git.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Patrick Steinhardt <ps@pks.im>
To: git@vger.kernel.org
Cc: karthik nayak <karthik.188@gmail.com>,
	 Junio C Hamano <gitster@pobox.com>,
	 Kristoffer Haugsbakk <kristofferhaugsbakk@fastmail.com>,
	 Christian Couder <chriscool@tuxfamily.org>
Subject: [PATCH v3 04/10] refs/files: support symbolic and root refs in initial transaction
Date: Mon, 25 Nov 2024 07:27:09 +0100	[thread overview]
Message-ID: <20241125-pks-refs-optimize-migrations-v3-4-17bc85e33ad7@pks.im> (raw)
In-Reply-To: <20241125-pks-refs-optimize-migrations-v3-0-17bc85e33ad7@pks.im>

The "files" backend has implemented special logic when committing
the first transactions in an otherwise empty ref store: instead of
writing all refs as separate loose files, it instead knows to write them
all into a "packed-refs" file directly. This is significantly more
efficient than having to write each of the refs as separate "loose" ref.

The only user of this optimization is git-clone(1), which only uses this
mechanism to write regular refs. Consequently, the implementation does
not know how to handle both symbolic and root refs. While fine in the
context of git-clone(1), this keeps us from using the mechanism in more
cases.

Adapt the logic to also support symbolic and root refs by using a second
transaction that we use for all of the refs that need to be written as
loose refs.

Signed-off-by: Patrick Steinhardt <ps@pks.im>
---
 refs/files-backend.c | 44 ++++++++++++++++++++++++++++++++++----------
 1 file changed, 34 insertions(+), 10 deletions(-)

diff --git a/refs/files-backend.c b/refs/files-backend.c
index 3ed18475a72aa4798d15b2912c37b4caabd47aac..116d4259697b20583cb2db34ed47025e8781cd42 100644
--- a/refs/files-backend.c
+++ b/refs/files-backend.c
@@ -2995,6 +2995,7 @@ static int files_transaction_finish_initial(struct files_ref_store *refs,
 	int ret = 0;
 	struct string_list affected_refnames = STRING_LIST_INIT_NODUP;
 	struct ref_transaction *packed_transaction = NULL;
+	struct ref_transaction *loose_transaction = NULL;
 
 	assert(err);
 
@@ -3040,6 +3041,7 @@ static int files_transaction_finish_initial(struct files_ref_store *refs,
 		if ((update->flags & REF_HAVE_OLD) &&
 		    !is_null_oid(&update->old_oid))
 			BUG("initial ref transaction with old_sha1 set");
+
 		if (refs_verify_refname_available(&refs->base, update->refname,
 						  &affected_refnames, NULL,
 						  err)) {
@@ -3048,26 +3050,48 @@ static int files_transaction_finish_initial(struct files_ref_store *refs,
 		}
 
 		/*
-		 * Add a reference creation for this reference to the
-		 * packed-refs transaction:
+		 * packed-refs don't support symbolic refs and root refs, so we
+		 * have to queue these references via the loose transaction.
 		 */
-		ref_transaction_add_update(packed_transaction, update->refname,
-					   update->flags & ~REF_HAVE_OLD,
-					   &update->new_oid, &update->old_oid,
-					   NULL, NULL, NULL);
+		if (update->new_target || is_root_ref(update->refname)) {
+			if (!loose_transaction) {
+				loose_transaction = ref_store_transaction_begin(&refs->base, 0, err);
+				if (!loose_transaction) {
+					ret = TRANSACTION_GENERIC_ERROR;
+					goto cleanup;
+				}
+			}
+
+			ref_transaction_add_update(loose_transaction, update->refname,
+						   update->flags & ~REF_HAVE_OLD,
+						   update->new_target ? NULL : &update->new_oid, NULL,
+						   update->new_target, NULL, NULL);
+		} else {
+			ref_transaction_add_update(packed_transaction, update->refname,
+						   update->flags & ~REF_HAVE_OLD,
+						   &update->new_oid, &update->old_oid,
+						   NULL, NULL, NULL);
+		}
 	}
 
-	if (packed_refs_lock(refs->packed_ref_store, 0, err)) {
+	if (packed_refs_lock(refs->packed_ref_store, 0, err) ||
+	    ref_transaction_commit(packed_transaction, err)) {
 		ret = TRANSACTION_GENERIC_ERROR;
 		goto cleanup;
 	}
+	packed_refs_unlock(refs->packed_ref_store);
 
-	if (ref_transaction_commit(packed_transaction, err)) {
-		ret = TRANSACTION_GENERIC_ERROR;
+	if (loose_transaction) {
+		if (ref_transaction_prepare(loose_transaction, err) ||
+		    ref_transaction_commit(loose_transaction, err)) {
+			ret = TRANSACTION_GENERIC_ERROR;
+			goto cleanup;
+		}
 	}
 
-	packed_refs_unlock(refs->packed_ref_store);
 cleanup:
+	if (loose_transaction)
+		ref_transaction_free(loose_transaction);
 	if (packed_transaction)
 		ref_transaction_free(packed_transaction);
 	transaction->state = REF_TRANSACTION_CLOSED;

-- 
2.47.0.274.g962d0b743d.dirty


  parent reply	other threads:[~2024-11-25  6:27 UTC|newest]

Thread overview: 54+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-11-08  9:34 [PATCH 00/10] refs: optimize ref format migrations Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 01/10] refs: allow passing flags when setting up a transaction Patrick Steinhardt
2024-11-11 10:30   ` karthik nayak
2024-11-11 12:53     ` Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 02/10] refs/files: move logic to commit initial transaction Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 03/10] refs: introduce "initial" transaction flag Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 04/10] refs/files: support symbolic and root refs in initial transaction Patrick Steinhardt
2024-11-11 10:42   ` karthik nayak
2024-11-11 12:53     ` Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 05/10] refs: use "initial" transaction semantics to migrate refs Patrick Steinhardt
2024-11-11 10:43   ` karthik nayak
2024-11-08  9:34 ` [PATCH 06/10] refs: skip collision checks in initial transactions Patrick Steinhardt
2024-11-11 10:53   ` karthik nayak
2024-11-08  9:34 ` [PATCH 07/10] refs: don't normalize log messages with `REF_SKIP_CREATE_REFLOG` Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 08/10] reftable/writer: optimize allocations by using a scratch buffer Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 09/10] reftable/block: rename `block_writer::buf` variable Patrick Steinhardt
2024-11-08  9:34 ` [PATCH 10/10] reftable/block: optimize allocations by using scratch buffer Patrick Steinhardt
2024-11-11 10:57 ` [PATCH 00/10] refs: optimize ref format migrations karthik nayak
2024-11-11 12:53   ` Patrick Steinhardt
2024-11-20  7:04     ` Junio C Hamano
2024-11-20  7:50       ` Patrick Steinhardt
2024-11-20 10:25         ` Christian Couder
2024-11-25  5:52           ` Patrick Steinhardt
2024-11-20  7:51 ` [PATCH v2 " Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 01/10] refs: allow passing flags when setting up a transaction Patrick Steinhardt
2024-11-20 10:19     ` Christian Couder
2024-11-20  7:51   ` [PATCH v2 02/10] refs/files: move logic to commit initial transaction Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 03/10] refs: introduce "initial" transaction flag Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 04/10] refs/files: support symbolic and root refs in initial transaction Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 05/10] refs: use "initial" transaction semantics to migrate refs Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 06/10] refs: skip collision checks in initial transactions Patrick Steinhardt
2024-11-20 10:21     ` Christian Couder
2024-11-25  5:52       ` Patrick Steinhardt
2024-11-20 10:42     ` Kristoffer Haugsbakk
2024-11-25  5:52       ` Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 07/10] refs: don't normalize log messages with `REF_SKIP_CREATE_REFLOG` Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 08/10] reftable/writer: optimize allocations by using a scratch buffer Patrick Steinhardt
2024-11-20 10:21     ` Christian Couder
2024-11-25  5:52       ` Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 09/10] reftable/block: rename `block_writer::buf` variable Patrick Steinhardt
2024-11-20  7:51   ` [PATCH v2 10/10] reftable/block: optimize allocations by using scratch buffer Patrick Steinhardt
2024-11-20 10:22     ` Christian Couder
2024-11-25  6:27 ` [PATCH v3 00/10] refs: optimize ref format migrations Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 01/10] refs: allow passing flags when setting up a transaction Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 02/10] refs/files: move logic to commit initial transaction Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 03/10] refs: introduce "initial" transaction flag Patrick Steinhardt
2024-11-25  6:27   ` Patrick Steinhardt [this message]
2024-11-25  6:27   ` [PATCH v3 05/10] refs: use "initial" transaction semantics to migrate refs Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 06/10] refs: skip collision checks in initial transactions Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 07/10] refs: don't normalize log messages with `REF_SKIP_CREATE_REFLOG` Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 08/10] reftable/writer: optimize allocations by using a scratch buffer Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 09/10] reftable/block: rename `block_writer::buf` variable Patrick Steinhardt
2024-11-25  6:27   ` [PATCH v3 10/10] reftable/block: optimize allocations by using scratch buffer Patrick Steinhardt
2024-11-25  6:29   ` [PATCH v3 00/10] refs: optimize ref format migrations Patrick Steinhardt

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=20241125-pks-refs-optimize-migrations-v3-4-17bc85e33ad7@pks.im \
    --to=ps@pks.im \
    --cc=chriscool@tuxfamily.org \
    --cc=git@vger.kernel.org \
    --cc=gitster@pobox.com \
    --cc=karthik.188@gmail.com \
    --cc=kristofferhaugsbakk@fastmail.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).