lkml.org 
[lkml]   [2014]   [May]   [16]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH 6/6] perf: Add dcacheline sort
From
On Tue, May 13, 2014 at 6:48 PM, Don Zickus <dzickus@redhat.com> wrote:
> In perf's 'mem-mode', one can get access to a whole bunch of details specific to a
> particular sample instruction. A bunch of those details relate to the data
> address.
>
> One interesting thing you can do with data addresses is to convert them into a unique
> cacheline they belong too. Organizing these data cachelines into similar groups and sorting
> them can reveal cache contention.
>
> This patch creates an alogorithm based on various sample details that can help group
> entries together into data cachelines and allows 'perf report' to sort on it.
>
> The algorithm relies on having proper mmap2 support in the kernel to help determine
> if the memory map the data address belongs to is private to a pid or globally shared.
>
> The alogortithm is as follows:
>
> o group cpumodes together
> o group entries with discovered maps together
> o sort on major, minor, inode and inode generation numbers
> o if userspace anon, then sort on pid
> o sort on cachelines based on data addresses
>
> The 'dcacheline' sort option in 'perf report' only works in 'mem-mode'.
>
> Signed-off-by: Don Zickus <dzickus@redhat.com>
> ---
> tools/perf/Documentation/perf-report.txt | 3 +-
> tools/perf/builtin-report.c | 2 +-
> tools/perf/util/hist.h | 1 +
> tools/perf/util/sort.c | 105 +++++++++++++++++++++++++++++++
> tools/perf/util/sort.h | 1 +
> 5 files changed, 110 insertions(+), 2 deletions(-)
>
> diff --git a/tools/perf/Documentation/perf-report.txt b/tools/perf/Documentation/perf-report.txt
> index bea1c83..9115d02 100644
> --- a/tools/perf/Documentation/perf-report.txt
> +++ b/tools/perf/Documentation/perf-report.txt
> @@ -100,7 +100,7 @@ OPTIONS
>
> If --mem-mode option is used, following sort keys are also available
> (incompatible with --branch-stack):
> - symbol_daddr, dso_daddr, locked, tlb, mem, snoop.
> + symbol_daddr, dso_daddr, locked, tlb, mem, snoop, dcacheline.
>
> - symbol_daddr: name of data symbol being executed on at the time of sample
> - dso_daddr: name of library or module containing the data being executed
> @@ -109,6 +109,7 @@ OPTIONS
> - tlb: type of tlb access for the data at the time of sample
> - mem: type of memory access for the data at the time of sample
> - snoop: type of snoop (if any) for the data at the time of sample
> + - dcacheline: the cacheline the data address is on at the time of sample
>
> And default sort keys are changed to local_weight, mem, sym, dso,
> symbol_daddr, dso_daddr, snoop, tlb, locked, see '--mem-mode'.
> diff --git a/tools/perf/builtin-report.c b/tools/perf/builtin-report.c
> index 89c9528..387459f 100644
> --- a/tools/perf/builtin-report.c
> +++ b/tools/perf/builtin-report.c
> @@ -702,7 +702,7 @@ int cmd_report(int argc, const char **argv, const char *prefix __maybe_unused)
> "sort by key(s): pid, comm, dso, symbol, parent, cpu, srcline,"
> " dso_to, dso_from, symbol_to, symbol_from, mispredict,"
> " weight, local_weight, mem, symbol_daddr, dso_daddr, tlb, "
> - "snoop, locked, abort, in_tx, transaction"),
> + "snoop, locked, abort, in_tx, transaction, dcacheline"),
> OPT_BOOLEAN(0, "showcpuutilization", &symbol_conf.show_cpu_utilization,
> "Show sample percentage for different cpu modes"),
> OPT_STRING('p', "parent", &parent_pattern, "regex",
> diff --git a/tools/perf/util/hist.h b/tools/perf/util/hist.h
> index 38c3e87..14d1dc1 100644
> --- a/tools/perf/util/hist.h
> +++ b/tools/perf/util/hist.h
> @@ -72,6 +72,7 @@ enum hist_column {
> HISTC_MEM_TLB,
> HISTC_MEM_LVL,
> HISTC_MEM_SNOOP,
> + HISTC_MEM_DCACHELINE,
> HISTC_TRANSACTION,
> HISTC_NR_COLS, /* Last entry */
> };
> diff --git a/tools/perf/util/sort.c b/tools/perf/util/sort.c
> index 635cd8f..0e91ba9 100644
> --- a/tools/perf/util/sort.c
> +++ b/tools/perf/util/sort.c
> @@ -1,3 +1,4 @@
> +#include <sys/mman.h>
> #include "sort.h"
> #include "hist.h"
> #include "comm.h"
> @@ -764,6 +765,102 @@ static int hist_entry__snoop_snprintf(struct hist_entry *he, char *bf,
> return repsep_snprintf(bf, size, "%-*s", width, out);
> }
>
> +#define CACHE_LINESIZE 64
I had something similar to your patch here in my original series for
perf mem, but I never pushed it.
I think this is a useful feature to have.
However, I don't think you can hardcode the cache line size to 64.
This is generic
code. There may be architectures where the line size is different from 64.
So I think you should add an option to change the default line size or provide
an arch-specific definition.

> +#define CLINE_OFFSET_MSK (CACHE_LINESIZE - 1)
> +#define CLADRS(a) ((a) & ~(CLINE_OFFSET_MSK))
> +
> +static int64_t
> +sort__dcacheline_cmp(struct hist_entry *left, struct hist_entry *right)
> +{
> + u64 l, r;
> + struct map *l_map, *r_map;
> +
> + if (!left->mem_info) return -1;
> + if (!right->mem_info) return 1;
> +
> + /* group event types together */
> + if (left->cpumode > right->cpumode) return -1;
> + if (left->cpumode < right->cpumode) return 1;
> +
> + l_map = left->mem_info->daddr.map;
> + r_map = right->mem_info->daddr.map;
> +
> + /* if both are NULL, jump to sort on al_addr instead */
> + if (!l_map && !r_map)
> + goto addr;
> +
> + if (!l_map) return -1;
> + if (!r_map) return 1;
> +
> + if (l_map->maj > r_map->maj) return -1;
> + if (l_map->maj < r_map->maj) return 1;
> +
> + if (l_map->min > r_map->min) return -1;
> + if (l_map->min < r_map->min) return 1;
> +
> + if (l_map->ino > r_map->ino) return -1;
> + if (l_map->ino < r_map->ino) return 1;
> +
> + if (l_map->ino_generation > r_map->ino_generation) return -1;
> + if (l_map->ino_generation < r_map->ino_generation) return 1;
> +
> + /*
> + * Addresses with no major/minor numbers are assumed to be
> + * anonymous in userspace. Sort those on pid then address.
> + *
> + * The kernel and non-zero major/minor mapped areas are
> + * assumed to be unity mapped. Sort those on address.
> + */
> +
> + if ((left->cpumode != PERF_RECORD_MISC_KERNEL) &&
> + (!(l_map->flags & MAP_SHARED)) &&
> + !l_map->maj && !l_map->min && !l_map->ino &&
> + !l_map->ino_generation) {
> + /* userspace anonymous */
> +
> + if (left->thread->pid_ > right->thread->pid_) return -1;
> + if (left->thread->pid_ < right->thread->pid_) return 1;
> + }
> +
> +addr:
> + /* al_addr does all the right addr - start + offset calculations */
> + l = CLADRS(left->mem_info->daddr.al_addr);
> + r = CLADRS(right->mem_info->daddr.al_addr);
> +
> + if (l > r) return -1;
> + if (l < r) return 1;
> +
> + return 0;
> +}
> +
> +static int hist_entry__dcacheline_snprintf(struct hist_entry *he, char *bf,
> + size_t size, unsigned int width)
> +{
> +
> + uint64_t addr = 0;
> + struct map *map = NULL;
> + struct symbol *sym = NULL;
> + char level = he->level;
> +
> + if (he->mem_info) {
> + addr = CLADRS(he->mem_info->daddr.al_addr);
> + map = he->mem_info->daddr.map;
> + sym = he->mem_info->daddr.sym;
> +
> + /* print [s] for shared data mmaps */
> + if ((he->cpumode != PERF_RECORD_MISC_KERNEL) &&
> + map && (map->type == MAP__VARIABLE) &&
> + (map->flags & MAP_SHARED) &&
> + (map->maj || map->min || map->ino ||
> + map->ino_generation))
> + level = 's';
> + else if (!map)
> + level = 'X';
> + }
> + return _hist_entry__sym_snprintf(map, sym, addr, level, bf, size,
> + width);
> +}
> +
> struct sort_entry sort_mispredict = {
> .se_header = "Branch Mispredicted",
> .se_cmp = sort__mispredict_cmp,
> @@ -856,6 +953,13 @@ struct sort_entry sort_mem_snoop = {
> .se_width_idx = HISTC_MEM_SNOOP,
> };
>
> +struct sort_entry sort_mem_dcacheline = {
> + .se_header = "Data Cacheline ",
> + .se_cmp = sort__dcacheline_cmp,
> + .se_snprintf = hist_entry__dcacheline_snprintf,
> + .se_width_idx = HISTC_MEM_DCACHELINE,
> +};
> +
> static int64_t
> sort__abort_cmp(struct hist_entry *left, struct hist_entry *right)
> {
> @@ -1023,6 +1127,7 @@ static struct sort_dimension memory_sort_dimensions[] = {
> DIM(SORT_MEM_TLB, "tlb", sort_mem_tlb),
> DIM(SORT_MEM_LVL, "mem", sort_mem_lvl),
> DIM(SORT_MEM_SNOOP, "snoop", sort_mem_snoop),
> + DIM(SORT_MEM_DCACHELINE, "dcacheline", sort_mem_dcacheline),
> };
>
> #undef DIM
> diff --git a/tools/perf/util/sort.h b/tools/perf/util/sort.h
> index 22cf912..460ec9c 100644
> --- a/tools/perf/util/sort.h
> +++ b/tools/perf/util/sort.h
> @@ -167,6 +167,7 @@ enum sort_type {
> SORT_MEM_TLB,
> SORT_MEM_LVL,
> SORT_MEM_SNOOP,
> + SORT_MEM_DCACHELINE,
> };
>
> /*
> --
> 1.7.11.7
>


\
 
 \ /
  Last update: 2014-05-16 16:41    [W:0.263 / U:0.540 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site