git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Patrick Steinhardt <ps@pks.im>
To: Karthik Nayak <karthik.188@gmail.com>
Cc: git@vger.kernel.org, jltobler@gmail.com, phillip.wood123@gmail.com
Subject: Re: [PATCH v4 7/8] refs: support rejection in batch updates during F/D checks
Date: Tue, 25 Mar 2025 13:31:17 +0100	[thread overview]
Message-ID: <Z-KiFeN0LXTR36lM@pks.im> (raw)
In-Reply-To: <CAOLa=ZRx8gt=_J2cRFEsUEi31ia7qNL7SZ-WBWuWSyU=8SRt+g@mail.gmail.com>

On Mon, Mar 24, 2025 at 05:48:32PM +0000, Karthik Nayak wrote:
> Patrick Steinhardt <ps@pks.im> writes:
> > On Thu, Mar 20, 2025 at 12:44:02PM +0100, Karthik Nayak wrote:
> >> diff --git a/refs/files-backend.c b/refs/files-backend.c
> >> index be758ffff5..1d50d4013c 100644
> >> --- a/refs/files-backend.c
> >> +++ b/refs/files-backend.c
> >> @@ -864,7 +868,9 @@ static enum ref_transaction_error lock_raw_ref(struct files_ref_store *refs,
> >>  		 * make sure there is no existing packed ref that conflicts
> >>  		 * with refname. This check is deferred so that we can batch it.
> >>  		 */
> >> -		string_list_append(refnames_to_check, refname);
> >> +		item = string_list_append(refnames_to_check, refname);
> >> +		item->util = xmalloc(sizeof(update_idx));
> >> +		memcpy(item->util, &update_idx, sizeof(update_idx));
> >>  	}
> >>
> >>  	ret = 0;
> >
> > Hm, so we have to allocate the `util` field now to store the update
> > index, which is a bit unfortunate because all of this is part of the hot
> > loop. We cannot store a direct pointer though because the array of
> > updates may be reallocated, which would invalidate any pointers pointing
> > into the array.
> >
> 
> Yes, your inference is on point.
> 
> > I was wondering whether we could abuse an `uintptr_t` and use it to
> > store the update index as a pointer. It does feel somewhat dirty though.
> >
> 
> We can do something like this, it would be _clever_ but does feel very
> hacky.
> 
> I did so some benchmarking here
> 
> Benchmark 1: update-ref: create many refs (refformat = files, refcount
> = 100000, revision = master)
>   Time (mean ± σ):      7.396 s ±  0.175 s    [User: 0.962 s, System: 6.312 s]
>   Range (min … max):    7.145 s …  7.688 s    10 runs
> 
> Benchmark 2: update-ref: create many refs (refformat = files, refcount
> = 100000, revision = b4/245-partially-atomic-ref-updates)
>   Time (mean ± σ):      7.514 s ±  0.144 s    [User: 0.919 s, System: 6.438 s]
>   Range (min … max):    7.297 s …  7.750 s    10 runs
> 
> Summary
>   update-ref: create many refs (refformat = files, refcount = 100000,
> revision = master) ran
>     1.02 ± 0.03 times faster than update-ref: create many refs
> (refformat = files, refcount = 100000, revision =
> b4/245-partially-atomic-ref-updates)
> 
> Overall the perf degradation is very minimal. I also looked at the
> flamegraph to see if there is something. But most of the time seems to
> be spent in IO (reading refs and creating locks).
> 
> So I think we should be ok here, wdyt?

Okay. I'm not a huge fan of using the `->util` pointer like this, but I
don't have a better idea either, and the performance regression seems
acceptable. So let's roll with it until somebody has a better idea.

Thanks for doing the benchmark!

Patrick

  reply	other threads:[~2025-03-25 12:31 UTC|newest]

Thread overview: 147+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2025-02-07  7:34 [PATCH 0/6] refs: introduce support for partial reference transactions Karthik Nayak
2025-02-07  7:34 ` [PATCH 1/6] refs/files: remove duplicate check in `split_symref_update()` Karthik Nayak
2025-02-07 16:12   ` Patrick Steinhardt
2025-02-11  6:35     ` Karthik Nayak
2025-02-07  7:34 ` [PATCH 2/6] refs: move duplicate refname update check to generic layer Karthik Nayak
2025-02-07 16:12   ` Patrick Steinhardt
2025-02-11 10:33     ` Karthik Nayak
2025-02-07  7:34 ` [PATCH 3/6] refs/files: remove duplicate duplicates check Karthik Nayak
2025-02-07 16:12   ` Patrick Steinhardt
2025-02-07  7:34 ` [PATCH 4/6] refs/reftable: extract code from the transaction preparation Karthik Nayak
2025-02-07  7:34 ` [PATCH 5/6] refs: implement partial reference transaction support Karthik Nayak
2025-02-07 16:12   ` Patrick Steinhardt
2025-02-21 10:33     ` Karthik Nayak
2025-02-07  7:34 ` [PATCH 6/6] update-ref: add --allow-partial flag for stdin mode Karthik Nayak
2025-02-07 16:12   ` Patrick Steinhardt
2025-02-21 11:45     ` Karthik Nayak
2025-02-11 17:03 ` [PATCH 0/6] refs: introduce support for partial reference transactions Phillip Wood
2025-02-11 17:40   ` Phillip Wood
2025-02-12 12:36     ` Karthik Nayak
2025-02-12 12:34   ` Karthik Nayak
2025-02-19 14:34     ` Phillip Wood
2025-02-19 15:10       ` Patrick Steinhardt
2025-02-21 11:50       ` Karthik Nayak
2025-02-25  9:29 ` [PATCH v2 0/7] " Karthik Nayak
2025-02-25  9:29   ` [PATCH v2 1/7] refs/files: remove redundant check in split_symref_update() Karthik Nayak
2025-02-25  9:29   ` [PATCH v2 2/7] refs: move duplicate refname update check to generic layer Karthik Nayak
2025-02-25  9:29   ` [PATCH v2 3/7] refs/files: remove duplicate duplicates check Karthik Nayak
2025-02-25  9:29   ` [PATCH v2 4/7] refs/reftable: extract code from the transaction preparation Karthik Nayak
2025-02-25  9:29   ` [PATCH v2 5/7] refs: introduce enum-based transaction error types Karthik Nayak
2025-02-25 11:08     ` Patrick Steinhardt
2025-03-03 20:12       ` Karthik Nayak
2025-02-25  9:29   ` [PATCH v2 6/7] refs: implement partial reference transaction support Karthik Nayak
2025-02-25 11:07     ` Patrick Steinhardt
2025-03-03 20:17       ` Karthik Nayak
2025-02-25 14:57     ` Phillip Wood
2025-03-03 20:21       ` Karthik Nayak
2025-03-04 10:31         ` Phillip Wood
2025-03-05 14:20           ` Karthik Nayak
2025-02-25  9:29   ` [PATCH v2 7/7] update-ref: add --allow-partial flag for stdin mode Karthik Nayak
2025-02-25 11:08     ` Patrick Steinhardt
2025-03-03 20:22       ` Karthik Nayak
2025-02-25 14:59     ` Phillip Wood
2025-03-03 20:34       ` Karthik Nayak
2025-03-05 17:38 ` [PATCH v3 0/8] refs: introduce support for partial reference transactions Karthik Nayak
2025-03-05 17:38   ` [PATCH v3 1/8] refs/files: remove redundant check in split_symref_update() Karthik Nayak
2025-03-05 21:20     ` Junio C Hamano
2025-03-06  9:13       ` Karthik Nayak
2025-03-05 17:38   ` [PATCH v3 2/8] refs: move duplicate refname update check to generic layer Karthik Nayak
2025-03-05 21:56     ` Junio C Hamano
2025-03-06  9:46       ` Karthik Nayak
2025-03-05 17:38   ` [PATCH v3 3/8] refs/files: remove duplicate duplicates check Karthik Nayak
2025-03-05 17:38   ` [PATCH v3 4/8] refs/reftable: extract code from the transaction preparation Karthik Nayak
2025-03-05 17:39   ` [PATCH v3 5/8] refs: introduce enum-based transaction error types Karthik Nayak
2025-03-05 17:39   ` [PATCH v3 6/8] refs: implement partial reference transaction support Karthik Nayak
2025-03-07 19:50     ` Jeff King
2025-03-07 20:46       ` Junio C Hamano
2025-03-07 20:48         ` Junio C Hamano
2025-03-07 21:05         ` Karthik Nayak
2025-03-07 22:54         ` [PATCH] config.mak.dev: enable -Wunreachable-code Jeff King
2025-03-07 23:28           ` Junio C Hamano
2025-03-08  3:23           ` Jeff King
2025-03-10 15:40             ` Junio C Hamano
2025-03-10 16:04               ` Jeff King
2025-03-10 18:50                 ` Junio C Hamano
2025-03-14 16:10                   ` Jeff King
2025-03-14 16:13                     ` Jeff King
2025-03-14 17:27                       ` Junio C Hamano
2025-03-14 17:40                         ` Junio C Hamano
2025-03-14 17:43                           ` Patrick Steinhardt
2025-03-14 18:53                           ` Jeff King
2025-03-14 19:50                             ` Junio C Hamano
2025-03-14 17:15                     ` Junio C Hamano
2025-06-03 21:29             ` Mike Hommey
2025-06-03 22:07               ` Junio C Hamano
2025-06-03 22:37                 ` Mike Hommey
2025-06-03 23:08                   ` Mike Hommey
2025-03-14 21:09           ` [PATCH v2 0/3] -Wunreachable-code Junio C Hamano
2025-03-14 21:09             ` [PATCH v2 1/3] config.mak.dev: enable -Wunreachable-code Junio C Hamano
2025-03-14 21:09             ` [PATCH v2 2/3] run-command: use errno to check for sigfillset() error Junio C Hamano
2025-03-17 21:30               ` Taylor Blau
2025-03-17 23:12                 ` Junio C Hamano
2025-03-18  0:36                   ` Junio C Hamano
2025-03-14 21:09             ` [PATCH v2 3/3] git-compat-util: add NOT_A_CONST macro and use it in atfork_prepare() Junio C Hamano
2025-03-14 22:29               ` Junio C Hamano
2025-03-17 18:00                 ` Jeff King
2025-03-17 23:53             ` [PATCH v3 0/3] -Wunreachable-code Junio C Hamano
2025-03-17 23:53               ` [PATCH v3 1/3] run-command: use errno to check for sigfillset() error Junio C Hamano
2025-03-17 23:53               ` [PATCH v3 2/3] git-compat-util: add NOT_CONSTANT macro and use it in atfork_prepare() Junio C Hamano
2025-03-18  0:20                 ` Jeff King
2025-03-18  0:28                   ` Junio C Hamano
2025-03-18 22:04                 ` Calvin Wan
2025-03-18 22:26                   ` Calvin Wan
2025-03-18 23:55                     ` Junio C Hamano
2025-03-17 23:53               ` [PATCH v3 3/3] config.mak.dev: enable -Wunreachable-code Junio C Hamano
2025-03-18  0:18               ` [PATCH v3 0/3] -Wunreachable-code Jeff King
2025-03-07 21:02       ` [PATCH v3 6/8] refs: implement partial reference transaction support Karthik Nayak
2025-03-07 19:57     ` Jeff King
2025-03-07 21:07       ` Karthik Nayak
2025-03-05 17:39   ` [PATCH v3 7/8] refs: support partial update rejections during F/D checks Karthik Nayak
2025-03-05 17:39   ` [PATCH v3 8/8] update-ref: add --allow-partial flag for stdin mode Karthik Nayak
2025-03-05 19:28   ` [PATCH v3 0/8] refs: introduce support for partial reference transactions Junio C Hamano
2025-03-06  9:06     ` Karthik Nayak
2025-03-20 11:43 ` [PATCH v4 0/8] refs: introduce support for batched reference updates Karthik Nayak
2025-03-20 11:43   ` [PATCH v4 1/8] refs/files: remove redundant check in split_symref_update() Karthik Nayak
2025-03-20 11:43   ` [PATCH v4 2/8] refs: move duplicate refname update check to generic layer Karthik Nayak
2025-03-20 11:43   ` [PATCH v4 3/8] refs/files: remove duplicate duplicates check Karthik Nayak
2025-03-20 11:43   ` [PATCH v4 4/8] refs/reftable: extract code from the transaction preparation Karthik Nayak
2025-03-20 11:44   ` [PATCH v4 5/8] refs: introduce enum-based transaction error types Karthik Nayak
2025-03-20 20:26     ` Patrick Steinhardt
2025-03-24 14:50       ` Karthik Nayak
2025-03-25 12:31         ` Patrick Steinhardt
2025-03-20 11:44   ` [PATCH v4 6/8] refs: implement batch reference update support Karthik Nayak
2025-03-20 20:26     ` Patrick Steinhardt
2025-03-24 14:54       ` Karthik Nayak
2025-03-20 11:44   ` [PATCH v4 7/8] refs: support rejection in batch updates during F/D checks Karthik Nayak
2025-03-24 13:08     ` Patrick Steinhardt
2025-03-24 17:48       ` Karthik Nayak
2025-03-25 12:31         ` Patrick Steinhardt [this message]
2025-03-20 11:44   ` [PATCH v4 8/8] update-ref: add --batch-updates flag for stdin mode Karthik Nayak
2025-03-24 13:08     ` Patrick Steinhardt
2025-03-24 17:51       ` Karthik Nayak
2025-03-27 11:13 ` [PATCH v5 0/8] refs: introduce support for batched reference updates Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 1/8] refs/files: remove redundant check in split_symref_update() Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 2/8] refs: move duplicate refname update check to generic layer Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 3/8] refs/files: remove duplicate duplicates check Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 4/8] refs/reftable: extract code from the transaction preparation Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 5/8] refs: introduce enum-based transaction error types Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 6/8] refs: implement batch reference update support Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 7/8] refs: support rejection in batch updates during F/D checks Karthik Nayak
2025-03-27 11:13   ` [PATCH v5 8/8] update-ref: add --batch-updates flag for stdin mode Karthik Nayak
2025-03-28 13:00     ` Jean-Noël AVILA
2025-03-29 16:36       ` Junio C Hamano
2025-03-29 18:18         ` Karthik Nayak
2025-03-28  9:24   ` [PATCH v5 0/8] refs: introduce support for batched reference updates Patrick Steinhardt
2025-04-08  8:51 ` [PATCH v6 " Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 1/8] refs/files: remove redundant check in split_symref_update() Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 2/8] refs: move duplicate refname update check to generic layer Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 3/8] refs/files: remove duplicate duplicates check Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 4/8] refs/reftable: extract code from the transaction preparation Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 5/8] refs: introduce enum-based transaction error types Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 6/8] refs: implement batch reference update support Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 7/8] refs: support rejection in batch updates during F/D checks Karthik Nayak
2025-04-08  8:51   ` [PATCH v6 8/8] update-ref: add --batch-updates flag for stdin mode Karthik Nayak
2025-04-08 15:02     ` Junio C Hamano
2025-04-08 15:26       ` Karthik Nayak
2025-04-08 17:37         ` Junio C Hamano
2025-04-10 11:23           ` Karthik Nayak

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=Z-KiFeN0LXTR36lM@pks.im \
    --to=ps@pks.im \
    --cc=git@vger.kernel.org \
    --cc=jltobler@gmail.com \
    --cc=karthik.188@gmail.com \
    --cc=phillip.wood123@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).