From nobody Tue Feb 10 03:36:48 2026 Received: from dggsgout11.his.huawei.com (dggsgout11.his.huawei.com [45.249.212.51]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 50AFD3382E6; Wed, 28 Jan 2026 08:06:22 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=45.249.212.51 ARC-Seal: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1769587585; cv=none; b=TyLVsWDu3KZi2GMws5sTwYJ0epb6q3wDWvdvCSUxRLAjr2RUtxu+UP3vEGiCJVyKiG3ZxAhRzkBwSYcyEekANtoJUEcKZxj1oPLFGIN+dV4WoPJlLtR9E00hBMO+cQAa+YNvBUy/xoJ80K8gFfGDBFDaufKWxpJk2GoGzrqlXxc= ARC-Message-Signature: i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1769587585; c=relaxed/simple; bh=cMAZKlzIKPR3uGJ6ezn7Dt5blppwOj18zFERVH8Yrl8=; h=From:To:Cc:Subject:Date:Message-Id:In-Reply-To:References: MIME-Version; b=Ufu6BvFOowYftn2h3yjuXMq9LwQW3I2he8329vERbXpWvCAfzMyDDbz1qRRLuRe4MtbU/kmYbcZY5xCpzyDFwDxtsKehmSDowthg0/0Lcf77dnJ5GiGNTba4YhXuMrp3JQAj/ChtePCnou3ekDMZpQZp2sBU+FonQk2P+tea6mI= ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=huaweicloud.com; spf=pass smtp.mailfrom=huaweicloud.com; arc=none smtp.client-ip=45.249.212.51 Authentication-Results: smtp.subspace.kernel.org; dmarc=none (p=none dis=none) header.from=huaweicloud.com Authentication-Results: smtp.subspace.kernel.org; spf=pass smtp.mailfrom=huaweicloud.com Received: from mail.maildlp.com (unknown [172.19.163.198]) by dggsgout11.his.huawei.com (SkyGuard) with ESMTPS id 4f1FFZ0v2ZzYQv2R; Wed, 28 Jan 2026 16:05:38 +0800 (CST) Received: from mail02.huawei.com (unknown [10.116.40.128]) by mail.maildlp.com (Postfix) with ESMTP id 1338640573; Wed, 28 Jan 2026 16:06:15 +0800 (CST) Received: from huaweicloud.com (unknown [10.50.87.129]) by APP4 (Coremail) with SMTP id gCh0CgBnFvd1w3lpPgyWFQ--.43207S18; Wed, 28 Jan 2026 16:06:14 +0800 (CST) From: linan666@huaweicloud.com To: song@kernel.org, yukuai@fnnas.com Cc: xni@redhat.com, linux-raid@vger.kernel.org, linux-kernel@vger.kernel.org, linan666@huaweicloud.com, yangerkun@huawei.com, yi.zhang@huawei.com Subject: [PATCH v2 14/14] md/raid1,raid10: fall back to smaller order if sync folio alloc fails Date: Wed, 28 Jan 2026 15:57:08 +0800 Message-Id: <20260128075708.2259525-15-linan666@huaweicloud.com> X-Mailer: git-send-email 2.39.2 In-Reply-To: <20260128075708.2259525-1-linan666@huaweicloud.com> References: <20260128075708.2259525-1-linan666@huaweicloud.com> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 Content-Transfer-Encoding: quoted-printable X-CM-TRANSID: gCh0CgBnFvd1w3lpPgyWFQ--.43207S18 X-Coremail-Antispam: 1UD129KBjvJXoW3XFykWr4xGFyxtw45ur1xXwb_yoW7CrWUpa 1UGrySv34rtFW3X3yfJr1DuF1Fk34fWayUAFnrWwn7u3WfWryDuF4UXay5WF1DZFn8AFy2 q3WDAr45uFs5JaUanT9S1TB71UUUUU7qnTZGkaVYY2UrUUUUjbIjqfuFe4nvWSU5nxnvy2 9KBjDU0xBIdaVrnRJUUUQF14x267AKxVWrJVCq3wAFc2x0x2IEx4CE42xK8VAvwI8IcIk0 rVWrJVCq3wAFIxvE14AKwVWUJVWUGwA2048vs2IY020E87I2jVAFwI0_JF0E3s1l82xGYI kIc2x26xkF7I0E14v26ryj6s0DM28lY4IEw2IIxxk0rwA2F7IY1VAKz4vEj48ve4kI8wA2 z4x0Y4vE2Ix0cI8IcVAFwI0_tr0E3s1l84ACjcxK6xIIjxv20xvEc7CjxVAFwI0_Gr1j6F 4UJwA2z4x0Y4vEx4A2jsIE14v26rxl6s0DM28EF7xvwVC2z280aVCY1x0267AKxVW0oVCq 3wAac4AC62xK8xCEY4vEwIxC4wAS0I0E0xvYzxvE52x082IY62kv0487Mc02F40EFcxC0V AKzVAqx4xG6I80ewAv7VC0I7IYx2IY67AKxVWUXVWUAwAv7VC2z280aVAFwI0_Gr1j6F4U JwAm72CE4IkC6x0Yz7v_Jr0_Gr1lF7xvr2IYc2Ij64vIr41lF7I21c0EjII2zVCS5cI20V AGYxC7M4kE6xkIj40Ew7xC0wCY1x0262kKe7AKxVWUAVWUtwCF04k20xvY0x0EwIxGrwCF x2IqxVCFs4IE7xkEbVWUJVW8JwC20s026c02F40E14v26r1j6r18MI8I3I0E7480Y4vE14 v26r106r1rMI8E67AF67kF1VAFwI0_Jw0_GFylIxkGc2Ij64vIr41lIxAIcVC0I7IYx2IY 67AKxVW8JVW5JwCI42IY6xIIjxv20xvEc7CjxVAFwI0_Gr1j6F4UJwCI42IY6xAIw20EY4 v20xvaj40_Jr0_JF4lIxAIcVC2z280aVAFwI0_Gr0_Cr1lIxAIcVC2z280aVCY1x0267AK xVW8Jr0_Cr1UYxBIdaVFxhVjvjDU0xZFpf9x0JUvhFsUUUUU= X-CM-SenderInfo: polqt0awwwqx5xdzvxpfor3voofrz/ Content-Type: text/plain; charset="utf-8" From: Li Nan RESYNC_BLOCK_SIZE (64K) has higher allocation failure chance than 4k, so retry with lower orders to improve allocation reliability. A r1/10_bio may have different rf->folio orders. Use minimum order as r1/10_bio sectors to prevent exceeding size when adding folio to IO later. Signed-off-by: Li Nan --- drivers/md/raid1-10.c | 14 +++++++++++--- drivers/md/raid1.c | 13 +++++++++---- drivers/md/raid10.c | 28 ++++++++++++++++++++++++++-- 3 files changed, 46 insertions(+), 9 deletions(-) diff --git a/drivers/md/raid1-10.c b/drivers/md/raid1-10.c index ffbd7bd0f6e8..e966d11a81e7 100644 --- a/drivers/md/raid1-10.c +++ b/drivers/md/raid1-10.c @@ -41,12 +41,20 @@ static void rbio_pool_free(void *rbio, void *data) } =20 static inline int resync_alloc_folio(struct resync_folio *rf, - gfp_t gfp_flags) + gfp_t gfp_flags, int *order) { - rf->folio =3D folio_alloc(gfp_flags, get_order(RESYNC_BLOCK_SIZE)); - if (!rf->folio) + struct folio *folio; + + do { + folio =3D folio_alloc(gfp_flags, *order); + if (folio) + break; + } while (--(*order) > 0); + + if (!folio) return -ENOMEM; =20 + rf->folio =3D folio; return 0; } =20 diff --git a/drivers/md/raid1.c b/drivers/md/raid1.c index 2253e65c5f03..5bee846f1534 100644 --- a/drivers/md/raid1.c +++ b/drivers/md/raid1.c @@ -149,6 +149,7 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *d= ata) int need_folio; int j; struct resync_folio *rfs; + int order =3D get_order(RESYNC_BLOCK_SIZE); =20 r1_bio =3D r1bio_pool_alloc(gfp_flags, conf); if (!r1_bio) @@ -182,7 +183,7 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *d= ata) struct resync_folio *rf =3D &rfs[j]; =20 if (j < need_folio) { - if (resync_alloc_folio(rf, gfp_flags)) + if (resync_alloc_folio(rf, gfp_flags, &order)) goto out_free_folio; } else { memcpy(rf, &rfs[0], sizeof(*rf)); @@ -193,6 +194,7 @@ static void * r1buf_pool_alloc(gfp_t gfp_flags, void *d= ata) r1_bio->bios[j]->bi_private =3D rf; } =20 + r1_bio->sectors =3D 1 << (order + PAGE_SECTORS_SHIFT); r1_bio->master_bio =3D NULL; =20 return r1_bio; @@ -2776,7 +2778,7 @@ static sector_t raid1_sync_request(struct mddev *mdde= v, sector_t sector_nr, int write_targets =3D 0, read_targets =3D 0; sector_t sync_blocks; bool still_degraded =3D false; - int good_sectors =3D RESYNC_SECTORS; + int good_sectors; int min_bad =3D 0; /* number of sectors that are bad in all devices */ int idx =3D sector_to_idx(sector_nr); =20 @@ -2858,8 +2860,11 @@ static sector_t raid1_sync_request(struct mddev *mdd= ev, sector_t sector_nr, r1_bio->sector =3D sector_nr; r1_bio->state =3D 0; set_bit(R1BIO_IsSync, &r1_bio->state); - /* make sure good_sectors won't go across barrier unit boundary */ - good_sectors =3D align_to_barrier_unit_end(sector_nr, good_sectors); + /* + * make sure good_sectors won't go across barrier unit boundary. + * r1_bio->sectors <=3D RESYNC_SECTORS. + */ + good_sectors =3D align_to_barrier_unit_end(sector_nr, r1_bio->sectors); =20 for (i =3D 0; i < conf->raid_disks * 2; i++) { struct md_rdev *rdev; diff --git a/drivers/md/raid10.c b/drivers/md/raid10.c index 030812f908ac..72c77db9957c 100644 --- a/drivers/md/raid10.c +++ b/drivers/md/raid10.c @@ -135,6 +135,7 @@ static void * r10buf_pool_alloc(gfp_t gfp_flags, void *= data) int j; int nalloc, nalloc_rf; struct resync_folio *rfs; + int order =3D get_order(RESYNC_BLOCK_SIZE); =20 r10_bio =3D r10bio_pool_alloc(gfp_flags, conf); if (!r10_bio) @@ -185,7 +186,7 @@ static void * r10buf_pool_alloc(gfp_t gfp_flags, void *= data) =20 if (!j || test_bit(MD_RECOVERY_SYNC, &conf->mddev->recovery)) { - if (resync_alloc_folio(rf, gfp_flags)) + if (resync_alloc_folio(rf, gfp_flags, &order)) goto out_free_folio; } else { memcpy(rf, &rfs[0], sizeof(*rf)); @@ -200,6 +201,7 @@ static void * r10buf_pool_alloc(gfp_t gfp_flags, void *= data) } } =20 + r10_bio->sectors =3D 1 << (order + PAGE_SECTORS_SHIFT); return r10_bio; =20 out_free_folio: @@ -3374,6 +3376,15 @@ static sector_t raid10_sync_request(struct mddev *md= dev, sector_t sector_nr, continue; } } + + /* + * RESYNC_BLOCK_SIZE folio might alloc failed in + * resync_alloc_folio(). Fall back to smaller sync + * size if needed. + */ + if (max_sync > r10_bio->sectors) + max_sync =3D r10_bio->sectors; + any_working =3D 1; bio =3D r10_bio->devs[0].bio; bio->bi_next =3D biolist; @@ -3525,7 +3536,15 @@ static sector_t raid10_sync_request(struct mddev *md= dev, sector_t sector_nr, } if (sync_blocks < max_sync) max_sync =3D sync_blocks; + r10_bio =3D raid10_alloc_init_r10buf(conf); + /* + * RESYNC_BLOCK_SIZE folio might alloc failed in resync_alloc_folio(). + * Fall back to smaller sync size if needed. + */ + if (max_sync > r10_bio->sectors) + max_sync =3D r10_bio->sectors; + r10_bio->state =3D 0; =20 r10_bio->mddev =3D mddev; @@ -4702,7 +4721,12 @@ static sector_t reshape_request(struct mddev *mddev,= sector_t sector_nr, r10_bio->mddev =3D mddev; r10_bio->sector =3D sector_nr; set_bit(R10BIO_IsReshape, &r10_bio->state); - r10_bio->sectors =3D last - sector_nr + 1; + /* + * RESYNC_BLOCK_SIZE folio might alloc failed in + * resync_alloc_folio(). Fall back to smaller sync + * size if needed. + */ + r10_bio->sectors =3D min_t(int, r10_bio->sectors, last - sector_nr + 1); rdev =3D read_balance(conf, r10_bio, &max_sectors); BUG_ON(!test_bit(R10BIO_Previous, &r10_bio->state)); =20 --=20 2.39.2