> + numa_hmat_lb = hmat_lb_info[hrchy][type];
> +
> + if (numa_hmat_lb) {
> + uint64_t start;
> + uint32_t *list_entry;
> + uint16_t *entry, *entry_start;
> + uint32_t size;
> + uint8_t m, n;
> +
> + start = table_data->len;
> + hmat_lb = acpi_data_push(table_data, sizeof(*hmat_lb));
> +
> + hmat_lb->type = cpu_to_le16(ACPI_HMAT_LB_INFO);
> + hmat_lb->flags = numa_hmat_lb->hierarchy;
> + hmat_lb->data_type = numa_hmat_lb->data_type;
> + hmat_lb->num_initiator = cpu_to_le32(num_initiator);
> + hmat_lb->num_target = cpu_to_le32(num_target);
use build_append_int_noprefix() like in previous patch to build parts of ACPI table
and drop all packed structures.
> +
> + if (type <= HMAT_LB_DATA_WRITE_LATENCY) {
> + hmat_lb->base_unit = cpu_to_le32(numa_hmat_lb->base_lat);
> + } else {
> + hmat_lb->base_unit = cpu_to_le32(numa_hmat_lb->base_bw);
> + }
> + if (!hmat_lb->base_unit) {
> + hmat_lb->base_unit = cpu_to_le32(1);
> + }
> +
> + /* the initiator proximity domain list */
> + for (i = 0; i < num_initiator; i++) {
> + list_entry = acpi_data_push(table_data, sizeof(uint32_t));
> + *list_entry = cpu_to_le32(initiator_pxm[i]);
> + }
> +
> + /* the target proximity domain list */
> + for (i = 0; i < num_target; i++) {
> + list_entry = acpi_data_push(table_data, sizeof(uint32_t));
> + *list_entry = cpu_to_le32(target_pxm[i]);
> + }
> +
> + /* latency or bandwidth entries */
> + size = sizeof(uint16_t) * num_initiator * num_target;
> + entry_start = acpi_data_push(table_data, size);
> +
> + for (i = 0; i < num_initiator; i++) {
> + m = initiator_pxm[i];
> + for (j = 0; j < num_target; j++) {
> + n = target_pxm[j];
> + entry = entry_start + i * num_target + j;
> + if (type <= HMAT_LB_DATA_WRITE_LATENCY) {
> + *entry = cpu_to_le16(numa_hmat_lb->latency[m][n]);
> + } else {
> + *entry = cpu_to_le16(numa_hmat_lb->bandwidth[m][n]);
> + }
> + }
> + }
> + hmat_lb = (AcpiHmatLBInfo *)(table_data->data + start);
> + hmat_lb->length = cpu_to_le16(table_data->len - start);
> + }
> + }
> + }
> +}
> +
> static void hmat_build_hma(GArray *hma, PCMachineState *pcms)
> {
> /* Build HMAT Memory Subsystem Address Range. */
> hmat_build_spa(hma, pcms);
> +
> + /* Build HMAT System Locality Latency and Bandwidth Information. */
> + hmat_build_lb(hma);
> }
>
> void hmat_build_acpi(GArray *table_data, BIOSLinker *linker,
> diff --git a/hw/acpi/hmat.h b/hw/acpi/hmat.h
> index f216e658c4..ffef9f6243 100644
> --- a/hw/acpi/hmat.h
> +++ b/hw/acpi/hmat.h
> @@ -32,6 +32,7 @@
> #include "hw/acpi/aml-build.h"
>
> #define ACPI_HMAT_SPA 0
> +#define ACPI_HMAT_LB_INFO 1
>
> /* ACPI HMAT sub-structure header */
> #define ACPI_HMAT_SUB_HEADER_DEF \
> @@ -46,6 +47,81 @@ enum {
> HMAT_SPA_RESERVATION_HINT = 0x4,
> };
>
> +/* the value of AcpiHmatLBInfo flags */
> +enum {
> + HMAT_LB_MEM_MEMORY = 0,
> + HMAT_LB_MEM_CACHE_LAST_LEVEL = 1,
> + HMAT_LB_MEM_CACHE_1ST_LEVEL = 2,
> + HMAT_LB_MEM_CACHE_2ND_LEVEL = 3,
> + HMAT_LB_MEM_CACHE_3RD_LEVEL = 4,
> +};
> +
> +/* the value of AcpiHmatLBInfo data type */
> +enum {
> + HMAT_LB_DATA_ACCESS_LATENCY = 0,
> + HMAT_LB_DATA_READ_LATENCY = 1,
> + HMAT_LB_DATA_WRITE_LATENCY = 2,
> + HMAT_LB_DATA_ACCESS_BANDWIDTH = 3,
> + HMAT_LB_DATA_READ_BANDWIDTH = 4,
> + HMAT_LB_DATA_WRITE_BANDWIDTH = 5,
> +};
> +
> +#define HMAT_LB_LEVELS (HMAT_LB_MEM_CACHE_3RD_LEVEL + 1)
> +#define HMAT_LB_TYPES (HMAT_LB_DATA_WRITE_BANDWIDTH + 1)
> +
> +struct AcpiHmatLBInfo {
> + ACPI_HMAT_SUB_HEADER_DEF
> + uint8_t flags;
> + uint8_t data_type;
> + uint16_t reserved1;
> + uint32_t num_initiator;
> + uint32_t num_target;
> + uint32_t reserved2;
> + uint64_t base_unit;
> +} QEMU_PACKED;
> +typedef struct AcpiHmatLBInfo AcpiHmatLBInfo;
> +
> +struct numa_hmat_lb_info {
> + /*
> + * Indicates total number of Proximity Domains
> + * that can initiate memory access requests.
> + */
> + uint32_t num_initiator;
> + /*
> + * Indicates total number of Proximity Domains
> + * that can act as target.
> + */
> + uint32_t num_target;
> + /*
> + * Indicates it's memory or
> + * the specified level memory side cache.
> + */
> + uint8_t hierarchy;
> + /*
> + * Present the type of data,
> + * access/read/write latency or bandwidth.
> + */
> + uint8_t data_type;
> + /* The base unit for latency in nanoseconds. */
> + uint64_t base_lat;
> + /* The base unit for bandwidth in megabytes per second(MB/s). */
> + uint64_t base_bw;
> + /*
> + * latency[i][j]:
> + * Indicates the latency based on base_lat
> + * from Initiator Proximity Domain i to Target Proximity Domain j.
> + */
> + uint16_t latency[MAX_NODES][MAX_NODES];
> + /*
> + * bandwidth[i][j]:
> + * Indicates the bandwidth based on base_bw
> + * from Initiator Proximity Domain i to Target Proximity Domain j.
> + */
> + uint16_t bandwidth[MAX_NODES][MAX_NODES];
> +};
> +
> +extern struct numa_hmat_lb_info *hmat_lb_info[HMAT_LB_LEVELS][HMAT_LB_TYPES];
> +
> void hmat_build_acpi(GArray *table_data, BIOSLinker *linker,
> MachineState *machine);
>
next prev parent reply other threads:[~2019-02-06 9:17 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-01-31 7:16 [Qemu-devel] [PATCH v3 0/8] Build ACPI Heterogeneous Memory Attribute Table (HMAT) Tao Xu
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 1/8] hmat acpi: Build Memory Subsystem Address Range Structure(s) in ACPI HMAT Tao Xu
2019-02-06 9:05 ` Igor Mammedov
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 2/8] hmat acpi: Build System Locality Latency and Bandwidth Information " Tao Xu
2019-02-06 9:17 ` Igor Mammedov [this message]
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 3/8] hmat acpi: Build Memory Side Cache " Tao Xu
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 4/8] Extend the command-line to provide memory latency and bandwidth information Tao Xu
2019-02-06 10:11 ` Igor Mammedov
2019-04-02 4:42 ` Tao Xu
2019-04-02 9:46 ` Igor Mammedov
2019-04-02 19:23 ` Eduardo Habkost
2019-04-02 19:12 ` Eduardo Habkost
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 5/8] numa: Extend the command-line to provide memory side cache information Tao Xu
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 6/8] hmat acpi: Implement _HMA method to update HMAT at runtime Tao Xu
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 7/8] hmat acpi: move some function inside of the caller Tao Xu
2019-02-06 10:49 ` Igor Mammedov
2019-01-31 7:16 ` [Qemu-devel] [PATCH v3 8/8] acpi: rewrite the _FIT method to use it in _HMA method Tao Xu
2019-02-06 9:35 ` Igor Mammedov
2019-02-01 4:43 ` [Qemu-devel] [PATCH v3 0/8] Build ACPI Heterogeneous Memory Attribute Table (HMAT) no-reply
2019-02-06 10:51 ` Igor Mammedov
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=20190206101702.3cf834bc@Igors-MacBook-Pro.local \
--to=imammedo@redhat.com \
--cc=danmei.wei@intel.com \
--cc=eblake@redhat.com \
--cc=ehabkost@redhat.com \
--cc=jingqi.liu@intel.com \
--cc=karl.heubaum@oracle.com \
--cc=mst@redhat.com \
--cc=pbonzini@redhat.com \
--cc=qemu-devel@nongnu.org \
--cc=rth@twiddle.net \
--cc=tao3.xu@intel.com \
--cc=xiaoguangrong.eric@gmail.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;
as well as URLs for NNTP newsgroup(s).