From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from smtp.kernel.org (aws-us-west-2-korg-mail-1.web.codeaurora.org [10.30.226.201]) (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 119D459B62; Tue, 16 Jan 2024 19:46:10 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.client-ip=10.30.226.201 ARC-Seal:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1705434370; cv=none; b=VL9INU5D18zDLF84mgUFy6/FHrjso4cImveZSFMNCcEammNPvYdHLV//gcI9QiGBMLG+DF5dXM82k11FsfgtIApS2ZBSz7IPkoTDN3jVWk0mSsfnsX20M1Y1B5BjHNKIReMoXSHY+pgG2ZUGacQg0csMa+7dXrZ1QaO77Ag7pfY= ARC-Message-Signature:i=1; a=rsa-sha256; d=subspace.kernel.org; s=arc-20240116; t=1705434370; c=relaxed/simple; bh=v1W0l4Sblp5LD2bIM4/cINPUeiALjJSbooXMHtzDdgY=; h=Received:DKIM-Signature:From:To:Cc:Subject:Date:Message-ID: X-Mailer:In-Reply-To:References:MIME-Version:X-stable: X-Patchwork-Hint:X-stable-base:Content-Transfer-Encoding; b=VhZxt7CoNHeSS/b6zDm0rlT8GY2ioYolbyEvw8ckav6Xg7iO2aWeR45Y3oJTefwDAOqqPvhyXTh1QImsxTyWrF/IsW1TAFgQdUuDZcosToSthBsDnQodWIDrBn3d+30RC5mG955xlexrzKDJWG8TL3jCvM8waKvVvLOnLpDjlYQ= ARC-Authentication-Results:i=1; smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b=JyL+bfyT; arc=none smtp.client-ip=10.30.226.201 Authentication-Results: smtp.subspace.kernel.org; dkim=pass (2048-bit key) header.d=kernel.org header.i=@kernel.org header.b="JyL+bfyT" Received: by smtp.kernel.org (Postfix) with ESMTPSA id 123B4C43394; Tue, 16 Jan 2024 19:46:08 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=k20201202; t=1705434369; bh=v1W0l4Sblp5LD2bIM4/cINPUeiALjJSbooXMHtzDdgY=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=JyL+bfyTGCVkFtZhr7MCc7qB+D4adADe2XCOa9/kXhMs45hlRevl7dC5PpXXS6gpV PgrbUzarWJ76yUaGDQyZA1gMq5W9DX/8pBWWnVBAVtAmhiK94eiUskd1B53aL5gPjZ KBRAqdhRynMAxqN0LsHVBO18iMovK7AEwZncnKph+pm9TI9ljQ0lJTrhc3MrZJEvsh f2cEfJFTg4UgYN9Hi+CskRuMXJ6vM38Yj8PpPwWr7ZilmLTFB6DHLm/QZWIxSWZYL/ 1xcxZJrhagGjocgtnuhzTz4b8WGCZIb6WdHyY9tcL6+voTQb+Kjq/yjBmMiCaDbT/Q P1KJMqMhPPQmg== From: Sasha Levin To: linux-kernel@vger.kernel.org, stable@vger.kernel.org Cc: Alex Lyakas , Song Liu , Sasha Levin , linux-raid@vger.kernel.org Subject: [PATCH AUTOSEL 6.7 081/108] md: Whenassemble the array, consult the superblock of the freshest device Date: Tue, 16 Jan 2024 14:39:47 -0500 Message-ID: <20240116194225.250921-81-sashal@kernel.org> X-Mailer: git-send-email 2.43.0 In-Reply-To: <20240116194225.250921-1-sashal@kernel.org> References: <20240116194225.250921-1-sashal@kernel.org> Precedence: bulk X-Mailing-List: linux-kernel@vger.kernel.org List-Id: List-Subscribe: List-Unsubscribe: MIME-Version: 1.0 X-stable: review X-Patchwork-Hint: Ignore X-stable-base: Linux 6.7 Content-Transfer-Encoding: 8bit From: Alex Lyakas [ Upstream commit dc1cc22ed58f11d58d8553c5ec5f11cbfc3e3039 ] Upon assembling the array, both kernel and mdadm allow the devices to have event counter difference of 1, and still consider them as up-to-date. However, a device whose event count is behind by 1, may in fact not be up-to-date, and array resync with such a device may cause data corruption. To avoid this, consult the superblock of the freshest device about the status of a device, whose event counter is behind by 1. Signed-off-by: Alex Lyakas Signed-off-by: Song Liu Link: https://lore.kernel.org/r/1702470271-16073-1-git-send-email-alex.lyakas@zadara.com Signed-off-by: Sasha Levin --- drivers/md/md.c | 54 ++++++++++++++++++++++++++++++++++++++++--------- 1 file changed, 44 insertions(+), 10 deletions(-) diff --git a/drivers/md/md.c b/drivers/md/md.c index 9bdd57324c37..da33270971ab 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -1209,6 +1209,7 @@ struct super_type { struct md_rdev *refdev, int minor_version); int (*validate_super)(struct mddev *mddev, + struct md_rdev *freshest, struct md_rdev *rdev); void (*sync_super)(struct mddev *mddev, struct md_rdev *rdev); @@ -1346,8 +1347,9 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor /* * validate_super for 0.90.0 + * note: we are not using "freshest" for 0.9 superblock */ -static int super_90_validate(struct mddev *mddev, struct md_rdev *rdev) +static int super_90_validate(struct mddev *mddev, struct md_rdev *freshest, struct md_rdev *rdev) { mdp_disk_t *desc; mdp_super_t *sb = page_address(rdev->sb_page); @@ -1859,7 +1861,7 @@ static int super_1_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor_ return ret; } -static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) +static int super_1_validate(struct mddev *mddev, struct md_rdev *freshest, struct md_rdev *rdev) { struct mdp_superblock_1 *sb = page_address(rdev->sb_page); __u64 ev1 = le64_to_cpu(sb->events); @@ -1955,13 +1957,15 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) } } else if (mddev->pers == NULL) { /* Insist of good event counter while assembling, except for - * spares (which don't need an event count) */ - ++ev1; + * spares (which don't need an event count). + * Similar to mdadm, we allow event counter difference of 1 + * from the freshest device. + */ if (rdev->desc_nr >= 0 && rdev->desc_nr < le32_to_cpu(sb->max_dev) && (le16_to_cpu(sb->dev_roles[rdev->desc_nr]) < MD_DISK_ROLE_MAX || le16_to_cpu(sb->dev_roles[rdev->desc_nr]) == MD_DISK_ROLE_JOURNAL)) - if (ev1 < mddev->events) + if (ev1 + 1 < mddev->events) return -EINVAL; } else if (mddev->bitmap) { /* If adding to array with a bitmap, then we can accept an @@ -1982,8 +1986,38 @@ static int super_1_validate(struct mddev *mddev, struct md_rdev *rdev) rdev->desc_nr >= le32_to_cpu(sb->max_dev)) { role = MD_DISK_ROLE_SPARE; rdev->desc_nr = -1; - } else + } else if (mddev->pers == NULL && freshest && ev1 < mddev->events) { + /* + * If we are assembling, and our event counter is smaller than the + * highest event counter, we cannot trust our superblock about the role. + * It could happen that our rdev was marked as Faulty, and all other + * superblocks were updated with +1 event counter. + * Then, before the next superblock update, which typically happens when + * remove_and_add_spares() removes the device from the array, there was + * a crash or reboot. + * If we allow current rdev without consulting the freshest superblock, + * we could cause data corruption. + * Note that in this case our event counter is smaller by 1 than the + * highest, otherwise, this rdev would not be allowed into array; + * both kernel and mdadm allow event counter difference of 1. + */ + struct mdp_superblock_1 *freshest_sb = page_address(freshest->sb_page); + u32 freshest_max_dev = le32_to_cpu(freshest_sb->max_dev); + + if (rdev->desc_nr >= freshest_max_dev) { + /* this is unexpected, better not proceed */ + pr_warn("md: %s: rdev[%pg]: desc_nr(%d) >= freshest(%pg)->sb->max_dev(%u)\n", + mdname(mddev), rdev->bdev, rdev->desc_nr, + freshest->bdev, freshest_max_dev); + return -EUCLEAN; + } + + role = le16_to_cpu(freshest_sb->dev_roles[rdev->desc_nr]); + pr_debug("md: %s: rdev[%pg]: role=%d(0x%x) according to freshest %pg\n", + mdname(mddev), rdev->bdev, role, role, freshest->bdev); + } else { role = le16_to_cpu(sb->dev_roles[rdev->desc_nr]); + } switch(role) { case MD_DISK_ROLE_SPARE: /* spare */ break; @@ -2890,7 +2924,7 @@ static int add_bound_rdev(struct md_rdev *rdev) * and should be added immediately. */ super_types[mddev->major_version]. - validate_super(mddev, rdev); + validate_super(mddev, NULL/*freshest*/, rdev); err = mddev->pers->hot_add_disk(mddev, rdev); if (err) { md_kick_rdev_from_array(rdev); @@ -3827,7 +3861,7 @@ static int analyze_sbs(struct mddev *mddev) } super_types[mddev->major_version]. - validate_super(mddev, freshest); + validate_super(mddev, NULL/*freshest*/, freshest); i = 0; rdev_for_each_safe(rdev, tmp, mddev) { @@ -3842,7 +3876,7 @@ static int analyze_sbs(struct mddev *mddev) } if (rdev != freshest) { if (super_types[mddev->major_version]. - validate_super(mddev, rdev)) { + validate_super(mddev, freshest, rdev)) { pr_warn("md: kicking non-fresh %pg from array!\n", rdev->bdev); md_kick_rdev_from_array(rdev); @@ -6833,7 +6867,7 @@ int md_add_new_disk(struct mddev *mddev, struct mdu_disk_info_s *info) rdev->saved_raid_disk = rdev->raid_disk; } else super_types[mddev->major_version]. - validate_super(mddev, rdev); + validate_super(mddev, NULL/*freshest*/, rdev); if ((info->state & (1<raid_disk != info->raid_disk) { /* This was a hot-add request, but events doesn't -- 2.43.0