From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-9.3 required=3.0 tests=DKIM_ADSP_CUSTOM_MED, DKIM_INVALID,DKIM_SIGNED,FREEMAIL_FORGED_FROMDOMAIN,FREEMAIL_FROM, HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_HELO_NONE,SPF_PASS,USER_AGENT_GIT autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 8B8E3C433FF for ; Thu, 15 Aug 2019 02:44:29 +0000 (UTC) Received: from lists.gnu.org (lists.gnu.org [209.51.188.17]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPS id 453D2208C2 for ; Thu, 15 Aug 2019 02:44:29 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=fail reason="signature verification failed" (2048-bit key) header.d=gmail.com header.i=@gmail.com header.b="Ph8XWJkK" DMARC-Filter: OpenDMARC Filter v1.3.2 mail.kernel.org 453D2208C2 Authentication-Results: mail.kernel.org; dmarc=fail (p=none dis=none) header.from=gmail.com Authentication-Results: mail.kernel.org; spf=pass smtp.mailfrom=qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Received: from localhost ([::1]:37710 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1hy5ka-0006Sn-7P for qemu-devel@archiver.kernel.org; Wed, 14 Aug 2019 22:44:28 -0400 Received: from eggs.gnu.org ([2001:470:142:3::10]:48592) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1hy5MP-0000a5-W4 for qemu-devel@nongnu.org; Wed, 14 Aug 2019 22:19:31 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1hy5MO-0001bV-75 for qemu-devel@nongnu.org; Wed, 14 Aug 2019 22:19:29 -0400 Received: from mail-qt1-x843.google.com ([2607:f8b0:4864:20::843]:35415) by eggs.gnu.org with esmtps (TLS1.0:RSA_AES_128_CBC_SHA1:16) (Exim 4.71) (envelope-from ) id 1hy5MO-0001b8-2F for qemu-devel@nongnu.org; Wed, 14 Aug 2019 22:19:28 -0400 Received: by mail-qt1-x843.google.com with SMTP id u34so950783qte.2 for ; Wed, 14 Aug 2019 19:19:28 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=from:to:cc:subject:date:message-id:in-reply-to:references :mime-version:content-transfer-encoding; bh=vErop0rGwk60sa3Qa4x2OiK1fh1r4aRc9p7r0AqR3oI=; b=Ph8XWJkK+6L+jNeXmHhkWiUCXrKNQYGI/LDLm5spgCRQW7kHFusJs/qpisw7IUmnM4 LZu91l43c8MTepbeJTSq99ZQT3FdwxJp52kR3MiYKx8NKwtFHqo5pl5ipD2YwzmRflGK 1lDjMc6XiNRZsbMhppebY2BPyW4hkh3RgHPwxX2Xj2PM0f2xOisePOTBn+BmLhgL1wiy f9anGDcAJaTrStzC1rTQPD6jRetQGf7YxIohFeES6F1XAcGin+wIZ+IBs3TFDz6XI8wW Z25AbPWE6a+kwkFiAEHcVicyN8zcaqBkqevAFXFBUsYhn2e4e+aDhRZKXF902yeYjfqR wR+A== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:from:to:cc:subject:date:message-id:in-reply-to :references:mime-version:content-transfer-encoding; bh=vErop0rGwk60sa3Qa4x2OiK1fh1r4aRc9p7r0AqR3oI=; b=Mm3djhYIQNHad/BPv05i+pM+lWKMw1JSJsgvFbiZ4mVu+VzBLOEu0HG7rcAFTgm0pO LX+s7W+yovWvQAZLnNUhKGToD3L61i4221PY9PyKu/7oU1hBFyoUJNWZzRPBIwNTDSxX Tk7kIXGijNkfGfwOpcICrBCY9i65KMkRyYyV2Jy2NSvOP5MLy0jVAJYOvr8ex2ukDUp7 fl/rfXGvaUzBiUnE3C7LyYifw0fOlnVTwAhC7YaQcNAbo5fyh1ewAd1kKSMkVR1p+gFi u76IX8WQkxvC2EQ+0TJqcyqLWdOo6ggxt3q367l9egqpNbmJkLyXoLKGsYx9knMwUaGM DGlw== X-Gm-Message-State: APjAAAXXqbRVghzrzhrHhDJLiWc0XDFYw8yTZYlkucb2QhIZPN+eGArM GXIYRHncqQC21YDBU30a74V2gGCZne0= X-Google-Smtp-Source: APXvYqxQhm32tXzVXJOo13OJKMd53oCQN4XSQOThEopKA2oS+nKUmkhvIO1oG+AAAaG/NiBw57whng== X-Received: by 2002:ac8:764f:: with SMTP id i15mr2039555qtr.194.1565835567208; Wed, 14 Aug 2019 19:19:27 -0700 (PDT) Received: from localhost.localdomain ([2804:14c:482:121::1]) by smtp.googlemail.com with ESMTPSA id o5sm757943qkf.10.2019.08.14.19.19.25 (version=TLS1_3 cipher=AEAD-AES256-GCM-SHA384 bits=256/256); Wed, 14 Aug 2019 19:19:26 -0700 (PDT) From: vandersonmr To: qemu-devel@nongnu.org Date: Wed, 14 Aug 2019 23:18:48 -0300 Message-Id: <20190815021857.19526-2-vandersonmr2@gmail.com> X-Mailer: git-send-email 2.22.0 In-Reply-To: <20190815021857.19526-1-vandersonmr2@gmail.com> References: <20190815021857.19526-1-vandersonmr2@gmail.com> MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-detected-operating-system: by eggs.gnu.org: Genre and OS details not recognized. X-Received-From: 2607:f8b0:4864:20::843 Subject: [Qemu-devel] [PATCH v5 01/10] accel: introducing TBStatistics structure X-BeenThere: qemu-devel@nongnu.org X-Mailman-Version: 2.1.23 Precedence: list List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Cc: Paolo Bonzini , vandersonmr , Richard Henderson Errors-To: qemu-devel-bounces+qemu-devel=archiver.kernel.org@nongnu.org Sender: "Qemu-devel" To store statistics for each TB, we created a TBStatistics structure which is linked with the TBs. TBStatistics can stay alive after tb_flush and be relinked to a regenerated TB. So the statistics can be accumulated even through flushes. The goal is to have all present and future qemu/tcg statistics and meta-data stored in this new structure. Signed-off-by: Vanderson M. do Rosario --- accel/tcg/Makefile.objs | 2 +- accel/tcg/perf/Makefile.objs | 1 + accel/tcg/tb-stats.c | 39 ++++++++++++++++++++++++ accel/tcg/translate-all.c | 57 ++++++++++++++++++++++++++++++++++++ include/exec/exec-all.h | 15 +++------- include/exec/tb-context.h | 12 ++++++++ include/exec/tb-hash.h | 7 +++++ include/exec/tb-stats.h | 43 +++++++++++++++++++++++++++ util/log.c | 2 ++ 9 files changed, 166 insertions(+), 12 deletions(-) create mode 100644 accel/tcg/perf/Makefile.objs create mode 100644 accel/tcg/tb-stats.c create mode 100644 include/exec/tb-stats.h diff --git a/accel/tcg/Makefile.objs b/accel/tcg/Makefile.objs index d381a02f34..49ffe81b5d 100644 --- a/accel/tcg/Makefile.objs +++ b/accel/tcg/Makefile.objs @@ -2,7 +2,7 @@ obj-$(CONFIG_SOFTMMU) += tcg-all.o obj-$(CONFIG_SOFTMMU) += cputlb.o obj-y += tcg-runtime.o tcg-runtime-gvec.o obj-y += cpu-exec.o cpu-exec-common.o translate-all.o -obj-y += translator.o +obj-y += translator.o tb-stats.o obj-$(CONFIG_USER_ONLY) += user-exec.o obj-$(call lnot,$(CONFIG_SOFTMMU)) += user-exec-stub.o diff --git a/accel/tcg/perf/Makefile.objs b/accel/tcg/perf/Makefile.objs new file mode 100644 index 0000000000..f82fba35e5 --- /dev/null +++ b/accel/tcg/perf/Makefile.objs @@ -0,0 +1 @@ +obj-y += jitdump.o diff --git a/accel/tcg/tb-stats.c b/accel/tcg/tb-stats.c new file mode 100644 index 0000000000..02844717cb --- /dev/null +++ b/accel/tcg/tb-stats.c @@ -0,0 +1,39 @@ +#include "qemu/osdep.h" + +#include "disas/disas.h" + +#include "exec/tb-stats.h" + +void init_tb_stats_htable_if_not(void) +{ + if (tb_stats_collection_enabled() && !tb_ctx.tb_stats.map) { + qht_init(&tb_ctx.tb_stats, tb_stats_cmp, + CODE_GEN_HTABLE_SIZE, QHT_MODE_AUTO_RESIZE); + } +} + +void enable_collect_tb_stats(void) +{ + init_tb_stats_htable_if_not(); + tcg_collect_tb_stats = TB_STATS_RUNNING; +} + +void disable_collect_tb_stats(void) +{ + tcg_collect_tb_stats = TB_STATS_PAUSED; +} + +void pause_collect_tb_stats(void) +{ + tcg_collect_tb_stats = TB_STATS_STOPPED; +} + +bool tb_stats_collection_enabled(void) +{ + return tcg_collect_tb_stats == TB_STATS_RUNNING; +} + +bool tb_stats_collection_paused(void) +{ + return tcg_collect_tb_stats == TB_STATS_PAUSED; +} diff --git a/accel/tcg/translate-all.c b/accel/tcg/translate-all.c index 5d1e08b169..b7bccacd3b 100644 --- a/accel/tcg/translate-all.c +++ b/accel/tcg/translate-all.c @@ -1118,6 +1118,23 @@ static inline void code_gen_alloc(size_t tb_size) } } +/* + * This is the more or less the same compare as tb_cmp(), but the + * data persists over tb_flush. We also aggregate the various + * variations of cflags under one record and ignore the details of + * page overlap (although we can count it). + */ +bool tb_stats_cmp(const void *ap, const void *bp) +{ + const TBStatistics *a = ap; + const TBStatistics *b = bp; + + return a->phys_pc == b->phys_pc && + a->pc == b->pc && + a->cs_base == b->cs_base && + a->flags == b->flags; +} + static bool tb_cmp(const void *ap, const void *bp) { const TranslationBlock *a = ap; @@ -1137,6 +1154,7 @@ static void tb_htable_init(void) unsigned int mode = QHT_MODE_AUTO_RESIZE; qht_init(&tb_ctx.htable, tb_cmp, CODE_GEN_HTABLE_SIZE, mode); + init_tb_stats_htable_if_not(); } /* Must be called before using the QEMU cpus. 'tb_size' is the size @@ -1666,6 +1684,34 @@ tb_link_page(TranslationBlock *tb, tb_page_addr_t phys_pc, return tb; } +static TBStatistics *tb_get_stats(tb_page_addr_t phys_pc, target_ulong pc, + target_ulong cs_base, uint32_t flags, + TranslationBlock *current_tb) +{ + TBStatistics *new_stats = g_new0(TBStatistics, 1); + uint32_t hash = tb_stats_hash_func(phys_pc, pc, flags); + void *existing_stats = NULL; + new_stats->phys_pc = phys_pc; + new_stats->pc = pc; + new_stats->cs_base = cs_base; + new_stats->flags = flags; + new_stats->tb = current_tb; + + qht_insert(&tb_ctx.tb_stats, new_stats, hash, &existing_stats); + + if (unlikely(existing_stats)) { + /* + * If there is already a TBStatistic for this TB from a previous flush + * then just make the new TB point to the older TBStatistic + */ + g_free(new_stats); + return existing_stats; + } else { + return new_stats; + } +} + + /* Called with mmap_lock held for user mode emulation. */ TranslationBlock *tb_gen_code(CPUState *cpu, target_ulong pc, target_ulong cs_base, @@ -1732,6 +1778,17 @@ TranslationBlock *tb_gen_code(CPUState *cpu, ti = profile_getclock(); #endif + /* + * We want to fetch the stats structure before we start code + * generation so we can count interesting things about this + * generation. + */ + if (tb_stats_collection_enabled()) { + tb->tb_stats = tb_get_stats(phys_pc, pc, cs_base, flags, tb); + } else { + tb->tb_stats = NULL; + } + tcg_func_start(tcg_ctx); tcg_ctx->cpu = env_cpu(env); diff --git a/include/exec/exec-all.h b/include/exec/exec-all.h index 16034ee651..24bd6a0a0c 100644 --- a/include/exec/exec-all.h +++ b/include/exec/exec-all.h @@ -22,21 +22,11 @@ #include "exec/tb-context.h" #include "sysemu/cpus.h" +#include "exec/tb-stats.h" /* allow to see translation results - the slowdown should be negligible, so we leave it */ #define DEBUG_DISAS -/* Page tracking code uses ram addresses in system mode, and virtual - addresses in userspace mode. Define tb_page_addr_t to be an appropriate - type. */ -#if defined(CONFIG_USER_ONLY) -typedef abi_ulong tb_page_addr_t; -#define TB_PAGE_ADDR_FMT TARGET_ABI_FMT_lx -#else -typedef ram_addr_t tb_page_addr_t; -#define TB_PAGE_ADDR_FMT RAM_ADDR_FMT -#endif - #include "qemu/log.h" void gen_intermediate_code(CPUState *cpu, TranslationBlock *tb, int max_insns); @@ -403,6 +393,9 @@ struct TranslationBlock { uintptr_t jmp_list_head; uintptr_t jmp_list_next[2]; uintptr_t jmp_dest[2]; + + /* Pointer to a struct where statistics from the TB is stored */ + TBStatistics *tb_stats; }; extern bool parallel_cpus; diff --git a/include/exec/tb-context.h b/include/exec/tb-context.h index feb585e0a7..3cfb62a338 100644 --- a/include/exec/tb-context.h +++ b/include/exec/tb-context.h @@ -23,6 +23,17 @@ #include "qemu/thread.h" #include "qemu/qht.h" +/* Page tracking code uses ram addresses in system mode, and virtual + addresses in userspace mode. Define tb_page_addr_t to be an appropriate + type. */ +#if defined(CONFIG_USER_ONLY) +typedef abi_ulong tb_page_addr_t; +#define TB_PAGE_ADDR_FMT TARGET_ABI_FMT_lx +#else +typedef ram_addr_t tb_page_addr_t; +#define TB_PAGE_ADDR_FMT RAM_ADDR_FMT +#endif + #define CODE_GEN_HTABLE_BITS 15 #define CODE_GEN_HTABLE_SIZE (1 << CODE_GEN_HTABLE_BITS) @@ -35,6 +46,7 @@ struct TBContext { /* statistics */ unsigned tb_flush_count; + struct qht tb_stats; }; extern TBContext tb_ctx; diff --git a/include/exec/tb-hash.h b/include/exec/tb-hash.h index 4f3a37d927..54c477fe79 100644 --- a/include/exec/tb-hash.h +++ b/include/exec/tb-hash.h @@ -64,4 +64,11 @@ uint32_t tb_hash_func(tb_page_addr_t phys_pc, target_ulong pc, uint32_t flags, return qemu_xxhash7(phys_pc, pc, flags, cf_mask, trace_vcpu_dstate); } +static inline +uint32_t tb_stats_hash_func(tb_page_addr_t phys_pc, target_ulong pc, + uint32_t flags) +{ + return qemu_xxhash5(phys_pc, pc, flags); +} + #endif diff --git a/include/exec/tb-stats.h b/include/exec/tb-stats.h new file mode 100644 index 0000000000..cc8f8a6ce6 --- /dev/null +++ b/include/exec/tb-stats.h @@ -0,0 +1,43 @@ +#ifndef TB_STATS_H + +#define TB_STATS_H + +#include "exec/cpu-common.h" +#include "exec/tb-context.h" +#include "tcg.h" + +typedef struct TBStatistics TBStatistics; + +/* + * This struct stores statistics such as execution count of the + * TranslationBlocks. Each sets of TBs for a given phys_pc/pc/flags + * has its own TBStatistics which will persist over tb_flush. + * + * We include additional counters to track number of translations as + * well as variants for compile flags. + */ +struct TBStatistics { + tb_page_addr_t phys_pc; + target_ulong pc; + uint32_t flags; + /* cs_base isn't included in the hash but we do check for matches */ + target_ulong cs_base; + /* current TB linked to this TBStatistics */ + TranslationBlock *tb; +}; + +bool tb_stats_cmp(const void *ap, const void *bp); + +void init_tb_stats_htable_if_not(void); + +/* TBStatistic collection controls */ +enum TBStatsStatus { TB_STATS_RUNNING, TB_STATS_PAUSED, TB_STATS_STOPPED }; +extern int tcg_collect_tb_stats; + +void enable_collect_tb_stats(void); +void disable_collect_tb_stats(void); +void pause_collect_tb_stats(void); +bool tb_stats_collection_enabled(void); +bool tb_stats_collection_paused(void); + +#endif diff --git a/util/log.c b/util/log.c index 1d1b33f7d9..393a17115b 100644 --- a/util/log.c +++ b/util/log.c @@ -31,6 +31,8 @@ int qemu_loglevel; static int log_append = 0; static GArray *debug_regions; +int tcg_collect_tb_stats; + /* Return the number of characters emitted. */ int qemu_log(const char *fmt, ...) { -- 2.22.0