From: Junio C Hamano <gitster@pobox.com>
To: "Utsav Shah via GitGitGadget" <gitgitgadget@gmail.com>
Cc: git@vger.kernel.org, Utsav Shah <ukshah2@illinois.edu>,
Utsav Shah <utsav@dropbox.com>
Subject: Re: [PATCH v4 1/1] unpack-trees: skip stat on fsmonitor-valid files
Date: Thu, 21 Nov 2019 13:15:49 +0900 [thread overview]
Message-ID: <xmqqd0dlamq2.fsf@gitster-ct.c.googlers.com> (raw)
In-Reply-To: <ea7880f2d07ef65f8b41f6786ff789f7a0ff8fd7.1574238737.git.gitgitgadget@gmail.com> (Utsav Shah via GitGitGadget's message of "Wed, 20 Nov 2019 08:32:17 +0000")
"Utsav Shah via GitGitGadget" <gitgitgadget@gmail.com> writes:
> From: Utsav Shah <utsav@dropbox.com>
>
> The index might be aware that a file hasn't modified via fsmonitor, but
> unpack-trees did not pay attention to it and checked via ie_match_stat
> which can be inefficient on certain filesystems. This significantly slows
> down commands that run oneway_merge, like checkout and reset --hard.
>
> This patch makes oneway_merge check whether a file is considered
> unchanged through fsmonitor and skips ie_match_stat on it. unpack-trees
> also now correctly copies over fsmonitor validity state from the source
> index. Finally, for correctness, we force a refresh of fsmonitor state in
> tweak_fsmonitor.
Instead of saying "also now correctly copies..." as if it started
working correctly by accident, be more assertive and actively make
it so ;-)
Check if a file is unchanged by fsmonitor in oneway_merge(), and
avoid unnecessary calls to ie_match_stat(). Copy the fsmonitor
validity state from the source index to the destination index in
unpack_trees(). Force a refresh of the fsmonitor state in
tweak_fsmonitor(), which is called after the index file is read
from the disk, for correctness.
perhaps.
> After this change, commands like stash (that use reset --hard
> internally) go from 8s or more to ~2s on a 250k file repository on a
> mac.
Good.
>
> Changes since the last version are:
> * The sanity checks around accessing the fsmonitor_dirty bitmap have
> been moved to another patch, which is in message id [1]
> * Unintended indentation changes in fsmonitor have been removed
> * A comment explaining what untracked->use_fsmonitor means has been
> re-added (it was dropped in the previous version)
> * A few "helped-by" entries have been added to the patch
>
> [1]: (xmqqzhh0d0ma.fsf@gitster-ct.c.googlers.com)
The above is for the cover letter or after the three-dash lines, and
not for the log message.
> Helped-by: Junio C Hamano <gitster@pobox.com>
> Helped-by: Kevin Willford <Kevin.Willford@microsoft.com>
> Signed-off-by: Utsav Shah <utsav@dropbox.com>
> ---
> fsmonitor.c | 23 +++++++++++++++++------
> t/t7519-status-fsmonitor.sh | 9 +++++++--
> unpack-trees.c | 6 +++++-
> 3 files changed, 29 insertions(+), 9 deletions(-)
>
> diff --git a/fsmonitor.c b/fsmonitor.c
> index 1f4aa1b150..0d270da80f 100644
> --- a/fsmonitor.c
> +++ b/fsmonitor.c
> @@ -189,13 +189,26 @@ void refresh_fsmonitor(struct index_state *istate)
> }
> if (bol < query_result.len)
> fsmonitor_refresh_callback(istate, buf + bol);
> +
> + /* Now mark the untracked cache for fsmonitor usage */
> + if (istate->untracked)
> + istate->untracked->use_fsmonitor = 1;
> } else {
> +
> + /* We only want to run the post index changed hook if we've actually changed entries, so keep track
> + * if we actually changed entries or not */
Multi-line comment style.
> + int is_cache_changed = 0;
> /* Mark all entries invalid */
> - for (i = 0; i < istate->cache_nr; i++)
> - istate->cache[i]->ce_flags &= ~CE_FSMONITOR_VALID;
> + for (i = 0; i < istate->cache_nr; i++) {
Lack of blank line between the last decl and the first stmt.
Probably the blank should go before "/* Mark all ...".
> @@ -257,9 +270,7 @@ void tweak_fsmonitor(struct index_state *istate)
> (uintmax_t)istate->fsmonitor_dirty->bit_size, istate->cache_nr);
> ewah_each_bit(istate->fsmonitor_dirty, fsmonitor_ewah_callback, istate);
>
> - /* Now mark the untracked cache for fsmonitor usage */
> - if (istate->untracked)
> - istate->untracked->use_fsmonitor = 1;
> + refresh_fsmonitor(istate);
> }
>
> ewah_free(istate->fsmonitor_dirty);
Looks good.
Thanks.
prev parent reply other threads:[~2019-11-21 4:15 UTC|newest]
Thread overview: 26+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-10-25 15:23 [PATCH 0/1] unpack-trees: skip lstat on files based on fsmonitor Utsav Shah via GitGitGadget
2019-10-25 15:23 ` [PATCH 1/1] unpack-trees: skip lstat " Utsav Shah via GitGitGadget
2019-10-28 3:37 ` Junio C Hamano
2019-10-28 6:39 ` Utsav Shah
2019-10-28 19:23 ` Kevin Willford
2019-10-29 19:06 ` Utsav Shah
2019-10-29 20:12 ` Kevin Willford
2019-10-29 23:50 ` Utsav Shah
2019-10-30 0:21 ` Junio C Hamano
2019-10-30 16:41 ` Utsav Shah
2019-11-04 6:02 ` Junio C Hamano
2019-11-05 15:27 ` [PATCH v2 0/1] unpack-trees: skip stat on fsmonitor-valid files Utsav Shah via GitGitGadget
2019-11-05 15:27 ` [PATCH v2 1/1] " Utsav Shah via GitGitGadget
2019-11-05 21:40 ` Kevin Willford
2019-11-06 4:36 ` Utsav Shah
2019-11-06 17:24 ` Kevin Willford
2019-11-06 4:54 ` [PATCH v3 0/1] " Utsav Shah via GitGitGadget
2019-11-06 4:54 ` [PATCH v3 1/1] " Utsav Shah via GitGitGadget
2019-11-06 10:46 ` Junio C Hamano
2019-11-06 22:33 ` Utsav Shah
2019-11-08 3:51 ` Utsav Shah
2019-11-08 4:11 ` Junio C Hamano
2019-11-06 10:16 ` [PATCH v3 0/1] " Junio C Hamano
2019-11-20 8:32 ` [PATCH v4 " Utsav Shah via GitGitGadget
2019-11-20 8:32 ` [PATCH v4 1/1] " Utsav Shah via GitGitGadget
2019-11-21 4:15 ` Junio C Hamano [this message]
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=xmqqd0dlamq2.fsf@gitster-ct.c.googlers.com \
--to=gitster@pobox.com \
--cc=git@vger.kernel.org \
--cc=gitgitgadget@gmail.com \
--cc=ukshah2@illinois.edu \
--cc=utsav@dropbox.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).