From: "Ævar Arnfjörð Bjarmason" <avarab@gmail.com>
To: Jeff Hostetler via GitGitGadget <gitgitgadget@gmail.com>
Cc: git@vger.kernel.org, peff@peff.net, jrnieder@gmail.com,
steadmon@google.com, Junio C Hamano <gitster@pobox.com>,
Jeff Hostetler <jeffhost@microsoft.com>
Subject: Re: [PATCH v2 5/7] trace2: report peak memory usage of the process
Date: Fri, 29 Mar 2019 23:16:21 +0100 [thread overview]
Message-ID: <87k1gh9vii.fsf@evledraar.gmail.com> (raw)
In-Reply-To: <86d8fe28cf2f51a494163f9cec3e3b0541288435.1553879063.git.gitgitgadget@gmail.com>
On Fri, Mar 29 2019, Jeff Hostetler via GitGitGadget wrote:
> From: Jeff Hostetler <jeffhost@microsoft.com>
>
> Teach Windows version of git to report peak memory usage
> during exit() processing.
>
> Signed-off-by: Jeff Hostetler <jeffhost@microsoft.com>
> ---
> common-main.c | 2 +-
> compat/win32/trace2_win32_process_info.c | 50 ++++++++++++++++++++++--
> trace2.c | 2 +
> trace2.h | 14 +++++--
> 4 files changed, 60 insertions(+), 8 deletions(-)
>
> diff --git a/common-main.c b/common-main.c
> index 299ca62a72..582a7b1886 100644
> --- a/common-main.c
> +++ b/common-main.c
> @@ -41,7 +41,7 @@ int main(int argc, const char **argv)
>
> trace2_initialize();
> trace2_cmd_start(argv);
> - trace2_collect_process_info();
> + trace2_collect_process_info(TRACE2_PROCESS_INFO_STARTUP);
>
> git_setup_gettext();
>
> diff --git a/compat/win32/trace2_win32_process_info.c b/compat/win32/trace2_win32_process_info.c
> index 52bd62034b..2a514caed9 100644
> --- a/compat/win32/trace2_win32_process_info.c
> +++ b/compat/win32/trace2_win32_process_info.c
> @@ -1,5 +1,6 @@
> #include "../../cache.h"
> #include "../../json-writer.h"
> +#include "lazyload.h"
> #include <Psapi.h>
> #include <tlHelp32.h>
>
> @@ -137,11 +138,54 @@ static void get_is_being_debugged(void)
> "windows/debugger_present", 1);
> }
>
> -void trace2_collect_process_info(void)
> +/*
> + * Emit JSON data with the peak memory usage of the current process.
> + */
> +static void get_peak_memory_info(void)
> +{
> + DECLARE_PROC_ADDR(psapi.dll, BOOL, GetProcessMemoryInfo,
> + HANDLE, PPROCESS_MEMORY_COUNTERS, DWORD);
> +
> + if (INIT_PROC_ADDR(GetProcessMemoryInfo)) {
> + PROCESS_MEMORY_COUNTERS pmc;
> +
> + if (GetProcessMemoryInfo(GetCurrentProcess(), &pmc,
> + sizeof(pmc))) {
> + struct json_writer jw = JSON_WRITER_INIT;
> +
> + jw_object_begin(&jw, 0);
> +
> +#define KV(kv) #kv, (intmax_t)pmc.kv
> +
> + jw_object_intmax(&jw, KV(PageFaultCount));
> + jw_object_intmax(&jw, KV(PeakWorkingSetSize));
> + jw_object_intmax(&jw, KV(PeakPagefileUsage));
> +
> + jw_end(&jw);
> +
> + trace2_data_json("process", the_repository,
> + "windows/memory", &jw);
> + jw_release(&jw);
> + }
> + }
> +}
> +
> +void trace2_collect_process_info(enum trace2_process_info_reason reason)
> {
> if (!trace2_is_enabled())
> return;
>
> - get_is_being_debugged();
> - get_ancestry();
> + switch (reason) {
> + case TRACE2_PROCESS_INFO_STARTUP:
> + get_is_being_debugged();
> + get_ancestry();
> + return;
> +
> + case TRACE2_PROCESS_INFO_EXIT:
> + get_peak_memory_info();
> + return;
> +
> + default:
> + BUG("trace2_collect_process_info: unknown reason '%d'", reason);
> + }
> }
> diff --git a/trace2.c b/trace2.c
> index 490b3f071e..6baa65cdf9 100644
> --- a/trace2.c
> +++ b/trace2.c
> @@ -213,6 +213,8 @@ int trace2_cmd_exit_fl(const char *file, int line, int code)
> if (!trace2_enabled)
> return code;
>
> + trace2_collect_process_info(TRACE2_PROCESS_INFO_EXIT);
> +
> tr2main_exit_code = code;
>
> us_now = getnanotime() / 1000;
> diff --git a/trace2.h b/trace2.h
> index 894bfca7e0..888531eb08 100644
> --- a/trace2.h
> +++ b/trace2.h
> @@ -391,13 +391,19 @@ void trace2_printf(const char *fmt, ...);
> * Optional platform-specific code to dump information about the
> * current and any parent process(es). This is intended to allow
> * post-processors to know who spawned this git instance and anything
> - * else the platform may be able to tell us about the current process.
> + * else that the platform may be able to tell us about the current process.
> */
> +
> +enum trace2_process_info_reason {
> + TRACE2_PROCESS_INFO_STARTUP,
> + TRACE2_PROCESS_INFO_EXIT,
> +};
> +
> #if defined(GIT_WINDOWS_NATIVE)
> -void trace2_collect_process_info(void);
> +void trace2_collect_process_info(enum trace2_process_info_reason reason);
> #else
> -#define trace2_collect_process_info() \
> - do { \
> +#define trace2_collect_process_info(reason) \
> + do { \
> } while (0)
> #endif
FWIW this is the "VmPeak" line in /proc/$$/status on Linux. I don't know
if we can/should parse that in practice (I've been bitten in the past by
things in /proc having high cost on some kernel versions in the past,
notably smaps).
next prev parent reply other threads:[~2019-03-29 22:16 UTC|newest]
Thread overview: 68+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-03-28 13:30 [PATCH 0/4] trace2: load trace2 settings from system config Jeff Hostetler via GitGitGadget
2019-03-28 13:30 ` [PATCH 1/4] trace2: refactor setting process starting time Jeff Hostetler via GitGitGadget
2019-03-28 13:31 ` [PATCH 2/4] trace2: add absolute elapsed time to start event Jeff Hostetler via GitGitGadget
2019-03-28 13:31 ` [PATCH 3/4] trace2: find exec-dir before trace2 initialization Jeff Hostetler via GitGitGadget
2019-03-28 13:31 ` [PATCH 4/4] trace2: use system config for default trace2 settings Jeff Hostetler via GitGitGadget
2019-03-28 14:36 ` Ævar Arnfjörð Bjarmason
2019-03-28 18:50 ` Jeff Hostetler
2019-03-28 21:28 ` Josh Steadmon
2019-03-29 17:04 ` [PATCH v2 0/7] trace2: load trace2 settings from system config Jeff Hostetler via GitGitGadget
2019-03-29 17:04 ` [PATCH v2 1/7] trace2: refactor setting process starting time Jeff Hostetler via GitGitGadget
2019-03-29 17:04 ` [PATCH v2 2/7] trace2: add absolute elapsed time to start event Jeff Hostetler via GitGitGadget
2019-03-29 17:04 ` [PATCH v2 3/7] trace2: find exec-dir before trace2 initialization Jeff Hostetler via GitGitGadget
2019-03-29 17:04 ` [PATCH v2 4/7] trace2: use system config for default trace2 settings Jeff Hostetler via GitGitGadget
2019-04-01 21:00 ` Josh Steadmon
2019-04-01 21:06 ` Jeff Hostetler
2019-04-03 0:01 ` Jonathan Nieder
2019-04-03 0:00 ` Jonathan Nieder
2019-04-09 15:58 ` Jeff Hostetler
2019-03-29 17:04 ` [PATCH v2 5/7] trace2: report peak memory usage of the process Jeff Hostetler via GitGitGadget
2019-03-29 22:16 ` Ævar Arnfjörð Bjarmason [this message]
2019-04-01 21:05 ` Jeff Hostetler
2019-03-29 17:04 ` [PATCH v2 7/7] trace2: make SIDs more unique Jeff Hostetler via GitGitGadget
2019-03-29 22:12 ` Ævar Arnfjörð Bjarmason
2019-04-01 21:16 ` Jeff Hostetler
2019-03-29 17:04 ` [PATCH v2 6/7] trace2: clarify UTC datetime formatting Jeff Hostetler via GitGitGadget
2019-04-01 21:02 ` [PATCH v2 0/7] trace2: load trace2 settings from system config Josh Steadmon
2019-04-11 15:18 ` [PATCH v3 00/10] " Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 01/10] config: initialize opts structure in repo_read_config() Jeff Hostetler via GitGitGadget
2019-04-12 3:52 ` Jonathan Nieder
2019-04-15 14:34 ` Johannes Schindelin
2019-04-11 15:18 ` [PATCH v3 03/10] trace2: add absolute elapsed time to start event Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 02/10] trace2: refactor setting process starting time Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 04/10] trace2: find exec-dir before trace2 initialization Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 05/10] config: add read_very_early_config() Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 06/10] trace2: use system/global config for default trace2 settings Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 07/10] trace2: report peak memory usage of the process Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 08/10] trace2: clarify UTC datetime formatting Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 09/10] trace2: make SIDs more unique Jeff Hostetler via GitGitGadget
2019-04-11 15:18 ` [PATCH v3 10/10] trace2: update docs to describe system/global config settings Jeff Hostetler via GitGitGadget
2019-04-12 2:29 ` [PATCH v3 00/10] trace2: load trace2 settings from system config Junio C Hamano
2019-04-12 13:47 ` Jeff Hostetler
2019-04-15 20:39 ` [PATCH v4 " Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 01/10] config: initialize opts structure in repo_read_config() Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 02/10] trace2: refactor setting process starting time Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 03/10] trace2: add absolute elapsed time to start event Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 04/10] trace2: find exec-dir before trace2 initialization Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 05/10] config: add read_very_early_config() Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 06/10] trace2: use system/global config for default trace2 settings Jeff Hostetler via GitGitGadget
2019-04-27 13:43 ` SZEDER Gábor
2019-04-29 19:03 ` Jeff Hostetler
2019-04-15 20:39 ` [PATCH v4 07/10] trace2: report peak memory usage of the process Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 08/10] trace2: clarify UTC datetime formatting Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 09/10] trace2: make SIDs more unique Jeff Hostetler via GitGitGadget
2019-04-15 20:39 ` [PATCH v4 10/10] trace2: update docs to describe system/global config settings Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 00/11] trace2: load trace2 settings from system config Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 02/11] trace2: refactor setting process starting time Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 01/11] config: initialize opts structure in repo_read_config() Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 03/11] trace2: add absolute elapsed time to start event Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 04/11] trace2: find exec-dir before trace2 initialization Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 05/11] config: add read_very_early_config() Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 06/11] trace2: use system/global config for default trace2 settings Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 07/11] trace2: report peak memory usage of the process Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 08/11] trace2: clarify UTC datetime formatting Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 09/11] trace2: make SIDs more unique Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 10/11] trace2: update docs to describe system/global config settings Jeff Hostetler via GitGitGadget
2019-04-29 20:14 ` [PATCH v5 11/11] trace2: fixup access problem on /etc/gitconfig in read_very_early_config Jeff Hostetler via GitGitGadget
2019-04-29 20:21 ` [PATCH v5 00/11] trace2: load trace2 settings from system config Jeff Hostetler
2019-05-07 1:18 ` Junio C Hamano
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
List information: http://vger.kernel.org/majordomo-info.html
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=87k1gh9vii.fsf@evledraar.gmail.com \
--to=avarab@gmail.com \
--cc=git@vger.kernel.org \
--cc=gitgitgadget@gmail.com \
--cc=gitster@pobox.com \
--cc=jeffhost@microsoft.com \
--cc=jrnieder@gmail.com \
--cc=peff@peff.net \
--cc=steadmon@google.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.
Code repositories for project(s) associated with this public inbox
https://80x24.org/mirrors/git.git
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 read-only IMAP folder(s) and NNTP newsgroup(s).